action research

PROCEEDINGS OF THE 16TH INTERNATIONAL FARM MANAGEMENT ASSOCIATION CONGRESS

A VIBRANT RURAL ECONOMY – THE CHALLENGE FOR BALANCE

UNIVERSITY COLLEGE CORK CORK, IRELAND 15 – 20 JULY 2007

PEER REVIEWED PAPERS

VOLUME I OF II

EDITORS: SEAMUS O’REILLY MICHAEL KEANE

PAT ENRIGHT

© 2007 International Farm Management Association ISBN: 978-92-990038-3-1 Volume I of II Printed by: Snap Printing Cork Ireland

IFMA16 SCIENTIFIC COMMITTEE Dr Michael Keane, Department of Food Business & Development, University College Cork. (Chairman) Dr Seamus O’Reilly, Department of Food Business & Development, University College Cork.(Secretary) Dr Duncan Anderson, Department of Agricultural and Food Economics, Queen’s University Belfast. Mr Tom Arnold, CEO, CONCERN. Mr Jim Beecher, Department of Agriculture & Food, Dublin. Professor Gerry Boyle, Head, Department of Economics, National University of Ireland Maynooth. Dr Diane Burgess, Department of Agriculture and Rural Development, Northern Ireland. Dr Anne-Marie Butler, Agriculture and Food Science Centre, University College Dublin. Dr Liam Connelly, Rural Economy Research Centre, Teagasc, Athenry. Professor Michael Cuddy, Department of Economics, National University of Ireland Galway. Dr John Davis, Head, Queen’s University Belfast. Mr Matt Dempsey, Editor, Irish Farmers Journal. Mr Trevor Donnellan, Rural Economy Research Centre, Teagasc, Athenry. Dr Pat Enright, Department of Food Business & Development, University College Cork. Dr Thia Hennessy, Rural Economy Research Centre, Teagasc, Athenry. Dr Mary Keeney, Research Division, Central Bank. Professor Alan Matthews, Head, Department of Economics, Trinity College. Dr. Mary McCarthy, Department of Food Business & Development, University College Cork. Mr Ian McCluggage, Head of Dairy and Pigs College of Agriculture, Food & Rural Enterprise, Greenmount, Co. Antrim. Mr James O’Boyle, College of Agriculture, Food & Rural Enterprise, Greenmount, Co. Antrim. Dr Declan O'Connor, Department of Mathematics, Cork Institute of Technology. Dr Cathal O’Donoghue, Head of Rural Economy Research Centre, Teagasc, Athenry. Professor Jim Phelan, Department of Agribusiness, Extension and Rural Development, University College Dublin. Dr Dermot Ruane, Department of Agribusiness, Extension and Rural Development, University College Dublin. Dr Laurance Shalloo, Moorepark Food Research Centre, Teagasc, Moorepark. Dr David Stead, Agriculture and Food Science Centre, University College Dublin. Dr Michael Wallace, Department of Agribusiness, Extension and Rural Development, University College Dublin.

IFMA 16 ORGANISING COMMITTEE JJ Harty, Clonakilty Agricultural College, Teagasc, Cork. (Chairman) Michael Keane, Department of Food Business & Development, University College Cork. (Vice Chairman) Joe Walsh, Congress Patron Malcom Stansfield, President, International Farm Management Association (IFMA). Tony King, Secretary/Treasurer, International Farm Management Association (IFMA). Colette Collins O’Sullivan, Clonakilty Agricultural College, Teagasc, Cork. John Donovan, Agricultural Science Association, Ireland. Pat Enright, Department of Food Business & Development, University College Cork. Ryan Howard, CEO, East Cork Area Development, Co Cork. Jerry McCarthy, CAO, Teagasc, East Cork. Mary McCarthy-Buckley, Food Industry Training Unit, University College Cork. James Moloney, Regional Education Officer, Teagasc. James O’Boyle, College of Agriculture, Food & Rural Enterprise, Greenmount, Co. Antrim. Mary O'Driscoll Murphy, Clonakilty Agricultural College, Teagasc, Cork. Seamus O’Reilly, Department of Food Business & Development, University College Cork. Aoife O'Sullivan, Department of Food Business & Development, University College Cork. Sean Shermin, Farm Managers Association, Ireland. IFMA 16 CONGRESS SECRETARIAT Aoife O'Sullivan, Department of Food Business & Development, University College Cork, Ireland.

i

TABLE OF CONTENTS

VOLUME I

Theme 1 – The Role of Agriculture in the Rural Economy

Changes In Agricultural Industries Bio-Fuel Chains – Analyzing Structure, Options And Impacts - Dautzenberg, Hanf and Jungklaus

1

Challenges Facing Agriculture In New Zealand - Gardner

9

Current Situation And Perspective Of The Horticultural Farms In Bulgaria – Case in the Plovdiv Region - Garnevska, Edwards and Vaughan

17

Food Security: When To Buy Derivative Instruments - Geyser and Cutts

26

The Determinants Of Entry And Exit Decision In Dutch Glasshouse Horticulture - Goncharova and Oskam

37

Policy Assessment And Development By Stakeholders: A Cross-Country Analysis Of National Recommendation On Organic Farming Policy In 11 European Countries - Häring, Vairo, Dabbert and Zanoli

50

The Use Of Technology Assessment (TA) In The Food-Chain From “Farm To Fork” - Larsen, Gylling and Pedersen

59

Changing Perceptions Of The Risk Environment Faced By Commercial Sugarcane Farmers In Kwazulu-Natal, South Africa - Mac Nicol, Ortmann and Ferrer

66

Improving Policy Coherence Between Agricultural And Development Policies - Matthews

74

Market Orientation, Social Embeddedness And Firm Profitability: An Empirical Exploration Of The Us Beef Industry - Micheels and Gow

84

Potential Cost Of Being Less Trade Distorting On U.S. Crop Farms - Raulston, Outlaw and Richardson

93

Farmers’ Behavioural Inclinations And Their Influence On The Anticipated Response To The Reform Of The Common Agricultural Policy - Rehman, Garforth, McKemey, Yates and Rana

100

Dairy Farm Business Analysis: Current Approaches And A Way Forward - Shadbolt, Newman and Lines

125

Europe’s Mansholt Plan Forty Years On - Stead

136

ii

Agricultural & Non-Agricultural Rural Employment In The EU: Issues And Strategies, With Special Reference To Accession & Candidate Countries - Turner and Wibberley

140

Farming In Eastern Germany: From Food To Energy Crop Production? - Zeller and Häring

157

Theme 2: Agrarian Vs Rural: Economies and Settlements

Recreating Location From Non-Spatial Data –Sample Size Requirements To Reproduce The Locations Of Farms In The European Farm Accountancy Data Network - Damgaard and Kjeldsen

164

Analysis Of Bean Marketing Channels In Kenya And Tanzania - Korir, Nyangweso, Serem, Kipsat And Maritim

177

Evaluation Of Strategies To Achieve Compliance With A Legal Risk Assesment Document By Farmers In Ireland - McNamara, Phelan, Griffin, Morahan and Laffey

186

Methodological Frameworks For Research And Development On Improving Linkages And The Competitiveness Of Supply Chains - Murray-Prior

195

Household Food Security In Vihiga District, Kenya: Key Determinants Using An Almost Ideal Demand System (AIDS) - Nyangweso, Odhiambo and Odunga

202

Drivers Of Agricultural Exports In Eastern Africa: Evidence From Kenya, Uganda, And Tanzania - Nyangweso, Odhiambo, Serem, Korir and Kipsat

211

South African Land And Market Reforms: Equity Versus Efficiency - Olubode-Awosola and Van Schalkwyk

221

Linking Rural Economies With Markets – An Institutional Approach - Van Schalkwyk, Kotze and Fourie

229

Review Of Quality Of Life Influential Factors Among Irish Farm Families Reporting Disability - Whelan, Ruane, McNamara and Kinsella

239

Theme 3: Farm Management

Case Study Analysis Of The Benefits Of Genetically Modified Cotton - Back and Beasley

247

Customized Commodity Derivatives; An Alternative To Reduce Financial Risks On Farms - Baltussen, van Asseldonk and Horsager

267

What Is The Potential For Precision Agriculture Based On Plant Sensing?

iii

- Biermacher, Epplin, Brorsen, Solie and Raun 276

A Feasibility Study Of Contract Finishing Of Hogs - Brown, Painter and Ferguson

288

Quantifying The Sources Of Dairy Farm Business Risk And Understanding The Implications for Risk Management Strategies - Chang, Boisvert and Tauer

297

Milk Components And Farm Business Characteristics: Estimation Of Production Functions Versus A Multiple Output Distance Function - Cho and Tauer

308

Productivity And Farm Size - Clark and Langemeier

321

A Model To Evaluate The Feasibility Of GM And Non-GM Co-Existence In Europe At Farm And Collection Firm Level For Maize. - Coléno

329

The Value of Pregnancy Testing Spring-Calving Beef Cows - Cook, Biermacher and Childs

338

The Development And Role Of New Farm Management Methods For Use By Small-Scale Farmers In Developing Countries - Dorward, Shepherd and Galpin

348

A Comparative Study Of Variability In Agricultural Enterprises And Fish Farming - Flaten, Lien and Tveterås

359

The Introduction Of A Supply Chain/Consumer Focus In Farmer Controlled Businesses In The UK - Gonzalez-Diaz, Newton and Alliston

370

Achieving Sustained Improvements In Profitability In Beef Enterprises And Regions In South Africa And Australia - Clark, Griffith, Madzivhandila, Nengovhela, Parnell, and Timms

379

Factors Influencing Ear Initiation And Ear Emergence Development Of Perennial Ryegrass Cultivars At Two Different Latitudes - Hurley, O’Donovan and Gilliland

393

Determining Labor Efficiency Of U.S. Row Crop Production - Ibendahl and Anderson

403

The Use Of Relevant Cost Analysis To Assess Production Viability Following The Decoupling Of Support Payments In England - Jones

412

Farm Income On Full And Part-Time Farms - 2005 - Kinsella and Moran

422

Strategies Of Polish Farmers – An Attempt Of Classification

iv

- Majewski and Sulewski 429

Farm Income Risk Assessment For Selected Farm Types In Poland - Majewski, Guba and Was

437

Estimating Input-Specific Recommendations For Technically Inefficient Crop Farmers - Matthews, Grové and Kundhlande

449

Economic Comparison Of Divergent Strains Of Holstein-Friesian Cows In Various Pasture-Based Production Systems. - McCarthy, Horan, Dillon, O’Connor, Rath, and Shalloo

459

VOLUME II

Early Spring Feeding Budget For Spring Calving Dairy Cows - McEvoy, O’Donovan, Murphy and Boland

474

An Economic Evaluation Of Four Fattening Strategies For Cull Dairy Cows - Minchin, O’Donovan, Kenny, Buckley and Shalloo

481

Factors Underpinning Improved Productivity In The WA Wheat Industry - Murray-Prior, Rola-Rubzen, Martin and Sirisena

488

Labour Productivity – Effects Of Scale, Capital Investment And Adoption Of Novel Technology - O’Brien, Shalloo, O'Donnell, Butler, Gleeson, and O’Donovan

497

Profile Of Labour Demand, Resources And Contribution On Irish Dairy Farms - O’Brien, Gleeson, Ruane, Kinsella and O’Donovan

508

Price Transmission From Market To Farm Gate: An Irish Dairy Study - O’Connor, Keane and Kenneally

518

What Are The Characteristics Of The Irish Dairy Farmers Who Use Profit Monitor? - O’Dwyer and Connolly

529

Understanding Deer Farmers’ Level Of Environmental Awareness - Payne and White

540

Dairy Farm Ownership And Management Structures: Focus Group Research - Payne, Shadbolt, Dooley, Smeaton and Gardner

547

Production Economics And Environmental Impact Of Improved Drip Irrigation And Fertilizer Management In Potatoes - Pedersen, Abrahamsen and Plauborg

556

Dairy Farm Ownership Structures And Their Management: Case Study Research - Reekers, Shadbolt, Dooley and Bewsell

565

An Initial Study Of The Use Made By Suckler Beef Farmers Of Agricultural Contractors In The Republic Of Ireland - Ruane, Fallon, Leahy, and O’Riordan

578

v

Sustainable Family Farming – A Balance Between Economy, Ecology And Commerce - Russell

588

The Historic Residual Return To Farm Land, Labour And Management, Saskatchewan, Canada: 1926-2005 - Schoney

597

Motives For Foreign Investment In Agriculture: Western European Investment In Ukraine - Stange

603

Efficiency Of Rice Farmers In Nigeria: Potentials For Food Security And Poverty Alleviation. - Umeh and Ataborh

613

Demand For Multi Peril Crop Insurance: What Role Should Public Policy Play? - Van der Meulen, Van der Meer and Van Asseldonk

626

Differences In Attitude Of Horticultural Entrepreneurs Towards The Introduction Of Reduction Techniques For Pesticide And Nutrients - Van Lierde, Taragola, Vandenberghe and Cools

633

A2 Milk, Farmer Decisions, And Risk Management - Woodford

641

Profitability Of Direct Marketing Farms In The Less Favoured Areas (LFAS): A Case From Northumberland, England - Yagi and Garrod

649

Theme 4: Environment – A Global Resource

Determining The Cost Effectiveness Of Solutions To Diffuse Pollution: The Case Of In-Field Mitigation Options For Phosphorous And Sediment Loss - Bailey, Quinton, Silgram, Steven and Jackson

657

Farm Management Implications Of Providing Wet Habitats To Improve Biodiversity - Bailey, Aquilina, Bradbury, Kirby, Lawson, Mortimer, Stoate, Szczur, Williams and Woodcock

665

Economic Effects Of On-Farm Nature Conservation For Dairy Farms - Berentsen

674

The Impact Of The Rural Environmental Protection Scheme On Irish Dairy Farm Performance And The Consequences In A Decoupled World - Breen

682

Recreational Leases As Means To Increase Landowner Income And Enhance Biodiversity: Observations From Illinois, USA - Eberle and Wallace

690

vi

Modelling Farm-Economic And Environmental Effects Of Reduced Tillage - With Focus On Pesticide Use - Jacobsen and Ørum

697

Measuring Environmental Performance And Value Added Using The Agri-Environmental Footprint Index - Mauchline, Park, Mortimer and Finn

706

Financial Returns From Organic V Conventional Cattle Rearing System - Moran and Connolly

712

An Investigation Into The Determinants Of Commitment To Organic Farming In Ireland - McCarthy, O’ Reilly, O’Sullivan and Guerin

718

Impacts Of CAP Designs On Rural Territories And Their Viability: Simulation Experiments On A Small German Region And Distributive Effects - Osuch, Damgaard, Sahrbacher and Happe

733

An Exploration Of Language For Biodiversity And Regeneration In Australian Agriculture - Scott and Watson

744

Optimal Nitrogen Fertilizer Application And Efficient Water Use - Teweldemedhin, Vijoen, Alemu and Anderson

753

Ecological Effects Of Payment Decoupling In A Case Study Region In Germany - Uthe, Sattler, Reinhardt, Piorr, Zander, Happe, Damgaard and Osuch

761

A Case Based Analysis Evaluating The Financial Contribution To Farm Income Of Entry Level Environmental Stewardship On Upland Farms In England - Wallis and Jones

771

Theme 5: Education and Training

Leadership Development In The UK Farming Industry: Outcomes Of A Leadership Programme - Alliston, Gonzalez-Diaz and Norman

783

Evaluation Of A Training Programme Designed To Improve The Entrepreneurial Competencies Of Dutch Dairy Farmers - Bergevoet, Giesen, Saatkamp, van Woerkum and Huirne

788

La Manera: Strategic Farm Management Under Uncertainty In Uruguay - Confore, Cameron and Shadbolt

797

Identifying Training Needs In New Zealand’s Sports Turf Industry - Haydu, Way, van Blokland, Hodges and Cisar

816

Behavioural Factors Affecting The Adoption Of Forward Contracts By Australian Wool Producers - Jackson, Quaddus, Islam and Stanton

826

Investigating Factors Affecting The Forward Pricing Behaviour Of Vaalharts Maize Producers: A

vii

Two-Step Econometric Approach - Jordaan and Grové

841

Dairy Farmer Clients’ Perceptions Of The Teagasc Advisory Service - O’Dwyer and Reidy

850

Strategic Management Of Farm Businesses: The Role Of Strategy Tools With Particular Reference To The Balanced Scorecard - Shadbolt

860

Defining New Clientele For University Outreach In The West - Tranel, Hewlett, Weigel, Ehmke, Rahman and Teegerstrom

871

Critical Success Factors Underpinning Industry Visits In Farm Management Education - Watson and Bone

881

viii

IFMA 16 – Theme 1 The Role of Agriculture in the Rural Economy

1

CHANGES IN AGRICULTURAL INDUSTRIES BIO FUEL CHAINS – ANALYZING STRUCTURE, OPTIONS, AND IMPACTS

Kirsti Dautzenberg, Jon H. Hanf, Sven-Oliver Jungklaus Leibniz Institute of Agricultural Development in Central and Eastern Europe, Germany

Email: [email protected]

Abstract Structural change in the agricultural sector as well as in the whole agricultural value chain is an ongoing dynamic process and affords a number of diverse phenomena. The EU Strategy for Biofuels (2006) and the Biomass action plan (2005) set a clear signal that the EU wishes to establish and to support the bio energy-industry. The perceivable aim of the Common Agricultural Policy (CAP) consists in reducing food production and in enlarging the non-food production. Another driver for the attractiveness of bio-energy and bio-fuel production is the price history of crude oil and natural gas in recent years. As a result the total production of biofuels in the EU is increasing rapidly. The EU`s production of liquid biofuels (bioethanol, biodiesel) amounted to a total of 2.4 Million tonnes in 2004 (EurObserver, 2005); an increase of more than 25 % compared with the previous year. Catalysts have been the adoption of Biofuels Directive (2003/30/EC) by the EU Commission as well as the urge that member states have to ensure that in 2005 biofuels account for at least 2 % of the total used transportation fuels. In 2010 a minimum stake of 5.75 % has to be met. In 2007 German enterprises are planning to enlarge the production capacity for bioethanol production for 330.000 t/a as well as for biodiesel 1.9 Million t/a. For example, on top of the already planned capital expenditures of about € 500 million in bioethanol production the German “Südzucker Group” plans within the next years to triple the production capacity to over one million tonnes in Germany, Austria, Belgium, France and Hungary. This development leads to structural changes in the agricultural sector as well as in the whole agricultural value chain. Beside price increase of commodities the formation of vertically organized structures along the value chain can be observed in order to guarantee on the one hand production efficiency (regional supply of raw material) and on the other hand safeguarding the high investments. The aim of the paper refers to the consequences of the rapid growth of the bio-energy sector and its diverse impacts on all stages along the whole chain and the agricultural sector. As generally pointed out the production is changing from an industry which is dominated by family-based, small-scale, relatively independent firms to one of larger firms that are more tightly aligned across the production and distribution value chain. Hence the aim is to elaborate on the impact of verticalisation as a main consequence on the management of agricultural enterprises. For example, in the value chain of bioethanol there are two types in which both farms are involved. On the one hand there are the central plants which are operated by companies and on the other hand there are the distilleries and cooperative distilleries which are ran by farmers or where they have a close regional connection with. In both types farms are providing the raw material. But the conditions differ between these types. In the first type, farmers account the advantages of companies as operators. The risk is lower and the safety of payments is more ensured than in smaller enterprises. Furthermore farmers can develop long term marketing possibilities. But for big producing sites there is the need of a high amount of raw material so the market power of the single farmer is going down and because of the heterogeneous group of farmers it will remain low. The single farm is replaceable. The smaller amount of suppliers in small producing distilleries makes a straight -and partially participatory- contractual framework necessary. Additionally in this context we will elaborate on the assessment of the implications for farm incomes and the rural economy. Interviews are conducted with the managers of enterprises which produce bioethanol and biodiesel in Germany. The focus will be laid on the vertical institutional structures between agricultural enterprises and the producers of biofuels, especially on embodiment of the formal and informal contracts between farmers and biofuel producers.


2

Introduction The EU Strategy for Biofuels (2006), the Biomass action plan (2005) and the adoption of Biofuels Directive (2003/30/EC) by the EU Commission set a clear signal that the EU wishes to establish and to support the bio energy-industry. As a result the total production of bio fuels in the EU is increasing rapidly. In addition, since 2005 bio fuels have to account for at least 2 % of the total used transportation fuels in the EU-member states. Moreover, in 2010 a minimum stake of 5.75 % has to be met. In Germany additionally the introduction of a bio fuel quota ensured that the mineral oil companies – starting at the 01 January 2007 – have to secure that 4.4 % of the sales of diesel are made of bio diesel as well as that 1.2 % (from 2008 2 %, from 2009 2.8 % and from 2010 3.6 %) of the sales of motor fuel are made of bio fuel (Bundestag resolution, 26.10.2006). Furthermore the perceivable aim of the Common Agricultural Policy (CAP) consists in reducing food production and in favour of enlarges the non-food production. Moreover, further drivers of the attractiveness of bio fuel production are the price history of crude oil and natural gas in recent years, international developments (such as climate change, pressure of environmental NGOs), technological advances and innovations, price development of commodities and substitutes, reduction of risks caused by harmful exhaust emissions and by greenhouse gas emissions, as well as free capacities (obligatory set-aside). Based on the drafted developments the aim of the paper is to clarify the structure of the biomass-based energy value chain exemplifying the production of bio ethanol and bio diesel. First, we will give an overview of the current sector developments and afterwards the questions “who is the initiator of the biomass-based energy value chain?”, “who coordinates the process of bringing biomass into final energy products?” and “how to organize it?” will be answered. Bio fuel production, potentials and future investments in the EU and Germany In 2005 the EU’s production of liquid bio fuels (bio ethanol, bio diesel) amounted to a total of 3.2 million tonnes; an increase of more than 30 % compared with the previous year (FNR, 2007). Whereas bio ethanol totalled for 0.5 million tonnes and bio diesel for 2.7 million tonnes (see figure 1). Figure 1: Bio fuel production in EU 25 and Germany in tonnes

• • • 2002 • 2003 • 2004 • 2005 • Biodiesel • EU 25 • 1,134,000 • 1,504,000 • 1,933,400 • 2,740,000

• • Germany • 450,000 • 715,000 • 1,035,000 • n.s. • Bioethanol • EU 25 • 388,200 • 424,750 • 491,040 • 500,000

• • Germany • n.s. • 0 • 20,000 • n.s. • Total • EU 25 • 1,522,200 • 1,928,750 • 2,424,440 • 3,240,000

• • Germany • 450,000 • 715,000 • 1,055,000 • n.s. Source: EurObserver 2005, European Biodiesel Board (EBB), FNR, 2007. Figure 1 shows that the German bio fuel production increased rapidly since 2003, especially for bio diesel. Also in 2007 an enlargement of the production capacity of bio diesel for additional 1.9 million tonnes and of bio ethanol for 430,000 tonnes is planned. In order to finance this expansion new resources and financing are taped e.g. the initial public offering of Crop Energies in September 2006 ,Verbio AG in October 2006 and, BKN BioKraftstoff Nord AG in February 2007. Bio Diesel Production, Capacities, and Potentials In Germany

Comparing the production and the sales of bio diesel in Germany for the last few years shows a balanced relationship. Furthermore, the bio diesel supply and the capacities of the plants are sufficient developed so that the bio fuel quota for diesel is already fulfilled in Germany within the next seven years. In order to


3

meet the compulsory blending of bio diesel of 7 % in Germany until 2015 a production of 2.4 million t/a bio diesel is demanded (FNR, 2007). Assuming that exclusively the raw material consumption for bio diesel would derive from German raw material (rapeseed) the demand will increase from 4.7 million t/a (2005) to 5.9 million t/a (2015). Therefore by 2015 the share of total agricultural area dedicated to bio diesel production will rise from 11.8 % (2005) to over 15 % (2015) (FNR, 2007). In the last years bio diesel production capacity has already increased continuously (see figure 2). Accordingly an increase in production is anticipated i.e. in 2007 an enlargement of the capacity for additional 1.9 million t. Figure 2: Bio-diesel production capacities in Germany 1998-2006 (1,000 tonnes/a)

• • 1999 • 2000 • 2001 • 2002 • 2003 • 2004 • 2005 • 2006 • Production

Capacity • 175 • 290 • 531 • 874 • 1,050 • 1,237 • 1,197 • 3,603

Source: FNR, 2007. In general for bio diesel we find a dual production structure. On the one hand there are smaller oil mills owned by single farmers or farmer bio fuel producing associations and on the other there are larger commercial mills with production capacities over 1,000 t/a. About 50 production facilities have even production capacities between 4,000 and 500,000 t /a (e.g. BKN BioKraftstoff Nord AG with a capacity up to 50,000t/a). Recently we can observe a change towards lager production facilities.

Bio Ethanol Production, Capacities, And Potentials In Germany

In Germany the bio ethanol production from cereals is with 330,000 t/a (2005) at a starting point (FNR, 2007). In 2006 there has been a production capacity of around 640,000 t. Figure 3 shows that the production capacities for bio ethanol in 2007 shall be increased for about 430,000 t. Probably, depending on political future developments (e.g. decisions about tax break) the production capacity of bio ethanol will rise up. Figure 3: Bio ethanol capacities in Germany in 2006 and 2007 in tonnes/a

• Operating Company • Location • Capacity in 2006

in t

• planned in 2007

• Capacity in 2007

in t • Crop-Energies (Südzucker

Group) • Zeitz • 260,000 • 100,000 • 360,000

• Verbio Vereinigte Bioenergie AG (NBE)

• Schwedt • 230,000 • • 230,000

• Verbio Vereinigte Bioenergie AG (MBE)

• Zörbig • 100,000 • • 100,000

• fuel 21 (Nordzucker Group) • Klein Wanzlebe

n

• • 130,000 • 130,000

• Bernhard Icking KG • Seyda • 7,500 • • 7,500 • WABIO Bioenergie • Bad

Köstritz • 8,400 • • 8,400

• NAWARO Chemie GmbH • Rostock • • 100,000 • 100,000 • PROKON Nord Energiesysteme

GmbH • Stade • • 100,000 • 100,000

• KWST • Hannover • 30,000 • • 30,000 • Total • • 635,900 • 430,000 • 1,065,900


4

Source: own source, FNR, 2007. The following example demonstrates the dimension of this development. On top of the already planned capital expenditures of about € 500 million in bio ethanol production the German “Südzucker Group” plans within the next years to triple the production capacity to over one million tonnes in Germany, Austria, Belgium, France and Hungary. Hence, the “Südzucker Group” will be the market leader in the EU with a market share of about 10 %. Taking a closer look on the example demonstrates the size of this plan. In Germany at the production location Zeitz 260,000 t/a bio ethanol are produced from 700,000 t of wheat since 2005. In future on the basis of sugar beets additionally 100,000 t/a bio ethanol will be produced. In Belgium “Südzucker AG” plans a bio ethanol production based on wheat and sugar beets with an annual capacity of 300,000 tonnes. Its Austrian affiliate AGRANA started with building a bio ethanol production plant in Pischelsdorf (Lower Austria). As from autumn 2007 it will operate with an annual capacity up to 240,000 t bio ethanol. In Hungary AGRANA already produces annually 50,000 t bio ethanol. It is planned to increase the capacity to 160,000 t/a. The example shows that due to the economies of size in the production of bio ethanol only large plants are profitable and high investments are needed. However, one factor that complicates the investments is that the tax break for bio ethanol only lasts until 2008. Between 2008 and 2012 the decision about a tax over compensation will be annually based on a new report. This aspect as well as the uncertainty about the development of the price of crude oil makes investments in large plants unsure. In addition the deferred WTO-Negotiations with MERCOSUR and the uncertainty about the import protection for bio ethanol are problems for investors. The WTO suggestions provide a reduction of the import protection for ethanol by 40 %. In this case the production would be no longer competitive in the EU. Mainly due to political pressure a significant increase of the production of bio fuels can be expected. Thus, we assume that the supply for raw materials for the bio fuel production in Germany and in the EU 25 will increase alike. Reasons for this prognosis are stagnating food-prices and simultaneously increasing yields. Additionally, the demand of land area for the increase of raw material for bio fuels can be made available without difficulties. For example, until 2010 more than 2.5 million ha agricultural crop land and until 2020 more than 5 million ha will be released from food production in Germany (Zeddies, 2006). Moreover the production of bio ethanol requires less land than of bio diesel, due to larger bio fuel yield per hectare from the crops-potential feedstock for bio ethanol (Kavalov, 2004). In respect to this aspect the potentials should be differentiate in bio diesel and bio ethanol production. Organisation of the Relations in the Value Chain After outlining the structure of the bio fuel sector and the recent developments we want to address in the following paragraphs the questions about “who is the initiator of the biomass-based energy value chain?”, “who coordinates the process of bringing biomass into final energy products?” and “how to organize it?”. Therefore, we sketch the organization of value chains in general and afterwards apply the knowledge on an example. Chain Organization

Value chains can be characterised as the collaborative interaction of independent enterprises (Brito and Roseira, 2005; Boehlje, 1999; Goerzen and Beamish, 2005). In agriculture most often they are supply chain networks. They are pyramidal-hierarchical structured so that they dispose over a focal company (Lazzarini et al., 2001). In the context of the organization of value chains on the one hand arising conflicts of individual interests must be managed and on the other hand interdependent actions of the actors must be coordinated (Gulati et al., 2005). Gulati (1998) emphasised that collaborations must be analysed not only from the perspective of the involved firms and the dyadic level of interaction but also out of an overall perspective. In the context of chain management, Duysters et al. (2004) named them firm level, dyadic level, and network level. Hanf/Dautzenberg (2006) combined these aspects in a general framework


5

of chain management. Although research on collaborations focuses on the interrelationships between firms, still single enterprises have to be regarded as the initial elements. Thus, in this paper we want to pay particular attention on the firm level. Even though it may sound self-evident we consider the willingness of the firms and the involved people to cooperate as the necessary prerequisite. Because cooperation demands that enterprises adjust their own actions with the ones of their partners, on the firm level general cooperativeness means that the enterprises have to be willing to abstain from some of their managerial freedom. Thus firms have to recognise collaborations as a means to overcome limitations of their resources. If a firm is participating in collaborations, it faces additional tasks and added work. Thus collaboration consumes resources of the firms e.g. time restraints of the managers and employees have to be reallocated. Therefore managing collaboration on the firm level demands particular managerial skills as well as resources (Duysters and Heimeriks, 2002; Dyer and Singh, 1998; Kale et al., 2002; Zaheer and Bell, 2005). Initiator and Focal Companies Of Value Chains

We pointed out that the initiation of the biomass-based energy value chain is a result of the EU Policy (EU Strategy for Bio fuels (2006), Biomass action plan (2005) and the concrete implementation in the resolution of 26 October 2006 in Germany). The key role for the embodiment and realisation of the process of bringing biomass into final energy products usually take the processor of bio fuels so that they can be considered to be the focal companies. Case Study “Crop Energies”

The processor Crop Energies (Südzucker Group) located in Zeitz (bio ethanol manufacturing plants for 260,000 t/a) has long-term contracts with agricultural enterprises in Germany to cover the grain supply. The main reason for using contracts is to safeguard their high investment costs. Actually, Crop Energies offers via local co-operatives or wholesalers contracts to local farmers. These contracts contain a price premium for protein poor bio ethanol-wheat a particular breed for energy production. Therewith, for the first time wheat with protein content less than 12 % receives a price premium (dlz 9/2006). The specific amount is not defined yet but it seems an interesting perspective for farmers. Backgrounds for this decision are different requirements for bio ethanol and food production. Whereas for food production usually wheat with high protein content and low starch content is necessary wheat for bio ethanol production needs high starch content combined with low protein content to ensure high crop of bio ethanol. In 2007 on the source of sugar beets Crop Energies will produce additionally 100,000 tonnes/a bio ethanol in Zeitz. On account of this about 600,000 tonnes of sugar beets are required for the supply for the bio ethanol plant every year. The decision about the investment in the new bio ethanol production plant has been dependent on the fact, that at least 80 % of the required sugar beets are produced under binding contracts with a term of 5 years. To supply bio ethanol beets farmers face the prerequisite that they have to subscribe the delivering right E. The delivering right E is a joint project of about 25,000 sugar beet farmers (SZVG, 2006). Having such a high number of farmers the total amount of the investment is divided upon them in order to share the risk. The amount of subscription of the delivering right E consists of fixed and variable components. The variable rate is coupled with the prices for bio ethanol. On the one hand the farmers get additions capital in rising markets and on the other hand in falling markets they will discharged. The fixed rate constitutes the own capital contribution of the farmer. The delivering right is also delivery commitment up to 2011. In 2006 farmers all over south Germany signed up production contracts. According to the association “Süddeutscher Zuckerrübenanbauer e.V.” and the regional associations the chances and risks were divided between farmers and “Südzucker” due to the fluctuating ethanol prices. Therefore the prices of bio ethanol beets will vary according to the changing ethanol prices. The price trend for ethanol is increasing on all markets, the world market price is also increasing


6

with 0.26 €/l in 2005. For Europe the actually import protection is € 0.192/l (Bayerische Landesanstalt für Landwirtschaft). Due to the great demand for bio ethanol beets on the part of the farmers every farmer is allowed to signed up for a maximum of 14,4 % of unabridged contract amount of sugar beets. Another important fact is the possibility of transfer the delivering rights for farmers until 2008. The above mentioned example of Crop Energies exemplifies the opportunity for farm enterprises to specialize in new markets. Even though the income risk is shared on both partners - due to the rather short-term nature of the contracts - investments into technology and training on side of the farmers can be regarded as risky. Because of this farmers might have only a limited cooperativeness. On the other hand the subscription of the delivering right safeguards the access to new production potentials. In the case of bio ethanol production the initiator is the processor, while only large plant production is profitable. Farmers have the opportunity to decide to invest in this new market with moderate risk. On the one hand production process knowledge about e.g. new breeds and cultivation and on the other hand knowledge about cooperation problems and developments of new markets are essential requirements. Conclusions Opportunities and Threats for Farmers and Processors

Farmers as well as producers can benefit from advantages but have to face also the flip side of the coin. The usage of contracts could be an advantage for farmers. They lower their income risks because they receive an ex ante defined price which is bound to defined qualities and quantities. But contracts have also a drawback because they bear the threat of dependencies. In the case of bio fuel production however this threat is not enormous because the production of bio-energy crops is at first sight not very different from the production of crops for food purposes. The common technology like machinery can be used as well as an akin production system with moderate adoptions. Minor differences will occur by getting optimal composition of input factors such as fertilizer and agro-chemicals. For instance, the above mentioned quality for wheat is no longer defined by high protein content (which can be achieved by additional nitrogen application) quite contrary low protein grains are demanded. Further advantage will arise to farmers in the different treatment of the commodity when the crop is no longer considered as food rather it is non-food. The benefit arises from lower standards for non food commodities. Especially this applies to transportation and storage. However, due to this development farmers face also danger. The non-food commodity will be transformed within progress in breeding and processing for its optimal use so that it is no more transferable for food-use which means that an alternative marketing is not possible. Thus, farmers get in a lock-in situation and their scope for tactical manoeuvre is narrowed, at least for short-term decisions. For processors the main motivations for integration in the value chain is to acquire secure supply i.e. they are interested to get a fixed demand with defined qualities and to reduce transportation costs. The example of Crop Energies demonstrates that processors find a way to get farmers actively involved in the supply chain. Processors, e.g. Choren are convinced that biomass supply will be scarce on the international market in medium term. In this case transportation costs will predominantly determine the price of biomass. Therefore bio fuel production will be profitable if processors can secure most of their biomass from local area. Summary The described changes in the biomass-based energy sector and the arising structures of vertical organized value chains raise a number of challenging issues. The main drivers for increasing future prospects for bio-energy in Europe are the Biofuels Directive (2003/30/EC) and the EU Strategy for Bio fuels (2006).


7

In total the EU’s production of liquid bio fuels amounted to a total of 3.2 million tonnes in 2005 which is an increase of more than 30 % compared with the previous year. Because solely in 2007 German enterprises are planning to enlarge the production capacity for bio ethanol production for 430,000 t/a as well as for bio diesel 1.9 million t/a we predict that the growth will continue. Because of this we discussed the emerging value chain organization and management requirements first in general and afterwards we applied it to a case study on bio ethanol production. Due to the chain’s characteristics in particular large companies such as the German Südzucker AG can be regarded as the initiator of these chains. In order to safeguard their high initial investments and to secure efficient supply these companies are relying rather on contract farming than on spot market interactions. On one hand the determination of the production by long-term contracts can lead to a restriction of the liberties of the farmers. On the other hand the integration in supply chains is becoming increasingly important and it will be essential for farmers to identify strategies for becoming compatible with such systems. References Bayerische Landesanstalt für Landwirtschaft, 2007. Zur Marktsituation von Bioethanol aus Zuckerrüben http://www.lfl.bayern.de/iem/agrarmarktpolitik/20256/. Boehlje, M., 1999. Structural changes in the agricultural industries: how do we measure, analyze and understand them? American Journal of Agricultural Economics, Vol. 81, No 5, 1028–1041. Brito, C. and C. Roseira, 2005. A model for understanding supply chain networks. Journal on Chain and Network Science, Vol. 5, 55-63. Commission of the European Communities, 2003. Directive 2003/30/EC of the European Parliament and of the Council of 8 May 2003 on the promotion of the use of biofuels and other renewable fuels for transport (OJEU L123 of May 2003). Commission of the European Communities, 2005. Biomass action plan, SEC (2005) 1573, Communication from the Commission, COM (2005) 628 final, Brussels, 7 December 2005. Commission of the European Communities, 2006. An EU Strategy for Biofuels, SEC (2006) 142, Communication from the Commission, COM (2006) 34 final, Brussels, 8 February 2006. dlz agrarmagazin Die Landwirtschaftliche Zeitschrift, 9/2006, p.22. Duysters, G.M. and K.H. Heimeriks, 2002. Alliance capabilities – How can firms improve their alliance performance? Paper at the 6. International Conference on Competence-based Management, IMD, Lausanne, Switzerland. Duysters, G.M., K.H. Heimeriks, and J.A. Jurriens, 2004. An integrated perspective on alliance management. Journal on Chain and Network Science, Vol. 4, 83-94. Dyer, J.H. and H. Singh, 1998. The relational view: cooperative strategy and sources of interorganizational competitive advantages. Academy of Management Review, Vol. 23, 660-679. EurObserver, 2005. Biofuels Barometer, June 2005: Paris, ObservÈr. Fachagentur für Nachwachsende Rohstoffe e.V. (FNR), 2007. http://www.fnr.de/. Goerzen, A. and P.W. Beamish, 2005. The effect of alliance network diversity on multinational enterprise performance. Strategic Management Journal, Vol. 26, 333-354.


8

Gulati, R. 1998. Alliances and networks. Strategic Management Journal, Vol. 19, 293-317. Gulati, R., Lawrence, P.R. and P. Puranam, 2005. Adaptation in vertical relationships: Beyond incentive conflicts. Strategic Management Journal, Vol. 26, 415-440. Hanf, J. and K. Dautzenberg, 2006. A theoretical framework of chain management. Journal on Chain and Network Science, Vol. 6, 79-94. Kale, P., J.H. Dyer, and H. Singh, 2002. Alliance capability, stock market response, and long term alliance success: the role of the alliance function. Strategic Management Journal, Vol. 23, 747-767. Kavalov, Boyan, 2004. Biofuel Potentials in the EU. European Commission, Joint Research Centre, Institute for Prospective Technological Studies. Lazzarini, S., Chaddad, F. and M. Cook, 2001. Integrating Supply Chain and Network Analysis: The Study of Netchains. Journal on Chain and Network Science, Vol.1, 7-22. Süddeutsche Zuckerrüben-Verwertungs-Genossenschaft eG (SZVG), 2006. Informationen zum Lieferrecht E, Beilage der SZVG zu Aussendung “Marktchancen nutzen – Bioethanol aus Zuckerrüben. Zaheer, A. and G.G. Bell, 2005. Benefiting from network position: firm capabilities, structural holes, and performance. Strategic Management Journal, Vol. 26, 809-825. Zeddies, Jürgen, 2006. Rohstoffverfügbarkeit für die Produktion von Biokraftstoffen in Deutschland und in der EU-25. Universität Hohenheim.


9

CHALLENGES FACING AGRICULTURE IN NEW ZEALAND

John Gardner Massey University, Palmerston North, New Zealand.


Abstract

The rural population of New Zealand as a proportion of the total is falling. An important consequence of this is diminished representation for the rural sector in the House of Representatives and on regional councils. The agricultural group within the rural sector is dominated by livestock farmers. Environmental legislation, passed in recent years, has important implications for livestock farmers. A major study on intensive farming and the environment questioned some common practices, in particular use of artificial nitrogen and some aspects of irrigation. The environment is now a major concern to many urban dwellers. In New Zealand, the bulk of the agricultural output is exported. Consumers in some markets, in particular Europe/UK, as well as some major retailers, are looking closely at environmental aspects associated with imported food as well as animal welfare, food safety and traceability. New Zealand livestock farmers face a number of challenges related to the environment and other issues. This paper addresses some of these. Keywords: New Zealand, livestock, farmers, exports, environment, consumers Introduction New Zealand has some important climatic features. The temperate climate allows pasture to grow year round. The livestock sector dominates New Zealand agriculture. Approximately 45 million sheep, 9 million cattle and 1.5 million deer graze around nine million hectares of pastures. The bulk of the output from this sector (lamb, mutton, wool, dairy products and venison) is exported to often far distant markets. The focus of this paper is on the livestock sector. In New Zealand this sector faces a number of challenges. The growth of the urban sector has diminished the political influence of the sector. Environmental groups are questioning some farming practices. Legislation is seen as raising compliance costs for farmers. “Property rights” are considered to be under attack. Some challenges confronting New Zealand farmers, for example those associated with the environment, are being faced by farmers throughout the world. Of particular importance to New Zealand farmers are the perceptions of overseas consumers towards farming practices in New Zealand. The environment is one issue but there are others, for example, animal welfare. This paper discusses some major challenges facing livestock farmers in New Zealand. Demographics and Political Representation The composition of the New Zealand population is changing. The urban population is growing and an increasing proportion of the population has been born outside New Zealand. Currently approximately 14% of the population live in a rural area and far fewer New Zealanders now have any experience of agriculture and rural life than a generation ago.


10

The shift in population balance from rural to urban has important implications for the rural sector in terms of representation both in the House of Representatives (Parliament) and on regional councils, an important tier of local government. Pastoral farming interests, are poorly represented in the House of Representatives, especially among the parties making up the current minority government. A search (www.parliament.nz/en-NZ/MPP/MPs/MPs1) of the published careers of minority government members of parliament identified two members who described themselves as “former farmers”, another had been a poultry farmer while a fourth is currently an organic farmer. There is currently no member of parliament on the government side who is a conventional pastoral farmer. There are twelve members of parliament on the opposition benches who list “small farmer” (1 member), “former farmers” (2) and “farmer” (10) in their careers. The information about members of parliament is that provided by the members themselves; it is possible that some could describe themselves as “farmers” but who have chosen not to do so. This seems unlikely however, on the government side. Regional Councils are important to farmers as they are involved with, inter alia, water and soil planning. There are twelve of these in New Zealand and farmers have a “reasonable” voice due to representation by a ward system. However recent legislation (Local Government Amendment Act, 2004) allows regional councils to elect councillors “at large” rather than on a ward basis. This is likely to diminish farmer representation as most regional councils have largely urban populations, making it more difficult for farmers to be elected as councillors. The national farmers organisation, Federated Farmers, has made representations to some regional councils to maintain a ward structure for councillor elections. This shift in political power provides important challenges for livestock farmers and those who act on their behalf. They need to put more resources into submissions to select committees (in Parliament) to ensure a farmer perspective is heard on proposed legislation important to farmers. At the regional council level, farmers must make submissions on regional council’s plans relating to soil and water issues where these impact on farmers. Legislation Recent labour related legislation, whilst not focused on farmers in particular, has nevertheless impacted on them principally as employers. Examples included the Employment Relations Act 2000, as amended in 2004, the Holidays Act 2003, the 2004 amendment to the Health & Safety in Employment Act 1992 and the 2002 and 2006 amendments to the Parental Leave and Employment Act to name but a few. These Acts have increased costs for employers. The test for the acceptable dismissal of an employee has been tightened, employees as from 1 April 2007 now get four weeks annual leave (previously three weeks) and penalties have been increased for failing to take “all practicable steps” to make the workplace safe. Farmers have however, gained the right to paid parental leave for the self employed, something that previously existed only for employees. Other legislation with important implications for farmers include the Biosecurity Act 1993, the Hazardous Substances and New Organisms Act 1996, the Resource Management Act 1991 (RMA), all of which are related to the environment, and the Animal Welfare Act 1999, again to give only some examples. These Acts impose obligations and costs on farmers. The RMA for example, requires resource consents for a range of activities. A current goal of Federated Farmers is to bring about changes to the legislation and the way it is implemented by regional and district councils. The Federation has identified the RMA


11

as its “top priority over the next two to three years”. It remains to be seen whether it will achieve its goal, given the strength of the environmental interests, in New Zealand. Farmers must now be “approved handlers” before being able to purchase and use a wide range of chemicals for control of weeds, pests and diseases. Furthermore some chemicals must now be “traced”, requiring records to be kept on purchases, usage and inventories. Farmers must be able to prove that chemicals have been applied at the correct rate. This requires accurate records keeping. For some farmers these obligations can be seen as a burden only, providing few if any tangible benefits. This is not necessarily so. Careless use of chemicals, for example on products exported, can lead to severe penalties. A beef farmer who supplied a meat processor with cattle contaminated by the chemical Endosulfan was traced after detection in South Korea. This led to a temporary halt on exports from the plant where the meat was processed. The farmer is being sued and is reported as saying that “the family farm in Northland is at stake” (The Dominion Post, Saturday, February 3 2007). The case is currently before the Court. Livestock Exports and the New Zealand Economy In relation to challenges facing the livestock sector in New Zealand, it is important to appreciate that New Zealand has a population of currently 4.17 m. Although the population is growing, domestic demand will always be limited. The livestock sector (dairying, sheep and beef, deer) are a major contributor to export receipts. Pastoral exports in the years 1998 to 2002 provided between 39.4% and 43.1% of total export income. If the returns from horticulture and arable, together with processed agriculture and forestry are included, total returns from primary exports ranged from 68% to 69.5% of total export receipts over the years 1998 to 2002. Most of New Zealand’s pastoral production (dairy products, lamb, mutton, beef and venison) is exported with only a minor proportion being consumed domestically (Table 1). Furthermore New Zealand has a high proportion of the product that is traded internationally. For example, New Zealand produces only about 2% of the total world milk production, but has about 40% of the world trade in dairy products. Table 1: New Zealand’s export production and share of world trade.

• Product • New Zealand production exported

(%)

• New Zealand share of world trade (%)

• • Wool • Lamb

• Mutton • Beef • Dairy • Deer

•

• • 90 • 90 • 79 • 78

• > 90 • 95

• •

• 75 • 53

• • 40

Source: Primary Economics (NZ) Limited, operating as Meat & Wool Economic Service of New Zealand, August 1999, Wellington.


12

Most of the livestock sector’s output will always need to be exported. Livestock farmers must compete in these markets with domestic producers and possibly other exporters. For New Zealand farmers to be competitive, costs on the farm, and in processing, transport, distribution and marketing are critical. In some markets pastoral farmers face trade barriers, for example tariffs and/or quotas, making it more difficult to compete. The challenge for New Zealand farmers remains, in principle, unchanged. New Zealand farmers must meet market demands more efficiently than their competitors. New Zealand farmers need to think about how to create a lasting competitive advantage in the market place by understanding and exceeding customers’ expectations. Customer expectations, particularly in Europe, are changing rapidly. Price alone is no longer enough. Food safety, traceability, environmental issues and animal welfare are all part of the mix. More recently “food miles” and “buy local” are surfacing as important issues. Food must also deliver a range of health and nutritional benefits. Environmental Challenges An important selling attribute of New Zealand in its overseas markets, particularly in Europe, is thought to be its clean green image. While New Zealand’s environmental record is relatively good, there are issues that need to be, and are, being addressed. The Parliamentary Commissioner for the Environment identified some of these in 2004 (Growing for Good, Intensive farming, sustainability and New Zealand’s environment, Parliamentary Commissioner for the Environment, 2004). Farming in New Zealand has intensified in recent years. In the dairy sector cows per hectare rose by 19% between 1994 and 2002, urea fertiliser per hectare went up by 162% between 1996 and 2002. In the sheep and beef sector, in intensive farms stock per hectare fell by 20% between 1981-2002 but lamb export carcass weights were up by 25% over approximately the same period. Tonnes of fertiliser on intensive farms in the sheep beef sector increased by 167%-263% over 1991-2002. The adverse environmental impacts of higher fertiliser and intensification on waterways, ground water and lakes was carefully documented by the Commissioner. The focus of the report is on synthetic nitrogen fertiliser and irrigation water as these have been the two key drivers lifting productivity in recent years but they also have the potential for adverse environmental impacts. The issues raised in the report are being addressed by the regional councils. In some instances councils have placed limits on nitrogen use. The Waikato Regional Council now requires a resource consent for pastoral farming in the catchment around Lake Taupo. New Zealand farmers face not only environmental challenges from fellow New Zealanders, but also increasing scrutiny from consumers overseas, because so much agricultural output is exported. The New Zealand Director General of Agriculture summed it up this way in his address to the AARES Annual Conference in 2007. “While other countries may find that they can’t be green if they are in the red, New Zealand can

not be in the black if we aren’t green”. An important environmental issue for New Zealand farmers that has surfaced in recent times is the so-called “food miles”. Proponents of “food miles” claim that the further food has to travel before it reaches the market place, the less sustainable or energy efficient it is and therefore the closer to the market food is


13

produced, the better it is for the planet. The concept of food miles is flawed as it fails to reflect the total energy used which includes production and processing as well as transport. Research at Lincoln University has found that the energy used in producing lambs and dairy products in New Zealand and then shipping them to the UK used less energy than that used by UK farmers. This finding has received extensive publicity in New Zealand and has also been reported overseas (Economist, December 9-15th, 2006). Although the food miles concept has been shown to be flawed, it nevertheless remains a concern in New Zealand. Thus the Minister of Trade was reported saying, New Zealand producers are:

“still threatened by the malicious use of food miles by protectionists and lobbyists seeking to shelter British producers from competition.”

Most will be aware that the major UK retailers (Tesco, Marks and Spencer, Asda and Sainsbury) have plans related to the environment for produce sold in their stores. Tesco intends to label all produce airfreighted to the UK. An even more ambitious proposal by the same company is to develop a “carbon footprint” labelling all products sold in their stores. It is the belief of the Minister of Trade that:

“measuring carbon footprints for New Zealand exports actually makes us look good”. Climate change, global warming and greenhouse gases have all been significant issues for some time, boosted in importance following the release of the Stern Report. For New Zealand farmers there are once again special challenges. Globally only 14% of greenhouse gases come from agriculture, but in New Zealand the estimate is 49%. The greenhouse gases in New Zealand consist of methane from livestock and nitrous oxide from animal waste and nitrogen fertiliser. Agricultural emissions have grown by one per cent per year since 1990 and it is anticipated that this rate will continue, at least for the medium term. However, productivity gains from farming animals more efficiently have resulted in lower emissions per unit of output. Forests play a vital role in lowering greenhouse gas emissions, absorbing carbon dioxide as they grow but releasing much of the carbon back into the atmosphere at harvest. Exports of wood products are an important part of the New Zealand economy constituting 10.4% of all merchandise exports in the year ended June 2006. However, in recent years deforestation (taking land out of forestry into another land use) of plantation forests has increased rapidly. Thus in the 2006 year about 12,700 ha were converted into pastoral farming but only 6000 ha were planted in forestry. The relatively poor returns from forestry in recent years have discouraged the conversion of land from pasture into forestry and encouraged the switch to livestock farming, following harvesting of the trees. Government in February 2007, released a document on sustainable land management and climate change. This outlined options and invited submissions on New Zealand’s future sustainable land management and climate change policies for the agriculture and forestry sectors. Currently meetings are being held throughout the country with farmers, growers and foresters to discuss the options foreshadowed in the document. These include possible taxes on nitrogen fertiliser and on land in forestry if that is converted to livestock farming following harvesting. Increasing Competition from Low Cost Producers New Zealand has built its agricultural export base on being a low cost producer of farm products of acceptable quality. However, competition from South America and other countries is challenging New


14

Zealand’s position. For example, in the last 15 years growth of exports of many dairy and meat exports from China has been twice that of New Zealand. The UN Food and Agriculture Organisation has reported that over the last 20 years meat production grew by 230 percent and dairy production by 200 percent. This growth has been largely due to adoption of technology and practices already developed and/or used in New Zealand. There is a quality gap between exports from New Zealand and those from South America, but it is clear that those countries can access the resources, skills and technology to make the necessary product improvement. Some farmers in New Zealand see South America as a threat to livestock farmers in New Zealand, but others see it as an opportunity. A small but unknown number of New Zealand farmers are known to have purchased farms in the “southern cone” countries of South America, of Chile, Argentina, Uruguay and Brazil. In 2006 a major agribusiness firm, PGG Wrightson, promoted “NZ Farming Systems Uruguay” with the intention of buying and developing farms for dairying and beef production in Uruguay using New Zealand technology and expertise in grassland management. A sum of $105 m NZD was raised. According to the Prospectus and Investment Statement,

“dry matter production in Uruguay can be boosted to New Zealand levels” and

“New Zealand style dairy farms can be established for about 25% of the cost of buying an established dairy farm in New Zealand”.

At 15 December 2006, the company owned nearly 6000 ha in Uruguay at various stages of development. Perceptions and Knowledge of New Zealand New Zealand is a small relatively isolated country and sometimes our agricultural policies are not understood by all, even when one might expect that they would be. For example, subsidies were removed from farming in the mid 1980s, the process being fully documented (Farming Without Subsidies, 1990). Many people, organisations, for example Federated Farmers and our own government, have put the message across at international conferences and meetings that New Zealand farmers are not subsidised. In my role as editor of the Journal of International Farm Management, I came across this sentence in a paper submitted for inclusion in the Journal:

“The US dairy industry, similar to other developed countries, is heavily supported through government programs that assist in stabilising milk production and farm incomes”.

The writer, not an American, holds a leading position at a university. I was very surprised at this sentence; either the writer was not aware that the New Zealand dairy industry is not subsidised or he/she considers New Zealand not to be a developed country. I exercised my editorial prerogative and changed the sentence! A group of New Zealand dairy farmers, suppliers of Fonterra (a major New Zealand dairy co-operative), when visiting Germany were told by lobbyists from that country that Fonterra should be broken up because it was driving up international prices. The New Zealand farmer who led the group claimed that a Mr Simon Michel-Berger from Copa-Copega, a lobby group representing Europe’s 11 million farmers, made the comment about Fonterra because (mistakenly) he believed that it was a state owned enterprise.


15

Accurate information for consumers will be extremely important in the marketing of New Zealand’s primary produce. It is vital that farmers, the major exporters and government ensure that consumers overseas are well informed on matters important to them. Well informed consumers however, may not be enough. Strongly held attitudes and beliefs are not easily changed. For example, the belief that one should “buy local” to support domestic producers might not change, even if it was shown that this leaves a greater carbon footprint than purchasing an imported item. Economics An important factor influencing the economics of livestock farming in New Zealand is the exchange rate. As over 80% of the output of an average sheep and beef farm is exported the exchange rate is a major determinant of farm gate prices. Table 2 shows for a typical sheep farm, a forecast of gross farm income, total farm expenses and net farm profit for 06-07 assuming exchange rates of US68.0 cents, UK 35.0 pence and 0.51 Euros. The implications of a 5% appreciation or depreciation of the New Zealand dollar on gross farm revenue, farm expenditure and net farm profit is also shown. Table 2: Sheep and beef farm revenue and expenditure and impact of variation in the exchange rate

• • • + 5%

• Exchange rate change

• - 5%

• Exchange rate change

• • Gross farm revenue

• • $297,500

• • $320,100

• • $2777,400

• Farm expenses • $235,000 • $236,570 • $233,570 • Net farm profit • $62,500 • $83,530 • $43,830

Source: Meat and Wool New Zealand, Paper No. PO 701, February 2007. It is clear from Table 2 why New Zealand farmers pay very close attention to the exchange rate. A devaluation of 5% over the whole season increases net profit by approximately 33%. Input costs would increase due to the rising cost of the imported component of inputs, for example, fuel, pesticides, plant and machinery and fertiliser. However, the increased gross farm revenue from a 5% currency depreciation ($22,600) farm exceeds higher input costs ($1570). The converse applies if the currency should appreciate. Conclusions New Zealand farmers face waning political power. In future it is likely a decreasing number of New Zealanders will have experience and knowledge of the rural sectors. The challenge for farmers is to reach out to the urban sector, to explain the issues confronting farmers, but most importantly, to try and understand the concerns of urban voters and to identify areas of common concern and shared interest. Farmers will find it difficult, if not impossible, to obtain legislative change if their wishes are strongly opposed by the urban sector. In some respects the challenges facing New Zealand livestock farmers now are similar to those confronting previous generations of livestock farmers. The bulk of the output from the livestock sector


16

must be exported, markets must be found and customers’ needs met. Farming profitability will always be an issue, as will the exchange rate because of its impact on farm gate prices. The principal challenge now facing livestock farmers in New Zealand will be, particularly in Europe, to meet the needs of far more discerning consumers. Environmental considerations, animal welfare, food quality, and traceability are just some of the issues that will progressively become more important. Importantly, New Zealand farmers must perform ahead of their competitors in an ongoing sustainable way. Only if they can achieve this will the long term future of New Zealand farmers be assured. References Sherwin, M. (2007). Opportunities, Threats and Sustainability: New Zealand’s Primary Industries.

Opening address to the Australian Agricultural and Resource Society’s 51st Annual Conference, Queenstown, 13-16 February, 2007.

Gardner, C. (2007). Affco supplier protests at meat-taint litigation. Business Day, The Dominion Post,

February 3, 2007. Primary Industry Economics (NZ) Ltd operating as Meat & Wool Economic Service of New Zealand,

August 1999, Wellington, New Zealand. Growing for good: Intensive farming, sustainability and New Zealand’s environment. Parliamentary

Commissioner for the Environment, October 2004, Wellington, New Zealand. Special Report: Food politics. Economist, 9-15th December, 2006, pp. 69-71. Sustainable land management and climate change. MAF Policy, Ministry of Agriculture and Forestry,

Wellington, New Zealand, 2006. NZ Farming Systems Uruguay Ltd. Prospectus and Investment Statement. ABN AMRO Craigs. Sandry, R. & Reynolds, R. (Editors). Farming without subsidies. New Zealand’s recent experience. A

MAF Policy Services Project, 1990. Sheep & Beef, Mid Season Update 2006-2007. Meat & Wool New Zealand Ltd, Paper No. PO 7001,

February 2007.


17

CURRENT SITUATION AND PERSPECTIVE OF THE HORTICULTURAL FARMS IN BULGARIA – CASE IN THE PLOVDIV REGION

Dr Elena Garnevska

Massey University, Palmerston North, New Zealand Email: [email protected]

Dr Jonathan Edwards and Prof. Roger Vaughan

Bournemouth University, Poole, UK

Abstract Agriculture/horticulture has traditionally been an important sector in the economy of Bulgaria. In the last two-three decades, agriculture has changed dramatically due to factors including economic reform from a centrally planned economy to a free market economy, political conflicts between the governing parties, agricultural reform, inefficient governmental decisions, poor legislation, lack of capital for investments, de-population of rural areas and the accession process towards the European Union (EU). This paper reviews the structural changes in Bulgarian agriculture since the period of Communism began (1944) and discusses the current situation for horticultural farms of different sizes in the Plovdiv region of Bulgaria. The respondents identified their cropping structure and land ownership patterns together with their marketing structure. Farm managers’ future vision is also discussed. The small-scale farms (less than 2 ha) were mainly subsistence farms that were primarily involved in vegetable production and their farmers (most often the land owners), perceived farming as a way of living and surviving in the transition towards a free market economy and joining the EU. The ‘medium’ farms (2-10 ha) were transitional and working under pressure for either survival or expansion. They mainly produced annual crops (vegetable and other agricultural crops) for the local market. The ‘big’ farms (farms over 10 ha) were more market and business orientated and were aiming at economic viability within the unstable and competitive environment. Together with their annual crops they also grew some perennials (fruits and grapes). Recent Bulgarian Ministry of Agriculture and Forestry (MAF) reports indicate that the number of farms over 10 ha has been increasing slowly and will likely represent the future of farming in Bulgaria. The dynamic external environments in Bulgaria over the last three-four decades did not provide stable conditions for farm modernisation, land expansion or establishment of new orchards and vineyards. Despite the difficult economic environment of the country, it can be argued that the horticultural farms have significant potential due to favourable natural and weather conditions together with the tradition of growing horticultural crops that has existed for centuries. Joining the EU will present new challenges and opportunities for the successful and sustainable future development of farm businesses in Bulgaria. Keywords: horticultural farms, farm characteristics, farm marketing, SWOT analysis, Bulgarian agriculture Introduction Agriculture has traditionally been an important sector in the economy of Bulgaria. In the last two decades, agricultural industry has undergone dramatic changes due to the economic reform from a centrally planned economy to a free market economy, political conflicts between the governing parties, agricultural reform, inefficient governmental decisions, poor legislation, lack of capital for investments, de-population of rural areas, accession process towards the EU and joining the EU in 2007 (OECD, 2000; Kostov and Lingard, 2002; MAF, 2002; Doichinova, 2003; Bachev, 2005; Bencheva, 2005). This paper evaluates the current situation of the different sized horticultural farms in the Plovdiv region of Bulgaria and is divided into five sections. The next section reviews the agriculture in Bulgaria. The


18

methodology is described in section three. The analysis of the data is reported in section four. The final section draws some conclusions.

Current Status of Agriculture/Horticulture in Bulgaria Bulgaria enjoys good natural conditions for agriculture/horticulture such as the fertile soils which, combined with a mild continental climate, provide a diversity of production systems (EC, 1998; OECD, 2000; Bencheva, 2005). In 1989, the transition towards a ‘free market’ economy began in Bulgaria. The reform in agriculture started with the introduction of a range of new regulations and laws that were developed in order to re-introduce private farming after 45 years of a Communist regime. The agricultural reform was characterised by the liquidation of the Agri-Industrial Complexes (AICs), the development of a private sector, land restitution, privatization and price liberalisation. As a result agriculture/horticulture was in a critical situation due to accumulated problems inherited from the period of Communism, the slow pace of reforms, lack of clear and consistent policies, reduced domestic demand and loss of the main export markets (EC, 1998; MAF, 2000; Georgieva, 2003). The farming structure that emerged after the liquidation of the AICs was a large number of private farms, (average size about 1.5 ha), private production co-operatives (average size of 600) and public partnerships. The majority of these agricultural enterprises (individual farms and co-operatives) are still transitional, in need of significant improvements and consolidation in order to be able to operate under the EU conditions (FAO, 1999; Georgieva, 2003; MAF, 2006). After 1998, a radical agricultural reform began in Bulgaria. Agricultural policies became more consistent with long-term goals to develop an efficient, competitive and export-orientated agricultural sector, to improve the incomes of those working in agriculture and to prepare the country for the EU accession. The Special Accession Programme for Agriculture and Rural Development (SAPARD) was introduced to prepare Bulgaria for the entry into the EU. In 2007, Bulgaria joined the EU and the impact of the CAP on Bulgarian agriculture is yet to be evaluated (EC, 2000; MAF, 2000; SENTER, 2000; Georgieva, 2003 MAF, 2006). Methodology

This study is one of the first to adopt strategic approaches to analyse agriculture/horticulture in Bulgaria. It is also one of the first to focus on the horticultural industry in Bulgaria and includes a sample of horticultural farms in the Plovdiv region. The Plovdiv region, one of the 28 regions in Bulgaria, is situated in central-south part of Bulgaria (Figure 1).


19

Figure 1: Map of Bulgaria

In this research horticulture includes fruits, vegetables and grapes. Data collection was undertaken during 2001. The research method used was structured face-to-face interviews as this took account of the farmers’ lack of experience with research interviews and the innovative nature of this study. Purposive sampling was employed due to the lack of an accurate and up-to-date list of the agricultural/horticultural farms in the Plovdiv region in 2001. The chosen sampling procedure (purposive) produced valid information for analysing the horticultural industry in the Plovdiv region. A total of 108 farmers were interviewed in their working places. A review of the literature suggested that size of the farm is a very important factor as it has a strong influence on the farm business performance and development. Farms in the sample were divided into the following groups: ‘small’ farms – less than 2 ha; ‘medium size’ farms – between 2-10 ha; and ‘big’ farms – more than 10 ha. The majority of the data collected was quantitative and was analysed using the Statistical Package for Social Sciences (SPSS). Some qualitative data derived from open ended questions. A range of descriptive analytical techniques were employed to determine patterns and relationships between variables.

Main results Cropping Structure of the Farms

More than half of the interviewees (53%) were cultivating fruit (Table 1). The most common fruits were apples, plums and cherries. The Plovdiv region is the biggest apple producer and second biggest producer of plums in Bulgaria (SENTER, 2000). The respondents cultivated fruit because they inherited their orchard/s as part of the land restitution process. They also stated that fruit were profitable during the transition period and have been traditionally grown in the Plovdiv region. The results also revealed that the farms of different size differed in their fruit orientation (χ2 = .023). The majority of the ‘big’ farms (76%) had fruit, whereas 60% of the ‘small’ farms did not cultivate any fruits (Table 1).


20

Table 1: Crop cultivation of the farms with different size

SIZE OF FARMS TOTAL

FRUITS SMALL MEDIUM BIG

Count % Count % Count % Count %

Yes 10 40 28 48 19 76 57 53

No 15 60 30 52 6 24 51 47 Total 25 100 58 100 25 100 108 100

Significance Value (χ2 = .023)

GRAPES

Yes 13 52 25 43 11 44 49 45 No 12 48 33 57 14 56 59 55

Total 25 100 58 100 25 100 108 100 Significance Value (χ2 = .747)

VEGETABLES

Yes 20 80 44 76 18 72 82 76

No 5 20 14 24 7 28 26 24 Total 25 100 58 100 25 100 108 100

Significance Value (χ2 = .803)

OTHER CROPS

Yes 15 60 46 79 22 88 83 77

No 10 40 12 21 3 12 25 23 Total 25 100 58 100 25 100 108 100

Significance Value (χ2 = .051)*

Note: * The validity of the chi-square test results is questioned because 20% of the cells have expected count of less than 5 and one or more cells have expected values less than 1

Grapes (table and wine) were cultivated by 45% of the respondents (Table 1). According to the interviewees, the rationale for cultivating grapes was very similar to those for the fruits, which were: inherited vineyards after the land restitution, profitability and increased demand from the increased number of private wineries. Grape production was largely stable during the transition period in the Plovdiv region, which is the second biggest in terms of area of vineyards after the Bourgas region (near the Black Sea) (MAF, 2002). One of the traditional varieties of wine grapes in Bulgaria ‘Mavrud’ is specific to the Plovdiv region and is a very popular crop for cultivation. There was no significant difference between the grape orientation of the farms with different size (χ2 = .747) (Table 1). The favourable natural conditions in the Plovdiv region, on the Thracian plain around the river Maritsa, has historically provided a sound basis for the development of the horticultural industry in the region and for growing vegetables in particular (MAF, 2002). This was confirmed by the respondents as the majority of them (76%) stated that vegetables were very important crops in their production system (Table 1). The reasons, according to the respondents, were that vegetables are annual crops that do not need big or long-term investments, they have traditionally been grown in the Plovdiv region and they were profitable having maintained relatively high prices. The most popular vegetables among these producers in the sample were tomatoes, peppers and potatoes. Farm size did not present any significant difference with the vegetable orientation of the enterprises investigated (χ2 = .803) (Table 1). A range of agricultural crops that were part of the production structure of some of the farms in the sample were collectively referred as ‘other’ crops and included arable crops, herbs, tobacco, etc. The majority of the farm managers that participated in this study (77%) cultivated together with their horticultural crops some ‘other’ crops. The main reasons for combining horticultural products with ‘other’ crops were: using resources available within the farm such as land, machinery and labour, profitability and necessity for


21

crop rotation. The results indicated that there were similarities between the farms with different size and the ‘other’ crop cultivation (χ2 = .051) (Table 1).

Marketing Structure

FAO (1999), SENTER (2000) and EC (2002) argued that the marketing structure in Bulgaria was poor due to the loss of the main international markets, reduced domestic purchasing power, the slow process of privatisation of the agri-food processing industry, lack of marketing skills among the farmers and limited marketing support by the Government. Prior to 1989, Bulgaria was a major exporter of agri-food products to the former USSR and other ex-socialist countries. Since then the country has not gained new market due to low competitive power, poor quality of products and increased competition from EU and other countries (OECD, 2000; MAF, 2002). The current markets of the farms within the sample in the Plovdiv region were investigated and the results revealed that 75% of them sold their production locally in the region. The national market was supplied by 21% of the farms and only 4% of them had international markets (Table 2). Farms of different size used different markets for their produce (χ2 = .004). The results revealed that 50% of the farms of more than 10 ha sold their production nationally. In comparison, the vast majority of the ‘small’ and ‘medium size’ farms (93% and 81%) were oriented towards their local market (Table 2). Table 2: The main markets of different types of farm in 2000

SIZE OF FARMS TOTAL

Current market SMALL MEDIUM BIG

Count

% Count

% Count

% Count

%

Local 13 93 34 81 4 33 51 75

National 1 7 7 17 6 50 14 21 International 0 0 1 2 2 17 3 4

Total 14 100 42 100 12 100 68 100 Significance Value

(χ2 = .004)*

Note: * The validity of the chi-square test results is questioned because 20% of the cells have expected

count of less than 5 and one or more cells have expected values less than 1

The farmers in the Plovdiv region have an advantage; one of the three established wholesale markets in Bulgaria is located in the Plovdiv region. However, according to FAO (1999) and Bachev (2005), the existing wholesale markets have functioned ineffectively and have been in need of significant improvement. FAO (1999) and SENTER (2000) argue that the distribution channels in Bulgaria have been under continuous development since the economic reform began and are still not well developed. After 1989, the large state monopolies in marketing and distribution in Bulgaria were dismantled. The wholesale and retail channels were privatised and that process resulted in the emergence of a large numbers of new private agents (suppliers, processors, intermediaries). This paper discussed the current distribution channels of the farms in the sample. About half of the respondents used a wholesale market for their products. The farms of less than 2 ha kept some of their production for self-consumption and sold the rest of it by themselves at the market. The ‘medium’ farms


22

mainly marketed their produce at the wholesale market (50%). The large farms, cultivating over 10 ha, mostly used wholesale markets and due to their bigger capacity also sold to distributors or processors. In comparison, some of the co-operatives used their previous contacts with processing factories to deliver their production or used distributors or the wholesale market (Figure 2). Figure 2: The distribution channels in the Plovdiv region of Bulgaria

By themselves at the market

Wholesale market

Distributors ‘middleman’

Processors

FARMING

INDIVIDUAL FARMS

CO-OPERATIVES

Small-scale farms Large farms Medium farms

30%

50% 20% 60%

30% 70% 30%

10%

50% 30%

20%

Source: (Author)

Both the secondary sources and the primary data suggested that the practice of growing under contract does not appear to be widely used among the respondents. However, a few commercial farms marketed relatively large amount of products trough advanced marketing channels (e.g. contract relationship with national or international companies).

SWOT Analysis of the Farms

Studying the internal capacity of the farms (strengths and weaknesses) provided helpful information for discussing the farm development. The results revealed that the key strengths of the farms within the sample were: possession of experience in agriculture (63%); availability of own machinery (48%); traditionally important sector in the Plovdiv region (41%); good natural conditions (37%) and independent management (24%) (Table 3). Farms of different sizes had different strengths. The vast majority of the farmers with ‘big’ farms (84%) identified the availability of their own machinery, while those with plots of less than 10 ha stated that their previous experience was their key strength. Another disparity observed was that 36% of the producers with a farm of more than 10 ha considered that independent management was one of their vital strengths compared to 16% of the growers with ‘small’ farms (Table 3). During the period of Communism, the government took all the managerial decisions and the role of the farm manager was to follow their directions. However, in the condition of a market economy, the farmer is responsible for all the business decisions, a challenging task that has been welcomed by some and frightened others.


23

Table 3: The top strengths, weaknesses, opportunities and threats of the farms with different size

SIZE OF FARMS TOTAL

STRENGTHS* SMALL MEDIUM BIG

Count % of cases Count % of cases Count % of cases Count % of cases

Having experience 17 68 37 64 14 56 68 63

Own machinery 4 16 27 47 21 84 52 48 Traditionally grown crops 16 64 24 41 4 16 44 41 Good natural conditions 15 60 19 33 6 24 40 37 Independent management 4 16 13 22 9 36 26 24

WEAKNESSES*

Lack of, or old machinery 20 80 39 68 18 73 77 72

Using old technologies 21 84 37 64 12 49 70 65 Having fragmented land 12 48 38 66 12 49 62 58 Having old plots of perennial crops

6 24 15 26 9 36 30 28

OPPORTUNITIES*

Planting new crops 9 36 26 45 9 36 44 41

Farm size expansion 7 29 27 47 4 16 38 36 Maintaining the same business 6 24 16 28 5 20 27 25 Applying new technologies 10 40 10 17 5 20 25 24 Market expansion 7 29 10 17 6 24 23 22

THREATS*

Unpredictable weather 20 80 44 76 19 76 83 77

Lack of or uncertain market 19 76 39 67 13 52 71 66 Unstable agricultural policies 15 60 31 53 17 68 63 58 Decreased consumer demand 7 28 17 29 7 28 31 29

Note: * This table includes only the most frequent answers given by the respondents. Percentages are

based on multiple response answers. They are the percentages of cases rather than responses therefore they do not sum to 100%

The key weaknesses stated by the respondents are presented in Table 3 and they were: lack of machinery or having obsolete machinery (72%); using old technologies (65%); having fragmented land (58%) and having old plots of perennial crops (28%). These findings were similar to those of FAO (1999), MAF (2002) and MAF (2006) which stated that after 1989 agriculture in Bulgaria had been characterised by a low level of technological innovation and this problem is yet to be solved. Although the farms within the sample inherited the same problems, accumulated over the periods of Communism and transition, there were some minor differences in terms of the weaknesses of the different sized farms. The results revealed that more than two thirds of the respondents with farms of more than 2 ha considered the lack of machinery or possession of obsolete machinery (more than 15-20 years) as their main weakness. However, the growers with farms of less than 2 ha stated their major weakness to be the use of old technologies (84%) (Table 3). As a result of the economic transition in Bulgaria after 1989 and joining the EU in 2007, the respondents confirmed that some opportunities has arisen and they identified the following: planting new crops (41%) to follow the new customers preferences and needs; expanding farm land (36%) assisted by the EU financial mechanisms; maintaining existing business level (25%) in a dynamic competitive environment; implementing new technologies (24%) and expanding new markets (22%). The results also revealed that the key opportunity for the ‘small’ farms was the application of new technologies (40%), whereas, the ‘medium size’ farms identified farm expansion in terms of their land (47%) and the ‘big’ farms were mainly oriented towards developing new crops (36%) (Table 3).


24

Changes in the external environment may either have beneficial or harmful effects upon the farm businesses, therefore any negative influences have to be avoided or overcome. Table 3 shows that the most important threats identified by the farmers were: unpredictable weather conditions (77%); lack of, or uncertain, market (66%); unstable agricultural policies, including high level of bureaucracy (58%); decline in consumer demand (29%). All farms irrespective of their size were threatened mostly by the unpredictable weather (Table 3). Conclusions Horticulture is an emerging field of research in Bulgaria and this study is one of the first to adopt strategic approaches. The research results suggested that the small-scale (often subsistence) farms (less than 2 ha) were involved primarily in vegetable production and their farmers, perceived farming generally as a way of living and surviving during the economic transition and joining the EU. The ‘medium’ farms (2-10 ha) were transitional and were working under pressure for either survival or expansion under the EU conditions. They produced mainly annual crops (vegetable and other agricultural crops) for the local market. The ‘big’ farms were more market orientated (farms over 10 ha) and were aiming at business viability within the competitive environment. Together with their annual crops they also grew some perennials (fruits and grapes). An investigation of the internal business capacity of the horticultural farms demonstrated that the key strengths were previous experience and owning machinery (although obsolete), while their major weaknesses were lack of machinery and the application of old technologies. The external environment both threatens and provides opportunities for the farm businesses in Bulgaria. The most noticeable opportunities and threats were the collapse of the Communist system and joining the EU in 2007. The main opportunities identified by the farmers were developing new products and land expansion while the key threats were the unpredictable weather conditions and uncertain markets. This research demonstrated that despite the difficult economic environment of Bulgaria, it can be argued that the horticultural farms have significant potential due to favourable natural conditions coupled with the tradition of growing horticultural crops has existed for centuries. Equally, joining the EU in 2007 has presented new challenges and opportunities for the successful future development of farm businesses in Bulgaria. References Bachev H., 2005. Assessment of sustainability of Bulgarian farms. 11th Congress of the EAAE, The

future of rural Europe in the global agri-food system, Copenhagen: August 24-27, 2005. Bencheva N., 2005. Transition of Bulgarian Agriculture: Present situation, Problems and Perspectives for

Development. Journal of Central European Agriculture, 6 (4), 473-480. Doichinova Y., 2003. Family farms in the transition period under Bulgarian conditions. Agricultural

Economics and Management, 48 (6), 35-39. EC, 1998. Agricultural Situation and Prospects in the Central and Eastern European Countries: Bulgaria.,

Brussels: European Commission. EC, 2000. EU and enlargement. Directorate General for Agriculture. Brussels: European Commission.


25

FAO, 1999. Strategy for agricultural development and food security in Bulgaria. Sofia: MAF and Food and Agricultural Organisation.

Georgieva M., 2003. Rural development in Bulgaria: Challenges of the accession. 40th Anniversary

conference, Rural development in Europe, London: 15-16 October 2003. Kostov P. and Lingard J., 2002. Subsistence farming in transitional economies: lessons from Bulgaria.

Journal of rural studies, (18), 83-94. MAF, 2000. National Agriculture and Rural Development Plan (2000 – 2006) for the Republic of

Bulgaria. Sofia: Ministry of Agriculture and Forestry (August 2000). MAF, 2002. Progress report on implementation of SAPARD in Bulgaria. Sofia: Ministry of Agriculture

and Forestry. MAF, 2006. National Agriculture and Rural Development Plan (2007 – 2013) for the Republic of

Bulgaria. Sofia: Ministry of Agriculture and Forestry. OECD, 2000. Review of Agricultural Policies: Bulgaria. Paris: Organisation for Economic Co-operation

and Development. SENTER, 2000. Bulgarian agriculture in transition: Prospects for co-operation of Dutch and Bulgarian

agribusiness. Hague: SENTER International


26

FOOD SECURITY: WHEN TO BUY DERIVATIVE INSTRUMENTS

Mariëtte Geyser University of Pretoria, South Africa Email: [email protected]

Michela Cutts

University of Pretoria, South Africa Email: [email protected]

Abstract Commodity prices are notoriously volatile which is a major source of instability and uncertainty for commodity-dependent developing countries. Commodity price volatility affects governments, producers (farmers), traders, processors, and local financial institutions financing production inputs in these countries. There have been several attempts to deal with commodity price volatility. A number and variety of international and national institutions and programs were designed for this purpose. Most of the earlier attempts concentrated in trying to stabilize prices through the use of buffer stocks, buffer funds, government intervention in commodity markets, and international commodity agreements. These schemes have not proven satisfactory in dealing with commodity price instability. Academics and policy makers began to emphasize the distinction between programs that attempted to alter the price distribution, either domestically or internationally, with programs that deal with market uncertainty using market-based solutions. The rise in market-based commodity risk management instruments has been significant since the development of derivative instruments. The aim of this study is to determine the optimal time for African countries to buy grain by making use of call option contracts. Chicago Board of Trade contracts and contracts traded on the South African Futures Exchange were compared to determine which exchange would be appropriate, and the optimal time in which the contract should be purchased. The article starts by looking at droughts in Southern Africa, ways available to insure and/or hedge against supply risk. Thereafter, agricultural commodity market variability and volatility were analyzed to end with the determination of an optimal hedging period. The study goes on to assess the optimal timing for Tunisa to hedge their supply risk using the South African Futures Exchange as compared to the Chicago Board of Trade. Keywords: Food security, call option, hedge, optimal time, volatility, supply risk Introduction Drought is a normal recurring event that affects the livelihoods of millions of people around the world, and especially the 200 million people living in southern Africa. Climate variability, including erratic and unpredictable seasonal rainfall, floods and cyclones, contributes to the risk of farming across most of southern Africa. Yield variability is a risk both farmers and governments have to contend with world wide when it comes to extensive course grain production. The African continent is particularly susceptible to high yield variability due to unpredictable weather patterns and frequent droughts (FAO, 1994). When it comes to maize, governments across Africa have used an assortment of intervention strategies to combat yield variability and its consequences on food security. These interventions range from single marketing channels to grain storage schemes and, more recently, crop insurance schemes as well as weather derivatives. These schemes have often proved to be very costly and inefficient in dealing with widespread drought and consequent famine (Gommes, 2006).


27

A number of factors may lead to agricultural drought, including, reduced rainfall, soil moisture levels, heat, and wind. Low rainfall does not necessarily lead to drought, nor is drought necessarily associated with low rainfall as poor timing of rainfall can lead to crop failure. Agricultural drought occurs when water supply is insufficient to cover crop or livestock water requirements (Abrahams, 1997). Much more than the occasional widespread and severe climatological drought which catches the attention of the media, it is this “invisible” agricultural drought which prevents farmers from achieving regular and high yields. The nature of agricultural drought makes it very difficult for governments and food aid programmes to know when and which intervention method would be most effective. Governments and non governmental organisations (NGO’s) such as the world food program, need to have some intervention strategy in place to counter the effects of a wide spread drought in a particular region Drought conditions frequently require government intervention in the form of emergency food relief, often supported by large amounts of donated food aid. Drought preparedness by governments has generally taken the form of creating food reserves (mainly maize) at national level to compensate for production shortfalls and provide for possible emergency relief. With the development of derivatives and agricultural insurance markets, governments have resorted to these as they are more cost effective than physical grain storage. While costly relief efforts have been perceived as a necessity, such short-term interventions have generally precluded support for longer-term development processes, particularly in those areas with dry climate conditions. As low and erratic precipitation is a key characteristic of these dryland areas, this fact of life must be reflected not only in the preparedness plans drawn up by governments, but also in the longer-term development strategies designed to prevent serious impact of future droughts on the environment and people’s livelihoods. This paper not only proposes an alternative view to crop insurance and weather derivatives, but also the optimal time in which this alternative intervention strategy should be implemented. Droughts in the Southern African Region According to the International Fund for Agricultural Development (IFAD), as cited by Benson, Thomson and Clay (1997), at least 60 percent of sub-Saharan Africa is vulnerable to drought and probably 30 percent is highly vulnerable. From 1980-2000, the Southern African Development Community (SADC) region was struck by four major droughts, notably in the seasons 1982/83, 1987/88, 1991/92 and 1994/95. This corresponds to an average frequency of once every four or five years, although the periodicity of droughts is not necessarily so predictable. FAO (1994) identified three drought cycles in the SADC region during the years 1960 to 1993 with lengths of 3.4, 7.1 and 5.8 years, respectively.

Drought is the most important natural disaster in southern Africa in economic, social and environmental terms (Buckland, Eele and Mugwara, 2000). A report by the United Nations Development Program (UNDP) states that drought is considered by many to be the most complex and least understood of all natural hazards, affecting more people than any other hazard (UNSO, 1999).

Benson and Clay (1998) reported that little research has been done on the macroeconomic impact of drought in SSA. The main reason is that drought is typically perceived as an agricultural or food supply problem. However, for most SADC countries drought represents the most important type of economic shock they are likely to experience. It is important for governments to understand the macroeconomic impacts of drought when developing drought management policies and programs.

Drought has primary and secondary (ripple) effects on a household or national economy. Primary or physical impacts include reduction in agricultural production, hydroelectric power generation, water intensive non-agricultural production (processing), and domestic availability of water, which has health implications. Secondary impacts are those that affect gross domestic product (GDP), e.g. reduction in


28

industrial output may lead to inflation and lay-off of labor, thus increasing unemployment. Collectively, these factors reduce demand, expenditure, savings and GDP.

Insuring and Hedging Against Supply Risks

Governments in the southern African region have a variety of policy options when it comes to drought intervention. The selection of a particular policy will largely be determined by the specific situation of the country and the cost versus benefit of the policy. Policies vary because governments are not only concerned about the financial survivability of the farmers given the drought, but also food security of the country’s people, in particular the more vulnerable, poorer segments of the population. Several reports and calculations show that physical grain storage is very costly, difficult to administer, and may have a significant impact on domestic prices (FPMC 2003). Dana et al. (2006) state that given the likely inefficiencies in the public storage sector, private sector storage, instead of public sector storage, could be subsidised. Continuous government grain storage is inefficient, due to the administrative requirements and costs of storing the grain, which is incurred irrespective of yield. On the positive side, government stored grain is readily available, however, government involvement could crowd out private sector initiatives. Consequently, Coulter (2005) suggests that it should be limited to no more then a small food security reserve.

Crop insurance has a long history, with various permutations of government support, and Hazel et al. (1986) and Skees (2000) explicitly indicate why multiple peril crop insurance programs have failed in developing countries. Skees (2000) states that the rainfall index can be used if the three major challenges of determining the correlation between critical rainfall periods and income, reliable rainfall measuring infrastructure, and the role government versus international reinsures have in protecting against systematic risk is resolved.

Although physical grain storage protects against price risk, crop insurance, and rainfall indexes only protect the farmer/government against crop failures or losses and the subsequent shortages. Given the lack of reliability of rainfall measuring infrastructure in most African countries, rainfall indexes become expensive due to basis risk and burn rates. Assuming a government has purchased insurance in the form of a rainfall index, should there be a drought and consequent crop failure, the onus is on the government to prove the loss before it can claim them. This would result in a time lag that would compound the delay caused by transportation of imported grains. Furthermore, as current grain prices are generally not included in the rainfall index, a lot of uncertainty would remain regarding whether the money received from the rainfall index would be enough to finance the necessary grain imports. A solution to both risks would be to purchase call options either on the Chicago Board of Trade (CBOT) or on the South African Futures Exchange (SAFEX), thus creating a “virtual storage facility”. The choice of where to purchase the call option depends on the individual country’s needs and transport infrastructure.

Governments and NGOs wanting to set up a “virtual grain storage facility” need to decide when is the most appropriate time to purchase the call options and when to exercise these options. The price of an option is largely influenced by two factors, volatility and time to expiry. Time to expiry of the option is fairly straight forward and requires little attention, volatility of commodity prices however warrants further discussion. Agricultural Commodity Market Variability and Volatility

To a world still recovering from the bursting of the internet bubble in 2001, the image most likely to be immediately conjured up by the word “volatile” might be that of an unstable stock market; or, in view of the balance-of-payments crises of the late 1990s, of unpredictable capital flows driven by fickle market sentiment to emerging market countries. But the adjective could equally be applied to the weather. In


29

India, for example, even though the share of agriculture in national output has dropped from one-half in the 1960s to one-quarter today, a good monsoon can still make a significant difference to GDP growth (Claessens et al., 1993). “Volatile” can also be used to describe a political climate, such as that prevailing in Iraq or Somalia; or the procyclical response of fiscal policy to fluctuations in the price of oil for an oil exporter such as Nigeria; or even the behaviour of a crowd in downtown Buenos Aires, Argentina, protesting the corralito or freeze on bank deposits in December 2001. Depending upon how one looks at it, volatility in mainstream economics has either been around for a long time or else is of more recent vintage. In common parlance, making a distinction among volatility, uncertainty, risk, variability, fluctuation, or oscillation would be considered splitting hairs; but, going back to Frank Knight’s classic 1921 work, Risk, Uncertainty, and Profit, there is a subtle difference in economics. Uncertainty describes a situation where several possible outcomes are associated with an event, but the assignment of probabilities to the outcomes is not possible (Eeckhoudt & Schlesinger, 2005). Risk, in contrast, permits the assignment of probabilities to the different outcomes. Volatility is allied to risk in that it provides a measure of the possible variation or movement in a particular economic variable or some function of that variable, such as growth rate. It is usually measured based on observed realizations of a random variable over some historical period (Hull, 2006). This is referred to as realized volatility, to distinguish it from the implicit volatility calculated, say, from the Black-Scholes (Black and Scholes, 1973) formula for the price of a European call option on a stock. To date there is no consensus on how volatility should be measured. Various authors follow different methods in calculating volatility. See Thurnsby and Thurnsby (1985, 1987), Bailey, Tavlas and Ulan (1986), Chowurdy (1993), Klein (1990), Koray and Lastrapes (1989), Nelson (1992), Engel and Russel (1998), Engel (2000), Zimmerman et al (2001), Szego (2002), Engel and Russel (2005, 2006), and Nwogugu (2005 and 2006). Determining the best method for calculating volatility is beyond the scope of this paper. The objective is rather to compare the volatility of the same commodity on two different markets. It is for this reason that any of the above methods, and many others, are suited for the task. The Chicago Board of Trade states that volatility is a measurement of the change in price over a given period of time. It is often expressed as a percentage and computed as the annualized standard deviation of the percentage change in daily price (CBOT 2006).

2

1111

1*250 ∑ =

−−

−

−=

n

ii

i

i

it

p

p

p

p

nV

Where pi is the closing spot price and n is the number of days over which the volatility is calculated. Because this method of determining the volatility of the commodity prices is used by one of the largest grain markets in the world it is the volatility calculation method of choice for this article. Agricultural commodity prices respond rapidly to actual and anticipated changes in supply and demand conditions. Demand and supply of farm products, particularly basic grains, are relatively price-inelastic (i.e., quantities demanded and supplied change proportionally less than prices) and weather can produce large fluctuations in farm production and therefore price.

Fundamental factors are primary drivers of price. On the South African Futures Exchange (SAFEX), the fundamental factors determining the price of maize and wheat are: supply and demand at the international level, as reflected in the Chicago Board of Trade (CBOT) price, domestic supply, demand and stock levels, as well as the Rand-Dollar exchange rates as it directly affects the import and export parity price


30

(FPMC, 2003). In light of the fact that the USA is by far the largest grain producer, it is logical that changes in supply and demand in the USA would not only affect the CBOT price but also in smaller grain producing countries, such as South Africa. Meyer et al (2006) state that the equilibrium price in the smaller market can be estimated as a function of the equilibrium price in the dominant market, the exchange rate and the transaction costs. Meyer tested the effect of a 10% increase in the world price on the South African producer price of yellow maize, resulting in an average percentage change of 7.3% indicating a strong link between the world price and the domestic producer price. In fact, converting the monthly average CBOT maize price to Rand terms using the Rand/US Dollar exchange rate, the CBOT price and the monthly average SAFEX maize price have a 0.911 correlation.

In light of the above, one therefore expects the SAFEX price to follow similar volatility patterns as CBOT and the exchange rate. A study conducted by Geyser and Cutts (2006) concluded that the SAFEX spot price, namely the yellow maize spot price (YMAZ) and the white maize spot price (WMAZ), is generally more volatile (61% of the time) than the CBOT price. CBOT and the exchange rate follow more or less the same up and down trends if average monthly volatilities are compared. The same is true for white and yellow maize on SAFEX. CBOT and SAFEX have periods where the same up and down trends occur, but there are also periods when the up and down trends do not correspond. What causes these differences? Fundamental factors, supply in particular, influence the price volatility of SAFEX maize prices, as indicated by Figure 1. Figure 1: Price volatility on SAFEX and ending stock levels

0

10

20

30

40

50

60

70

80

90

100

Dec

-01

Feb-0

2

Apr-0

2

Jun-

02

Aug-0

2

Oct-0

2

Dec

-02

Feb-0

3

Apr-0

3

Jun-

03

Aug-0

3

Oct-0

3

Dec

-03

Feb-0

4

Apr-0

4

Jun-

04

Aug-0

4

Oct-0

4

Dec

-04

Feb-0

5

Apr-0

5

Jun-

05

Aug-0

5

Oct-0

5

Dec

-05

Feb-0

6

Apr-0

6

Jun-

06

Vo

lati

lity

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

1000 t

on

s

Sagis Wmaz Sagis Ymaz Sagis total Wmaz YMAZ

From figure 1 one can see that price volatility tends to be higher in periods with low stock1 (SAGIS total) levels and vice versa. South African stocks tend to be low between February and June, this is also when there is a lot of uncertainty regarding the current crop. The differences in volatility between SAFEX and CBOT still need to be explained.

1 Stock levels were obtained from the South African Grain Information Service (SAGIS) http://www.sagis.org.za


31

The difference between the CBOT and SAFEX maize price volatilities can be explained when planting and harvesting seasons are taken into account. Figure 2 reports the monthly average volatilities for CBOT and SAFEX maize prices with planting and harvesting seasons taken into account.

Figure 2: Average monthly volatility on SAFEX and CBOT markets per season

0

10

20

30

40

50

60

Vo

lati

lity

Wmaz YMAZ CBOT R/$ Gold

Planting time Harvest time

From figure 2 it is clear that volatility of white maize (Wmaz), yellow maize (Ymaz) and Chicago Board of Trade (CBOT) yellow maize number 22 is generally high during the planting season, gradually decreasing as the value of the current crop becomes more certain. This understanding of the different levels of volatility can be utilised when deciding when to purchase a call option.

Determining Optimal Hedging Period

Options can give investors the flexibility to hedge market exposure, speculate on a specific market move, or allow investors to put on simple to complex option positions called spreads. The question is, given the volatile nature of commodity markets, when would it be the advisable for a government, organization, or user of maize to hedge his/her exposure and protect him/herself against future price increases?

The Black and Scholes (Black & Scholes, 1973) model is often used to determine the price of an option. Based on this model, the average monthly annualized volatilities, the time to expiration, the assumption that the interest rate is adjusted to zero, the average price for May and July option contracts on SAFEX and CBOT were calculated. May and July contracts were selected because countries purchasing these options would know if they need to exercise the option due to a crop failure or other, as it coincides with the end of their growing seasons. Based on the above, an assessment was then made on the cheapest time to buy an option.

The premium cost for various ATM options traded on SAFEX were calculated and are shown in figure 3.

2 CBOT nr 2 are used for export purposes, therefore the usage of CBOT nr 2 maize


32

Figure 3: Call premium cost/ton/month for SAFEX option contracts

0.00

50.00

100.00

150.00

200.00

250.00

Dec

-01

Feb-0

2

Apr-0

2

Jun-

02

Aug-0

2

Oct-0

2

Dec

-02

Feb-0

3

Apr-0

3

Jun-

03

Aug-0

3

Oct-0

3

Dec

-03

Feb-0

4

Apr-0

4

Jun-

04

Aug-0

4

Oct-0

4

Dec

-04

Feb-0

5

Apr-0

5

Jun-

05

Aug-0

5

Oct-0

5

Dec

-05

Feb-0

6

Apr-0

6

Jun-

06

R/t

on

May WM May YM Jul WM Jul YM

From the above graph, it is clear that the most expensive time to purchase a July call option is January, a typical weather month. From the above it is also clear that the cheapest time to buy a call option is just before harvest time, this is however impractical, and therefore the cheapest time to purchase a call option is just before planting, and in the South African case this would be in September. Figure 4 indicates the call premium cost for ATM options traded on CBOT for the various contract months. Figure 4: Call premium cost/ton/month for CBOT option contracts

0

50

100

150

200

250

Dec

-01

Feb-0

2

Apr-0

2

Jun-

02

Aug-0

2

Oct-0

2

Dec

-02

Feb-0

3

Apr-0

3

Jun-

03

Aug-0

3

Oct-0

3

Dec

-03

Feb-0

4

Apr-0

4

Jun-

04

Aug-0

4

Oct-0

4

Dec

-04

Feb-0

5

Apr-0

5

Jun-

05

Aug-0

5

Oct-0

5

Dec

-05

Feb-0

6

Apr-0

6

R/t

on

CBOT May CBOT Jul


33

As with the SAFEX market, the cheapest time to purchase a call option on the CBOT-market is just before planting, which in the American case is in March. The decision that still remains is which market should be used for hedging purposes for food security. This decision will be based on a number of factors including among others, transport costs. For this reason, cost comparisons between the CBOT and SAFEX market were calculated for Tunisia. Transport costs from the USA to Tunisia were obtained from the International Grains Council, while those from South Africa to Tunisia were obtained from Cargill South Africa. Table 1: Total cost of at the money call option and transport costs of maize to Tunisia

• • Tunisia

• • CBOT • SAFEX

• Contract month • May • July • May • July

• Call option R/ton • 35.68 • 45.36 • 73.46 • 82.11 • Transport $/ton • 45 • 45 • 33 • 33 • Average exchange

rate R/$ • 6.05 • 6.41 • 6.05 • 6.41

• Total cost $/ton • 50.90 • 52.08 • 45.14 • 45.81 Table 1 clearly shows that purchasing a call option in March or April on the CBOT market is cheaper than purchasing the same option in September on the SAFEX market. This difference in price is mainly due to the different times to expiry of the option and the underlying volatility. The above allows for two different policy options. The first option is the annual purchase of a call option on the SAFEX market in September, as a form of insurance against crop failure as price and volumes would be secured prior to the coming planting season. The second option is a reactive measure, when is appears that there is going to be a drought or crop failure for whatever reason, then a call option can be purchased on the CBOT market. This option would be purchased in March or April, requiring policy makers to have a good understanding of the crop situation throughout the year. Conclusion Production of maize in southern Africa is dominated by South Africa, which accounts for the bulk of production in this region. Both poor harvests and bumper crops in South Africa will have a major impact on price formation and trade flows in southern Africa. Food security in the face of frequent droughts and other natural disasters is a common problem on the African continent. Governments across the continent have used a variety of policies to mitigate the effects of substantial crop losses, but very few have been sustainable in the long term due to their high cost and administrative requirements. An alternative policy option available to the policy maker that has not been widely used is the purchase of call options for course grains as a way of guaranteeing a certain volume at a predetermined price, while lowering administrative costs. This paper further suggests that the optimal time of the year in which governments and NGOs should purchase May and July call options, based on seasonal volatility and the time to expiry of the contracts is September on the SAFEX market and March or April on the CBOT market. Stocks, insurance schemes, and forward markets or other derivatives all impose known costs to reduce unpredictable risks. For each, if the costs are low enough, it may be possible for countries (or producers) to make their own arrangements; if the costs are high (and for poor developing countries, costs may be considered ‘high’ at a lower level than for developed), it may be necessary to share some of these costs with donors.


34

Although this paper presents a strategy to reduce the effect of yield and price risk, research still needs to determine how to finance the exercising of the call option. Stocks, insurance schemes, and forward markets or other derivatives all impose known costs to reduce unpredictable risks. Commodity risk management needs to fit into a country’s overall strategy for managing external risk and liability. In some countries, financing can be linked to the price of a commodity and financial instruments can serve a financing and hedging function. They have the advantage of relying on market determined prices and shifting risk away from the government to entities better able to manage and willing to assume risks. References Abrams, L. 1997. Drought policy - water issues (available at http://www.thewaterpage.org/). Bailey, M.J., Tavlas, G.S. and Ulan, M. (1986). Exchange rate variability and trade performance:

Evidence for the big seven industrial countries. Weltwirthschaftliches Archiv, 122, 466 – 477. Benson, C. & Clay, E. 1998. The impact of drought on sub-Saharan African economies: a preliminary

examination. World Bank Technical Paper No. 401. Benson, C., Thomson, A. & Clay, E. 1997. The macroeconomic impact of drought. In Proc. Highlevel

Regional Drought Policy Seminar, SADC, 1-19 November. Gaborone. Black, F. & Scholes, M. 1973. The Pricing of Options and Corporate Liabilities. Journal of Political

Economy, 81:3, 1973, pp. 637-654. Buckland, R., Eele, G. & Mugwara, R. 2000. Humanitarian crises and natural disasters: a SADC

perspective. In E. Clay & O. Stokke, eds. Food aid and human security. European Association of Development Research. London, Frank Cass Publishers.

Cargill South Africa CBOT 2006 :http://www.cbot.com/cbot/pub/page/0,3181,1237,00.html Chowdhury, A.R. (1993). Does exchange rate volatility depress trade flows? Evidence from error-

correction models. The Review of Economics and Statistics, 76, 700-706. Claessens, Stijn., Kneafsey, Devin, P. and Kroner, Kenneth F. (1993) Forecasting volatility in

commodity markets. Policy Research Working Paper Series, no. 1226 : The World Bank. Coulter, J. P., 2005 “Making the transition to a market-based grain marketing system”, paper prepared for

the World Bank-DfID workshop, “Managing Food Price Risks and Instability”, Washington DC, 28 February-1 March 2005.

Dana, J., Gilbert, C. L., Shim, E. 2006 Hedging grain price risk in the SADC: Case studies of Malawi and

Zambia. Food Policy 31, 357-371 Eeckhoudt, L. and Schlesinger, H. (2005). Putting risk in its proper place. CESifo Working Paper

Series No. 1462 Engel R. and Russel P. (2005). A discrete-state continuous-time model of financial transactions prices

and times: the autoregressive conditional multinomial-autoregressive conditional duration model. Journal of Business Economics & Economic Statistics, 23 (2):166-180.


35

Engel R. and Russel P. (1998). Autoregressive conditional duration: a new model for irregularly spaced

transaction data. Econometrica, 66(5): 1127-1162 Engel R. (2000). The econometrics of ultra-high frequency data. Econometrica 68(1): 1-22 FAO. 1994. Rainfall variability and drought in sub-Saharan Africa since 1960, by R. Gommes & F.

Petrassi. Agrometeorology Series Working paper No. 9, FAO, Rome. Food Price Monitoring Committee final report, 2003 National Agricultural Marketing Council, Pretoria Geyser, J.M. and Cutts, M. 2007. SAFEX maize price volatility scrutinized. Agrekon (to be published

in September 2007). Gommes, R. 2006. Data issues in climate related risk and impact assessments for food security. Living

with Climate conference 2006. Hazel, P., Pomareda, C., Valdes, A., 1986. Crop Insurance fro Agricultural Development: Issues and

Experience. The John Hopkins University Press, Baltimore. Hull, John, C. (2006). Futures, options and other derivatives. 6th ed. London : Prentice Hall. International Grains Council, www.icg.org.uk Klein, M.W.(1990). Sectoral effects of exchange rate volatility on United States exports. Journal of

International Money and Finance, 9, 299-308. Koray, F., and Lastrapes, W.D. (1989). Real exchange rate volatility and U.S. bilateral trade: A VAR

approach. The Review of Economics and Statistics, 71, 708-712. Meyer F., Westoff P., Binfield J., and Kirten J.F. (2006). Model closure and price formation under

switching grain market regimes in South Africa. Agrekon. 45(4):369-380 Nelson D. (1992). Filtering and Forecasting with mis-specified ARCH models I: getting the right variance

with the wrong model. Journal of Econometrics, 52: 347-370 Nwogugu, M. (2005). Towards multifactor models of decision making and risk: a critique of prospect

theory and related approaches, part one & two, Journal of Risk Finance 6(2). Nwogugu, M. (2006). Further critique of GARCH/ARMA/VAR/EVT Stochastic-Volatility models and

related approaches. Applied Mathematics and Computation, 182: 1735-1748. Skees J., R. 2000. ARole for capital markets in natural disasters: a piece of the food security puzzle. Food

Policy 25, 365-378 Szego G. (2002). Measures of Risk. Journal of Banking & Finance, 26 (5): 1253-1272. Thurnsby, J.G., and Thurnsby, M.C. (1985). The uncertainty effects of floating exchange rates: Empirical

evidence on international trade flows. In Ardt et al. (Eds.), Exchange rates, trade and the U.S. economy (Year: 153-166). Cambridge: Ballinger.


36

Thurnsby, J.G., and Thurnsby, M.C. (1987). Bilateral trade flows, the Linder hypothesis and exchange rate risk. Review of Economics and Statistics, 69, 488-495

United Nations Special Office (UNSO). 1999. International workshop on coping with drought: best use of

climate information for farmer decision making. Kadoma Ranch, Zimbabwe, 4-6 October. Zimmerman H., Neuneier R., and Grothmann R. (2001). Multi-agent market modelling of foreign

exchange rates. Advances in Complex Systems, 4(1): 29-43


37

THE DETERMINANTS OF ENTRY AND EXIT DECISION IN DUTCH GLASSHOUSE HORTICULTURE

Natalia Goncharova* and Arie J. Oskam Wageningen University, The Netherlands


Abstract

Entry of new and exit of existing firms are two crucial decisions and they have a special meaning for development of economy and the particular sector or industry. It is a way of introducing new technologies, products and management approaches. The Dutch glasshouse horticulture is a business card of Dutch agriculture due to quick adjustment to new technologies and implementation of many now-how. The unique way of trading of horticulture production through auction system implies a strong competition with horticulture firms not only inside of the Netherlands but also outside. The evolution and adaptation of the sector to such changes are reflected in the process of firm entry and exit. Therefore it is important to examine the firms’ entry and exit decisions. By considering the entry as an investment decision and exit as a disinvestment (negative investment) decision, the findings in investment theory can be applied for explanation of observed changes in population of operating firms. The economic literature on investments and entry-exit decision suggests different possible theoretical models to explain choices of entry, exit and size of firms. This article uses the model developed by Dixit to model the entry and exit decision in Dutch glasshouse horticulture. These decisions can be considered in contents of Real option theory as one of the options of the firm to “act” that has an alternative to “wait”. The recognition of waiting option value is the fundamental concept underpinning the Real Options theory that was developed by Dixit and Pindyck (1994). Any decisions taken now has an opportunity cost, in the sense that it kills off the option of waiting for further information and the possibility of making better decision. In the evolving environment, time brings more information about the future prospects of the project and they should be considered in today’s menu of choices (Dixit and Pindyck, 1994). In his earlier article Dixit (1989) derived exit and entry trigger prices of investments and examined effect of their changes on entry and exit decisions in numerical examples. Developing this idea it is possible to say that output and input prices (and their expectations) are driving the investment decisions in the way that they change cash flow (Dixit, 1992). Changes in prices (and in expected cash flow) can attract firms to the sector or to push them away. In conventional economics firms are induced to enter if current revenue exceeds sunk costs (“Marshallian trigger point”) and to exit if revenue falls below sunk cost. However it is often observed that farmers prefer to wait with entry or exit decision, expecting that prices and revenue can change in the future. In model of Dixit (Dixit, 1989; Dixit, 1992) a wedge between the Marshallian trigger point and “observed” trigger point produces zone of “hysteresis” in which firms do not respond to price signals. Then we can formulate research questions: What are the trigger points for exit and entry firms in horticulture? What is an impact of these trigger points on observed number of firms? What is the dynamic of entry and exit barriers for glasshouse horticulture? Another important issue is the high heterogeneity of entry and exit and absence of clear classification of entry and exit. The different way of entry can affect the entry decision itself and the length of surviving of firm after entry. The different types of exit can indicate about different processes in a sector, which provide a better understanding of reallocation of sources. The goal of this study is to develop empirically applicable classification of entry and exit and to investigate the impact of real option trigger points on number entering and exiting firms for Dutch glasshouse horticulture. Keywords:Dutch agriculture, waiting option value, Marshallion trigger point


38

Introduction Decisions about entry or exit are accompanied by investments that are likely to be irreversible. These two decisions, which are crucial for the firm, have profound implications for economic growth, because the entry and exit of firms can be beneficial for productivity growth, technological upgrading and employment generation. According to the OECD (2003), the entry and exit of firms accounts for 20-40% of total productivity growth in eight selected OECD countries. By considering entry as an investment decision and exit as a disinvestment (negative investment) decision, the findings in investment theory can be applied to explain industry dynamics. The economic literature suggests different theoretical and empirical approaches to explain choices of entry, exit and size of firms ( for an overview see, for example, Siegfried and Evans, 1994). This article is based on Marshall’s model of long-run and short-run equilibriums, which assumes that firms are induced to enter if current revenue exceeds sunk costs (“Marshallian trigger point”) and to exit if revenue falls below sunk costs. However, it is observed that firms sometimes prefer to delay an entry or exit decision, in the expectation that prices and revenue can change in the future. The real option theory postulates that uncertainty will affect the entry-exit investment decisions in such a way that it will change trigger points. In the model of Dixit (Dixit, 1989; Dixit, 1992), a wedge between the Marshallian trigger point and “observed” trigger point produces a zone of “hysteresis” in which firms do not respond to price signals. The goal of this study is to investigate the impact of investment trigger points on the number of entering and exiting firms for Dutch glasshouse horticulture. Dutch glasshouse horticulture can be characterised as a dynamically changing, highly competitive, and capital intensive sector. The evolution and adaptation of the sector to new technologies, to consumer preferences and to market requirements are reflected in the process of firms’ entry and exit. For this reason, it is suitable data for studying the firms’ entry and exit investments. In the next section, the theoretical model is presented. It also includes the specification of empirical models of entry and exit; the negative binomial econometric model is used for estimation. Section 3 discusses the data, and provides an analysis of changes in trigger points over time as well as the comparison of different types of trigger points. Section 4 provides the estimation results of the different specifications of econometric models indicating the effect of trigger points on entry and exit. Finally, Section 5 closes with some concluding and qualifying remarks.

Modelling of Entry and Exit Investment Decisions Theoretical Model

The long-run competitive equilibrium is determined not only by the price and output levels of the firms but also by the number of operating firms. Following MsCollel et al. ( 1995, p. 335 ) the central assumption is: “A firm will enter the market if it can earn positive profits at the going market price and will exit if it can make only negative profits at any positive production level given this price.” The long-run equilibrium price (p*) equates demand with long-run supply, where the long-run supply takes into account firms’ entry and exit investment decisions. Consider an industry initially in a long-run equilibrium position, which assumes number N0 of operating firms and long-run cost c (Figure 1, a). Suppose that demand shifts upward, then the industry will immediately move to a new short-run equilibrium position. The shock in demand causes an increase in prices to pS and the output per firm increases to qS; this can influence the investment decision of firms. Because firms’ profits increase, operating firms earn more in the short-run (due to pS>c) and can even be induced to make investments to expand; inactive firms can be induced to invest in entry.


39

Figure 1: Impact of trigger points on Entry and Exit

New Long-Run Equilibrium

Initial Long-Run Equilibrium

Short-Run Equilibrium

SqN 0

qN 1

cp =*

b ) Exit

p

Sp

qN 0

Q

New Long-Run Equilibr ium

Init ial Long-Run Equilibrium

Short-Run Equilibrium

SqN 0

qN 0

cp =*

qN1

p

a ) Entry

Sp

Q

In the long run, firms enter in response to the increase in profit, with the number of firms increasing to N1>N0; the industry will then move to the right along a new demand curve until it reaches the new long-run equilibrium. The graph (b) demonstrates the change in the number of firms as a result of the exit of firms as an adjustment to the new long-run equilibrium. In the long run, firms exit in response to the decrease in profit, with the number of firms falling to N1<N0. Now, consider that a firm’s profit-maximising investment decision is to enter or to remain inactive. A firm has to invest a lump sum k and will have a variable cost w for the production of output. In the case of an exit decision, it must also pay a lump sum l, which it loses due to the exit of the firm, and a variable cost w will be saved. The goal of the firm is to maximise expected net present value (NPV). The standard Marshallian theory (Marshall, 1920) postulates that a firm will invest (and enter) if expected NPV is greater than zero, and in the case of an operating firm a decision to exit will be undertaken when NPV is negative. Then for the entry investment of a firm, the trigger point HW is Marshall’s long run cost (when NPV>0), which is the sum of the variable cost and the interest on the sunk costs:

kwWH ρ+≡ (1) where ρ is interest rate. The Marshallian trigger point for exit disinvestment of a firm (NPV<0) becomes:

lwWL ρ−≡ (2) The recent developments described in articles of Dixit (1989), Dixit and Pindyck (1994) introduced a discussion concerning a difference between Marshallian trigger points and Real Option trigger points. The difference can be explained by the presence of uncertainty that causes a firm to consider the option of waiting. In Dixit (1989) we find the following relationships for the Real Option entry PH and exit PL trigger points:

HH WkwP ≡+> ρ (3)

LL WlwP ≡−< ρ (4)


40

In the same article, the author analytically derives a closed form solution for trigger points that take into account uncertainty and the effect of changes in expectation of output prices ( µ ), uncertainty ( 2σ ) and interest rate ( ρ ) on trigger points.

Empirical Model

From equations (1-2) we can numerically calculate Marshallian entry and exit thresholds. In the case of Entry, firms consider parameters of a potential sector to enter, consequently ρ is an average value indicating the current profitability of the sector as perceived by a potential entrant. kw, are operating costs in the first year and the costs of capital; they represent the sunk costs of entrant firms. These individual characteristics of a firm are also important, because when the firm decides on entry it takes into account the level on which it is going to operate. In the case of Exit, ρ is the same as for entry firms, but w and l are operating costs of the previous year and irreversible costs of capital; this represents sunk costs of the exit of firm j. To calculate losses l due to exit, we also include loss of profit because the firm no longer operated. The changes in the number of entering or exiting firms indicate investment (or disinvestment) decisions of firms. According to the empirical model represented in Equations 5-6, we estimate the impact of investment trigger points on entry (5) or exit (6) decisions:

t

i

t

iHi

t TREntry ηγ += ,,1 (5) t

jjLj

t TRExit ηγ += ,,1 (6)

where t

iHTR , is the calculated threshold of a firm i, that entered in time t, and t

jLTR , is the calculated

threshold of a firm j, that exited in time t. Marshallian trigger points ( HW and LW ) are calculated as

shown in Equations 1-2; Real Option trigger points ( HP and LP ) are calculated as shown in Dixit (1989). Additional variables, following Real Option theory, have an impact on trigger points and perception about the profitability of the sector. They are the trend rates of growth of the market price of output µ and its

variance 2σ . tEntry is the number of firms entering in the year t; tExit is the number of firms that were previously

observed to be in operation in the year t. tη - is an error term, a subscript i indicates an entering firm, j indicates an exiting firm, and γ - is the parameter to be estimated. As a possible modification of the model based on Marshallian trigger points, we include ρσµ ,, 2 as additional variables in the Equations 5-6, thereby assuming that these parameters have an impact on the firm’s decision concerning entry/exit, but have no impact on the threshold as assumed by Real Option theory. Econometric model Since the dependent variable in the entry (exit) equation is the number of firms entering (exiting), this can take only nonnegative integer values. A count is understood as the number of times an event occurs. The ordinary least squares (OLS) method for even count data results in biased, inefficient, and inconsistent estimates (Long, 1997). Thus, various nonlinear models that are based on the Poisson distribution were developed for this type of “count data”.


41

The Poisson regression is )(~ ii Poissony µ (7)

)exp( ii x=µ (8)

for observed count iy with covariates for the i-th observation.

The Poisson model assumes that its mean is equal to its variance, which is unlikely in reality. This leads to a problem of overdispersion, i.e. that the observed variance is greater than the mean ( )()var( ii yEy > ). One reason for this is the omission of relevant explanatory variables. Estimates of a

Poisson model for overdispersed data are unbiased, but inefficient with standard errors biased downward (Cameron and Triverdi, 1998; Long, 1997). The most common alternative is the Negative Binomial model, which introduces an individual, unobserved effect into the conditional mean.

)(~ *ii Poissony µ (9)

)exp(*iii ux += βµ (10)

),/1(~ λλGammae iu λ is the overdispersion parameter. The larger α is, the greater the overdispersion. If λ =0 then the model converges to the Poisson model. A more detailed description of the Poisson model and the negative binomial model can be found in Cameron and Triverdi (1998: p. 59), Greene (2003: p. 744). 3. Data This section first gives a description of the data used in estimation and then presents an analysis of calculated trigger points, which are used as independent variables in the model. We combine two data sets: FADN (Farm Accountancy Data Network) and “Meitelling” data1, provided by the LEI. The variables used for estimating thresholds, and the econometric specification of the model are represented in Table 2. “Meitelling” data provide us with information about all firms in the glasshouse horticulture sector during the period 1975-2004. If a firm exited and entered during these time periods then we have the complete record of the “firm’s life”: from “birth” to “death”. Although the coverage of glasshouse horticulture firms is good, the data content is fairly small. Basically, only the land and the numbers of employees are available.

1 Meitelling is the Register of Enterprises and Establishments of agriculture firms in the Netherlands. The register covers all firms with a size equal to or bigger than 2 nge (Dutch Size Units). www.lei.nl


42

Table 2: Descriptive Statistics for Glasshouse Firms, Thresholds and Number of Entry and Exit • Variable • Description of Variable • Mean • Standard

Deviation • • • • • Ha_tot • Land per firm, ha • 2.31 • 0.33 • Ha_glass • Land under glass per firm, ha • 0.62 • 0.11 • Profit_ha • Profit per ha, 1000 Euros* • 59.0 • 17.7 • Cost_mat_ha • Material cost per ha, 1000 Euros* • 234.8 • 44.3 • Lab_tot • Number of workers per firm, annual workers • 3.4 • 5.4 • Cost_lab • Labour cost per annual worker, 1000 Euros* • 20.3 • 0.5 • Inv_ha • Investments per ha, 1000 Euros* • 26.9 • 8.3 • µ • Trend rate of growth of output prices • 0.06 • 0.01

• σ • Standard deviation of output prices • 0.14 • 0.02 • ρ • Interest rate, % • 7.63 • 1.67

• • • • • EntryK • Number of entering firms • • • • K=1 as real entry • 194.4 • 62.1 • • K=2 as entry in horticulture • 767.9 • 143.5 • • • • • ExitK

• Number of exiting firms • • • • K=1 as real exit • 339.0 • 73.6 • • K=2 as exit from horticulture • 278.8 • 89.8 • • • • • WH,K • Marshallian entry threshold, calculated for entering

firm, 1000 euros* • •

• • K=1 as real entry • 437.3 • 153.1 • • K=2 as entry in horticulture • 190.1 • 77.8 • • • • • WL,K • Marshallian exit** threshold, calculated for exiting

firm, 1000 euros* • •

• • K=1 as real exit • -235.6

• 61.1

• • K=2 as exit from horticulture • -66.1 • 23.3 • • • • * Monetary values are normalised by 1985 prices ** Exit thresholds were used for estimation as absolute values for the simplicity of the interpretation of results of the econometric model

The FADN is an unbalanced panel data set, amongst others, on glasshouse horticulture firms during the period 1975-1999. Due to the rotation of firms, firms stay in the sample for an average of 3-5 years. These data provide a wide range of individual characteristics of firms such as revenue, capital, investments, variable costs, which we used for the estimation of the annual level of these variables. For the calculation of the trigger points, we used variables from both data sets; however, due to the time period of FADN data, the further estimation is limited by the period 1975-1999. We distinguish and use for the analysis two different types of entry and exit: 1) the genuine (or real2) entry and exit, 2) the entry and exit by changing specialisation (e.g., when an existing firm starts with horticulture production).

2 We use terms “genuine” and “real” interchangeably for the definition of one of the types of entry or exit


43

The variables represented in Table 2 are used for the calculation of trigger thresholds3. These variables characterise the average glasshouse firm, which earns 59,000 euros profit through the use of 2.3 ha of land (0.6 ha under glass) and employs 3.4 workers per year. The average firm invests 26900 Euros per ha in capital (such as land, glasshouses and installations). The salient characteristic of Dutch glasshouse firms is that they remain small-scale family firms (68.8% of family labour) with respect to labour and land, but they are highly capital-intensive, with an average capital per firm of 383,000 euros (at 1985 price levels). The next step, as an extension of the conventional approach, will be to calculate Real Option trigger points and compare them with Marshallian ones. As can be seen, the investment thresholds (Table 3) vary over the years with the common tendency of growth. The gap between Marshallian and Real option trigger points varies and becomes bigger: if at the beginning of the analysed period the difference for entry was about 5,000 euros and for exit about 2,000 euros, then at the end it had risen to 30,000 and 14,000 euros respectively. Following the discussion in Dixit (1989), the difference between thresholds is caused by uncertainty. So the years with the biggest gap, namely 1981, 1987, 1993, and 1996 possibly exhibit the effect of “hysteresis”, when firms prefer to wait and would need to overcome a higher threshold to make investments (in the case of entry) or disinvestments (in the case of exit). It can be also noted that the difference between entry trigger points is bigger than for exit trigger points; although in both cases the difference between Marshallian and Real Option thresholds is affected in the same years.

Table 3: Marshallian and Real Option trigger points

• Real Entry

• Trigger Points,

• 1000 euros

• Real Exit

• Trigger Points,

• 1000 euros

• Horticulture

• Marshallian

• Trigger Points,

• 1000 euros

• Year

• Marshallian • Real Option • Marshallian • Real Option • Entry • Exit

• • • • • • • • 1976 • 201.6 • 206.4 • na • na • 17.1 • na • 1977 • 222.8 • 228.6 • -117.0 • -119.2 • 91.8 • -48.5 • 1978 • 224.7 • 230.3 • -154.1 • -156.6 • 110.3 • -52.9 • 1979 • 274.1 • 280.9 • -179.8 • -182.9 • 140.4 • -58.4 • 1980 • 431.3 • 441.4 • -243.0 • -247.5 • 178.3 • -68.1 • 1981 • 544.5 • 557.5 • -275.6 • -280.8 • 164.3 • -70.7 • 1982 • 315.8 • 324.0 • -242.0 • -246.8 • 206.6 • -86.3 • 1983 • 344.3 • 354.4 • -243.3 • -248.9 • 175.8 • -87.7 • 1984 • 475.5 • 488.2 • -179.0 • -182.3 • 173.8 • -64.8 • 1985 • 342.6 • 352.5 • -209.7 • -213.7 • 184.6 • -53.7 • 1986 • 358.0 • 369.0 • -251.0 • -255.7 • 181.1 • -41.5 • 1987 • 385.0 • 400.1 • -176.4 • -181.8 • 191.8 • -63.1 • 1988 • 305.0 • 317.0 • -168.1 • -173.2 • 161.8 • -43.8 • 1989 • 366.2 • 380.6 • -207.1 • -213.9 • 235.3 • -69.1 • 1990 • 429.4 • 443.9 • -158.5 • -162.9 • 220.8 • -16.9 • 1991 • 521.9 • 539.7 • -279.3 • -287.4 • 243.4 • -84.6 • 1992 • 555.9 • 575.5 • -354.7 • -365.1 • na • na • 1993 • 666.1 • 696.4 • -284.1 • -295.5 • 312.1 • -43.9 • 1994 • 659.2 • 688.4 • -264.2 • -274.8 • 42.9 • -72.8 • 1995 • 600.0 • 626.5 • -254.8 • -265.1 • 241.3 • -65.3 • 1996 • 762.1 • 797.7 • -344.2 • -358.7 • 284.2 • -90.3 • 1997 • 388.7 • 407.8 • -252.2 • -263.3 • 196.1 • -60.7

3 A detailed description of the calculation of trigger points by combining of two data sets can be provided upon request


44

• 1998 • 590.2 • 621.8 • -292.8 • -306.4 • 310.9 • -134.5 • 1999 • 529.7 • 558.1 • -286.9 • -300.4 • 306.9 • -76.8 • • • • • • • • Total • 437.3 • 453.6 • -235.6 • -242.7 • 190.1 • -66.1

- Trigger points represent the annual average level - na – not possible to calculate due to the absence of reliable data on horticulture entry/exit An existing firm that enters (exits) glasshouse horticulture has to overcome lower impediments compared to the real entry (exit). This is demonstrated by the difference in the investment trigger points: an existing firm that enters the horticulture sector should invest (on average, over the years) 190.1 thousand euros, but for a real entry a firm should invest almost twice as much, on average 437.3 thousand euros. For the real exit, a firm should overcome (on average) losses of 235.6 thousand euros, which is three times the threshold for the exit from the horticulture sector (loss of 66.1 thousand Euros). Results Of Estimation Econometric Models The change in the level of trigger points can encourage or discourage exit and entry into glasshouse horticulture, as is shown in Tables 4-5. These tables give the negative binomial estimation results for entry and exit. The results lend support to the negative binomial model, since the λ parameter is significantly different from zero. This is confirmed by the Likelihood-ratio test. The significance of overdispersion parameter λ confirms the presence of an individual, unobserved effect that means non constant mean and variance in the data. By this fact, the outperforming level of Log-Likelihood for Negative binomial regression over the Poisson model can be explained. The exit barriers were included in the model as the positive values for the purpose of easier interpretation. The difference among models is in the explanatory variables: Model 1 includes Marshallian trigger points, Model 2 includes Real Option trigger points, which are corrected for the effect of expectation of prices, uncertainty, and interest rate; and Model 3 explicitly incorporates the expectation of prices, uncertainty and interest rate in Model 1, that deviates from the specification of Dixit (1992). Based on Pseudo R2, it can be concluded that the Model 3 provides the best explanation of the variation of entry and exit out of three specifiations. As can be seen from the estimation results, a higher level of entry thresholds has a negative impact on the number of firms that decide to enter. Increasing exit thresholds deters firms from exiting the sector. In agreement with the theory, positive expectations about the trend of output prices induce more firms to enter and fewer firms to cease operation.


45

Table 4: Effect of Trigger Points on Real Entry and Exit

• Real Entry • Real Exit

• Model 1 • Model 2 • Model 3 • Model 1 • Model 2 • Model 3

• Variable

• Trigger

point WH,1

• Trigger point PH,1

• Trigger

point WH,1

• Trigger point WL,1

• Trigger point PL,1

• Trigger

point WL,1

• Dependent variable:

• Entry1 • Exit1

• Independent variables:

• • • • • •

• TR • -0.002*** • (0.0004)

• -0.002*** (0.0004)

• -0.001*** • (0.0006)

• -0.001* • (0.0006)

• -0.001* • (0.0006)

• -0.002***

• (0.001) • µ • • • 12.269*

• (6.776) • • • -

19.020***

• (3.785) • σ • • • 1.459

• (5.087) • • • -

7.300*** • (2.554)

• ρ • • • 0.095** • (0.046)

• • • 0.012 • (0.023)

• Constant • 5.372***

• (0.203) • 5.371***

• (0.203) • 3.402***

• (1.142) • 5.253***

• (0.154) • 5.245***

• (0.152) • 7.405*** • (0.629)

• λ • 0.093 • (0.028)

• 0.091 • (0.027)

• 0.057 • (0.018)

• 0.034*** • (0.011)

• 0.034*** • (0.011)

• 0.014 • (0.357)

• Likelihood-ratio test of

λ = 0:

Chi2(01)

•

• 334.79***

•

• 324.01***

•

• 183.13***

•

• 198.40***

•

• 199.20***

•

• 70.52***

• Log likelihood:

• • • • • • •

• - Poisson model

• -299.12 • -293.37 • -217.81 • -227.44 • -227.87 • -154.24

• - Negative binomial regression

• -131.72 • -131.37 • -126.24 • -128.24 • -128.27 • -118.98

• Pseudo R2 • 0.06 • 0.06 • 0.10 • 0.01 • 0.01 • 0.08

• N • 24 • 24 • 24 • 23 • 23 • 23 1) t-statistics in parentheses 2) *** denotes coefficient significant at 1% level, ** at 5% and * at 10% level

Higher interest rate, which is an indicator of the profitability of a sector, has a positive connection on entry, and a negative one for exit (except a real exit, which is not significant). Uncertainty (σ ) has a positive (and not significant) result for real entry, but a negative one for entry into horticulture.


46

Table 5: Effect of Trigger Points on Entry into and Exit from Horticulture

• Entry into Horticulture • Exit from Horticulture

• Model 1 • Model 2 • Model 3 • Model 1 • Model 2 • Model 3

• Variable

• Trigger

point WH,2

• Trigger point PH,2

• Trigger

point WH,2


• Trigger point PL,2



• Entry2 • Exit2

• Independent variables:

• • • • • •

• - TR • -0.002*** • (0.0007)

• -0.002*** (0.0006)

• -0.001 • (0.001)

• -0.007*** • (0.003)

• -0.007* • (0.003)

• -0.007*** • (0.002)

• - µ • • • -5.950 • (4.682)

• • • -17.941***

• (7.048) • - σ • • • -3.115

• (3.958) • • • -

15.797*** • (4.265)

• - ρ • • • 0.049* • (0.028)

• • • -0.066* • (0.039)

• Constant • 7.639***

• (0.148) • 7.626***

• (0.144) • 6.433***

• (0.813) • 6.600***

• (0.210) • 6.597***

• (0.205) • 9.070*** • (1.186)

• λ • 0.039***

• (0.012)

• 0.039***

• (0.012) • 0.025 • (0.008)

• 0.093***

• (0.030)

• 0.092***

• (0.030) • 0.042 • (0.014)

• Likelihood-ration test

of λ = 0:

Chi2(01)

• 577.62*** • 574.98*** • 347.79*** • 487.20*** • 483.66*** • 210.08***

• Log likelihood:

• • • • • •

• - Poisson model

• -424.10 • -422.76 • -304.86 • -360.69 • -358.84 • -214.43

• - Negative binomial regression

• -135.29 • -135.27 • -130.97 • -117.09 • -117.00 • -109.38

• Pseudo R2 • 0.03 • 0.04 • 0.05 • 0.02 • 0.02 • 0.07

• N • 21 • 21 • 21 • 20 • 20 • 20

This can be explained by the statement of Wennberg et al. (2007) that the negative effect of uncertainty on the likelihood of entry will turn positive at a high level of uncertainty for real entry but not for the entry of existing firms. Therefore the results can be understood as an indication of higher uncertainty for the real entry, compared to the entry into horticulture. The higher variation of input prices deters firms from exits; this effect is larger for exiting due to a change in specialisation. This means that firms prefer to delay the decision to exit, because of expectations of positive changes in prices. The presence of investment thresholds predetermines a certain number of firms that are able to overcome these thresholds and that decide to invest and enter (or to disinvest and exit). Changes in investment thresholds affect firms and change their behaviour in such a way that an additional number of firms will enter or exit. This effect of changes in trigger points can be demonstrated by analysing elasticises (Table 6).


47

Table 6: Elasticities for trigger points after Negative Binomial Estimation (Model 34)

• • Real Entry

• Real Exit

• Entry in Horticulture

• Exit from Horticulture


• En1 • Ex1 • En2 • Ex2

• Independent variable:

• TRH,1 • TRL,1 • TRH,2 • TRL,2

• - trigger point W • -0.270 • (0.11)

• -0.530

• (0.18)

• -0.733 • (0.82)

• -1.977 • (0.64)

The establishment of a new firm can be expected if the real entry threshold decreases by 3,700 euros. The real exit investment threshold should decrease by 1,900 Euros to induce an additional firm to cease trading. The difference in elasticises demonstrates the fact that existing firms respond more to changes in trigger points, because it is easier for these firms to overcome investment barriers. The changes in entry barriers should be bigger than for exit barriers to have an impact on a firm’s decision as can be seen from smaller values of elasticities for entry compared to exit thresholds. Another observation from the table is that the existing firms that enter or exit the horticulture sector are more sensitive to the changes in investment thresholds. It can be expected that with a 2,700 Euro decrease in the horticulture investment threshold (TRH,2), two more firms will enter the horticulture sector, while to encourage the establishment of the two additional firms the threshold (TRH,1) should decrease by 7,400 Euros. The same holds true for the exit: we can expect the exit from the horticulture sector of the two additional firms if the investment threshold (TRL,2) is bigger in absolute value by an amount of 1,000 euros; but for real exit TRL,1 should change by 3,800 euros.

Table 7: Predicted and Actual mean of Number of Entry and Exit firms

• • Real Entry

• Real Exit

• Entry into Horticulture

• Exit from Horticulture

• Number of Entry or Exit:

• • • •

• - actual • 194.4 • (62.1)

• 339.0 • (73.6)

• 767.9 • (143.5)

• 278.8 • (89.8)

• - predicted by: • • • • • Model 1 • 197.6

• (46.1) • 339.6 • (37.5)

• 803.7 • (133.5)

• 289.4 • (56.7)

• Model 2 • 197.4 • (46.4)

• 339.5 • (36.7)

• 802.5 • (129.8)

• 289.1 • (55.8)

• Model 3 • 194.8 • (46.9)

• 339.0 • (57.5)

• 785.1 • (86.7)

• 277.7 • (51.3)

By analysing the Table 7, we can compare how close the prediction can be compared to the actual average of events. It can be seen that real entry and exit events have closer predicted values than

4 Model 3 is represented in Table 6, because, as is shown in Tables 4,5 and 7, Model 3 outperforms other specifications


48

horticulture entry and exit. This can be related to the slower reaction to changes in investment thresholds, as discussed above. As a comment to the discussion about the real option approach, we can see that the use of RO trigger points only slightly improves the prediction of entry and exit, while assuming that characteristics of the sector influence the firm’s decision instead of changing trigger points (Model 3) gives the most accurate prediction. The preference for Model 3 can be also supported by the differences in values of Log-likelihood and Pseudo R2 provided in Tables 4-5. Discussion We have examined empirically the entry-exit process in Dutch glasshouse horticulture as an investment decision of a firm that should overcome an investment threshold. This chapter has demonstrated that investment trigger barriers have an impact on a firm’s decision to invest and enter, or to disinvest and exit. An increase in the barriers discourages firms from taking any action; they prefer to delay the decision, which is associated with irreversible investments. The models that include Marshallian and Real Option trigger points were compared. The explicitly calculated investment thresholds provide insights into the barriers that a firm should overcome and shows the increase of competition in the sector, partially due to the use of capital-intensive technology in glasshouse horticulture. We distinguished two types: real (or genuine) entry-exit; glasshouse horticulture sector entry-exit. The heterogeneity of entry and exit investments has two consequences. First, firms will overcome different thresholds that can induce or deter firms from entry or exit. Second, the change in thresholds results in a different number of entering or exiting firms, e.g. existing firms whose specialization changes, resulting in them entering horticulture are more sensitive to the change in investment thresholds compared to firms, which potentially can enter the sector and which are considering establishing a new business. The difference in degree of irreversibility of the different types of entry and exit can be one of the reasons for this. The impact of thresholds can be a confirmation of the effect of irreversibility on an investment decision: if a threshold (as a sum of operational and fixed costs) is possible to be reversed, a firm will not take it into account. The empirical results do not provide reasonably strong support to real option theory, while the model that suggests the direct impact of the sector-characterizing variables, such as expectation of output prices, uncertainty and interest rate, explains entry-exit decision better. The effect of these variables is larger for the real entry and exit compared to the change in specialization entry-exit. Moreover, uncertainty has a negative impact on exit and entry into horticulture, but turns out to be positive for the real entry. One of the possible suggestions, which can be further explored in future research, is that for a higher level of uncertainty, the negative effect of uncertainty on the likelihood of entry can turn positive. Further research can be conducted on deepening the knowledge of the individual firm’s decision for entry and exit which differentiates the heterogeneity of entry and exit. Thus it can have an important impact on the length of survival of firms, and on their post-entry performance. The entry-exit investments associated with changes in management or ownership of a firm (classified as “transferred entry-exit”) needs further investigation and assumes acquiring additional (qualitative) information. References Cameron, A.C. and Triverdi, P.K. (1998). Regression Analysis of Count Data. Cambridge: Cambridge

University Press.


49

Dixit, A. (1989). Entry and Exit Decisions under Uncertainty. The Journal of Political Economy. No 97(3): 620-638

. Dixit, A. (1992). Investment and Hysteresis. Journal of Economic Perspectives. No 6(1): 107-132. Dixit, A.K. and Pindyck, R.S. (1994). Investment under Uncertainty. Princeton: Princeton University

Press. Greene, W.H. (2003). Econometric Analysis. Upper Saddle River, New Jersey: Prentice Hall. Long, J.S. (1997). Regression Models for Categorical and Limited Dependent Variables. Advanced

Quantitative Techniques in the Social Sciences. Sage Publications. Marshall, A. (1920). Principles of Economics. Macmillan, New York. Mas-Colell, A., Whinston, M.D. and Green, J.R. (1995). Microeconomic Theory. Oxford University

Press, Oxford. OECD (2003). The Sources of Economic Growth in OECD Countries. Paris, OECD. Siegfried, J.J. and Evans, L.B. (1994). Empirical Studies of Entry and Exit: A Survay of the Evidence.

Review of Industrial Organization. No 9: 121-155. Wennberg, K., Folta, T.B. and Delamr, F. (2007). Real Option Model of Stepwise Entry into Self-

Employment. Proceedings of the conference: Entrepreneurship Research Conference, University of Maryland.


50

POLICY ASSESSMENT AND DEVELOPMENT BY STAKEHOLDERS: A CROSS-COUNTRY ANALYSIS OF NATIONAL RECOMMENDATION ON ORGANIC

FARMING POLICY IN 11 EUROPEAN COUNTRIES

Anna Maria Häring University of Applied Sciences Eberswalde, Eberswalde, Germany


Daniela Vairo

DIIGA – Polytechnic Univeristy of Marche, Ancona, Italy

Stephan Dabbert, 3 University of Hohenheim, Stuttgart, Germany

Raffaele Zanoli

DIIGA – Polytechnic Univeristy of Marche, Ancona, Italy Abstract

There is no single 'best way' of policy development. Bottom-up approaches to policy design and a broad debate among stakeholders facilitate policy learning and innovation. A novel approach of a bottom-up policy design process involving stakeholders is introduced. First results obtained by this methodology are presented. The outcomes of an international effort for a development of policies for organic food and farming in Mai 2004 in Europe are analyzed: the synthesized results from 11 European countries (AT, CH, CZ, DE, DK, EE, GB, HU, IT, PL, SI) on the current situation of policies related to the organic food and farming sector in Europe are highlighted and policy recommendations for the development of the sector formulated. Specifically, strengths, weaknesses, opportunities and threats of policies related to organic food market are identified and policy instruments to address these aspects are developed.

Keywords: Multi-stakeholder process, policy learning, policy transfer, organic farming policy

Introduction Organic Farming has become an inherent part of European agriculture in the EU. The first policy involvement in the organic farming was the EU-wide harmonisation of the definition of organic farming by Council Regulation (EC) 2092/91 as to ensure market transparency and consumer protection. Government support, largely made under Council Regulations (EC) 2078/92 and 1257/99 based on the organic farming definition of Council Regulations (EC) 2092/91 and 1804/99, has played a significant role in stimulating an increase in organically managed farms and land area. These policies were developed by agricultural policy makers legitimated by democratic processes or institutional background, e.g. national “consultative groups” for the implementation of the agri-environmental measures within the the Accompanying Measures and the Rural Development Programmes (Council Regulations (EC) 2078/92, 1257/99). Representatives of organic farming assocications or informal groups were involved – if at all – deliberatetly through informal communication with members of these consultative groups. In part this was due to the origin and development of the organic farming sector as a private sector social movement. Organic farming organsiations were considered with the principles of organic farming and their justification, and not so much worried about lobbying for policy support (DABBERT 2001). In addition, due to its’ relatively small size, the organic farming sector did not produce a very strong lobbying power (finances and personpower). This resulted in a very low intensity of lobbying by organic farming organisations in most Member States (DABBERT ET AL. 2004) despite


51

support by environmental organisations. At the EU level, not until October 2003 the IFOAM established a permanent office for lobbying activities close to the European Commission. For the first time, stakeholders were consulted on organic farming issues in a conference on organic farming in 1999 in Baden (Austria) under the lead of the Austrian government. This consultation was continued in a similar conference at Copenhagen, Denmark in 2001. Although both conferences were not formal policy consultation processes, they were organised as to provide input to the policy development at the EU level. However, these consultations followed a top-down approach. Goals and topics to be addressed as well as invited stakeholders were defined by the organisers. Since 2001, the European Commission follows principles of good governance (EC 2001). Governance refers to the process of decision-making and the process by which decisions are implemented. This includes the mechanisms, processes and institutions through wich citizens and groups articulate their interest, exercise their legal rights, meet their obligations and mediate their differences. The objective of the EC is to more strongly involve citicens in legislative processes, i.e. by participation, and to speed up the adoption of a common policy framework in all European Member States. One of the five principles of good governance is participation in the formulation of policies and their implementation The first EU wide effort of stakeholder participation in the development of policies concerning organic farming was the 'European Hearing on Organic Food and Farming - Towards a European Action Plan' in Brussels in 2004 (EC 2004), followed by an online-consultation. The main purpose of this hearing was to listen to the views of the widest possible range of stakeholders of the agricultural, environmental and consumer field (EC 2004). Over a 100 stakeholder organisations, Agricultural Ministers from Member States, Acceeding and Candidate Countries participated in this conference. As a result, the Commission prepared an Action Plan in the form of a Communication to the European Council and Parliament, including a list of possible actions to boost organic farming. Again this hearing was organised top-down, only allowing participation of certain, invited stakeholders. The resulting European Action Plan for Organic Food and Farming was never accompanied by specific policy measures or budget for specific policy goals, but left it to the member States to come up with policy measures to address organic farming within their Rural Development Programmes. Nevertheless, the Action Plan Document provided justifcations for a range of measures and a list of ideas for national implementation. By today all Member States have opted to address organic farming by specific support measures (HÄRING ET AL. 2004). In the Member States only in some cases, a formalised involvement of stakeholders in organic farming policy development has been intiatied from the legitimated bodies of governance (e.g. Germany, UK). Bottom-up approaches to policy design with a broad debate among stakeholders can contribute to an increased understanding of policy practices and their impact. There is no single 'best way' of policy development. However, to design policies or to assess the transferability of "good practices" from one country to another it is essential to understand the specific national environments, policy practices and their impact. The objective of this research was to contribute to the development of organic food and farming policy in Europe by assessing existing agricultural policies and their impact on the organic food and farming sector together with the most important stakeholders of the organic farming sector in the European Union. This contribution presents a methodological approach of stakeholder involvement designed as to contribute to a scientifically based formulation of policy recommendations, and the results from a large international effort which has applied this methodology in order to develop policies supporting the development of the organic food sector at the Member State (MS) and EU level (HÄRING et al., 2005).


52

Methodology Bottom-up approaches to policy design require multi-stakeholder involvement in order to achieve policy learning by collaborative working and the creation of networks. Multi stakeholder processes intend to bring together all major stakeholders to participate in a new form of communication, decision-finding (and possibly decision-making) on a particular issue (HEMMATI, 2002). Mutual collaboration of stakeholders with different experiences and competences are considered an enrichment opportunity for the policy design process. Action research or interactive social research approaches, based on the interaction between social subjects (TODHUNTER, 2001), and collaborative policy learning procedures (DOLOWITZ, MARSH, 2000, ROSE, 1991) generally are promising to stimulate stakeholders to co-produce knowledge. The collaboration inside a group is considered one of the more favorable moments of learning, as collaboration implies synergy, a common effort to the realization of a particular objective. Collaborative working or learning favors the development of a critical thought; it increases the abilities to problem solving and contributes to the development of cognitive abilities (DE KERCKHOVE, 2004). Policy learning and policy transfer strongly depend on knowledge and spread of information (DE

KERCKHOVE, 2004, ROSE, 1991). Policy transfer can take place across time, within countries and across countries. For the example of agricultural policy, all Member States (MS) may benefit from learning from other MS how to best develop and implement policies supporting organic farming, e.g. the New from the Old Member States of the European Union. However, even if ‘trans-national policy learning’ is facilitated, the countries involved in the enlargement process need to verify if all conditions to transfer crucial elements of what made the policy or institutional structure a success in the originating countries. Thus, the creation, management and transfer of knowledge are crucial. In the present case the aim was to assess existing agricultural policies and their impact on the organic food and farming sector, by identifying relevant policies in other Member States which can be transferred through emulation, adaptation or simply more or less coercive acquisition (EVANS, DAVIES, 1999). A structured form of participation of and consultation with policy stakeholders was developed to contribute to a scientifically based formulation of policy recommendations at the national and EU level (HÄRING et al., 2004B). Stakeholder involvement is achieved through two national and one EU level workshop which were managed as to facilitate policy learning among stakeholders of a country and across countries. 1) At the national level, there is an opportunity to facilitate policy learning among stakeholders of a

country, to create a national network, and to create agreement able to produce future actions. 2) At the trans-national level, there is an opportunity for the MS to learn from each other (e.g. New and

Old MS), to create transnational networks, and to reduce the differences in national policies and policy innovation.

3) A link between national and transnational stakeholder networks and the EU commission can be created as these workshops are an EU-wide “experiment” in developing organic farming policy recommendations.

The developed bottom-up approach to policy design may result in policy transfer: knowledge and information generated and transferred by these workshops favor the establishment of national networks and the consolidation of international consensus. National and trans-national networks potentially created may facilitate participant’s building of alliances and developing a common language. With the active participation and involvement of stakeholders, these networks have the potential to influence decision-makers in the policy implementation. Thus participants were chosen cautiously as to represent a good representation of stakeholder perspectives: participants from four groups were involved in the process: policy makers, organic sector representatives, non-organic sector representatives and third parties.


53

In April 2004 the first series of national workshops was conducted in 11 European countries (AT, DE, DK, CH, CZ, EE, HU, IT, PL, SI, UK) according to common guidelines (HÄRING, VAIRO, 2004). The objective of these workshops was to assess the effectiveness of different policy instruments in each country, and to develop suggestions for ‘future’ policy instruments to positively influence the development of the organic farming sector in the respective country (HÄRING, VAIRO, 2004). The workshop group discussion was structured in 3 phases: 1) Definition of SWOT: The analysis of organic farming policy was based on the methodological

approach of SWOT analysis. On the one hand, participants analyzed their country’s specific policy instruments’ strengths and weaknesses. On the other hand, looking at the external (uncontrollable) environment of the organic farming sector, participants identified those areas that pose opportunities for organic farming in their own country, and those that pose threats or obstacles to its performance.

2) WOT rating: Participants assessed which weaknesses were most relevant in the organic farming policies of their country (criteria: high impact and high importance), which opportunities could be exploited for Organic Farming in their country (criteria: high attractiveness and high probability) and which were the threats from which the sector needs to defend itself (criteria: high seriousness and high probability).

3) Identification of policy instruments: Participants were asked to elaborate possible policy instruments to address weaknesses, opportunities and threats through a brainstorming. This lead to a list of recommendations for national policy makers and provided the basis for the discussion of a EU policy frame-work for organic farming during an EU level workshop in February 2005 (VAIRO et al., 2005).

A large number of strengths and weaknesses of organic farming policy related to the organic food market and opportunities and threats for the organic food sector where identified by the 11 national workshop groups. Results from all 11 countries’ workshop groups were analyzed by iterative coding as to achieve a cross national analysis with the objective to identify the most relevant WOT concepts and policy instruments (HÄRING et al., 2005). To structure these codes further, groups of codes were summarized under headings which are used to present the information in the following. The separation into strengths, weaknesses, opportunities and threats, rating of WOT and design of policy instruments for the identified WOT was applied mainly to provide a common framework for discussion in the 11 involved European countries. Thus, the final step of the synthesizing analysis ignored this methodological separation and grouped the obtained information according to topics. The presented results are the synthesized assessment of policy instruments by stakeholders of very different professional backgrounds and cultural settings. Results neither represent a group consensus nor conclusions of the synthesis of the whole series of workshops. Results Organic Farming has become an inherent part of European agriculture in the Old and New EU Member States (MS). EU enlargement has combined two very different patterns of organic farming development under one market and policy framework. Specific policy support for organic farming has been developed in all MS and a range of measures supporting organic farming exist (LAMPKIN et al., 1999, HÄRING et al. 2004, PRAZAN et al., 2004). As part of the most recent reform of the Common Agricultural Policy (CAP), the CAP Reform 2003 MS have the chance to revise their Rural Development Programs, within which policies for organic farming are implemented, by mid 2006. The introduced first series of national workshops resulted in an assessment of the current situation of organic farming policy in Europe and has provided policy recommendations for the development of organic farming. These results can provide a valuable input on how to consider organic farming and food in the revision process of the Rural Development Plans.


54

The External Environment of the Organic Farming Sector

The environment for organic farming is characterized by two important aspects. On the one hand, natural conditions are considered favorable for the conversion of existing agricultural production systems to organic production methods, despite the less favorable farming structure in terms of efficiency and organization of farms in some countries. On the other hand, rising wealth and the level of education in the enlarged EU have created societal trends such as concerns about the environment, health, wellness and food quality, creating demand for organic products. Policy Design Issues for the Development of the Organic Farming Sector

In several countries an opportunity for the development of the organic farming sector is seen in an increasingly favorable political climate in the future. For example, the most recent reform of the Common Agricultural Policy has had a positive impact on organic faming. New development opportunities for organic farming were expected from modulation, regionalization and financial resource transfer from the Common Market Organizations to the Rural Development Programs. Nevertheless, an expressed general sympathy of policy makers for organic farming has not yet lead to the implementation of many concrete actions pro organic farming. Public budgets are increasingly tight and decreasing financial support for the agricultural sector also relates to the organic farming sector. Stakeholders demand more political commitment towards the support of organic farming and, consequently, a coherent design of policy measures with clear quantitative targets and concrete actions for their achievement. An efficient implementation of policies and the development of organic farming seem to be the lacking coherence of the existing policy framework with regard to organic farming and a lacking integration of organic farming policy with other policy areas (e.g. rural development, environmental, health and food policy). With regard to policy design, especially an imbalance of support measures for different policy goals was criticized. In some countries, only the agri-environmental measures provide options to support the development of the organic farming sector and other measures implemented within the Rural Development Programs focus too little on the potential integration of the organic sector in other policy areas. Additionally, an inappropriate difference between organic and conventional agri-environmental area payments on the other hand was mentioned. Stakeholders also proposed to improve the financial framework of organic farming by prioritizing environmentally friendly farming systems in the CAP and by prioritizing organic farming in the second pillar of the CAP and nature protection legislation. According to stakeholders, financial funds to finance these efforts could come from non-agricultural sources or from funds for conventional agriculture. An option to efficiently integrate organic farming policy with all agricultural and other policy areas (e.g. nature protection, health policy or tourism) is seen in the development of an Organic Action Plan (OAP). This OAP is to be implemented by a national organic farming committee at the ministry in charge of planning and policy design, supported by an alliance of organic associations which cooperate closely with institutions of other policy areas. National Organic Action Plans should include links to an EU Action Plan and regional Action Plans. This could include options to develop regional projects and the formation of regional organic clusters. Measures relating to general agricultural legislation but with a potentially positive impact for organic farming proposed by stakeholders were stricter nitrogen levels in agriculture. Specific Policy Areas to be Developed in Support of Organic Food and Farming

Financial support to organic farming is still made mainly as area payments within the agri-environmental measures. On the one hand a reduction or abolishment of area payments was proposed in favor of other


55

measures (e.g. market support). On the other hand, in improvement of the design of area payments was proposed in several aspects (difference to conventional or between uses, land types and regions). The current certification system is considered rigid and the required documentation for control authorities complicated hampering the structural development of organic farming and conversion. A simplification and harmonization of standards was demanded to reduce required data collection, to coordinate farm inspections of different control systems, to establish special regulations for small scale production and to introduce IT technology management in the inspection system. All stakeholders should be included in these revisions, linking regional, national and EU level efforts to simplify and harmonize standards. On the one hand, these revisions must focus on conserving the quality differential between organic and conventional farming. On the other hand, the definition of high standards and a robust organic certification system is considered necessary to conserve consumers confidence and avoid scandals in organic farming. A range of measures on how to achieve this were proposed. These constant efforts of improving standards should be communicated to consumers to strengthen the credibility of organic farming. Consumer confidence in organic food quality is considered a very important factor for the future development of organic farming. In the conventional sector scandals and food quality seem to discredit conventionally produced food. Consumers believe in the credibility of organic producers and organic product quality due to its certification and control. Rising consumers’ awareness of healthy nutrition, food quality and the benefits of organic farming increase consumers’ acceptance of organic products. However, in some countries a weak interest and willingness to pay of consumers is still observed due to a high price sensibility of consumers in times of declining economic growth and a high percentage of unemployment. A great opportunity is seen in a better communication with consumers on organic product quality. A better engagement of consumers either directly or indirectly through education and local authorities is expected to increase the demand for organic food by raising consumers’ awareness, eradicating negative attitudes and developing special market segments. For a better communication with consumers a range of elements for public information and promotion campaigns and educational programs were proposed. These efforts should focus on consumers’ expectations and on creating new target groups. As labels are an important element of communicating with consumers a range of elements to improve the transparency of labeling to demonstrate the added value of organic food were developed by workshop groups. According to stakeholders, these efforts on consumer communication should be financed at the EU level but managed by an alliance of organic associations. The contamination with GMO is considered the greatest threat for the organic farming sector. If GMO are registered and certified for conventional production they will contaminate organic production, as coexistence is difficult. However, if GMO residues are found in organic products, trust in organic farming is undermined. Nevertheless, consumers are becoming more interested in organic products as they are afraid of GMO contaminated products. Several measures to avoid the contamination of organic production are proposed. A high competition on markets due to the increased EU, emerging countries, globalization, and the power of large food retailers is perceived a severe threat for the organic sector. To face this situation, stakeholders propose the development of new markets and marketing channels, especially the development of distribution technologies and trade possibilities outside the usual retailers. Stakeholders have identified a lack of support measures for marketing initiatives, especially in New Member States. To improve the market situation stakeholders proposed to: a) increase the cost of conventional production by applying a tax on pesticides, fertilizers and nutrient outputs (internalize external costs); b) reduce the cost of organic products; c) harmonize the comparative costs and quality of organic products from different


56

countries. Furthermore, stakeholders proposed around 20 different options to support the development of organic marketing structures. Capacity building measures in organic farming are considered insufficient, mainly due to perceived insufficient financial resources. Similarly, educational offerings on organic farming in agricultural universities and schools are scarce. Around 10 different policy strategies and measures were proposed to tackle the observed deficits in capacity building. The beneficiaries of these measures should be, apart from farmers, all public sector employees, particularly policy implementers. To encourage participation among farmers, training courses should be – according to stakeholders - free of charge and linked to area support for organic farming. Scientific research and development on organic farming seems to be supported weakly by policy as a core research strategy does not exist. Thus, financial support for research on organic farming does not meet the current needs. Research activities tackling organic farming could be improved by creating a research institute specialized in organic farming, e.g. a governmental research institution, or by emphasizing organic farming in national research funding. A list of topics to be tackled urgently by research was compiled and ranged from research on the comparative advantage of organic farming to scientifically based policy analyses. Workshop participants evaluated the internal organization of the organic sector in two different ways. Some countries considered the networking of organic actors as productive, while other countries still consider their organic sector networking as insufficient, particularly with regard to lobbying. The dialogue of policy makers with organic stakeholders is considered insufficient, especially in two New Member States. Despite the sustained efforts on behalf of non-governmental initiatives to enter in a dialogue with policy-makers, no common institutions have been established to make such joined efforts work and participation in more informal efforts lack participants from the ministries. An improved institutional setting for organic farming was proposed to support the communication of policy makers and organic stakeholders. A productive organic actor network (EU and national) helps to build the sectors capacity to communicate with policy makers. Measures to improve networking at different levels are proposed. Conclusions There is no single “best way” of policy innovation in Europe. However, a broad political debate among stakeholders is essential. A bottom-up approach to stakeholder involvement in agricultural policy design was developed, consisting of a series of three workshops with stakeholders in agricultural policy. The developed series of national workshops were a first step to policy learning, innovation and transfer for the organic farming sector in the EU. Normative approaches to policy design would have obtained very different results. Nevertheless, the presented approach to policy design has provided interesting insight to the necessities of the specific sector and stakeholders viewpoints. A range of policy instruments for the long-term development of organic farming were developed and have spread widely. Results have fed into and provided the base for a discussion at the EU level in a second workshop with EU level stakeholders and representatives from national workshop groups in February 2005 and the second series of national workshops which was conducted in all participating countries in Mai/June 2005. Furthermore, a series of discussion papers outlining policy recommendations on the consideration of organic farming in the design of the national Rural Development Plans (e.g. HÄRING et al., 2005B; SLABE et al., 2005) was disseminated to all participants of all three workshops as well as the most common dissemination channels for the organic farming sector in Europe. Feedback from stakeholders from several countries has demonstrated the high impact the described policy development and transfer process had: in severals countries results have fed into the development of national Rural Development Plans and national strategic documents on


57

organic farming policy, i.e. national Action Plans for Organic Farming. Especially in countries were policy networks are not yet strongly developed and mainly centered on a few stakeholders and organisations (MOSCHITZ AND STOLZE, 2005) such approaches to policy development, learning and transfer were effective and transfer networks were developed. Obviously, normative approaches to policy design would have obtained very different results, however, not taking the principles of good governance, proposed by the EC (2001) into account: openness, participation, accountability, effectiveness and coherence. Each principle is considered important for establishing more democratic governance. The described policy development process contributed primarly to the principle of participation “throughout the policy chain – from conception to implementation”. Acknowledgement Research funding provided through the project ‘Coherence between Ireland’s Official Development Cooperation Activities and other Policy Areas in Particular Agricultural Trade and Support Policies’ funded by the Advisory Board for Irish Aid is gratefully acknowledged. See www.tcd.ie/iiis/policycoherence for further details. References Dabbert, S. (2001): Der Öko-Landbau als Objekt der Politik. In: Reents: Beiträge zur 6.

Wissenschaftstagung zum Ökologischen Landbau, Freising-Weihenstephan (6.-8.3.01), S. 39 – 43. Dabbert, S., A. M. Häring, R. Zanoli (2004): Organic Farming: Policy and Prospects. Zed Books. ISBN:

1-842773-26-7. De Kerckhove, D. (2004): Lessons on collaborative working. Polytechnic University of Marche

http://www.del.univpm.it/LO_zucchermaglio/LO4/pagina1b.htm Dolowitz, D., Marsh., D. (2000): Learning from Abroad: the role of Policy transfer in contemporary

policy-making, Governance 13 (1), 5-24. EC (2001): Weissbuch Europäisches Regieren. Evans, M., Davies, J. (1999): Understanding Policy Transfer: a multi-level, multi-disciplinary

perspective, Public Administration, 77 (2): 361-85. Häring, A. M., Dabbert, S., Aurbacher, J., Bichler, B., Eichert, C., Gambelli, D., Lampkin, N.,

Offermann, F, Olmos, S., Tuson, J., Zanoli, R. (2004): Organic farming and measures of European agricultural policy. Organic Farming in Europe: Economics and Policy, 11.

Häring, A. M., Vairo, D. (2004): Multi-stakeholder integration in the Identification of a new Organic

Farming Policy: Workshop manual 1. EU-CEEOFP. Häring, A. M., Vairo, D., Zanoli, R., Dabbert, S. (2005a): Assessment of policies and development of

policy recommendations for organic farming: a cross-country synthesis of national policy workshops in 11 European countries. EU-CEEOFP project report.


58

Häring, A.M., Stolze, M., Zanoli, R., Vairo, D., Dabbert, S. (2005b): The potential of the new EU Rural Development Programme in supporting on Organic Farming. Discussion paper. EU-CEE-OFP.

Hemmati, M. (2002): Multi-Stakeholder Processes for Governance and Sustainability – Beyond deadlock

and conflict. Earthscan. Lampkin, N., Foster, C., Padel, S., Midmore, P. (1999): The policy and regulatory environment for

organic farming in Europe. Organic Farming in Europe: Economics and Policy, 1. Moschitz, H., Stolze, M., Michelsen, J. (2004): The development of political institutions involved in

policy elaborations in organic farming for selected European states. EU-CEEOFP project report. Moschitz, H.; M. Stolze (2005): Institutional dimensions of the elaboration of organic farming relevant

policies at the European Union. Organic Farming in Europe: Economics and Policy, Vol. 12. Prazan, J., Koutna, K., Skorpikova, A. (2004): Development of Organic Farming and the Policy

Environment in Central and Eastern European Accession States, 1997-2002. Report EU-CEEOFP. Rose, R. (1991): What is lesson-drawing? Journal of Public Policy 11 (1): 3-30. Slabe, A., Häring, A.M., Hrabalova, A.. (2006): Specific needs of organic farming sectors in the new EU

Member States and Candidate Countries to be addressed by the Rural Development Programmes 2007-2013. Discussion paper. EU-CEE-OFP.

Todhunter, C. (2001): Undertaking Action Research: Negotiating the Road Ahead, Social Research

Update, 34. Vairo, D., Häring, A. M, Zanoli, R., Dabbert, S. (2005): Multi-stakeholder integration in the identification

of a new Organic Farming Policy: Concept & Outline of the EU workshop. EU-CEEOFP project report.


59

THE USE OF TECHNOLOGY ASSESSMENT (TA) IN THE FOOD-CHAIN FROM “FARM TO FORK

Anders Larsen, Morten Gylling, Søren Marcus Pedersen

Institute of Food and Resource Economics, KVL, Denmark Email: [email protected]

Abstract The introduction of GMO crops in the EU has been and is a matter of severe controversy. First and foremost consumers, often represented by NGO’s, have been known to be extremely negative towards the introduction. With the EU legislation in place allowing import of and tillage with GMO it is only a matter of time before farmers and food manufacturing companies are forced to relate their strategies to this new technology. This paper focuses on Technology Assessment (TA) in a food chain perspective. So far Technology Assessments have mainly been applied for targeted cases with partial stakeholders such as farmers or consumers but seldom in a holistic chain perspective from “farm to fork”. Trough the food chain a main focus for enterprises and farmers are cost verses benefits but also assessment of opportunities for competitive advantage and outright threats will be of major importance in relation to this new GMO technology. Also the managers of the 21st century will need to be more aware of, and sensitive to, the social concerns related to developing and implementing gene technology within food products. If these matters are not addressed by decision makers GMO’s in the EU is likely to fail as a commodity. By making an overall Technology Assessment, enterprises stand to make better informed choices regarding their optimal actions to secure their own future and gain a competitive advantage. If, for instance, the enterprise Monsanto had conducted holistic Technology Assessments in a chain perspective prior to a market introduction of GMO crops, it might have faired better. In matters concerning GMO’s, failure to incorporate such aspects as social acceptance and politics can at best be expensive for the enterprise or individual farmer - in worst case it can be detrimental to the very existence of a company. The outcome of this paper is an overview of suitable methods that has been or in the future can be used for Technology Assessment in a food chain perspective. By combining the most suitable methods the concept is moved towards a dedicated common framework for Technology Assessment in food chains from “farm to fork”. This framework includes social, ethical and political issues and can be utilised by farmers and enterprises as a tool in order to quickly and efficiently assess the consequences of actions and relate to food chains in an improved way. By conducting Technology Assessment in a chain perspective which integrates economy with social and political issues we believe that a powerful new and highly useful tool is provided to the farmers and enterprises of tomorrow. Keywords: GMO crops, NGOs, Technology assessment , food chains Introduction The food chain is no longer local or even domestic. Several national and international stakeholders are involved. Decisions made by one stakeholder in the food chain may influence the possibilities and limits for the next stakeholder. The food market has become more globalised with significant competition on a world scale. In the search for new products with lower cost of production or high value added, different types of genetically modified ingredients have been utilised in the food industry - and the development of functional foods with health benefits or improved taste are new product innovations that could be expanded. The development of such products are however slow and consumer acceptance seems to be limited. Developing new food products with low consumer acceptance is expensive for companies and for society


60

at large. The steering of the development needs to a larger extent to focus on the entire food chain and involve not only the economic aspects in each link of the chain, but also the social and ethical consequences. In this paper it is argued that the concept of Technology assessment (TA) conducted in a chain perspective can be used by governments as well as private companies to steer the developments of new biotechnology related food products. The objective of this study was to present different approaches for assessing the technical, ethical and socio-economic implications of biotechnology in a food chain perspective. Technology Assessment as a methodology approach New technology has played a central role in the western world for many decades. However not all technological developments have proved beneficial. The diffusion of the plant pesticide DDT1 was for instance followed by severe and largely unknown negative consequences for the animal habitants - also the development of highly cost efficient animal production systems may cause negative consequences like mad cow diseases and avian bird flue. As a result of the growing awareness of the consequences of new technologies in US, the Office of Technology Assessment (OTA) was established in 1972. The concept of conducting TA was originally conceived as an analytic activity, aimed at providing decision makers with an objective approach to analyse the consequences of a new technology (van Eijndhoven 1997). In the very beginning the main purpose of TA was to act as an early warning for unwanted effects. A common expectation on TA has been that it should reveal future consequences of technologies that otherwise would not have been recognised (Palm & Hansson 2006). In its most basic form TA’s can roughly be grouped into 3 categories;

1. Reactive TA 2. Proactive TA 3. User oriented TA

The reactive TA relates to already known problems or perceived problems with an emerging or already introduced technology. In that case a major contribution for the TA is to present different pathways to eliminate or reduce the negative effects of the technology in society. The proactive TA involves the assessment of technologies that are in its pre-commercial state and not yet fully implemented. In that case the assessment usually focuses on different technologies and pathways that can bring about the predetermined goal in an optimal way. In order to make proactive TA it is necessary to procure an objective for the needs in the future. Finally the user oriented approach narrows the scope of the TA down to a few very specific needs. By focusing on theses needs the technology should then be shaped according to these requirements. An example of such a TA could be to find a vaccine against diabetes. User oriented TAs are usually applied on major specific problems and are very costly to carry out. Technology assessments have usually been conducted when new innovations cause “value dissent” among citizens. Value dissent expresses the sometime strong divergence between official public policies and the perception of the individuals. Such value dissents have been obvious in the case of GM crops but also on several other areas including the use of IT, surveillance, etc. In such cases, it is usually in a states

1 Dichloro-Diphenyl-Trichloroethane. Hihgly effective pesticide that turned out to have significant negative effect on the environment.


61

interest to include the divergence in the public policy making. Otherwise technologies may fail in society. Technology assessments can however be conducted on several levels from governmental over corporate chains and down to single a company. Private companies have a great interest in assessing the introduction of new technologies. In fact, the performance of new technologies in a company may be detrimental to it survival in the marketplace. TA conducted by private companies will mainly focus on economic performance of technologies including risk-assessments, impact assessment, feasibility study and different types of economic modelling. The main difference between TAs conducted by governments or even at an over-national level is the focus on the broader welfare economic issues. In developing new technologies private companies do not have an incentive to include all the costs that may accrue to society because of its new introduction. Out of economic reasons companies can have a strong incentive to carry on with imposing new technologies even though they are unwanted in society. However from the lessons of GM crops companies will have to pay closer attention to public wants before developing new products if chances of failure are to be minimised. Related to technology failure the timing is crucial in TA. The optimal timing for conducting TA is at the point of time where the knowledge about the technology is sufficient enough to discover the negative impacts and before the costs of changing direction is too high. A possibility is to build in TA at the very beginning of the development phases in food products. It has not been possible to develop a single uniform technology assessment methodology (Smith 1990). TA is multidimensional in nature and it may involve several aspects. With the wide scope of possible impacts it is not possible to present one uniform method that can be applied to all new technologies. The kind of TA to be conducted depends on the technology in question and who will be affected by it. The first TA was highly expert oriented meaning the assessment was conducted by scientific experts that would pass judgement on the pros and cons of new technologies. Such assessments are known ass “Classical TA”. For the purpose of this paper details of the various discourses in TA and methods that has been developed will not be discussed. It shall merely be conclude that today TA to a large extent includes the general public. As stated by (Skorupinski 2002) “If TA is expected to come up with answers about how technological options should be handled by society, these cannot be given by scientific experts alone”. Examples of such TA methods include participatory TA, aiming at involving the public in the development phase of new technologies. Lately (Palm & Hansson 2006) suggested the use of Ethical TA to make sure that new technologies were ethically sound. The exact method to be applied varies with the technology in question and at the level it is conducted. Private companies will tend to focus strictly on economic performance. However looking at the developing of new biotechnological products the scope needs to be widened. If ethics and the perception of the public (consumers) are not involved in the developments, new products will fail at large cost. Closer attention will have to be paid to other issues apart from strict economic performance in the development and implementation of new products. TA in a food chain perspective Up till this point, TA has not been widely used in the food industry. In the 80s some TAs were carried out in the food industry. Cronberg (1996) mentions “the good work” and the “the good slaughter house”. Here labour unions in corporation with researchers carried out TA projects to reorganise work and utilize technology for the benefits of the workers (Cronberg 1996). The concept of Participatory TAs was set up in order to get a better feeling with the public opinion. This resulted in several so-called Consensus Conferences that managed to spin off some major reflections on


62

the use of biotechnology. Such conferences revealed for instance that the concept of risk is very different in the minds of the public compared to the scientific experts. Such assessments also played a role in understanding where the use of biotechnology to a larger extent is acceptable. The use of biotechnology to produce medicine is several times more acceptable compared to the use in agriculture. Consumers are simply not willing to accept new technologies that serve only to provide benefits to the industry. A scientific and industrial sector that fails to engage with and take home messages from the public will not be able to guide research and development in a direction that is acceptable to the public; hence it will not be able avoid disputable applications of gene technology (Lassen, Madsen, & Sandoe 2002). In the food industry the concept of LCA (Life Cycle Assessment), which is a method involved in TAs, been used quite frequently. With the major impact food production has on the environment it seems natural to assess the impact of a product over its whole life time. Such LCA has procured interesting results clearly showing what products are presenting the heaviest burden on the environment. Such assessments have been the starting point of a proactive TA where different strategies are tested in order to reduce the environmental impact. Not surprisingly farmers are keen to point out the virtues of biotechnology which has been done on several occasions. One outspoken possibility is to breed animals that can reduce negative agricultural effects on the environment. The most known example of these is the so called EnviropigTM: a pig that has the capacity to digest plant phytate, leading to less phosphate in the manure and thus less environmental pollution (Vajta & Gjerris 2006). In the area of introducing new genetic traits into animals it seems indeed to be only the imagination that limits the range possibilities. The economic potential of such new technologies seams just as high as the consumers’ acceptance seems low. Such commodities are likely to fail because they are unacceptable to most consumers. The point is that it is not sufficient that only parts of the food chain can se the benefit in a new technology. If consumers cannot see the benefit the commodity will fail. It can be argued that in order to develop new successful products, they have to present net benefits for every single stakeholder in the food chain from farm to fork. If the cost for one stakeholder exceeds the benefits, the product should either be dropped or the stakeholder should be compensated by the potential winners (see Kaldor Hicks criteria). Such a concept requires an understanding of the stakeholders and their perceptions of costs and benefits. Stakeholders in the food chain To successfully introduce a new product, every stakeholder has to see a benefit and be motivated to take part of the chain. In agriculture, biotechnology has this far been a “push strategy” where only the first part of the food chain has been presented with benefits in the form of reduced costs. There has been a lack of benefits for the consumers. Therefore the coherence in the food chain is week and developed products are likely to fail. Understanding and assessing the impacts in terms of costs and benefits for each stakeholder is crucial to the development of new products. In the remains of this section every stakeholder in the food chain is presented see figure 1.


63

Figure 1. Simple presentation of the central actors in food supply chain and the framework in which it operates.

The figure presents a simplification of the system in which the agri-food industry operates. Politics and law are the main frame in which the agri-food industry operates. With the significant influence food production has on the environment and rural job creation the food industry plays a special role compared to other industries. Within this framework politicians have to procure legislation that is in concordance with criterias for sustainable development balancing the need for a prosperous growing industry with environmental concerns. In this balancing act, politicians have to pay attention to the public and cannot ignore the opposition against certain products. The moratorium in the EU against the GM crop serves to prove this point. The consumers obviously play a significant role in the food chain. If there is no interest in buying a product the product will fail on all levels. Consumers quite naturally have an interest in the price of a food product. But there is also a growing concern over ethical issues related to food production which has also been very clearly demonstrated with the introduction of GM food. As food production depends on the exploitation of living resources the ethical issues in this industry are not hard to find. Several of the products that have been or are being developed are not in concordance with the ethical standard of many western consumers. Food producers will in the future have to pay close attention to ethical standards and realise that not every technology that is possible should be pursued. New developments have to be in line with the consumers’ wants and needs. Stopping or altering unacceptable products early in the development phase is important not only for ethical reasons but also for economic reasons. Retailers are the first to feel the unhappiness of consumers and they have to pay close attention to consumers wants to stay in business. The retailers are very powerful stakeholders in the food chain. Any new product has to be accepted by them before a product can find its way to the consumer. If they are not willing to spend sufficient time to market or sell a product, it will fail. In developing new products the industry has to take into consideration that the retailers have strict demands to logistics. (Jørgensen 1993) describes how a new organically grown bread product failed as a product because raising time required was too long to fit into the existing food chain. Such problems have to be thought of early in the developing phrase. Similar logistical problems are generally known for fresh food products. Using biotechnology to stop the ripening process seams genius from an industry point of view. A later part of the food chain, the consumers, may not share this view and thus the product will fail.

Politics and law

New Biotechnology

Food products

Farm supply company Farmers

Food industry Retailers

Consumers


64

The Food industry is also experiencing the challenges of globalisation. Competing on labour intensive work is not a viable strategy for many western food companies. This has led to a strong focus on knowledge intensive, high value added products such as GM crops with improved health quality or functional foods etc. These products involve higher profits and are seen as a plausible path in the future to gain business and stay competitive. There has been a strong lobbyism for the introduction of these new products into the EU, so far without much success. The primary agriculture is under increased pressure from several sides. The sector has experienced falling prices on agricultural products and subsidies are reduced in many countries. At the same time agriculture is moving away from a production orientation towards market orientation based upon consumer wants and needs (Boon 2001). With the increased international competition a common strategy to stay competitive is to maximise production and reduce costs. The farmer has seldom an influence on the price, so he will often have to reduce costs to be competitive. For the agricultural sector the use of GM crops seems attractive because it presents an option to produce crops at lower costs. Other products of interest is medical or other non-food products with high value characteristics as a strategy for the future (Bedsted 2005). In the farm supply business there is a close connection between the primary farmer and this industry but not much attention has been paid to the end-users of agricultural products. This has led to unacceptable products that are only lead to benefits for the farm supply and the farmers. Such products are not likely to succeed in the market. Further attention will have to be paid to the remaining part of the food chain if new developments are to succeed. Technology Assessment in a Chain Perspective To be successful in the food industry it is important to be innovative and continuously seek new markets. But being innovative does not guarantee that a new product is successful. A success may depend on whether or not the retailer will give a product sufficient time on the shelves to be noticed by costumers. In other words a producer is highly dependent on the decisions being made downstream in the food chain. A succesful product introduction allows all stakeholders in the whole food chain to make a profit which satisfies the ethical and moral standards of the consumers. Just one weak link in the chain and the product may fail. Many new products never enter the market because it does not fit into the current food supply chain. When developing new products or processes it is increasingly important to have a thorough understanding of the whole chain and the motives of each part of the chain. In many TA’s the process leads up to a conclusion on which one decision maker will have to act. In a chain perspective there is not only one decision maker, every partner or potential partner in the chain makes a decision. This decision has an influence on all the other decisions. This makes the TA both more complicated but also more realistic in the big picture. Global firm decisions depend not only on their own decisions but also the decisions made by other actors in the whole supply chain. By conducting TA in a chain perspective new developments can be steered in a better cost-efficient way. New products can be assessed in each part of the chain and if week links are spotted it can be dealt with quickly. This is a benefit for policy making purpose for the governments as new technologies can be aligned with the wants and needs in society. For private companies such an assessment can be used to stay competitive by having better control of the elements in the food chain. This is a benefit for strategic purposes when designing and introducing new products. Furthermore understanding the consequences of new biotechnological products in a chain perspective may be a good starting point for a public debate on of biotechnology.


65

Conclusion The agrifood market is undergoing globalisation. The food chains are no longer local or domestic in character, they operate worldwide. The globalisation contains positive aspects as well as challenges that have to be overcome. On the positive side, globalisation has created new markets for the industry and cheaper products for the consumers. On the downside the agrifood industry in the western world cannot compete on cost of labour and operates under strict environmental regulations. The industry is forced to develop high value added products in order to stay competitive on the world market. This has led the attention to the use of biotechnology. This area seams to hold great potentials for the industry but the technology needs to be effectively steered because several aspects of the technology is unacceptable to many stakeholders in the food chain from. Any chain is only as strong as its weakest link. In this paper a thorough assessment for costs and benefits in each part of the chain supplemented by an ethical and environmental assessment is suggested. Such and assessment will reveal week or strong spots in the chain upon which responsible decisions can be made. It is also argued that such an assessment can be a beneficial tool in the public debate. Technology Assessment in a chain perspective will not guarantee a successful introduction but it will greatly enhance the chances. References Bedsted, B. 2005, Nye GM-planter - ny debat. slutdokument og ekspertindlæg fra borgerjury om GM-

planter afholdt fra d. 28 april til d. 2. maj 2005 Teknologirådet, Kbh. Boon, A. 2001, Vertical coordination of interdependent innovations in the agri-food industry

Handelshøjskolen i København, Det økonomiske fakultet, København. Cronberg, T. 1996, "European TA-discourses - European TA?", Technological Forecasting and Social

Change, vol. 51, no. 1, pp. 55-64. Jørgensen, S. M. 1993, "Some experiences with proactive technology assessment in the Danish food

sector", Technology and Democracy.The use and impact of technology assessment in Europe., vol. Vol. 2.. - S.487-506.

Lassen, J., Madsen, K. H., & Sandoe, P. 2002, "Ethics and genetic engineering - lessons to be learned

from GM foods", Bioprocess and Biosystems Engineering, vol. 24, no. 5, pp. 263-271. Palm, E. & Hansson, S. O. 2006, "The case for ethical technology assessment (eTA)", Technological

Forecasting and Social Change, vol. 73, no. 5, pp. 543-558. Skorupinski, B. 2002, "Putting precaution to debate - About the Precautionary Principle and participatory

technology assessment", Journal of Agricultural & Environmental Ethics, vol. 15, no. 1, pp. 87-102.

Smith, R. E. H. M. State of the art Technology Assessment in Europe. 1990. The second European

Congress on Technology Assessment, Milan 14-16 November 1990. Ref Type: Generic

Vajta, G. & Gjerris, M. 2006, "Science and technology of farm animal cloning: State of the art", Animal Reproduction Science, vol. 92, no. 3-4, pp. 211-230.

Van Eijndhoven, J. C. M. 1997, "Technology assessment: Product or process?", Technological

Forecasting and Social Change, vol. 54, no. 2-3, pp. 269-286.


66

CHANGING PERCEPTIONS OF THE RISK ENVIRONMENT FACED BY COMMERCIAL SUGARCANE FARMERS IN KWAZULU-NATAL, SOUTH AFRICA

R Mac Nicol,

Masters student, University of KwaZulu-Natal,

Pietermaritzburg, South Africa

GF Ortmann Professor in Agricultural Economics ,

School of Agricultural Sciences and Agribusiness, University of KwaZulu-Natal, Pietermaritzburg, South Africa.

[email protected]

SRD Ferrer Lecturer in Agricultural Economics,

School of Agricultural Sciences and Agribusiness, University of KwaZulu-Natal, Pietermaritzburg, South Africa.

Abstract This study identifies sources of risk that commercial sugarcane farmers in the province of KwaZulu-Natal (KZN), South Africa, presently perceive to pose the greatest threat to the viability of their businesses. Data obtained in 2006 via structured personal interviews of 76 large-scale sugarcane farmers from a stratified random sample of 110 farmers in two separate mill-supply areas of KZN were used to elicit farmers’ perceptions of various sources of risk. The most important risk sources were found to be the threats posed by land reform, minimum wage legislation and the variability of the sugar price, in that order. Land reform and minimum wage legislation did not feature prominently in past studies of KZN farmers during the 1990s. Factor analysis identified additional risk dimensions that exist within the remaining risk sources. Recommendations include that government improve accessibility to information regarding future plans for land and labour policies, and that farmers become more proactive in terms of obtaining information to reduce uncertainty and resultant efficiency barriers. Keywords: commercial sugarcane farmers, land reform, minimum wage legislation Introduction On average, 22 million tons of sugarcane are produced seasonally in 14 mill supply areas of South Africa, by approximately 50,940 growers (SACGA 2006; SASA 2006). Sugarcane contributes approximately 82% of the income from field crops in the province of KwaZulu-Natal (KZN) (STATSSA 2006), with 72% of the crop planted by large-scale growers, 19% by small-scale growers and nine percent by sugar millers (SACGA 2006). Approximately 87% of the gross farming income earned by South African (SA) sugarcane farmers in 2002 was by producers in KZN (STATSSA 2006). SA farmers are faced with many challenges attributable to their uncertain and complex decision making environment. In addition to dealing with the deregulation of domestic agricultural markets in the 1990’s, farmers have also had to adapt to changes such as a dynamic global economic and trade environment and a dynamic local political environment. More specifically, other challenges that SA farmers are continuing to face include land reform, AgriBEE (Agricultural Black Economic Empowerment in Agriculture), new labour legislation and minimum wages, property taxes, skills levies, uncertain water rights, HIV/Aids, a


67

volatile exchange rate, and high transport and communication costs (Ortmann 2005). SA sugarcane farmers also had to deal with a highly variable sugar price in recent years (Illovo Sugar 2006). Between January and March 2006 sugar prices averaged 37.43 US cents per kilogram, 91% higher than in the same period in 2005 (FAO, 2006). Following a rise to almost 44 US cents per kilogram in early 2006, the price declined to about 26 US cents per kilogram by November 2006. SA studies where farm-level data sets were used to identify the perceived importance of multiple risk sources include those by Swanepoel and Ortmann (1993), Bullock et al. (1994), Woodburn et al. (1995), Stockil and Ortmann (1997) and Hardman et al. (2002). These studies identified mainly price and production risks as the most important perceived risk sources, although there was a trend towards the increasing importance of government legislation risks by the late 1990s. This is evident in the study by Stockil and Ortmann (1997) where changing labour laws and land reform policies were found to be the fourth and sixth most important risk sources, respectively. Results of this study are briefly compared to previous studies in South Africa and KZN to analyse farmers’ changing risk perceptions. This study will help to identify those sources of risk that are currently perceived to be the most important by large-scale commercial sugarcane farmers in KZN and aims to use factor analysis to examine the dimensions of these perceived risks. This research will facilitate a better understanding of the risks facing commercial sugarcane producers. Findings could assist policy-makers, consultants, extension officers and financial institutions in designing appropriate risk management products and strategies for this group of farmers. Data Source The sample of producers for this study was drawn from a list of commercial sugarcane farmers from two separate mill-supply areas in KZN, namely the Noodsberg mill-supply area in the KZN Midlands and the Umfolozi mill-supply area on the Zululand Coast. Large-scale operations were defined by the South African Cane Growers’ Association (SACGA) representatives as those responsible for annual sugarcane deliveries exceeding 10,000 tons. Large-scale producers are studied in this research because they account for 72% of the area planted to sugarcane compared to small-scale farmers who account for only 19% of the total area planted. The remaining 9% is planted by sugar millers (Eweg 2005; SACGA 2006). Size economies and higher average education levels of large-scale farmers result in these farmers using a wider range of risk management strategies (Barry 2003). Therefore, large-scale farmers are better suited to the objectives of this study. Furthermore, land reform policies pose risks to mainly large-scale farmers. Budgetary constraints due to the personal interview approach limited the maximum size of the sample to 110 respondents. Interviews consisted of structured questionnaires completed in the presence of the main author. Fifty-five farming operations were randomly selected from complete lists of large-scale growers supplied by SACGA regional managers in each mill-supply area. Four responses from the Zululand region were excluded on the grounds that sugar cane contributed less than 30% to gross farm income (GFI). A total of 76 usable responses (38 from each study area) were obtained (69% response rate). Respondents (principal farm decision-makers) were, on average, 47 years of age and had 22 years of sugarcane growing experience. University degrees were held by 42% of respondents. Average farm size was 417 hectares, with sugarcane contributing 77% to GFI. Sources of risk as perceived by survey respondents Respondents were asked to rate sources of risk for their farm businesses, from a list of 14 potential sources, on a Likert-type scale ranging from one to five – where five and one indicate “highly important” or “not particularly important”, respectively. Mean ratings of risk sources are shown in Appendix 1. Respondents could include additional risk sources that they deemed to be important; however, no additional risk sources were included. Respondents were also asked to rank their top five most important risk sources from the list.


68

The three most important sources of risk as rated by all respondents were land reform, minimum wage labour legislation and crop price variability. These had mean overall ratings on the Likert-type scale of 4.31, 4.14 and 3.68, respectively. The risk sources that were perceived to be the next most important were: changes in input costs (3.56), crop yield variability (3.43), the threat of HIV/AIDS (3.41), changes in the cost of capital items (3.33) and changes in land tax legislation (3.24). Compared to previous SA and KZN studies, these findings confirm that government legislation risks (particularly relating to agrarian reform) have become increasingly important, relative to price and production risks. The remaining sources of risk included in the survey questionnaire (unionisation of labour, variability in interest rates, changing water rights, changing credit availability, farm operator illness or death, and changes in family relationships) received mean overall ratings of less than three, indicating that most respondents regarded them as less than moderately important. Concerns among respondents regarding the land reform process in South Africa had become more pertinent leading up to the time of this survey, considering threats by the SA government to discard the willing seller, willing buyer principle due to the perceived slow pace of land reform (Farmer’s Weekly 2006; Democratic Alliance 2006; Afrol News 2006). Subsequent to the survey, the Restitution of Land Rights Act 22 of 1994 has been changed to allow the Minister of Land Affairs to expropriate land, for the purpose of awarding it to a claimant who is entitled to the restitution of a land right, on behalf of the state without being ordered to do so by the court. Effectively, should negotiations over a new market value for claimed land fail, the government will issue farmers with notices of appropriation allowing a period of 30 days for reconsideration, after which final letters of expropriation will be issued and farmers compensated at government-determined “market values” (Nailana and Gotte 2006). The Sectoral Determination (an amendment to the Basic Conditions of Employment Act 75 of 1997) required farmers to meet new minimum wage requirements from March 2003 (Department of Labour 2006), creating uncertainty and increasing the costs of managing permanent labour (i.e., those who work more than 27 hours per week). Many survey respondents speculated during the interview process that minimum wage legislation could be extended to include casual labour. Considering the relatively high demand for this form of labour in the sugar industry (during planting and harvesting) (SACGA 2006), respondents perceive the potential higher costs involved to pose the second most important threat to their business’ viability. Uncertainties, therefore, may be due to recent changes in land and labour legislation creating expectations that further changes are likely. Overall, 79% and 75% of respondents included land reform and minimum wage legislation, respectively, in their top five list of risk sources most important to their farm businesses. These two risk sources were considered to pose the greatest threat to farm businesses in both survey areas. Compared to findings of previous studies (Swanepoel and Ortmann (1993); Bullock et al. (1994); Woodburn et al. (1995); Stockil and Ortmann (1997); Hardman et al. (2002)), these risk sources have become more prominent. Crop price variability was included in the top five list by 45% of all respondents. This may be explained by the high degree of fluctuation of the sugar price during the time leading up to the survey. Product price variability was previously found to be among the three most important perceived risk sources by Bullock et al. (1994) and Woodburn et al. (1995). Changes in input costs (53%) and crop price variability (45%) were, respectively, the fourth and fifth most likely risk sources to be included in the top five list. Compared to Midlands respondents, double the number of respondents from Zululand (67%) included changes in input costs as one of the five most important risks faced by their farm businesses, whereas more than double the number of respondents from the Midlands (47%) included the risk of unionisation of labour in their top five. This is most likely due to respondents in the Midlands region facing threats of labour union strike action shortly prior to the interview process.


69

Factor Analysis of Risk Sources All 14 sources of risk initially considered were included in a factor analysis incorporating all sample respondents. The multivariate technique of Principal Component Analysis (PCA) was used to determine the number of factors to be included in the analysis. The main aim of PCA is to reduce the dimensionality of a data set, while retaining as much of the variation present in that data set as possible (Jolliffe 1986, p.1). Principal components were extracted using the covariance matrix. The first seven factors had initial eigenvalues greater than one and collectively explained 78% of the variance in all 14 risk sources. Ten of the 14 risk sources had factor loadings exceeding 0.40 in absolute value in more than one factor and therefore a varimax rotation with Kaiser Normalisation was used in order to obtain factors that are easier to interpret. The rescaled communalities for risk sources all exceeded 0.62 with the exception of that for changes in the cost of capital items (0.565), indicating that most of the variance in the perceived importance of risk sources was accounted for by the first seven common factors (Manly 1986). The first five of the seven factors are shown in Appendix 1 and had interpretations that provide further insight into this analysis. These factors are discussed in this section (risk sources with absolute factor loadings <0.40 are excluded from the equations below): Factor 1: “Crop Gross Income Index” = (0.926) crop yield variability + (0.781) crop price variability – (0.518) land reform. Factor 1 indicates that the ratings for crop yield and price variability were positively correlated and displayed a high degree of variability. This factor suggests that respondents who are concerned with price and yield variability are less concerned with the threat posed by land reform and vice versa. This may be due to farmers with significant liquidity stress being less concerned about losing their farms to land reform. It may also suggest that some farmers have more confidence in the government’s land reform policies than others. A comparison of group means for this factor indicates that farmers in both regions are similarly concerned with Crop Gross Income variability. The reason that land reform seemed to be more of a concern for respondents from the Midlands (negative mean value) may be explained by a larger proportion of respondents from the Midlands (44.7%) facing land claims in line with the land redistribution program, as compared to respondents from Zululand (9.5%). Mean factor scores for each region were estimated for each factor and comparisons conducted using a two-tailed t-test for independent samples, with equal variances not assumed (Steel and Torrie 1980). Factor 2: “Macroeconomic and Political Index” = (0.710) changing credit availability + (0.655) changing capital item costs + (0.591) land reform + (0.542) interest rate variability. Mean factor scores show that Midlands respondents are more concerned with the four “Macroeconomic and Political” risk sources. This can be explained by the larger number of land claims lodged for farmland in this area, and Midlands respondents had relatively more capital investment (e.g., for forestry enterprises) than respondents from Zululand. Forestry enterprises contribute, on average, 22% of gross farm income (GFI) in the Midlands compared to 0.5% in Zululand. Mean factor scores for the two regions are statistically significantly different at the 10% level of probability. Factor 3: “Legislation Index” = (0.916) land tax legislation + (0.681) minimum wage legislation + (0.432) interest rate variability. Mean factor scores for the two regions in this factor (which are statistically significantly different at the 5% level of probability) show that the three risk sources with the highest factor loadings are more important to Midlands respondents. This could be due to respondents in the Midlands employing larger labour forces on average, using extra labour capacity mainly for their timber enterprises. The fact that respondents in this area considered the threat of a land tax to be relatively more important than respondents from Zululand could be due to increased familiarity of this issue among Midlands respondents. The higher level of information on land tax issues by Midlands respondents can be attributed to legal precedents involving the initial implementation of this legislation in the region.


70

Factor 4: “Labour and Inputs Index” = (0.929) labour unionisation + (0.526) minimum wage legislation – (0.450) changing input costs. The negative loading attached to changes in input costs suggests that respondents who are concerned with labour unionisation and minimum wage legislation are less concerned with changes in input costs and vice versa. This may be due to substitution between labour and other inputs. Zululand respondents are more concerned with changing input costs due to the more intensive nature of sugarcane farming in the coastal region. Sugarcane is normally harvested annually in the Zululand region compared to every 20 months in the Midlands. Midlands respondents consider minimum wage legislation and the threat of labour unionisation to be relatively more important. This can be attributed to respondents in the Midlands employing larger labour forces on average. Mean factor scores are statistically significantly different at the 1% level of probability. Factor 5: “Human Capital and Credit Access Index” = (0.903) HIV/AIDS + (0.512) illness or death of farm operator + (0.469) changes in credit availability. The fact that illness or death of the farm operator and changes in credit availability occur together in this factor may be due to the effects of the death of the farm operator on borrowing capacity. Mean factor scores were similar for the two study regions. The threat of HIV/AIDS, illness or death of the farm operator and changing credit availability are, therefore, considered equally important by respondents from both areas. Discussion and Conclusions This study shows that the most important risk sources as perceived by large-scale commercial sugarcane farmers in KwaZulu-Natal are the threat of land reform, the uncertainty involved with minimum wage labour legislation and the variability of the sugarcane price, in that order. With the exception of crop price variability, the relative ranking of risk factors differs from those of previous studies. Clearly, this is due to farmers now facing a new set of challenges such as continued land reform, property rates legislation and minimum wage legislation, none of which were perceived by farmers to be important in the past. The fact that the perceived importance of risk sources within dimensions has changed compared to previous studies indicates that current government land and labour legislation in particular are raising levels of uncertainty amongst commercial sugarcane producers. It is important that the government’s land and labour legislation processes are conducted in as transparent a manner as possible, with improved information made available concerning specific objectives and timeframes, in order to reduce the uncertainty involved in decision making for farmers. For the SA sugarcane industry to remain competitive in a continually globalising market environment, policy makers need to create an enabling business environment that will reduce risk and uncertainty for producers. Although recent developments regarding the land restitution process have offered farmers some certainty regarding the willing seller, willing buyer principle, further uncertainty has been created amongst farmers in terms of the accuracy and reliability of the government’s land valuation process. Government should also consider making labour legislation reform more flexible in order to avoid raising the costs associated with permanent labour to inhibitory levels. This has important implications for levels of unemployment due to the presence of substitutes for permanent labour, such as mechanisation and casual labour. Farmers also need to develop risk management strategies that reduce existing barriers to improved efficiency. To achieve this, farmers require relevant and reliable information; for example, by engaging third parties such as SACGA extension officers and other private consultants and by using published information. This study has contributed toward ongoing research into risk management amongst commercial sugarcane farmers by describing the changes in perceived risk by a representative sample of sugarcane producers in two regions of the SA sugar industry. It has identified that the threats posed by land reform and minimum wage legislation have become more relevant and are currently perceived to pose the greatest risks to


71

business viability. Further research could be aimed at quantifying the various responses to these sources of risk, and at identifying the extent to which producers consider multiple sources and responses to risk simultaneously. Acknowledgements The authors gratefully acknowledge funding from the National Research Foundation (NRF) in South Africa. The NRF supported this research under the “Making South African firms and farms competitive” project (GUN 2054254). All views, interpretations, recommendations and conclusions expressed in this paper are those of the authors and do not necessarily reflect those of the NRF. References Afrol News, 2006. Renewed Focus on South Africa Land Reform. [Online]. Available from: http://www.afrol.com/articles/18026 [cited 05 December 2006]. Barry, P.J., 2003. Major Ideas in the History of Agricultural Finance and Farm Management. Department

of Applied Economics and Management, Cornell University, Ithaca, New York. Bullock, W.I., Ortmann, G.F. and Levin, J.B., 1994. Farmer characteristics, risk sources and managerial

responses to risk in vegetable farming: Evidence from large- and small-scale commercial vegetable farmers in KwaZulu-Natal. Agrekon, 33 (3), pp.103-112.

Department of Labour, 2006. [Online]. Available from: http://www.lab.gov.org/ [cited 18 March 2006]. Democratic Alliance, 2006. [Online]. Available from: http://www.da.org.za/DA/Site/Eng/campaigns/DOCS/DAPOS%20willingbuyer%20seller.doc [cited 5 December 2006]. Eweg, M.J., 2005. The changing profile of small-scale ‘sugarcane’ farmers in South Africa. Unpublished

paper, South African Sugarcane Research Institute, Mount Edgecombe, South Africa. FAO, 2006. Sugar Commodity Note: World sugar prices have increased significantly since FAO’s

preliminary forecast in December 2005. [Online] Available from: http://www.fao.org/ES/ESC/en/20953/21032/highlight_108726en_p.html [cited 5 December 2006]. Farmer’s Weekly, 2006. Officials hatch secret land plan. 13 October 2006, pp.24-27. Hardman, P.A., Darroch, M.A.G. and Ortmann, G.F., 2002. Improving cooperation to make the South

African fresh apple export value chain more competitive. Journal on Chain and Network Science, 2 (1), pp.61-72.

Illovo Sugar, 2006. Interim Report for the six months ended 30 September 2006, Durban. [Online]. Available from:


72

http://www.illovosugar.com/financial/pdf2006/Interim/Interim%20Report.pdf [cited 5 December 2006]. Jolliffe, I.T., 1986. Principal Component Analysis. New York: Springer-Verlag New York Inc.. Manly, B.F.J., 1986. Multivariate Statistical Methods: A Primer. Bristol: J.W. Arrowsmith Ltd.. Nailana, K. and Gotte, S., 2006. Expropriation - your questions answered. AgriReview, Standard Bank,

fourth quarter 2006. [Online]. Available from:

http://www.standardbank.co.za/za/vgn/images/portal/cit_4931/41/47/24740676aGRIq42006.pdf [cited 23 March 2007]. Ortmann, G.F., 2005. Promoting the competitiveness of South African agriculture in a dynamic economic

and political environment. Agrekon, 44 (3), pp.286-320. SACGA, 2006. [Online]. Available from: http://www.sacanegrowers.co.za [cited 14 March 2006]. SASA, 2006. [Online]. Available from: http://www.sasa.org [cited 14 March 2006]. STATSSA, 2002. Census of commercial agriculture, Report No. 11-02-01 (2002). [Online]. Available from: http://www.statssa.gov.za/publications/Report-11-02-01/Report-11-02-012002.pdf [cited 1 November 2006]. Steel, R.G.D. and Torrie, J.H., 1980. Principles and Procedures of Statistics, a Biometrical Approach. 2nd

ed., McGraw-Hill Book Company. Stockil, R.C. and Ortmann, G.F., 1997. Perceptions of risk among commercial farmers in KwaZulu-Natal

in a changing economic environment. Agrekon, 36 (2), pp.139-156. Swanepoel, V. and Ortmann, G.F., 1993. Sources and management of risk in extensive livestock farming

in the North-Western Transvaal Bushveld. Agrekon, 32 (4), pp.196-200. Woodburn, M.R., Ortmann, G.F. and Levin, J.B., 1995. Sources and management of risk: Evidence from

commercial farmers in KwaZulu-Natal. South African Journal of Economic and Management Sciences, 17 (summer 1995), pp.46-63.


73

Appendix 1: Perceived importance and factor loadings of risk sources, commercial sugarcane farmers in KwaZulu-Natal, South Africa, 2006.

Risk sources

Mean rating(a)

Factors (b)

1 2 3 4 5

Land reform 4.31 -0.518 0.591

Labour legislation 4.14 0.681 0.526

Crop price variability 3.68 0.781

Changing input costs 3.56 -0.450

Crop yield variability 3.43 0.926

HIV/AIDS 3.41 0.903 Changes in capital item costs 3.33 0.655 Changes in land tax legislation 3.24 0.916 Unionisation of labour 2.89 0.929 Variability in interest rates 2.60 0.542 0.432 Changing water rights 2.26 Changing credit availability 2.13 0.710 0.469 Farm operator illness/death 1.98 0.512 Changing family relationships 1.79

Mean factor scores: Zululand: 0.146 -0.218 -0.245 -0.360 -0.047 KZN Midlands: -0.146 0.218 0.245 0.360 0.047 Means comparison (significance)(c)

0.207

0.057 *

0.033 **

0.001 ***

0.688

Note: (a) Where 1 = “not particularly important” and 5 = “highly important”

(b) Only factor loadings >0.40 in absolute value are included.

(c) *, **, *** indicate means statistically significantly different at the ten, five and one percent

levels of probability, respectively.


74

IMPROVING POLICY COHERENCE BETWEEN AGRICULTURAL AND DEVELOPMENT POLICIES

Alan Matthews

Trinity College Dublin, Ireland Email: [email protected]

Abstract There is now a strong political commitment to policy coherence for development (PCD) in many OECD countries. Agriculture is at the heart of much of the debate about possible incoherence between trade and development policy. This paper reviews the evidence on the impact which OECD country agricultural policies have on developing countries. Policies to promote coherence between agricultural policy reform in OECD countries and food security and agricultural development objectives in developing countries must take account of the need not only for improvements in market access opportunities for developing countries, but also their responsibility to integrate trade objectives as a central component of their national development strategies, as well as increased and effective international financial and technical assistance for developing production and trade capacities. Keywords: Agricultural policy, trade, developing countries, policy coherence Introduction Recent years have seen more attention paid by the development community to the pursuit of greater policy coherence in order to promote the achievement of the Millennium Development Goals (OECD, 2003). Policy coherence for development is a process whereby a government, in pursuing its domestic policy objectives, makes an effort to design policies that, at a minimum, avoid negative spillovers which would adversely affect the development prospects of poor countries and, more positively, seeks to maximise synergies. There is now widespread recognition that the impact of the transfer of resources alone by the industrialised countries through aid – the cornerstone of traditional development co-operation – will not have the desired impact – indeed, may well be undermined – if these same countries or their development partners adopt conflicting policies in other areas, such as trade, migration, investment, and so on. Agricultural trade and support policies are an oft-quoted example of policy incoherence (OECD, 2005). By limiting market access to the food markets of developed countries, while subsidising the export of surpluses to developing countries, it is argued these agricultural trade and support policies undermine markets for rural producers in developing countries and make it more difficult for these countries to trade their way out of poverty. The negotiations on further agricultural trade liberalisation in the Doha Round provide an opportunity to tackle this example of policy incoherence. Indeed, for developing countries and NGOs, achieving a high level of ambition in the agricultural negotiations has become the lynchpin by which progress in the overall talks is judged. Nonetheless, in more recent years there has been a growing sense that agricultural trade liberalisation by developed countries may not make as substantial a contribution to policy coherence as was first thought. The reasons for this are varied. Partly, it has been fed by an awareness that not all developing countries, and perhaps not even all farmers in these countries, necessarily stand to benefit from multilateral trade liberalisation. At the country level, the problem of net food importers, which could face an adverse terms of trade shock if world food prices increase as a result of liberalisation, had already been recognised in the Uruguay Round Agreement. The Marrakesh Decision was an attempt to put in place policies which could


75

help to alleviate any adverse impacts.1 The problem of net food importers arises because the consumer interest in importing countries is larger than the producer interest; producers still gain from higher world food prices but these gains are outweighed by the losses to consumers. Recently, more attention has been paid to the potential losses to those developing countries which benefit from preferential access to developed country markets, where it is the producers who are the losers. This argument has gathered force in line with new empirical results from simulation models which appear to identify a number of countries that could be made worse-off as a result of a Doha Round agricultural agreement (Ackerman, 2005; Bouët, 2006; Polaski, 2006). Even where simulation results appear to show positive gains for farmers in developing countries as well as overall gains, scepticism is evident. Some question whether the postulated increases in trade flows would in fact take place given the potential for various non-tariff barriers, both formal and informal, not captured in the model specifications to hinder this. The trade-restraining role of sanitary and phytosanitary standards is often mentioned in this context. The growing concentration in retail markets particularly in developed countries, and the related emergence of global supply chains with their potential to exclude particularly smaller producers from the benefits of formal market access, is another cause for concern. Yet another reason for scepticism concerns the ability of the poorest developing countries to take advantage of improved market access. One of the consequences of the renewed interest in preferences has been to highlight that many developing countries have failed to maintain their market share in developed country markets despite significant preferential advantages. The limitations of preferences as a way to encourage trade are well known: they are arbitrary and uncertain, their value is undermined by restrictive rules of origin, and the preferences themselves are often limited in precisely those commodities which developing countries could export. Nonetheless, the apparent lack of response to preferences suggests that increased market opportunities do not necessarily translate into increased market access.2 These arguments that the impacts of a Doha Round agricultural agreement which led to a reduction in developed country tariffs might bring more limited gains to developing countries than initially foreseen have also resulted in doubts about its likely impact on poverty alleviation. For example, an agricultural exporting country may benefit from OECD country liberalisation, but may also be required to reciprocate by reducing tariffs on import-competing food crops. If the tariff reduction outweighs the impact of higher world prices, and if import-competing food producers are relatively poorer than other households, then poverty may increase even if aggregate welfare indicators suggest that the country as a whole is better off. It is obviously important not to let the pendulum swing too far. Even if unrealistic expectations of the gains from OECD country agricultural trade liberalisation for developing countries have built up, it remains an essential ingredient in any Development Round. What is important to recognise is that there will be winners and losers from this policy change, and the gains to the winners will not come automatically. Awareness of these issues has led to a growing interest in trade-related development assistance (TRA). TRA covers technical assistance, trade capacity building, adjustment assistance and support for trade-related infrastructure. OECD countries have indicated their support to further increase TRA, including at the G8 Summit in Gleneagles in July 2005 and at the Development Committee meeting of the IMF and World Bank in September 2005. The WTO Hong Kong Ministerial Declaration in December 2005 invited

1 Its full title is the Marrakesh Ministerial Decision on Measures Concerning the Possible Negative Effects of the Reform Programme on Least-Developed and Net Food-Importing Developing Countries. 2 For example, in 1962 Africa’s share of world exports of groundnuts was 83%, by 2002 this had fallen to 3%. This collapse was not due to external trade barriers but to domestic supply difficulties.


76

the Director-General to create a task force to provide recommendations on how to operationalise Aid for Trade which has now reported (WTO, 2006). The pursuit of policy coherence for development through the reform of OECD country agricultural policies is explored in this paper. Section 2 of the paper reviews recent assessments of the likely gains from a successful Doha Round agreement on agriculture and the distribution of these gains between different developing countries. The specific problems of developing countries with preferential access to OECD country markets are discussed in Section 3. The need to accompany greater market opening with the aid for trade agenda is discussed in Section 4. Section 5 concludes that the agricultural policy coherence agenda needs to be broadened to focus not just on removing barriers to developing country exports but also to ensure that the necessary complementary policies to provide adjustment and capacity-building assistance are put in place. The Effects of Trade Liberalisation Table 1 shows that tariff barriers against developing country exports remain significant even after the Uruguay Round. The average tariff on agricultural imports by high-income countries from other high-income countries is 8.4 per cent. By contrast, the average tariff on developing country exports to high-income markets is nearly twice as high at 15.9 per cent. Developing country agricultural exports to other developing countries face even higher average tariffs at 18.3 per cent. As these figures take preferences for developing countries into account, they underline the continued barriers facing developing countries pursuing trade as a route to poverty alleviation.3

3 These figures are taken from the GTAP 6 database which in turn builds on the MacMap tariff database maintained by the Centre d’Etudes Prospectives et d’Informations Internationales (CEPII) in Paris. The higher average tariffs on developing country agricultural exports despite preferences are partly explained by their concentration on products with particularly high tariffs (sugar) and partly by the frequency of specific tariffs in developed country tariff schedules which weigh more heavily on the lower-value products typically exported by developing countries within a tariff category (Hertel and Keeney, 2006).


77

Table 1. Average applied import tariffs, by sector and region, 2001 (per cent, ad valorem equivalent)

Importing region

Exporting region High-income

economies Transition economies

Developing economies

Agriculture High-income 8.4 16.8 18.8

Transition 10.3 10.3 17.4 Developing 15.9 17.2 18.3

Other primary High-income 0.2 0.8 4.8


Textiles and apparel High-income 3.4 6.4 18.2


Other manufactures High-income 1.0 3.7 9.9


Source: Hertel and Keeney, 2006

There are now numerous studies which have simulated what would happen as a result of further reducing these trade barriers. We present a selection of the headline numbers from four of the most recent and careful studies in Table 2.4 At first glance, comparing only the projected global welfare gains from global merchandise liberalisation, the numbers vary considerably, ranging from US$84 billion (Hertel and Keeney, 2006) to US$287 billion (Anderson et al., 2006). There can be many reasons why model results differ, including differences in the way scenarios are specified, differences in the way results are presented (for example, in 2015 values for dynamic models compared to 2001 values for static models), differences in model specification (for example, whether perfect or imperfect competition is assumed, whether models are static or dynamic, and whether resources are assumed in fixed supply or not). Despite these differences, some common themes emerge (see also Bouët, 2006).

4 For surveys of empirical model results and reasons why they differ, see Ackerman, 2005; FAO, 2005a; Bouet, 2006 Chapter 4; Polaski, 2006, Chapter 4. For a critique of the computable general equilibrium methodology which underlies these model results, see Taylor and von Arnim, 2006.


78

Table 2. Recent assessments of the impact of further trade liberalisation

Anderson et al. 2006 (World Bank)

Anderson et al. 20061 (World Bank)

Hertel and Keeney, 2006

Hertel and Keeney, 2006

Bouët 2006 (IFPRI)

Polaski 20062 (Carnegie)

Scenario Global merchandise trade liberalisation

Doha Round merchandise liberalisation

Global merchandise trade liberalisation

Global agricultural trade liberalisation

Global merchandise trade liberalisation

Doha Round merchandise liberalisation

Model used Linkage 6.0

Linkage 6.0

GTAP-AGR

GTAP-AGR

MIRAGE GTAP-mod

Static or not Dynamic Dynamic Static Static Dynamic Static Aggregation (region x sector)3

27 x 25 27 x 25 29 x ? 29 x ? 20 x 17 24 x 27

Data 2001 2001 2001 2001 2001 2001 World welfare, US$bn

287 17.7 84 56 99.6 59

World welfare, %

0.04 0.33% 0.19%

of which: Share of DC gains in total

30% -3% 26% 21% 26% 51%

Share due to agricultural lib.

63% 45%3 67% 100% 9%

Share due to DC liberalisation

45% 89%

Losers None Hong Kong China Mexico Russia MENA Rest of Europe Rest of SSA

Philippines Bangladesh Other LA Mozambique Rest of SSA

Philippines Bangladesh Mozambique Rest of SSA Vietnam Other MENA4

Canada EU Argentina Mexico SACU [OECD liberalisation only - plus] Rest of SSA Zambia

Bangladesh East Africa Rest of SSA [Agricultural liberalisation alone - plus] China MENA Mexico Vietnam India Rest of SSA

Notes: 1 Scenario 2 in Anderson, Martin and van der Mensbrugghe, 2006. In this scenario, 2% of developed country and 4% of developing country agricultural tariff lines can be subject to smaller tariff cuts as a result of Sensitive Product and Special Product treatment. 2 The central Doha scenario in Polaski, 2006. GTAP-mod is the basic GTAP model but with an altered labour market specification as discussed in the text. 3 The question mark in this row indicates that the aggregation level is not specified in the paper. 4 MENA = Middle East and North Africa. The table is based on a structure which was originally devised by Bouët, 2006.


79

• First, all studies underline the global gains from further trade liberalisation, although more recent studies tend to show lower overall gains than earlier studies.

• Second, studies tend to show that, under full merchandise liberalisation, while the largest gains in absolute terms accrue to OECD countries, in proportionate terms trade reform is ‘development friendly’, i.e., the percentage gains are higher for developing countries (DCs) and highest for the LDCs.

• Third, studies tend to show that the largest proportion of gains arise because of agricultural trade liberalisation.

• Fourth, in studies which simulate a more realistic Doha scenario compared to full liberalisation, the magnitude of the estimated gains falls dramatically and a much smaller proportion accrue to developing countries.

• Fifth, while most studies show that developing countries in aggregate will benefit from further trade liberalisation, they also agree that some of the poorest countries, and particularly countries in Sub-Saharan Africa, are likely to lose particularly in the context of a more limited Doha Round outcome.

We now turn to the mechanisms proposed to turn ‘losers’ into ‘winners’ and to help ensure that ‘winners’ really win. Preference Erosion

Much greater attention has focused on the potential problems facing preference recipients in the Doha Round negotiations as compared to the Uruguay Round. All OECD countries implement preference schemes which provide developing countries with preferential access at lower than most-favoured-nation (MFN) tariffs to OECD markets. Lowering MFN tariffs will erode the value of this preferential access for beneficiary countries. Because tariffs are generally higher on agricultural and food products with more tariff peaks, preferences for these products tend to be more valuable. Not surprisingly, the consequences of preference erosion are likely to be more significant for beneficiaries with preferences in agri-food products. Two widely quoted IMF studies assuming a 40 per cent reduction in the preference margin enjoyed by LDCs and middle-income countries found an insignificant impact overall for these groups (e.g., less than two per cent of exports for all LDCs). But eight middle-income countries (where sugar and banana preferences account for the vast majority of benefits) and seven LDCs could lose 4–12 per cent of total export revenues (Subramanian, 2003; Alexandraki and Lankes, 2004; see also Low et al, 2006; Amiti and Romalis, 2006). A variety of responses have been suggested to this problem. Some authors point to the continued significance of tariff barriers even for preferred exporters, and argue that market access gains from MFN tariff reductions (either in the preference-giving country or other countries’ markets) could offset the loss of preferences. Another suggested response is to maintain nominal margins of preference to the maximum extent possible. This is clearly impossible when preferred countries already face zero tariffs. Some WTO members have proposed that tariff reductions in OECD countries for products where preferences are significant might be smaller or phased in over a longer period than might otherwise be the case under any general tariff-cutting formula that might be agreed. Yet others sought ways in which the erosion of existing preferences might be offset by the extension of new preferences. For LDCs this was achieved at the WTO Hong Kong Ministerial Meeting in December 2005, where it was agreed that all developed country members (and developing countries in a position to do so) would extend duty-free and quota-free access to LDCs by 2008, although up to 3 per cent of tariff lines can still be excluded (WTO 2005). In any event, middle-income developing countries are not affected by this offer. Led by Mauritius, which faced significant losses due to preference erosion on both sugar and clothing, there were calls for a


80

compensation mechanism for countries adversely affected by preference erosion (Commonwealth Secretariat, 2004; Hoekman and Prowse, 2005). However, assistance for trade adjustment where this is due to preference erosion is contentious. There are many sources of negative shocks that create the need for adjustment, both trade and non-trade related. Focusing on just one of these while ignoring others is difficult to justify. Trade reforms by countries which do not currently grant preferences can help to attenuate the negative impact effects of erosion. Gains from trade reforms in non-related sectors (for example, in manufacturing trade) may also balance potential losses in agriculture. This raises the difficult question, if compensation were to be made, whether this should be related to the gross value of specific preferential access arrangements, or whether it should depend on the net adverse effects of MFN liberalisation overall. A related issue is whether compensation for preference erosion is a bilateral or multilateral responsibility. Because the most important preferences originate in unilateral trade policy decisions by OECD countries, it is argued that it is those countries whose preferences are being undermined who should bear the responsibility to put in place alternative mechanisms to assist the recipient countries. On the other hand, proposals for a multilateral preference erosion compensation fund have been justified on the grounds that trade liberalisation can be seen as a global public good. The limited number and small size of most of the economies concerned imply that measures to help mitigate the impact of preference erosion need to be closely focused on the countries at risk. Aid for Trade One of the reasons that many developing countries feel they will not benefit from further liberalisation of access to OECD agri-food markets is ubiquity of supply-side constraints. Low-income countries, in particular, face many constraints in taking advantage of improved market access. They may be land-locked countries facing high transport and transit costs across neighbouring countries. They may have difficulty in complying with increasing stringent sanitary and phytosanitary standards. They may simply lack the trading infrastructure and market contacts in developed countries to exploit new market opportunities. Thus, many academics as well as WTO members have called for increased financial assistance to developing countries to accompany any market liberalisation package (Hoekman and Prowse, 2005; Charlton and Stiglitz, 2005). The scope of such aid for trade is potentially broad, covering implementation of new standards, social safety nets, support for negotiating capacity, overcoming supply side capacity constraints such as poor infrastructure, and trade facilitation and services, as well as adjustment and implementation costs for any Doha Round agreement, compensation for fiscal revenue losses, compensation for food price increases for net food importers, and compensation for preference erosion. Aid for trade has become part of a final Doha agreement since the Hong Kong Ministerial Council. The Task Force on Aid for Trade set up at that meeting reported in July 2006 (WTO, 2006). It proposed a narrower focus for aid for trade activities, including technical assistance for trade policy formulation and negotiation, trade development, trade-related infrastructure, building productive capacity, trade-related adjustment and other trade-related needs. The idea of providing compensation, whether for higher food prices, preference erosion or loss of fiscal revenues, remains contentious. But even with this narrower scope, questions remain. Does it make sense to differentiate aid for trade from development aid in general? Given that it is often difficult to distinguish the two, is it sensible to complicate the aid system by creating separate frameworks and structures for trade-related assistance? There are already a variety of new channels to deliver this assistance, including the IMF’s Trade Integration Mechanism, various bilateral donor programmes as well as multi-agency programmes such as the Integrated Framework for Trade-related Technical Assistance to Least Developed Countries. At a minimum, trade-related assistance should be disbursed in the context of the “new aid framework” which emphasises the need for


81

coordination between donors and coherence with national policies and priorities. The relationship between aid for trade and policy conditionalities which may be associated with other forms of assistance also needs clarification. Should aid commitments be brought under the WTO umbrella and formalised as part of a Doha Round Agreement, thus making them subject to the dispute resolution mechanism? The Food Aid Convention which seeks to guarantee a minimum level of food aid deliveries is a previous example of such an agreement, which also serves to underline its possible limitations. The big potential for disillusionment lies in the fact that aid is fungible, and that new ‘commitments’ for trade-related assistance may simply repackage aid flows that would otherwise go to other sectors. A proposal that all countries agreeing to increased aid for trade should subscribe to a Maintenance of Effort Commitment that current aid levels would not be reduced has been made to deal with this concern (Charlton and Stiglitz, 2006). Conclusions There is now a strong political commitment to policy coherence for development in many aid donors. This reflects the growing understanding in development circles that increased development aid resources to help developing countries to achieve the Millennium Development Goals may well be nullified by the non-developmental policies pursued by donor countries in areas such as trade, migration, agriculture and fisheries policies, investment and debt. This paper has reviewed the evidence on the impact which OECD country agricultural policies have on developing countries, and the impact which reform of these policies would have on global poverty. Recent model simulation results highlight that not all developing countries are likely to benefit from further trade liberalisation, particularly in agriculture, as a result of a successful Doha Development Round. The paper proceeds to discuss the mechanisms proposed to turn ‘losers’ into ‘winners’ and to help ensure that the ‘winners’ really win. Ensuring additional market access through ambitious reductions in both agricultural and non-agricultural trade barriers is part of the story, but only one part. Other elements are also needed: trade rules must support and not undermine food security; the fears of net food importers need to be addressed; solutions must be found to preference erosion at the country level; developing countries need assistance to improve their capacity to trade and to ensure a positive supply response to enable them to take advantage of increased market opportunities; and there needs to be a greater attention to understanding the impact on the poor of further agricultural trade liberalisation. These issues underline the importance of broadening the policy coherence agenda. First generation policy coherence policies sought reform of OECD country agricultural policies because of the way they make it more difficult for developing countries to trade their way out of poverty. Second generation policies must take account of the need not only for improvements in the international trade regime, but also ensure that developing countries integrate trade objectives as a central component of their national development strategies, as well as provide increased and effective international financial and technical assistance for developing production and trade capacities. References Ackerman, F., 2005. The Shrinking Gains from Trade: A Critical Assessment of Doha Round

Projections, Working Paper No. 05-01, Global Development and Environment Institute, Massachusetts, Tufts University.

Alexandraki, K. and Lankes, P., 2004. The Impact of Preference Erosion on Middle-Income Countries,

Working Paper WP/04/169, Washington, International Monetary Fund.


82

Amiti, M. and Romalis, J., 2006. Will the Doha Round Lead to Preference Erosion, Working Paper

WP/06/10, Washington, International Monetary Fund. Anderson, K., Martin, W. and van der Mensbrugghe, D., 2006. Market and Welfare Implications of Doha

Reform Scenarios, in Anderson, K. and Martin, W., eds., Agricultural Trade Reform and the Doha Development Agenda, Washington and New York, World Bank and Palgrave Macmillan.

Bouët, A., 2006. What Can the Poor Expect from Trade Liberalization? Opening the “Black Box” of

Trade Modelling, MTID Discussion Paper No. 93, Washington, D.C., International Food Policy Research Institute.

Charlton, A. and Stiglitz, J., 2006. Aid for Trade: a report for the Commonwealth Secretariat, London. Commonweath Secretariat, 2004. Preference-Dependent Economies and Multilateral Liberalization:

Impacts and Options, London. Department for International Development, 2003. Agriculture and poverty reduction: unlocking the

potential, DFID Policy Paper, London, DFID. FAO, 2005a. Trade Policy Simulation Models: Estimating Global Impacts of Agricultural Trade Policy

Reform in the Doha Round, FAO Trade Policy Technical Note on Issues Related to the WTO Negotiations on Agriculture No. 13, Rome, FAO.

Hertel, T. and Keeney, R., 2006. What is at Stake: The Relative Importance of Import Barriers, Export

Subsidies and Domestic Support, in Anderson, K. and Martin, W., eds., Agricultural Trade Reform and the Doha Development Agenda, Washington and New York, World Bank and Palgrave Macmillan.

Hoekman B. and Prowse S., 2005. Economic Policy Responses to Preference Erosion: From Trade as Aid

to Aid for Trade, Policy Working Paper No. WPS3721, Washington, World Bank. Patrick Low, Roberta Piermartini and Jurgen Richtering, 2006. Non-Reciprocal Preference Erosion

Arising From MFN Liberalization in Agriculture: What Are the Risks? Staff Working Paper ERSD-2006-02, Geneva, World Trade Organisation.

OECD, 2003. Policy Coherence: Vital for Development, OECD Observer Policy Brief, Paris, OECD. OECD, 2004. Non-Tariff Measures on Agricultural and Food Products: The Policy Concerns of

Emerging and Transition Economies, Paris, OECD. OECD, 2005. Agriculture and Development: the Case for Policy Coherence, Paris, OECD. Polaski, S., 2006. Winners and Losers: Impact of the Doha Round on Developing Countries,

Washington, Carnegie Endowment for International Peace. Subramanian, A., 2003. Financing of Losses from Preference Erosion, WT/TF/COH/14, Geneva, World

Trade Organisation. Taylor, L. and von Arnim, R., 2006. Modelling the Impact of Trade Liberalisation: A Critique of General

Equilibrium Models, Oxford, Oxfam International.


83

WTO, 2005. Doha Declaration, Ministerial Declaration adopted following the Hong Kong Ministerial Conference, WT/MIN(05)/DEC 18 December 2005, Geneva, World Trade Organisation.

WTO, 2006. Recommendations of the Task Force on Aid for Trade, WT/AFT/1 27 July 2006, Geneva,

World Trade Organisation.


84

MARKET ORIENTATION, SOCIAL EMBEDDEDNESS AND FIRM PROFITABILITY: AN EMPIRICAL EXPLORATION OF THE US BEEF INDUSTRY

Eric Micheels and Hamish Gow

University of Illinois Urbana-Champaign, USA Email: [email protected]

Abstract The U.S. beef industry is currently in a transition phase. As with most agricultural commodity sectors, beef producers generally perceive themselves as anonymous price takers in the market. As a result, most producers view increased operational efficiency as their only viable strategy available for improving farm profitability, as the socially embedded structure of the current marketing system coupled with the currently high cyclical prices provide insufficient incentives for producers to change their marketing orientation. Consequently industry consolidation continues as producers resist responding to consumers’ non-price market signals instead preferring to pursue operational efficiencies and scale economies. This has resulted over time in a growing divergence between the consumer and marketing channels’ (feedlot, packer, and end-user) expressed needs and the beef producers’ product offering. There is however a growing interest by beef producers towards adopting a more market orientated strategic direction as demonstrated by trade magazines now regularly discussing and providing real-world examples of various farmer -owned consumer-driven ‘value-chain’ initiatives (Tatum, 2005). This stronger market orientation and focus on consumer needs should allow firms to better identify consumer needs, implement production practices to meet those needs and more appropriately tailoring their offering to meet the consumers’ expressed and latent needs (food safety, traceability, and quality) and thereby increase firm profitability. Several studies have examined relationship between market orientation and firm performance and found there to be a significant positive effect (Narver & Slater, 1990; Slater & Narver, 2000; Olson, Slater & Hult, 2005). Market orientation is defined as the process of acquiring knowledge about customers expressed as well as latent needs and diffusing this knowledge throughout the company and channel partners (Jaworksi & Kohli, 1993). The knowledge generated during this process allows the firm to become aware of the specific characteristics and attributes customers are seeking in products or services. The goal is to meet both the customers expressed as well as latent needs while possibly augmenting the product or service with additional yet to be appreciated attributes so its perceived value is higher. In this study we examine the effect of market orientation on farm income and economic performance within the US beef industry. The beef industry was chosen as it offers different production alternatives (organic, natural, grass-fed, and grain-fed) and various alternative marketing arrangements (auction, schedule sales, contract production/marketing, specialty marketing and direct sales) coupled with various expressed and latent consumer identified product and market needs including leanness, traceability, animal identification, and certification. We believe a market oriented firm will be better able to meet the customers’ needs regarding preferences as well as their needs involving trust. Based upon a survey of 200 beef producers we empirically determine their level of market orientation and test if there are significant differences in profitability due to this increased customer focus. If so, these findings would build off previous studies which examined the market orientation effect in more traditional industries, while also encouraging agricultural producers to take advantage of this information. Keywords: beef industry, transition phase, market orientation


85

Introduction The U.S. beef industry is currently in a transition phase. Beef producers are confronted with ever-changing tastes and demanded attributes from consumers while facing stronger competition for market share by foreign producers and other protein products. At question is whether producers are willing to incur the costs associated with meeting the consumer demands. While these changes would add value and allow producers to differentiate their product their competitors, it is not required and contrary to the cost-minimization mind-set they generally operate under. In order for current producers to survive and stave off consolidation, they are going to have to find a way to provide products the consumer is demanding, even at the expense of raising costs of production. Along with the shift in demanded attributes, increased household disposable income has brought about other wholesale changes in eating habits of households. As shown in McCracken & Brandt, food expenditures away from home are positively affected by household income (1987). This follows Becker’s theory which states that as income increases there is ‘a shift away from earnings-intensive commodities and towards goods intensive ones (1965). Since 1955, the percentage of food consumed away from home has nearly doubled as a share of income (table 1). This understates the actual shift, however, as this does not consider the increase in disposable income over the same period. As shown in table 2, the percentage of disposable income that was spent on food away from home has remained fairly stable over the past 50 years, while disposable income has increased more than 30 fold over the same period. In the U.S., the rise in second-earner households has increased the value of time for the food preparer, making it more affordable time-wise for the household to eat away from home rather than spending the time to prepare meals themselves. This also has the second-order effect of increasing expectations for the consumers. As income increases and more meals are consumed away from home, expectations on consistency, taste, tenderness, and the safety of the food product also increase. Table 1: Percent of food consumed away from home

Year Food away from home as a share of food expenditure

1955 25.5 1960 26.3 1965 29.8 1970 33.4 1975 35.8 1980 39.0 1985 41.3 1990 44.9 1995 46.4 2000 47.9 2005 48.5

Source: ERS-USDA While the increase in income has positively affected beef demand, it has also led to substitution between beef products. As household income increases, people shift away from lower-value cuts or products to higher-value products. However, there has also been evidence of a substitutability of quality for time-


86

savings. McCracken & Brant point out “…eating away from home in fast-food places depends less on income than on the value of the food preparer’s time” (1987). When consumers do eat at home, there has been a shift to quickly prepared meals which can be ready in less than 10 minutes. The reason being parents and children have less time available to sit down and eat due to a lack of synchronization of schedules caused by work, sports and other activities. The main protein source of these quick-prep meals is generally not beef, but rather chicken or pork. As a result, the chicken and pork industries have increased market share by capitalizing on this consumer need for a safe and consistent product. According to the USDA, this trend is likely to continue in the next decade (2005-2015) as they predict poultry to increase its market share of meat expenditure by 1.46% over the next decade while beef market share is expected to decline by roughly the same amount (USDA-ERS). Table 2: Food expenditures by families and individuals as a share of total disposable personal income

Expenditures for food

Year

Disposable personal income At home1 Away from home2 Total3

Billion dollars

Billion dollars Percent



1955 283.3 42.9 15.1 9.8 3.5 52.7 18.6 1960 365.4 51.5 14.1 12.6 3.4 64.0 17.5 1965 498.1 58.4 11.7 16.9 3.4 75.4 15.1 1970 735.7 75.5 10.3 26.4 3.6 102.0 13.9 1975 1,187.4 117.4 9.9 45.9 3.9 163.3 13.8 1980 2,009.0 180.8 9.0 85.2 4.2 266.0 13.2 1985 3,109.3 234.0 7.5 128.6 4.1 362.6 11.7 1990 4,285.8 299.7 7.0 178.0 4.2 477.7 11.1 1995 5,408.2 345.6 6.4 225.8 4.2 571.5 10.6 2000 7,194.0 420.0 5.8 290.1 4.0 710.2 9.9 2005 9,038.6 524.3 5.8 369.8 4.1 894.1 9.9

1Food purchases from grocery stores and other retail outlets, including purchases with food stamps and WIC vouchers and food produced and consumed on farms (valued at farm prices)

because the value of these foods is included in personal income. Excludes government-donated foods.

2Purchases of meals and snacks by families and individuals, and food furnished to employees since it is included in personal income. Excludes food paid for by government and business, such as donated foods to schools, meals in prisons and other institutions, and expense-account meals.

3Total may not add due to rounding. Source: ERS-USDA One reason for the shift is the ‘no fail’ characteristics of injected pork and chicken. The meal can therefore be prepared quickly while preserving the qualities which the consumer demands, e.g. consistency, tenderness, safety. Beef, though, does not have a ‘no fail’ reputation. As people spend less time buying and preparing food, they do not gain the necessary knowledge on cooking techniques or quality characteristics of food products. A recent study showed that consumers negatively related fat content of beef cuts to quality5, as choice cuts were rated to be of less quality than select cuts (Mennecke

5 Beef quality is graded by USDA using a composite score measuring the palatability of meat. Factors included in this measure are degree of marbling, color of lean meat, texture and firmness as well as carcass maturity.


87

et. al). The knowledge gap, as well as the lack of consistent production practices, does not currently allow beef to meet the consumer need for a consistent, tender protein source in their quickly prepared meals. Unlike the beef industry which is largely independent, chicken and pork processors specify how the animals will be produced through contracting. By providing correct incentives of certain marketing opportunities at profit-making prices, processors force the grower to play by the processor’s rules, or not play at all. This is dissimilar to the beef marketing channel which does not provide incentives to change production practices to meet consumer demand. Through contracting, pork and poultry processors have a consistent commodity to which they can add value by providing consumers with safe and dependable products, as well as the ability to change as market forces dictate. Compounding the income effects and shift in eating habits was the discovery of BSE in Washington in 2003, which along with the subsequent findings both in the U.S. and Canada, had an impact on beef consumption. The loss of consumer confidence abroad led to several countries closing their borders to U.S. beef. This allowed Australian and New Zealand beef producers to enter those markets and provide the consumers the product attributes they desired. Consequently, the U.S. beef producer and marketing channel as a whole is struggling to come to a consensus on the methods to meet these needs. An ad hoc rule requiring export-bound cattle to be less than 20 months of age was instituted to ease safety concerns regarding BSE6. The reason for the age limit was the lack of a national animal identification system (NAIS) which would allow for the tracking of animals throughout the marketing channel. A traceability program would increase confidence in food safety not only abroad, but domestically as well. A survey of 1,000 consumers conducted by Global Animal Management, Inc showed consumer confidence would increase from 6.5 currently to 7.5 (10-point scale; 1 = not confident, 10 = very confident) under a mandatory animal ID system (Heinle, 2005). Yet, in its current form the NAIS will allow producers to choose their level of participation in the system. In this voluntary system, however, consumer confidence would only rate at 5.8 out of 10 (Heinle, 2005). Concerns regarding confidentiality on the farm level were the major sticking point for many producers, and these concerns have led to the NAIS being a completely voluntary system following mandatory premises registration. While the voluntary nature of the program will cause the USDA’s goal of 100% participation to be unattainable, it will allow those producers who choose to participate to differentiate themselves in the marketplace. By participating in phases II and III of the NAIS, producers will be able to provide the information feedlots, packers, and consumers are asking for relative to age, vaccinations, etc. While there will be additional costs associated with participation in the program (ear tags, labor/management costs), price premiums for the information provided should make participation worthwhile. In a recent issue of Drovers magazine, a Colorado chef explained the effects information has on his willingness to pay for quality beef. His need for ‘beef with a story’ stems from his customers access to more information concerning their food, thus increasing expectations. In order to meet this need, he is now sourcing some of his beef through local producers who can verify production practices which contribute to higher quality beef. An example of the returns to the information, “A 16 oz. prime ribeye from a small farm, raised all natural with total paperwork, is worth an easy $50 plus” (Maday, 2006). What do these changes in the consumer landscape mean for U.S. beef producers? With the voluntary nature of animal ID participation, innovative beef producers could take advantage of premium pricing if a

6 Bovine Spongiform Encepholopathy is a neurodegenerative disease in cattle. Transmission occurs through contact with infected tissues from tainted cattle.


88

majority of producers do not adopt this practice. This would allow the early adopters to receive profits from their ability to change and meet consumer needs. Adopting a market orientation would permit producers opportunities for profit making while also allowing them to build a reputation as a market leader and producer of a quality product. Market Orientation Recent studies have examined the effect a market orientation has on firm performance (Narver and Slater, 1990; Kohli and Jaworski, 1990; Deshpande, Farley and Webster, 1994). They found that firms with a higher degree of market orientation have higher profitability as measured by return on assets (ROA). Kohli and Jaworksi, however, did not find a link from a higher degree of market orientation to a higher market share in the industry (1990). However, the authors point out that market share may not be the best indicator of performance and cite instances where low market share firms have outperformed high market share firms. Furthermore, larger market share is generally may not concern individual producers who are too small to have any discernable impact on market prices. There are several different definitions of a market orientation, but they all share similar qualities. Narver and Slater define a market orientation as consisting of three main components; a customer orientation, a competitor orientation, and interfunctional coordination (1990). Jaworksi and Kohli define a market orientation to be comprised of intelligence generation, its dissemination across departments, and the overall responsiveness to the information that was gathered (1993). While the definitions are slightly different, their focus is explicitly on the customer’s needs and how the firm can best respond to meet these needs. A market orientated firm is by definition utilizing an external focus. This is orthogonal to traditional agricultural production where an internal focus on cost control or efficiency has been dominant. We are not stating an external focus is superior simply that it allows producers to meet the needs of the consumer and/or marketing channel and thus earn increased returns for this activity. However, they also must maintain an internal focus; once the innovation stops, they again must remain efficient or else risk being forced from the market. The move from the known production practices that a producer is familiar with to the unknown is entrepreneurial in nature. Naman and Slevin define the entrepreneurial firm as one with the capability to innovate, initiate change and react to changes in the market (1993). Returns to entrepreneurial behaviors are uncertain, but Ross and Westgren have found there to be positive returns to such behavior in a hog processing system (2006). In their study using a simulation approach, they find returns to segregated early weaning7. Beef producers could possibly earn similar returns from new production processes. The risk of failure or low product success, however, leads many to simply stick with the status quo. However, this is to the detriment of the consumer and ultimately the producer as efficiency gains from others force the price of the undifferentiated product ever lower. In order to compete at a small-scale, producers must find their niche, be it in production, information, or relationship building in order to survive in this market. How does a firm become market oriented? First it must focus on the needs of its consumer, possibly by working backwards through the channel. It should view the product from the consumer’s perspective so they may see where they can create the most value (Ravald and Gronroos, 1996). Value is created by taking the generic product, in our case beef, and augmenting it with attributes that consumer’s are demanding. Examples could be age and source verification, animal ID participation, or a switch to

7 A production process which enables firm to increase efficiency by increasing the number of hogs produced


89

organic production practices. How the producer chooses to create value is up to them, but they must always examine this through the eyes of the consumer. The capacity to discover and meet these consumer needs is based on the firm’s ability to observe what consumer needs are currently not being met. In the beef industry, this translates into market-sensing and information gathering as to where cattlemen can improve or change production practices to fill this void. Examples could include a move to grass-fed or natural beef production along with direct marketing as well as preconditioning steers to increase performance at the feedlot. As Slater and Narver point out, the ability to learn faster than one’s competition could be the firm’s only source of sustainable competitive advantage (1995). The capability to continually learn about the market at a rapid pace is essential to the firm’s ability to maintain its advantage. Using this information, firms then use innovation to provide consumers with augmented products which meet their expressed needs. Competitors begin to notice the increased returns to the entrepreneurial activity and enter the market thus compete these profits away. This is similar to Cochrane’s treadmill theory (1958) where the source of increased profits is the efficiency gains due to technological adoption (figure 1). Figure 1. The Innovation Treadmill

Innovation

Imitation

Reduced Profits

The firm’s decision to focus on a market orientation enables it to position itself as a provider of quality products to the market and earn a return for doing so. However, these returns can be moderated by several factors. These include, but are not limited to channel choice, commitment, trust, and technological and market turbulence. Naman and Slevin found the concept of fit to be positively related to firm performance where ‘fit’ is a measure of organizational style and strategy (1993). Foley and Fahy also propose market orientation of an organization, or in this case a marketing channel, can be affected by the ‘shared vision’ of the participants in the channel (2004). In terms of beef production, it would seem that the producer must ‘believe in the cause’ in order for a market orientation to have a significant positive effect. In other words, a strong market oriented firm who operates in an organic marketing channel, while not fully seeing the benefits or getting behind the idea, may limit the effect a market orientation may have on firm performance. Commitment has been shown to be highly correlated with market orientation, trust, cooperation, and satisfaction (Baker et al. 1999). When a firm commits to a marketing channel, all its energy is used to utilize the channel to create value for the consumer. The level of commitment put forth by channel members both up and downstream depend on the amount of trust that they have in their channel partners. Bigne and Blesa show that a firm’s level of market orientation has a significant impact (positive) on their distributor’s trust in the Spanish ceramic industry (2003). In the beef industry, this would be analogous to


90

the feedlot’s trust in the relationship increasing as the level of market orientation of the producer increases. The level of turbulence has also been found to lessen the impact a market orientation has on firm performance. Kohli and Jaworski posit in that the greater the market turbulence a firm faces, the more important a market orientation is regarding firm performance (1990). Here market turbulence refers to the changing consumer base and their evolving tastes. We see this clearly as discussed in the first section of this paper. Consumers are now earning more money, increasing beef’s customer base. Along with the increased income, though, come increased expectations relating to the eating experience. A market oriented firm may be better positioned to meet the needs of this ever-changing market. Along with market turbulence, Kohli and Jaworski examine the effect technological turbulence has on market orientation. They suggest the relationship is opposite that of market turbulence where increased technological turbulence lessens the relationship between a market orientation and firm performance (1990). Jaworksi and Kohli test these two theories using a sample of the top 500 companies (by sales revenue) in the U.S. as well as 500 members of the American Marketing Association. They find that a market orientation is key to performance regardless the level of turbulence (technological or market) faced by the firm (1993). Building from the learning organization which utilizes innovation and entrepreneurial activities to earn excess profits, we built a theoretical model to demonstrate the key processes and moderating factors which influences the effect a firm’s level of market orientation has on performance (figure 2). Figure 2. The Theoretical Model

Innovation Entrepreneurship

Channel Choice Trust Commitment Tech/Mkt Turbulence

Learning Organization

Market Orientation

Firm Performance

Implications and Discussion This paper examined the theoretical framework of employing a market orientation and its possible impact on firm performance in the beef industry. While the impact of a market orientation has been studied thoroughly in the marketing literature, little has been done in terms of how this relates to agriculture. This void in research has left firms socially embedded and unwilling to change production practices or marketing strategies for an uncertain return. Three C’s of this model seem to be of greatest significance. First is communication, where a firm meets with its channel partners and end-users to determine what needs they have and to the degree that they are


91

being met. Using this information, the firm then chooses a marketing channel in which to operate. There are several choices to make, one being to modify their existing production method in order to provide an augmented commodity product. In a more entrepreneurial move, they may choose to move to an all-natural, grass-fed, or organic marketing channel so they may fill a niche which is currently not being met. The final C is commitment. In order to fully realize the benefits of a market orientation, the firm needs to be a believer in the cause. Without the commitment to being market orientated, firms risk falling into the U-shaped curve that can lower profits for commodity firms relative to firms with both a higher and a lower degree of market orientation (Narver and Slater, 1990). This comes from firms not devoting all their attention to their market orientation. It is important to note that a market orientation does not insulate firms from inefficiency forever. They must remain efficient to defend themselves from the imitators who will enter and reduce prices and profit opportunities for all players in that specific marketing channel. Relationship building through the channel does several important things. It allows the firm to see what needs are not being met, it builds trust throughout the channel, and it leads to the more seamless flow of information between parties so changes in the market landscape can be adapted to more quickly and easily. In terms of a firm utilizing a direct marketing approach, building relationships can lead to higher retention rates of customers, as well as increasing trust in the products and services being provided. Further research would empirically examine these relationships to determine the effect a market orientation has on an agricultural setting. As agriculture is similar to other industries in many respects, it would seem an examination of market orientation on firm performance would be applicable. The degree to which a market orientation has a significant impact may depend heavily on the commodity chosen. One may suspect that a market orientation may have little impact on a typical corn/soybean rotation producer, but it may have a tremendous impact on the wine industry. Implications for positive results could be a shift in focus in extension education to help producers raise their market awareness. Overall, we feel that a strong market orientation will benefit firms, regardless of their industry. References Baker, T. L. et al. 1999. The Impact of Suppliers’ Perceptions of Reseller Market Orientation on Key

Relationship Constructs. Journal of the Academy of Marketing Science. 27 (1) 50-57. Becker, G. S. 1965. A Theory of the Allocation of Time. The Economic Journal. 75(299) 493-517. Bigne, E. and A. Blesa. 2003. Market Orientation, Trust and Satisfaction in Dyadic Relationships: A

Manufacturer-Retailer Analysis. International Journal of Retail & Distribution Management. 31 (11) 574-590.

Cochrane, W. W. 1958. Farm Prices Myth and Reality. Minneapolis, MN: University of Minnesota

Press. Deshpande, R. et al. 1993. Corporate Culture, Customer Orientation, and Innovativeness in Japanese

Firms: A Quadrad Analysis. Journal of Marketing. 57 (1) 23-37. Foley, A. and J. Fahy. 2004. Towards a Further Understanding of the Development of a Market

Orientation in the Firm: A Conceptual Framework Based on the Market-Sensing Capability. Journal of Strategic Marketing. 12 (Dec) 219-230.


92

JaworskI, B. J. and A. K. Kohli. 1993. Market Orientation: Antecedents and Consequences. Journal of Marketing. 57 (3) 53-70.

Heinle, J. 2005. Building Consumer Confidence. Drovers. Accessed online on 11/29/2006.

http://www.drovers.com/printFriendly.asp?ed_id=3244. KohlI, A. K. and B. J. Jaworksi. 1990. Market Orientation: The Construct, Research Propositions, and

Managerial Implications. Journal of Marketing. 54 (2) 1-18. Maday, J. 2006. What’s Your Beef? Drovers. Accessed online on 01/12/2006. http://www.drovers.com/news_editorial.asp?pgID=731&ed_id=3939. Mccracken, V. A. and J. A. Brandt. 1987. Household Consumption of Food-Away-From-Home: Total

Expenditure and by Type of Food Facility. American Journal of Agricultural Economics. 69(2) 274-284.

Mennecke, B. et al. 2006. A Study of the Factors that Influence Consumer Attitudes Toward Beef

Products Using the Conjoint Market Analysis Tool. Card Working Paper 06-WP 425. Accessed online Feb 24, 2006.

http://www.econ.iastate.edu/research/publications/viewabstract.asp?pid=12650 Naman, J. L. and D. P. Slevin. 1993. Entrepreneurship and the Concept of Fit: A Model and Empirical

Tests. Strategic Management Journal. 14 (2) 137-153. Narver, J. C. and S. F. Slater. 1990. The Effect of a Market Orientation on Business Profitabiltiy. Journal

of Marketing. 54 (4) 20-35. Ravald, A. and C. Gronroos. 1996. The Value Concept and Relationship Marketing. European Journal of

Marketing. 30 (2) 19-30. Ross, R. B. and R. E. Westgren. 2006. Economic Returns to Entrepreneurial Behavior. Journal of

Agricultural and Applied Economics. 38 (2) 403-419. Slater, S. F. and J. C. Narver. 1995. Market Orientation and the Learning Organization. Journal of

Marketing. 59 (3) 63-74. Usda. Economic Research Service. Paul Wescott, contact. USDA Agricultural Projections to 2016. No.

(OCE-2007-1) February 2007. http://www.ers.usda.gov/publications/oce071/


93

POTENTIAL COST OF BEING LESS TRADE DISTORTING ON U.S. CROP FARMS

J. Marc Raulston, Joe L. Outlaw, James W. Richardson Agricultural & Food Policy Center, Texas, USA


Abstract The objective of this research is to evaluate impacts of moving to a less trade distorting commodity program. An optimal control stochastic simulation model is teamed with primary representative farm data and a whole farm simulation model to evaluate the impacts of shifting government payments from countercyclical payments (CCPs) to direct payments (DPs) on U.S. crop producers. The actual difference in total government expenditures can be sizable when switching from an uncertain payment dependent on prices that fluctuate to a fixed payment that is paid regardless of prevailing market conditions. Results indicate producers historically experiencing prices high enough to exclude them from receiving substantial CCPs require very little or no increase in DPs to make them as financially viable as before removal of the CCPs. Cotton and rice farms, historically receiving significant levels of CCPs, will require a larger cash outlay in the form of DPs to maintain financial viability. Keywords: WTO, decoupled payments, trade, simulation model Objective The objective of this research is to evaluate impacts of moving to a less trade distorting commodity program. That is, how do different methods of shifting current countercyclical payments (CCPs) to direct payments (DPs) impact agricultural producers at the farm level? This study utilizes a two step methodology to measure and compare the impacts of shifting government payments from CCPs to DPs on crop producers in the United States. An optimal control stochastic simulation model is teamed with primary representative farm data and a whole farm simulation model to measure financial impacts of shifting government program payments at the firm level. Background Government program payments will certainly be more heavily scrutinized than ever as the upcoming farm bill debate gets underway. Negotiations during the now “paused” Doha Round negotiations have focused on reducing amber box payments. For the U.S. that would include the loan deficiency payments (LDPs), and the costs of the dairy and sugar support programs among others. Thus far, the U.S. has not reported expenditures for the CCP program, but it very likely could be considered amber box as it is only partially decoupled. Partially fueled by the panel rulings regarding the Brazilian cotton case, pressure is mounting for U.S. policymakers to shift government payments even further away from coupled payments tied to current production and market conditions to decoupled payments that do not depend on current production or prices. One way of accomplishing this goal is to shift expected future coupled payments to fixed, decoupled payments. Different methods for achieving this exist; however, each method has very different, specific


94

impacts on various sectors of production agriculture. In this study, a stochastic optimal control model was used to determine increases in DP rates given a budget neutral policy change through two different methods. One method estimated increases in DP rates by a fixed percentage across all crops to maintain current spending levels, but changed how the payments are awarded. A second method maintained total payments for each crop by increasing DP rates sufficiently to offset losses in coupled payments. Data and Methods A two step approach was utilized to quantify and compare the impacts of alternative methods of shifting CCPs to completely decoupled, fixed DPs. In the first step, a model for projecting annual farm program payments to nine major crops is used to determine the DP rates necessary to offset government support forfeited through elimination of CCPs. The second step in the methodology calls for simulating representative crop farms with the DP rates identified in the first step for each policy alternative in order to determine the farm level impact of these potential changes.

Stochastic Optimal Control Model

The March 2006 Congressional Budget Office (CBO) Baseline for CCC and FCIC provides a projection of annual CCP, DP, and LDP program payments for feed grains, wheat, rice, upland cotton, soybeans, and peanuts. The CBO Baseline was used to develop a stochastic simulation model that calculates annual payments for these program crops over 2007-2016. The model uses the same stochastic framework as CBO to calculate program payments over the complete range of possible crop prices and weighing these costs by the probability of price falling in the associated range. The model is naive in that it does not allow a production response to changes in target prices, DP rates, and loan rates. Given that CCPs and DPs are decoupled from production, this assumption is not viewed as a limitation to the model. The lack of a production response to reductions in loan rates is not a significant limitation if the loan rate reductions are small in percentage terms and mean prices are greater than the loan rates. Extensions in the author’s model beyond the CBO model used to develop the CBO Baseline include an update of the probability distributions for prices based on the January 2006 FAPRI Stochastic Baseline and the inclusion of minor feed grains, comprised of sorghum, barley, and oats. These minor feed grains were added to the model using the January 2007 FAPRI Baseline projections of prices, acres, yields, DPs, CCPs, and LDPs for these crops. The CBO Baseline reports total payments to the three minor feed grains. The proportion of payments in FAPRI’s Baseline paid annually to each crop was used to apportion CBO’s projected payments to the minor feed grains. The mix of payments (CCP, DP, and LDP) to the minor feed grains was estimated using the fraction of payments for these programs in the FAPRI Baseline. An optimal control mechanism (Solver in Microsoft® Excel) was used to estimate unreported price wedges, LDP wedges, and program participation fractions implicit in the CBO Baseline. After calibrating the model to the March 2006 CBO Baseline, the difference in total payments (error) for the nine program crops over the 2007 to 2016 period between the two models was $0.907 billion, or less than one percent, on a $104 billion budget forecast. Total government expenditures for 2007-2016 CCPs are calculated given current and projected market conditions assuming the January 2007 FAPRI Baseline and assuming continuation of current farm program provisions (2002 farm bill). The model was used to estimate the increase in DP rates necessary to offset an elimination of the CCP program. The optimal control mechanism was used to estimate the DPs assuming there are no CCPs over the 2007-2016 period. The DP rates were calculated two ways:


95

Equity DP -- the DP rates for all crops were increased the same regardless of which crop had generated the CCPs. No Equity DP -- the DP rates were only increased for a crop to offset its loss in CCPs. The DP rates for the Base DP scenario and the calculated DP rates for the Equity DP and No Equity DP alternatives are reported in Table 1. Table 1: Direct payment rates for Base situation and two policy alternatives and average probability of receiving CCP by crop, 2008-2012.

• • Base DP • Equity DP • No Equity DP • P (CCP > 0)

• • --$/unit-- • --$/unit-- • --$/unit-- • --%--

• Cotton • 0.0667 • 0.0815 • 0.1547 • 91.4 • Wheat • 0.52 • 0.64 • 0.53 • 3.2

• Sorghum • 0.63 • 0.76 • 0.63 • 1.6 • Corn • 0.28 • 0.34 • 0.28 • 0.8

• Barley • 0.24 • 0.29 • 0.24 • 2.2 • Oats • 0.02 • 0.03 • 0.03 • 2.0

• Soybeans • 0.44 • 0.54 • 0.44 • 7.7 • Rice • 2.35 • 2.87 • 3.10 • 38.8

• Peanuts • 36.00 • 44.00 • 36.00 • 59.2 For the first option, the DP rates were increased the same amount (22.2 percent) for all crops to offset the 10 year expected CCPs of $11.35 Billion. Under the second method, each crop’s DP rate was solved for on a crop by crop basis using the optimal control mechanism in Excel. The DP rate for wheat increased 2.8 percent and the DP rate for corn remained unchanged because these crops have very low projected CCPs in the CBO baseline (Table 1). On the other hand, the DP rate for cotton increased 131.9 percent and the rice DP rate increased 32.1 percent as these crops have projected CCPs of $7.96 Billion and $1.33 Billion, respectively, over the next 10 years. For the first option, the DP rates were increased the same amount (22.2 percent) for all crops to offset the 10 year expected CCPs of $11.35 Billion. Under the second method, each crop’s DP rate was solved for on a crop by crop basis using the optimal control mechanism in Excel. The DP rate for wheat increased 2.8 percent and the DP rate for corn remained unchanged because these crops have very low projected CCPs in the CBO baseline (Table 1). On the other hand, the DP rate for cotton increased 131.9 percent and the rice DP rate increased 32.1 percent as these crops have projected CCPs of $7.96 Billion and $1.33 Billion, respectively, over the next 10 years. Representative Farm Analysis The simulation step utilizes primary representative farm data paired with a whole farm simulation model to examine the effects of alternative farm policies on agricultural producers. The representative farms were created through a focus group interview process. Variables including commodity prices, crop yields, production costs, equipment complement, and government program data are collected in initial meetings with the focus groups and are periodically updated through face to face meetings with the panels of producers. The Agricultural & Food Policy Center (AFPC) representative crop farms are categorized into four commodity groups based on percent of total receipts earned from particular commodities. Impacts of policy changes were evaluated on representative farms located in major production regions throughout the United States using the farm level income and policy simulation model (FLIPSIM)


96

developed by Richardson and Nixon (1986). The FLIPSIM model draws random crop yields, livestock production variables, and prices from a multivariate empirical probability distribution allowing projections to incorporate production and price risk using the procedures described by Richardson, Klose, and Gray (2000). This study analyzed two farms from each of four commodity groups including feedgrain (Iowa and Nebraska), cotton (Texas Middle Coast and Georgia), wheat (Northwest Kansas and Montana), and rice (Texas and Arkansas). Government program variables for the eight farms are reported in Table 2. Results All representative farms within a commodity classification exhibited consistent preferences. Table 3 reports total government payments for the Base DP scenario and the percentage change resulting from implementing the Equity DP and No Equity DP alternatives. Average annual net cash farm incomes (NCFI) for the representative farms show the impacts of eliminating the CCP program and increasing the DP rates on the farms’ financial situations. The two representative feedgrain farms prefer the Equity DP scenario based on increases in NCFI. The second choice for the feedgrain farms is the Base DP situation. The least preferred alternative is the inequitable distribution of former CCPs in the No Equity DP scenario, as this scenario results in a decrease in NCFI for the Iowa and Nebraska representative farms of 1.6 percent and 1.5 percent, respectively. Although the feedgrain farms have considerably better CCP payment yields due to updating in response to the 2002 farm bill legislation, projected corn and soybean prices are high enough that a CCP is expected to be paid only 0.8 percent of the time for corn and 7.7 percent of the time for soybeans over the 2008-2012 period.


97

Table 2: Government program base acres and program payment yields for AFPC Representative Farms.

• • • Base

Acres • DP Yield • CCP

Yield

• Feedgrain • • • • Iowa • • • • Corn • 675 • 127 • 154

• Soybeans • 675 • 37 • 45 • Nebraska • • •

• Corn • 1470 • 130 • 170 • Soybeans • 300 • 42.3 • 56.3

• Wheat • • • • Kansas • • • • Wheat • 1200 • 37 • 37

• Sorghum • 450 • 37 • 37 • Corn • 450 • 70 • 70

• Montana • • • • Wheat • 2295 • 41 • 41 • Barley • 1260 • 42 • 42 • Cotton • • • • Texas • • •

• Sorghum • 495 • 39.4 • 41.4 • Cotton • 720 • 548 • 632 • Corn • 495 • 86 • 90 • Rice • 90 • 56.3 • 57.6

• Georgia • • • • Cotton • 1495 • 833 • 880 • Corn • 230 • 82.2 • 88.25

• Peanuts • 575 • 1.9 • 1.9 • Rice • • •

• Texas • • • • Rice • 1280 • 60 • 60

• Sorghum • 160 • 43.7 • 43.7 • Soybeans • 50 • 23 • 23 • Arkansas • • •

• Rice • 1620 • 55.4 • 59.4 • Wheat • 235 • 44 • 44

• Soybeans • 1620 • 29 • 36


98

Table 3. Average total government payments and net cash farm income for Base situation and percent change from Base for two policy alternatives, 2008-2012.

• • • Base DP • Equity

DP • No Equity

DP

• • -- $1000 -- • -- % Change from Base --

• Government Payments

• • •

• Feedgrain • • • • Iowa • 57.4 • 5.6% • -4.1%

• Nebraska • 99.2 • 6.2% • -3.9% • Wheat • • •

• Kansas • 37.8 • 11.7% • -1.7% • Montana • 55.3 • 14.6% • -0.3%

• Cotton • • • • Texas • 95.9 • -22.0% • -5.4%

• Georgia • 293.4 • -37.8% • -16.6% • Rice • • •

• Texas • 181.2 • 4.4% • 10.8% • Arkansas • 180.7 • 3.5% • 7.0%

• • • • • Net Cash Farm

Income • • •

• Feedgrain • • • • Iowa • 207.5 • 1.2% • -1.6%

• Nebraska • 429.9 • 0.8% • -1.5% • Wheat • • •

• Kansas • 87.4 • 4.4% • -1.9% • Montana • 203.0 • 4.0% • -0.1%

• Cotton • • • • Texas • 117.2 • -33.5% • -18.0%

• Georgia • 288.8 • -56.2% • -31.6% • Rice • • •

• Texas • -341.3

• 0.2% • 4.1%

• Arkansas • 134.2 • 1.0% • 6.3%

The representative wheat farms exhibited similar preferences to the feedgrain farms; however, the preference of the Base DP scenario over the No Equity DP scenario is very slight, as the NCFI for the No Equity DP alternative is only 1.9 percent lower for the Kansas farm and 0.1 percent lower for the Montana farm. Wheat is only expected to experience a CCP 3.2 percent of the time over the 2008-2012 period. The preference for the Equity DP situation over the Base DP case is a much stronger preference. The representative wheat farms did not update their payment yields during the 2002 farm bill base and yield updating period, so their payment yields are equal between CCP and DP. Both representative cotton farms prefer the Base DP situation. Cotton producers view CCPs favorably as they are expected to be paid on average 91.4 percent of the time over the 2008-2012 period. Increasing


99

the DP to compensate for lost CCPs inequitably in the No Equity DP scenario is the second preference for the representative cotton farms, although it results in sharp declines in NCFI for the Texas and Georgia farms of 18.0 percent and 31.6 percent, respectively. The Base DP scenario is preferred over the No Equity DP scenario because the associated increases in DP rates are not enough to compensate for the high CCPs already expected to be paid out over the period on a higher CCP yield. The equitable shift of former CCPs (Equity DP) is the least favorable scenario for cotton producers. On average, rice is expected to experience prices low enough to trigger a CCP payment 38.8 percent of the time over the 2008-2012 period. It is interesting to note rice producers comprising the Texas representative farm did not update CCP payment yields during the 2002 farm bill base and yield update period as they would have lost valuable rice base acres in the process. Planted acres of rice in Texas have decreased sharply in recent history, and updating payment yields would have required assigning base acres on the basis of plantings over the 1998-2001 period. Both representative rice farms prefer the inequitable distribution of DPs in the No Equity DP scenario. The second preference for rice producers is the equitable shift of DPs in the Equity DP scenario, thus preferring the Base DP situation last. Implementation of the Equity DP scenario results in modest increases in NCFI over the Base DP situation of 0.2 percent and 1.0 percent for the Texas and Arkansas farms, respectively. Rice producers prefer both methods of shifting their CCPs, a risky form of government support, to DPs, a payment that is guaranteed regardless of prevailing market conditions because of little or no differences in payment yields. In summary, producers historically experiencing prices high enough to exclude them from receiving CCPs of any consequence require very little or no increase in DPs to make them as financially viable as before removal of the CCPs. Cotton and rice farms, historically receiving substantial CCPs, will require a larger cash outlay in the form of DPs to maintain financial viability. Discussion The actual difference in total government expenditures can be sizable when switching from an uncertain payment that is dependent on prices that fluctuate to a fixed payment that is paid regardless of prevailing market conditions. All producers will not necessarily be affected equally. In addition, the process of updating farm program yields and base acres associated with the 2002 farm bill affects how a farm is impacted by the policy change. Many rice farms held off on updating program yields in many areas as they would have lost valuable rice program acres to improve DP yields, and, as a result, would suffer from converting CCPs to DPs. Cotton producers prefer CCP because it is essentially guaranteed money under the current FAPRI price projections. Rice producers prefer to shift risky CCP payments into higher guaranteed fixed payments. Feedgrain, oilseed, and wheat farmers were expected to receive very little or no CCPs anyway, so the transfer of expected CCPs to DPs is favorable for them. References

Food and Agricultural Policy Research Institute. February 2007. FAPRI US Baseline Briefing Book. University of Missouri, Columbia, MO. FAPRI – UMC Report #02-07.

Hull, D., J. Langley, and G. Hitz. 2006. “CBO March 2006 Baseline for CCC and FCIC Outlays.” U.S.

Government Congressional Budget Office. Richardson, J. W. and C. J. Nixon. July 1986. “Description of FLIPSIM V: A General Firm Level Policy

Simulation Model.” Texas Agricultural Experiment Station, Bulletin B-1528. Richardson, J.W., S.L. Klose, and A.W. Gray. 2000. “An Applied Procedure for Estimating and

Simulating Multivariate Empirical (MVE) Probability Distributions in Farm-Level Risk Assessment and Policy Analysis.” Journal of Agricultural and Applied Economics, 32:2: pgs. 299-315.


100

FARMERS’ BEHAVIOURAL INCLINATIONS AND THEIR INFLUENCE ON THE ANTICIPATED RESPONSE TO THE REFORM OF THE COMMON AGRICULTURAL

POLICY

Tahir Rehman, Chris Garforth, Kevin McKemey, Chris Yates and Ram Rana School of Agriculture, Policy and Development

University of Reading, Early Gate, PO Box 237, Reading RG6 7RP UK Email: [email protected]

Abstract Recently the University of Reading has completed a project on behalf of Defra (Department of Food, Environment and Rural Affairs) to understand the behaviour and motivation of farmers in adjusting to the reform of the Common Agricultural Policy (CAP), particularly to the Single Payment Scheme. This research provides interesting insights into how farmers can be expected to use the Single Payment (SP). In the literature on goals and objectives, the main interest is in ascertaining farmers’ motivations for being in farming. The Reading project has created an ‘influence’ model to identify the factors that are likely to determine farmers’ responses, in a differentiated way, to the unprecedented event of the SP. The Reading typology of farmers is a refined set of behavioural types, capable of providing insights into farmers’ intentions with regard to the SP. The project has used data from a survey, which used a postal questionnaire with a stratified (by region and farm type) random sample of 3,000 farmers in England in January 2006. Some 683 useable responses to 25 statements on “objectives” in farming, and 26 statements on “values” were generated. The questionnaire also elicited farmers’ attitudes and likely responses to the introduction of the SP. A set of six behavioural responses were identified through discussion with farmers including a general response of changing one’s farming system and practices in the next five years, and five specific ways of applying the SP. The analysis of farmers’ responses shows that of the five potential methods of using the SP, the most likely to be adopted is to regard it as a substitute for the previous production-linked subsidies. The respondents felt that family members, business partners, accountants and the farming press would strongly support changing the farming system and practices as a result of the SP, while Defra, land agents and other farmers would be indifferent or against the idea. Amongst all five farmer types the family is the strongest influence. Referents fall into three distinct categories: referents external to the farm business, farming peers, and those that are internal to the business (including family members and business partners). Attitudes, perceived behavioural control and the views of others all have a significant influence on farmers’ behavioural intentions with respect to the use of the SP.

Introduction

The University of Reading has recently completed a project for Defra: “Research to Understand and Model the Behaviour and Motivations of Farmers to Policy Changes (England)”; the project has explored and assessed the possibility of incorporating data on farmers’ motivations and the influences on their behaviour into Defra policy analysis models. The objectives of the research were to:

i review existing literature on farmers’ motivations and behavioural influences;

ii review existing predictive models intended to simulate or forecast farmers’ responses to policy changes or market price changes, drawing out their strengths and weaknesses and identifying implicit assumptions;

iii gather and analyse appropriate data on farmers’ motivations and behavioural influences relevant to their farm management decisions;

iv identify and describe the main factors found to influence farmers’ behaviour;


101

v characterise different groups of farmers found to have distinct behavioural patterns;

vi where possible relate any such groups to existing conventional systems for classifying farm types including the Farm Business Survey, and to farm income/return on capital;

vii construct and, where possible, parameterise an “influence model” of farmer behaviour capable of describing the behaviour of the full range of groups identified at (v) above; and,

viii make recommendations for using the outputs of this research in conjunction with existing and possible new quantitative economic models used by Defra, and as far as possible specify in detail the techniques and construct the model design.

The study1 began with a review of international academic literature on farmers’ motivations, values, objectives and behavioural influences. For objective (ii), a review of policy modelling literature was followed by a series of interviews with key informants working on the models most relevant to the present study. Empirical data for objectives (iii) to (vi) came from two surveys: the ADAS 2005 Farmers’ Voice survey, and a stratified random sample survey of farm holdings in England drawn by Defra from the June Census database. These data were used in three main ways: (a) for an analysis based on the Theory of Planned Behaviour of the influences on farmers’ behavioural responses to the Single Payment Scheme; (b) to identify, through Principal Component Analysis (PCA) and Cluster Analysis, distinct farmer types in respect of values and behavioural objectives; and (c) to contribute to the construction and parameterisation of “influence models” (objective (vii)). The proposition tested in our research is that it is possible to extract strata of behavioural types from empirically collected data. These strata can be merged with predictive models for policy analysis to generate differentiated predictions of responses to policy changes.

A postal questionnaire survey of a stratified (by region and farm type) random sample of 3,000 farmers in England in January 2006 generated 683 responses to 25 statements of “objectives” in farming, and 26 statements of “values”, on nine point Likert scales. Factor Analysis, using Principal Component Analysis followed by a two-step Cluster Analysis, identified five distinct behavioural types: family orientation, business / entrepreneur, enthusiast / hobbyist, lifestyler, and independent / small farmer. The objectives and most of the values statements used in the postal survey represent long term, enduring aspirations and, therefore, the behavioural types derived from them can, in turn, be expected to remain robust through changes in the policy and business environment in which farmers operate. Farmers’ attitudes and likely responses to the introduction of the Single SPS were explored within the conceptual framework of the Theory of Planned Behaviour (TpB), which postulates that behavioural intention is determined by a combination of attitudes towards the outcomes of the behaviour, perception of the views of others towards the behaviour (subjective norm), and the degree of control one thinks one has over a decision to carry out the behaviour (perceived behavioural control).

Six behavioural responses were identified through discussion with farmers: a general response of changing one’s farming system and practices in the next five years, and then five specific ways of applying the Single Payment (SP) that farmers’ will receive under the SPS. Data on key TpB variables were collected through a two stage process. A series of focus groups identified a set of outcomes that farmers believed may or may not occur as a result of the SPS (“outcome beliefs”), plus a list of people and organisations (“referents”) to whom farmers might turn for advice in respect of SPS. Farmers’ behavioural intentions, and their assessment of the outcomes and how they would react to the views of referents, were measured on rating scales through our postal survey.

1 The final report and all the associated material with this project is available at: http://statistics.defra.gov.uk/esg/reports/Farmer%20Behaviour/


102

Of the five potential methods of using the SP, the most likely to be adopted is to regard it as a substitute for the previous production-linked subsidies. Analysis of the influence of ten referents showed that respondents felt family members, business partners, accountants and the farming press would strongly support a decision to change farming system and practices as a result of the SPS, while Defra, land agents and other farmers would be neutral or against the idea. Respondents are least motivated to comply with what Defra, consultants and land agents suggest. They also do not feel inclined to follow the views of other farmers, apart from those with whom they associate in farmers’ clubs. This suggests they think the farming population generally is as uncertain as they are about the implications and future consequences of the SPS. For all five farmer types, the family is acknowledged as the strongest influence. Farming press and farmers’ clubs have a stronger influence on the independent farmers than on other types, while enthusiast/hobbyists are least likely to be influenced by accountants. Cluster analysis shows that referents fall into three distinct categories: referents external to the farm business, farming peers, and those that are internal to the business (including family members and business partners). Correlation of the TpB parameters with intention shows that attitudes, perceived behavioural control and the views of others all have a significant influence on farmers’ behavioural intentions with respect to SP. Further analysis through ordinal regression showed that farmers’ attitudes towards the impact of SPS on farming in general have a separate and significant influence.

After this introduction, this paper deals with: (i) the behavioural typology of farmers; (ii) the influences on behavioural intentions of farmers with regard to Single Payment; (iii)observations emerging out of the research undertaken.

Treatment of Behaviour in Agricultural Policy Models The social psychology theories of value expectancy, such as the Theory of Reasoned Action or its extension the Theory of Planned Behaviour, offer promise in understanding and modelling behaviour (not necessarily defined exclusively in terms of profit or utility maximisation) when they are combined with traditional economic analysis (Lynn 1995). As the reform and restructuring of the Common Agricultural Policy proceeds, the demand for behavioural studies and models is likely to increase, particularly in the context of participation in environmental management schemes to mention one case in point (Burton 2004). A policy model is expected to extrapolate and predict from a sample for the population as a whole. If a pragmatic approach could be devised to group farmers (Edward-Jones et al. 1998) into strata where members of each stratum are sufficiently similar to be taken as one ‘type’ and at the same time they are sufficiently distinct from other types, then a policy model can be built that takes a step towards the inclusion of behavioural influences that act on individual decision-makers. Theoretical approach – Theory of Planned Behaviour The Theory of Planned Behaviour (TpB) was developed by Ajzen (1988; 2005) and it is an extension of the Theory of Reasoned Action (TORA) originally proposed by (Ajzen and Fishbein, 1980). Both TORA and TpB provide the conceptual framework for exploring farmers’ attitudes and intentions. According to TORA, the intention to adopt a particular behaviour is a function of attitudes towards the behaviour and the subjective norm – the extent to which one is influenced by the views of other people regarding the behaviour. Attitudes are a product of the extent to which one expects the behaviour to result in specified outcomes and the perceived importance attributed to those outcomes. The subjective norm is a function of the perceived support of important referents toward the performance of the behaviour and the motivation to comply with those referents. TORA claims that the intention to undertake a particular behaviour is a


103

reliable indicator of future behaviour, if the expressed attitude toward this behaviour and/or the perceived social pressure to do so correlate closely with the stated intent. A comparison of the strength of correlation of the stated attitude (SA) and subjective norm (SN) with the stated intent (I) to apply the Single Payment (SP) in a particular way indicates which of the two components has greater influence on the subjects' decision to apply the SP for the suggested purpose. The quantitative components of TORA are stated as: TpB extends TORA by introducing an additional component, perceived behavioural control (PBC). PBC is an assessment of the actor’s perceived ability to perform a particular behaviour and his/her capability to do so. TpB states that PBC can also predict behavioural intent. The contribution of PBC is assessed by comparing the strength of correlation with intent with that of the other two causal components, attitude (SA) and the subjective norm (SN). TpB is generally seen as a more appropriate conceptual framework when studying behaviours which are not fully under ‘volitional control’: i.e. where an individual might want to carry out a particular behaviour but feels he or she is constrained from doing so. The quantitative componets after the modification stand as: TpB was used as the conceptual framework and has been applied to predicting farmers’ intentions to change their farming systems over the next 5 years as a result of the introduction of the SP. Figure 1 is a graphic presentation of the Theory of Planned Behaviour. To understand farmers’ attitudes toward the SP in general, the salient outcome attitudes (OAs) have been combined to form a ‘reasoned’ or calculated attitude (CA)2. The OAs do not relate to the specific six behavioural intentions3 considered and are therefore not associated directly with these. Rather, the strengths of the specific OAs are used to gain a deeper understanding of the beliefs and values underpinning the farmers’ stated attitudes toward the SP. To enhance the understanding of the farmers’ attitude to the SP a further measure of attitude is also taken, the general or emotive measure (GA). This is

2 An OA is the product of ‘strength of belief’ that an outcome will result (b) and evaluation of how good or bad that outcome would be (e). Thus OAi = ei*bi, and CA is the sum of all the OAs. 3 That is, change in farming system and practice in the next five years; and five behaviours related to how the farmer intends to use the SP.

comply toMotivation

belief Normative

evaluation Outcome

strength Belief

Norm Subjective

Attitude

behaviour theperform toIntention

actionor Behaviour

where

21

211

mc

nb

oe

bs

SN

Aact

BI

B

SNAact

mcnboebsBIB jjii

γγ

γγ

+=

∑+∑=≈

belief control ofPower

belief Control

control lbehavioura Perceived

where

321

3211

pb

cb

PBC

PBCSNAact

pbcbmcnboebsBIB kkjjii

∑++=

∑+∑+∑=≈

γγγ

γγγ


104

arrived at by taking the mean of the farmers’ evaluation of the necessity, helpfulness and constructiveness of the SP, each measured on a 5 point bi-polar scale. In contrast, specific referent social norms (RSNs) are related to the farmers’ intention (I) to change their farming practices in the next five years. Therefore, those RSNs that are found to correlate closely with the stated intent to change indicate which salient referents are likely to have greatest influence on the subjects' decisions.4 Figure 1: Schematic representation of the Theory of Planned Behaviour

The non-parametric Spearman Rank Order Correlation (rs) has been applied to identify the differences in the contribution or influence of the attitude and subjective norm on the intention (I). Similarly a non-parametric equivalent of t test, the Mann Witney U test, is applied to identify significant differences in the TORA variable readings between the comparative categories such as size of holding, type of farm enterprise and type of farmer, tenure, level of education, gender etc. The research involved data gathering in two interdependent stages. Initially the salient outcome beliefs, social referents and probable investment strategies regarding the SP were identified through focus group discussions with farmers in three different areas of England. The second stage incorporated the identified salient outcome beliefs and pertinent referents in a structured questionnaire, which was then posted to a random sample of 3000 English farmers. Focus Group Discussions Three focus groups for farmers were held in May 2005 – in Devon, Norwich and Reading – and a fourth for students studying agriculture at first degree level, and planning to go into farming on completion of their studies, to capture the views of the next generation of farmers. Between four (for the students) and

4 RSN is the product, for a specified social referent, of the respondent’s motivation to comply with that referent (m) and the respondent’s subjective belief (sb) – i.e. how likely the referent is to approve or disapprove of the respondent carrying out the behaviour in question.

Outcome belief

Outcome evaluation

Subjective belief

Motivation to comply

Control belief

Power of control

Attitude

Subjective norm

Perceived behavioural control

Intention Behaviour


105

ten participants took part in each discussion. In each case, participants were identified by consultants in the area with a deliberate attempt to represent a broad range of farm types and sizes. Each discussion was structured around three main issues:

a) the reasons why the participants are farming, or were intending to farm in the future

b) how they intend to apply the SP and what they expect its impact to be on farming

c) what sources of advice they would turn to on the SP and its application.

As part of the discussion on a), participants were asked if they could identify themselves in any of three farmer types referred to in the literature on farmer objectives and motivations as related to their behavioural orientations: dedicated producer, flexible strategist and environmentalist or lifestyler. These behavioural types were taken from the existing literature (Fairweather and Keating 1994) and their existence in the target population was corroborated by a Cluster Analysis of the ADAS Farmers’ Voice survey data collected in 2006. Postal Survey A representative random sample of 3,000 was drawn by Defra statisticians, which was stratified by farm type and by region. Defra also supplied data from the June Census returns for each of the years 2001 to 2005 for all the farm holdings in the sample. These data were needed to build and parameterise the models as described in the section below on farm models. Questionnaires were sent out with a covering letter between 11th and 13th January 2006. A reminder was sent to those who had not yet responded on 3rd February; and a second reminder with a copy of the questionnaire two weeks later. An effective response rate of 25% was achieved. The questionnaire for the postal survey comprised seven main parts with a total of 36 questions. Part 1 asked for some basic information about the farm and farm business that was not included in the June Census data. Part 2 comprised the statements of objectives and values which respondents were asked to rate on scales of importance and agreement. Part 3 focused on the main building blocks of the TpB framework, asking respondents to indicate their intentions, attitudes, perceived difficulty, and perceptions of the views of others, in respect of changing their farming practice as a result of the introduction of the SP. Part 4 asked for stated intentions, attitudes and social norms in respect of using the SP in five specific ways identified from the focus groups. Part 5 listed 15 salient outcome beliefs about SP (obtained from the focus groups) and asked respondents to indicate their agreement / disagreement with, and their evaluation of each statement on five point scales. Part 6 asked how supportive they think ten specified organisations and people would be if they were to change farming practice as a result of the introduction of the SP and how motivated they would be to follow their advice. Part 7 covered information about the respondent. The questionnaire was piloted by mailing it with an adapted version of the covering letter to a sample of 100 farmers selected randomly from the “farmers” listing at yell.com. Ten replies were received. The completed questionnaires suggested that respondents who decided to take part did not find the questions and format of responses difficult to deal with. A few minor modifications were made to the questionnaire. In total, 742 usable questionnaires were returned. Those reporting farm areas less than 4 ha were excluded from the TpB analysis and modelling, to avoid the results being skewed by people who are not farming as a business.


106

Behavioural Typology of Farmers Derivation of farmer types from the postal survey was done in two stages. First, a set of factors was identified through Principal Component Analysis (PCA) of responses to the objectives and values statements in questions 13 and 14 of the survey questionnaire (Appendix I). Then a two-step Cluster Analysis was performed on these factors to identify distinct clusters of respondents. Those farming less than four hectares were excluded from the analysis, leaving a dataset of 683 survey respondents. The sixteen factors that emerged from the PCA are listed in and Table 1 and 2 below, together with the ‘objectives’ and ‘values’ statements that are associated significantly with them. The first step of the subsequent Cluster Analysis identified two distinct clusters, one reflecting a mostly positive outlook on a majority of factors whereas the other was just the opposite. The former cluster had 379 cases (55.5%) and the latter had 304 cases (44.5%). In the second step, separate CAs were conducted for the two sub-samples using the same set of factors as employed previously. From the first sub-sample a further two clusters emerged, labelled family oriented and business orientation/entrepreneur with 202 and 177 cases respectively. Similarly, from the second sub-sample three further clusters, enthusiast/hobbyist, lifestyler and independent/small farmer with 113, 147 and 44 cases respectively, were identified. Table 4 shows the distinguishing features of each type, in terms of the factors identified in the PCA. The family orientation type score highly on environmental aspects and such farmers tend to be very sensitive to environmental issues. Considerations such as “stewardship”, “working alongside family” and “passing on viable business to the next generation” receive priority over other factors and this group tends to be content with the institutional and communal outlook on farming and they don’t feel neglected. The business orientation /entrepreneur behavioural type records high scores in several of the factors that determine their success in business. This group views farming as a business and approach it professionally, scoring highly on “quality of achievement”, “expansion”, “investment”, “debt avoidance” and “staff management”. The members of this group however feel that they have been marginalised despite doing a worthwhile job in the community, leading to dissatisfaction with the present state of affairs. The behavioural label enthusiast/hobbyist would suggest that to such farmers farming is a hobby activity, with the main occupation being something different. This group has high scores on “diversification” combined with low scores on “profit” and the financial aspects. The simultaneously high scores on “quality of life” and “leisure” indicate that farmers are more concerned about reducing work load and spending more quality time with family and friends away from the farm. Such farmers farm because of the intrinsic values attached to farming as reflected in the “job satisfaction” factor. Not being full-time farmers, they do not record a high score on “independence”. The lifestyler behavioural orientation scores highly on “family standard of life”, suggesting that the objective for being in farming is to increase family income to maintain and/or increase “family’s standard of living”. At the same time there are high scores for “quality of life” and “leisure”, indicating that such farmers balance high income with reduced work load and quality time with family and friends. This group scores highly also on future “security /investment” and “staff management”. A low level of job satisfaction is expressed and there is an awareness of and a concern for the uncertainty associated with farming. A high score for “marginalisation” might suggest that this group feels let down by the government and society at large. The independent/small farmer group also records high scores on “family standard of living”, but unlike the lifestyler group, low scores for “quality of life” and “leisure” contrast with the high scores for “job satisfaction” and “independence”, indicating the emotive value of farming and the intrinsic nature of these influences. This group is rather indifferent to “profit” and “financial” aspects, reinforcing the impression


107

that their reasons for farming are more intrinsic rather than instrumental. Interestingly though, this group does not feel marginalised. Table 1: Factors extracted by PCA from ‘objective’ statements

• Factor • Factor label • Statements with significantly high factor loadings • 1 • Family standard of

life • Increase family income • Maintain my family’s standard of living • Improve my family’s standard of living

• 2 • Quality of achievement

• Produce the best quality output on my farm • Be the best farmer I can be • Contribute to the farm in order to achieve something

• 3 • Environmental concern

• Be sensitive to the environmental impacts of farming by reducing input on my farm

• Do everything to be environmentally aware • 4 • Expansionist • Buy more land

• Rent/contract more land • 5 • Quality of life and

leisure • Reduce work load and improve quality of life • Make more time to spend on activities away from the

farm • 6 • Debt avoidance • Reduce debts

• Keep my ordinary business borrowing and mortgages below 50% of my farm ‘net worth’

• 7 • Stewardship • Have my family work with me • Pass on a viable business to the next generation

• 8 • Investment • Increase my ‘net worth’ • Make farm investments that will pay for themselves

quickly • 9 • Diversification • Diversify my business by investing both on-farm and

off-farm

The Family Oriented Farmer This group are the most likely, of the five categories considered, to have identified a successor and are most likely to still be farming in five years (62%). However, average age of this category of farmer is the same as the whole group. The group reported the highest percentage attending technical college (42%) but the proportion that has received a university education is below average (17%). The category reported above average economic dependency on the farm, tend to be farming larger than average areas (median 77 hectares) and reported above average annual agricultural sales (mean £122,224). The family oriented farmer group’s opinion regarding the impact the SP would have was equal to the average opinion when the whole sample was considered, only 30% considering it would make a ‘great difference’. Similarly the general attitude (GA) to the SP was similar to the overall sample mean GA. Overall, this group is most likely to have their succession secured and to be more economically dependent on the farm, though farming a larger area than the average. The majority have a tertiary level of education though mainly at a technical level.


108

Business Orientation / Entrepreneur The entrepreneurs’ dependency on the farm is above average but not the highest. They are also more likely than average to have identified a successor for the farm (39%). Just over half (55%) thought they would still be farming in 5 years, equivalent to the overall mean. Table 2: Factors extracted by PCA from ‘value’ statements

• Factor • Factor label • Statements with significantly high factor loadings • 1 • Job satisfaction • Farming allows the expression of special abilities and

skills • Farming gives self-respect for doing a worthwhile job • Farming provides a chance to be creative and original • Farming is about meeting a challenge, forging one’s

character and achieving one’s objectives

• 2 • Marginalisation • Bad press has undermined farmers’ standing in the community

• Local authorities do not understand farmers and their needs

• Local residents are not sympathetic to farmers and their needs

• Central government does not appreciate farmers and their needs

• 3 • Profit and financial

• In running a farm as a business, planning and financial management are the most important parts

• Farming is about maximising profits from the farm business

• Paying attention to details is crucial in making a success of running a farm

• 4 • Staff management

• Farmers should provide congenial working conditions, hours, security and surroundings for

themselves and their staff • Farmers should maintain good relations with staff

• 5 • Technology • To survive in farming, a farmer has to adapt to changing and new technologies

• Survival in farming depends upon being technically efficient

• 6 • Self-reliance • Farming makes one independent, free from supervision and gives one the chance to gain control

in a variety of situations • By choosing to be a farmer, one is expressing a preference for a clear purpose and value in hard work

• 7 • Uncertainty of control

• Farmers have always to bear in mind that any decision they take will affect their farm and their

family • Farming today depends on forces beyond farms’

control, all they can do is to adjust to the situation


109

Table 3 Farmer behavioural types derived from the Defra sample

• Farmer types • n • % of Total

• Family orientation • 202

• 29.6

• Business/entrepreneur

• 177

• 25.9

• Enthusiast/hobbyist • 113

• 16.6

• Lifestyler • 147

• 21.5

• Independent/small farms

• 44

• 6.4

• Total • 683

• 100

Although the median area farmed is similar to the overall average, they tend to be farming better land, i.e. they registered the lowest proportion farming upland (12%). However, the average annual farm income is slightly below average. They are also less likely than average to have received environmental grants in the past year. They have the largest proportion of those over 66 years of age of any of the categories of farmer (25%) and are less likely than average to have received a tertiary education. Their opinion regarding the impact of the SP differs from most other groups in that they have the lowest proportion (11%) who thought that it would have a negative impact. Similarly the group has the lowest proportion expressing a negative attitude to the SP (21%). The entrepreneurs expressed a more positive GA towards the SP than average. The entrepreneurs tend to be older and less educated than the average and still highly dependent on the farm and agriculture, although with a lower income than average. However, they have a positive attitude to the SP while recognising it will make a difference to their business in future.


110

Table 4 Farmer behavioural types with their characteristic features

• Behavioural type • Characteristic features (based on factors from PCA) • Family orientation • High environmental score

• High stewardship • Low marginalisation

• Business/entrepreneur • High quality of achievement

• High expansion • High investment • High independence • High staff management

• High job satisfaction • High debt avoidance • High technical efficient • High marginalisation

• Enthusiast/hobbyist • High quality of life and leisure

• High diversify • High job satisfaction

• Low independence • Low profit and financial

aspects

• Lifestyler • High family standard of life • High quality of life and

leisure • High future security • High staff management

• High uncertainty of control • Low job satisfaction • High marginalisation

• Independent/small farmer

• High family standard of living

• High independence • High job satisfaction

• Low marginalisation • Low quality of life and

leisure • Indifferent to profit and

finance

Note: The use of the terms, "high" and "low" in the context of the formation of these clusters does not mean a measurement along a scale for a particular score, but instead it implies that in any cluster an aspiration, say "job satisfaction", is rated highly as compared to other clusters and thus this observation becomes a defining characteristic of that cluster. Business Orientation / Entrepreneur The entrepreneurs’ dependency on the farm is above average but not the highest. They are also more likely than average to have identified a successor for the farm (39%). Just over half (55%) thought they would still be farming in 5 years, equivalent to the overall mean. Although the median area farmed is similar to the overall average, they tend to be farming better land, i.e. they registered the lowest proportion farming upland (12%). However, the average annual farm income is slightly below average. They are also less likely than average to have received environmental grants in the past year. They have the largest proportion of those over 66 years of age of any of the categories of farmer (25%) and are less likely than average to have received a tertiary education. Their opinion regarding the impact of the SP differs from most other groups in that they have the lowest proportion (11%) who thought that it would have a negative impact. Similarly the group has the lowest


111

proportion expressing a negative attitude to the SP (21%). The entrepreneurs expressed a more positive GA towards the SP than average. The entrepreneurs tend to be older and less educated than the average and still highly dependent on the farm and agriculture, although with a lower income than average. However, they have a positive attitude to the SP while recognising it will make a difference to their business in future.

Enthusiast / Hobbyist This grouping are farming on average only a slightly smaller area (median 67 hectares) than the overall median area registered for the whole sample of 69 hectares. They have the lowest dependency on the farm income, only 44% indicating that they relied on the farm to provide more than 50% of their income. However, their mean annual agricultural sales are only slightly lower than the median of the whole sample. They registered the highest proportion of any category involved in non agricultural enterprises (40%). They also registered the highest proportion receiving environmental grants in the last year (41%). This category of farmer has achieved a higher level academic education than any other category considered, 41% claiming to have gone to university whist only 26% have not had post-secondary education of some kind. Of the five categories, this group is least likely to feel that the SP will make a difference to the way they manage their farms, only 16% indicating that it would make a great difference, whilst 24% felt it would make none. The enthusiast / hobbyist category of farmer tends to be the most educated. The fact that the majority of their income is not tied to agricultural production may lead them to feel that the SP will have less impact on the way they farm. They are also most likely to be taking advantage of available environmental grants and to be involved in non-agricultural enterprises.

Lifestyler This category indicated a comparatively lower dependency on the farm as a source of income, although they register the highest mean income from agricultural sales (£127,320) and claim to be farming on average the largest area (90 hectares). A higher proportion than average is also involved in non agricultural enterprises (39%). Therefore as a category, the lifestylers appear to be the most economically secure, with enterprises diversified between agricultural and non agricultural activities. They tend to be younger, registering the largest proportion (20%) under 40 years of age of any of the five categories. They also tend to be more highly educated than the average. Only 18% claim to have identified or probably identified a successor, the lowest of the five categories. This might reflect the fact that they have the largest proportion still less than 40 years of age. However, only 51% think they will still be farming in five years time, a lower proportion than the overall average. Although their opinion regarding the future impact of the SP on the way they farm does not differ from the norm, they expressed the most negative attitude (GA) toward the SP of all (37% negative).


112

Independent / Small This category of farmer is the most dependent on income from the farm, 37% indicating that they are totally dependent. They tend to farm the smallest area of the five categories (median 40 hectares) and have the smallest mean income from agricultural sales (£49,706) – less than half of the overall average. As a group they also registered the lowest proportion with non agricultural enterprises. They are the group with the highest proportion farming upland. However, they are the least likely to have received an environmental grant in the last year (21%). The group is mainly middle aged, it registered the lowest proportion both under 40 years of age (5%) and over 66 years old (12%). This indicates that the young are not entering this category and the old are not remaining. As a group they also registered the lowest level of educational achievement with only 37% having received a tertiary education. This group was also the most likely not to have identified a successor (50%). However, only 37% thought that they would still be farming in five years, indicating that they are the group most likely to withdraw from farming. Of all the categories, the independent / small category felt that the SP would have the greatest impact on they way they farm. However, they are also the farmers that expressed the most positive attitude toward the SP. Given the high proportion of upland farmers, this may be because they see the SP as offering them a viable route out of an insecure future in farming. The independent /small category of farmer is the least economically secure with the highest dependency on smaller than average farms. They tend to be farming more marginal land and are the group most likely to withdraw from farming. Table 5 shows those characteristics, other than values and objectives, on which there are significant differences between the farmer types. Influences on Farmers’ Behavioural Intentions Regarding SP Attitude to Single Payment Scheme

General Attitude General attitudes towards SP, based on perceptions of the extent to which it is necessary, constructive and helpful, were slightly positive, with ‘necessary’ recording the most positive score. ‘Constructive’ was the only element to register a slightly negative response while ‘helpful’ received an overall neutral score. Around half of all respondents gave a neutral response on all three elements (Figure 1), suggesting that, while farmers on the whole have neither dismissed SPS out of hand nor accepted it with enthusiasm, they are waiting to see how the scheme affects them before coming to a judgement Table 6 shows differences between the five types on the three separate measures and the overall general attitude score. Overall, there is a significant difference between the farmer types regarding the general attitude. The independent /small farmers registered the most positive attitude across the three components that were assessed to form the general attitude, in particular regarding the perceived necessity of the SP. In contrast, the lifestylers registered consistently the most negative or weakest opinions across the three components, the only positive aspect being with regard to perceived necessity. The only component where a significant difference (p = 0.012) was registered was regarding the perceived constructive nature of the SP. In this regard the independent/small farmer was the only farmer


113

type, of the five considered, to register a positive, though only slightly so, attitude regarding the constructiveness of the SP. Although all the other farmer types registered negative opinions regarding this aspect, the lifestyler group registered a relatively stronger negative opinion than the other groups. Interestingly, this measure of general attitude is more sensitive to farmer type than to either type of enterprise or farm size. Given that the lifestylers tend also to be the larger farmers it may have been assumed there would be a significant difference in general attitude to the SP based on the size of holding. However, no significant differences were observed. Also no significant differences were observed between the general attitudes toward SP expressed by different farm enterprise types. This suggests that the general emotive response to the SP is more sensitive to the farmer typology than to either size of holding or type of enterprise. Figure 1 Assessment of general attitude to the SP (whole sample)

Table 6 Mean general attitude to the SP by farmer type

• • All • Family • Business • Hobbyist • Lifestyler • Independent • KW Sig.

• range -2 to +2 • Mean • Mean • Mean • Mean • Mean • Mean • p <0.05

• Necessary • 0.41 • 0.37 • 0.49 • 0.45 • 0.30* • 0.60 •

• Helpful • 0.06 • 0.09 • 0.12 • 0.14 • -0.14* • 0.15 •

• Constructive • -0.10 • -0.03 • -0.02 • -0.10 • -0.32* • 0.05 • 0.012

• General mean • 0.13 • 0.15 • 0.19 • 0.18 • -0.05* • 0.29 • 0.046

Notes: Shaded cells indicate the highest value per descriptor and * indicates the lowest value; only p values <0.05 presented.

0

10

20

30

40

50

60

strongly disagree disagree neither agree nor disagree

agree strongly agree

Perc

en

tag

e o

f re

sp

on

den

ts

necessary helpful constructive

Overall Attitude

SPS is ….

Response


114

Table 5:

Significant

differences

between the identifi

ed farmer

types

Descriptors Family oriented

Business / entrepreneur

Enthusiast / hobbyist

Lifestyler Independent / small

Overall means

P value

Dependency on farm income - 100% 31% 33% 19% 15%* 37% 26%

Dependency on farm income >50% 66% 67% 44%* 54% 70% 50%

0.000

Age (proportion <40) 12% 7% 13% 20% 5%* 13%

Age (proportion >66) 18% 25% 18% 13% 12%* 18%

0.000

Education: University (%) 17% 17% 41% 31% 14%* 24%

Education: Technical college (%) 42% 29% 32% 38% 23%* 35%

Education: Secondary school (%) 42% 54% 26%* 30% 63% 41%

0.000

Successor: (%) of those likely & definitely

42% 39% 20% 18%* 25% 32%

Successor: (%) definitely not 17%* 21% 30% 34% 50% 26%

0.000

Land area: Median hectares 77 70 65 90 40* 67 0.001

Those with non agricultural enterprises 27% 36% 40% 39% 16%* 33% 0.005

Agricultural sales (mean annual value) £122,224 £110,323 £101,937 £127,320 £49,706* £112,275 0.005

Difference SP will make –great difference

30% 31% 16%* 35% 43% 30%

Difference SP will make –no difference 14% 11%* 24% 14% 18% 15%

0.016

Those likely to be farming in 5 years 62% 55% 59% 51% 34%* 56% 0.016

Attitude to SP Positive 39% 44% 46% 33%* 49% 40%

Attitude to SP Negative 23% 21%* 28% 37% 23% 26%

0.034

Land type: those farming mainly or all upland

20% 12%* 16% 14% 35% 17% 0.037

Received environmental grants 34% 26% 41% 30% 21%* 31% 0.043


115

Calculated Attitude The “calculated attitude” (CA) is based on responses to fifteen statements about specific possible impacts or “outcomes” of SPS on farmers and farming. For each outcome, respondents indicated their strength of agreement / disagreement that SPS would lead to the outcome (their outcome belief, (b)) and how good or bad they think this outcome would be (outcome evaluation, (e)), both on scales from -2 to +2. For each statement, the Outcome Attitude (OA) is calculated as (b) * (e) with a range of -4 to +4; and the CA is the sum of the OAs, giving a possible range for the CA of -60 to +60. The CA shows a similar picture to the general attitude, with the mean for various categories based on farmer and farm type, scale, enterprise, etc. falling within a limited range. The overall mean score was slightly negative, while the range in scores for all the different categories was between -13.11 and -2.66. This demonstrates that overall there was little difference in the CAs registered across all the groups. There are small, though statistically significant, differences between farmer types: lifestylers hold more negative attitudes than other farmer types, while independent / small farmers express the least negative attitudes. The most strongly expressed OAs are all negative, resulting from strong agreements with the statements (b) and equally negative attributed values (e). These OAs in rank order according to their strength of expression relate to:

• Loss of national food self sufficiency (-1.94) • Loss of skilled rural labour (-1.87) • Smaller farmers being forced out of farming (-1.79) • Loss of pride due to being seen as park keepers paid for by the state (-1.69) • Reduced long term investments in farming (-1.55).

Intentions to Change Farming System and Practices as a Result of SP

Overall Intentions, Attitudes, Subjective Norm and Perceived Behavioural Control

Theories of reasoned action (TORA) and planned behaviour (TPB) state that the intention (i) to undertake a particular behaviour is the immediate precursor of that behaviour. The survey sample was asked how strong their intention (I) was to change their own farming system and practices as a result of the SP in the next 5 years. The respondents’ attitude (SA) and perceived social pressure (SN) regarding changing their farming system and practices in the next 5 years were also assessed. They were also asked how difficult it would be to make the proposed change and how confident they felt in their ability to make the intended changes to their farming system and practices over the next 5 years. All the responses were measured on bi-polar 5 point scales. The sum of the last two responses, difficulty and ability, was taken to represent the perceived behavioural control variable (PBC). Therefore each variable corresponded regarding proposed activity, its target, the context and the time in which it should take place. An alternative measure of the subjective norm was arrived at by presenting the respondents with a list of 10 salient social referents. They were asked to indicate how motivated (m) they would be to follow the advice of each referent regarding changing their farm system and practices in the next 5 years as a result of the SP. Also the subjective belief (sb) regarding each referent was assessed by asking how strongly they felt each referent would support their adopting the proposed change. The individual referent subjective norms (RSN) were calculated by taking the product of the respective measures of motivation and subjective belief, i.e. (RSNj) = (mj*sbj). The calculated subjective norm is arrived at by taking the sum of the RSNs giving a possible range of -40 to +40.


116

Figure 3 shows the distribution of the main TpB variables for the whole sample. Only 26% of respondents indicated that they intended to change their farming system and practices as a result of the SP in the next 5 years, while the largest proportion, 44%, were still uncertain. Figure 3: TpB variables regarding changing farming systems and practices in the next 5 years (whole sample) Four out of ten feel that those they respect most in farming would be in favour of their changing their farming system and practices in response to the SP, while only 12% indicated they would disapprove. Although 46% are uncertain as to what their most respected others would think, a large proportion do feel a pressure to change their farming system. Differences between Farmer and Farm Types The mean values of the main TpB variables for each of the farmer types are shown in Figure 4. Statistical analysis shows there is no overall significant difference between the five on any of the variables. However comparison between each pair of farmer types does show a significant difference in intention between lifestyler and both family oriented and enthusiast/ hobbyist farmers. The ‘lifestyler’ is the only farmer type to indicate a positive although weak intention to change their farming system and practices over the next 5 years in response to the SP. It is interesting that the lifestyler feels the most able to make that change when compared to the other types. The lifestyler also registered the most positive SN of all the farmer types. The ‘business oriented’ farmer registered the second least negative intention to change (-.05) also registered the second most positive SN of the five farmer types. The ‘independent/ small’ farmer registered a weak negative intent (-0.10) but expressed the least negative attitude (SA). However, this seems to have been offset to some degree by the weakest SN of the five farmer types being recorded by this group. The ‘family oriented’ farmer recorded the second most negative intention (-0.16) to change their farming system during the next five years. This group is noted for registering the most negative PBC (-0.42) of the five farmer types indicating that they consider a change to their farming system and practices most difficult to achieve.

0

10

20

30

40

50

60

-2 -1 0 1 2 Values

% Intention Attitude SN PBC


117

The ‘enthusiast/hobbyist’ type of farmer registered the most negative intention (-0.24) to change of the five types. They also expressed the most negative attitude to this behavioural change. Therefore of the five farmer types, the ‘lifestyler’ is the most likely to change their farming system while the ‘enthusiast/hobbyist’ is least likely to change. Figure 4 TpB variables (means) regarding changing farming system and practices (means) by farmer type

Potential Methods of Using the Single Payment

When the five investment strategies are compared, the use of the SP to ‘substitute’ for the previous production linked payment is the most likely strategy to be adopted as shown in Fig 5. This strategy of, in effect, “recoupling” may be a reflection of many farmers not yet having worked out what SP means for them and waiting to see how the scheme works and what it means for their business. One could anticipate the proportion expressing this intention to fall once the details and implications of SPS become clearer after full implementation. The only other strategy to register a positive intent (I) was the use of the SP as a ‘pension’ or income supplement. Both of these strategies are also supported by positive SA and SN responses. The concept of using it as a supplement / pension registered the most positive SA and SN scores across all the strategies considered. The least likely strategy to be adopted was to invest the SP in establishing non-agricultural enterprises. This non agricultural strategy was also the only one to register a negative attitude (SA).

-0.50

-0.40

-0.30

-0.20

-0.10

0.00

0.10

0.20

0.30

0.40

0.50

Valu

e (

-2 t

o +

2)

Intention (I) Attitude (SA) Subjective norm (SN) Perceived Control (PBC)

Family Independent Lifestyler Hobbyist Business


118

Figure 5: Intention regarding Single Payment investment strategies (whole sample)

The option of investing SP in ‘new agricultural enterprises’ is also unlikely to be adopted. However, in this case a positive though weak attitude was expressed. The concept of using the SP so as to achieve a ‘less intensive system’ also registered a negative intent. However, the most positive attitude was expressed toward this option (equal with the ‘pension’ strategy), though relatively weak. The SN was also slightly positive. This may suggest that the concept of adopting a less intensive system is considered positively, but the idea of investing to achieve this objective is negatively perceived. For four of the five uses of the SP (i.e. all except use as a pension or income supplement), there are significant differences in intent between the five farmer types; only for maintain current farming system is there a difference between farm types. For attitude and social norm also, there are more differences between farmer types than farm types. Salient Referents and Subjective Norms

Table 7 presents the motivation, subjective beliefs and resulting referent subjective norms (RSN) regarding each of the salient social referents identified in the focus groups. The calculated subjective norm (CSN) is neutral to slightly positive (mean 6.97 on a -40 to +40 scale). It does not correlate significantly with the intention to change nor with the stated subjective norm SN. Neither are any significant (I vs. RSN) correlations observed with correlation coefficients greater than 0.2. For the whole sample, respondents are most motivated to comply with farmers clubs, accountants and family regarding changing their farming system and practices in the next 5 years. They are least motivated to comply with land agents, Defra and consultants and, surprisingly, ‘other farmers’.

-1.00

-0.80

-0.60

-0.40

-0.20

0.00

0.20

0.40

0.60

Less intensive

Valu

e -

2 t

o 2

Intention (I)

Attitude (SA)

Subjective norm (SN)

New Agric Enterprise

Non Agric Enterprise

Substitution Pension


119

Table 7: Subjective norm with respect to changing farming system and practices as a result of the SPS over the next 5 years

• n=676 • Motivation to comply (m)

• (-2 to +2)

• Subjective belief (sb)

• (-2 to +2)

• RSN

• (-4 to +4)

•

• Correlation between

• RSN & intent (I)

• Social referents (RSN -4 to +4)

• mean • mean • mean • rs

• Farmers clubs • 0.75 • 0.79 • 0.42 • ns

• Accountants • 0.39 • 0.54 • 0.70 • ns

• Family • 0.24 • 0.25 • 1.14 • ns

• Farming press and literature • 0.14 • 0.36 • 0.45 • ns

• NFU • -0.12 • 0.00 • 0.73 • ns

• Business partners • -0.18 • 0.14 • 0.78 • ns

• Consultants • -0.41 • 0.01 • 0.80 • ns

• Defra • -0.45 • -0.03 • 0.85 • ns

• Land agents • -0.46 • -0.07 • 0.63 • ns

• Other farmers • -0.89 • -0.26 • 0.51 • ns

• Own experience • 1.67 • 1.08 • 1.67 • ns

• CSN (-40 to +40) • • • 6.97 • ns

• Alpha coefficient • • • • 0.812

With respect to the issue of changing farming system and practices due to the SP, therefore, the farmers in the sample clearly do not like to comply with those sources which one might expect to have the most accurate information regarding the SP. This could mean that they feel challenged by these referents to go against their own desires regarding future farming practice. Their negative inclination to comply with other farmers, other than those with whom they associate closely in farmers clubs, is interesting as it suggests that there is an underlying recognition that their peers are similarly ignorant regarding the future consequences of the SP. A possibility to explore in future research is that respondents are implicitly distinguishing between referents who are sources of information and knowledge, and those who they regard as sources of guidance, wisdom and support. The strongest subjective beliefs are expressed regarding the perceived opinions of farmers clubs, accountants, the farming press, other farmers and family. Of these only other farmers are believed not to support a change in farming system and practices. A positive mean RSN was recorded for all referents. The strongest RSNs in rank order were for the family, Defra, consultant and business partners. For family referents, the high positive RSN comes from a generally positive motivation to comply and positive subjective belief. In the case of Defra, it is the result of a negative to neutral subjective belief (i.e. most respondents think either that Defra does not want them to change or they do not know) and a negative


120

motivation to comply. The NFU also registered a stronger than average RSN. Although none of these RSNs when correlated with intent produced a correlation coefficient greater than 0.2, two (I vs. RSN) correlations were significant at p<0.01: those for consultants and business partners. Across all the categories these two referents appear to be the most commonly influential, i.e. demonstrating significant (I vs. RSN) correlations. It is interesting to note from the final row of Table 7 that the highest positive (m) and (sb) values are attributed to following their ‘own experience’. However, the resulting RSN does not correlate significantly with their intent. This also suggests that farmers may wish to change, given the strong positive (sb) registered in this case but that other factors such as their stated attitude and perceived behavioural control regarding changing their farming system may subdue this underlying desire. Table 8 compares referent subjective norms for the five farmer types derived from the main survey. For five of the ten referents, there are no significant differences in RSN between farmer types, indicating that these referents’ influence is similar across all types. Among these five, the family registers the highest RSN for all farmer types. Farmers clubs have a higher degree of influence on the independent / small farmer type than the others, and accountants have a lower influence on the hobbyist than on others. The independent / small farmer type also seems to be more influenced by the farming press and literature than other farmers, an observation which seems to underline their independence. Table 8: Referent subjective norms (means) by farmer type

• • Farmer type

• Social referent • Family • Business • Hobbyist • Lifestyler • Independent

• • RSN (m*sb)

• RSN (m*sb)

• RSN (m*sb) • RSN (m*sb) • RSN (m*sb)

• Farmers clubs • 0.39 • 0.40 • 0.37 • 0.37a • 0.98a

• Accountants • 0.78a • 0.65 • 0.46a,b • 0.77b • 0.86

• Family • 1.20 • 1.12 • 0.98 • 1.24 • 1.02

• Farming press and literature

• 0.51 • 0.38a • 0.42 • 0.30b • 1.02a,b

• NFU • 0.76a • 0.76b • 0.52a,b • 0.72 • 1.02

• Business partners

• 0.80 • 0.69 • 0.55a • 0.98a • 0.91

• Consultants • 0.81 • 0.88 • 0.59 • 0.84 • 0.86

• Defra • 0.81 • 0.92 • 0.68 • 0.89 • 0.98

• Land agents • 0.59 • 0.65 • 0.49 • 0.67 • 0.93

• Other farmers • 0.53 • 0.59 • 0.41 • 0.49 • 0.50

Note: pairs of farmer types with the same superscript letter in a row have significantly different (p<0.05) referent subjective norms for the given social referent. Influences on Behavioural Intentions

Correlation between Intent and Other Tpb Components


121

For the general behaviour of changing farming system or practices in the next five years as a result of SP, the SA, SN and PBC for each farmer type correlate significantly with stated intent (I). The (I vs. SA) is the dominant correlation in each case followed by the (I vs. PBC) correlation in all cases but the ‘enthusiast/hobbyist’ (Table 9). The ‘hobbyists’ appear to be more likely to take into account the opinions of their ‘respected others’ than the difficulty posed by the change and their ability to achieve it - the PBC. There are similar significant correlations for each of the six farm types, with the exception of specialist cereals where only SN correlates significantly with intent. Except for specialist cereal growers, all farm and farmer types show a stronger correlation between intent and attitude than between intent and subjective norm. This implies that the respondents’ decisions regarding changing farming system and practices in response to the introduction of the SP will be governed more by their own experience and values than by perceived social pressure. For beef and sheep, general cropping and ‘other’ farm types, perceived behavioural control correlates with intent more strongly than attitude. As the mean PBC for these farm types is negative, this suggests that their perception that it would be difficult for them to make a change will be more influential than their attitudes or social pressure. For the five investment strategies, only in the case of using the SP as a pension or income supplement did the SN produce a stronger correlation with intent than the SA at the level of the whole sample (Table 10). This suggests that in this one instance the farmers are more likely to be influenced by the opinions of their respected referents over and against their own experience. Though this is the most likely strategy to be adopted, the influence of the SN in this case indicates a degree of uncertainty. This is only natural given that the consequences of the SP were still unknown at the time of the survey. Therefore the stronger positive intent expressed regarding this strategy reflects a tendency to ‘stick with the known’ – rather than accepting the need for change. This could mean that the SP could initially have a detrimental impact on farmers until the need for adjustment is recognised and an alternative strategy adopted. Table 9 Correlations of TpB variables with intent to change system: whole sample and farmer types

• Main TpB variables

• All • Family

• Business

• Hobbyist

• Lifestyler

• Independent

• n • 674 • 200 • 172 • 112 • 146 • 44 • • rs • rs • rs • rs • rs • rs • Attitude (SA) • .48

3** • .41

2** • .587

** • .48

7** • .438

** • .537**

• Subjective norm (SN)

• .325**

• .318**

• .276**

• .371**

• .327**

• .377*

• Perceived Behavioural Control (PBC)

• .403**

• .361**

• .552**

• .217*

• .381**

• .461**

** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed).

For the five farmer types, however, the relative influence varies, with the business / entrepreneur, hobbyist and independent / small types recording higher (I vs SN) than (I vs SA) correlations for four, one and two respectively of the five strategies, as shown in Table 10). Overall, the strong correlations observed between intention and the other TpB variables suggest that stated intent with respect to the six behavioural responses to SPS is a reliable predictor of actual future behaviour. However, on a lot of the measures, large proportions of farmers gave neutral or non-committal responses, indicating a continuing degree of uncertainty about how the scheme will impact on their farm and on farming in general.


122

Table 10 Correlations of TpB variables with intent to use SP in various ways, by farmer type • TpB

variables • All • Family • Business • Hobbyist • Lifestyler • Independent

• • rs • rs • rs • rs • rs • rs • Use SP as pension or income supplement • Attitude

(SA) • .462** • .467** • .380** • .570** • .481** • .458**


• .477** • .445** • .458** • .549** • .467** • .529**

• Use SP to compensate for loss of previous subsidy, to maintain current farming system

• Attitude (SA)

• .488** • .489** • .371** • .598** • .624** • ns


• .470** • .403** • .442** • .474** • .562** • .465**

• Use SP to invest in non-agricultural activities • Attitude

(SA) • .514** • .515** • .423** • .545** • .558** • .511**


• .428** • .432** • .429** • .427** • .404** • .485**

• Use SP to develop new farming enterprises • Attitude

(SA) • .517** • .567** • .334** • .520** • .571** • .602**


• .486** • .427** • .439** • .599** • .478** • .524**

• Use SP to make farming system less intensive

• Attitude (SA)

• .646** • .540** • .677** • .738** • .612** • .797**


• .529** • .399** • .562** • .579** • .529** • .704**

The above analysis shows distinct patterns of response by the five behavioural types in respect of SPS. The behavioural typology is potentially useful for policy analysis, complementing the other typologies that differentiate the farming population on scale, enterprise and economic status. Concluding Observations This project leads to make the following observations: • economic drivers are not necessarily paramount for all farmers - environmental, family, lifestyle and

stewardship motives are equally, and sometimes even more, important for many farmers

• these non-economic drivers are long term goals while the economic drivers in current policy models reflect shorter term objectives - more research is needed on how to integrate these into a common modelling framework

• identifying different behavioural types leads to more accurate predictions about responses to specific policy changes

• to induce change in behaviour, a blanket “one size fits all” policy is not appropriate - different farmers will respond to a new policy initiative in different ways


123

• uncertainty engendered by a new policy makes it difficult for farmers to plan how to adapt to policy change.

References Ajzen, I. and Fishbein, M. (1980). Understanding Attitudes and Predicting Social Behaviour. Englewood

Cliffs, New Jersey: Prentice Hall. Ajzen, I. (1988) and (2005). Attitudes, Personality and Behaviour. Milton Keynes: Open University

Press. Burton, R. J. F. (2004) Re-conceptualising the ‘behavioural approach’ in agricultural studies: a socio-

psychological perspective. Journal of Rural Studies 20:359-371. Edwards-Jones, G., Deary, I. and Willock, J. (1998). Incorporating psychological variables in models of

farmer behaviour: does it make for better predictions? Etud. Rech. Syst. Agraires Dév. 31:153-173.

Fairweather, J.R. and Keating, N.C. (1994). Goals and management styles of New Zealand farmers.

Agricultural Systems 44: 181-200. Lynne, G. D. (1995). Modifying the neo-classical approach to technology adoption with behavioral

science models. Journal of Agricultural and Applied Economics 27:67-80. Acknowledgement The authors are grateful to Defra for the financial support for the research report here for their project EPES 0405/17: Research to Understand and Model the Behaviour and Motivations of Farmers in Responding to Policy Changes (England). Appendix Statements on values and objectives use as Question 13 and 14 of the Questionnaire “what you are trying to achieve as a farmer”: (a) “not at all important” ….. “most important” (on nine point rating scale)

1. Produce the best quality output on my farm 2. Be the best farmer I can be 3. Contribute to the farm in order to achieve something 4. Reduce work load and improve quality of life 5. Diversify my business by investing both on-farm and off-farm 6. Concentrate on farm work and not be sidetracked by outside activities 7. Be sensitive to the environmental impacts of farming by reducing input use on my farm 8. Do everything to be environmentally aware 9. Have my family work with me 10. Buy more land 11. Rent/contract more land


124

12. Avoid borrowing money 13. Reduce debts 14. Make more time to spend on activities away from the farm 15. Increase my ‘net worth’ 16. Keep my ordinary business borrowing and mortgages below 50% of my farm ‘net worth’ 17. Invest part of my profits for retirement 18. Save for children’s education 19. Make farm investments that will pay for themselves quickly 20. Increase family income 21. Maintain my family’s standard of living at its current level 22. Improve my family’s standard of living 23. Gain recognition among my peers 24. Be my own boss 25. Pass on a viable business to the next generation

(b) “do not agree at all” … “agree completely” (in nine point rating scale) 1. In running a farm as a business, planning and financial management are the most important parts 2. The present level of development of my farm is satisfactory and I do not intend to develop it further 3. Farming is about maximising profits from the farm business 4. Farm work is a chore and it has no joy 5. Paying attention to details is crucial in making a success of running a farm 6. Farmers have always to bear in mind that any decision they take will affect their farm and their family 7. Farming today depends on forces beyond farmers’ control, all they can do is to adjust to the situation 8. Farm work and tasks must come before family obligations 9. Working with nature is difficult and unrewarding 10. Farming is not just about making maximum profit 11. Farmers should provide congenial working conditions, hours, security and surroundings for themselves and their staff 12. Farmers should maintain good relations with staff 13. Farming allows expression of special abilities and skills 14. Farming gives self-respect for doing a worthwhile job 15. Farming provides a chance to be creative and original 16. Farming is about meeting a challenge, forging one’s character and achieving one’s objectives 17. To survive in farming, a farmer has to adapt to changing and new technologies 18. Farming makes one independent, free from supervision and gives one the chance to gain control in a variety of situations 19. By choosing to be a farmer, one is expressing a preference for a clear purpose and value in hard work 20. Survival in farming depends on being technically efficient 21. Being a farmer I am a respected member of local community 22. Bad press has undermined farmers’ standing in the community 23. Farmers should promote farming interests more actively 24. Local authorities do not understand farmers and their need 25. Local residents are not sympathetic to farmers and their needs 26. Central government does not appreciate farmers and their needs


125

DAIRY FARM BUSINESS ANALYSIS; CURRENT APPROACHES AND A WAY FORWARD

Nicola Shadbolt IFNHH College of Sciences, Massey University, Palmerston North, New Zealand


Matthew.Newman Dexcel Ltd

Ivan.Lines

Agribusiness Ltd

Abstract In 2003 a voluntary-based industry group, calling itself the KPI Working Group, formed to discuss and address the fragmented approach to measurement of business performance that existed in the dairy industry. The objective they set themselves was to develop a coordinated approach to provide sound, robust data and consistent benchmark calculations which would provide increased clarity of data for the dairy industry and benchmarks that could be relied upon. Group discussion related to the need to provide farmers and wider industry players with timely information on liquidity, profitability and wealth creation/loss as it occurs on farm from year to year. Critical areas that required consistency in how they were determined included the value of family labour and management, changes in feed inventory and the value of land and buildings. Indicators of success for both the property and the farming businesses was needed to ensure a holistic evaluation was made of overall investment strategy. The research provides a useful example of how inter-disciplinary groups can work towards a common goal and suggests a framework for farm analysis that could be used internationally. Keywords: liquidity, profitability, wealth creation, dairy farm analysis Introduction In 2003 a voluntary-based industry group, calling itself the KPI Working Group, formed to discuss and address the fragmented approach to measurement of business performance that existed in the dairy industry. It was recognized that not only was the data fragmented and not always robust there were also inconsistencies in both terminology and calculation of key performance indicators (KPIs). The objective they set themselves was to develop a coordinated approach to provide sound, robust data and consistent benchmark calculations which would provide increased clarity of data for the dairy industry and benchmarks that could be relied upon. The purpose of this research was to document and define the variety of methods used to analyse farm businesses that existed, both in New Zealand and overseas, and to determine through group consensus the method and the indicators that would be most beneficial to all stakeholders in the New Zealand dairy industry. The methodology included a review of both the literature and current practice amongst rural professionals to define the methods used and how they delivered to common business and industry goals. This documentation of the various approaches was followed by rigorous debate and discussion by the group to determine the indicators of most relevance to the dairy industry. This paper summarises that documentation and presents the results of the group consensus.


126

History The voluntary working group consisted of representatives from the NZ Institute of Chartered Accountants, Dexcel, NZ Institute of Primary Industry Management, Massey University, Fonterra and trading banks. It is of interest first to note how industry standards have developed in New Zealand and what role the various organizations have played in this development.

The NZ Institute of Chartered Accountants (formerly New Zealand Society of Accountants) has always played an active role in farm management accounting. McEwen (1965) documents the process by which an Agricultural Development Conference resulted in the following recommendations:

1. That the NZ Society of Accountants (NZSA) convene a committee to revise the form of accounts and code of terminology in the 1961 Research Report of Farm Accounting to provide forms for use by the farmer to record essential management and financial information during the year.

2. That to ensure the widest possible adoption of the recommendations regarding minimum standards for farm accounting a publicity campaign among farmers and accountants be sponsored by the NZ Society of Accountants, the Government Producer boards, Federated Farmers and others including lending institutions and farm improvement clubs.

At that time NZ’s 73,000 farmers earned over 90% of NZ’s total overseas earnings and it was noted with concern how there was a serious lack of information on the economic aspects of farming.

The result of implementing the above recommendation was the publication of Farm Accounting in New Zealand (commonly referred to as the “Green Book”) in 1968 in which an agreed chart of accounts was presented as were recommended formats for various accounting reports including a cash flow statement. It is of interest to note that this publication outlines an Economic Farm Surplus statement as the method by which to provide comparison between one farm with another and between years on the same farm. The publication recommends three major reporting statements as critical to business analysis:

� The farm working account (known now as the Statement of Financial Performance)

� The Cash Flow Statement

� The Economic Farm Surplus

In the preface to this NZSA (1968) publication it is stated “…no longer is it sufficient for the accountant to produce only historical records and taxation returns – he must be looking ahead and fulfilling his role as his client’s financial adviser”. It also notes how the changeover to decimal currency in 1967 and the increasing use of computers “…presages a climate of change and progress and the need for more precise planning of farming operations”.

In 1977 the NZSA produced a subsequent publication “Management Accounting for the New Zealand Farmer” (NZSA, 1977). In this they stated that accounts prepared on a purely historical cost basis are misleading to the user and that there was an increased emphasis on accounting to provide information for business management essential to sound decision making. They recommended a move away from accounts drawn up largely for tax assessment and the adoption of net current values for assets and the abandonment of tax values for livestock. The committee preparing this work drew heavily on work completed by the Queensland (Australia) Joint Committee on Standardization of Farm Management Accounting. Of interest is the absence of the Economic Farm Surplus as a recommended key measure and the emphasis on budgeting (cashflow, partial, parametric and gross margin) and enterprise accounting. Many of these concepts draw on the economic approach of separating variable and fixed costs espoused


127

in the farm management literature from Australia, the UK and the USA at that time that was used to determine optimal enterprise combinations on mixed enterprise farms.

A subsequent NZSA publication ‘Financial reporting for Primary Producers’ was produced in1989 to update members on the continuing changes in financial reporting requirements (Clarke, 1989). Its purpose was to recommend accepted accounting principles for primary producers ‘..with a view to providing guidance on financial reporting and valuation policies and techniques for primary producers and their financial advisers’. Again a sample set of statements is presented including cash flows but no chart of accounts is included this time and, again, no mention is made of Economic Farm Surplus. It presents financial reporting as being primarily historical but suggests a sound accounting and financial reporting system provides a greater degree of precision that will enable better assessment of unprofitable areas and areas where economies can be made. 1984 was when subsides were removed from NZ agriculture so it is not surprising that it suggests producers’ ability to make sound financial decisions as becoming increasingly more important as they deal with variable input costs and volatile market conditions, debt levels and interest rates. Similarly experience in the US during the ‘Farm Debt Crisis’ years of 1983 to 1987 pointed out that methods used at that time to determine, measure and analyse the financial position and financial performance of agricultural producers were either totally inadequate of seriously underutilized. (FFSC, 1997).

A consistent theme throughout these publications has been the recommendation that accountants produce a cashflow statement in conjunction with other financial statements but this has never become a legislated requirement. McEwen (1965) identified the cashflow statement as a restatement of the accounts in the form of total sales and expenses ignoring the profit concept of accounting; he believed it was in the cashflow form that his farmers thought about finances. He also pointed out how the farm budgets used are simply a projection of the cashflow statement for the following year so providing a cashflow statement of the year that has been assists in the farmer’s projection of the year to come. Clark (1989) defines the task of the cashflow statement is to provide information about the operating, financing and investing activities of an entity and the effects of those activities on cash resources.

However despite this early work and subsequent recommendations by the NZSA Angus (1991) identified that the conventional presentation of accounts was still failing to communicate clearly a meaningful cash result. Angus (1991) states that while most farming clients are well served by their accountants in the area of legitimately minimizing tax the “simple objective of defining if earnings exceed spending has been lost sight of”.

Since 1965 a dedicated group of farm accountants has developed in NZ; this group has put in to practice many of the recommendations of the various NZSA publications and many of them have also developed various forms of benchmarking for their clients, analyzing the cash result, the profitability and the equity change of their clients and comparing each result with group averages.

In parallel with these developments in the accounting profession and perhaps because of them other rural professionals have also developed various methods of financial reporting. Bankers tend to focus very closely on the cash position of their clients, often using Change in Net Indebtedness (fixed plus current liabilities less current assets) as a key measure. They link this to changes in stock numbers and capital purchases to determine if their clients risk status has changed. They also monitor asset values to determine client debt to asset ratios and, inversely, the increasing or reducing risk of their lending portfolio. Students targeting a banking career have traditionally been expected to have both farm management and valuation qualifications and registration to enable such valuations to be carried out.

Farm consultants commonly assist farmers with their cash budgets so also require details on the cash position of previous years. In the absence of meaningful cashflow statements both they and farm financiers must complete accounts analyses (cash reconciliations) to determine historical cash results from which to base or compare projections. As farm consultants are also often involved in benchmarking


128

for a group of clients they have tended to calculate economic farm surplus (various versions based on the NZSA (1968) recommendation) and other efficiency ratios (Return on Assets, Return on Equity and various per hectare, per stock unit and per kg output measures).

The Ministry of Agriculture developed FMAS and provided an accounts analysis service with a mainframe computer throughout the 70s before personal computers and spreadsheets made it redundant. They also provided pre-coded sheets for manual cash books to farmers that were based on the NZSA 1968 recommended chart of accounts. The analysis provided by FMAS and subsequent farm extension and consultancy software programmes provided liquidity, profitability and efficiency measures. Over time the definitions of such measures altered at the whim of the people involved and the connection with a common standard or definition was lost. Their varying academic backgrounds (accountancy, farm management or valuation) largely determined the emphasis they placed on liquidity, profitability, efficiency, taxation and equity and the reliability and accuracy of each calculation.

In the US the Farm Financial Standards Council was established in 1989 in order to develop some standardization in financial reporting and financial analysis. The first edition of their report ‘Financial Guidelines for Agricultural Producers’ was issued in 1991. In it they recommended a list of measures that addressed liquidity, solvency, profitability, repayment capacity and financial efficiency. They made the distinction between net income (taxable income) and operating profit (economic farm surplus) and defined the latter as including an estimated value for family labour and management (FFSC, 1997).

Boehlje(1994) defined operating profit and operating profit margin as critical measures of revenue generation and cost control and added further measures for reinvestment rates and cost containment. Using the Du Pont business model as his base he emphasised the three drivers that impact bottom line performance, as measured by return on investment equity, as operating profit margins, capital turnover and leverage. Each of these drivers are affected by specific decisions on cost control, efficiency and productivity, as well as marketing choices, business structures and management systems.

Operating profit, often termed Economic Farm Surplus in New Zealand, is calculated for both dairy and sheep and beef cattle farms in annual statistics collected by the respective industries (The Economic Survey, 2006, Sheep & Beef Economic Survey, 2006). In Australia it is termed Profit at full Equity and is available for broadacre and dairy farms from ABARE (2005).

The Process

Despite the wide range of measures and definitions used by the various members of the group and a high level of ‘patch protection’ the group made good progress in the first 12 months deciding on key performance indicators and their standardisation. Most members of the group provided a type of benchmarking service to their clients in which considerable investment had been made in data collection, analysis and interpretation. However all members saw the benefit in pooling their skills and the farm data to enable a national service to be developed. In October 2004 funding for the project was granted by Dairy Insight. This allowed the working group to proceed with the development of the software, the web interface, the reports and database systems and procedures to establish DairyBase. The buy-in and contribution from all members of the group has been the key reason for the project’s success to date. Ultimately the project will only be successful if rural professionals use the database and adopt the calculations and terminology as the industry standard. It is critical that the benchmarks are produced from a system which has integrity and will allow meaningful comparisons. The group determined that integrity resulted from having trained individuals entering standardised and verified data that meets specified quality standards. The volume of data, or number of data sets entered from different


129

farms, must be high enough to ensure an accurate representative sample. The target was to process 1,800 dairy farm businesses in the 2006 year, building to 5,000 sets of accounts by 2010. The key objectives of DairyBase are to:

• Standardise terminology, calculations and reporting of key KPIs.

• Provide sufficient volumes of reliable data for farm comparisons

• Develop a National Database for the dairy industry that will provide robust national and regional data for different farm types. This includes producing an annual publication of industry trends.

• Provide improved aggregate data to measure industry progress and for R&D purposes

Accredited rural professionals enter farm physical and financial data. It is anticipated that accountants will enter most of the data as they finalise each year’s Annual Financial Statements. If accountants do not enter the data it can be entered by accredited consultants or bankers. Rural professionals are be permitted to enter data without authorisation from the farm business owner. The farm business owner is able to authorise any one or more rural professionals to enter data into the system. The initial data is entered over the internet to a validation or scratch pad area. Once the data has been validated or passed through a series of checks it is transferred into the actual database. Reports are generated after data has been validated and committed to the database. The reports produce data for the individual farm business and the data for a chosen benchmark group. A sample of the available reports is attached as Appendix 1. Reports of aggregate (not individual farm business) data will be made available to industry bodies as requested. Market research carried out at the commencement of this project confirmed that a National Database for the dairy industry to provide information to industry for research & development and planning purposes, and also provide a basis for benchmarking, was supported by the majority. Figure 1: The Dairy Base process

Farm Business

Owner

Rural Professionals

Training & Accreditation

Obtain Data

Industry

Database

Data Entered into Scratchpad

Validate & Commit

ReportsAggregates


130

Level One Physical and Financial Reports These reports focus on a physical summary then key performance indicators in the three critical areas: Cash (liquidity) Profit Wealth creation. The emphasis on cash noted by McEwen in 1965 is as valid today for many farmers and is an essential financial management skill at the operational level (Shadbolt & Gardner, 2005). The focus on profit and efficiency includes the operating profit, return on assets and return on equity as well as the key Du Pont drivers of operating profit margin and asset turnover. Results are stated also on a per hectare, per cow and per kg milksolid. Delivery to these measures is the result of good financial management at the tactical level as managers make revenue generation and cost control decisions as the season unfolds. Wealth creation is recognized as a key financial outcome at the strategic level for many farm businesses and is reliant on a realistic estimate of asset values at opening and closing (Shadbolt & Rawlings, 2001). The important distinction is also made between wealth created from profit retained and invested in the business and that achieved as a result of changing asset (land and shares) values. Various solvency and debt servicing capacity measures are also included to ensure the vulnerability of the business is understood. More in depth ‘Level Two’ physical data can also be collected to provide more in-depth analysis of the farming system. Summary Essentially discussion related to the need to provide farmers and wider industry players with information on liquidity, profitability and wealth creation/loss as it occurs on farm from year to year. No one area was more important than another and each provided relevant information useful for both off- and on-farm decision making. Critical areas that required consistency in how they were determined included the value of family labour and management, changes in feed inventory and the value of land and buildings. Indicators of success for both the property and the farming businesses was needed to ensure a holistic evaluation was made of overall investment strategy. A timely method of ensuring the analyses were carried out as close to the end of the financial year as possible and to provide comparisons with chosen benchmark groups was also devised. The research provides a useful example of how inter-disciplinary groups can work towards a common goal and suggests a framework for farm analysis that could be used internationally. References ABARE (2005), Australian Bureau of Agriculture and Resource Economics, Australian Broadacre

Agriculture. Agricultural and Grazing Industries Survey, ABARE, Canberra. Angus (1991) The Cashflow Statement in Farm Accounts – Lets make it meaningful. A project in

fulfilment of the requirements of the NZ Society of Farm Management Study Award. Boehlje, M. (1994). Evaluating farm financial performance. Journal of the American Society of Farm

Managers & Rural Appraisers 58 (1):109-115. Clarke, M. (1989) Financial reporting for Primary Producers; for the Primary Sector Accounting Sub-

Committee. Published by The New Zealand Society of Accountants, Wellington, NZ


131

FFSC (1997) Financial Guidelines for Agricultural producers. Recommendations of the Farm Financial

Standards Council. (Revised) December, 1997. Gardner J., Shadbolt N.M. (2005) Financial Management. In: Farm Management in New Zealand. Editors

Shadbolt N.M. & Martin S. Oxford University press, Melbourne. McEwen (1965) Farm Management Accounting. An address presented to the New Zealand Society of

Accountants National Convention, Christchurch, NZ, March 19th, 1965. NZSA (1968) Farm Accounting in New Zealand. Prepared by The Farm Research Committee of the New

Zealand Society of Accountants’ Board of Research and Publications. Published by The New Zealand Society of Accountants, Wellington, NZ

NZSA (1977) Management Accounting for the New Zealand Farmer. Published by The New Zealand

Society of Accountants, Wellington, NZ Shadbolt, N.M., Rawlings M. (2001) Successful benchmarking by balanced planning and identifying key

performance indicators for goal attainment in dairy farming. DRDC Australia Project MUNZ001. Shadbolt N.M. (1997) Key Performance Indicators. Massey Dairy Farmers Conference, Palmerston

North. The Economic Service (2006) The New Zealand Sheep and Beef Farm Survey. The Economic Service of

New Zealand, Wellington. Dexcel (2006) Economic Survey of New Zealand Dairy Farmers 2004-2005. Dexcel Ltd, Hamilton. Appendix I: Sample DairyBase reports


132


133


134


135


136

EUROPE’S MANSHOLT PLAN FORTY YEARS ON

David R. Stead Agriculture and Food Science Centre, University College Dublin,

Belfield, Dublin 4, Ireland. Email: [email protected]

Abstract Forty years ago, Europe’s Agriculture Commissioner Sicco Mansholt headed a unique attempt to transform the rural economy of the then six-member European Economic Community. The Commission published a provocative memorandum proposing policies to accelerate structural change in agriculture, including providing financial incentives to encourage about half of the farming population to leave the sector. Unsurprisingly ‘the Mansholt Plan’ produced passionate protests, and the Directives eventually adopted were far less ambitious than the original proposals. The fate of the Mansholt Plan is a classic example of the difficulties faced in overcoming the status quo bias of agricultural policy. Juxtaposing the plan with today’s policy environment highlights the dramatic changes that have occurred in the European agricultural sector, and the re-orientation of the CAP. The counterfactual scenario of wholesale implementation of the 1968 memorandum programme would, in hindsight, have almost certainly featured repercussions and recriminations exacerbating the problems of European integration. Hence the case study of the Mansholt Plan also looks like a cautionary tale against introducing ‘big bang’ policy reform in the future. Keywords: Agricultural policy reform, CAP, Mansholt. Introduction The year 2008 will mark the fortieth anniversary of the unique attempt to transform the European rural economy by the Commissioner for Agriculture Dr Sicco L. Mansholt. The approaching anniversary is an opportune moment to briefly reflect upon Mansholt’s Plan, and the lessons its failure has for agricultural policy today − including the appropriate pace of future CAP reform, which has been the subject of ongoing academic debate (e.g. Ackrill and Kay, 2006; Harvey, 2006a, b) as the 2008 ‘Health Check’ of the CAP draws near. Halving the numbers in farming It is salutary to recall that the CAP’s erstwhile open-ended system of price support, which despite high costs did not resolve the problem of low incomes for small farmers (Baldwin and Wyplosz, 2006), came under serious criticism from its very inception. In December 1968, the Commission of the then six-member European Economic Community published a provocative memorandum which proposed policy principles to greatly accelerate structural change in agriculture (CEC, 1968). Although many aspects had been trailed in advance (e.g. Mansholt, 1968), the blueprint still sent shockwaves through countryside communities. The key recommendations of what was immediately dubbed ‘the Mansholt Plan’ included: Co-financed monetary incentives to encourage about half of the farming population to leave the sector during the 1970s by taking early retirement or retraining to obtain alternative work, ideally locally (the latter to be facilitated by regional job creation schemes); Much of the land thereby released to be added to farms with approved development plans for expansion to a specified minimum efficient scale; Investment aid to be granted only to such farms, or to farmers who chose to combine with others to create a large jointly-managed holding;


137

A target of removing at least five million hectares from agricultural production and using the land largely for forestry; Slaughter premiums and other payments to reduce the Community’s dairy herd by three million head. These policy interventions, the Commission argued, would resolve the problems of commodity surpluses, free up institutional prices to be set more in line with demand and costs, support non-viable farmers who would have to leave the industry anyway, and ensure that those remaining obtained a quality of life comparable to that available in other occupations. ‘I do not think we have any alternative’, asserted Mansholt.1 ‘The peasant killer’

Unsurprisingly the plan to convert the Community’s traditional structure of small farms into larger, more efficient units within ten years produced passionate protests. Farmers, some erroneously fearing the imposition of Soviet-style collectivism, demonstrated against Mansholt ‘the peasant killer’. Equally unsurprisingly, national governments – especially West Germany – generally had strong reservations on the grounds of cost, even though the memorandum promised an eventual reduction in FEOGA spending, and also on the grounds that structural policy was their remit, not the Community’s. Practically only young farmers’ groups were wholeheartedly supportive. The European Council of Ministers took the line of least resistance by barely discussing the memorandum. Mansholt shrugged off ‘the most vulgar accusations’ and worked hard to persuade his opponents, likening the annual cost of the CAP to what the US was spending on the moon landings. But the first detailed legislative proposals issued in April 1970 showed that the Commission had realised the need to sharply rein in their ambitions. After more debate, delay and protest, with Mansholt linking a decision on structural policy to urgently-required decisions on price levels, in May 1971 the Council resolved on an even more diluted structural reform package. ‘The original proposals have undergone such radical and extensive changes’, wrote one public official, ‘that the “Mansholt Plan” is today a misnomer for the reform programme now proposed’.2 The three Directives eventually adopted eleven months later − comprising provision of financial assistance for farm modernisation, for the early retirement of farmers aged 55 to 65, and for training and advice (Directives 72/159-61) − were even further reduced in scope, with the total commitment from FEOGA just a ninth of the sum first suggested. Hence this so-called ‘mini-Mansholt’ had very little impact in leading structural change (see, especially, Neville-Rolfe, 1984; Fennell, 1997). Reflection Political economists use the fate of the Mansholt Plan as a classic example of the difficulties in achieving significant CAP reform in the absence of strong external pressure from Europe’s trading partners (e.g. Fennell, 1997). The Commission tried shock tactics to overcome the well-known status quo bias of agricultural policy (‘the only way in which a full and thorough-going debate will be generated is by posing the problem starkly and by suggesting root and branch solutions’), but even this strategy did not elicit deep reform.3

1 Interview with Corriere della Sera, 20 February 1969, translated by Centre Virtuel de la Connaissance sur l’Europe, www.ena.lu [last accessed 15 February 2007]. 2 This subsection, and the previous one, are based on the sources cited in the list of references given below, together with the following recently-released government documents: National Archives of Ireland, Department of Foreign Affairs files 2003/1/199, 2003; UK National Archives, Foreign and Commonwealth Office files 30/317-8. 3 See de Gorter and Swinnen (2002) for a survey of status quo bias and methods to overcome it.


138

Of course, juxtaposing the plan with today’s policy environment highlights the dramatic changes that have occurred in the agricultural sector, and the re-orientation of the CAP. Food safety, animal welfare and the rural environment had no place in the Mansholt debate. And although measures encouraging (for example) early retirement are still in place, European agricultural policy in the twenty-first century is based on the multifunctionality model, keeping farmers on their land to provide public goods (Fischer Boel, 2005). Current concerns over climate change and energy security, with the positive contribution farmers can make by producing biomass for biofuels (De La Torre Ugarte, 2005), make it even harder to see another Mansholt Plan appearing in the next forty years. Finally, the counterfactual scenario of wholesale implementation of the 1968 memorandum programme would, in hindsight, have almost certainly featured repercussions and recriminations exacerbating the problems of European integration. The world economy took an unforeseen downturn in the 1970s as the post-war recovery ended, a process aggravated by the oil price shocks (Foreman-Peck, 1995). The ‘stagflation’ of the 1970s was not the macroeconomic situation needed to successfully implement the original proposals, if only in terms of creating and sustaining the non-agricultural jobs required to absorb the two million or so workers targeted to be redeployed from farming, thereby vindicating another of the objections to the plan. Mansholt (1979) himself recognised this. ‘Should we continue to encourage migration from the farms?’ he asked. ‘I say no ... This would only swell the army of the unemployed.’ Hence the case study of the Mansholt Plan also looks like a cautionary tale against introducing ‘big bang’ policy reform in the future – although it should not be taken as an excuse for inertia. Acknowledgements The author is grateful to Kingidila Mwaba Daba for research assistance, and to Cormac Ó Gráda and Jacqueline O’Reilly for valuable criticism and suggestion. References Ackrill, R. W. (2000). The Common Agricultural Policy. Sheffield: Sheffield Academic Press. Ackrill, R. W. and Kay, A. (2006). The EU budget and the CAP: an agenda for the review? EuroChoices,

5(3): 20-2. Baldwin, R. E. and Wyplosz, C. (2006). The Economics of European Integration. London: McGraw-Hill

(second edition). Commission of the European Community (CEC) (1968). Memorandum on the Reform of Agriculture in

the European Economic Community and Annexes, 18 December. Commission of the European Community (CEC) (1969). The Commission’s Memorandum on the reform

of agriculture in the Community. Newsletter on the Common Agricultural Policy, 1 (January). de Gorter, H. and Swinnen, J. (2002). Political economy of agricultural policy. In Gardner, B. L. and

Rausser, G. C. (eds) Handbook of Agricultural Economics Volume 2B: Agricultural and Food Policy. Amsterdam: Elsevier.

De La Torre Ugarte, D. (2005). The contribution of bioenergy to a new energy paradigm. EuroChoices,

4(3): 6-13.


139

Fennell, R. (1997). The Common Agricultural Policy: Continuity and Change. Oxford: Clarendon Press. Fischer Boel, M. (2005). Delivering on the potential of the new CAP. EuroChoices, 4(2): 6-11. Foreman-Peck, J. (1995). A History of the World Economy: International Economic Relations since

1850. Harlow: Pearson (second edition). Harvey, D. R. (2006a). The EU budget and the CAP: an agenda for the review? EuroChoices, 5(1): 24-9. Harvey, D. R. (2006b). Reply to Ackrill and Kay. EuroChoices, 5(3): 22-5. Ingersent, K. A. and Rayner, A. J. (1999). Agricultural Policy in Western Europe and the United States.

Cheltenham: Edward Elgar. Mansholt, S. L. (1968). The future shape of agricultural policy. Newsletter on the Common Agricultural

Policy, 1 (January). Mansholt, S. L. (1970a). The Mansholt Plan. Studies, 59 (Winter): 404-18. Mansholt, S. L. (1970b). Farm reform. European Community, 11 (November): 7-9. Mansholt, S. L. (1979). The Common Agricultural Policy: Some New Thinking from Dr Sicco Mansholt.

Stowmarket: Soil Association. Mayhew, A. (1970). Structural reform and the future of West German agriculture. Geographical Review,

60(1): 54-68. Neville-Rolfe, E. (1984). The Politics of Agriculture in the European Community. London: Policy

Studies Institute. Potter, C. et al. (1991). The Diversion of Land: Conservation in a Period of Farming Contraction. London

and New York: Routledge. Pugliese, E. (1972). The Mansholt Plan and the Mezzogiorno. English translation published in Pinto, D.

(ed.) Contemporary Italian Sociology: A Reader. Cambridge: Cambridge University Press (1981). Ritson, C. and Harvey, D. R. (eds) (1997) The Common Agricultural Policy. Wallingford: CAB

International (second edition). Rosenthal, G. G. (1975). The Men Behind the Decisions: Cases in European Policy-Making.

Massachusetts: Lexington. Tracy, M. (1989). Government and Agriculture in Western Europe 1880-1988. New York: Harvester

Wheatsheaf (third edition). van Lierde, J. (1969). The Mansholt Plan. London: Fertiliser Society.


140

AGRICULTURAL & NON-AGRICULTURAL RURAL EMPLOYMENT IN THE EU: ISSUES AND STRATEGIES, WITH SPECIAL REFERENCE TO ACCESSION & CANDIDATE

COUNTRIES

Martin M. Turner University of Exeter,Devon, UK.


E. John Wibberley Royal Agricultural College, Cirencester, UK


Abstract Rural unemployment and rural depopulation are common concerns of European countries, especially the new EU accession and candidate countries. The context for this paper is the substantial change in labour requirements, opportunities and responses of the labour force following the collapse of communism, coupled with the current emphasis on market liberalisation. Associated changes in the structure of farming are reviewed, with special reference to a commonly observed polarisation between small, semi-subsistence farms (sometimes operating as ‘hobby’ farms) and the far fewer emergent large, commercial farms. Diversification into alternative farm-based economic activities - such as agri-tourism and various types of ecotourism, on-farm processing of raw farm products – is considered. Using recent data, the paper explores the current issues involved in rural employment in Europe. A digest of trends in a number of countries is presented and an exploration of common and contrasting elements follows, together with diagrams/illustrations. In particular, these data are reviewed in relation to The Lisbon Strategy of 2000 and the actual trends measured since that date in a selection of countries. Issues of unemployment, hidden unemployment, under-employment and ‘the grey economy’ are covered. Changes in labour migration patterns are discussed in relation to the social, contractual and economic consequences of these changes. Contrasts are presented between employment opportunities for rural and urban, male and female, oldest and youngest, well-educated and less educated (especially in regard to appropriate rural vocational education and training). Shortcomings of The Lisbon Strategy are examined and a case is made for some alternative strategy elements in the light of environmental, livelihood and international relations imperatives. These concomitant matters demand management with ingenuity, determination and long-term vision. The paper concludes by suggesting recommended management approaches for both policy-makers and rural enterprise practitioners. Keywords: rural depopulation, labour, diversification Introduction ‘Rural Vitality’ is a comprehensive term that aggregates the economic, environmental and social factors which go to make a dynamic, sustainable countryside. Rural vitality requires enough farmers and farm staff in place ‘there to care’ for the countryside as heritage asset as well as present and future resource. It also needs the integration of sufficient non-agricultural employment (NAE) of a rurally-compatible kind (i.e. non-urbanising and operating on a modest scale). Growing displacement and disconnection are key rural and agricultural concerns. Farming integrates the delivery of rural vitality in practice. Thus, a viable agriculture with local food and locally-determined farm environmental management is crucial. The problems faced by rural areas in terms of economic development arise from a complex mix of issues, including social and historical patterns of land use. In the present world, many businesses in rural areas


141

experience problems which are related to the spatial characteristics of the region, in particular to remoteness – from the centres of population, from markets, from infrastructural and trading links. These problems may be exacerbated by a low population density and limited local markets, as well as by inferior educational, training and technology transfer opportunities. Rural areas, then, typically face numerous and serious economic challenges and often carry a disproportionate share of national poverty. Even though, in some respects at least, the global market is increasingly and actively extending into many rural areas, it remains true that young workers are leaving for urban centres, thus further disadvantaging the future economic and social vitality of rural areas. In an EU context, the policy debate about rural economic development has moved into a new era with enlargement. The EU enlarged from 15 to 25 member countries in 2004 and added two more – Bulgaria and Romania – in January 2007, with further candidate countries such as the Balkan States and Turkey queuing to join. EU rural development policy (Pillar II of the EU budget) has three ‘axes’, economic, environmental and social:-

• Competitiveness of agricultural and forestry sectors; • Improving the environment and countryside; • Improving the quality of life in rural areas and encouraging diversification.

While it is not clear to what extent these objectives are mutually compatible, or where compromises or ‘trade-offs’ may result in a re-focussing of one or more of them, the EU’s commitment to a three-pronged vision for rural development represents an endorsement of the concept of rural integration within the broader national and super-national economy and society. As such, it is to be welcomed by all who are concerned for those who live and work in rural areas. The proportion of the rural population employed in agriculture in CEECs varies widely between countries, and even between regions within countries (Baum and Weingarten, 2004), but in many regions, agriculture with its associated upstream and downstream sectors still plays a very important role in rural labour markets (European Commission, 2005). Moreover, it is often the case that its significance in employment terms is rather greater than its importance in terms of total gross value added (GVA) terms, reflecting the relatively lower productivity of labour in agriculture (European Bank for Reconstruction and Development, 2003). This feature of agricultural employment and relative factor productivity is not confined to the transition economies, of course. Initiated and supported by the European Commission (DG Research and DG Agriculture), the CEEC AGRI POLICY project which was funded for two years to April 2007 aimed to create a network of experts involved in agricultural policy analysis in the New Member States (NMS), in the Candidate Countries (CC) and in the countries of the Western Balkans. Its overall aim was to support the EU Commission and other policy-makers in the formulation of Community agricultural policies, and its main focus is on agricultural markets and rural development. This paper draws principally on the third rural development study conducted as part of this project (CEECAP, 2007) which focused on rural employment. Rural Development for Rural Vitality The objective of rural development has been defined as achieving ‘…an overall improvement in welfare of rural residents and in the contribution which the rural resource base makes more generally to the welfare of the population as a whole’ (Hodge, 1986). Post World War II, rural development was viewed largely as improving the economic conditions of agriculture and, later, as assisting economically depressed regions. Now, however, the emphasis is broader and encompasses achieving greater equity for those who live and work in rural areas, in terms of income, housing, health care, and access to other goods and services. Viewed from this perspective, rural development may be defined as making rural


142

Europe a better place in which to live and work. The emphasis is on the overall well-being of people, not merely on economic growth and development. The concerns of rural development range widely, involving issues of rural poverty, population demographics, rural housing, public services and creative employment opportunities, as well as economic development. Leon (2005) has argued that the development of rural areas is complex and involves using a wide range of perspectives to integrate and exploit complementary insights. Viable farming is central to rural vitality (Wibberley & Turner, 2006). This is because of the role agriculture plays in the production of public goods such as environmental quality and rural amenity, as well as because it remains the principal user of rural land (McInerney, 1999). In the UK, a CPRE/NFU (2006) survey calculated that 85% of the time for managing the countryside was effectively given freely by farmers rather than from funded agri-environment schemes. The Lisbon Agenda & Rural Vitality At the meeting of the European Council at Lisbon in March 2000 an action plan to deal with the EU’s low productivity and stagnant economic growth led to the formation of numerous policy initiatives. The Lisbon Strategy forms an over-arching framework for policy development in the EU during the decade to 2010, with the ultimate aim of making the EU ‘the most competitive and dynamic knowledge-based economy in the world, capable of sustainable economic growth with more and better jobs and greater social cohesion’. At the European Council held in Spring 2005 EU leaders put economic growth and employment at the top of Europe’s political priorities, and the renewed Lisbon Strategy represented a fresh commitment to mobilise and implement a positive reform agenda. In the EU, therefore, the competitiveness agenda has been defined by the Lisbon Strategy, which set out three main goals:

• An increased employment rate (from 61% in 2000 to 70% in 2010); • Regional cohesion; and • An average economic growth rate of 3%.

In January 2006, the Strategy was re-launched in order to capitalise on the new momentum for growth, with a strong focus on national reform programmes aimed at improving the competitiveness of the EU in global markets (European Commission, 2006). The focus now is on two main areas, productivity and employment. Already structural changes are considerable. In 2006 in the UK, for instance, the 60,000 biggest farms generated 96% of total production, while 65% of ‘diversification income’ on farms came from renting out redundant farm buildings.1 The Lisbon Agenda has caused some general concern. The key concern expressed is that the Lisbon Strategy should not be reduced simply to an economic goal, but that the social objectives originally identified should be at the heart of the implementation process over the next few years. The outcome of this debate about the meta-policy shaping the rural development of the EU over the coming years will be central to the future rural vitality of much of Europe. One of the challenges to policy makers at all levels is to develop a better appreciation of the range of alternative development trajectories of rural areas, in the context of the intrinsic strengths and possibilities such areas possess. There has been a widespread perception of rural areas as relatively passive recipients of an essentially urban-centred development agenda, and this has to change before real progress can be made. In tropical areas, it has long been recognised that people must participate fully in their own rural development if it is to be owned and sustained (Batchelor, 1993). In the UK in the 1950s, Evans(1956) proposed ‘ask the fellows who cut the hay’ in order to ascertain the realities faced by rural workers. Farmer participatory research and extension has proven effective in practice (Wibberley, 1988; Chambers et al 1989). Such approaches merit wider adoption within the CEEC countries.

1 Porter, C. (2007) Farm Business Vol.6 (2) 2nd February.


143

Management & Rural Vitality The contributions of management to the attainment of rural vitality operate at several levels:- Strategic thinking ‘outside the box’ simultaneously to integrate the complex components Marshalling of the facts and trends concerning contributory factors in the rural context Focus on rural employment – both in farming and non-agricultural work Review of the provision of training, extension and advisory services to enable it Individual enterprise management within a business Collaboration in learning and earning, including group formation and co-operation Change management at enterprise, business, regional, national & international levels Integrated rural development in a locality/region linking businesses & service providers The present paper is principally concerned with rural employment. Global & EU Rural Employment Trends For the first time in history, international statistics published by the UN in February 2007 show fewer people employed in farming and land-based work world-wide (38% of the global workforce) than in the service sector (40%). Furthermore, the urban population overtook the rural one globally for the first time ever recorded. The rural-urban exodus, and especially the loss of farmers represents a considerable upheaval for the management of natural resources, let alone the plight of many stressed farm families.2 Of the estimated 191 million migrants in the world - some of them refugees - many originate from rural areas and have previously been subsistence farmers. In China, there has been recently an active government policy promoting migration to the cities with at least four mega-cities being built for the purpose and an associated, albeit relatively short-term, enormous economic boom. Concern is growing that this policy and this boom is unsustainable. In Europe too, there is considerable recent migration on an unprecedented scale. Much of this migration represents a ‘brain drain’ from rural areas – particularly, perhaps, of entrepreneurial talent and spirit since those prepared to migrate may be less risk-averse. However, in the short-term, remittances sent back home by migrants can provide very strategic means for survival of those left behind, including enabling them to acquire hardware and so stimulate demand in local shops. In the EU, annual remittances are significant for some of the countries here studied (EPW, 2007) including Poland ($m 2,347), Bosnia ($m1,312) and Turkey ($m 804). Of the study countries, Turkey, Poland, Hungary, Croatia and the Czech Republic have the largest numbers of tourists – though tourism is predominant within the small Cyprus economy, and significant for countries like Slovenia and Slovakia. Both Bulgaria and Romania have high hopes to develop their tourist potential.

2 In the UK during 2006, there was a 60% increase in the number of distress calls to FCN (Farm Crisis Network)


144

Figure 1:. A comparative overview of EU Accession and Candidate Countries since 2004 COUNTRY Area

000 km²

Pop. M

% of pop. rural

Density Pop./km²

HDI #

GDP per capita : $ PPP

Ag. % of GDP

% who work in Ag.

% unemp.

Estonia 45.2 1.32 31 29.26 85.3 14,560 4.0 6 9.2 Latvia 63.7 2.30 34 36.21 83.6 11,650 4.0 14 8.8 Lithuania 65.2 3.44 31 52.82 85.2 13,110 6.0 18 5.3* Poland 312.6 38.58 37 123.43 85.8 12,970 2.8 18 7.3** Hungary 93.0 9.87 35 106.20 86.2 16,810 4.0 5 7.1 Czech Republic

78.8 10.23 25 129.80 87.4 19,410 3.4 4 9.1

Slovakia 49.0 5.40 42 110.24 84.9 14,620 6.0 6 11.5*** Slovenia 20.2 1.98 51 98.21 90.4 20,940 3.0 8 9.8 Romania 237.5 22.33 45 94.03 79.2 8,480 13.1 35 6.5 Bulgaria 110.9 7.89 32 71.20 80.8 8,080 10.1 36.4~ 11.5 Bosnia - Hercegovina

51.1 4.16 56 81.38 - 7,030 ‘grey’ 34.8 45.4

Croatia 56.5 4.42 - 78.33 84.1 12,190 8.0 - 15.7 Serbia 88.5 9.30 - 103.10 - 2,700

est. 17 - - ?

Cyprus 9.2 0.80 30 86.70 89.1 22,810 3.8 - 3.2 Turkey 779.4 71.32 34 91.51 75.0 7,750 12.0 34 10.3 Sources: Derived from Collins Handy World Atlas 2004; Whitaker’s Almanack 2007(139th edn.); Data submitted from National Reports of countries; The Economist Pocket World (2007) for PPP [ Note: $ PPP = Purchasing Power Parity, adjusting for cost of living differences based on a basket of goods and services - relative to USA at index100]. ~ In Bulgaria, this figure includes the 26.4% who are farm owners. * The Economist Pocket World 2007 states significant variances:- * = 12.8% for Lithuania; ** = 19.0% for Poland; *** =18.1% for Slovakia. # HDI = Human Development Index, which the UNDP launched in 1990 factoring together income levels, adult literacy and life expectancy data. In Bulgaria, 97% of farms are <5ha in size, and only <1% exceed 50ha. The social status of rural household members is shown in Fig.2.


145

Figure 2: Social status of the household members in rural areas 2005

There is considerable variation in population density among the new EU States (Fig.3). Even in the least densely populated country, Estonia, there is an absolute decline in population (Fig.4). Figure 3: Population density in New EU Member States (2004) Source: EUROSTAT

Unemployed, 8.4%

Students, 13.5%

Hired workers, 13.7% Retired people, 24.8%

Others, 8.3%

Farm owners, 26.4%

0

20

40

60

80

100

120

140

CZE POL SLK HUN SLO CYP LIT LAT EST

Inh

ab

itan

ts p

er

km

2

Density of population


146

Figure 4: Population dynamics in Estonia 1971-2005

Source: Estonian Statistical Office, 2006. Rural unemployment ranges from ‘the biggest social problem’ in Bosnia & Hercegovina and the Balkans generally to virtually nil in Cyprus. There is a significant informal employment (‘grey economy’) sector in many countries, notably in Serbia. Older rural residents are sometimes more likely to be unemployed (e.g. Cyprus) while in many places they provide the ‘social buffer’ continuity which underpins rural society e.g. in Romania, agriculture is the second earner for 95% of those in NAE. There are great regional differences in many countries, such as in Croatia, and at the periphery everywhere. Migrant labour is moving between the countries studied e.g. from Romania into Serbia for seasonal, casual farm work. Out-migration into the western EU and elsewhere is causing rural depopulation in the Baltic States and Poland. Many farm families face an uncertain future. (Fig.5). Figure 5: Serbian Farmer: a troubled past; an uncertain future. (Source: R.McCurrach)

1.2

1.25

1.3

1.35

1.4

1.45

1.5

1.55

1.6

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

mil.

Population (Mil.)


147

In Slovenia, settlement patterns are particularly dispersed and there is a particularly strong attachment to place such that growing numbers commute from rural areas into small towns to work. In general, the standard of living (space and fresh air notwithstanding) is greater in urban than in rural areas of the countries studied, and rising expectations cannot be met out of farm incomes; this is marked in Poland. The most strongly rurally dependent are the economies of Romania and Bulgaria with consequent expected impacts as they integrate into the EU following their accession in January 2007. There are particular ethnic issues, such as the gipsy population in Bulgaria and Romania. There is a rural exodus in Turkey but, because of population growth rate at 2% or so, the absolute rural population is maintained and infrastructure consequently strained (Fig.6). Figure 6: Population Trends in Turkey

Census Years

Total Agric. Population: villages + towns

Agricultural Population (%)

Urban Population

Urban Population (%)

Total Population

Population Growth Rate (%)

1980 25,091.950 56.1 19,645.007 43.9 44,736.957 2.07

1990 23,146.684 41.0 33,326.351 59.0 56,473.035 2.17

2000 23,797.653 35.1 44,006.274 64.9 67,803.927 1.83

Source: DİE, National Censuses, Turkey Agricultural Employment A number of factors are relevant in any consideration of the nature, level and trend of agricultural employment:- Active versus inactive labour force – many on farms are underemployed (Fig.7). ‘employed in farming’ status – this may fail to count owner-occupiers in some cases Registered and unregistered workers: The ‘Grey Economy’ large in e.g. Bosnia Disguised unemployment and underemployment categories Part-time employment & self-employment as potentially very good, not always negative The ‘Circulatory Economy’ of migrant labour, learning & remitting cash from abroad Entry incentives, assistance into employment, retention, and retirement schemes Professionalism in agriculture:- recognition, and ways of enabling CPD and LLL Targeting of vulnerable or disadvantaged groups – women, elderly, disabled… SME start-up incentives and conditions Planning policies facilitating adding value to farm produce by processing in situ Figure 7: Latvia: Main indicators of employment in rural territory (000 population)

2004 2005 Rural territory

Total Men Women Total Men Women Persons aged 15 to 74 years 565.7 281.4 284.3 566.2 282.5 283.8 Active population 332.1 186.0 146.1 331.5 186.1 145.4 Employment rate 53.8% 60.1% 47.5% 54.2% 60.5% 47.9% Unemployment rate 8.4% 9.1% 7.5% 7.4% 8.1% 6.5% Economically inactive people 233.6 95.4 138.2 234.7 96.4 138.4


148

Source: CSB of Latvia, Labour Force Survey Meanwhile, farmer loss is a common feature (Fig.8.) as agriculture, essentially a primary industry, releases people for employment elsewhere in the economy or, as in the case of some of the post-centrally planned economies, for little productive contribution. Figure 8: Estonian employment trends in rural areas and agriculture (1991-2005)

207.5

177.5 176.2 181.7 184.8184188.4

237.3

128.9

91.9

55.547 43.5

28.4 25.4 23

0

50

100

150

200

250

1991 1993 1995 1997 1999 2001 2003 2005

(000 e

plo

yed)

Rural employment (000) Agricultural eployment (000)

Source: Estonian Statistical Office. NOTE: Non-Agricultural Employment (NAE) increased by 50% between 1991 & 2005. There is a lack of rural pensions generally so loss of farm livelihood is serious everywhere. In the Czech Republic, the farm workforce declined by 73% between 1989 and 2005, and dramatic losses occurred elsewhere too. Perhaps for cultural reasons, women have a greater farm involvement in some countries than men (e.g. in Turkey) but men do more farm work in most European countries than women. Men are paid much less than women for farm work in some countries, e.g. Cyprus. Foreign workers account for 70% of all farm staff in Cyprus. Part-timers feature in many countries and 70% of all farm work is done by part-timers in Slovenia, 81% in Latvia. In Turkey, half of all farm labour is unpaid family members, especially women. As well as between countries, there are also huge regional differences in the percentage employment in agriculture e.g. in Poland, Silesia has 9% while Podlaskie has almost 40% in farm work; considerable variation also exists in Hungary (Fig.9), especially in the proportion engaged in agriculture. Figure 9: Hungary: Regional variation in employment indices (2003)

Region Activity rate (%) Unemployed (%) Employment (%) Agric.work (%)

Central Hungary 57.5 4.0 55.1 1.6 Central Transdanubia 58.0 4.6 55.3 4.9 Western Transdanubia 57.7 4.6 55.1 4.8 Southern Transdanubia 51.3 7.9 47.2 9.4 Northern Hungary 49.8 9.7 45.0 4.6 Northern Great Plain 49.3 6.8 45.9 7.9 Southern Great Plain 50.3 6.5 47.0 11.6


149

Source: CSO, Budapest. The Hungarian farm population declined by 50% between 1991 and 2003. The average farm size is 3 ha but the social importance of farming far exceeds its financial contribution in the economy. Non-Agricultural Employment (NAE) NAE includes other primary employment sector work – in forestry (e.g. it is the leading NAE in Latvia), fishing, hunting, mining and quarrying. Agricultural processing and ‘adding value’ to farm products is often important. Other sectors include construction, manufacturing, services and tourism, including agri-tourism. Niche markets are key, including local handicrafts and other local products e.g. Turkish carpets, Bulgarian garments, Slovenian electrical and electronic goods. Hotels and restaurants are important in established and expanding resorts such as in Croatia, Cyprus, Turkey and Bulgaria. However, education and training is often inadequate for NAE, many areas lack micro-credit sources, and many rural populations are risk-averse e.g. in Romania. Poor infrastructure, e.g. bad roads, impairs NAE in some countries e.g. Poland. Advisory Services are often lacking, though Estonia has remedied this with useful results, together with consultations on business start-up; there, NAE is growing faster than the EU average. Conversion and rental of farm buildings for other purposes can become an increasingly important source of income, as in the UK (Turner, et al, 2006). It is reported that the ‘grey economy’ produces as much as 40% of Serbia’s ‘social product’ – though overburdening the taxpayers - and is important in most countries studied. Emigrant remittances are also vital, especially in the Balkans, e.g. in Bosnia & Hercegovina it is reckoned that 25% of its GDP originates in this way. New Rural Employment Opportunities Farm multifunctionality is seen as pivotal, with considerable scope to develop farm and forest product processing and value-adding, particularly for niche markets (Fig.10). Renewable energy is seen by many as a major upcoming opportunity, not only for biofuels but also for windfarms and other technologies. Rurally compatible factories - such as those making clothing in Bulgaria – have further potential. Figure 10: Slovenia: Family farms by supplementary activities; 2003 & 2005

Number of farms Index Share (%) 2003 2005 2005/03 2003 2005 TOTAL 2.867 3.146 109.7 100.0 100.0 Food processing - meat 101 189 187.1 3.5 6.0 Food processing - milk 115 185 160.9 4.0 5.9 Food processing – fruits and vegetables 354 390 110.2 12.3 12.4 Food processing – others 104 200 192.3 3.6 6.4 Wood processing 508 449 88.4 17.7 14.3 Services with agricultural machinery 905 796 88.0 31.6 25.3 Tourism on the farm 675 628 93.0 23.5 20.0 Cottage industry 130 171 131.5 4.5 5.4 Public utility services 149 297 199.3 5.2 9.4

Source: Statistical Office of Republic of Slovenia


150

SMEs, services, crafts and tourism are varyingly developed – with tourism, and agri-tourism seen as hopeful in many places, though with great regional differences. Nearer to cities, such as Riga in Latvia, opportunities in general are much greater. Seasonal work is important in many areas, as is circulatory migration and the remittances it brings, though it could be argued that their arrival diminishes the need for innovative business ventures. In some areas e.g. Slovakia, already there is diminished demand for agricultural graduates. In Slovenia, many commute from rural areas to work in small towns where opportunities are greater, perceiving the countryside as a preferred place of residence, recreation and sports. These last social changes offer new job opportunities for some in providing the associated services, including retirement homes and healthcare for often growing numbers of elderly rural residents. Telecottages and e-businesses both offer new jobs but have yet to be significantly developed in most areas surveyed. It is a concern everywhere that primary sector jobs languish, and those in the secondary sector have declined as far as heavy manufacturing industry is concerned. Simply to switch all hope to the tertiary sector may leave economies with a vacuum of solid primary production; not everyone can produce computers, speculate financially or become social workers! However, policy makers can influence opportunities significantly e.g. in Lithuania, it is reported that every third job promoted by the Labour Exchange is in rural areas. Education, Skills & Rural Employment Opportunities Rural people are generally less well-educated than urban residents, with commonly as many as one-fifth to one-third having no formal education. Rural schools are often ill-equipped. This tendency is usually greatest for women and for older people. Some countries are making big improvements but from a very low base. However, the Baltic States are focusing on improved relevant agricultural training, combined in the case of Latvia especially with free business consultations and new advisory services. Discerning employers prefer experience to qualifications per se, though a combination of both is increasingly sought. However, it is difficult for many graduates to gain relevant experience, and some employers stage long apprenticeships in these circumstances to cheapen their wage bills – e.g. cited in Slovakia as an issue, where there is quite a supply of educated agriculturalists (Fig.11). Targeted training schemes have been successful in some places e.g. Bosnia & Hercegovina. There is also evidence from Turkey that trained agriculturalists are much more likely to own more land. Figure 11: Slovakia: Education of those employed in Agriculture (%)

Level of education 2000 2001 2002 2003 2004 2005

Elementary 19.7 16 16.4 14.8 13.8 14.8 Full secondary vocational 51.8 56.7 57.2 54.5 53.7 50.5 Full secondary specialised with leaving exam 24.3 22.5 22.8 23.7 25.4 27.7 Higher specialised 0.2 0.4 - 0.3 0.2 - University 3.7 4.4 3.6 6.7 6.9 7

Source: Slovak Statistical office, 2006 Many countries report a mismatch between labour market needs and actual vocational training offered, together with a shortage of CPD/LLL. Farm labour is poorly paid, poorly educated and often may be slower to seek educational opportunities – as reported in Romania, for instance. In Bulgaria, rural illiteracy is double that in urban areas and there is concern regarding the danger of regional educational imbalances too.


151

Other Factors & Rural Employment

There are a number of other factors which have a bearing on the level and robustness of rural employment, principal among which are the following: Rural infrastructure is often poor – notably roads – and regional differentials can be great. Low wages in farming plus a low proportion of farm jobs which are waged at all make for rural poverty in many areas. The ‘brain drain’ is depleting able, pioneering types from rural areas. Gender issues are still acute in some areas e.g. for women in parts of Turkey. Isolation is an issue for many, though improved information systems can help e.g Estonia’s new rural newspaper ‘Good Advice’ Foreign workers are socio-economically significant in some countries e.g. Cyprus, while for others (Romania and Bulgaria) the resident though nomadic Gipsies present particular challenges. For many countries, such as Poland, the key factor is that there is still a high proportion of surplus people in the rural economy leading to many underemployed or high hidden unemployment. Underemployment in Rural Areas

Under-employment affects especially women, older men and those less educated. Employers get skilled people cheaply because of it. Many countries do not really try to record it, as admitted by the Czech Republic. Under-employment appears as part-time work, the ‘grey economy’ and a high proportion of family farmers who may work only part-time on their small farms. One big reason for agricultural under-employment is the lack of development to date of ‘value adding’ to farm products in many places e.g. in Slovenia GVA for agriculture is 20% of that for the economy as a whole. However viewed positively, the combination of flexible labour contracts, self-employment and multiple part-time work by which people survive may portray a truly sustainable future. Other responses to rural under-employment include commuting to towns for work (as done by one-third of Estonia’s rural residents), early retirement schemes (as introduced in Romania in 2005 for farmers >62) and re-skilling training. The Lithuania Report complains that social welfare grant policies are inimical to progress – including rural people’s willingness to do seasonal farm work despite high under-employment - and it advocates re-skilling instead of ‘dole’. The place of Semi-subsistence Farming Methods of description vary but, in general, there is growing polarisation between many, small farms and very few large farms e.g. In Romania in 2002, 76% of farms were classed as ‘subsistence only’ (selling nothing but eating all their produce), 21.7% ‘semi-subsistence’ (i.e. selling some of their produce) and only 2.3% ‘commercial’ ; farms >100ha occupied only 0.23% of all Romanian farms but 48% of the nation’s UAA (Utilised Agricultural Area). During restructuring there, as elsewhere, many returned to small farms as their only means to escape poverty. Small farms are often described as providing a ‘social buffer’ e.g. in Croatia, 40% of the rural population are poor and depend on small farms to survive. Retaining yet enlivening as many as possible of these small farms may well be the key challenge for truly sustainable development. Rural family living costs are smaller than urban ones both financially and in global energy terms.


152

Examining the Lisbon Strategy in relation to Rural Employment The ‘Lisbon Strategy’ of 2000 emphasises achieving fuller employment and social cohesion as well as raising workplace quality standards. To attain these, its priority is economic growth (albeit sustainable growth) by creating new job opportunities. The Lisbon Strategy’s objectives for 2010 include having at least 70% of the labour force employed (at least 55% of the labour force aged 55-64 years, and 60% of the female labour force to be in work). Participating countries are exhorted to pursue :- knowledge-based societies, improved internal markets, better business environments, more dynamic labour markets and sustainable development. According to a PriceWaterhouseCoopers Report (Daily Telegraph Business, December 11th 2006), skilled workers are not moving about Europe as freely as anticipated by The Lisbon Strategy. With the exception of the Nordic countries, Ireland and the UK, mobility of skilled workers is said to remain ‘disappointingly low’. Barriers to greater labour mobility that are cited include language differences, incompatible or non-transferable health-care benefits and different tax systems. On the other hand, rural depopulation is excessive in some countries owing to particular out-migration e.g. UK Office for National Statistics data show that over 0.4 million arrived in Britain from Eastern Europe between 2004-2006, originating as follows :- Poland 264,560 Lithuania 50,535 Slovakia 44,300 Latvia 26,745 Czech Rep. 22,555 Hungary 12,870 Estonia 5,110 Slovenia 420 It should be noted that numbers arriving from Romania and Bulgaria are expected to increase in the future. There is also the phenomenon of circulatory migration whereby people move across borders for seasonal work; this has increased following suspension of the visa for the Schengen space on January 1st 2002, and in Romania accounted for some 62,000 people during 2002. Alternative considerations are proposed towards a sustainable context for rural employment :- opportunities should be reviewed in relation to actual trends both locally and globally, ‘semi-subsistence farming’ should not be used exclusively perjoratively. Part-time farming can be the least bureaucratically complex way of ensuring a future stake in the land for many people while enabling them to earn income outside traditional farming ventures without the social stigma of being ‘only part-time farmers’3 (i.e. ‘not to be taken seriously’ or, in some cases, seen as ‘mere hobby farmers’ – albeit ‘hobby farmers’ are a legitimate category, often self-labelling). The Lisbon Strategy might be significantly questioned and modified – not simply accepted as being comprehensively and finitely appropriate. For instance, some of the more serious criticism made of it include the following :- it is weak on sustainability and integration of development in the light of energy-efficiency; it does not sufficiently address environmental management issues; it does not address the growing interest in the role of land in relation to climate change; it does not address the ‘citizen acceptability’ in civil society of EU CAP & Rural Policy; it is lacking in relation to retention of people in rural areas ‘there to care’ for land/heritage;

3 This was a conclusion of a 7-country study: Wibberley, E.J.(1990) Survival of the family farm: family-worked dairy farms &

the viability of rural communities. NSch Report , Nuffield Farming Scholarships Trust/ Trehane Trust, 59 pp.


153

it omits that many small-scale, private entrepreneurs make for a strong agrarian structure; it places insufficient priority on local food strategies and national/regional food security and it is silent on food sovereignty, and the need to ‘build the middle’ in the Food Chain everywhere. However, realpolitik suggests that rural unemployment and rural depopulation are and will remain common concerns of European countries, especially the new EU accession and candidate countries. Thus, targeting more funds towards rural development is close to the real needs of these countries. Appropriate rural development ought to take proper account of the factors noted above, not to separate production and beauty conceptually as well as via the ‘pillars’ through which EU policy addresses them. Countries should be encouraged to articulate their concerns about the shortcomings of The Lisbon Strategy rather than simply being deferential towards it in order to appease current EU policy-makers, or to comply for EU entry. Only the Poland Report (on p.11; Wibberley, 2006b) alluded directly to any critical appraisal :- ‘Experts think that the Lisbon Strategy was prepared well but is realised in a wrong manner, among other things due to the intra-country barriers – weakness of the political leadership and lack of the social acceptance for increasing the market’s role and individual responsibility, and limitation of the welfare role of the State’. This is taken as a plea for integrated realism not a return to the shortcomings of socialism. Most countries seek to comply with ‘Lisbon’ policies to favour: economic growth via private property and enterprise, social inclusion, eliminating job discrimination, stimulating the labour market and improving labour mobility. Some countries e.g. Romania, admit their inability to comply with ‘Lisbon’ targets as yet. Only Poland questions somewhat the feasibility of the ‘Lisbon’ agenda. Bulgaria is more concerned with its own national stable development than with new jobs per se. The role of foreign and domestic investment is crucial, together with reduction of regional differentials within countries. A range of policies have been devised in the attempt to deliver the ‘Lisbon agenda’, including :- More diversification and the encouragement of SMEs Increased R&D spending on rural job creation and related topics (raising this to 3% of GDP) National Action Plan for Employment – in Croatia National Strategic Rural Development Plans 2007-13: e.g. Czech Republic, Cyprus, Slovakia More NGOs: Civil Society duly motivated towards self-help and enterprise, not ‘hand-outs’. Internet access and other communications improved. Local resource development being fostered for renewable energy, crafts, local produce… Gender equality addressed where necessary e.g. for women in Turkey ‘Multi-professionality’ for most rural people is advocated in Poland. Conclusions & Recommendations The studies on which this paper has drawn have identified a wide range of conditions and an almost equally wide range of policy initiatives, as might be expected (CEECAP, op cit). Most countries report poor rural infrastructure and weak demographic structure (many outside the active working age range). The agricultural workforce is hampered by poor education, low farm wages, few paid jobs, low mobility of workers, ageing and much under-employment. Remoter rural areas are losing many young people. Countries aim to attract investment and EU grants by compliance with EU policies to reform farm structure etc. Diversification, training and harnessing advice are seen as keys to rural development. The huge continuing importance of agriculture is highlighted especially in some countries e.g. Romania has 64% of its rural workforce farming with 32% of its total workers; Turkey has 34% of its workforce engaged in agriculture. However, it is reported that commuting to town jobs is increasing e.g. in Estonia, Hungary. Nevertheless, food security and the ‘social buffer’ of small farms remains of crucial importance for the rural population in particular.


154

There needs to be more concerted effort towards Integrated Rural Development, with its in-built diversification and appropriate ‘rural hubs’ (one-stop advice, sales and information points). These should logically include ‘rural development forestry’ as a long-term strategy, incorporating heritage and leisure-based business opportunities. Rural living costs are lower than urban ones, both in terms of family finance and in energy costs of the whole system. The ‘local resource management’ theme needs better, more overt linkage into upcoming global issues, notably energy security as well as food security and water security. Bosnia & Hercegovina recognises an outstanding fact - which some established EU members have been inclined to ignore to their peril – by stating, ‘Ensuring food security remains the first role of the farm sector’. In some countries at least there appears to be an ambivalence towards the ‘grey economy’ and this is conflated with a general note of seemingly ‘expected’ disapproval of part-time employment for the eyes of those in EU circles. However, ‘multiprofessionality’ in rural areas, with a solid core of self-employed and part-time employed people is wisely seen as a key hope for the future by the authors of the Polish Report. The essentials for sustainable rural vitality – vibrant ecology, economy, employment, energy-efficiency, equity and ethics – do not simply fortuitously coincide. They need to be simultaneously conceived, pursued and managed within an integrated vision. The role of government in this is to signal that food security matters in each nation/region, to ensure that laws deter bad practices and penalise where necessary, and then to minimise bureaucratic interference in creative rural enterprise. The Lisbon Strategy is essentially about improving the competitiveness of the EU, but it is widely acknowledged that achieving this is difficult. Competitiveness is an idea that is widely used but usually understood in only its simplest sense and, in particular, its relevance to a nation, region or locality is sometimes disputed. Typically, the policy focus often becomes one of improving labour productivity rather than sustainable, integrated rural development. References & Further Reading Batchelor, P.G. (1993) People in Rural Development. (2nd edn, Paternoster, Carlisle, UK, 228 pp.) Baum, S. & Weingarten, P. (2004). ‘Developments of rural economies in Central and Eatern Europe: an

overview’, in Changing functions ofrural areas in the Baltic Sea Region. Institute of Agricultural and Food Economics and the Institute of Geography and Spatial Organisation, Polish Academy of Sciences.

CEECAP (2007). Rural employment in the context of rural development. Third Rural Vitality study (15

countries). Published on: www.agripolicy.net CFG (2006) National Food Security. The Commercial Farmers Group, UK : [email protected] Chambers, R, Pacey, A. & Thrupp, L.A. (1989) Farmer First: farmer innovation and agricultural

research.(Intermediate Technology Publications, London, 219 pp.). Curry, D.T.Y. et al (2002) Farming & Food: a sustainable future. Report of the Policy Commission on the

Future of Food & Farming. (Crown Copyright, London, 150 pp.) EPW (2007) The Economist Pocket world in Figures 2007.(London, 254 pp.). European Bank for Reconstruction and Development (2003). Transition Report 2003: Integration and

regional cooperation. (London: EBRD).


155

European Commission (2005). Agriculture in the European Union: statistical and economic information

2004. (Luxembourg: Office for Official Publications of the European Communities).European Commission (2006) Time to Move up a Gear: the new Partnership for Growth and Jobs (http://europa.eu.int/growth and jobs/ ).

Evans, G.E. (1956) Ask the Fellows who cut the Hay. (Faber & Faber, London, 262 pp. 1972 reprint) Gore, A.(2006) An Inconvenient Truth. (Bloomsbury, London, 325 pp.) & www.climatecrisis.net HM Treasury/Defra (2005) A Vision for the Common Agricultural Policy. (69 pp, Dec.’05). Hodge, I. D. (1986) The Scope and Context of Rural Development. European Review of Agricultural

Economics Vol. 13, 271-282. Houghton, J. (2004) Global Warming (Cambridge UP, UK, 3rd edn.) Leon, Y. (2005) Rural development in Europe: a research frontier for agricultural economists. European

Review of Agricultural Economics Vol. 32 (3) pp.301-317. Lewis, J.S. (1954) Fairer Shares. (Staples Press Ltd., London, 244 pp.). Lobley M., Reed, M. & Butler, A. (2005) The Impact of Organic Farming on the Rural Economy in

England. Final Report to Defra, Centre for Rural Research, University of Exeter. Lovelock, J. (2006) The Revenge of Gaia. (Penguin, 177 pp.). McInerney, J. P. (1999) Agriculture at the Crossroads. Journal of the Royal Agricultural Society of

England 160, 8-27. Morison, J., Hine, R. and Pretty, J. (2005) Survey and Analysis of Labour on Organic Farms in the UK

and Republic of Ireland. International Journal of Agricultural Sustainability Vol.3, No.1, pp. 24-43.

Nell, W.T. & Napier, R.J. (2005) Strategic approach to Farming Success. (Nell, Bloemfontein, RSA, 323

pp. ISBN 0-62033428-2). Padda, Z. (2006) Ethical First (21st Century Gangmastering) : [email protected] Thomas, R. & Wibberley, E.J. (2001) Integrated Rural Development: Agriculture & Rural Development

Forestry. Journal of the Royal Agricultural Society of England 162, 89-96. Turner, M.M., Whitehead, I. & Millard, N. (2006) Embedding farm diversification: contemporary trends

and policy development in England. Agricultural Economics Society, 80th Annual Conference, Paris, 17 pp.

Turner, M.M. & Wibberley, E.J. (Eds.) (2005). Study of Rural Vitality: First Report for the CEECAP

Project (4 volumes). Published at: http://www.europartnersearch.net/agri-policy/index.php?page=expertcontributions

Wibberley, E.J. (1988) Developments in arable management through farmer groups. Journal of the Royal

Agricultural Society of England, 149, 133-147.


156

Wibberley, E.J. (2005a) Agriculture in Place. Journal of the Royal Agricultural Society of England Vol. 166, 96-104. (www.rase.org)

Wibberley, E.J. (2005b) Agro-economic policy analysis of the new Member States, the Candidate States

and the countries of the Western Balkans: Overview & Recommendations. Report for Centre for Rural Research, University of Exeter, UK, 23 pp.

Wibberley, E.J. (2005c) Leadership Values & Sustainable Trading Management for Food Security,

Biodiversity & Equity. Vol.1 pp. 333-345 In Developing Entrepreneurship Abilities to feed the world in a sustainable way. International Farm Management Association 15th World Congress, Campinas, Brazil.

Wibberley, E.J. (2006a) The ‘Farming Tsunami’: the crisis of farm livelihoods. Tropical Agriculture

Association Newsletter 26(1) 20-21 (www.taa.org.uk). Wibberley, E.J. (2006b) Rural Employment in the context of Rural Development in the new EU Member

States, the Candidate States and the countries of the Western Balkans. Report for Centre for Rural Research, University of Exeter, UK, 34 pp.

Wibberley, E.J. (2007) Recognising Professionalism in Agriculture. The Farmers’ Club Journal 206, 6-7

(www.thefarmersclub.com London). Wibberley, E.J. & Turner, M.M. (2006) Farming & Rural Vitality : connections in the enlarging EU.

Journal of The Royal Agricultural Society of England 167, 85-93. (www.rase.org)


157

FARMING IN EASTERN GERMANY: FROM FOOD TO ENERGY CROP PRODUCTION?

Heiko Zeller and Anna Maria Häring University of Applied Science, Eberswalde, Germany


Abstract

Renewable resources are of importance in our modern society due to their positive effects on agriculture, the environment and the economy. To support renewable energy from biomass the EU promotes the cultivation of energy crops. This creates alternative income sources for farmers as primary or energy producers and strengthens added value and employment. For the analysis a model is developed to assess the potential impacts of energy crop production on cropping activities. A basic quadratic version of PMP is used to maximize total gross margin in two regions, which allows one to simulate farmers` behavior under different conditions. Different scenarios show, that the bio-energy boom partially contributes to crop substitution effects. Potential energy crops like rye, rape seed or silage maize are more profitable. This has an impact on crop rotations because less profitable crops are substituted. However, these tendencies approach a limit in terms of limited area and crop rotational aspects. Keywords: energy crops, land-use change, positive mathematical programming (PMP) Introduction The development of renewable energy has for some time been a central aim of energy policy within the European Union. There are two reasons. Firstly, the dependency on energy imports is already 50% and is expected to rise over the next years if no action is taken. Secondly, the EU has recognized the need to tackle the climate change issue to reduce greenhouse gas emissions. Consequently, the expansion of energy production from biomass becomes more important. Biomass is a key source because of its potential to limit CO2 emissions. Energy crops will be used to produce a broad spectrum of fuels including bio-diesel, ethanol and the new Biomass-To-Liquid (BTL) fuels. To support renewable energy from biomass in the EU promotes the cultivation of energy crops with area payments and allows the cultivation on set-aside land. As a result energy crop production has come to offer an alternative for agricultural enterprises as it opens new income sources for farmers besides food production and simultaneously strengthens added value and employment particularly in rural areas. In the scope of multifunctional agriculture farmers may act as primary food or energy producers in the future. Across Germany almost 1.4 million ha, that is 8% of land under cultivation, were being used for energy crop production in 2006.

Rising demand for food and the increasing area for energy utilisation as biogas or transport fuel have positive effects on prices for agricultural commodities. Food versus energy production is at present a commonly used "slogan". In light of the decoupling of area payments and the gradual liberalization of agricultural trades, market oriented production structures gain in importance and determine the economic success of farm enterprises. Concerning the cost-value ratio this has an impact on crop rotations which are mainly influenced by monetary and phytosanitary concerns. The study assesses the possible impacts of these aspects on cropping portfolios and provides information about future trends of biomass production.

Methodology For policy analyses in the agricultural sector Linear Programming (LP) is one of the main instruments used to analyse the effects on production output, land use and farm income. By assuming optimum


158

production combinations and profit maximising behaviour the approach is based on simulation models that reflect farmers reactions and allows an analysis of policy changes. However, most of these studies refer to a more or less tangible and empirical application because this kind of approach requires comprehensive data and field-work (Varela et al. 1998). Another drawback is the phenomenon of over-specialization; the number of constraint functions is smaller than the number of activities observed in the base period. For that reason the modeller is obliged to extent the set of constraints to avoid overspecialization with the intention to calibrate results to the observed situation. Both characteristics limit the information value of usual farm models. Models should reproduce base-run results according to observed production activities and should react reasonably.

In response to this problem and to analyze different spatial units in a regional agricultural production model alternative approaches were developed, to overcome the lack of accuracy and to ensure greater analytical capacity for agricultural policy problems.

The Positive Mathematical Programming method (PMP), originally developed by Howitt and Mean (1983), is designed to tackle the above mentioned problems of traditional linear programming and has become widely used to calibrate agricultural production models at various levels e.g. farm, region or sector. In revisions the methodology was improved by Paris (1988) and Howitt (1995). A recent development is to combine PMP with the method of maximum entropy (Paris and Howitt 1998; Heckelei and Britz 1999). The basic concept of PMP is that it is easier to get information on farm output data compared with data on production costs. Output levels are the result of complex decision making processes by farmers. These are based on total cost functions, which are difficult to measure externally and only known to the farmer. In the end it is possible with this information to develop models that can accurately represent farmers behaviour (Arfini 2001). Further advantages of the approach are exact representation of the reference situation as well as a smooth response of model results to changes in exogenous parameters.

PMP methodology is a three step procedure. In the first phase a conventional LP model is defined and solved to provide activity based dual values. These are used in the second stage to derive calibration coefficients with the aim of specifying a nonlinear objective function of the calibrated model in stage three. The new calibrated programming model reproduces almost exactly observed crop allocations compared to the base-run.

Although the calculation of PMP coefficients requires lower data requirements a minimum amount of a priori information such as supply elasticity or expected yield variation is needed to identify the cost or yield function of the marginal crop. This is a disadvantage as these data are sometimes not easily available with regard to time and financial restrictions.

A PMP approach developed by Paris (1988) is a suitable option to specify all crops, while the above mentioned a priori information is not needed. The corresponding dual formulation of the initial LP version is used to derive equations for calculating the PMP coefficients.

The calibration model can be compactly written as follows:

max TGM with

[ ]∑ −+=i

iiiiiiiXγXvcPRp(yTGM 2

1 (1)

1 The dual formulation of the model appears as :

∑ ∑+=i i

iilandiXXTC )()( λλ

0≥

≥+

iland

iilandGM

λλ

λλ

with TC =Total Cost λland = shadow price of land λi = dual value of calibration constraint GM = Gross Margin


159

subject to

)ˆ()( ∑≤∑i

ii

iXX resource constraint to calculate (

landλ ) (2)

)1(ˆ ε+≤ii

XX calibration constraint to calculate (i

λ ) (3)

0≥∑i

iX non-negativity condition (4)

where TGM denotes the objective function value of Total Gross Margin; y is a vector of crop yield; vc is a (n x 1) vector of variable cost per production activity; X is a (n x 1) vector of production activity levels; PR is a vector of area payments for energy crops; i denotes the crop type and ε is a perturbation coefficient with a small positive number.

By assuming the optimum production combination the coefficients for λi can be calculated due to the major condition that marginal gross margins of each activity are identical in the base run. The marginal gross margin results from the shadow price of the resource constraint which is identical with the difference of the gross margin and shadow price of the calibration constraint for each crop (compare note 1).

iLPilandiiiiii

i

i GMXvcPRpyX

GMλλγ −==−+=

∂

∂)(

ˆ2 (5)

Now γi can be isolated by extending the equation with the variable cost term on both sides and rewritten in the following way:

ii

ii

iXvc

vc

ˆ2

λγ

+=

(6)

It this expression the right hand side represents the slope coefficient of the cost function. The term includes variable costs for the considered cropping activities and the shadow price of the calibration constraint (λi). The composition of the term demonstrates why there is no need for a special operation to calculate the PMP coefficients for the marginal crop. Even if there is no value of the calibration constraint the numerator is in either case greater than zero since vci is typically positive. Hence, a further step to calibrate the marginal crop is not necessary.

The approach developed by Paris (1988) has its limitations as well. Especially noteworthy is the assumption that marginal costs are equal to zero at an activity level of zero, which means that the marginal cost curves intersect the origin. This leads to an overestimation of the gross margins at least for the marginal crop.

In our case a basic quadratic version was applied instead of a more recent development of PMP because of the bioenergy boom and its implications on prices and yields. For the period under consideration market prices have changed significantly and did not allow to calculate reliable supply elasticities. The problem with yield functions has its origin in the alternative ways of biomass utilization. For bioenergy purposes either the whole plant or the seeds are used as substratum. Depending on utilization different quality parameters are important such as starch for bioethanol production. For that reason it is difficult to get data about yield variations.

For the study variable cost data were available for both ways of utilization from the Ministry for Rural Development, Environment and Consumer Protection of the Federal State of Brandenburg (LVLF 2005) and the KTBL (2006).

subject to


160

Empirical application Using this methodological framework an initial model with a comparative-static approach was defined which allowed easy application. The model operates on regional level and provides information on land use changes resulting from different scenarios. On the basis of regional statistics, typical farms and their crop rotations are defined.

For each region six different cropping activities are differentiated. The crops include wheat, rye, barley, triticale, rape seed and maize silage. On the basis of different land categories information about crop prices, yields, input use per crop, variable costs per activity and area payments were available. 2 In addition it is assumed that 40% of rape seed, rye and corn silo are calculated with area payments for energy crops.

Specifically, the study uses the following scenarios:

(1) Market price changes for different crops (∆ price); the price comparison refers to prices from November 2005 and 2006. The price increases for the differing crops were different ranging from 5% for triticale to almost 25% for wheat (see Table 2).

(2) Technical progress (∆ yield); in this scenario a 10 % increase in yields is assumed for all crops to show how changes in quantity may affect the cropland allocation on the basis of current prices.

(3) Elimination of area payments for energy crops (∆ premium); it is assumed that this area payment is abolished as the bioenergy boom has stimulated the market prices.

The different scenarios refer to the selected regions Barnim and Uckermark in the Federal State of Brandenburg. The study determines the effects on crop production without accounting for animal husbandry response. The crop activities for the base-run are illustrated in Table 1.

Table 1. Crop activities in the base-run

• Region • Crop activities • Total area

• • Rape seed

• Wheat • Rye • Barley • Triticale • Maize silage

•

• Barnim [ha]

• 3.700 • 4.000 • 6.500 • 2.900 • 5.300 • 3.200 • 25.600

• Uckermark [ha]

• 25.300 • 46.000 • 16.100 • 11.500 • 6.900 • 8.000 • 113.800

Source: ATKIS (2005)

Table 2. Market prices for crops within the observation period

• • Rape seed

• Wheat • Rye • Barley • Triticale • Maize silage

• Market price 2005 [€/t]

• 205 • 92 • 75 • 82 • 79 • 21

• Market price 2006 [€/t]

• 256 • 115 • 94 • 90 • 83 • 25

2 Organic farming areas, pasture lands and other crops with a minimum percentage rate are not included. The documented

share of acreage amounts to 70%.


161

Source: ZMP (2006)

Figures 1 and 2 show percentage changes of cropping portfolios for the described scenarios. Results show that rising market prices induce farmers to change cropping patterns. In the Barnim region area increases for rape seed, wheat and rye are most pronounced. The growth of rape seed arises from the increasing demand for biodiesel resulting in rising prices of about 25% within a year. The changes for wheat profitability can not only be attributed to the bio-energy discussion but rather increasing world market prices. Interestingly enough, maize silage did not react in the same way although prices increased by 20%. However maize silage competes with other crops for limited area even though it is one of the main substratum for biogas plants. Rye is an important crop in this region due to the poor land quality. Recently, the crop has gained importance as not far away from this region a large bioethanol factory with a production capacity of about 600,000 tons/year has been established. Prices for barley and triticale did not react in the same way as neither crop is used for bioenergy purposes nor plays a major role in the food sector. The area of these crops may be substituted by rye and where the land is suitable by wheat. The considered technical progress for the second scenario (∆ yield) would enhance the effects of the price scenario with the exception of rye.

An elimination of the area payments may induce a 3% reduction for the considered energy crops compared to the price scenario. However, significant effects on cropland allocation are still observed (see Figure 1). Thus the elimination of the energy crop premium has a minor effect on crop activities on the basis of current prices.

Figure 1: Area changes in Barnim compared to the base-run in %

-25,0% -20,0% -15,0% -10,0% -5,0% 0,0% 5,0% 10,0% 15,0% 20,0%

[%]

Rape seed

Wheat

Rye

Barley

Triticale

Maize silage

Cro

ps

∆ premium

∆ yield

∆ price

Wheat and rape seed are the main crops in the Uckermark indicating a better land quality. The dominating crop rotation in this area is rape seed, wheat and barley or rye depending on market prices (see Table 1 and 2). The results of the scenario runs are presented in Figure 2. Areas of rape seed and wheat increase but are different from the results of Barnim. Area changes in the first scenario (∆ price) are lower demonstrating that there is little scope left to extent the current crop rotation. The area increases for rye and corn silo are negligible and show that these crops are not competitive at current prices. The elimination of the area payment (∆ premium) leads to cropland substitution.


162

Technical progress has a significant impact on rape seed production. Considering crop rotational aspects rape seed production will approach a limit in the near future. Areas for triticale and barley decline most compared to other crops due to the very small activity levels of these crops (see Figure 2).

Figure 2: Area changes in Uckermark compared to the base-run in %

-35,0% -30,0% -25,0% -20,0% -15,0% -10,0% -5,0% 0,0% 5,0% 10,0% 15,0%

[%]

Rape seed

Wheat

Rye

Barley

Triticale

Maize silage

Cro

ps

∆ premium

∆ yield

∆ price

Conclusions

Renewable resources are of importance in our modern society due to their positive effects on agriculture, the environment and the economy. The EU promotes the cultivation of energy crops. Germany has introduced further incentives like the Renewable-Energy-Law which guarantees fixed energy prices for electricity produced from biomass. This kind of promotion creates alternative income sources for farmers as primary or energy producers and strengthens added value and employment. The question is whether it starts to change the landscape of agriculture ?

Concerning this issue a model is developed to assess the potential impacts of energy crop production on cropping activities. A basic quadratic version of PMP is used to maximize total gross margin in two regions which allows to simulate farmers` behavior under different scenarios. Although the approach is simple and has its limitations, it is used because there is a lack of adequate data because the potential to utilize crops as sustainable energy source has lead to a bioenergy boom where cropped areas and market prices are in constant flux.

Results show increasing cropping areas of rape seed and wheat for all scenarios. Rye and maize silage increase or decrease depending on the scenario and region. Barley and triticale are substituted by more profitable crops. The area expansion of rape seed production can be attributed to the increased demand for biodiesel, whereas high yield prices are the result of rising world market prices.

Consequently, the bioenergy boom has partially contributed to crop substitution effects. Potential energy crops like rye, rape seed or silage maize are more profitable. This has an impact on crop rotations because less profitable crops are substituted. However, these tendencies approach a limit in terms of phytosanitary and crop rotational aspects. For instance, area changes in the Uckermark scenario are lower demonstrating that there is little scope left to extent the current crop rotation. Thus energy crops provide an opportunity as alternative income source besides food production in the scope of multifunctional agriculture.


163

References Amtlich Topographisch- Kartographisches Informa-Tions-System (ATKIS) (2005): Basis-DLM, ATKIS

® Landes-vermessung und Geobasisinformation Brandenburg, Germany. Arfini, F. (2001): Mathematical Programming Models Employed in the Analysis of the Common

Agricultural Policy. Working Paper n. 9, Department of Economics, University of Parma, Parma, Italy.

Heckelei, T. and Britz, W. (1999): Maximum Entropy Specification of PMP in CAPRI. CAPRI Working

Paper, University of Bonn, Germany. Howitt, R.E. and Mean, P. (1983): A Positive Approach to Micro-economic Programming Models.

Working Paper 6, Department of Agricultural Economics, University of California, Davis, CA, USA.

Howitt, R.E. (1995): Positive Mathematical Programming. American Journal of Agricultural Economics,

77, pp. 329-342. Kuratorium Für Technik und Bauwesen in der Landwirt-Schaft (KTBL) (2006): Energiepflanzen – Daten

für die Planung des Energiepflanzenbaus. Darmstadt, Germany. Landesamt Für Verbraucherschutz Landwirtschaft und Flurneuordnung (LVLF) (2005): Datensammlung

für die betriebswirtschaftliche Bewertung landwirtschaftlicher Produktionsver-fahren in Brandenburg. Frankfurt (Oder), Germany.

Paris, Q. (1988): PQP, PMP, Parametric Programming and Comparative Statics. Chap.11 in Notes for AE

253. Department of Agricultural Economics, University of California, Davis, CA, USA. Paris, Q. and Howitt, R.E. (1998): An Analysis of Ill-Posed Production Problems Using Maximum

Entropy. American Journal of Agricultural Economics, 80, pp. 124-128. Varela Ortega, C.; SUMPSI, J.M.; GARRIDO, A.; BLANCO, M.; IGLESIAS, E: (1998):Water Pricing

Policies, Public Decision Making and Farmers` Response: Implications for Water Policy. Agricultural Economics, 19, pp. 193-202.

Zentrale Markt- und Preisberichtstelle (ZMP) (2006): Jahres-bericht 2006/2007 Rückblick und Vorschau

auf die Agrarmärkte. Bonn, Germany.

IFMA 16 – Theme 2 Agrarian Vs Rural: Economies and Settlements

164

RECREATING LOCATION FROM NON-SPATIAL DATA –SAMPLE SIZE REQUIREMENTS TO REPRODUCE THE LOCATIONS OF FARMS IN THE EUROPEAN FARM

ACCOUNTANCY DATA NETWORK

Martin Damgaard IAMO (Leibniz-Institute of Agricultural Development in Central and Eastern Europe),

Germany Email: [email protected]

Abstract Individual farm accountancy data sources such as the European Farm Accountancy Data Network (FADN) include no specific information on the spatial location of farms. However, spatial characteristics and site conditions determine the farms’ production potential and its influence on the surrounding environment. Spatially explicit models that make use of the FADN data need to be able to recreate a landscape including the location of the farms in a plausible way. This paper investigates the minimum sample size of farm locations required to insure the ability to reproduce a reliable map of a given region. This is done by analysing relative locations between all the 1871 farms present in the Danish river Gudenå watershed. As we have detailed information about each of the farms we can categorize the farms in groups in a way similar to what one would be able to do with farms from a FADN sample. By utilising the rich information that the FADN sample contains to create a multidimensional spatial set of requirements that the farms on average have to meet it is possible to reduce the number of available locations to a minimum. This investigation is divided into the following two-step procedure: First the variability of an individual farms spatial relationship is investigated with regard to variation in sample size and composition. Secondly is the average values investigated with regard to variation in sample size and composition. Keywords: FADN, spatial location, methodology

Introduction To recreate a reliable representation of the complex reality is one of the fundamental challenges in creating empirically founded models. Numerous models are based on abstract representations of the underlying system and do not need the empirical foundation for investigating the characteristics of the object of study. However once the findings from the models are used for policy recommendations, realistic and empirical founded models are preferred. Obtaining sufficiently empirical data for large regional models through personal field studies are seldom possible. Most models are instead relying on available data from databases or other collectively gathered information. The accuracy of these data differs however. Many of the most adequate economic data are collected by the local authorities indirectly through the assessment of taxes or similar administrative issues. This means however, that the most reliable data are at times restricted to insure personal privacy. The European Farm Accountancy Data Network (FADN) is one of these large but restricted data collections. Every year a large sample of farm accounts is collected in each of the member states in the European Union. From this base sample a number of so-called “representative” farms are found. Each with an extrapolation factor constructed in such a way that the farms provide a representative sample for the commercial farms in a given region. The extrapolation factor incorporates the regional characteristics, the economic size and type of farming found in the whole collection. The term “representative” as well as the


165

accuracy of the methodology is up to debate within the scientific community (Beer et al. 2001; Meier, 2004; Meier, 2005) however will not be questioned here. The sensitive nature of the micro-economic data within the FADN sample means, that the data comes with no other specific geographical reference than which region/ country the collective sample represents. However the spatial nature of agricultural production means that both the farms’ production potential as well as its impact on the surrounding environment makes it vital for a potential modelling application based on FADN-data to recreate the plausible spatial locations of the farms in the sample. A few attempts based on indirect statistics have previously been published (Fais et al., 2005; Fais & Nino, 2004). One of the most ambitious attempts is undoubtedly the work done by the Seamless project (Elbersen et al. 2006). The methodology developed here is also making use of statistics and remotely sensed data. However the restricted nature of the FADN data sample makes it difficult to validate the findings. The present analysis is therefore taking a novel approach. Rather than working directly with the FADN-sample and thereby not knowing the underlying reality that the sample describes, this study is using a sample of 1871 farms located in the Danish watershed to river Gudenå. Both the exact location as well as production data for all the 1871 individual farms are known with similar categories as offered in the FADN sample, with the exception of the economic data present in the FADN sample. Throughout the rest of this paper we are assuming that the “representative” farms found in the FADN sample and their extrapolation factors create a perfectly fitting description of the 1871 farms found in the river Gudenå watershed. Although this assumption is rather unrealistic it is similar to the normal confidence one has to have in the FADN sample, when no other information is available. This perfect sample consists of the production data in our database of the 1871 farms, with the exception of the geographical references. Our task is to investigate the sample size of farm locations required to insure the ability to reproduce a reliable map of the region. This is done by analysing the relative location between all the 1871 farms present in the Danish river Gudenå watershed. As we have detailed information about each of the farms we can categorise the farms in groups similar to what one would be able to do with farms from a FADN sample. By utilising the rich information that the FADN sample contains to create a multidimensional spatial set of requirements (such as the distance to the nearest dairy farm or to the 2nd nearest farm between 0 ha. and 20 ha.) that the farms on average have to achieve it is possible to reduce available locations down to a minimum. In this case we utilize data of the farms size, production type and number of animals units. The rest of the paper is structured as follows: in section 2 an introduction to the study area as well as introduction to the empirical data is given. The difficulties in relying only on remotely sensed data and landscape characteristics, such as topographical maps, soil maps and road system maps for this particular case are presented. In section 3 the outline of the analysis is presented. In section 4 the results are presented and in the following section further applications as well as the difficulties in constructing the maps practically is shortly discussed. Finally a conclusion is made. Introduction to the study area and the empirical data

The valleys of “Nørreå” and “Gudenå” are located in the central part of Jutland between three major cities: Aarhus, Viborg and Randers. The area covers over 76600 ha. 1871 farms on 72089 ha of arable land and 5089 ha of grassland on an average size of 41 ha are for most of them (62%) performing field crop farming. The other farms are then quite equally distributed among dairy farming (11%), grazing livestock farming (6%), granivores (14%) and mixed farming (7%).


166

Figure 1. Map of the study area. The dark marked area shows the involved farms and their fields. All fields belonging to farmers in the area are included even if the location of the field is outside the watershed. Note that due to Danish area requirements fields very far from the farmstead can still be favourably owned.

The study area was chosen partly due to the data availability and partly due to the landscape characteristics. In contrast to a large number of other areas is this region lacking strong spatial indicators by which the available space for farm locations could be deduced. This becomes apparent when one compares the Danish river Gudenå watershed region with other regions where the landscape characteristics can help in locating the farms through e.g. the topography. The outline of the analysis

The location of a farm in space can be defined as an individual event independently of all other farms or structures in the vicinity. However such an analytical framework would not only reduce the historical process in creating the present agricultural structure out of the empirical data it would at the same time also reduce a large part of the knowledge we have of the present farms. Even the freedom of action of the present farms will to a varying degree be determined by its history to its actual state as well as by the actions and history of other agents in the area. So although it would be reckless to claim that the location of a given farm will directly tell us much about the neighbouring farms it can still reveal some elements of an indirect relationship between the farms. Often local experts will be able to locate a given farm type to a small part of the region simply because farms are not randomly distributed in space but tend to cluster around certain areas. This means that we should be able to utilise this information, when we are going to recreate the distribution of farms in a given region. In the case of FADN farms however the difficulty is that we always start off with a sample seldom know what characterizes this particular selection. Therefore this investigation is conducted in such a way, that influence of both the sample size as well as the composition of the sample is analysed. The incomplete knowledge one has in working with FADN data makes some kind of up-scaling or extrapolation of the location of all farms from the initial sample unavoidable. We will here make use of a similar framework of thought as used in resampling techniques, such as jackknife or bootstrap as we investigate the possible level of error such extrapolations could lead to. At the same time will we utilise the rich information that the FADN sample contains to create a multidimensional spatial set of requirements that the farms on average have to achieve and thereby


167

exploit the possibility to reduce the available farm locations down to a minimum. The procedure will therefore draw upon interrelationships between the farms rather than the spatial characteristics of the individual farm. It is often beneficial to include such spatial characteristics. For reasons of simplification will these characteristics not be included in the following. Here only the interrelationship between the farms spatial location is utilized as the location of the farmstead is viewed as a network of interrelated points in space. The network is represented as graphs that consist of a set of vertices (or nodes) connected by a set of edges (links). Here the vertices represent the farmsteads, and the links between the points represents the Euclidean distance between those farms. Each farmstead holds information about the farm size and the production system. This information is used to categorize a given farms relationship to the 1870 other farms (such as the 2nd nearest dairy farm or the nearest farm between 51 ha. and 100 ha.). Therefore the edges are directed lines, as the interrelationship is not symmetrical. This means that the investigated network consist of 3498770 (or 1871*1870) links. The investigation of the network is divided into the following two-step procedure: First the variability of an individual farms spatial relationship is investigated with regard to variation in sample size and composition. Secondly is the average values investigated with regard to variation in sample size and composition. To understand the chosen procedure it is important to remember that our enterprise is to investigate the minimum sample size of farm locations required to reproduce a reliable map of a given region. Therefore we will mimic the situation, where one is collecting data in the field by varying the sample size and this has been repeated with different order of the farms at least ten times. The latter is done as we can’t be certain in what order the farms are chosen if one is collecting the data in the field. Though the number of different selections of farms from a combinatory point of view hardly scratches in the surface of possible orderings, the sample size will still provide us with some insights into the variation one normally will encounter.

The Analysis We look at an individual farm by investigating the variability in the statistical properties in its relative location to all the other farms due to sample size and composition. This is done by taking approximately 10% of all the farms and for each of these farms calculating the Euclidean distance to all the 1871 farms in the region. For each of the 188 selected farms has the most commonly used descriptive statistics (including: mean, median, standard error, 95% confidence level, standard deviation) been calculated for sample sizes varying from 11 farms (the selected farm and 10 additional farms) and up to all the 1871 farms. This is done with an interval of 10 farms. In addition it is done for 11 different successions of the farms. The values are calculated based on the distance and no further categories have been made. The reliability of the values for each individual farm can be assessed through this calculation. This is important as the further analysis eliminate the uncertainty each individual farm constitutes in an incomplete sample. This uncertainty will however unavoidably be included in a sample solely building upon FADN data. In figure 2. a plot of the relative deviation of the mean as a function of the sample size is presented. In figure 4 a plot of the relative deviation of the median as a function of the sample size presented. In the case of the relative deviation of the mean are the first plot supplemented by an additional plot (figure 3.) of the frequency by which the different relative deviations occur. Please note the scale of the frequency plot, as the scales are not made with equal intervals.


168

Figure 2. The relative deviation of the mean as a function of the sample size. Own calculations

Figure 3. The frequency of the values of relative deviation of the mean. Own calculations.

0-10 11-50 51-200 201-500 501-1000 >1000


169

Figure 4: The relative deviation of the median as a function of the sample size. Own calculations.

Please note the difference in scales used in the plot for the relative deviation of the mean and the plot of the relative deviation of the median. Looking at figures 2-4 one can see that once the sample size is around 20% of the full sample (≈400 in this case) the individual farm values are generally reliable. Even earlier are the majority of values within a 10% span in the case of the mean. The median values are naturally fluctuating within a larger span, however otherwise show similar structural characteristic. When working with samples of less than 10% of all the farms the fluctuations within both the mean and median values are so large, that one hardly can trust ones findings to any significant degree. The first part of the investigation has shown the reliability of the values for each individual farm, while varying the sample size and composition. In the real world this variability would be a part of the uncertainty entering into the average values now to be investigated. Here it would however only blur our findings. The entire network is therefore used in the second part of the investigation. Each of the 1871 farms knows now the Euclidean distance to all others. That means that the distance each individual farm contributes with is founded on perfect information.


170

Table 1: List of categories used in this study

Categories

All farms 0-20ha farms

21-50ha farms 51-100ha farms

101-200ha farms More than 200ha farms Plant production farms

1-50 animal unities More than 50 animal unities

Pork Dairy

The variations in the average values are only due to the size and composition of the selected sample. The further procedure is making use of the included production related data. As we know the production category for both the farm working as our point of reference as well as all the other farms we have created a 2-D matrix with the categories seen in table 1 on each side. In the case of the point of reference only the categories that the particular farm fulfils are in use. For all the other farms the scheme is expanded by the

subcategories distance to the 1st , 2nd, 3rd , 4th and 5th nearest farm of the category as well as the average distance. Below are two examples (figure 5-6 and 7-8) of what the ten different successions of farms produce. The two chosen examples are the distance to the nearest farm (figure 5-6) and the average distance to all other farms (figure 7-8). In figure 5 and 7 are the nominal values presented. The percentile deviation from the full sample are presented in figure 6 and 8. The examples reveal mainly two general characteristics. First of all one can see the modifications that the selections produce. Secondly and more importantly is that the precision of course depend upon the number of farms falling into a given category. Only a fraction of the farms will influence the value of the nearest farm, where as all other farms will affect the average value. This simple fact makes a large number of the possible categories, one can make for a given region, questionable for the purpose considered here. If only a few farms fall into a given category the fluctuations for this group will simply be too large for one to rely on the results. However instead of dismissing such findings altogether the different categories should be supplemented with a weight factor expressing the reliability. Such a weight factor can of course only be an estimate and may be based on studies similar to this one.


171

Figure 5: The deviation of the distance to the nearest farm of all other farms for ten different successions of farms. Figure 6: The relative deviation of the distance to the nearest farm of all other farms for ten different successions of farms. Figure 7: The deviation of the average distance to all other farms for ten different successions of farms.

250

300

350

400

450

500

20 120

220

320

420

520

620

720

820

920

1020

1120

1220

1320

1420

1520

1620

1720

1820

sample size

Euclid

ean d

ista

nce [m

]

-40%

-30%

-20%

-10%

0%

10%

20%

30%

40%

20 120

220

320

420

520

620

720

820

920

1020

1120

1220

1320

1420

1520

1620

1720

1820

sample size

% d

evia

tion

10000

12000

14000

16000

18000

20000

22000

20 120

220

320

420

520

620

720

820

920

1020

1120

1220

1320

1420

1520

1620

1720

1820

Sample size

Euclid

ean d

ista

nce [m

]


172

Figure 8. The relative deviation of the average distance to all other farms for ten different successions of farms. As one can see from the above plots the different ordering of farms will fluctuate around the values for the complete region with a spread that diminish with the larger sample size. This spread have we used to pass on the reliability of a number of the different categories (presented in table 1) and results are shown in appendix 1. For the category “All Farms” as well as the five field size categories are the relative difference between the maximum and minimum values presented for the sample size 20, 100, 400 and 1000. This is done as a function of the average value for all the 11 categories used in this study. From the shown values in appendix 1. one can see that the size of the fluctuations to a far larger degree depend on the farms chosen as the point of reference than the different categories, that the rest of the farms are categorized under. Because the differences between the tables are much larger than the deviations between the categories. Once more this is due to the number of the individual farms that fulfils a given type description. This is apparent when the values of the most common groups are compared with the less common groups, such as the 24,97% spread for the sample size 20 for 21-50 ha farms against the category “All farms” where as for the group >200 ha is the value 466,76% for the same. At the same time can one also see that some groupings such as the group 0-20 ha farms and 21-50 ha farms produce better results than the “All farms” group. This demonstrates that some of the sub-groupings one can make of the FADN sample can actually reveal better insights to the spatial distribution of the farms in a region than using only averaged considerations. Discussion Neither the presented method nor the more commonly used methods based on indirect statistics and remotely sensed data will ever be able to recreate a 100% accurate location of the farms in a region as long as “representative” farms from the FADN sample are used. The challenge is to find the most reliable method. Each methodology has its strengths and weaknesses that differ from location specific settings. The actual procedure of using the data holds another set of challenges. Guiding the location of farms by average values will of course produce false locations. The question is however whether it reduces mistakes to a larger degree than a random location procedure would produce. A question we hope to investigate in the near future.

-40%

-30%

-20%

-10%

0%

10%

20%

30%

40%

20 120

220

320

420

520

620

720

820

920

1020

1120

1220

1320

1420

1520

1620

1720

1820

sample size

% d

evia

tion


173

Conclusion The ability to find values through supplementing field studies to help the location of farms for FADN-based spatial models has been demonstrated. Both the variability of an individual farms spatial relationship as well as the average values of farm categories found in the FADN sample has been investigated with regard to variation in sample size and composition. Acknowledgement The research for this paper has received funding from the European Commission’s 6th Framework Programme (MEA-Scope, STReP No. 501516). This publication reflects only the views of the author. The Community is not liable for any use that may be made of the information therein. We are grateful for the data supplied by Chris Kjeldsen, DIAS. References Beers, G. et al. (2001): Pacioli 8 –Innovation in the FADN, Report 8.01.02 Agricultural Economics

Resarch Institute (LEI), The Hague Elbersen, B et al. (2006): Protocols for spatial allocation of farm types, SEAMLESS No.010036 Deliverable number: PD4.7.1, Wageningen, NL Fais,A & Nino,P.(2004): Mapping the Spatial Distribution of Plant Diseases, 24th Annual ESRI

International User Conference Proceedings: pap1820. San Diego, California USA. Fais,A et al. (2005): Microeconomic and GEO-Physical data integration for Agri-environmental analysis,

georeferencing FADN data: A case study in Italy, Paper prepared for the XIth seminar of the EAAE “The Future of Rural Europe in the Global Agri-Food System”, Copenhagen, Denmark, 24-27 August, 2005

Meier, B (2004): The role of cash flow indicators in understanding farm households, Paper prepared for

OECD “Workshop on Information Needs for the Analysis of Farm Household Income Issues”, Paris, France 29-30 April 2004

Meier, B (2005):“Organic” Sampling and Weighting in Farm Accountancy Data Networks –A Discussion

Note on Standard Gross Margins and Calibration, from “Towards a European Framework for Organic Market Information” Proceedings of the Second EISfOM European Seminar, Brussels, November 10-11, 2005


174

Appendix 1:

All farms

Summary: 20 100 400 1000

All FARMS 50.57% 12.53% 3.38% 2.25%

0-20ha farms 50.34% 12.72% 3.54% 2.30%

21-50ha farms 50.46% 12.07% 3.98% 2.38%

51-100ha farms 51.07% 12.03% 3.20% 2.04%

101-200ha farms 51.03% 14.23% 2.44% 1.90%

more than 200ha farms 50.56% 14.08% 4.01% 2.53%

plant_production farm 50.69% 13.83% 3.00% 2.17%

1-50 animal unities 50.46% 12.06% 3.70% 2.30%

more than 50 animal unities50.47% 11.57% 3.78% 2.28%

pock 50.56% 12.42% 3.93% 2.41%

dairy 50.49% 11.58% 3.78% 2.23%

0-20ha farms

Summary: 20 100 400 1000

All FARMS 33.32% 7.55% 4.18% 2.21%

0-20ha farms 34.33% 7.66% 4.16% 2.21%

21-50ha farms 34.94% 6.80% 4.44% 2.22%

51-100ha farms 29.12% 7.08% 4.60% 2.18%

101-200ha farms 29.34% 11.01% 5.03% 2.22%



1-50 animal unities 34.04% 6.98% 4.27% 2.20%


pock 36.01% 7.56% 4.56% 2.32%

dairy 33.42% 6.01% 4.52% 2.15%


175

21-50ha farms

Summary: 20 100 400 1000

All FARMS 24.97% 10.89% 7.06% 2.44%

0-20ha farms 26.72% 11.25% 6.59% 2.32%

21-50ha farms 25.13% 10.39% 6.66% 2.10%

51-100ha farms 20.97% 10.98% 8.77% 2.87%

101-200ha farms 28.92% 14.62% 7.79% 3.35%



1-50 animal unities 23.56% 10.10% 7.06% 2.28%


pock 26.96% 11.05% 6.75% 2.19%

dairy 20.51% 9.37% 8.12% 2.38%

51-100ha farms

Summary: 20 100 400 1000

All FARMS 304.90% 76.24% 62.59% 18.08%

0-20ha farms 305.09% 76.24% 62.65% 18.04%

21-50ha farms 299.00% 76.07% 62.82% 18.01%

51-100ha farms 305.41% 76.22% 62.35% 18.23%

101-200ha farms 322.47% 77.18% 61.93% 18.28%



1-50 animal unities 301.14% 75.95% 62.69% 18.05%


pock 301.58% 76.05% 62.90% 18.07%

dairy 297.17% 75.69% 62.72% 18.13%


176

101-200ha farms

Summary: 20 100 400 1000

All FARMS 132.39% 87.32% 97.66% 45.47%

0-20ha farms 132.21% 86.80% 97.75% 45.21%

21-50ha farms 133.25% 87.21% 97.77% 45.55%

51-100ha farms 132.51% 88.68% 97.09% 46.39%

101-200ha farms 129.74% 87.42% 97.80% 45.12%



1-50 animal unities 133.27% 87.41% 97.64% 45.67%


pock 133.08% 86.84% 98.09% 45.56%

dairy 135.11% 88.61% 97.54% 46.58%

more than 200ha farms

Summary: 20 100 400 1000

All FARMS 466.76% 281.75% 91.46% 50.22%

0-20ha farms 468.18% 283.22% 91.94% 50.19%

21-50ha farms 482.32% 282.26% 92.78% 49.96%

51-100ha farms 459.67% 276.07% 88.91% 50.16%

101-200ha farms 419.13% 282.50% 88.99% 51.50%



1-50 animal unities 477.63% 281.32% 92.09% 50.05%


pock 481.55% 282.93% 92.82% 49.82%

dairy 486.15% 277.33% 91.33% 49.75%


177

ANALYSIS OF BEAN MARKETING CHANNELS IN KENYA AND TANZANIA

Korir, M.K*, Nyangweso, P.M., Serem, A.K., Kipsat, M.J Department of Economics and Agricultural Resource Management,

Moi University, Eldoret, Kenya Email: [email protected]

Maritim, H.K School of Business and Economics,

Moi University, P.O. Box 3900-30100,

Eldoret, Kenya

Abstract The common bean is a major staple food crop in Eastern and Southern Africa, providing dietary protein and calories. This study identified lack of adequate information on the marketing channels of the bean marketing systems in Kenya and Tanzania. The objective of this study, therefore, was to define the bean marketing channels across the borders of Northern Tanzania and Southern Kenya. The study hypothesized that the average bean price in terminal markets is approximately equal to estimated marketing costs. Purposive sampling was used to select two out of five districts of Kilimanjaro province and four out of ten districts of Arusha province. However, systematic random sampling procedures were used to select bean farmers and traders. Structured questionnaires were used to collect primary data from 64 farmers, 78 retailers and 51 bean wholesalers. The gross marketing margins and marketing costs analyses were used to evaluate the beans marketing system. The results show that the dominant marketing channel was from the farmer to upcountry assemblers to wholesalers/long distance wholesalers to wholesaler/retailer to retailer and finally to the consumer. Majority of the farmers (92.1%) produce dry beans for local markets, while 7.9% produce for the export market. A large proportion of the farmers (81.4%) sold their dry beans to upcountry assemblers and farm gate markets. In Arusha market, there was no significant difference between the average marketing cost and the average market price. This indicated that the average market prices approximated the marketing costs. The analysis of the marketing margins showed that the farmers’ share (producer participation) in the price paid by the consumer is 45.65%, while those of the Nairobi long-distance wholesaler and Nairobi wholesaler were 14.88% and 9.65% respectively. These indicate that margins varied with the nature of marketing costs incurred by the various participants. The study concludes that producer participation should be increased by reduction of marketing costs through the removal of Horticultural Crops Development Authority (HCDA) levy and 3.5% import duty by the Kenya government Key words: Bean marketing, marketing channels, costs and margins Introduction The common bean (Phaseolus vulgaris L.) is a major staple in Eastern and Southern Africa, where it is recognized as the second most important source of dietary protein and the third most important source of calories (Wortmann, 1998). Animal protein is seldom affordable by the poor in developing countries, so the bean provides the chief and sometimes the only source of protein. Beans are specifically important as a component to carbohydrate staples such as rice, corn, plantains, cassava, and other cereals, root and tuber crops. The combination of legumes and cereals provides a very good balance of amino acids. Maize and beans together provide a well-balanced protein, beans supplying the lysine deficient in maize and


178

maize providing the sulphur amino acids, cystine and methionine, which are lacking in beans (Mukoko, 1989). Bean consumption in Eastern and Southern Africa exceeds 50 kilograms per person per year, reaching 66 kilograms per person in parts of Kisii, Kenya (Wortmann, 1998). The bean is a readily available and popular food to both the urban and rural population in Uganda. In 1987, the Food and Agriculture Organization (FAO) estimated Uganda’s bean consumption as 29.3 kg. per capita (Kirkby, 1987). However, recent studies show that the per capita bean consumption in Uganda’s Nabongo area is about 58 kg per year (David, 1999). Dry beans can be consumed, boiled alone or mixed with cereal grains, especially maize (to form a meal known as ‘githeri’ in Kenya or ‘makande’ in Tanzania), sorghum or rice. Beans are also mixed with cooking banana especially in Kagera, Tanzania and Uganda, or mashed with Irish potatoes to form ‘mataha’ dish for Kenyans. Green shelled beans, tender leaves and immature pods are some of the forms in which beans are consumed (Korir, 2005; Kosgei, 1998; Ouedraogo et al, 1994 and Rugambisa, 1990). Apart from its primary role of supplying essential nutrients, the common bean is also commercially important. Though the primary objective of small farmers in producing beans is home consumption, the surplus is sold whereupon marketing becomes a major consideration. In central Ethiopia, farmers grow the white pea bean for export as their cash crop (Abebe and Kefene, 1989; Abebe, 1987). This study recognized the importance of the common bean both for domestic consumption and for commercial purposes. The Problem The problem is that little is known about the nature of the bean marketing systems in Tanzania and Kenya. Stakeholders have insufficient knowledge as pertains the marketing channels that beans pass through, from the producer to the consumer; and the nature of marketing margins that accrue to various market participants. Objectives The general objective of the study was to analyse the bean marketing system in northern Tanzania as a surplus area, and Nairobi as a deficit area, and how the two areas are linked by cross border marketing channels. The study aimed at determining the bean marketing channels and evaluating the performance of the bean marketing system by use of marketing costs and marketing margins analyses. Hypothesis The average bean price in terminal markets is approximately equal to estimated marketing costs i.e. Ho: µ = ĉ Methodology The Study Area

The study was done between February 2002 and June 2002 in the northern zone of Tanzania and Nairobi area of Kenya. The northern zone of Tanzania includes Arusha and Kilimanjaro regions (provinces). This zone lies between 350 43' and 380 28' East and 1045' and 600' South, and is an important area in the production of beans, contributing 13% of Tanzania’s national production (Kamau et. al., 2000). Four


179

administrative districts in Arusha region, namely Arumeru, Monduli, Karatu and Mbulu and two districts in Kilimanjaro region, namely Moshi and Hai were covered. Kenya was studied mainly as the area of common bean destination for consumption, with particular focus on Nairobi, the capital city. Within these two countries, the main wholesale and retail markets of Arusha, Moshi, Himo, Namanga, Taveta and Nairobi were surveyed. Arusha and Moshi markets are the largest two markets of the northern zone while Namanga, Himo and Taveta markets are the major exit points of beans from Tanzania into Kenya. Type of Data Primary data relating to the bean varieties traded, quantities offered for sale, selling prices, sources of bean stocks, and transaction costs (handling, duties and levies, transportation, and storage costs), units of measure, and the exchange rates were collected for analysis. Data Analysis

The Statistical Package for Social Scientists was used to generate descriptive statistics on trader characteristics. These included the various sources of bean stocks by traders, the bean market outlets, transport costs, mean marketing costs and mean market prices. The Microsoft Excel program was used to generate the gross marketing margins for each of the market participants. The marketing margins and marketing costs analyses were used to evaluate the performance of the bean marketing system in place. Results and Discussions Bean Marketing Channels in the Study Area The bean commodity was established to flow from the northern zone of Tanzania hinterland into the regional market centres of Arusha and Moshi. From these markets, the beans flow northwards to Nairobi, through Namanga border point. However, other stocks flow to Mombasa (via Taveta), Tanga, Dar es Salaam and Zanzibar. The majority of the farmers (92.1%) produce dry beans for local markets, while 7.9% produce for export market. The farmers who produced for export were those contracted by Pop Vriend to produce Mexican 142, for the canning industry. The survey revealed that these beans were locally cleaned in Tanzania and exported for canning abroad. In the year 2001, 81.4% of the farmers sold their dry beans to upcountry assemblers and farm gate markets. The long-distance wholesalers sourced their bean stocks from assemblers and wholesalers and did the bulk of cross border bean export trade. The dominant marketing channel that the beans were established to flow was from the farmer to upcountry assemblers to wholesalers/long distance wholesalers to wholesaler/retailer to retailer and finally to the consumer. Regional Transportation Costs The transport costs for beans from the farm gate to terminal markets and within regional markets were inquired. These costs are shown in table 1.


180

Table 1: Regional Route Specific Transport Costs Per 110 Kg Bag

Rail Road Route Distance (KM)

Tsh. Ksh. US$ Cost/KM (US$)

Tsh. Ksh. US$ Cost/KM (US$)

Mbulu- Arusha

200 - - - - 2,000 156.25 2.08 0.0104

Arusha- Moshi

70 - - - - 1,000 78.10 1.04 0.0149

Arusha- Nairobi

250 - - - - 5,504 430.00 5.73 0.0229

Taveta- Nairobi

486* 1,536 120 1.6 0.0033 - - - -

Source: Authors’ Survey, 2002 *This is the distance by rail. The cost per bag consists siding charges charged at the rate of Ksh. 960 per 36 ton wagon, Ksh. 35,640 fixed charge per wagon and a value added tax of Ksh. 6,588 (18% VAT). The Kenya Railways Corporation charges lower rates of transportation for agricultural commodities than for industrial goods. The cost is based on the distance, wagon capacity, siding or terminal charges and value added tax. At the time of survey, the railway connection from Taveta to Arusha was not operational. The figures show that, in Kenya, rail transport is cheaper than road transport by lorries (it would cost US$ 0.0033/km to transport a bag of beans by railway, and US$ 0.0229/km by road along the Arusha-Nairobi highway). The table also shows that the shorter the distance, the more expensive it is to transport beans. Route Specific Marketing Costs for Beans Imported from Tanzania to Kenya

The analysis of route specific marketing costs enables the judgment of whether bean prices reflect marketing costs and hence gauge the performance of the bean marketing business. The major route specific marketing costs encountered by the traders as the beans are passed through the marketing system, from the farmer to the urban market centres are shown in table 2. The following are the notes that explain the costs in this table: The exchange rate: At the time of the survey, the rate was 960 Tanzania shillings to 1 United States dollar to 75 Kenya shillings. The unit of measure: In Tanzania, bags of beans are sold in ‘100’-kg units. However, weighing scales are not used; rather, an approximation is done by the use of 6-debe tins to imply 100 kg. It was found out that on weighing, this actually yielded 110 kg of beans. This unit (110 kg bag) is therefore used in this analysis. Transport cost: The quoted transport cost per bag from Arusha to Nairobi was Ksh. 430. This, the traders said, included duties at border points. Long distance wholesalers paid this much to the transporters. It was therefore the duty of the transporter to pay any incidental duties at the border point. On arrival at the border point, however, the beans are transported across the border by head portarage, one bag after the other, a practice that does not attract duties. There may, however, be certain unknown


181

unofficial payments to government authorities for importation of goods. Transporters confided that they were often faced by certain embarrassments as they performed their duties, a fact that confirmed this. A certain portion of the Ksh. 430-transport cost may go to these payments. This cost is very high and forms 17.2% of all the marketing costs. At the time of the survey, maize traders sourced their maize from Nakuru and Eldoret towns of Kenya and transported it to Taveta, via Nairobi and Voi, a distance of about 450 km. They hired Kenya Railway wagons, whose capacity is 400 bags at a cost of Ksh. 70,000 (US$ 933 or Tsh. 896,000) per wagon. This represents a unit cost of Ksh. 175 (US$ 2.33 or Tsh. 2,240) per bag. This is evidently a far much cheaper mode of transport. With the revival of the East African Community, this means should be explored, especially so by the long distance wholesalers operating between Arusha and Nairobi. Levies and duties: Tanzania’s export duty of $ 2 per consignment and the cost of phytosanitary certificate of $ 15 per consignment translate to Ksh. 12.75/ bag, if the long distance wholesaler transports a consignment of 100 bags. Exporters can, however, exploit the economies of scale by trading in larger consignments. Kenya’s HCDA levy charged by Kenya’s customs authorities, at the tare of 3.5% of the value of produce translates to Ksh. 63/bag.


182

Table 2: Estimated Route Specific Marketing Costs in the Study Area

Karatu – Arusha

Cost/110 kg Bag

Arusha – Nairobi

Cost/110 kg Bag

No. Item

T.Sh. US$ K.Sh. T.Sh. US$ K.Sh.

1 Purchase 20,607.14 21.47 1,609.93 25,600 26.67 2,000.00

Handling

Reweighing and

rebagging

64.00 0.07 5.00

Loading 300.00 0.31 23.44 256.00 0.27 20.00

2.

Unloading 300.00 0.31 23.44 256.00 0.27 20.00

Tax

District cess/tax 200.00 0.21 15.63

Market tax 400.00 0.42 31.25

Export duty &

phytocertificate

163.20 0.17 12.75

HCDA levy (Ksh.

1/kg produce)

1,280.00 1.33 100.00

3

3.5% Import duty 806.40 0.84 63.00

4 Transport 2,000.00 2.08 156.25 5,504.00 5.73 430.00

5 Storage 200.00 0.21 15.63

6 Lodging and meals

50.00 0.05 3.91 256.00 0.27 20.00

7 Total cost 24,057.14 25.06 1,879.46 31,936.00 33.27 2,495.00

8 Selling price 25,600.00 26.67 2,000.00 36,678.40 38.21 2,865.5

9 Marketing Margin/Bag

1,542.86 1.61 120.54 4,742.40 4.94 370.50

Source: Authors’ Survey, 2002 In this analysis, duties and levies are not added to the total marketing costs because the marketing is largely informal and therefore, does not attract them. This study hypothesized that the bean prices in terminal markets reflected the marketing costs. A sample of 10 wholesalers at Arusha had a mean wholesale price of Tsh. 25,600 with a standard deviation of 3,627.05. By use of a two tailed test, the marketing costs and mean market prices were tested for significant difference at 95% confidence level. In Arusha market, there was no significant difference, hence the hypothesis that market prices reflect marketing costs was accepted.


183

Marketing Margins Analysis

Marketing margins for the different market participants at different levels in the marketing chain were calculated. Table 3 shows the marketing margins per bag and the corresponding % margins.

Table 3: Marketing Margins for Various Participants in the Study Area

Price /110 kg bag Margin/110 kg bag Market participant

Ksh. Tsh. US$ Ksh. Tsh. US$

% Margin

Farmer (Tanzania)

1,610 20,607 21.46 1,610 20,607 21.46 45.65

LD W/Saler* (Arusha)

1,850 23,680 24.66 240 3,072 3.20 6.81

W/Saler (Arusha)

2,000 25,600 26.66 150 1,920 2.00 4.20

LD W/Saler (Nairobi)

2,525 32,320 33.66 525 6,720 7.00 14.88

W/Saler (Nairobi)

2,866 36678 38.21 341 4,358 4.54 9.65

Retailer (Nairobi)

3,526 45,140 47.02 661 8,462 8.81 18.70

Source: Authors’ Survey, 2002 *LD: Long distance; W/Saler: Wholesaler From the analysis above, the farmer’s share (producer participation) in the price paid by the consumer is about 45.65%.; i.e. the farmer is getting 45.65% of the price that the final consumer pays. This margin is rather low. Mendoza 1995, studied the marketing margins for potatoes grown in the North of Chuquisaca, Bolivia, and found out that the producer participation was 54%. This result indicates that there is need to look into ways of reducing the marketing costs, so that the producer’s share can be increased. The long distance wholesaler in Nairobi is getting 14.88% of the consumer’s price. Although this looks large, the transportation cost forms the bulk of this share. These shares generally reflect the kind of marketing functions and services the market participants have performed. For example, the wholesaler at Arusha has a share of only 4.2%. This wholesaler gets this little because he just buys the beans, stores, and sells it, with no transportation or sorting costs. In contrast, the Nairobi retailer gets a higher share of 18.7% for she is involved in a thorough sorting and cleaning exercise, transportation, and payment for watchmen and city council license fees, which the consumer has to pay for. To gain further insight into the margins of various participants, typical farmers’ and traders’ gross margins were analysed. The results show that it costs a farmer Tsh. 17,976.80 to produce a 110 kg bag of beans. This bag sells for Tsh. 20,607 at the farm gate. The farmer therefore gets Tsh. 2,630.20/110 kg bag for his management, which is equivalent to 5.83% of the price the consumer pays at the retail market.


184

Conclusions and Recommendations The major markets for beans from northern Tanzania are Arusha, Tanga, Dar es Salaam, and Zanzibar towns of Tanzania, and Nairobi and Mombasa towns of Kenya. The dominant marketing channel that the beans were established to flow was from the farmer to upcountry assemblers to wholesalers/long distance wholesalers to wholesaler/retailer to retailer and finally to the consumer. Although the bean marketing system is generally efficient with market bean prices reflecting the marketing costs, the producer participation is low. The study recommends the reduction of marketing costs, and thereby increasing the producer participation, by the removal of HCDA levy and the 3.5% import duty by the Kenya government. Acknowledgements We acknowledge the International Centre for Tropical Agriculture (CIAT) and the East and Central African Bean Research Network (ECABREN) for funding this research. References Abebe, A. (1987). Bean Production and Research in Ethiopia. In Kirkby, R.A. Proceedings of a Workshop on Bean Research in Eastern Africa, Mukono, Uganda, 22-25th June 1987.

CIAT African Workshop Series No. 2, pp 22. Abebe, A. and Kefene, H. (1989). Country Reports-Eastern Africa: Ethiopia. In Smithson, J.B. Proceedings of a Workshop on Bean Varietal Improvement in Africa, Maseru, Lesotho,

30th January-2nd February 1989. CIAT African Workshop Series No. 4, pp 114-119. David, S. (1999). Beans in the Farming System and Domestic Economy of Uganda: A Tale of Two Parishes. Network of Bean Research in Africa. Occasional Publications Series

No. 28. CIAT. Kampala. Kamau, M.W., Ndakidemi, P.A., and Muimu, J. (2000). A Bean Market Survey in Arusha and Kilimanjaro Regions of Tanzania: A Comparison of Production, Trading and Consumption

of Bean Varieties. Unpublished paper. Kirkby, R.A. (1987). Proceedings of a Workshop on Bean Research in Eastern Africa, Mukono, Uganda, 22-25th June 1987. CIAT African Workshop Series No. 2, pp 2. Korir, M.K. (2005). Cross Border Bean Marketing Between the Northern Zone of Tanzania and Nairobi, Kenya. M.Phil. Thesis. Moi University. Eldoret. Kosgei, D.K. (1998). The Marketing of Beans: An Assessment of the Structure and Conduct of Beans Marketing System in Nandi District, Kenya. M.Phil. Thesis. Moi University.

Eldoret. Mendoza, G. (1995). A Primer on Marketing Channels and Margns. In Scott, G.J. Prices, Products and People: Analyzing Agricultural Markets in Developing Countries. Lynne

Rienner Publishers, London. Mukoko, O. (1989). Country Reports-Southern Africa: Zimbabwe. In Smithson,


185

J.B. Proceedings of a Workshop on Bean Varietal Improvement in Africa, Maseru, Lesotho, 30th January-2nd February 1989. CIAT African Workshop Series No. 4, pp 201.

Ouedraogo, I., Kere, P., Osore, J., and Matheka, F. (1994). Dry Beans Sub- Sector in Kenya: A Rapid Appraisal with Emphasis on Market Information Needs and Extension Issues.

Nairobi Market Information System Report No. 94-02. Government of Kenya Market Information System.

Rugambisa, J. (1990). Marketing of Beans in Sub-Saharan Africa and Impact of Market on New Cultivars. In Smithson, J.B. Progrss in Improvement of Common Bean in Eastern and

Southern Africa. Proceedings of the Ninth SUA/CRSP and Second SADCC/CIAT Bean Research Workshop, Sokoine University of Agriculture, Morogoro, Tanzania, 17-22 September, 1990. CIAT Africa Workshop Series No. 12.

Wortmann, C.S., (1998). Atlas of Common Bean (Phaseolus vulgaris L.) Production in Africa. CIAT Publication No. 297.


186

EVALUATION OF STRATEGIES TO ACHIEVE COMPLIANCE WITH A LEGAL RISK ASSESSMENT DOCUMENT BY FARMERS IN IRELAND

.John McNamara,

Teagasc – Agriculture and Food Development Authority, Kildalton College, Piltown, Co. Kilkenny, Ireland


James Phelan University College Dublin, Ireland.

Patrick Griffin and Anthony Morahan Health and Safety Authority, Ireland.

Frank Laffey

Farm Safety Partnership Advisory Committee, Ireland.

Abstract Recent legislation in Ireland permits farmers who are self-employed or who employ three or less employees to meet legal duties regarding safety and health management by complying with the terms of a Code of Practice and completing a Risk Assessment Document. A three year National Initiative commenced in 2005 to develop the Code of Practice and Risk Assessment Documents and to evaluate strategies to assist farmers to effectively complete and implement their legal requirements. Preliminary findings of an evaluation of the initial phase of the Initiative are presented in this paper. The evaluation was conducted among farmers who attended a half-day training course on completing and using the Risk Assessment Document are compared with a group of farmers who completed it without training. The study findings indicate that 74% of farmers who returned the document for evaluation completed it satisfactorily. Satisfactory completeness rates were similar whether a training course was or was not attended. However, 100% of participants stated that attendance at the training was worthwhile. The on-farm evaluation found that just over 24% of the farms were not achieving a satisfactory standard of safety and health management and this was unrelated to the level of completeness of the document or attendance at a training course. Further research is required to determine what further assistance is required by farmers who either do not complete the Risk Assessment Document or achieve a satisfactory standard of safety and health management. Key Words: Farm Safety; Health; Risk Assessment; Training; Safety Legislation. Introduction In Ireland the issue of improving the safety and health record of farmers presents a major on-going challenge. There are 270,000 persons employed in agriculture on 143,000 Irish farm holdings (CSO, 2001). Approximately 27% of workplace deaths in Ireland occur in agriculture (including forestry) (HSA, 2007), even though just 6.0% of the working population is employed in the sector (CSO, 2003). Regarding non-fatal accidents in Ireland a recent estimate indicates an injury rate per 100,000 farms of 1,800 (McNamara et al, 2007). Health, generally, is neglected by farmers in Ireland (Hope et al, 1999) and farmers have above average mortality rates arising from accidents and ill health (O’Shea, 1997). At an international level, farming has long been ranked as a hazardous occupation and is ranked in the top three occupations with the highest incidence rates of injuries in the United States (DeRoo and Rautiainen, 2000). While injury and ill health leads to tragedy, human suffering and disability, it also has the


187

potential to impact adversely on the farm business (McNamara et al, 2007). McNamara et al, 2007 also reported that farm incomes were reduced, on average, by 15% on farms where the farm operator reported a disability. Physical and mental health also correlates highly among farmers. Melberg, 2003 reported that severe injury, illness or disablement negatively influenced farmer’s mental health. Thus strategies to assist farmers to effectively incorporate injury and ill health prevention into farm management are urgently needed both in Ireland and internationally. A three-year national initiative commenced in Ireland in 2005 to assist farmers to effectively manage safety and health. The commencement of this initiative coincided with the enactment of the Safety, Health and Welfare at Work Act, 2005. This legislation updated previous legislation enacted in 1989 and strongly emphasizes the requirement for active management of safety, health and welfare at all workplaces. Small scale enterprises whose owners are self-employed or which employ three or less employees are allowed to meet legal duties regarding safety and health management by complying with the terms of a Code of Practice and completing a Risk Assessment Document prepared for a specific sector, such as Agriculture (Section 20(8) of the legislation). The initiative is being undertaken jointly by the Health and Safety Authority, the state agency responsible for ensuring compliance with this legislation and Teagasc, the state agency responsible for provision of research, training and advice to the agriculture sector. Farmer’s representatives provided an input into the development of the initiative through membership of a statutory advisory committee to the Health and Safety Authority, known as the Farm Safety Partnership Advisory Committee (FSPAC). Development of the Initiative The overall aim of the National Initiative is to develop a Code of Practice and Risk Assessment document, evaluate strategies to assist farmers to effectively complete and implement the documents and then commence a national programme to assist farmers to comply with the legislative requirements. A Teagasc Health and Safety Officer was appointed as Project Manager to develop and manage implementation of the Initiative within Teagasc and a Health and Safety Authority Inspector had overall charge of the legislative aspects of the project. These officials formed a steering committee to implement the Initiative. The National Initiative had the following phases: Phase 1, (2005 -2007): Develop Risk Assessment Document and evaluate its use and implementation by a sample of 1,000 farmers. Phase 2 (2005- 2006): Develop the Code of Practice Document and conduct the statutory required consultation process for the documents developed in phases 1 and 2. Phase 3 (2007 -2008): Commence a national training programme to assist farmers to comply with the legislative requirements. Development of the Risk Assessment Document and evaluating the effectiveness of its use by farmers was prioritized in Phase 1 of the Initiative as this document is central to assisting farmers to manage safety and health. Previously, in 2003 a non-statutory Farm Safety Self Assessment Document was circulated, by post, to all farmers by the Health and Safety Authority (Health and Safety Authority, 2003). However, this document was completed by only 28.5% of farmers, nationally (McNamara et al, 2006) and no evaluation was undertaken of how satisfactorily this document was completed or the subsequent actions undertaken by farmers to improve safety and heath on the their farms. To initiate the National Initiative in 2005, Phase 1 commenced with development of a pilot Risk Assessment Document, taking into consideration the views of organizations represented on the FSPAC.


188

The pilot Risk Assessment Document consisted of a series of sections where the safety and health control measures were outlined in question format. Farmers were required to identify controls both in place and missing on their farms. Each section was accompanied by a page giving fatal farm accident data in pie chart format related to the specific area. Pictures showing the necessary control measures were also included. A Farm Safety Action List Page was provided where farmers were required to list the control measures not in place and to set a time schedule for their completion. The final Risk Assessment Document developed can be viewed at www.hsa.ie. It was decided to evaluate completion and implementation by farmers of the Pilot Risk Assessment Document following circulation by post. Additionally, as the completion rate of the previously circulated Farm Safety Self Assessment Document was low, it was decided to include in Phase 1 of the initiative provision of a half-day training course on Farm Safety and Health with particular reference to providing training on completion of the pilot Risk Assessment Document and then to evaluate document completion and the implementation of subsequent control measures on farms by course participants. Each half-day training course included: a short introduction on the objectives of the course, a discussion exercise on the causes of farm accidents and a presentation on the causes of fatal farm accidents, viewing of DVD clips where victims described their accident occurrence, a presentation on the key requirements of safety and health legislation and a session of about 3 hours where each section of the Risk Assessment Document was explained and a short DVD on the content of the Risk Assessment Document being considered was shown. Farmer participants were then given time to consider the questions asked in the Risk Assessment Document as they related to their own farm. Each course had an attendance in the 40 to 50 farmers and was facilitated by at least two Teagasc training and advisory staff members. A Farm Safety Handbook already published by the Health and Safety Authority was distributed to farmer participants as an information source on safety and health to accompany the course. Implementation of the Initiative During November/December 2005, training was provided to Teagasc staff in six counties chosen to implement the pilot Initiative. The counties were chosen regionally on the basis of having a high level of fatal farm accidents and because farmers were involved in a range of farming enterprises. Training was provided to the Teagasc County Manager, an Education Officer and approximately six Agricultural Advisers in each county. The role of the Teagasc manager was to manage implementation of the Initiative in each county. Education Officers have a specialized role at county level in providing training and have particular expertise in safety and health training. The role assigned to Education Officers was to present the training courses. Each Adviser provides advice to an average of 120 clients and the role assigned to each adviser was to promote farmer involvement in the Initiative, to assist the Education Officer in the delivery of the course by such means as stimulating discussion and to assist farmers individually and in small groups to complete the pilot Document. Advisers also had the role of providing follow-up advice to clients following the course. A detailed Memorandum of Implementation was developed and supplied to all Teagasc staff involved in the Initiative to ensure that courses were delivered consistently (Teagasc, 2005). During January and February 2006, farmers in five counties were invited to attend a half-day course by their adviser. The invitation indicated that the course would assist farmers to comply with their legal duty of completing a Risk Assessment under the new legislation. It also pointed out that following attendance at a course, participants would be exempted from a routine inspection by HSA inspectors (other than where an accident or dangerous occurrence was reported) for 2006. Courses were free of charge and were advertised jointly by Teagasc and the Health and Safety Authority in the local farming press and on local radio. Approximately 1,500 farmers participated in these half-day courses in Spring 2006. Farmers in the


189

sixth county received no training, but 800 were sent the Risk Assessment Document and requested to complete it. The purpose of this paper therefore is to present initial findings of an evaluation of Phase1 of the National Initiative Objectives of the Evaluation The objectives of the evaluation were: 1) to evaluate the impact of participation at a half-day training course on completion of the Risk Assessment Document; and 2) to evaluate the role of completion of the Risk Assessment Document in assisting farmers to manage safety and on their farms. Methodology The following methods were used to complete the evaluation:

1) An evaluation of farmers and advisors (who participated in the training) perceptions of the utility of the training courses, 290 completed farmer evaluation questionnaire and 27 from advisers were analyzed. 2) An examination of a sample of Risk Assessment Documents completed by farmers after their participation on the training courses. Following attendance at the courses approximately 600 farmers were supplied with a pre-paid envelope and requested to return the document to Teagasc when they had completed it, 336 completed documents were received. 3) Approximately 300 farmers who did not attend a training course (the sixth county) were written to with the request to return the document to Teagasc for evaluation. This process resulted in 137 completed Documents being returned. These Risk Assessment Documents were analysed and a Completeness Score of Satisfactory or Unsatisfactory recorded for each document. A satisfactory score was allocated when all sections of the document were properly completed, with no major inconsistencies. Otherwise documents were regarded as unsatisfactory. 4) An on-farm follow up evaluation was carried out on 66 farms, 49 of whom had obtained training and 17 who had not received training. The purpose of the farm visits was, firstly to check the accuracy of completion of the Risk Assessment Document with the actual safety and health situation on the farm. A check-list was developed and used for the purpose of checking completeness. This was done by modifying and expanding one developed by Teagasc (Teagasc, 1997). An Accuracy Score of Satisfactory or Unsatisfactory was again recorded for each farm. Secondly an assessment of the management of safety and health on each farm was conducted. A Safety Score was allocated to each farm. This score estimated the overall level of management of safety on the farm. Thus this score reflects the long term management of the farm. The following two point Safety Score was used: 1 – Satisfactory; 2- Unsatisfactory. Farms in compliance with Safety, Health and Welfare at Work legislation and where safety and health were being managed on a satisfactory basis were allocated a satisfactory score. Otherwise they were recorded as unsatisfactory. Farm visits were undertaken by two persons, a Teagasc Health and Safety Officer and a Health and Safety Authority Inspector both of whom were qualified and experienced in occupational safety and health as it relates to farms.

All evaluations conducted should be regarded as preliminary due to the short time period between completion of the training and the preliminary evaluation. Further evaluations are planned for 2007.


190

Results Results presented relate to a preliminary evaluation of Phase 1. Because of the lag between training and the subsequent impact of training, evaluations will also be carried out in 2007 and later. Evaluation of the Training Courses

The principal reason given by farmers for attendance (290 responses) were: to comply with legislation (43%); improve safety and health on own farm (47%) and invitation from Adviser (10%). Particularly strong positive responses were received to questions about the importance of safety management, motivation to implement safety and health controls and plans to make safety improvements following attendance at the course. In relative terms, the least positive responses were obtained to questions on the adequacy of discussion among participants, the overall length of the course being about right and the number of participants. Table 1: Farmers Opinions (%) on the Adequacy of the Training Courses (N=290)

5 4 3 2 1 Attendance worthwhile 47 53 0 0 0 Overall length of course about right 25 65 6 4 0 Helped me understand legal duties 42 56 2 0 0 Number of participants satisfactory 33 59 7 1 0 Adequate discussion among participants 19 61 12 8 0 Will complete Document within 2 weeks 37 59 4 0 0 Motivated me to implement measures 50 47 3 0 0 Will make safety improvements. 42 58 0 0 0 Worthwhile to offer to all farmers 66 32 1 0 0 Safety management is important 82 17 0.5 0 0.5

5=Strongly Agree: 4=Agree: 3=Neither: 2=Disagree: 1 = Strongly Disagree. Participant’s perceptions of the training methods used and on the completion of the Risk Assessment Document also indicate very positive responses (Table 2). Use of visual approaches including use of the DVD clip of accident victims describing their accident and those showing safety and heath controls received the highest scores. Aspects receiving the least positive comments included: level of discussion on accidents locally; number of questions in the document and ease of understanding of questions in the document. Notably, a medical doctor’s message (delivered on DVD) on farmer’s health was described as OK or poor by 27% of respondents.


191

Table 2: Farmers Perceptions (%) about Training Methods and the Pilot Risk Assessment Document (N=290)

5 4 3 2 1 Good Use of Power Point 51 47 2 0 0 DVD of victims describing accidents worthwhile 65 33 2 0 0 Discussion on accidents worthwhile 30 55 13 2 0 Completion of document in groups useful 35 59 5 1 0 DVD showing safety and health controls worthwhile 55 43 2 0 0 Number of questions in Document right 17 69 12 2 0 Questions in Document easy to understand 22 65 11 2 0 Pictures aided communication 23 73 4 0 0 Pie Charts showing data useful 41 56 2 1 0 Adequate information provided 24 69 5 2 0 Doctors DVD message about health* 28 45 22 4 1

5=Strongly Agree: 4=Agree: 3= Neither: 2=Disagree: 1= Strongly Disagree. *5=Excellent: 4= Very Good: 3=OK: 2=Poor;1=Very Poor. Almost 22% of farmers stated that they felt they would have difficulty implementing controls in the Risk Assessment Document and two thirds of these respondents outlined their concern. The concerns related to: electrical installations (25%); upgrading of machinery and buildings (20%); costs associated with implementing controls (13%); using a bull chain (13%); chemical container disposal and chemical storage (11%); health/older people/children (9%) and other (9%). Positivity towards the issue of safety and heath is further indicated by the fact that 93% of farmers stated that Teagasc should include health and safety as a topic at seminars and farm walks, while 86% of course participants stated that they were willing to attend a practical on-farm demonstration as a follow-up to the half–day training course. Regarding Advisers perceptions of the training course, 89% considered that their clients considered the course worthwhile and 90% considered the course was well structured. Advisers were requested not to actively raise Safety and Health issues with farmers on the course so that the level of follow-up queries could be gauged. However, just 12% of Advisers reported receiving a high or very high level of queries from farmers. Evaluation of Documents

Document Completeness was recorded as satisfactory for 74% of all documents assessed (473). The documents assembled originated from farms with the following principal farm enterprises: drystock (43%); dairying/drystock (41%) and tillage (16%). Little difference in level of completeness across enterprises was noted. Similarly no significant differences were noted for level of completeness between those who did and did not attend the half-day training session. On-Farm Risk Assessment Document Evaluation and allocation of a Safety Score The On-Farm Evaluation revealed that the Accuracy Score assigned to a document was accurate in 92.4% of cases. This gives confidence that the Risk Assessment Document Completeness Score findings reflect the actual situation on farms. No significant relationship was found for the following variables when related to Document Completeness: farm safety score; number of controls specified in the document;


192

farm enterprise; full or part time farmer; farm size; farmer age category; regular health check undertaken; plans for farm development, if controls had been implemented, or if a Self Assessment Document had been completed previously. Regarding the Safety Score of farms, 74.2% of farms were allocated a satisfactory score. No significant relationship was found between Farm Safety Score and number of controls specified, full or part time farmer, farm size, regular health check undertaken, if farm development was planned or if a Self Assessment Document had been completed previously. A significant relationship was observed between Safety Score and enterprise (P=0.039), with dairy farms on average performing worse than drystock or tillage farms. A significant relationship was also observed between Safety Score and controls implemented (P=.000). Those who received an unsatisfactory Safety Score had implemented no improvements compared to 70% of farms receiving a Satisfactory Score. Safety Scores for Individual farm elements are presented in Table 3. Farmyard/Buildings and Farmer Behavior (77.3% and 78.1% respectively) were the areas causing most problems. Other elements receiving a low percentage of satisfactory scores included safety related to Livestock (82.1%), Electrical facilities (84.9%) and Machinery (89.1%). Where an unsatisfactory Safety Score was allocated an unsatisfactory Farmer Behavior Score was allocated in 93.8% of cases. Table 3. Number of Farms and % receiving a Satisfactory Safety Score for Individual Farm Elements

Safety Score Number Percentage Satisfactory Tractors 63 95.5 Machinery 57 89.1 Livestock 56 82.1 Farmyard/Buildings 66 77.3 Electrical 66 84.9 Chemicals 60 85.0 Health Issues 66 92.5 Protective Equipment 64 87.5 Children 27 92.7 Older Farmer 18 89.0 Farmer Behavior 64 78.1 Safety Score 66 75.8

Discussion The study findings indicate that 74% of farmer respondents completed the Risk Assessment Document satisfactorily. Satisfactory Completeness rates were similar whether a training course was or was not attended. It is worth noting that only, 28.5% of farmers, nationally, completed a similar document when circulated previously (McNamara et al, 2006), so the response to training suggests it has a positive role in assisting farmers’ with Risk Assessment Document completion. The on-farm evaluation found that 24.2% of the farms were not achieving a satisfactory standard of safety management. These findings suggest that circulation of the Risk Assessment Document on its own or when explained as part of a half day course, as outlined, has limitations for a significant number of farmers with regard to completing the document satisfactorily or motivating them to implement controls in the immediate aftermath of the course. This finding regarding implementation of control measures is not unexpected in the light of relevant Irish and International literature. In Ireland, converting farmers’


193

high levels of awareness for safety into high levels of adoption has been identified as the key challenge to improving safety standards (Finnegan, 2007). Regarding implementation of management practices by farmers in Ireland a time lag tends to occur between gaining knowledge of its adoption and the subsequent use of this knowledge to improve practices (Phelan, 1985). At an International level, there are few evaluations of interventions available to determine what types of programmes are most effective in reducing farm injuries (DeRoo and Rautiainen, 2000). The so called Three-E Method of accident prevention involving Engineering (implementation of physical controls), Education and Enforcement (internally within organization or externally) has had success in industrial settings but not in agriculture (Murphy, 1992). The reason for this difference, according to Murphy, was that a greater level of control exists in the industrial workplace. Further initiatives with this research will assess the impact of more comprehensive training in improving satisfactory health and safety levels on farms as well as examining the impact of training over a longer time period. Acknowledgements The input of participating farmers and Teagasc training and advisory staff to the Initiative is acknowledged. References CSO – Irish Central Statistics Office (2001). Statistical Yearbook of Ireland, Stationery Office, Dublin,

Ireland. 373pp. CSO – Irish Central Statistics Office (2003). Quarterly National Household Survey, Stationary Office,

Dublin, Ireland. 52pp. DeRoo, L.A., and Rautiainen, R.H., (2000). A Systematic Review of Farm Safety Interventions.

American Journal of Preventative Medicine. 18 (4S) pp 51-62. Health and Safety Authority (2004). Annual Report for 2003. Health and Safety Authority Publication, 82

pp. Health and Safety Authority (2007). Fatal Statistics by Economic Sector 2000-2007.Available at

www.hsa.ie Finnegan, A., (2007). An Examination of the Status of Health and Safety on Irish Farms. Unpublished

Ph.D., UCD, Dublin. Hope, A., Kelleher, C., Holmes L., & Hennessy, T. (1999). Health and safety practices among farmers

and other workers: a needs assessment. Occupational Medicine. 49 (4), 231-235. McNamara, J., Moran, B., & Cushion, M. (2007). National Survey of Farm Accidents in Ireland. Paper

presented at Irish Agricultural Research Forum. p 125. McNamara J., Connolly L., Cushion, M., & Laffey F. (2006). Progress with completion of safety

statements by farmers in Ireland. Paper presented at Irish Agricultural Research Forum. p 92.


194

McNamara, J., Ruane, D.J., Whelan, S., & Connolly, L. (2007). A Preliminary Investigation of the Incidence and Impact of Disability on Irish Farms. Journal of International Agricultural and Extension Education 14 (2). Paper accepted.

Melberg, K. (2003). Farming, Stress and Psychological Well-Being: The Case of Norwegian Farm

Spouses. Socilogia Ruralis, 43(1) pp 56-76. Murphy, D.J. (1992). Safety and Health for Production Agriculture. St Joseph, MI, the American Society

of Agricultural Engineers. O’ Shea, E. (1997). Male Mortality Differentials by Socio-Economic Group in Ireland. Social Science

and Medicine. 45 (6), 803-809. Phelan, J. (1987) An Analysis of Factors Associated with Decisions Related to Property and Management

Transfer on Irish Farms. Unpublished Ph.D., National University of Ireland. University College Dublin.

Teagasc (1997). Certificate in Farming Host Farmer Safety Inspection Protocol. Teagasc internal

document. Teagasc (2005). Memorandum for implementation of joint Health and Safety Authority – Health and

Safety Pilot Project on Farm Health and Safety at County level. 9pp.


195

METHODOLOGICAL FRAMEWORKS FOR RESEARCH AND DEVELOPMENT ON IMPROVING LINKAGES AND THE COMPETITIVENESS OF SUPPLY CHAINS

Roy Murray-Prior

Department of Agribusiness, Muresk Institute, Curtin University of Technology,

Northam WA 6401, Australia Email: [email protected]

Abstract This paper outlines methodological frameworks for conducting research and development with agribusiness supply chains in transitional economies where the objective is to improve the competitiveness of the supply chains in a global environment. The key difficulty when operating with supply chains is the complexity of the issues involved because constraints can occur from production by small farmers through to the relationship between retailers and consumers. Failure to address key constraints can lead to failure to have any impact. This creates problems for integrating research, because researchers are generally trained in disciplines that cover only a small proportion of the issues and operate from differing epistemologies. The other difficulty is to incorporate a framework for change management, since good research is not much use unless it leads to positive outcomes. A framework for managing these problems is outlined, which has been developed and trialled in work with vegetable supply chains in the Philippines and coffee supply chains in PNG. The framework incorporates a dualistic agribusiness systems model for mapping the chain issues and combines this with a pluralistic framework derived from Checkland’s soft systems methodology for research analysing the system. This is integrated with a participatory action research methodology for change management. Keywords: rural development, dualistic agribusiness systems, pluralistic research, participatory action learning

Introduction Until the last two decades, much research and development work in agriculture focussed on transferring production technology to small farmers in the hope that this would lead to improvements in their productivity and would enable them to compete with farmers from other regions and countries. Such top-down approaches have been widely critiqued because of their perceived failure (e.g. Tully 1963; Chambers 1983). More recently there has been a shift towards more ‘bottom-up’ or participatory models of development, although people have been advocating these models for over half a century. Participatory models have promoted more farmer-centred approaches rather than focussing on particular innovations or commodity specific activities. Such models have relied on building capacity of small farmers and their communities to enable them to compete in globalised world markets. Ladders of participation (e.g. Arnstein 1969; Pretty 1995) are represented hierarchically and imply that more participation is better and that the ultimate method for achieving change is to adopt the top level of participation. Hayward et al. (2004) have challenged the idea that participation is necessarily a solution to complex social problems. For different reasons, Gladwin et al. (2002) argues that participatory research methods are necessary but not sufficient for conducting development work. Another view is that what is required is a partnership between farmers, extension, researchers and industry in order to develop effective solutions to industry problems (Schulz et al. 2004). Such a


196

partnership model is consistent with level 6 of the 8 levels in Arnstein’s typology or level 6 of the 7 in Pretty’s typology. We can extend the partnership view of participation to development work with farmers from transitional farmers in the increasingly globalised market place for food because the constraints to small farmers competing in such markets are complex and varied. Part of the issue is that participation and empowerment is only one part of the solution to development problems, just as technology is another part. Other researchers (e.g. Mingers 2001; Harriss 2002; Kanbur 2002; Madsen & Adriansen 2004) have focussed on the need to combine disciplines when tackling complex problems. Researchers have suggested various approaches and names for combining disciplines or research methods including: multidisciplinary, cross-disciplinary, interdisciplinary, multi-methods, multi-methodology, methodological pluralism and pluralistic methodology. Such approaches are becoming more widely used in development work because of the multidimensional and complex nature of the social, economic and technical problems faced. Each discipline has its strengths and weaknesses and the partnership of these disciplines can lead to richer and more reliable solutions to complex problems. The difficult issue is the framework and processes used to combine the disciplines and their various philosophical paradigms while retaining the ability of the disciplines to maintain their scientific rigour. Since the ultimate aim of most development work with small farmers is to improve the economic well-being of them and their communities, the focus of research and development has a need to focus on those constraints and their causes limiting farmers’ ability to achieve this. The complexity of the causes for these constraints requires them to be addressed at different levels of the causal relationship (Mikkelsen 2005). In order to identify the causes and their linkages, some holistic or systems framework is required to guide the investigation. In this paper, one method for addressing these issues is outlined that combines a dualistic agribusiness systems model with a pluralistic research framework and a participatory learning model. It is developed from work conducted with small farmers in South Africa, in the vegetable industry of the Philippines and the coffee industry in Papua New Guinea. A Dualistic Agribusiness Systems Model Murray-Prior and Ncukana (2000) developed the concept of a dualistic agribusiness systems model to help with analysing the issues facing resource poor farmers in South Africa, particularly from the former homelands, in their struggle to raise their standard of living in a globalised agribusiness system. A key issue faced by small-scale producers from many industries in transitional economies is how to change their production and marketing systems to enable them to shift from supplying their produce to low-priced markets to supplying the needs of the growing higher-priced institutional markets. World Trade Organization and bilateral trade agreements have opened up markets in and to transitional economies. These changes create both opportunities and threats for small farmers because the demands of the new markets require them to significantly increase the quality of their produce. The issues involved in achieving this are complex and failure to address a number of key constraints can lead to failure to have any impact. In the face of this complexity, a dualistic agribusiness systems model has proved useful in conceptualising the issues associated with enhancing the profitability and competitiveness of vegetable supply chains in the Philippines (Murray-Prior et al. 2004; Murray-Prior et al. 2006) and coffee supply chains in Papua New Guinea (Murray-Prior & Batt 2006). It is derived from a simple agribusiness systems model (Murray-Prior et al. 2003) that incorporates the elements of a supply chain, logistics and information flows along the chain, chain management, waste, and elements external to the system such as the socio-economic and political environment and the agro-climatic-ecological environment.


197

PNG has two coffee chains that can be conceived as being two separate (or dualistic) agribusiness coffee systems that are a remnant of colonial occupation (Murray-Prior & Batt 2006). The plantation system produces higher quality coffee for the speciality market, while the smallholder system produces coffee for the soluble market. While PNG Arabica coffee has the potential to be sold into the speciality market the current smallholder chain is highly unlikely to achieve this in its current form. Figure 1 is an example of the dualistic model based using the PNG coffee industry. It illustrates the complexity of the issues facing smallholder coffee producers in their efforts to produce coffee suitable for the speciality coffee market rather than the soluble coffee market. As can be seen the agribusiness system model provides a guide for representing supply chains as well as a checklist for research and development into the problems faced by smallholders in their attempts to produce product suitable for higher-priced markets. Figure 2: Dualistic model of coffee supply chains in PNG including some of the constraints to improving its competiveness A Pluralistic Research Framework Complexity also creates problems for conducting and integrating research into agribusiness systems. Generally, researchers’ training is in disciplines that cover only a small proportion of the issues and they can operate from differing epistemologies. Therefore, while multi-disciplinary teams of researchers are required, in order to be effective some process is required to integrate the various discipline-based research projects. In fact, the process needs to start earlier than this, in that we need to identify the problems to address from a systems or holistic perspective, not from a disciplinary perspective. Murray-Prior et al. (2004) developed and implemented a pluralistic framework based, in part, on Jackson’s (1999) call for a meta-methodology to deal with complex problems. Jackson suggests using a soft-systems paradigm based on the initial processes developed by Checkland (1999) to gain initial understanding of the system and to follow the learning cycle implicit in the soft systems methodology. Murray-Prior et al. (2004) refined this process to include six steps: 1. Analyse the system with stakeholders. 2. Structure the problem statements & determined what methodologies are appropriate to research

each of the problems. 3. Formalise understanding of the problem – may involve hard and soft systems research on problems

that have been identified. 4. Verify understanding with reality – involves comparing and discussing the findings from the

various methodologies and then discussing them with stakeholders. 5. Debate desirable and feasible change.

Higher-priced speciality market

Lower-priced Y grade market

Inputs &services

Smallholder & plantation

cherry

Plantation & exporter wet

factoryDry factory Exporter

Niche roaster

Speciality coffee

consumer

Smallholder parchment

Organic or FT

cooperative

Dry factory ExporterMass

roaster

Instant coffee

consumer

Roadside trader

Cooperative

Smallholder wet process

Lack of finance

Law & order

Poor roads

& telecom

Variable

quality

Poor knowledge & application

Uncertain land tenure

Misdirected

policy

Social resistance to

change Social distance between farmers

& exporters


198

6. Take action to improve the situation. The soft-systems framework has proven useful in providing a clearer picture of system boundaries, the relationships among chain participants, the institutional frameworks within which actors operate and most importantly the key constraints to improvements in the system. When combined with the agribusiness systems model it helps maintain focus on the whole picture rather than taking a reductionist approach to the problem. Another advantage of this approach is that it enables a systems approach to the whole problem while allowing researchers to remain consistent to the theoretical foundations of their discipline. Methods, models and techniques are not separated from their theoretical foundations and consequently improvements can be made within particular theoretical frameworks. However, it does challenge researchers, because sometimes methodologies may be employed side-by-side to investigate particular problems and may give inconsistent or diametrically opposite results (Murray-Prior et al. 2003). This forces researchers to question the validity of the assumptions of their theories and to examine problems from different theoretical perspectives. Researchers are therefore educated about other ways of looking at a problem and gain a greater understanding of the strengths and weaknesses of their own and other disciplines. A Participatory Action Learning Process with Chains and Industry While the pluralistic research framework outlined above implies consultation with actors and stakeholders involved with chains in planning research, its focus is on the research activity and does not explicitly address the issue of facilitating the change process at the farm, chain or industry level. The concept of a partnership model (Schultz et al. 2004) was extended from relationships and participation at the farm systems level to relationships and participation at the chain systems level and with industry and government institutions. Therefore, the agribusiness systems model helps guide the selection of the actors and stakeholders to involve in this process, but it is implemented through a participatory action learning process. At the farm level, this occurs with farmer groups, beginning with a Participatory Rural Appraisal process. At the chain level, a similar process occurs with selected actors from the chain, including representatives from the farmer group. This addition to the approach came about as a result of perceived failures or weaknesses in our project with vegetable farmers in Mindanao, Philippines and from a need to integrate with a Participatory Rural Appraisal and Planning Process being implemented by the Coffee Industry Corporation in PNG. In the former case we recognised that our strategy for change in Mindanao was ad hoc and while it did involve consultations with farmers, traders and retailers the process for change was not formalised or guided by a coherent process. Part of the answer was provided by recognition that in the case of the PNG coffee industry, more research on its own was not the answer, and that we needed to involve farmers in a learning process so they could learn more about the constraints to improving their profitability. Another weakness in our method we had identified was that we had not done enough to encourage linkages and understanding along the chain. Consequently, we believed that a participatory learning process would be the best solution to this issue. We use the participatory action learning process to help structure and prioritise research problems and to identify and prioritise learning needs (see Figure 2). The prioritised research needs are key inputs to Steps 1 & 2 of the pluralistic research framework and the chain actors are partners in this process. The formalisation of this process also provides a feedback mechanism for Steps 4, 5 and 6 to verify understanding of the outcomes of the research, debate desirable changes and to take action to improve the situation.


199

Figure 3: Participatory action research, development & learning cycle with farmers, chain & industry At the same time, the process helps identify and prioritise learning needs for farmer and chain actors, which are then addressed through organised learning activities. Outcomes from research are also fed into the learning activity cycle and outcomes or observations from the learning activities can be fed into the research and development cycle. This is an ongoing process, where reflection on leaning activities and experiences from implementation of change are discussed and provide input to revise research, development and learning activities. Conceptually and in practice, this is a multi-level action learning process; one level with farmer groups, one level with chain actors, and another level with industry and political institutions. The focus of the research project also necessarily influences and constrains the focus of the learning and research activities. In the case of the PNG coffee industry project, the focus is on improving the price received by farmers through increasing the proportion of coffee that achieves the standard necessary for sale in the speciality market. Consequently, research and development effort concentrates on marketing and chain relationship issues, although some of the learning activities relate to production and processing issues. However, information from the participatory processes inputs into other research projects dealing with pest and disease and post-harvest problems. In the case of the Philippines vegetable chain projects, the projects’ foci are more holistic and research and development activities conducted by the projects were and will have a more paddock to plate scope. Consequently, while the project can address some of the issues, we endeavour to influence and involve other actors with influence or resources that could benefit the agribusiness system to address the issues that are beyond the scope and resources of the project to address. Conclusion Conducting research and development work in transitional economies to deal with the issues faced by small farmers and local businesses who are struggling to compete in globalised world markets is difficult and complex. Many issues constrain their ability to compete and focussing on just one of these issues is generally unsuccessful because farmers and businesses may not be in a position to implement any recommendations dealing with this issue because of the other constraints. In this paper, a methodological framework for conducting research and development with agribusiness supply chains is outlined which

Farmer &chainPRAs

Learningneeds

Planlearning

Observe& reflect

Plan newlearning

Conductlearning

Conductlearning

Observe& reflect

Research/development

needsDesign

research

Conductresearch

Discuss& reflect

Developoutcomes

RedesignresearchDiscuss

& reflect

Conductresearch

Redevelopoutcomes

Industrydiscussions

Learning activities

R&D activities


200

suggests a series of processes and models for dealing with this complexity that may increase the chances of achieving a positive impact. It consists of three components: • A dualistic agribusiness systems model that helps guide investigation of the system so that

important elements are less likely to be omitted. • A pluralistic research framework to help identify which issues need to be researched, what

methodologies are appropriate for that research and to integrate research conducted by a multi-disciplinary team of researchers.

• A participatory action research, development and learning process to involve actors and stakeholders, enhance their ownership of project activities, and increase their capacity to change and overcome the constraints to their involvement in higher value markets.

Experience from a range of projects in transitional economies has led to the development of this framework, which is still in the process of development and evaluation. As is obvious from the reference to projects over time, each element in the framework was developed as part of an ongoing learning process, in an effort to overcome weaknesses identified with our research activities at various stages in these projects. Acknowledgements I acknowledge financial support from the Australian Centre for International Agricultural Research for projects in the Philippines and PNG, which led to the development of many of the ideas presented here. I would also like to acknowledge members of the research teams in these projects who were involved in the development of some of the ideas or commented on the models I presented. In particular, I would like to acknowledge: Murray McGregor, Peter Batt, Sylvia Concepcion, Fay Rola-Rubzen, Larry Digal and Nerlie Manalili. References Arnstein, S. 1969, ‘A ladder of citizen participation’, American Institute of Planners Journal, July, 216-

224. Chambers, R. 1983, Rural development: Putting the last first. Longman, Harlow, England. Checkland, P. 1999. Systems thinking, systems practice: Includes a 30-year retrospective, Wiley,

Chichester, England. Gladwin, C.H., Peterson, J.S. & Mwale, A.C. 2002, 'The quality of science in participatory research: A

case study from Eastern Zambia', World Development, 30(4), 523-543. Harriss, J. 2002, 'The Case for Cross-Disciplinary Approaches in International Development', World

Development, vol. 30, no. 3, 2002/3, pp. 487-496. http://www.sciencedirect.com/science/article/B6VC6-44TCXS3-1/2/96b795a1a6a99fdfa4a4b310e38f588c.

Hayward, C., Simpson, L. & Wood, L. 2004, 'Still left out in the cold: problematising participatory

research and development', Sociologia Ruralis, 44(1), 95-108. Jackson, M.C. 1999. ‘Towards coherent pluralism in management science’, Journal of the Operational

Research Society, 50(1), 12-22.


201

Kanbur, R. 2002, 'Economics, Social Science and Development', World Development, 30(3), 2002/3, 477-486.

Madsen, L.M. & Adriansen, H.K. 2004, 'Understanding the use of rural space: the need for multi-

methods', Journal of Rural Studies, 20, 485-497. Mikkelsen, B. 2005, Methods for development work and research: a new guide for practitioners, 2nd edn,

SAGE Publications, New Delhi. Mingers, J. 2001, 'Combinins IS research methods: Towards a pluralistic methodology', Information

Systems Research, 12(3), September, 240-259. Murray-Prior, R. & Batt, P. 2006, 'Emerging possibilities and constraints to PNG small-holder coffee

producers entering the speciality coffee market', Paper presented to: International Symposium on Fresh Produce Supply Chain Management, Lotus Pang Suan Kaeo Hotel, Chiang Mai, Thailand, 6-10 December.

Murray-Prior, R. and Ncukana, L. 2000, Agricultural development in South Africa - A dualistic

agribusiness systems perspective, Paper presented to the ‘African Studies Association of Australasia and the Pacific 23rd Annual and International Conference: African Identities, St Marks College, University of Adelaide, North Adelaide, South Australia, 13-15 July 2000.

Murray-Prior, R., Batt, P.J., Rola-Rubzen, M.F., McGregor, M.J., Concepcion, S.B., Rasco, E.T., Digal,

L.N., Montiflor, M.O., Hualda, L.T., Migalbin, L.R., Manalili, N.M. 2006, ‘Global value chains: a place for Mindanao producers? in P.J. Batt (ed.), Proceedings of the 1st International Symposium on Improving the Performance of Supply Chains in the Transitional Economies, Acta Horticulturae 699, January 2006, ISHS, Belgium, 307-315.

Murray-Prior, R.B., Concepcion S., Batt, P., Rola-Rubzen, M.F., McGregor, M., Rasco, E., Digal, L.,

Manalili, N., Montiflor, M., Hualda, L. & Migalbin, L. 2004, ‘Analyzing supply chains with pluralistic and agribusiness systems frameworks’, Asian Journal of Agriculture and Development, 1(2), 45-56.

Murray-Prior, R.B., Rola-Rubzen, M.F., McGregor, M., Batt, P., Concepcion S., Rasco, E., Digal, L.,

Manalili, N., Moran, A., Ellson, A., Montiflor, M., Hualda, L. & Migalbin, L. 2003, ‘A pluralistic methodology for analysing supply chains’ in Proceedings of the Australian Agricultural and Resource Economics Society Conference, Fremantle, WA, 11-14 February 2003.

Pretty, J.N. 1995, Regenerating agriculture: policies and practice for sustainability and self-reliance,

Earthscan Publications, London. Schulz, L., Murray-Prior R.B., Storer, C.E. and Walmsley, T., 2004, ‘Overcoming difficulties with

outsourcing in partnership extension models: lessons learned from TOPCROP West’, Australian Journal of Experimental Agriculture, 44 (3), 223-31.

Tully, J. 1964, 'Operational research in agricultural extension in Queensland', Agricultural Progress,

XXXIX, 7-11.


202

HOUSEHOLD FOOD SECURITY IN VIHIGA DISTRICT, KENYA: KEY DETERMINANTS USING AN ALMOST IDEAL DEMAND SYSTEM (AIDS)

Nyangweso P.M,

Economics and Agricultural Resource Management Moi University P.O. Box 1125

ELDORET, KENYA E-mail: [email protected]

Odhiambo M.O

Economics and Agricultural Resource Management Masinde Muliro University of Science and Technology

P.O. Box 190 KAKAMEGA, KENYA

E-mail: [email protected]

Odunga P. Economics and Agricultural Resource Management

Moi University P.O. Box 1125


Abstract Vihiga, one of the poorest and densely populated districts in Kenya is perpetually food deficit (GOK, 2005). Rising population pressure coupled with intense competition for limited resource endowments has curtailed efforts to improve household food production in the district. To make matters worse, unfavorable poverty indicators hinder attainment of food security, in the district, through the demand side. About 57.6 percent of the population and more than 50 percent of households live below absolute poverty line while 57 percent of the population and households live below food poverty line. Poor welfare indicators for Vihiga district underscore the importance and urgency for addressing the basic needs of its residents. Understanding determinants of food security in Vihiga district will improve targeting, the focus and success of policies for addressing food insecurity. This paper examines determinants of food security in Vihiga district using an Almost Ideal Demand System (AIDS) to determine the demand side constraints using household survey data. Cluster sampling was used with divisions forming the main clusters in the district. Using systematic random sampling, 50 households were selected from each cluster resulting in a sample of 300. Results show that household income, dependency ratio, gender of household head, household savings/transfers characteristics, ethnicity, education, market access and nutrition awareness significantly influence household food security. Food programmes in Vihiga should pay special attention to household structure, preferences and decision dynamics for successful implementation. Key Words: Food security, Almost Ideal Demand System (AIDS), Vihiga, Kenya. Introduction

Despite having the potential to meet domestic food demand, Kenya has continued to grapple with persistent food deficits over the last two decades. Over the last six years the annual demand for maize in the country rose from 29.5 million bags to 32.9 million bags (GOK, 2004). However, production in the same period ranged between 25 and 30 million bags per year thus necessitating importation of food to meet the deficit.


203

Vihiga, one of the poorest and densely populated districts in Kenya with an average household land size of less than 0.4 hectares is perpetually food deficit (GOK, 2004). This has been attributed to limited land, high poverty levels, limited off-farm income, and non-adoption of recommended farm technologies. Vihiga district is a perfect case of why the Kenyan government will be unable to meet millennium development goals especially as regards eradication of extreme poverty and hunger (UN, 2005). Maize is the main staple food for residents of Vihiga district thus its insufficiency is synonymous with food insecurity. Over the last decade, the district maize demand outpaced local production worsening the already bad food deficit situation.

Food security describes a situation in which people do not live in hunger or fear of starvation. According to FAO (2003), food security exists when all people, at all times, have access to sufficient, safe and nutritious food to meet their dietary needs and food preferences for an active and healthy life. This study defines household food security as access to nutritionally adequate and safe foods by all households at all times to meet their dietary needs and food preferences for an active and healthy life.

As poverty levels rise, household food insecurity in the district worsens. Families with the financial resources to escape extreme poverty rarely suffer from chronic hunger; while poor families not only suffer the most from chronic hunger, but are also the segment of the population most at risk during food shortages and famines (FAO, 2003). Vihiga district has unfavorable poverty indicators as measured by food poverty, absolute poverty and hard-core poverty. About 57.6 percent of the population in Vihiga district lives below the absolute poverty line, which is set at Kshs. 2648, and Kshs. 1238 per month for urban and rural areas respectively (GOK, 2004). Similarly, more than half of the households in Vihiga, which is one of the worst hit districts in Kenya, fell below the absolute poverty line. To make matters worse, about 57 percent of both individuals and households in the district live below the food poverty line. While 45 percent of the households live in hard-core poverty, more than half of the individuals in these households live in hard-core poverty. Poverty has a twin impact on household food security. It not only reduces the capacity of households to access farm inputs due to capital limitations thus hindering expanded food production, but also prevents households from accessing food due to their low or non-existent purchasing power. Consequently, malnutrition among households has become a big issue since if basic food needs can not be met very few household would care about the quality of food they eat. Poor welfare indicators for Vihiga district underscore the importance and urgency for addressing the basic needs of its residents. Understanding determinants of dietary diversity presents an opportunity for improving targeting, the focus and success of policies for addressing food insecurity. The paper examines the major demand side constraints to food security among households in Vihiga district of Kenya. The paper is subdivided into five sections. In section one, an introductory exposition of the problem is presented. Section two reviews theoretical considerations and presents the model used for estimation. In sections three and four, methods and materials followed by results and discussions are presented. Finally, conclusions and recommendations are presented. Theoretical Considerations Modeling Consumption Behavior

The objective of analyzing consumer behavior is to explain the level of demand for the commodities an individual consumes given the structure of relative prices faced, real income, a set of individual characteristics such as age, education, professional status, type of household to which he belongs and the geographical environment (De Janvry, 1993). Knowledge of the demand structure assists in (1) definition of policy interventions for improving nutritional status of individuals or households ; (2) formulation of a country’s strategy on food subsidies; and (3) sectoral and macroeconomic policy analysis. The theory of consumer behavior basically explains how a rational consumer chooses what to consume when confronted with various prices and limited income (Varian 1992, De Janvry, 1993, Mas-colell et al,


204

1995). Considering a consumer whose utility function is u (x, z), with x a vector of quantities and z individual characteristics, the consumer maximizes utility with respect to quantity, x, subject to a budget constraint px=m. This can be re-written as:

Max u (x, z) + λ (m-px) (1) x, λ

Where p, m and λ are vectors of prices, income and the Lagrange multipliers respectively. The first order condition which shows that the gradient of the Lagrangian function must be equal to zero is used to derive the optimal solution for utility maximization problem, which occurs when the marginal rate of substitution (MRS) between goods i and j is equal to the rate of exchange of the two goods (Varian, 1993, Mas-colell, 1995, Jehle and Reny, 1998). This can be expressed as (2) below:-

L(x*,λ*) = 0 (2) The solution to this maximization problem is a set of n-demand equations:

xi = xi(p, m, z), i = 1, ….n (3) The second order condition for utility maximization is satisfied when the bordered Hessian is positive semi-definite. Since:

2L(x*,λ*) = -λ∂2u(x,z) = -λ ∂2u(x,z) (4) ∂xixj ∂xixj

Expenditure function, which is the dual of the utility function, when minimized yields the same result as maximization of the utility function. Considering a consumer whose expenditure function is e (u, p, z), with u targeted utility, p a vector of commodity prices and z a vector of individual characteristics, the consumer’s objective function is to minimize expenditure with respect to quantity, x, subject to a targeted utility constraint u (x) = u. This can be specified as: -

Min px + γ (u - u (x)). (5) x, γ

Where γ is the Lagrange multiplier. To confirm that expenditure minimization is a dual for utility maximization first order conditions for expenditure function are derived to prove that at optimality the marginal rate of substitution (MRS) between goods i and j is equal to the price ratio of the two goods. Sufficiency conditions for expenditure minimization when f (.)=px and g(.)=u - u (x), are twice differentiable and vectors x* є Rn , λ* є Rm exist require that such that

L (x*, γ*) = 0 (6)

g(x*) = 0 occur for p =2,3,…,n, if the bordered Hessian of the second derivative of the Lagrange function is negative semi definite. The expenditure approach is adopted in this paper because it is practically feasible to deal with consumer expenditure behavior when doing empirical evaluation. Model selection A variety of models have been used to describe the allocation of consumers’ expenditure that is compatible with consumers behaving according to well-defined preferences. Such models include linear expenditure system (Stone, 1954), direct and indirect translog system (Lau, 1984), quadratic expenditure System (Matsuda, 2006), Price independent generalized linearity (PIGL) demand System (Muellbauer, 1980, De Janvry, 1993), Price independent generalized logarithm (PIGLOG) demand systems (Muellbauer, 1980, 1986) Almost Ideal Demand System (AIDS)(Deaton and Muellbauer, 1980, Chalfant, 1987 Rossi, 1988, Nyang. 1999, Andersson et al, 2006). Many of the demand models are not well suited to survey data. Linear expenditure system is overly restrictive. Direct and indirect translog systems are expensive to estimate using extensive survey data. The AIDS model suffers from neither of these drawbacks (Cheser and Rees, 1987). Further, the AIDS model can be easily estimated by inexpensive non-iterative methods and be used to examine expenditure allocation within a broad food group and also between broad food groups. The AIDS model expresses the share of total expenditure allocated to good i, wi = yi/x, as a linear function of the logarithm of total expenditure, x, and of prices, pj, j= 1…m, thus:

wi =α

*i + βi log(x/P) + ∑γijlnPj + ui (7)


205

j=1

Where: m m m Log P = α*

0 + ∑ α*k log pk + ½∑ ∑γkj logpk logpi (8)

k=1 k=1 i=1

P is a price index. Homogeneity and symmetry restriction of demand theory require that: α*

i, γ, and βj which are easily imposed and tested to meet the following conditions:

∑ α*i = 1 ∑γij = ∑γji =0 ∑βj = 0 γij = γji

i i j i

With prices constant, as they are approximately for many foods within one survey period, the model yields an income -expenditure relationship of the form:

wi =αi + βi log(x) + ui , where αi = α*i +∑γkj logpj - βj logP (9)

This is the form of the Engel curve used by Working (1943), later developed by Leser (1963, 1976), Deaton and Muelbauer (1980) and found to perform well when faced with cross-section data. Following Deaton and Muellbauer (1986) income is expressed per capita using a simple headcount of household members and the intercept in the model is augmented to allow for influence of household composition. The estimated model is specified as (10):-

W=α + β log (X/n) + γZ + θV + δY + ε (10)

Where; n = number of household members X = household monthly total expenditure on food. W = a vector of ratio of survey month expenditure on each food item to household monthly

total food expenditure. Z, V, and Y are vectors of household characteristics, environmental factors and ethnic, savings/transfer characteristics while α, β, γ, θ and δ are corresponding vectors of parameters to be estimated, and ε is a normally distributed random error term. The specific variables contained in the vectors Z, V, and Y are shown in the appendix. Foods for which β < 0 are necessities and as total expenditure increases become inferior once β + w < 0.

Materials and Methods Sampling Design

The study targeted all farm households in Vihiga district. Cluster sampling was adopted on the basis of the six divisions. Using systematic random sampling procedure, 50 households were selected from each cluster generating a total sample of 300 respondents. Data Types and Sources

Both primary and secondary data were used. The data encompassed expenditure on various commodity groups, commodity prices, household characteristics (education, age, family size, gender of head of household, employment, business income, ethnic origin, savings/ transfer behavior, highest education level, market access, geographical location, monthly household consumption. Primary data was collected through a survey while secondary data was acquired by perusal of annual agricultural reports, economic surveys, statistical abstracts and development plans.


206

Data Collection Methods

Both interviews and questionnaires were used as instruments for data collection. Interviews were used to supplement questionnaires. To validate survey instruments, 10 questionnaires were pre-tested in one of the divisions, revised and forwarded to enumerators. Trained enumerators were used to administer the questionnaires. Focused group discussion was used to elicit information from key informants who included district agricultural officer, district development officer, heads of district non-governmental organizations, divisional agricultural extension officers, field extension workers and local administration. Observation was used to countercheck some of the findings.

Data Analysis

Descriptive statistics such as bar charts, cross tabulations and measures of central tendency were used to describe emerging relationships between variables. Multiple regression analysis was used to estimate a system of budget share equations from the survey data using Statistical Package for Social Sciences (SPSS) version 11.5. Multi-collinearity was tested using Pearson’s correlation coefficient. Results and Discussion Results (table 1) show that a bundle of food necessities for residents of Vihiga district consist of maize grain, sugar, cabbage, kale, oranges and vegetable oil. However, none of the foods become inferior as levels of income for the residents improve. That shows that levels of income recorded in the district even when they improve only manage to improve accessibility to basic foods, but is not good enough for locals to start perceiving some of the foods as inferior. On the other hand a bundle of normal goods whose consumption is significantly influenced by the level of income includes maize meal, wheat meal, bread, rice, sorghum, millet, Irish potatoes, peas, green grams, black night shade(sucha), spider plant(saka), fish and beef. This clearly shows how households will exclude some food items from their budgets if they are non-affordable even when such actions result in escalation of malnutrition among households. It is therefore critical to fight poverty as a means of ensuring households access food not only in the right amount but also in the appropriate quality. The dependency ratio, number of adults, gender of household head, education, employment and ethnicity significantly influence budget shares for some commodities with mixed results. As the dependency ratio and number of adults increase more budget share is allocated to basic foods which tend to be cheaper. On the contrary more expensive foods such as beef get less allocation. Male household heads significantly influence choice of more expensive foods such as beef while female counterparts tend to go for lower- priced commodities so long as diversity of the food stuffs is achieved. Education and employment which are associated with status positively influence consumption of goods that go with status and negatively influence foods that are considered inferior. Different ethnic groups such as Banyore, Maragoli, Tiriki and others exhibit different preferences for different food commodities. Consequently, while one ethnic group might increase budgetary allocation for one commodity the other group may be doing the reverse. The results show consistency since they satisfy homogeneity and symmetry requirements of demand theory. Results further show that savings/transfers, market access and nutrition awareness influence budget shares for an assortment of commodities. While savers are likely to go for status commodities, non-savers are likely to go for more traditional basic food stuff. Transfers received by some households from relatives and friends boost their ability to access food and in some cases can result in change of consumption behavior. Nutrition awareness of dietary requirements has both positive and negative impact on consumption. For households with enough resources awareness results in higher consumption. However, for resource poor household’s, awareness results in concentration on basic food stuff thus reducing budget shares of some commodities. Market access is crucial for commodities that have to be purchased from the market.


207

Table 1: Estimated budget share parameters

Income

Depend.

ratio

No. of

adults H.Head

Business

income Educat. Employ. Ethnicity Savings Transfers

Urban-

rural

Market

access

Nutr.

awareness

Commodity i αi βi γi1 γi2 γi3 γi4 γi5 γi6 Φi1 Φi2 Φi3 δi1 δi2 δi3 R R2

Maize grain 0.693a -0.078 a 0.003 -0.002 0.011 -0.019 b -0.004 0.001 -0.011 b 0.001 -0.004 0.012 0.017 b -0.015 0.57 0.32 Maize meal -0.038 0.012b 0.002 0.003 -0.006 (-0.005) 0.001 -0.009 -0.006 -0.009 -0.009 -0.003 -0.018 a 0.005 0.30 0.09 Wheat meal -0.056 a 0.008 a -0.001 0.001 0.003 0.004 0.002 0.008 b -0.001 0.001 0.002 0.002 0.005 0.005 0.39 0.15 Bread -0.098 a 0.018 a -0.001 (0.000) 0.002 0.011 b 0.001 0.014 b -0.008 a 0.001 -0.004 0.012 a 0.003 -0.006 0.42 0.17 Rice -0.038 0.006 b -.002 0.001 -.002 .004 0.005 a -.002 0.003 0.01 a 0.005 0.001 0.005 -0.002 0.44 0.19 Sorghum -0.017 0.004 b 0.000 0.000 -0.002 -0.001 0.002 b -0.003 0.002 b -0.001 0.000 -0.003 0.002 -0.006 b 0.31 0.10 Millet -0.01 0.004 b 0.001 0 -0.007 a -0.001 0 0.001 0.001 0.001 0.001 -0.002 -0.002 -0.004 0.25 0.10 I/potatoes -0.023 0.005 b 0 0.001 -0.001 0.004 b 0.002 b -0.005 b 0 0.003 -0.003 0.002 -0.002 -0.002 0.31 0.10 S/potatoes 0.039 (-0.002) 0.005 -0.002 0.005 -0.009 -0.006 -0.017 b 0.010 0.002 0.001 0.003 0.007 0.014 0.26 0.10 Sugar 0.369 a -0.026 a -0.005 -0.001 0.003 -0.005 0.003 -0.004 -0.01 b -0.012 0.001 -0.005 -0.01 -0.016 0.33 0.11 Dry beans -0.016 0.006 0.004 b 0.001 -0.013 a -0.002 -0.005 b 0.002 0.008 a 0.001 -0.008 -0.009 -0.004 0.021 a 0.38 0.15 Peas -0.007 b 0.001 b 0.001 a 0.00004 0.000 -0.001 0.00002 -0.001 0.000 0.000 0.001 a 0.0001 .000 0.000 0.33 0.11 Green grams -0.069 a 0.009 a 0.003 b 0.001 b -0.007 a 0.005 b 0.002 0.003 0.005 a 0.001 -0.003 0.002 -0.002 -0.002 0.45 0.20 Cabbage 0.085 a -0.008 b 0 -0.001 0.003 0 0.001 -0.001 0.003 -0.006 -0.001 -0.005 0.002 -0.012 b 0.26 0.10 Kale 0.085 a -0.008 b 0 -0.001 0.003 0 0.001 -0.001 0.003 -0.006 -0.001 -0.005 0.002 -0.012 b 0.35 0.12 Fresh cowpeas -0.005 0.006 0.004 0.001 -0.003 -0.003 -0.001 -0.004 0 -0.003 0.001 -0.008 0.005 b 0.004 0.25 0.10 Sucha -0.032 b 0.006 b 0.001 0.001 -0.002 -0.002 0.001 -0.001 0 -0.005 b 0.002 -0.001 -0.000 0.000 0.26 0.10 Saka -0.046 a 0.008 a 0.003 a 0.001 b -0.003 -0.002 0.0001 0.006 b -0.001 -0.009 a 0.003 -0.001 -0.002 0.003 0.35 0.12 Miro 0.01 0 0.002 0 -0.006 a -0.001 -0.001 0 0.001 -0.001 0.003 0.002 0.001 0 0.29 0.10 Bananas 0.092 0.002 -0.007 -0.002 0.01 0.001 -0.005 -0.018 b -0.012 a -0.004 -0.005 0.01 0.005 0.002 0.29 0.10 Oranges 0.059 a -0.007 a -0.002 -0.001 b -0.001 -0.001 0.002 0 0 0.007 b 0.001 0.001 0.001 0.006 0.30 0.10 Avocadoes 0.037 -0.002 -0.002 0 -0.001 0.002 -0.004 a 0.001 0.004 -0.003 -0.002 0.006 b -0.003 0.007 0.26 0.10 Mangoes -0.003 0.005 -0.001 -0.001 -0.001 -0.002 0 0.002 0 0.002 0.002 -0.001 -0.003 -0.01 0.24 0.10 Fish -0.055 0.014 b 0.001 0.002 0.003 -0.003 -0.001 -0.007 -0.005 0.005 0.002 -0.005 0.008 0.002 0.22 0.05 Beef -0.201 a 0.029 a -0.007 b 0.002 0.02 b 0.012 b 0.003 0.004 0.013 a 0.013 -0.015 a 0.003 0.001 0.017 0.50 0.25 Chicken 0 0.004 -0.001 -0.003 b -0.007 0.003 0.001 0.012 b 0.009 a -0.007 0.009 a -0.016 a 0.016 a -0.006 0.37 0.14 Milk 0.123 a -0.007 -0.001 -0.001 0.002 0.009 0.003 0 -0.002 0.007 0.005 0.097 0.003 -0.014 0.24 0.06 Vegetable oil

0.072 a -0.006 b 0 -0.001 0 -0.003 0 0.006 -0.002 0.002 0.003 -0.002 -0.002 0.012 a 0.29 0.08

Total 1.005 0.005 0.005 0.001 -0.009 0.003 -0.0009 -0.003 -0.009 -0.011 -0.017 0.087 0.018 -0.015

Source: Author’s compilation from cross-sectional survey, 2007. a-significant at 1 percent, b-significant at 5 percent.


208

Conclusion and Recommendations Vihiga, one of the poorest and densely populated districts in Kenya is perpetually food deficit. Poor welfare indicators for district underscore the importance and urgency for addressing the basic needs and the need to document determinants of food security in Vihiga district to improve targeting, focus and success of food programmes. An attempt is made to evaluate demand side constraints to food security in Vihiga district using an Almost Ideal Demand System (AIDS) through a household survey. Household income, dependency ratio, gender of household head, household savings/transfers characteristics, ethnicity, education, market access and nutrition awareness are critical when addressing household food security. It is therefore recommended that food programmes in Vihiga should pay special attention to household structure, preferences and decision dynamics for successful implementation. This is critical since structure and preferences of household reflect the level and diversity of consumption, while decision dynamics reflect how consumption decisions are made whether through negotiated or dictatorial consensus. References Andersson M., A. Envall and A. Kokko (2006): “Determinants of poverty in Lao PDR”. Working paper

223, march 2006. Stockholm School of Asian Studies. Stockholm School of Economics. Chalfant J.A (1987): “A Globally Flexible, Almost Ideal Demand System”.

Journal of Business & Economic Statistics, Vol. 5, No. 2 (Apr., 1987), pp. 233-242. Chesher A. and H. Rees (1987): “Income elasticities of Demands for Foods in Great Britain. Poverty,

Food Security Status and Farmland allocation to various crops”. Journal of Agricultural Economics, Vol. 38 No.3.

De Janvry S. and Alain De Janvry, (1993): Demand Analysis. Quantitative Development Policy Analysis. GOK (2005): Economic Survey. Ministry of Planning and National Development. _____(2004): Economic Survey. Ministry of Planning and National Development. _____(2004): Strategy for Revitalizing Agriculture. Ministry of Agriculture. _____ (2004): Vihiga District Development Plan. Ministry of Planning and National Development. _____ (2005): Vihiga District Development Plan. Ministry of Planning and National Development. _____ (2004): Statistical Abstract. Ministry of Planning and National Development. Jehle G.A. and P J. Reny(1998): Advanced Microeconomic Theory. Addison Wesley Longman, Inc. Mas-colell A, M. D. Whinston and J.R. Green(1995): Microeconomic Theory. Oxford University Press. Matsuda T. (2006): “Linear approximations to the quadratic almost ideal demand system”. Journal of

Empirical Economics Volume 31, Number 3 / September 2006. Muellbauer J. (1974): “Household composition, Engel curves and welfare comparisons between

households. A duality approach”. European Economic Review Volume 5, Issue 2 , August 1974, Pages 103-122.

Nyang F.O (1999): “Responsiveness in household demand for fuels: an application of the almost ideal

demand system”. Household Energy Demand and Environmental Management in Kenya.


209

Rossi N. (1988): “Budget share demographic translation and the aggregate Almost Ideal Demand System”. European Economic Review Volume 32, Issue 6, July 1988, Pages 1301-1318.

Varian H.R (1992): Microeconomic Analysis 3rd Ed. W.W. Norton & Company, New York. Walpole R. E and R.H Myers (1978): Probability and statistics for Engineers and Scientists. 2nd Ed. UN (2005): Website development: UN Web Services Section, Department of Public Information, United

Nations.


210

Appendix I Table 2: Variables and definitions

Variable Definition and comment

Dependent variable

Consumption per capita

Expenditure share which is proxy for real per capita consumption Independent variables

Household characteristics

Dependency ratio

Adults

Gender of head of household

Household income

Household business

Highest education index

Employment status

Ratio of dependents, below 18 years and above 59, versus adults 18-59

Number of adults in household

1 if male head of household; 0 if female head of household

Total household food expenditure as proxy for income

1 if household run business, 0 otherwise

0 if highest educational attainment in household is pre-primary, 1 if primary, 2 if secondary, 3 if vocational training, 4 if university/college

1 if head of household is employed, 0 otherwise

Ethnic/savings/transfer characteristics

Ethnic origin

Savings

Transfers

1 if Tiriki, 2 if Mnyore, 3 if Maragoli and 4 if non-Luhya

1 if make any savings from salary or business, 0 otherwise

1 if get any transfers from relatives or friends, 0 otherwise

Environmental factors

Rural /urban cluster

Access to muddy season road

Awareness of nutritional needs

1 if household is in urban center, 0 otherwise

1 if village is accessible by truck during rain season, 0 otherwise.

1 if aware of balanced meal, 0 otherwise.


211

DRIVERS OF AGRICULTURAL EXPORTS IN EASTERN AFRICA: EVIDENCE FROM KENYA, UGANDA, AND TANZANIA.

Nyangweso P.M,

Economics and Agricultural Resource Management Moi University P.O. Box 1125


Odhiambo M.O

Economics and Agricultural Resource Management Masinde Muliro University of Science and Technology

P.O. Box 190 KAKAMEGA, KENYA

E-mail: [email protected]

Odunga P. Economics and Agricultural Resource Management

Moi University P.O. Box 1125


Abstract Agriculture contributes substantially to the overall economic growth of East African countries. This sector alone accounts for 25%, 31.1%, and 43.2% of the GDP for Kenya, Uganda and Tanzania, respectively. More than 70 percent of the population in Eastern Africa live in rural areas and rely heavily on agriculture for their survival. Agricultural exports have continued to earn Eastern Africa the much-needed foreign exchange for financing imports for import dependent domestic industries. In 2005, export earnings in Kenya, Uganda and Tanzania, respectively were, US$3.173 billion, $768 million, and $1.581 billion. Out of the total export earnings, agricultural exports contributed more than 70 percent. This paper examines the key determinants of agricultural exports in Eastern Africa. It also evaluates the impact of regional integration and differences arising across countries in the region. It uses Nerlovian Partial adjustment model to fit data for 1974-2004. Results indicate that key factors influencing agricultural exports in the region are exchange rates, regional integration and technological progress. However, there are conflicting results regarding the role of International prices and country specific policies. While some similarities are noted, some differences are recorded across countries. We recommend policies that are outward looking with governments in the region creating conducive business environment to facilitate more foreign investment in export oriented agriculture. Key Words: Drivers, Agricultural Exports, Eastern Africa.


212

Introduction Background on Agricultural Sector

Agriculture contributes substantially to the overall economic growth of East African countries. It accounts for 25-43 percent of the Gross domestic products of Kenya, Uganda and Tanzania (GOK, 2005, FAO, 2005). More than 70 percent of the population in Eastern Africa live in rural areas and rely on agriculture for their survival. Kenya’s Agricultural Sector

The Kenyan agricultural sector has been branded the backbone of the economy. It accounts for 25 percent of GDP, employs over 80 percent of the Kenyan population either directly or indirectly and earns more than 28 percent of visible exports (FAO, 2005). In the last four decades there was a decline in agriculture of 3.5 percent and a corresponding decline in the overall economy by 4.6 percent (GOK, 2002). However, the recovery of the sector was followed by a recovery in the Kenyan economy in the last 4 years. The major food commodities produced in Kenya include: - maize, wheat, cassava, Sorghum, and finger millet. While tea continues to contribute the largest share of total visible exports, other agricultural exports include: - horticulture (flowers, vegetables and fruits), coffee and processed foods and vegetables. In 2005, export earnings in Kenya amounted to US$3.173 billion, (GOK, 2005). The major contributor to these earnings is agricultural exports consisting mainly of primary products. The leading export partners of Kenya are Uganda, UK, US, Netherlands, Egypt, Tanzania and Pakistan who account for 13.9%, 10.5%, 9.5%, 8.2%, 5.1%, 4.7%, and 4.5% of Kenya’s total exports respectively (FAO, 2005). Uganda is the leading single importing country of Kenyan products both globally and regionally. The European market and the US consume 28 percent of the total Kenyan exports. Uganda’s Agricultural Sector Uganda’s economy is predominantly agrarian with 31.1 percent of the GDP, 81 percent of the employed labour force, and 31 percent of export earnings being derived from the agricultural sector (FAO, 2006). Only one third of the total land area is under cultivation with subsistence production representing 70 percent of the area under cultivation. Women provide over half of agricultural labour, traditionally focusing on food rather than cash crop production. The major food commodities produced in Uganda include: - plantains, cassava, sweet potatoes, and bananas. The monetary value of marketed agricultural commodities continues to fall way below estimated value of subsistence agriculture. While coffee continues to remain the primary export earner for Uganda, other export crops include: - cotton, raw sugar, tobacco, roses and carnations. In 2001 coffee earned Uganda an estimated US $51.3 million accounting for 11 percent of the total exports (FAO, 2004). Even though the Ugandan economy is depended on agriculture only a small proportion of agricultural production is export oriented and is mainly primary products. In 2005, export earnings in Uganda were estimated at US$768 million (FAO, 2005). Out of the total export earnings, agricultural exports contributed more than 70 percent. The leading export partners of Uganda are Kenya, Belgium, Netherlands, France, Germany and Rwanda with each country accounting for 15.1%, 9.9%, 9.7%, 7.1%, 5.1%, and 4% of the total exports respectively (FAO, 2005). While Uganda is Kenya’s leading single trade partner both globally and regionally, the European market consumes more than 32 percent of Uganda’s aggregate exports. Tanzania’s Agricultural Sub Sector

Tanzania’s economy depends heavily on agriculture, which accounts for almost half of GDP, provides 85% of exports, and employs 80% of the work force (FAO, 2006). Topography and climatic conditions,


213

however, limit cultivated crops to only 4% of the land area. Industry traditionally featured the processing of agricultural products and light consumer goods. The World Bank, the International Monetary Fund, and bilateral donors have provided funds to rehabilitate Tanzania's out-of-date economic infrastructure and to alleviate poverty. In 2005, Tanzania realized export earnings to the tune of US$1.581 billion with agricultural exports contributing more than 85 percent of the total earnings (FAO, 2005). The major export commodities in Tanzania are coffee, cashew nuts, manufactures and cotton. The leading export partners of Tanzania are China, Canada, India, Netherlands, Japan, Kenya, and Germany which account for 10.2%, 8.6%, 7.3%, 5.2%, 4.5%, 4.4%, and 4.3% respectively of the country’s total exports (FAO, 2005). Just like in Kenya and Uganda, subsistence farming and export of primary agricultural products hinder rapid industrialization. Regional Integration

Kenya, Tanzania and Uganda have had a history of co-operation dating back to the early 20th century, including the Customs Union between Kenya and Uganda in 1917, which the then Tanganyika joined in 1927, the East African High Commission (1948-1961), the East African Common Services Organization (1961-1967) and the East African Community (1967-1977). In 1977, the East African Community collapsed after ten years, amid disagreements caused by dictatorship in Uganda, socialism in Tanzania, and capitalism in Kenya. This resulted in the three member states losing over sixty years of co-operation and the benefits of economies of scale. Each of the three states had to embark upon the establishment of services and industries that had previously been provided at the Community level at a great cost. The EAC made such political and economic sense that it was inevitable that its revival would be touted once the political climate in the region stabilized. It was no surprise, therefore, when the Treaty for East African Co-operation was signed in Arusha, Tanzania, on November 30, 1993, and a Tri-partite Commission for Co-operation established. A process of re-integration was embarked on, involving tripartite programmes of co-operation in political, economic, social and cultural fields, research and technology, defense, security, legal and judicial affairs. The East African Community was finally revived on 30 November 1999, when the Treaty for its re-establishment was signed. It came into force on 7 July 2000, twenty-three years after the total collapse of the defunct Community and its organs. The reinvigorated East African Community (EAC) articulates itself as based on the principles of good governance deemed to include adherence to democratic principles, the rule of law, accountability, transparency, social justice, equal opportunities, gender equality and most pertinently in this context, “recognition, promotion and protection of human and peoples’ rights in accordance with the provisions of the African Charter on Human and Peoples' Rights (ACHPR). Pertinent Issues in Regional Agricultural Sub Sector

Review of the status of the agricultural sub sector and regional integration in the three East African countries raises some pertinent issues. What options are there for expanding agricultural exports? Is transformation of the large subsistence sector into a commercial sector an option? Has regional integration positively contributed to enhanced agricultural exports? What role do the exchange rates, international prices play? Policy options depend on the evaluation and isolation of the drivers of agricultural exports. This paper examines the determinants of agricultural exports in Eastern Africa as well as the impact of regional integration and differences arising across countries in the region.


214

Theoretical Considerations Aggregate supply response

The agricultural sector, unlike the industrial sector, is considered in many countries to be non-responsive to policy incentives. Raising taxes from the sector, while it provides resources for investment in the industrial sector, it leaves the level of agricultural production unchanged. Growth of agricultural exports translates into a higher contribution of the agricultural sector to Gross domestic product. In Eastern Africa expansion of the agricultural sector is associated with economic growth. Therefore agro based countries aiming at spurring economic growth should ignore the agricultural sector at their own peril. The efficacy of export policy depends on the responsiveness of the agricultural sector. In general, policies biased against agriculture have done more harm than good, reducing growth in the agricultural sector and consequently in the economy as a whole (Bautista et al, 1993). Modeling Supply Response

Many studies have evaluated aggregate agricultural or individual crop supply response both in developed and developing countries using time series data. The commonly used approaches are Nerlove (1956, 1958, and 1979) model, Griliches (1960) model, and error correction and co-integration models (Hallam and Zanolli, 1992, Banerjee et al, 1993, Townsend and Thirttle, 1995, Abdulai and Rieder, 1995, Townsend, 1996, Ahmed, 2000). Both the Griliches and Nerlovian models are single equation and partial equilibrium models since they do not characterize the linkages between commodities or groups of commodities via a matrix of cross price elasticities and also ignore the interaction between agricultural and non-agricultural sectors (McKay et al, 1997). However, Nerlovian models are still able to pin point inherent policy implications in the agricultural sector. Nerlovian model describes dynamics of agricultural supply by incorporating price expectations and/or adjustment costs. The general form of the Nerlovian supply model can be specified as:-

X*t = a + bpe

xt, + µt (1)

Where X*t is the “desired” or equilibrium output X at time t and pe

xt is the expectations of price Px in time t formed at time t-1. When the dynamics of agricultural supply is driven by price expectations only, the desired output equals the actual output (X*

t = Xt). The Nerlovian price expectations model is assumed to be adaptive since producers revise their price expectations for the current period in proportion to the error in the previous period. This model can be expressed as:- Pe

xt - Pex, t-1 = β (Px, t-1 - P

ex, t-1)

Pext = β Px, t-1 + (1 - β) Pe

x, t-1 + µt (0 < β <1) (2)

T i-1

Pext = ∑ (1 - β) Px, t-1 + µt

i=1

Substituting (2) in (1) and application of Koyck’s reduction procedure (Johnston, 1984) gives a finally derived Adaptive expectations equation (3).

Xt = Φ0+ Φ1 Px, t-1 + Φ2X, t-1+ νt (3) Where Φ0 = βa Φ1= βb Φ2 = (1 - β) νt = µt - (1 - β) µt-1 and µt, µt-1 are error terms. Adjustment costs resulting from moving factors across sectors economy wide can cause lag in output response to price changes. Majority of cited studies on aggregate supply response ignore farmers price expectations and concentrate on partial adjustment hypothesis (McKay, 1997). In the formulation the change in actual output is a fraction of the discrepancy between the desired output(X*

t-1) and actual output (Xt-1).


215

Xt - Xt-1= λ (X*t-1- Xt-1)

Xt = λ X*t-1 + (1 - λ) Xt-1 (0 < λ < 1) (4)

Assuming the expected price is the lagged price, substituting (4) in (1) followed by Koyck’s reduction procedure (Johnston, 1984) gives a finally derived Partial Adjustment equation (5).

Xt = ψ0+ ψ 1 Px, t-1 + ψ 2X, t-1+ ψ 3X, t-2 + ξt (5)

Where ψ0= βλa ψ 1= βλb ψ 2 = [(1 - β) + (1-λ)] ψ3 = [(1-β) (1-λ)] ξt = λ µt - (1 - β) λ µt-1 and µt, µt-1 are error terms. The “Adaptive Expectations” model emphasizes price uncertainty as the determinant of production lags while “Partial Adjustment” model stresses technological uncertainty as the key determinant of these lags. There are conceivable circumstances when both forms of uncertainty are present (Johnston, 1984). Under such circumstances a “mixed model” is used, but presents estimation problems. A choice has to be made between the two Nerlovian models. In situations of price uncertainty the Adaptive Expectations model is preferable, whereas in situations where price uncertainty is removed by government guaranteeing of producer prices, the partial adjustment model is applicable (Griliches, 1967). Even though the world market is turbulent, international trade in agricultural commodities is associated with contractual arrangements through either bilateral or multilateral agreements between trading partners. Such arrangements tend to improve certainty of international transactions favoring choice of partial adjustment model in this study. The estimated partial adjustment model is as follows:-

Xti = β0i + β1i X (t-1) i + β2i V (t-1) i + β3i E (t-1)i + β4iI (t-1) i + β4iTi + µti (6)

Where Xti and X (t-1) i are current and previous year’s exports supply for country i, i= 1, 2, 3. V (t-1) i = Real value per unit of exports for country i. E (t-1) i = Country i’s exchange rate in period t-1 I (t-1) i =Dummy for integration in country i. Ti = Time trend as a proxy for technological change in country i. β0i...β4i and µti are country specific parameters to be estimated and error terms. Other variables such as regional weather patterns, even though considered crucial for agricultural production, were omitted in the model due to non-availability of data. Methods Data Types and Sources

The types of data used in the study include volumes and values of aggregate agricultural exports, unit value of aggregate agricultural exports, value of exchange rates, participation in the East African community and budgetary allocation to the agricultural sectors in the three east African countries. The study used mainly time series data for the period 1974-2004. The data was retrieved from a variety of data banks. Volumes and values of agricultural exports were retrieved from the FAO trade statistics data bank. Country specific exchange rates data was retrieved from the IMF statistical data bank. Other sources of data were annual reports of Central banks and Ministries of finance from in the respective countries.

Data Analysis

Descriptive statistics were used to describe emerging trends of the key determinants of agricultural exports. The study used a linear form of Nerlovian Partial adjustment model to fit data for 1974-2004 because it exhibited a better fit than its log linear form. Multiple regression analysis was used to estimate the aggregate agricultural export supply function. Durbin-Watson’s D-statistic was used to scan for serial correlation in the time series data. Multicollinearity was tested using Farrah-Glauber test and Klein’s rule.


216

Results and Discussion Fig 1 shows the trend of agricultural exports in Kenya, Uganda and Tanzania. After recovering from the oil crisis of the early 1970’s, agricultural exports in the three East African countries grew steadily over the last three decades, with Kenya recording higher exports for the better part of the period than its two neighbors. Figure 1: Agriculture export trend across countries 1976-2004

Tanzania

Kenya

Uganda

0

100000

200000

300000

400000

500000

600000

700000

800000

900000

1000000

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

Years

Export

volu

me(M

T)

A steady growth in agricultural exports in Kenya is attributed to a mixture of macroeconomic stability, policy incentives and government restraint. Macroeconomic stability ensured a stable exchange rate thus stabilizing exports income. Policy incentives have also been used in the same period to expand exports. Such incentives came in the form of input subsidies, crop and livestock development loans, market and other infrastructure development programmes and export compensation schemes. Success of the horticultural sector, on the other hand, has been attributed more to government restraint than motivation. It is one sector that has seen rapid expansion with limited interference from the government. Fig 2 shows the exchange rate fluctuation in Kenya, Uganda and Tanzania over the last three decades. Kenya has comparatively experienced a relatively stable exchange rate regime in the region. Uganda’s currency recorded prolonged stability between 1975 and 1987 with its value comparing strongly with the value of the Kenyan currency, but relatively stronger than the value of the Tanzanian currency. However, the post 1987 period witnessed drastic fluctuation in Uganda’s exchange rate with its value falling way below the value of Tanzanian and Kenyan currencies. Even though the Kenyan exchange rate appears comparatively stable in the post 1987 period when plotted against the other two regional currencies, some fluctuation is recorded when the currency is analyzed individually.


217

Figure 2: Regional real exchange rate fluctuations by country 1876-2006

Tanzania

Kenya

Uganda

-1000

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Years

Exchange r

ate

Table 1: Estimated Agricultural Export Coefficients for Eastern Africa.

Variables Country

Kenya Tanzania Uganda

R 0.729 .843 0.965

R2 0.531 0.710 0.931

Adjusted R2 0.421 0.649 0.916

Durbin -Watson Statistic 1.94 2.2 1.9

Constant -155997 (-0.711) 21647.60 (0.263) 18407.664b (2.046)

Lag agric. exports 0.104 (0.268 ) .335b (1.881) 0.409a (3.024)

Real unit value 18.7 (0.69 ) 2.413b (1.737) 7.352a (3.272)

Real exchange rate 10995a (2.65) 45.976 (0.667) 17.249a (2.923)

Integration -19173 (-1.368) 101693.036a (2.705) 809.910 (0.203)

Time trend 568.821 (0.063 ) 7817.323a (2.531) -1964.613a (-3.375)

a- Significant at 1 % level of significance

b- Significant at 5 % level of significance

t-values are in parentheses Results indicated that previous periods exports, integration in the East African community, real unit value of agricultural exports, a proxy for international prices, and technological progress over the last three decades significantly influenced agricultural exports in Tanzania. Such results can be used to convince Tanzania to enthusiastically embrace the East African Community since its economy is already reaping


218

the benefits of integration. The real exchange rate is not so crucial in Tanzania as far as agricultural exports are concerned. Results showed further that in Kenya real exchange rate is the most significant factor influencing agricultural exports. As discussed earlier (fig 2) real exchange rate in Kenya has comparatively undergone prolonged periods of stability over the last three decades. However, considered individually the Kenyan currency has exhibited some fluctuations though not as drastic as witnessed in Uganda and Tanzania in the post 1987 period. After devaluation of the Kenyan currency during implementation of the structural adjustment programmes the value of the Kenyan shilling fell to about 90 shillings a dollar before stabilizing between 65 and 80 shillings a dollar. Liberalization of the exchange rate and removal of import licensing and export taxes boosted export crop values and earnings (Odhiambo et al, 1998). Stability in the Kenyan currency has tended to create predictability in expected export earnings thus stabilizing agricultural export growth in Kenya. Appreciation of the Kenyan Shilling causes a lot of uneasiness in the tea, coffee and cut flower industries due to the accompanied erosion of profits. In Uganda, previous periods exports, international prices and exchange rate significantly influenced growth of agricultural exports in the last three decades. This shows that agricultural exporters in Uganda formulate their export plans based on their previous period’s experiences. Therefore review of past behaviors of export firms may yield some insight into their future behavior. Since international prices are crucial determinants of agricultural exports in Uganda, and are known to be fragile, contractual arrangements between Uganda’s agricultural exporters and trading partners can cushion the country from unpredictable world markets. Devaluation of the local currency (fig 2) led to drastic loss in value of the Ugandan shilling resulting in massive gains by agricultural exporters in the country. This explains why agricultural exporters in Uganda take depreciation of the local currency as a sign of better things to come. In addition, results point towards a technological decline in the agricultural sector over the last three decades. This could be attributed to the dictatorial regime in the early 1970’s which expelled a number of foreign investors from Uganda and the upheavals that followed after it was toppled hindering expanded investment in the agricultural sector. Consequently, a large subsistence agricultural sector continues to persistence in Uganda. The lack of clear benefits from integration in Kenya and Uganda could be attributed to measurement errors during data collection at the various government agencies. However, this calls for a more empirical evaluation of the benefits and costs of regional trade among the East African countries. Concluding Remarks Agriculture contributes substantially to the overall growth of Eastern Africa. A large proportion of the population in the region, who live in rural areas survive on agriculture. Agricultural export earnings in Eastern Africa, apart from providing foreign exchange for financing import dependent domestic industries, constitute a substantial proportion of the total exports. In an effort to evaluate drivers of agricultural exports in Eastern Africa, the study isolates a number of issues. Some of these issues include exchange rates, regional integration, technological progress, international prices and previous experience. While some similarities are noted, some differences are recorded across countries. Some factors are more significant in one country but less significant in another country. We recommend policies that are outward looking with governments in the region creating conducive business environment to facilitate more foreign investment in export oriented agriculture. It is crucial to tailor measures for tackling country specific peculiarities.


219

References Abdulai A. and P. Reider( 1995): “ The Impact of Agricultural Price policy on Cocoa Supply in Ghana:

an Error Correction Estimation. Journal of African Economies, 4(3): 315-35. Ahmed N (2000): “Export response to trade liberalization in Bangladesh: a co-integration analysis”.

Applied Economics, 32(8): 1077-1084(8). Banerjee A. J Dolado and J. Glabraith and D. Hendri (1993): Co-integration, Error Correction and the

Econometric Analysis of non-stationary data, Oxford. Oxford University Press. Bautista, R. and A.Valdes (1993): The Bias Against Agriculture. San Francisco: ICS. FAO (2006): Food and Agriculture Organization Trade Statistics. FAO (2005): Food and Agriculture Organization Trade Statistics GOK (2005): Economic Survey. Ministry of Planning and National Development. GOK (2002): Statistical Abstract. Central Bureau of Statistics. Ministry of Planning and National

Development. Griliches Z. (1960): “Estimates of the Aggregate US Farm Supply Functions”. A Journal of Farm

Economics, 42(2): 282-293. Griliches Z. (1967): “Distributed Lags: A survey”. Econometrica, 35: 16-49. Hallam D. and R. Zanoli (1992): “Error Correction Models and Agricultural Supply Response”,

European Review of Agricultural Economics, 2: 111-20. Johnston J. (1984): Econometric Methods. McGraw Hill Book Company, New York. Mckay A., O. Morrissey and Vaillant C. (1997): Aggregate Export and Food Crop Supply Response in

Tanzania. DFID-Trade and Enterprise Research Programme (TERP) Credit discussion paper 4(CDP04). Center for Research in Economic Development and International Trade. University of Nottingham.

Nerlove M. (1956): “Estimates of the Elasticities of Supply of Selected Agricultural Commodities”.

Journal of Farm Economics: 38: 496-512. Nerlove M. (1958): “Distributed Lags and Estimation of Long Run Supply and Demand Elasticities”.

Journal of Farm Economics: 40: 301-314. Nerlove M. (1979): “The Dynamics of Supply: Retrospect and Prospect”. American Journal of

Agricultural Economics. Odhiambo M.O, H.K. Maritim and Kidane W. (1998): “The Impact of Reforms of Parastatals on

Agricultural Development and Food Security in Eastern and Southern Africa”. Proceedings of Sub-Regional Expert Workshop. FAO Sub-Regional Office for Eastern and Southern Africa


220

Townsend R and C. Thirtle (1995): “Dynamic Acreage Response an Error Correction Model for Maize and Tobacco in Zimbabwe, University of Reading”. Discussion Paper in Development Economics, Series G 2(20)

Townsend R. (1996): “Price Liberalization, Technology and Food Self-Sufficiency. An Analysis of

Summer Grains in South Africa”, mimeo.


221

SOUTH AFRICAN LAND AND MARKET REFORMS: EQUITY VERSUS EFFICIENCY

Olubode-Awosola OO Dept. of Agricultural Economics

Faculty of Agricultural and Natural Sciences University of the Free State

P.O. Box 339, Bloemfontein 9300, South Africa E-mail: [email protected]

H.D. Van Schalkwyk,

Department of Agricultural Economics, University of the Free State, Africa,

E-mail: : [email protected]

Abstract This study makes a contribution to the land redistribution policy, which is presently not only one of the most definitive political and development issues, but perhaps the most intractable in South Africa. The study develops and uses a mathematical model for regionalised farm-level resource use and output supply response to show that the current policy requires more economic imperatives, as it tends towards smallholder agriculture that cannot produce adequate yields to meet either domestic demand or a tradable volume. Given the challenges of a free market and the fact that the settled small-scale, resource-poor (mainly black) farmers are less efficient compared to large-scale (mainly white) farmers from whom government transfers land, the study compares and prescribes land redistribution strategy that considers equity with efficiency. The study further suggests that agricultural land may act as a safety net for the poor, where the efficiency argument does not hold. Keywords: farm-level, supply response, equity, efficiency, land and market reform, policy analysis Introduction In an economy, factors that determine agricultural supply include resource availability which itself is a function of the climate. Government can enhance agricultural supply with the use of policy that encourages effective allocation of existing resources, increases the rate of use of the existing resources and encourages the competitive industry structure, amongst others. South Africa is sub-tropical along the east coast and is characterised by prolonged droughts. The climate determines the spatial distribution of farm resource use and output supply. Subsequently, agricultural supply varies from region to region and within each region (DWAF, 2002). In addition, during the greater part of the twentieth century, the government, through a number of policy measures, supported commercial large-scale (mainly white) farmers. The economy was protected from the uncertainty in world prices. On the other hand, smallholder (mainly black) farmers did not have access to information, support services and improved technology. Therefore, the difference between black (mainly subsistence) and white (mainly commercial) farmers is huge in relation to farm resource use and output supply (NDA, 2004). Since 1994, one policy measure meant to correct the imbalance of resource use and output supply in South Africa is Land Redistribution for Agricultural Development (LRAD). LRAD projects are meant to assist the portion (black, women and youth) of the population in land acquisition, thereby settling and supporting small-scale commercial farmers. The LRAD programme involves transferring 30% of


222

farmland under large-scale commercial farmers to settle a number of small-scale commercial farmers before 2015 (DLA, 2006). In addition, trade is liberalised, the market deregulated and subsidies and price supports to large-scale commercial agriculture are removed to enhance competitiveness. However, the literature shows that some of the problems in the farm industry may be attributed to the market and land reform implementations. For example, land reform and its implementation are raising, among other problems, uncertainty about property rights, insecure land tenure, free-rider problems, land invasion and crime in the farming communities (Ortmann & Machethe, 2003). Trade liberalisation and market deregulation expose the farmers to risks associated with the vagaries in world prices and exchange rate volatility. For instance, while exports have grown rapidly since 1990, imports have grown even faster in some sub-sectors of agriculture because of tariff reductions (Kirsten, 1998). At one point, the rate of farm sequestration increased due to a rising debt/asset ratio resulting not only from the effect of bad weather, but also through market deregulation, the elimination of government support to commercial farmers and relatively high nominal interest rates (Van Zyl, 2001). Land and market reforms may affect agricultural supply and consequently regional and national food self-sufficiency in the short term and the near future, as is theoretically plausible to expect farmers to respond to the changes in the agricultural policy by changing their level of resource use and output supply, in an effort to maximise farming profits. This insight is based on the argument by Just (1993) that farmers do respond to changes in exogenous variables such as price or policy variables by changing land allocation and/or cropping patterns. Moreover, production and price risks might affect these farmers to different degrees, since they have a distinct efficiency level, resource endowment and experience. The effects may also differ on different farming enterprises. This study estimates risks in the revenues of selected production activities, simulates ‘representative’ farmers’ risk attitudes and incorporates the risks and risk attitudes into the model. The model is applied to simulate potential changes in resource use and output supply as a result of the implementation of land redistribution, given the challenges of a free market. Methodology In this study, a case study of Free State province was undertaken because an analysis of the effects of changes in policy and development strategies, on resource use and output supply response, might be complex at national level. Agriculture happens to be very important in a number of contexts to the Free State province. So also is Free State agriculture important to the South African agriculture as a whole. The study explores the advantages of mathematical modelling and as much as possible, minimises the problems of such methodology. For example, to avoid over-specialisation, which is a common problem in mathematical modelling, the study uses the Positive Mathematical Programming (PMP) calibration approach (Howitt, 1995). Efforts were also made to make the model’s specification and calibration as rich and realistic as possible by incorporating risk and farmers’ risk attitudes into the model. Previous trends in regional output producer prices and yields were used to estimate the risk in production revenues. The model was also calibrated to an a priori supply response that was estimated with econometric models as reported in the literature (BFAP, 2006). The model features constraints due to resource availability and land quality distribution. Data used At regional level, regional data such as hectares allocated to crops, numbers of animal breeding stocks and output levels of some farming activities were used. These data are taken as farmers’ or farm types’ decision variables; models are often calibrated to these variables (Paris and Arfini, 2000). In this study


223

however, judicious use of regional and farm-level data was made as allowed by the PMP modelling approach. At farm-level, enterprise budget data for each production activity, namely the unit costs of resources, resource requirement per activity level, yield, output prices and average activity level were used. These data were sourced from Combud Enterprise Budgets. The Combud Enterprise Budgets are compiled and updated from time to time by the Provincial Department of Agriculture (PDA) for the homogenous production sub-regions in each province. Time series data between 1994 and 2004 on farm gate output prices, producer price index (output) and yields were used to formulate the probability distribution of the revenues. Base year (2004) data on resource use and output supply at regional and farm-type levels were used as variables in the model according to the PMP modelling approach. The data include those necessary for accounting equations and resource constraints, activity levels, policy variables such as the proposed rural land tax, water quotas, farmland and irrigation water availability at farm- and regional-level, farmland prices and rents, the number of farm units, crop and animal products supply levels, etc. These data were sourced from the reports of censuses of commerical agriculture, reports on the survey for the drought relief programme in the 5 zones of the province, the agricultural information database at the Free State PDA, the database of LRAD projects from the Department of Land Affairs (DLA), the national register of water use from the national Department of Water and Forestry Affairs (DWAF), etc. Data and model validation The data were validated in consultation with resource persons (extension officers, agricultural economists, agronomists, etc.) from the PDA, DLA, DWAF, etc. using their knowledge and experience to validate the data. Additionally, data from other sources were used to cross-check the base source data. The model is validated in its capability to reproduce observed data. The model reproduces almost exactly the base activity levels. It also reproduces exactly the observed base period allocation of land among cropping activities and among farm types at regional level. It is also calibrated to an a priori supply response at all levels. Policies on land redistribution and market deregulation were conceptualised into some scenarios. The model was used to simulate the possible impacts of these scenarios. The effects on farm-level supply curves were examined. The farm type supply curves were aggregated into a regional supply response. Results and discussions Output supply response based on Scenario I

In Scenario I, the effects of the risk in the marginal revenues of the selected production activities and the trends in the number of farm units are combined. The technical progress in the farm industry, as found in the literature, is also assumed. The cost of production is assumed as constant. The number of farm units in the large farm type has been decreasing at the arithmetic mean of 129 farm units per year from 1994 to 2004. This trend is assumed to continue to decrease from 8,531 in 2004 to 7,112 in 2015. This is a decline of about 17%. However, the number of farm units in the small farm type is assumed to continue increasing from 495 in 2004 to 8,635 in 2015 at the arithmetic mean of 740 farm units per year. It is acknowledged that this increase is very high as such an increase has never been recorded in the Free State LRAD programme. However, this assumption is based on the proposal by the Ministry of Agricultural and Land Affairs which hopes to transfer land at the rate of 2.2 million ha per year from 2006 to 2015, in order to reach the target of transferring 30% of farmland from commercial agriculture by 2014.


224

Table 1 shows that for the large farm type, despite that the expected marginal revenues of white maize, yellow maize and wheat may be higher in 2015, the supplies decrease for all crops and animal products by an average of 15.23%. The general decrease in the supply of the crops and animal products could be explained by the decrease in the number of large farm units. The decline could have been more pronounced if technical progress had not been incorporated into this scenario. The differences in the decline in crop and animal products could be attributed to different risk levels and expected marginal revenues. Table 1: Base level and % changes in supply as a result of Scenario I

Base 27.50% land transfer

Large farm type Small farm type Region Large farm type

Small farm type

Region

No of farm units 8531 495 7112 8640 Crop (ton) (%) White maize 2718395 6055.29 2724448.07 -15.03 1657.51 -11.32 Yellow maize 1617420 12557.90 1629976.83 -15.12 1657.14 -2.24 Wheat 517674.00 280.78 517956.78 -14.56 1658.85 -13.65 Soya beans 30508.34 - 30508.34 -15.33 - -15.33 Sorghum 162899.11 - 162899.11 -15.14 - -15.14 Sunflower seed 269342.58 - 269342.58 -15.54 - -14.54 Livestock (ton/litre/unit) (%) Beef-cattle 60255.18 336.18 60591.36 -15.41 1657.81 -6.13 Mutton 30000.00 150.92 30150.88 -15.38 1657.98 -7.00 Pork 10233.02 259.77 10492.79 -15.31 1658.36 26.13 Broilers-chicken 74360.98 8.14 74369.12 -15.18 1659.02 -15.00 Layers-eggs 515199600 568755.24 515768700 -15.36 1658.07 -13.52 Dairy milk 357984700 274069600 360725400 -15.38 1657.99 -3.66

Source: Own simulation results from the model


225

For the small farm type, as expected, the supply curves for crops and animal products are shifted to the right at an average of about 1,658%. This could mainly be explained by a massive increase (1,645%) assumed in the number of farm units. It confirms that percentage change in the number of farm units could lead to a more or less proportional change in the supply curves. These data show the overwhelming effects of the increase in the number of farm units and technological progress. An increase in the number of small farm units shifts the regional supply curves to the right. This looks promising with respect to establishing more developing farm units as a means of redressing the imbalance in the industry. However, these results should be interpreted with caution. It is obvious that some LRAD farms will be established. However, one burning issue in agrarian reform remains the productivity and efficiency among the LRAD farms, as the government lacks enough resources to provide integrated support services which would enhance productivity amongst the LRAD farmers. Additionally, the decreasing effects that the declining number of large farm units has on the regional supply curves crowd-in the increasing effects that the increasing number of small farm units has on them. This happens for all crops, but is especially pronounced in the supply responses for soya beans, wheat, sorghum, sunflower seeds, broilers-chicken and layers-eggs. These are relatively capital-intensive production activities. The net positive change in the regional supply curve for pork production is as a result of the observation that a relatively high proportion of the small farm type is engaged in rearing pigs for pork production. It is the only production activity where small farm units produced about 2.28% of the regional production. This result shows the implication of small farm types not having enough capital and other resources necessary to engage in capital-intensive production activities. Output supply response based on Scenario II In this scenario, the assumed technical progress for the small-scale farmers in Scenario I was dropped, as an indication exists that most LRAD farms have not been able to make a substantial contribution to the regional supply. This may, among other things, be attributed to a lack of capital, experience, and the like, which are necessary for large volume production. Columns 2, 3 and 4 in Table 2 show that the assumed technical progress has negligible effects on the simulated production levels of the small farm type. It may therefore be inferred that multiplying the number of small farm units has a negative implication on regional resource use and output supply, especially where land is being transferred from the large farm type to settle small farm units. The decline in the supply response is, on average, about 0.25%. These results raise concern about the policy objective to settle a large number of small farm units which are less efficient compared to large farm units.


226

Table 2: % changes in supply as a result of Scenarios II and III

Scenario II Scenario III

Large farm type

Small farm type

Region Region Region Region

Crop (%) 1% 5% 10% White maize -15.03 1645.45 -11.34 7.32 14.60 24.04 Yellow maize -15.12 1645.45 -2.33 17.95 25.69 35.69 Wheat -14.56 1645.45 -13.66 5.92 15.20 27.34 Soya beans -15.33 - -15.33 3.78 9.48 16.76 Sorghum -15.14 - -15.14 -32.73 -29.15 -24.58 Sunflower seed -15.54 - -14.54 -30.86 -24.47 -16.10 Livestock (%) (%) Beef-cattle -15.41 1657.81 -6.13 -2.78 1.13 6.02 Mutton -15.38 1657.98 -7.00 -1.24 2.91 8.13 Pork -15.31 1658.36 26.13 -11.47 -7.89 -3.40 Broilers-chicken -15.18 1659.02 -15.00 -1.52 3.33 9.51 Layers-eggs -15.36 1658.07 -13.52 2.29 6.78 12.45 Dairy milk -15.38 1657.99 -2.66 2.07 6.49 12.07

Source: Own simulation results from the model Output supply response based on Scenario III This scenario presents a picture of a pursuit of equity with more economic imperatives. This is not an objection to the land reform process in the South African context, but the presentation, from another perspective, of a more efficient method of agrarian reform. Government may target the farmland of inefficient large farms for redistribution, to settle black farmers who have a proven commitment to farming as an economically viable activity. This may assist the settled farmers to gain economies of scale, as compared to small farm types that are numerous in number but with low productivity, as observed in most under-developed and developing countries. It has been established earlier in this study that the means by which a nation may be more productive and thus become wealthier, is to allocate the existing resources efficiently and to increase the rate of use of such resources. From the previous scenario, the yearly decrease of about 129 farm units in the large farm type will result in about 7,112 farm units by the year 2015, when 30% of the transfer will have been achieved. The farmlands of 1419 large-scale farmers, who cease to be active, may be transferred to settle large-scale committed (black) farmers. Expected marginal revenues for 2015 are assumed. Each of the activity levels has different level of revenue risk. Technical progress is also assumed. Taking size as an indication of efficiency, an increase in technical progress at 1%, 5% and 10% was simulated in this scenario. Among and in addition to other changes, the increase in technical progress is expected to shift the supply curves to the right for each crop and animal product. Furthermore, the economics, the risk in the marginal revenues of each crop and animal product, coupled with the risk attitudes of the farm types, are expected to have effects on the supply curves. The last three columns in Table 2 show possible impacts of the assumed increase in technical progress. It is observed that at a 1% increase in technical progress, the supplies of white maize, yellow maize, wheat, soya beans, layers-eggs and dairy milk would increase, while the others would decrease. This may be explained by the relative level of risk in the marginal revenues and the magnitudes of the gross margins, which resulted in substitution between the enterprises. At a 5% increase in technical progress, more of


227

the enterprises have positive responses. Therefore, these results show that having more farmland, which may imply reduced lack of capital constraints as assumed in this model, would not necessarily imply a general increase in all the production activities. At a 10% increase in technical progress, there is an increase in response. Conclusions In Scenario I, it may be inferred that a decline in the number of farm units shifts the farm-type supply curves by almost the same proportion. It is important to note that the decreasing effects that the declining number of large farm units has on the regional supply curves, crowd-in the increasing effects that the increasing number of small farm units has on the regional supply curves. This happens for all crops, but is particularly pronounced in relatively capital-intensive production activities namely soya beans, wheat, sorghum, sunflower seeds, broilers-chicken and layers-egg productions. Scenario II indicates that multiplying the number of small farm units has a negative implication for regional resource use and output supply, especially where land is being transferred from a more efficient large farm type. Scenario III shows a possible picture of an agrarian reform that allows the emergence of a larger farm unit and assisting a previously disadvantaged portion of the population, who have a proven commitment to farming as a business. Policy implications Policy needs to discourage settling small-scale farmers presently observed in the LRAD projects. Land reform may limit the production of the large-scale farm sub-sector, especially if the farmland is transferred from a large-scale farmer to proliferate a number of small-scale farmers who are less efficient. This could also lead to a poverty trap for the settled farmers and land fragmentation, which has consequences for large production necessary for export surplus. The land of inactive and less-successful large-scale farmers who are bankrupt, can be targeted for redistribution to settle black farmers with a proven commitment to farming on a large enough farmland. This will enhance such farmers’ competitiveness. Small-scale farmers may be settled, but only on very intensive projects with high-valued crops such as vegetables on irrigation projects. Such farm units may be small in size but big in turnover. However, more research is necessary on this approach. It is in the interest of all the stakeholders in the farming industry to implement land transfer and capacity development of intended beneficiaries simultaneously and quickly, especially in the use of risk-hedging mechanisms and the art of enterprise diversification. It is noted that rural land may act as a safety-net. References Bureau for Food and Agricultural Policy (BFAP), 2006. The grain, livestock and dairy sector model.

University of Pretoria, Pretoria, South Africa Department of Land Affairs (DLA), 2006. Television interview with President Thabo Mbeki with SABC

2, Broadcast on Sunday 5 February, 2006, South Africa. Department of Water and Forest Affairs (DWAF), 2002. Proposal for the establishment of a catchment

management for the Olifants Water Management Area. Version 2.2 BKS (Pty) Ltd for DWAF, South Africa.

Howitt, R.E., 1995. Positive Mathematical Programming. American Jounal of Agricultural Economics,

77(2):329-342.


228

Just, R.E., 1993. Discovering production and supply relationships: Present status and future opportunities.

Reveiew of marketing and agricultural economics 61(1):11-40. Kirsten, J., 1998. How government policy is taxing South African agriculture: outlook. Farmers Weekly

Iss 88049, Dec. 4, pp.30-31. National Department of Agriculture (NDA), 1995. National Department of Agriculture White Paper on

Agriculture, ISBN 0-621-16111-x National Department of Agriculture (NDA), 2004. AgriBEE, Broad-Based Black Economic

Empowerment Framework for Agriculture, Pretoria, South Africa. Ortmann, G. and C. Machethe, 2003. Problems and opportunities in South Africa Agriculture. In L

Niewouldt and Jan Groenewald (eds). The challenge of change, agriculture, land and the South African economy. South Africa, Pietermaritzburg: University of the Natal Press, pp. 47-62.

Paris, Q. and F. Arfini, 2000. Funzioni di costo di frontiera, auto selezione, Rischio di prezzo, PMP e

Agenda 2000. in Rivista di Economia agrarian, 2:211-242. Van Zyl, J., 2001. Writing on the wall for many farmers; these industries are optimistic: agribusiness:

opportunities, trend and ideas. Finance Week. ISSN:0256-0321 p 42.


229

LINKING RURAL ECONOMIES WITH MARKETS – AN INSTITUTIONAL APPROACH

H.D. Van Schalkwyk, Professor, Department of Agricultural Economics,

University of the Free State, Africa, E-mail: : [email protected]

N.A. Kotze

Master students, Department of Agricultural Economics, University of the Free State, Africa


P. Fourie

Master students, Department of Agricultural Economics, University of the Free State, Africa


Abstract For many developing countries, the agricultural sector is still the main employer, especially for women, and particularly in sub-Saharan Africa. The causes of poverty are complex and often superficially understood. As a result, efforts to resolve these problems are frequently fragmented and development interventions become severely limited in focus and reach. The development of the agricultural sector can ensure integration of the region’s economies and the resultant upliftment of rural communities. Market access seems to be one of the most limiting factors which have been identified that is hindering growth in rural agriculture. Factors influencing market access are lack of information, training and extension services, tenure systems, transport and credit. Resolving the South African problem requires a concerted, holistic, innovative and integrated approach through partnerships between the civic, public and private sectors. The market access problem should be addressed based on linkages between the small producers and markets by addressing the constraints. The fact is recognise that small farmers do not exist in isolation but is part of a large market system. Interventions are therefore well grounded in understanding the business development service markets within which small producers operate, as well as enhancing win-win linkages between rural-based service providers and small producers. Keywords: Market access, rural farmers, institutions. Background The dualistic agricultural sector creates many challenges for South Africa. The established commercial sector and the areas, in which commercial agriculture preponderates, are served by a sophisticated agricultural marketing system, with good infrastructure. This infrastructure includes roads, railways, telecommunication, postal services, marketing and financial services. Circumstances for small-scale farmers is different, marketing services to these small farmers are often poor (sometimes non-existent), as are roads, telecommunication, physical marketing infrastructure and financial services (Van Schalkwyk, Groenewald and Jooste, 2003). According to Magingxa, (2006) market access is one of the most critical factors that influence small scale farmer’s potential for success. Jooste and Van Rooyen (1996) also concluded that the transition of the small-scale livestock sector towards commercial production will ultimately be determined by its access to markets. Market access can be considered to be the reason why those farmers who produce surpluses


230

remain trapped in the poverty cycle. Smallholder access to markets is considered central in the commercialisation debate of small-scale agriculture in general because it is implicit in the view that development of smallholder agriculture can influence economic growth. Market access is an important factor in improving rural livelihoods. In addition, within the South African context, market access is also seen as feeding directly into the government’s development objectives that include poverty alleviation and economic growth. In this regard, it becomes important to understand the factors that influence market access in smallholder production. Understanding the factors influencing market access will improve the knowledge of those involved in rural development in general and smallholder management in particular. By understanding the factors one can pay the necessary attention to them when designing successful smallholder projects (Magingxa, 2006). A substantial number of studies in the recent history have put market access as one of the main ingredients to successful irrigation management. The core argument in recent studies such as Gabre-Madhin and Haggblade (2001), Hau and von Oppen (2002), Foremen and Livezey (2003) and Muhammad, Tegegne, and Ekanem (2004) is that market access needs to receive more attention in studies dealing with smallholder farmers. However, as Kherallah and Kirsten (2002) argue, the frequent occurrence of market failure in developing countries requires an institutional analysis. Markets Types of Markets There are two distinctive market types in South Africa, i.e informal and formal markets. These two differ dramatically. The informal market forms an essential part of South Africa’s marketing power and is very important to small-scale farmers. There are over 90 000 stores in the formal sector, whereas the informal sector is impossible to quantify. The informal market tends to escape the attention of policy makers and planners. This is not only due to the difficulty associated with an informal market and the complexity of gathering data, but also because the socio-economic groups involved in the market are frequently the weakest ones in the society (Marttin, 2001). Its size, volumes and revenues cannot be easily determined, as the outlets are not registered as retailers, nor do they pay rent or taxes. According to RocSearch (2004) and Wilson (2004) as quoted by Botha (2006) the informal retail market had an estimated turnover of R34 billion during 2004 and estimated that there could be over 6 000 spaza shops in South Africa. According to Botha (2006) the most common store types throughout South Africa are the smaller types, such as rural shops, urban counter shops and urban self-services shops. Figure 1 below indicates the dispersal of various outlets throughout South Africa.


231

Figure 1: Numbers of different food retail store types per region in South Africa in 2004.

0

5,000

10,000

15,000

20,000

25,000

30,000

Eastern Cape Free State /

Northern Cape

Gauteng Kw azulu-Natal Limpopo /

Mpumalanga /

North West

Western Cape

Total Majors Branded Superets Forecourt Rural Urban Counter Urban Self

Source: Botha (2006) Circumstances that Influence Markets There is growing evidence that many smallholder farmers can benefit from market-oriented agriculture. However there are fundamental barriers small-scale farmers’ needs to bridge before they can access markets. No development can be expected in large areas of South Africa, and no upliftment of the rural poor can occur in the absence of significant improvements in the marketing set-up serving poor areas and people (van Schalkwyk, Groenewald and Jooste, 2003). A study done by Cichello, Fields and Leibbrandt (2005) found that crime can be identified as the single most dominant hindrance for small-scale farmers in the informal market while other severe hindrances are the continual risk of business failure, a lack of access to start up capital, transport costs, uncertainty over profits before one starts the business and jealousy faced in the community if an individual is successful. These are all essential factors which need to be considered before market entrance can be finalised by small scale farmers in an informal market. A study by Chandra, Nganou and Noel (2002) found that lack of credit, low demand and variability of income streams, high cost of infrastructure (public transport) and services (water, electricity, and telephone) and poor access to business support centers, poor access to training, lack of storage spaces/permanent stalls, lack of transport facilities, and inadequate business space are all constraints for small scale farmers to link to markets. Factors considered as restraining small-scale farmers ability to market their cattle include lack of market information (Nkosi and Kirsten; 1993), large distances to the market place (Mahanjana, Esterhuizen and Van Rooyen; 2001), marketing infrastructure (Fraser, 1991), lack of diversity of the market outlets (Lyster; 1990), cultural and subsistence type of farming (Ainslie, Kepe, Ntsebeza, Ntshona and Turner, 2002), and Makhura (2001) mentions that small-scale farmers contribute inadequately to the mainstream market because of low production and poor access to other options for obtaining a livelihood. Other factors include, farmer training, herd size, household characteristics and support services, (Lapar, Holloway and Ehui, 2003; Coetzee, Montshwe and Jooste, 2004; Bellemare and Barrett, 2004; Nkhori, 2004). By comparing the factors influencing market access of Magingxa (2006) study on irrigation farmers and Montshwe (2006) study on the cattle industry a comparison is found in factors influencing market access in a broad spectrum of agricultural production. These factors include information, training and extension


232

services, transport and credit, all of which have a significant impact on market access for small-scale farmers. The unstandardised and standardised regression coefficients results between cattle and irrigation small scale farmers are presented in table 1. Only the variables which occurred in the cattle as well as in the irrigation framework are discussed. The interpretation of results focus on the standardised coefficients. Table 1: Standardised and unstandardised logistic regression analysis for cattle and irrigation farming

Cattle Irrigation

Variables Unstandardised Standardised Unstandardised Standardised

Tenure system -0.4517 -0.1429 2.57217 -1.04574

Information 0.8908** 0.3483 3.318946 1.617806

Training 2.2789*** 0.93 3.349972 1.108553

Transport 0.0274*** 1.3009 3.089256 1.077983

Extension services -0.1758 -0.0829 0.032404 0.835903

Credit 0.9862*** 0.3453 -0.35272 -0.84365

The * represents the significance level: 1%=***; 5% = ** and 10% = *

Information According to the standardized coefficients in table 1, information has the biggest impact on irrigation small scale farmers. Significant and sensible information is very important to improve success potential. Smallholder farmers are seldom in a position to understand what to produce, when and in what quantities or quality requirements. Small-scale farmers usually lag behind in terms of technology as a result of this and make it difficult to enter the more profitable markets. Information is crucial for any enterprise not just cattle or irrigation smallholders, access to market information, and the use thereof, results in increased participation of small-scale farmers in the mainstream markets. Since the deregulation of marketing boards in South Africa acquiring information which is accurate and timely has proofed to be an expensive and difficult task. Furthermore, the publicly available information is more often than not historical, and its validity is sometimes questionable. Availability of prompt and reliable market information regarding the movements in markets and what prices are quoted for different commodities considerably improves the decision making capability of the farmers and strengthens their bargaining power. Different institutions currently involved in the supply of information to the agricultural sector should be invited or be nominated to exercise certain information gathering, processing and dissemination duties on behalf of the agricultural sector. Institutions, which can play a major role in this regard and who may become member institutions of the Institute, are for example:

• National Department of Agriculture (Market and production data) • Agricultural Research Council (Technical data and information) • The South African Weather Bureau (Climate data)


233

• Agri SA (General information on policies etc.) • SAGIS (Field crop data) • Agrimark Trends / University of the Free State (Market and production information - value added)

All of the above institutions should have good linkages with the future users of their information, but this is especially true with small-scale farmers. In order for this to happen an Advisory Board consisting of e.g. 5 officials from producer organisations (representing field crops, livestock, vegetables, fruits and secondary products), 5 from agribusinesses (representing the agribusiness chamber, banking sector, input suppliers, traders and processors) and 3 from the consumer sector (representing wholesalers, retailers and the consumer union) should be appointed. The objective of this Advisory Board should be to identify priorities for the Institute and to find the necessary funds to finance the identified priorities (Van Schalkwyk and Frick, 2002). As small-scale farming has been identified as a priority by the government as a vehicle for growth in the rural areas, it is necessary to incorporate these farmers with the most updated information for them to make informed decisions. Information should reach small-scale farmers by means of extension officers. Training and Extension Services

Comprehensive Agricultural Support Programme (CASP) represents an important step in the Department of Agriculture’s strategy to promote agricultural production among previously disadvantaged individuals and communities (NDA, 2005). In the past, the focus was mainly on providing access to agricultural land but since 2004 CASP has a new dimension in the form of post-settlement support. This includes improved access to financing and credit for small-scale farmers and co-operatives, as well as six pillars of non-financial support services. The pillars are information and knowledge management, technical and advisory assistance, training and capacity building, market and business development, on and off-farm infrastructure services and regulatory services. There is a positive relationship between market access and training for both cattle and irrigation farming practises, as proved by the studies of Magingxa and Montshwe. This positive relationship is expected since marketing through the mainstream markets requires knowledge in terms of product specification, price determination and timing. The main problems small-scale farmers are faced with are that; farmers are not trained and therefore often apply low technology systems that do not make full use of their resources. This result in low yields per animal/land unit and high production costs per unit of product. Small-scale farmers also usually buy their inputs individually and in small quantities at the last link of the market chain and, as a result, they pay elevated prices. Farmers are not organized, they sell individually without adding value to their products and they often sell to the first buyer. Their sale prices are low and their profits are shared with intermediaries. All of these limiting and constraining factors can be addressed by the proper training and extension. Extension officers have a responsibility towards small-scale and emerging farmers in the sense that they need to provide proper support to them regarding management practises and marketing information. The obstacle which faces the extension officers is that they are not always capacitated to the necessary level in which they can help these farmers. It thus becomes necessary to rethink the training of these extension officers for them to be able to face the modern day challenges of the rural farmers and in the end become trainers for the small-scale and emerging farmers. Extension officers need the most up to date information available to them in order to provide the rural farmers with price increases, market changes, new production practices, and management expertise. A


234

range of training of trainers and organisational development courses must be offered to the extension officers to address gaps in the fields of all major agricultural commodities in their respective areas. It is clear that without improved education of extension and farmer training it is very difficult to promote change. Extension should be strengthened to increase its effectiveness and ensure that its training programmes are adapted to the resource needs and capabilities of smallholder rural families. This is where government as well as private organisations can become role players. Government can create the necessary policies or incentives for farmers to receive training. This has already been initiated through different Skills Education Training Authorities (SETA) programmes in South Africa. The AgriSETA cover all the economic sub sectors, like all types of farming, slaughtering houses, fibre processing, exporting and importing, sales and distribution, agricultural research, and marketing. Transport The third positive standardised variable for both cattle and irrigation farming is transport. The results suggest that market distances have a positive effect on participation of small-scale farmers in the mainstream markets. Transport availability is critical in accessing both input and output markets. Increased participation will take place when the real distance to markets is reduced by bringing buyers closer to the small-scale farmers, i.e. visits by speculators with the necessary transport to move animals/grain. Government have the responsibility to economic advancement in the rural areas to create the necessary roads and transportation infrastructure to support the rural farmers. Logistics refer to that part of the supply chain process that deals with the transportation, wharehousing, inventory carrying, and administration and management of physical products between the point of production and the point of delivery to the final consumer. According to the Council for Scientific and Industrial Research (CSIR, 2005) it was calculated that transportation costs make up 75% of logistics costs in South Africa, making it the biggest part of the supply chain. If the specialist rail export lines are excluded, the tonnage transported by rail has declined by nearly 20% over the past decade. In contrast road transport increased by more than 50% over the same period (CSIR, 2005). As transport is one of the most limiting factors for the small-scale farmers, methods to bring markets to them are an option which needs to be considered. Mobile banking is an option in which credit could be combined with the transport issue to increase the farmers’ access to banking facilities. A truck with an Automatic Teller Machine (ATM) on the back where withdrawals and payments can be made should run on a scheduled route through the rural areas. This could reduce transaction cost considerably regarding banking for these rural farmers. Other ways in which transportation cost to markets could be decreased or eliminated is for auctioning houses to go mobile. This will entail that once a month or less an auction is held in a different central point in the rural farming communities, where the pens are set up for the auction and afterwards broken down to transport it to the next auctioning place. This will decrease the actual distance to the market for most rural livestock farmers, and a spill over effect is that livestock will not lose as much condition before being auctioned as would have happened when they where to be transported over vast distances.


235

Tenure system Tenure systems were in both cases influenced negatively, this is because the lack in proper tenure systems decreased the farmers’ ability to access markets adequately. Roth and Haase (1998) points out that increases in market access can show dramatic increases in food production in the short to intermediate run, these gains are typically achieved under low capital intensity. Roth and Haase (1998) further states that for output to increase, tenure security become a binding constraint. At some point of production, farmers will demand high tenure security before undertaking fixed land improvements or investing in capital intensive technology. Credit supply by informal lenders becomes limiting, while formal lenders will require clear and transferable title before lending. It is doubtful whether the transition to high value crops and a high capital/labour ratio can be achieved without land tenure that confers right of sale, mortgage, and low cost transaction in the eyes of creditors. The conclusion thus is that accelerated commercialisation of smallholder agriculture will require careful attention to both issues of land tenure institutions and market access. According to Tietenberg (2003) the structure of property rights that could produce efficient allocations in a well-functioning market economy structure has three main characteristics: • Exclusivity – All benefits and costs accrued as a result of owning and using the resources should

accrue to the owner, and only the owner, either directly or indirectly by sale to others. • Transferability – All property rights should be transferable from one owner to another in a

voluntary exchange. • Enforceability – Property rights should be secure from involuntary seizure or encroachment by

others. According to Keyfitz and Dorfman (1991) there are 14 institutional and cultural requirements necessary for the operation of an effective private market, and one of these are security of persons and property. They further state that as a requisite for a properly functioning marketing system, property rights should be clearly established and demarcated as well as the procedures for establishing property rights and transferring them. It is imperative that clearly defined property rights system and land ownership be established in the rural areas. Well defined property rights are needed in order for these farmers to not only efficiently use their resources, but also to gain access to credit and become market players. Credit A study done by Batha (2003) indicated that about 90% of the people living in developing countries lack access to financial services from institutions, either for credit or saving. This includes nearly all the poor of the developing world. While not all the poor can use microfinance, there remain a massive gap between the low level of commercial microfinance available from financial institutions and the extensive worldwide demand from such financial services among low-income people. Moreover, large scale sustainable microfinance helps to create an enabling environment for the growth of political participation and of democracy. Such services are rarely accessible through the formal financial sector: credit is however widely available from informal commercial money lenders at very high cost to the borrower-especially poor borrowers. Banks generally assume that providing small loans and deposit services would be unprofitable. According to Bahta (2003) rural financial systems help to promote economic growth through the mobilization of resources and by providing financially and economically viable investments and economic activities; to further efficient resource allocation; to contribute to better income distribution and


236

poverty reduction by enabling access to financial markets also for poor; to support the building of sustainable self support rural financial institutions. As mentioned, mobile banking can increase the access that rural farmers will have to credit and other banking facilities. By using the post offices in the rural areas as banking facilities one immediately increase the reach of credit availability to the rural areas. Conclusion Present trade, marketing and institutional policies in this county make limited provision for the unique interest of emerging small-scale farmers. Furthermore, the current institutions involved in promoting market access are not co-ordinated. The various constraints to market access can be addressed through a combination of public interventions and private sector involvement. Potential solutions include overcoming infrastructure backlogs especially in the transporting sector. Small-scale farmers have limited access to transport and mobile auctioning houses should be considered, decreasing transportation cost. Other potential solutions are the improvement of market information, institutional reform (with specific reference to land tenure), supporting small-scale food processing and value-adding, and encouraging closer links between small-scale and established commercial farmers. Makhura and Makoena (2006) came to the same conclusions and indicated that Farmer-Public-Private-Sector Partnership is crucial in addressing the problem of market access of small-scale farmers in South Arica. It is important that Government recognises the diverse roles the state can play: as provider, facilitator, or partner. The state needs to use resources wisely to maximise their impact. One of the critical decisions which government needs to take is who can provide the services most cost-effectively. Currently government is providing many services that it cannot provide cost effectively. It is important to work out which are the core processes of the organisation, where the organisation can really make an impact. For other areas the state should seek either to leave a space for the private sector or Non Governmental Organisations to take up those services, or if it will not happen without subsidy, then to outsource those activities. References Ainslie, A., Kepe, T., Ntsebeza, L., Ntshona, Z., and Turner, S. (2002). Cattle ownership and production

in the communal areas of the Eastern Cape, South Africa. Research Report no. 10. University of the Western Cape.

Bahta, Y.T. (2003). Village banks, group credit, farmers’ domestic saving mobilization in Eritrea.

Department of Economics, University of the Free State, Bloemfontein, South Africa Bellemare, M.F and Barrett, C.B (2004). An ordered tobit model of market participation evidence from

Kenya and Ethiopia. Unpublished report, Cornell University, Ithaca Botha. L (2006). The evolving food retail industry and the buying behaviour of consumers in developing

areas: a Qwa-Qwa case study. Thesis in partial fulfilment of a M.Sc. Agric Agricultural Economics degree, Department of Agricultural Economics, University of the Free State

Coetzee, L., Montshwe, B.D. and Jooste, A. (2004). Livestock Marketing: Constraints, challenges and

implications for Extension Services. Paper presented at the 38th Conference of the South African Society for Agricultural Extension. May 6, 2004.


237

Chandra, V., Nganou, J., and Noel, C. (2002) Constraints to growth in Johannesburg;s black informal sector: Evidence from the 1999 informal sector survey. Discussion paper no 17. World Bank Southern Africa Department, Washington, DC

Cichello, P, Fields, G.S and Leibbrandt, M (2005) “Earnings and employment dynamics for Africans in

Post-Apartheid South Africa: A panel study of Kwazulu Natal.” Journal of African Economics 14 no 2: 143-90

CSIR, (2005). On the path to creating efficient logistics for South Africa. Export SA, April 2005 page 14-

15. Doward, A. & Kydd, J. (2005). Making agricultural market systems work for the poor: Promoting

effective, efficient and accessible coordination and exchange. Imperial College London. February 2005

Foreman, L.F. and Livezey, J.S. (2003). Factors contributing to financially successful Southern rice farms. Selected paper presented at the Southern Agricultural Economics

Association annual meeting. February 1-5, 2003, Mobile, Alabama. Frazer, G.C.G. (1991). Agricultural Marketing in Less Developed Countries with special reference to

Ciskei. Unpublished PhD Thesis. University of Stellenbosch. Stellenbosch. Gabre-Madhin, E.Z. and Haggblade, S. (2001). Successes in African Agriculture: Results of an Expert survey, September 2001. IFPRI, Washington D.C. Hau, A.M. and Von Oppen, M. (2002). Market Access and Agricultural Productivity in Doi Inthanon Villages. International Symposium: Sustaining Food Security and Managing

Natural Resources in Southeast Asia - Challenges for the 21st Century - January 8-11, 2002 at Chiang Mai, Thailand

Jooste, A. and Van Rooyen, C.J. (1996). Constraints to small farmer development in the red meat

industry. Invited paper, Red Meat Producers Conference, Roodevallei, Pretoria, 27 March 1996. Keyfitz, N. and Dorfman, R.A. (1991). The Market economy is the best but not the the easiest. Kherallah, M. and Kirsten, J.F. (2002). The New Institutional Economics: Applications for Agricultural Policy Research in Developing Countries, Agrekon, 41 (2) 111- 134. Lapar, M.L, Holloway, G. and Ehui, S. (2003). Policy options promoting market participation among

smallholder livestock producers: a case study from the Phillipines. Food Policy, Vol 28: 187 – 211.

Lyster, D.M (1990). Agricultural marketing in KwaZulu: a farm-household perspective. Unpublished

MSc Agric thesis, University of Natal, Pietermaritzburg. Magingxa, L.L. (2006). Smallholder irrigators and the role of markets: a new institutional approach.

Submitting in partial fulfilment of the requirements for the degree of PhD. South Africa, Bloemfontein.

Mahanjana, A.M., Esterhuizen, D. and Van Rooyen, J. (2001). Farming for profit. National Emerging

Red Meat Producers Organisation, Pretoria. South Africa.


238

Makhura, M. (2001). Overcoming transaction costs barriers to market participation of smallholder farmers in the Northern Province of South Africa. Unpublished Ph.D. thesis, University of Pretoria, Pretoria

Makhura. M and Mokoena. M, (2003) Market access for small scale farmers in South Africa. The

Challenge of change. Chapter 8. Agriculture, land and the South African economy. University of Natal Press, Pietermaritzburg. 137-148.

Marttin, F. (2001). Informal market, local consumption and bartering of inland fisheries. Website

available at: http://www.oceansatlas.org (Accessed on 28 February 2007) Montshwe, B.D. (2006). Factors affecting participation in mainstream cattle markets by small-scale cattle

farmers in South Africa. Submitting in partial fulfilment of the requirements for the degree of M.Sc. South Africa, Bloemfontein.

Muhammad, S., Tegegne, F. and Ekanem, E. (2004). Factors contributing to success of small farm operation in Tennessee. Journal of Extension Vol. 42 (4), August 2004. NDA (2005). National Department of Agriculture: Annual report. South Africa Nkhori, P.A. (2004). The impact of transaction costs on the choice of cattle markets in Mahalapye

District, Botswana. Unpublished M.Inst.Agrar thesis, University of Pretoria, Pretoria. Nkosi, S.A. and Kirsten, J.F. (1993). The marketing of livestock in South Africa’s developing areas: A

case study of the role of speculators, auctioneers, butchers and private buyers in Lebowa. Agrekon, Vol 32 (3).

North, D.C. (1997). The new institutional economics and third world development. Pfeffermann, G. (2002). The State, Institutions and the Market Economy. Institutions for the private

sector in transition economies. Study of the Cuban Economy Coral Gables, August 1, 2002 Roth, M. and Haase, D. (1998). Land tenure security and agricultural performance in Southern Africa.

Land Tenure Center and Department of Agriculture and Applied Economics, University of Wisconsin-Madison. June 1998.

Tietenburg, T. (2003). Environmental and Natural Resource Economics, International Edition. Peason

Education, Inc. Van Schalkwyk, H D. and Fick, A. (2002). Agricultural information in South Africa: The way forward.

Department of Agricultural Economics, University of the Free State, Bloemfontein, South Africa. Van Schalkwyk, H.D; Groenewald J and Jooste, A (2003). The challenge of change. Agriculture, Land

and the South African economy. Agricultural Marketing in South Africa. Chapter 7. p131-132. University of Natal Press. Pietermaritzburg.


239

REVIEW OF QUALITY OF LIFE INFLUENTIAL FACTORS AMONG IRISH FARM FAMILIES REPORTING DISABILITY

Shane Whelan

School of Agriculture, Food Science, and Veterinary Medicine, Agriculture and Food Science Centre, University College Dublin, Belfield, Dublin 4, Ireland.


Dermot J Ruane School of Agriculture, Food Science, and Veterinary Medicine, Agriculture and Food Science Centre,

University College Dublin, Belfield, Dublin 4, Ireland.


John McNamara Teagasc - Agricultural and Food Development Authority, Kildalton College, Piltown, Co. Kilkenny,

Ireland. Email: [email protected]

Anne Kinsella

Teagasc Rural Economy Research Centre, Athenry, Co. Galway, Ireland. Email: [email protected]

Abstract Quality of Life is influenced by a number of key influential factors (happiness; family life; health, and finances). The current literature pertaining to these Quality of Life factors were reviewed to examine if any variation existed among farm families experiencing disability relative to the general farm population. Almost 10% of Irish farm families experience disability. The principal cause of disability among farm families is often health-related. While farm families feel that farming is a good way of life, the experience of disability can add considerable strain on family life and relationships. Farm families experiencing disability recorded lower family farm incomes and lower participation in off-farm employment. Service and support provision was a concern for farm families experiencing disability. It is imperative the restrictive nature of disability is minimised for farm families. Improved service/support provisions for farm families reporting disability are required for this to be achieved.

Keywords: Disability, Quality of life, farm families Introduction Success in farming is a dynamic process, full of challenges and opportunities. The modern farmer, to be successful, needs to adapt swiftly and accurately to changes in the immediate and global environments (Nell and Napier, 2005). The development of a strategic plan to maximise the competitive strengths of the farm business provides focus for the farm operator, but cannot be perceived as an end-point in the management of the farm. Certain challenges and opportunities will present themselves to the farm operator and the farm family in the course of the farm family life-cycle. Some of these challenges will considerably alter the sustainability of an existing farm business plan and place the farm business in potential jeopardy in particular circumstances. The impact and experience of disability in the farm household is one such event.


240

Farm operators, despite romantic notions of operating within a slow-paced idyllic country setting (Gerrard, 1998) are also exposed to health and life-threatening circumstances (Zeida et al, 1993). While the farming industry claims a number of fatalities each year (Gerrard, 1998, Finnegan & Phelan, 2003), many other farm operators and farm family members are left experiencing a range of disability outcomes (Doyle, 1988, Strong & Maralani, 1999) following incidents or accidents occurring on the farm. Disability may have an impact across many facets of life (Hosain et al., 2002) including the psychological, physical and social aspects (Bishop, 2005). Self-identity may need to be re-defined following disability, as the outcome of disability in an individual case may dramatically challenge or alter an individual’s sense of self (Bishop, 2005). Roles, habits and routines may also be significantly altered following a disability event (Molyneaux-Smith et al., 2003). The impact of disability on a person is difficult to measure; however, Quality of Life, adequately defined, represents such a measure (Bishop, 2005).

Quality of Life is a term that has gained prominence in recent years, yet it has not been adequately defined (Bogue, 2004). The World Health Organisation (WHO) defined quality of life as an “individual’s perception of their position in life in the context of the culture and value systems in which they live and in relation to their goals, expectations, standards and concerns” (WHO, 1997). Bishop (2005) suggested that to measure quality of life adequately, one must understand not only the level of impact experienced in different areas of life, but the personal importance placed upon these life areas. A recent Irish study (Amarach, 2004) postulated four key factors that impact and influence quality of life. These factors, in order of importance were: 1) Happiness – as measured by how happy people say they are, 2) Family life – as measured by how satisfied people say they are with their family lives, 3) Health – how people reported their state of health, and 4) Finances – how prosperous people say they are financially. The Agricultural industry in Ireland is an important indigenous sector accounting for 9% of Gross National Product and employing 240,000 people (Scully, 2007). The sector reports the highest level of people with disabilities relative to any other occupational group (CSO, 2002), yet little is known on the key influential factors on the Quality of Life among farm families reporting disability. Disability may create a substantial impact upon an individual’s life, yet in the occupational area of agriculture and related fields, there has been relatively little research on farm-based disability.

Purpose The purpose of this paper is (a) to review current literature on the key influential factors as identified in the Amarach (2004) study, namely: 1) Happiness, 2) Family Life, 3) Health, and 4) Finances, and (b) to investigate the extent of any variation in the differences among farm families experiencing disability and those not experiencing disability. The definition of disability used in this article is that derived from the 2005 Disability Act of the Oireachtas (Irish Parliamentary system), which stated: “Disability, in relation to a person, means a substantial restriction in the capacity of the person to carry on a profession, business or occupation in the State or to participate in social or cultural life in the State by reason of an enduring physical, sensory, mental health or intellectual impairment” (Disability Act, 2005, p6)


241

Irish Agriculture The Irish agricultural sector is diverse in many respects, with farms possessing unique individual characteristics such as scale of enterprise, system of farming, farm mechanisation along with regional and demographic characteristics (McNamara & Reidy, 1992). Over 60% of the land area of Ireland is used for agricultural purposes, with 79% devoted to grass (3.4 million hectares), 11% to rough grazing (0.5 million hectares) and 10% to crop production (0.4 million hectares) (Dept. of Ag & Food, 2006). Beef production is the primary enterprise among Irish farms, followed by dairy, mixed grazing and sheep production respectively (Census of Agriculture, 2000). The majority of Irish farms are owner-occupied (Bogue, 2004; McNamara & Reidy, 1992) with a rich inter-generational transfer tradition. However the agricultural sector has experienced significant consolidation in recent years (Finnegan, 2007). The number of farms decreased by 17% between 1991 and 2000. The average size of farm increased by 21% in the same period of time. The sector also experienced a 17.5% reduction in the numbers of regular farm operatives (Crowley et al., 2004). Primary agriculture accounts for 5.7% of the labour in the total labour market (Dept. of Ag. & Food, 2006), yet the sector contributes approximately 30% of work related fatalities with children and elderly farmers (>65 years) primarily affected. A conservative estimate of 3,000 accidents is reported for the occurrence of farm accidents each year (Finnegan, 2007), frequently resulting in premature death or disability (Doyle, 1988).

Irish farm families reporting disability It is estimated that 8.8% of farm families experience a disability (CSO, 2002). The farm operator reported the highest incidence of disability among farm families, accounting for approximately 40% of farm family disability (McNamara et al., 2003). The incidence of disability and the number of disabilities experienced increased with age, as older farmers (>65yrs) experienced a higher incidence of disability than their younger counterparts (McNamara et al., 2003; CSO, 2002).

The most commonly reported type of disability among farm families were physical disabilities, with the primary cause being health-related. Arthritis and cardio-vascular problems were the main health-related causes among farm operators. Physical injury, 70% of which resulted from farm work, was another reported cause of disability among farm families. The remaining 30% of physical injury was attributed to non-farm causes, most notably vehicular and industrial accidents. The highest incidence of non-physical disability among farm families was recorded in children experiencing learning and intellectual disability (McNamara et al., 2003). The occurrence of disability was independent of the type of farming system in that study. However, where the farm operator experienced disability, a higher proportion of these farms had cattle rearing/other cattle systems as their principal system of farming compared to farms nationally (McNamara et al., 2003). There were proportionately fewer specialist dairy and sheep farm operators with disability than for all farms nationally and proportionately more in tillage farming in that study. Finnegan et al (2005) reported from case studies that the system of farming may be altered due to the presence of farm operator disability. In that study, it was not feasible for farm operators experiencing disability following injury to maintain the scale and system of farming in which they had previously engaged. It follows that the labour requirements and commitments in addition to the intensity of work activities in different farm enterprises may preclude many farm operators experiencing significant disabilities from fully participating in these enterprises unless significant and appropriate adjustments are made to their individual farm situations. Quality of Life

The issue of Quality of Life was reviewed in the current literature under each of the influential factors as identified in the Amarach (2004) study. These were identified above as Happiness, Family life, Health and Finances. In this paper each factor will be discussed separated.


242

On the issue of happiness, farming for many of those involved is far more than a set of tasks to be performed; it also forms part of identity (Molyneaux-Smith et al., 2003) and lifestyle (Finnegan, 2007). Bogue (2004) and Connolly et al (2007) found that farmers, generally, were happy with their farm situation in Ireland. The majority of farmers (98%) indicated that farming is a good way of life (Connolly et al., 2007), and Bogue (2004) reported that farmers considered their life to be relaxed or balanced. In spite of the changed nature of farming, where farmers must now cope with an increasing amount of regulation, being your own boss was identified as being important to farmers (Connolly et al, 2007) in 93% of cases. Molyneaux-Smith et al (2003) reported similar sentiments among farm operators with disabilities. On the issue of family life, the family home is an integral part of farm life. Up to three family generations often live on the same farm (Finnegan, 2007). Despite the long hours typically operated by farm operators, farm operators generally feel that they have adequate time to spend with their family and friends (Bogue, 2004). In the same study, younger farmers (<35years) were more likely to report adequate family time than older farmers. However, as farm size increased, farm operators tended to have less time for family and friends. In addition, farm operators who were engaged in off-farm employment were less likely to report adequate family time than those not engaged in off-farm employment (Bogue, 2004). Almost half of the farm operators in that study had an off-farm job, but they spent their vacation time from their off-farm employment position catching up upon farm work that required attention. The family offer the main source of assistance to farm operators experiencing disability in their efforts to remain farming (Molyneaux-Smith et al., 2003). However, the experience of disability can place a considerable strain on family life, with family members often having to make significant personal sacrifices to keep the farm functioning when the farm operator experiences disability (Finnegan et al., 2005). Family relationships can be affected as individuals with disability experience inherent frustration and impatience in their attempts to cope with disability. Particular strain can be placed on the spouse. The spouse may have to balance off-farm employment, work on the farm in addition to attending to the needs of the farm family member experiencing disability (McNamara et al., 2003; Finnegan et al., 2005). Farm operators experiencing disability, together with the family may “experience a shrinking social world due to the extra time required for occupational performance, appointments with health professionals, physical barriers, or societal attitudes” (Molyneaux-Smith et al., 2003)

On the issue of health it was suggested that farm operators generally take less care of their health when compared against other occupational groups (Hope et al., 1999). The physical nature of farm work, coupled with exposure to numerous hazards and stress caused from long working hours, isolation, and financial pressures may lead to restrictive health conditions among farm operators (McNamara & Reidy, 1997; Finnegan, 2007; Stepanyan & Blasoni, 2005; Zeida et al., 1993). The principal cause of disability among farm families, as indicated above, is often health related (McNamara et al., 2003). Seventy five percent of farm operators that experienced disability felt that farm related ill-health was the primary cause of their disability (McNamara & Reidy, 1997). In particular, respiratory problems were identified as being the dominant health-related disability, with back problems and allergies being frequently reported (McNamara & Reidy, 1992). The provision of adequate services and supports is important to provide stability and quality of life for rural inhabitants (Halseth & Ryser, 2006). Rural communities were reported as often being disadvantaged in terms of accessing health care services (Fitzgerald et al., 2001). The absence of professional support and community-care facilities continues to be an issue in rural Ireland, and this leads to personal anxiety and distress among farm families experiencing disability (Finnegan et al., 2005; McNamara et al., 2003). McNamara et al (2003) reported that almost 75% of farm families reporting disability did not avail of any State or voluntary services and on farms where the farm operator experienced disability, only 7.5% of them availed of any State or voluntary service. Lack of awareness on what services and supports are


243

available to farm families, coupled with the requisite assessment procedures for accessing certain services (Finnegan et al., 2005) may explain, to some extent, why the uptake of these services was so low. In Ireland, farm families experiencing disability must satisfy a means-test before they may be considered eligible for any form of financial disability payment. The means-test is a complex process, with virtually all sources of income taken into account in the eligibility assessment (Comhairle, 2005). Where an individual is married, or living with another person as husband and wife, the means of the spouse or partner are taken into account as well as the claimant’s own means (Dept of Social and Family Affairs, 2006). The means-test varies, depending on whether the claimant experiencing disability can or cannot continue farming. In the situation where the farm operator experiences disability, the value of the agricultural land is taken into account when calculating the means of the farm operator when the farm operator cannot continue farming. However, it is not taken into account in the means test where the farm operator continues to farm. The prospective income from the farm in the following 12-month period is used where the farm operator continues farming in assessing eligibility (Dept of Social and Family Affairs, 2006). While the above illustrates that ill-health is a contributory cause of disability, it does not infer that all farm operators, or farm family members, experiencing disability experience lower levels of health than the general farming population. As Susman (1994) stated “some disabled people are not sick”. On the issue of finances, farm operators were reported as being less willing to share information about their economic or financial situation to the wider agricultural community (Prospect Management Services, 2006) compared to physical farm information. In spite of the rapid economic growth experienced by the Irish economy during the mid 1990’s (Powell, 2003), “the buying power of farm incomes declined by 17% between 1995 and 2004” (CORI Justice, 2006). Direct payments (subsidies) presently constitute 94% of farm incomes (O’Donoghue, 2007) with farm families becoming every reliant on off-farm employment to generate income. Presently an estimated 52% of Irish farm families have off-farm employment (Connolly et al., 2004). These figures represent a dramatic change since Ireland entered the European Union, where farm income constituted 70% of the total household income. By 2000 however, this proportion had fallen to 41% (Finnegan, 2007). Farm families with disabilities experienced the same financial conditions and social challenges as other farm families. However, they have additional costs and challenges related to the requisite specialised tools and equipment, and the hiring of extra assistance in carrying out farm tasks (Molyneaux-Smith et al., 2003). These research workers observed that modification costs in excess of $100,000 have being reported in the United States and often with no financial assistance from State or other sources. McNamara et al (2003) and Finnegan et al (2005) reported financial losses in farm income when the farm operator experienced disability. McNamara et al (2003) quantified the reduction in income suffered by farm families at a rate of €24/ha of farm income compared to farms where disability was not experienced. Finnegan et al (2005) observed that the loss of spousal income was a common financial outcome of farm operator disability. The family farm income (FFI) was lower on farms where the impact was “major” by €5,098 compared to farms with no disability, or €3,678 compared to farms with “some disability” (McNamara et al., 2003). At present, off-farm employment makes an important contribution to farm household income as outlined above. McNamara et al (2003) reported that farm operators experiencing disability were two and a half time less likely to be involved in off-farm employment than farm operators not experiencing disability. The proportion of spouses with an off-farm job where the farm operator experiences disability was lower than on non-disability farms while the corresponding proportion in respect of farms where a person other than the farm operator experienced disability was higher (McNamara et al., 2003).

Conclusion Disability affects almost 10% of Irish farm families with the farm operator reporting the highest incidence among family members, accounting for almost 40% of reported cases. The primary cause of disability


244

among farm families is health-related, followed by personal injury. The review of the literature highlighted important issues experienced among farm families. The experience of disability has significant effects on various aspects of family life, with family members often having to make significant personal sacrifices to keep the farm functioning when the farm operator experiences disability (Finnegan et al., 2005). Relationships among family members can also be affected as individuals with disability experience feelings of frustration and impatience in their attempts to cope with disability. Farm families tend to experience a reduced social network following disability, due largely to social attitudes, physical barriers, and the increased time required to complete farm tasks. The lower participation levels in off-farm employment, coupled with reduced operations and considerable modification costs, may result in lower family farm incomes among farm families experiencing disability. The lack of professional advice to address specific needs of farm families experiencing disability and the assessment methods currently used to comply with State supports tend to exacerbate this issue. However despite the provisions for farm families experiencing disability, they report similar happiness sentiments as farm families where disability is not experienced. Ongoing research work at University College Dublin, and Teagasc Rural Economy Research Centre aims to identify the service and support requirements of farm operators experiencing disability and develop strategies for the implementation of such services. References Amarach. (2004). Quality of Life in Ireland – 2004 Report. Amarach Consulting.28pp Bishop, M. (2005). Quality of Life and psychosocial adaptation to chronic illness and acquired disability:

a conceptual and theoretical synthesis. Journal of Rehabilitation. 13pp. Retrieved 12/10/06 from http://www.findarticles.com/p/articles/mi_m0825/is_2_71/ai_n13820423/print

Bogue, P. (2004). The Quality of Life of Farm Families. Planning Post Fischler Programme Action

Research Project – Report 7. 66pp Census of Agriculture. (2000). Central Statistics Office, Dublin. Retrieved 14/03/07 from:

http://www.cso.ie/releasespublications/pr_agrifishpubshardcopies.htm Central Statistics Office (CSO). (2002). Census of Population of Ireland – Volume 10: Disability and

Carers. Central Statistics Office, Dublin. Retrieved 16/03/07 from: http://www.cso.ie/census/documents/vol10_entire.pdf

Comhairle. (2005). Entitlements for people with disabilities. Comhairle, Dublin. Connolly, L., Moran, B., & Cushion, M. (2007). A survey of farmers attitudes to farming as an

occupation. Paper presented at the Agricultural Research Forum, 12-13 March. CORI Justice (2006). Farm Incomes: more money less buying power. Policy Briefing. July 2006. Dublin.

ISSN: 1649-4954 Crowley, C. Meredith, D., & Walsh, J. (2004). Population and Agricultural Change in Rural Ireland,

1991-2002. Teagasc. Received 14/03/07 from http://www.teagasc.ie/publications/2004/2004030/paper03.html

Department of Agriculture and Food. (2006). Factsheet on Irish Agriculture. Economics and Planning

Division, Department of Agriculture and Food. Retrieved 21/03/06 from http://www.agriculture.gov.ie/publicat/factsheet/Feb2006.doc


245

Department of Social and Family Affairs. (2006). Assessment of Means. Department of Social and family Affairs. Retrieved 15/01/07 from http://www.welfare.ie/foi/meansassess.html

Disability Act (2005). As Enacted, Government of Ireland. Stationary Office, Dublin. Retrieved 14/03/07

from: http://www.oireachtas.ie/documents/bills28/acts/2005/a1405.pdf Doyle, Y. (1988). How Safe Are Irish Farms? A One Year Survey of Farming Accidents And Their

Medical Outcome Finnegan, A. (2007). An Examination of the Status of Health and Safety on Irish Farms. PhD thesis,

School of Biology and Environmental Sciences, University College Dublin.232pp Finnegan, A., & Phelan, J. (2003). A Survey of Health and Safety on Irish Farms – Implications for

Extension and Education. Paper presented at the 19th Annual conference of the Association for International Agricultural and Extension Education. Raleigh, North Carolina, USA. April 8-12th. Retrieved 06/08/06 from: http://www.aiaee.org/2003/Finnegan271-281.pdf

Finnegan, A., Ruane, D., & Phelan, J. (2005). A Study of the Impact of Farmer Disability on Farm

Households – Case Studies from the Republic of Ireland. Paper presented at the 17th European Seminar on Extension Education. Turkey, August 30th – September 3rd.

Fitzgerald, M., Pearson, A., & McCutcheon, H. (2001). Impact of Rural Living on the Experience of

Chronic Illness. Australian Journal of Rural Health. Vol. 9, 235 -240. Gerrard, C.E. (1998). Farmers’ occupational health: cause for concern, cause for action. Journal of

Advanced Nursing. 28 (1), 155-163. Halseth, G. & Ryser, L. (2006). Trends in service delivery: Examples from rural and small town Canade,

1998 to 2005. Journal of Rural and Community Development. 1, 69-90. McNamara, J. & Reidy, K. (1992). Survey Of Farm Safety And Health On Irish Farms. Teagasc

Publication, Dublin. 75pp. McNamara, J. & Reidy, K. (1997). Survey for Farm Safety and Health on Irish Farms. Teagasc, Health

and Safety Authority Publication, 61pp. McNamara, J., Ruane, D., Connolly, L., Reidy, K., & Good, A. (2003). A Study of the Impact of

Disability in Farm Households on the Farm Business in Ireland. Paper presented at the 19th Annual conference of the Association for International Agricultural and Extension Education. Raleigh, North Carolina, USA. April 8-12th. Retrieved 06/08/06 from: http://www.aiaee.org/2003/McNamara425-436.pdf

Molyneaux-Smith, L., Townsend, E., & Guernsey, J.R. (2003). Occupation Disrupted: Impacts,

Challenges, and Coping Strategies for Farmers with Disabilities. Journal of Occupational Science. 10 (1), 14-20.

Monawar Hosain, G.M., Atkinson, D., & Underwood, P. (2002). Impact of Disability on Quality of Life

of Rural Disabled People in Bangladesh. Journal of Health, Population and Nutrition. 20 (4), 297-305.

Nell, W.T. & Napier, R.J. (2005). Strategic Approach to Farming Success. Paper presented at the 15th

International Farm Management Association Congress, Campinas, Brazil. 14-19 August.


246

O’Donoghue, C. (2007). Presented at Teagasc National Rural Development Conference - Towards Sustainable Rural Livelihoods, Mullingar, 1st February.

Powell, B. (2003). Economic Freedom and Growth: The Case of the Celtic Tiger. The Cato Journal. Vol.

22. Retrieved 14/03/07 from: http://www.questia.com/PM.qst?a=o&se=gglsc&d=5009094163&er=deny

Prospect Management Services. (2006). Global Best Practice in Agricultural Benchmarking. Global

Management Services, North Yorkshire. 62pp. Scully, G. (2007). Presented at Teagasc National Rural Development Conference - Towards Sustainable

Rural Livelihoods, Mullingar, 1st February. Stepanyan, M. & Blasoni, B. (2005). Stress Among Farmers in Brittany (France): Myth or Reality?” – An

Exploratory Study. The National school of Public Health, Rennes, France. Retrieved 12/12/06 from: http://www.europhamili.org/protect/media/30.pdf

Strong, M.F. & Maralani, V.F. (1999) Farmworkers and disability: results of a national survey. Journal of

Vocational Rehabilitation. 12, 45-57. Susman, J. (1994). Disability, Stigma and Deviance. Social Science & Medicine. 38 (1), 15-22. World Health Organisation. (1997). WHOQOL – Measuring Quality of Life. Division of Mental Health

and Prevention of Substance Abuse, World Health Organization. Zejda. J.E., McDuffie, H.H., Dosman, J.A. (1993). Epidemiology of Health and Safety Risks in

Agriculture and Related Industries – Practical applications for Rural Physicians. The Western Journal of Medicine. 158 (1), 56-63.

IFMA 16 – Theme 3 Farm Management

247

CASE STUDY ANALYSIS OF THE BENEFITS OF GENETICALLY MODIFIED COTTON

William Back and Steve Beasley School of Natural & Rural Systems Management The University of Queensland (Gatton Campus),

Gatton QLD 4343, Australia [email protected]

Abstract Research in north and south New South Wales (NSW), Australia was conducted to assess the benefits of genetically modified (GM) cotton. Gross margins from 20,263 hectares on two properties for a three to five year period were analysed. A phone survey of cotton growers in the target districts was also used to determine grower opinions on benefits of cotton type. This also allowed for comparison between growing regions. Performance of cotton types is extremely variable, with no cotton type having a clear economic advantage. In years with high weed and/or Heliothis pressure, the financial return of GM cotton should be better than that of conventional cotton. However, findings indicate that GM cotton displays significant environmental and social benefits, due to reduction in chemical use and easier management. Although not as profitable, southern growers have adopted management practices to improve profitability and prefer GM Cotton because it is “easier to grow”. Keywords: genetically modified cotton, environmental benefits, social benefits

Introduction Transgenic cotton has been available to Australian growers for ten years. Ingard was initially released expressing resistance to Heliothis while Roundup Ready (RR) was later released with resistance to Glyphosate herbicide. This has impacted on the way cotton is grown, but it has raised questions regarding the benefits of genetically modified (GM) cotton. Prior research has lacked conclusiveness on the economic benefits of GM cotton. This paper investigates the benefits of GM cotton in terms of economic, environmental and social benefits on two irrigated properties in New South Wales – one in the northern growing district and the other in the southern growing district. The properties under examination are privately owned and specific data is confidential. However, generic findings, together with the results of a phone survey of farmer attitudes, will be provided. The properties experience a different climate and growing season, therefore the effect of climate on profitability is examined. Background Cotton is an important global crop which has uses ranging from its natural fibre to vegetable oil and animal feed. Currently Monsanto offers two biotech traits commercially in Australian cotton seed varieties, Bollgard II (BII) and Roundup Ready (RR). Cotton varieties incorporating these traits have the same lint and oil quality as conventional cotton lines. However, Monsanto claims they offer significant production and environmental benefits including fewer chemical applications, easier crop management and increased profitability (Monsanto Cotton 2006). RR cotton has a licensing fee of $51 per ha with only approved RR1 herbicide to be applied - BII currently carries a licensing fee of $300 per hectare while

1 Active ingredient- Glyphosate


248

the combined technology of Bollgard II/ Roundup Ready (BR) has a licensing fee of $351per ha (McDonald, J. 2006. pers com). In Australia, conventional cotton can require 10 to 11 insecticide sprays in a season to control Heliothis and other secondary pests. Also, because cotton is a broadleaf plant like many weeds, conventional growers have had to rely on residual herbicides, inter-row cultivation, chipping and careful fallow management to manage broadleaf weeds. Climate Comparison The properties in this research are situated in different regions of NSW. One is in the Gwydir region while the second is in the south-west Riverina region. Climatic data for Collarenebri and Hay from the Australian Bureau of Meteorology (BOM) is used to represent the growing areas. Collarenebri is situated on the western edge of the North West Slopes and Plains. Hay is located in the Riverina region of southern NSW. Collarenebri has a higher maximum temperature (3.2ºC on average) than Hay (Figure 1). It is 3.5 to 4.5ºC hotter during the cooler months. Figure 4: Mean Daily Maximum Temperature (BOM 2004a&b)

Mean daily minimum temperatures (Figure 2) indicate that Collarenebri has more variability than Hay. Collarenebri experiences minimum temperatures similar to Hay in the winter, but for a shorter period. In summer Collarenebri maintains minimum temperatures approximately 4ºC warmer.

Mean Daily maximum Temperature

0 5

10 15 20 25 30 35 40

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month

Tem

per

atu

re (

deg

C)

Hay Collarenebri


249

Figure 5: Mean Daily Minimum Temperature (BOM 2004a&b)

Figure 3 shows that Hay experiences a longer, cooler winter, whilst Collarenebri experiences more days over 30ºC. Figure 6: Mean No. of Days where Max Temp >=30 deg C (BOM 2004a&b)

Collarenebri experiences more summer dominant rainfall, with the majority falling between the months of November and March (Figure 4). Hay experiences its highest rainfall in the winter months; but its rainfall is more uniform throughout the year.

Mean Daily Minimum Temperature

0

5

10

15

20

25


Tem

per

atu

re (

deg

C)

Hay Collarenebri

Mean no. of days where max Temp>=30 deg C

0 5

10 15 20 25 30


tem

per

atu

re (

deg

C)

Hay Collarenebri


250

Figure 7: Mean Monthly Rainfall (BOM 2004a&b)

Methodology Gross Margins Gross Margins were calculated for the cotton crops (20,263 hectares) on the 2 properties over 3 year (south) and 5 year (north) periods. A total of 173 gross margins were generated (142 – north, 31 - south). The average paddock size on the northern property was 130 hectare compared to 60 hectare. These gross margins were also compared with industry standards from the ‘Australian Cotton Comparative Analysis’ for each relevant year. The aim was to analyse gross margins from the previous five growing seasons for each property across Conventional, Bollgard II, Roundup Ready and stacked gene Bollgard II/Roundup Ready. Gross margins for the northern property were able to be obtained from the years 05/06, 02/03, 01/02, 00/01, and 99/00 making a total of five years data. Three years of data were able to be collated for the southern property. These were, 05/06, 04/05, and 03/04. In addition to the four cotton types, Ingard2 and Ingard/ Roundup Ready were also present from earlier years. Telephone Survey To understand social and environmental benefits of GM cotton, the opinions of farmers in surrounding districts were researched through interviewing using a phone survey. Surveys from within the Riverina region were classed as southern, while all surveys north of the Riverina were classed as northern. The survey was written to be short, concise; and to be completed in a maximum of five minutes at the very most. It is recognised that the public is increasingly not accepting of unfamiliar phone calls. A short sharp survey was thought to be the best way of securing time from busy growers. This method was used to allow for analysis between northern and southern NSW, in an effort to understand potential differences in the two growing areas. Ethical clearance for the survey was obtained from the School of Natural & Rural Systems Management’s Ethics Committee. Participants were asked if they wanted a summary of the results. Because of time constraints and privacy laws, cotton grower contact details had to be sourced from the Yellow Pages Online. Phone numbers were found by searching for cotton in different regions in NSW. The inclusion of farmer in any searches immediately listed all farmers in that area. There was a very

2 Predecessor to Bollgard II, carries only one Bt gene.

Mean Monthly Rainfall

0 10 20 30 40 50 60 70 80


Ra

infa

ll (

mm

)

Hay Collarenebri


251

limited number of cotton grower numbers able to be found, as a small number of growers listed themselves under cotton. Cotton is a new crop in southern NSW which made it very difficult to secure any phone numbers in that region. The Yellow Pages returned no property phone numbers searching for cotton, so general farmers had to be called in order to find cotton growers. This was extremely time-consuming. For this reason only eleven surveys were able to be completed for that region compared to twenty-six in northern NSW. Because of the lack of phone numbers every contact was phoned on the list. The survey was conducted between May and September 2006. The interviewer contacted growers between 11:30 am and 2:00pm during weekdays to try and reach growers when they were home for lunch. He also contacted growers between 7:30pm and 8:45pm in the evenings, starting just after ABC news and weather. The evening timeslot had the most success, with more growers home at this time, and willing to part with five minutes of their time. Results Gross Margins

Northern Property 142 gross margins totalling 18,398 hectares over 5 growing seasons were analysed. Within year comparisons of gross margins found that: 1. in the 99-00 season, Ingard had a significantly higher gross margin than conventional cotton; and 2. in the 00-01 season both Ingard and Ingrad/Roundup Ready had significantly higher gross margins than conventional cotton. Table 1 shows the significant differences in gross margins in each year reviewed. Figures in brackets are the number of gross margins. Table 1. Differences between Gross Margins – Northern Property

Year Significant Difference No Significant Difference

99-00 I/RR(1), Conv(22) and Ingard (11) 00-01 Ingard (18) > Conv (24) I/RR (1) and Conv (24)

I/RR (1) & Ingard (18) 01-02 Ingard(9) & I/RR(3) > Conv (21) I/RR (3), Ingard (9) & RR (8)

Conv (21) & RR (8) 02-03 BG(1), Conv(3), I/RR(2) & RR(3) 05-06 BG (2), BR (6), Conv (5) & RR(2)

More detailed statistical work using the combined data showed that planting configuration and planting date had no significant effect. A General Linear Model using type, year and area found that gross margins increased by 0.1375% per hectare for each extra ha in field area. Further testing showed however, that area had no significant effect on costs. A mean area was established for all fields of 130 hectare to develop the final model adjusted to take into account all effects. Hence, means for cotton type were adjusted for differences between years and the fact that not all types were grown every year. Bacillus thuringiensis (Bt) gene cotton performed the best with BII, I/RR and Ingard showing the three highest returns, conventional cotton had the second lowest return with BR lowest. However the model only explained 53% of the variance in the data and the only significant finding was that Ingard had a


252

higher gross margin than conventional cotton. (Ingard is no longer available being replaced by BG and BII.) Southern Property 31 gross margins totalling 1,865 hectares over 3 growing seasons were analysed. Within year comparisons of gross margins found that in the 03-04 season, Ingard/Roundup Ready had a significantly higher gross margin than Roundup Ready cotton. Table 2 shows the significant differences in gross margins in each year reviewed. Figures in brackets are the number of gross margins. Table 2. Differences between Gross Margins – Southern Property

Year Significant Differences No Significant Differences

03-04 I/RR (3) > RR (6) BR (1) & RR (6) BR (1) & I/RR (3)

04-05 BR (6), Conv (4), & RR (4) 05-06 BR (3), Conv (2), & RR (2)

More detailed statistical work using the combined data, found the only significant difference was between the gross margin for IRR and all other cotton types. IRR is no longer available having been superseded by BR. Industry Comparisons Gross margins were averaged for each year to compare with industry standards. The industry gross margins steadily increased throughout the seven years of data (Figure 5). Gross margins for the two properties examined were highly variable. Northern property gross margins were on a downward trend, but then produced an excellent return in 05/06. Southern property gross margins are shown to be improving. Industry margins outperformed the northern property in three of the five years and the southern property in all three years. Figure 5: Gross Margin comparison between Industry and Case Study Properties (Industry data from Doyle et al. 2005a)

Industry Comparison

-500

0

500

1000

1500

2000

2500

3000

99/00 00/01 01/02 02/03 03/04 04/05 05/06

Year

Gro

ss M

arg

in (

$/h

a)

Industry North South


253

Telephone Survey A total of 37 surveys were completed. In order to achieve this, approximately 148 phone numbers were used and 255 phone calls made. 25 surveys were conducted on growers who grew GM cotton in northern NSW and 10 in southern NSW. Two respondents did not grow GM cotton. One “used to, but because of health problems and no financial benefit I changed to organic”. The other grew conventional because “couldn't get hold of any GM cotton in that year”. Cotton Type Comparison Cotton growers were asked the percentage they grow of each cotton type. For northern NSW it is shown in Figure 6 that the majority of farmers grew only one or two different types of cotton. Stacked gene BR cotton was found to be most predominant, with all but one grower planting a portion of his farm to the cotton type. The graph also shows that 8 out of the 25 respondents grew 100% BR. Ten of the growers planted BII cotton and 11 planted conventional cottons. Ten growers planted RR cotton; but it was less than 20% of the farms’ planted area. Figure 6: Percentage of Cotton Type grown in Northern NSW

Cotton Type Grow- North NSW

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1 3 5 7 9 11 13 15 17 19 21 23 25

Respondent no. (1-25)

Perc

en

tag

e (

%)

Bollgard II/ Roundup Ready

Roundup Ready

Bollgard II

Conventional

Figure 7 shows the percentage of each cotton type grown through northern NSW. On average BR cotton represents 63% of the cotton planted. This is significantly higher than the remaining three cotton types which make up between 7% and 16% of the cotton planted. On average, 84% of the cotton grown on farms in northern NSW is GM.


254

Figure 7: Percentage Grown of Cotton Type in Northern NSW- Average

Cotton Type North NSW- Average

0%

20%

40%

60%

80%

100%

Cotton Type

Perc

en

tag

e (

%)

Conventional

BII

RR

BR

Southern findings followed the same trend as the north, with all but one grower planting BR cotton. BII, RR and conventional cotton are shown in Figure 8 to represent a small minority of the cotton planted. Half the growers interviewed planted their whole farm to BR cotton. Figure 8: Cotton Types Grown in Southern NSW

Cotton Type Grown- Southern NSW

0%

20%

40%

60%

80%

100%

1 2 3 4 5 6 7 8 9 10

Respondent no. (1-10)

Perc

en

tag

e (

%)

BR

RR

BII

Conventional

On average 80% of the cotton planted on properties in southern NSW is BR. The remaining three cotton types individually make up 10% or less of the cotton area planted (Figure 9).


255

Figure 9: Cotton Type Grown in Southern NSW- Average

Cotton Type Grown South NSW- Average

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Cotton Type

Perc

en

tag

e (

%)

Conventional

BII

RR

BR

When the findings for north and south NSW are compared, it is evident (Figure 10) that both are similar. However, southern growers plant nearly 20% more BR cotton. Figure 10: Cotton Type Comparison of North and South NSW

Cotton Type Comparison

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

Conventional BII RR BR

Cotton Type

Perc

enta

ge (%

)

Southern NSW

Northern NSW

Financial Return As the majority of growers only grew one or two cotton types, there was not a good indication of the profitability of all four cotton types. As shown in Figure 11, BR was believed to have the best financial return by 11 of the 25 respondents in the northern district. Interestingly, 6 of the respondents believed conventional cotton provided the best financial returns. Four of the 25 interviewed commented that financial performance depended on the field the cotton was grown in and the season.


256

Figure 11: Cotton Type with Best Financial Return- Northern NSW

Best Financial Return

0

5

10

15

20

25

Conv

entiona

lBII

RR

BR

unde

cided

Cotton Type

Nu

mb

er

of

Resp

on

dan

ts

Northern NSW

Only half of the cotton farmers interviewed were able to indicate the second best performing cotton type. The second best performing cotton types were believed to be BII and BR (Figure 12). Figure 12: Cotton Type with Second Best Financial Return- North NSW

Second Best Financial Return

0

2

4

6

8

10

12

Conventional BII RR BR

Cotton Type

Nu

mb

er

of

Resp

on

dan

ts

Northern NSW

Many southern farmers were unable to comment on the most profitable type as they were new to cotton growing. BR was still selected as most profitable by those able to comment (Figure 13).


257

Figure 13: Cotton Type with Best Financial Return- South NSW

Best Financial Return

0

2

4

6

8

10

Convent

iona

lBII

RR

BR

unde

cided

Cotton Type

Nu

mb

er

of

Resp

on

dan

ts

Southern NSW

Social Benefits In both the north and south regions, the majority of respondents said social benefits influenced their selection of cotton type (Figure 14). Figure 14: Do GM Social Benefits Influence Cotton Type Selection

Do GM Cotton Social Benefits Influence Selection?

0

5

10

15

20

25

Yes No Yes/No

Response

No

. R

esp

on

den

ts

Southern NSW

Northern NSW

Northern farmers presented nine common social or managerial benefits: Less Labour Less Spraying/ Chemical Easier Management Timeliness of Applications Better Public Perception Sensitive Areas Management Better Lifestyle Less Stress/ Pressure No Benefits


258

The most common response was that GM cotton did allow for easier management (Figures 15 & 16). Less spraying/chemical use and better timelines in making farming applications were the next most reported benefits. One grower reported that GM cotton ‘takes the pressure off for timeliness of applications and the amount of applications’. One fifth of those interviewed believed less labour, a better lifestyle and less stress or pressure were major benefits of the technology. Another grower commented that it is ‘time saving, easy to manage and takes out the guesswork’. Figure 15: Perceived Social Benefits of GM Cotton- North NSW

Perceived Social Benefits of GM Cotton- North NSW

0

5

10

15

20

25

Percieved Benefit

No

. R

esp

on

dents

(0-2

5)

Less Labour

Less Spraying/ Chemical

Easier Management

Timeliness of Applications

Better Public Perception

Sensitive Areas Management

Better Lifestyle

Less Stress/ Pressure

No Benefits

A small percentage of respondents believed GM cotton provided good sensitive areas management and better public perceptions of cotton. This benefit was best summed up by one grower who remarked ‘we are the first farm from town, and can only spray with certain wind directions. There is a better perception from people in town; not so many planes flying around’. Figure 16: Perceived Social Benefits of GM Cotton- North NSW

Perceived Social Benefits of GM Cotton- North NSW

Less Labour


Easier Management


Better Public Perception


Better Lifestyle

Less Stress/ Pressure

No Benefits

Southern growers reported easier management and less spraying or chemical use (Figure 17). Interestingly, one fifth of those interviewed said they wouldn’t grow cotton if they didn’t have access to the GM traits. One farmer said ‘cotton used to be hard to grow but GM has made it easier, hence its adoption down here’.


259

Figure 17: Perceived Social Benefits of GM Cotton- South NSW

Perceived Social Benefits of GM Cotton- South NSW

0

2

4

6

8

10

Percieved Benefit

No

. R

es

po

nd

en

ts (

0-1

0)


Easier Management


Public Perception

Wouldn't grow cotton w ithout

GM traits

Figure 18 shows that the public perception of GM cotton is a significant benefit perceived by southern growers. Figure 18: Perceived Social Benefits of GM Cotton- South NSW

Perceived Social Benefits of GM Cotton- South

NSW


Easier Management


Public Perception

Wouldn't grow cotton

w ithout GM traits

Environmental Benefits When growers were asked if the environmental benefits of GM cotton influenced their selection of cotton types, the responses across both growing regions was predominately yes (Figure 19).


260

Figure 19: Do GM Cotton Environmental Benefits Influence Selection of Cotton Type?

Do GM Cotton Environmental Benefits Influence

your Selection of Cotton Type?

0

5

10

15

20

25

Yes No Undecided

Response

No

. R

esp

on

ses

Southern NSW

Northern NSW

Less chemical and/or spraying was the biggest environmental benefit perceived as indicated by 23 of the 25 surveyed (Figures 20 & 21). One farmer reported “‘Smart Rivers’ monitoring has proven it (less chemical)”. Sensitive areas management was also a significant benefit to 10 of the respondents. Six people believed less aerial spraying or drift to be a major benefit. Only one person saw no benefit. Figure 20: Perceived Environmental Benefits of GM Cotton- North NSW

Perceived Environmental Benefits of GM Cotton- North NSW

0

5

10

15

20

25

Percieved Benefit

No

. R

es

po

nd

en

ts (

0-2

5)

Less Chemical/ Spraying

Less Aerial Spraying/ Drift

Less load on environment


Increased Biodiversity

No Benefits


261

Figure 21: Perceived Environmental Benefits of GM Cotton- North NSW

Perceived Environmental Benefits of GM Cotton-

North NSW

Less Chemical/ Spraying


Less load on environment



No Benefits

For southern NSW, the results on environmental benefits were similar, but not the same. Less chemical was the most common response reported by 70% of respondents (Figures 22 & 23). Sensitive areas management was a big benefit, reported by 50% of respondents. One grower said, ‘we are near rivers, so you don’t have to worry about planes’. Once again, increased biodiversity and no benefit were only reported by one grower each. Less aerial spraying or drift was a significant benefit reported by 30% of growers. Figure 22: Perceived Environmental Benefits of GM Cotton- South NSW

Perceived Environmental Benefits of GM Cotton- South NSW

0

2

4

6

8

10

Percieved Benefit

No

. R

es

po

nd

en

ts (

0-1

0) Less Spraying/ Chemical


Less Load on Environment



No Benefits


262

Figure 23: Perceived Environmental Benefits of GM Cotton- South NSW

Perceived Environmental Benefits of GM Cotton- South NSW

Reduced Chemical

Less Aerial Spraying/ Drif t

Less Load on Environment



No Benefits

Discussion Survey results showed almost 95% of respondents grow GM cotton. Huesing & English (2004) highlighted the significant amount of farming country planted annually to GM crops (more than 67 million hectares (ha) in 18 countries worldwide); growing at 10% per year. Huesing & English (2004) state that “Genetically modified crops are most often associated with high-input industrial economies, but farmers in the developing world are rapidly adopting them. Surprisingly, nearly one third of all GM crop hectares are now grown in developing nations.” Survey results indicate that almost all growers believe there is a place in their farming system for GM cotton. As southern NSW is an emerging cotton growing area, the phone survey was to a degree limited by the availability of growers to interview. Contacting sufficient growers was extremely difficult, compounded by privacy laws preventing the sourcing of contacts. Survey results revealed that GM cotton, specifically BR, has been adopted at an incredible rate. Growers in both districts plant the majority of their country to BR. Only two respondents who grew transgenic cotton were found not to grow BR, with over 37% devoting their entire farm to BR. Growers evidently believe GM cotton offers an alternative with more benefits than conventional cotton. Economic Benefits

It is difficult to draw conclusions on the economic performance of the cotton types. The return on both properties was shown to be extremely variable. The season experienced had a big impact on the performance of each cotton type. The results indicated the importance of the findings by Marra et al. (1998) and Bryant et al. (1999a) that there is a required level of pest infestation before Bollgard II technology becomes profitable, due to the high licence cost of GM technology which must be paid regardless of insect pressure. Greater yields can be contributed to continual protection from Heliothis by GM cotton, as opposed to spraying when threshold levels are reached in conventional cotton. Most growers interviewed believed that BR had the best financial return. This was not shown from the case study nor from the literature; however Doyle et al. (2005) found of those surveyed, 66% grew BII for economic benefit in comparison to conventional cotton.


263

Of the northern growers interviewed, just under a quarter of respondents believed conventional cotton to posses the best return. This is similar to the findings of Bryant (1999b), who found that in years of lower insect pressure, conventional cotton was able to return the highest profit as it did not have to cover licence fee costs. It is far from definitive which cotton type is most profitable. However, it is evident that there needs to be a level of weed and/or insect pressure before GM cotton will prove profitable in comparison to conventional. Studies by Marra et al. (1998) showed transgenic cotton had no overall benefit as the savings and increased revenues did not outweigh the higher seed and technology costs. Klotz-Ingram et al. (2001) concluded that herbicide tolerant and insecticide-resistant crops do require a certain level of infestation to break- even. In years of high level infestation, there were several reported benefits of GM cotton which reduce spending. Less labour, easier management, less spraying/ chemical, better timeliness of applications and improved sensitive area management are all factors which contribute to cost savings. From the literature and survey findings, it could be concluded that the profitability of each cotton type is heavily related to the specific weed and Heliothis pressures faced in the growing season. Financial return is in no way fixed for the type of cotton grown, but dependent on the variables experienced during the growing season. Social Benefits Social benefits reported were consistent with the literature. 74% of growers indicated that social benefits offered by GM cotton influenced selection of cotton type. The best social benefit was simply easier management. Less spraying/ chemical and a better timeliness of applications all indicate that GM cotton is easier to manage. These findings were consistent with those of Fernandez-Cornejo et al. (2000) who reported increased flexibility and a reduction in the number of operations. Social benefits of GM cotton indicated that RR and BII cotton can be a very effective management tool. RR can be planted in paddocks known to have a heavy weed infestation allowing for greater control of weed problems. This reduces management and can minimise costs like chipping. RR is currently limited by its short application window (Doyle 2005b), but the release of Roundup Ready Flex is expected to further reduce reliance on residual herbicides. Due to its easier management, GM cotton would be expected to reduce the difficulty of growing cotton in remote parts, or small or irregular paddocks. Perhaps the biggest social benefit of transgenic cotton is its use in managing sensitive areas. This is consistent with literature with Edge et al. (2001) indicating the ability to grow cotton near more heavily populated areas because of a reduction in the reliance on insecticide. GM cotton allows growers to manage areas close to rivers, houses, livestock, neighbours, towns and highways better. One grower indicated that as they were close to town and their neighbours had cattle, there was less time lost waiting for the right wind directions. It meant for one grower they could plant cotton in areas which before were too environmentally sensitive. These benefits reduce the time and frustration required to grow cotton in sensitive areas. Environmental Benefits From this project, it became apparent that the perceived environmental benefits of GM cotton are undeniable. The reduction in chemical or sprays is by far the greatest environmental benefit of GM


264

cotton. Growers are concerned with the amount of chemical required to grow conventional cotton, with nearly 90% saying it influenced their decision when selecting cotton types. Research indicated that BII has an environmental impact value of 23% of that of conventional cotton (Knox et al. 2006). Survey findings confirmed a significant drop in total volume of residual herbicide and insecticide. Insecticide use on average was shown to be reduced by 50-80%. A considerable number of growers believe GM cotton has reduced the load on the environment. Climatic Considerations The most limiting climatic factor of the southern property is the shorter cotton growing season it experiences. Southern NSW has a hot summer, but temperatures reduce considerably after February. Northern growing regions maintain summer temperatures, especially daily minimum temperatures, with a less noticeable cut-out in temperature in early autumn. Collarenebri was shown to receive similar rainfall to Hay in the autumn cotton picking season, however Hay’s dominant winter rainfall in a normal season would be expected to have a greater chance of affecting cotton harvest. Cotton growing techniques in southern areas have adapted to improve production despite the colder climate. However, southern regions such as the Riverina are still not as profitable as northern NSW. Narrow row cotton is shown from the literature to lead to earlier maturity3 and reduce the chance of rain setting in during picking. BII assists with the drive for early maturity with higher retention rates and less tipping out. Narrow row cotton’s higher plant population and increased ability to utilise available sunlight was found to lift yields, and ultimately the profitability of southern farming systems. (Millyard, 2003; McDonald, 2004; Barber, 2005) The different climate and cropping systems meant southern growers experience different Heliothis pressure. Adapting IPM strategies and making use of available technologies all make management of Heliothis in southern growing areas practicable and viable (Lawrence 2004). By adapting to the different climatic conditions and developing management practices suitable for southern NSW, it was apparent that growers can improve profitability. Fibre quality issues were reported by Millyard (2003) to have plagued the southern growing area; however changing harvest methods from stripping to picking was reported to have overcome the problem. Findings by McDonald (2004), that many rice growers are considering growing cotton were reinforced by the survey. Growers reported that they were looking for alternatives to rice that used less water. This, along with the reduced workload of GM cotton and potential for returns similar to rice, is the driving factor behind the adoption of cotton in southern areas. Conclusions Neither the case study nor the literature showed consistently superior economic returns from GM cotton. However the survey found that BR was perceived as having the best economic return. As well as the perceived economic benefit, GM cotton provides social and environmental benefits leading to easier management and the ability to successfully grow cotton in more closely settled and environmentally sensitive areas. For these reasons, GM cotton has been widely adopted especially in the southern district where farms are smaller and more closely settled.

3 Up to 21 days earlier


265

References Australian Bureau of Meteorology, 2004a, ‘Averages for Hay (Miller Street)’, Climate Averages for

Australian Sites, medium- government website, viewed 3 April, http://www.bom.gov.au/climate/averages/tables/cw_075031.shtml

Australian Bureau of Meteorology, 2004b, ‘Averages for Collarenebri (Albert Street)’, Climate Averages

for Australian Sites, medium- government website, viewed 3 April, <http://www.bom.gov.au/climate/averages/tables/cw_048031.shtml>.

Barber, J 2005, ‘15-inch cotton: Why, where and hows it going? The Australian Cottongrower, vol. 25,

no. 1, pp. 25-26. Bennett, R Ismael, Y Morse, S and Shankar, B 2004, ‘Reductions in insecticide use from adoption of Bt

cotton in South Africa: impacts on economic performance and toxic load to the environment’, Journal of Agricultural Science, vol. 2004, no. 142, pp. 665-674.

Bryant, K Allen, C Kharboutli, M Smith, K Bourland, F Earnest, L 1999a, Cost and Return Comparisons

of Transgenic and Conventional Cotton Systems in Arkansas, AAES Special Report, vol. 198, pp. 172-175.

Bryant, K Robertson, W Lorenz, G Allen, C Bourland, F Earnest, L 1999b, Economic Evaluation of

Transgenic Cotton Systems in Arkansas, AAES Special Report, vol 198, pp. 38-43. Doyle, B 2005(b), ‘Weed Management and Roundup Ready Cotton 2005’, report prepared by Cotton

Consultants Australia for Cotton Catchment Communities CRC and The Cotton Research and Development Corporation , IRF Cotton Research, University of New England.

Doyle, B Reeve, I Coleman, M, December 2005(a), ‘The Cotton Consultants Australia 2005 Bollgard II

Comparison Report’, report prepared by Cotton Consultants Australia for Cotton Catchment Communities CRC, IRF Cotton Research, University of New England.

Edge, J Benedict, J Carroll, J and Reding, K 2001, ‘Contemporary Issues- Bollgard II Cotton: An

Assessment of Global Economic, Environmental, and Social Benefits’, The Journal of Cotton Science, vol. 5, issue 2, pp.131-136, viewed- 17 May, <http://journal.cotton.org>.

Fernandez-Cornejo, Jorge and McBride with contributions from Klotz-Ingram-Ingram, C Jans S and

Brooks, N April 2000, ‘Genetically Engineered Crops for Pest Management in U.S Agriculture: Farm-Level Effects’, U.S Department of Agriculture, AER-786, Economic Research Service.

Huesing, J English, L 2004, ‘The impact of Bt crops on the developing world’, AgBioForum, vol. 7, no.

1&2, pp.84-95. Klotz-Ingram, C Jans, S Fernandez-Cornejo, J McBride, W 2001, ‘Farm-level production effects related

to the adoption of genetically modified cotton for pest management’, Journal of Agrobiotechnology Management and Economics, Vol. 2, no. 2, Article 3, viewed- 12 April, <http://www.agbioforum.org>.

Knox, O Constable, G Pyke, B Gupta, V 2006, ‘Environmental impact of conventional and Bt insecticidal

cotton expressing one and two Cry genes in Australia’, Australian Journal of Agricultural Research, vol. 2006, no. 57, pp. 501-509.


266

Lawrence, L 2004, ‘Managing Helicoverpa in southern NSW crops’, The Australian Cottongrower, vol. 25, no. 4, pp. 17-19.

Marra, M Carlson, G Hubbell, B 1998, Economic ‘impacts of the First Crop Biotechnologies’, North

Carolina State University, Department of Agriculture and Resource Economics, viewed- 11 May, <http://www.ag-econ.ncsu.edu/faculty/marra/FirstCrop/sld001.htm>.

McDonald, C 2004, ‘Germinating Ideas’, The Australian Cottongrower, vol. 25, no. 2, pp. 79-80. Millyard, J 2003, ‘Narrow row cotton gives third option in the south’, The Australian Cotton Grower,

Vol. 24, no. 3, pp. 8. Monsanto Cotton, 2006, Monsanto cotton, medium- Product website, viewed 6 June 2006,

<www.Monsanto.com.au>.


267

CUSTOMIZED COMMODITY DERIVATIVES; AN ALTERNATIVE TO REDUCE FINANCIAL RISKS ON FARMS.

W.H.M. Baltussen Agricultural Economics Research Insititute (LEI B.V.)

P.O. Box 29703 2502 LS, Den Haag, Wageningen University and Research Centre, The Netherlands [email protected]

M.A.P.M. van Asseldonk

Institute for Risk Management in Agriculture (IRMA), Wageningen University, The Netherlands

K. Horsager Compass Strategic Investments, LLC, USA

Abstract Customised derivatives are developed for the agriculture industry to decrease the volatility of input or output prices. These derivatives can be attractive for agriculture producers because a substantial part of the business risk in agriculture is caused by fluctuating commodity input and output prices. The aim of the paper is to provide information on customised derivatives, their background and contemporary applications for natural gas procurement in the Dutch horticulture sector. To research the added value of customized commodity derivatives (a maximum price contract and a collar contract) a simulation model is developed. With this model mean and variation of the natural gas costs are calculated and compared with buying on the spot market and a fixed price contract. Our findings show that the use of the maximum price contract in the period 2000-2005 in the Netherlands helps producers to decrease expected costs and lower the variability of natural gas prices as well. Keywords: customized commodity derivatives, natural gas, horticulture, financial risks Introduction Risks can be classified into various types of risk. For most agricultural risks the classification of Hardaker et al. (2004) can be used, who distinguish first between business risks and financial risks. Business risks include production risks, price risks, personal or human risks and institutional risk. The other type of risk, financial risk, refers to the risks related to the way a farm is financed. Price risks are expected to become more important for European agricultural farms because of changing WTO, EU regulation and national policies. Because of the protection of the European market many producers are not used to coping with these price risks. In the Netherlands, new products are developed by financial institutions (e.g. Rabobank and ABN-AMRO) to decrease the price risks of certain costs. One of these products is customized commodity derivatives. The aim of this article is to give insight into the value that customized commodity derivatives have for agricultural producers. In section 2 a short explanation is given about derivatives. In section 3 a model is introduced to compare different alternatives to reduce price risks of natural gas for glasshouses in the Netherlands. In section 4 the obtained results for the period 2000-2005 for a Dutch flower greenhouse are given. In section 5 a list


268

of conditions is offered which ought to be present for the development of additional customised derivatives. Also some suggestions for further research are listed. In section 6 the conclusions are listed. Customized derivatives A derivative is an instrument whose characteristics and value depends upon the characteristics and value of an underlying instrument. Derivatives are generally designed to manage or hedge price risk, or to swap cash flows (Hull, 2002). Some derivatives are standardised and traded on regulated exchanges while others are customised bilateral agreements. An exchange traded derivative is a derivative that is traded on an organised and regulated exchange. These derivatives are most commonly standardised by quantity, grade, delivery location and expiration date. Exchange traded derivatives are typically cleared and settled through the multilateral clearing process used by regulated exchanges. An example of an exchange traded derivative is a futures contract on natural gas. Exchange traded natural gas contracts are actively traded on organised exchanges such as the New York Mercantile Exchange in New York or the Intercontinental Exchange which is electronic and offered for trading globally. Options traded on such exchanges which settle directly to cash or to a futures contract are also considered exchange traded derivatives. A customised derivative is one that is designed specifically for a user or a group of users or for a specific application. Customised derivatives are also known as over the counter ('OTC') derivatives. These instruments generally contain some unique characteristics such as the size, grade, delivery location or settlement benchmark, expiration date, pay-off matrix or counter party characteristics. Customised derivatives are generally bilateral and can not be directly offset against other customised or exchange traded derivatives. An example of a customised derivative would be a contract offered by a gas company to purchase natural gas at a fixed price for the next year designed to meet a customer's requirements. There are numerous other examples of customised derivatives where the supplier customises some aspect of a standardised contract to meet his customer's needs. Model and data A stochastic simulation model is developed to compare different strategies for decreasing the volatility of gas prices for greenhouses in the Netherlands. Four price strategies are compared (see also figure 1): Variable price contract. A variable price contract is a contract where two parties have a relationship to supply and receive a good, but the price floats until the customer takes delivery of the physical commodity. For this case study the Dutch situation, where prices are established each quarter is simulated; Fixed price contract. A fixed price contract is a contract where the buyer and seller fix the price of a good for a certain period. For this case study the price is fixed for the upcoming calendar year (it is assumed that the market is neither in contango nor in backwardisation);


269

Maximum price contract. A maximum price contract is a contract where the supplier (or a third party) agrees to charge no more than a predetermined price for a commodity. In this case the buyer will pay a premium up front for the maximum price contract. The maximum price is known as the strike price. If the market price is below the strike price, the greenhouse operator is charged the prevailing market price. However, if the market price is above the strike price, the supplier will only charge the contract maximum price; Collar contract. A collar (sometimes also called a fence) contract establishes both a maximum price and a minimum price between the supplier and the buyer. If the market price is between the minimum and the maximum prices the user pays the prevailing market price. If the market price is above the maximum price, the user pays only the contract maximum price. If the market price is below the minimum price, the user pays the agreed upon contract minimum price. Figure 1: Natural gas contract comparison (prices in eurocents per m3).

For pricing the derivatives (maximum price contract and collar price contract) Black’s Option Pricing Model for Futures and Forwards is used (Black 1976, see appendix 1). The maximum price contract utilises an at the money call option. This means that on October 1 of each year, which is the time the derivative is agreed upon by both parties, the average price of the previous four quarters (a proxy for the forward price) and the contract maximum price are the same. The collar contract strategy uses the purchase of an out of the money call option and the sale of an out of the money put option. In this analysis the collar has a net capital outlay of zero. The purpose of this method of structuring a collar is to create a realistic collar price contract where the derivative premiums are approximately the same for the call option and the put option so that initial capital outlay is approximately zero (premium for the collar is zero).

0

5

10

15

20

25

30

35

40

45

2 6 10 14 18 22 26 30 34 38 42

Natural Gas Market Price

Natural Gas Procurement Price

Variable Price Contract

Fixed Price Contract

Maximum Price Contract

Collar or Fence Contract


270

The price strategies are compared for a fairly typical flower greenhouse in the Netherlands. The characteristics of this farm are listed in table 1. Table 1: Characteristics of a flower greenhouse in 2005

Production (ha) 1.9 Natural gas usage (m3/m2) 44 (top in January 7 m3 and low in July 1 m3) Income ( €) 1,005,100 Expenses ( €) 1,052,600 Profit ( €) (47,500) Total Energy costs 222,300 Natural gas costs 188,955

Source: Farm Accountancy Network LEI Table 1 shows that the greenhouse enterprise has a turn over of about €1 million and a negative profit (loss) of about €50,000, 85 % of the energy costs are costs for natural gas. The expenses for natural gas constitute more than 20% of the total expenses. Only the labour costs (about 30% of total expenses) are more important. Figure 2 shows the volatility of the gas prices in the Netherlands in the period January 2000 till January 2006. In this period the gas prices are fluctuating round an increasing trend. In fact the price per m3 doubled from about 12 eurocent per m3 in 2000 to 24 eurocent per m3 in 2005. For the greenhouse enterprise this means an increase of expenses of about 100.000 Euro annually or 10% increase in total costs. Figure 2: Natural gas price (quarterly prices in eurocents per m3 from 2000 to 2006).

Results Table 2 shows the results of the historical case study analysis. If the farmer used only variable price contracts for procuring natural gas his average procurement price would have been €8.38 per square

10 12 14 16 18 20 22 24 26 28

Jan 2000/ 2001

Jan 2001/ 2002

Jan 2002/ 2003

Jan 2003/ 2004

Jan 2004/ 2005

Jan 2005/ 2006

Price


271

meter of production from 2002 through 2005. By enlisting the use of fixed price purchase contracts he would have paid €7.91 per square meter. The maximum price strategy would have resulted expenses of on average €7.72 per square meter and by using collar price contracts €8.02 per square meter. The standard deviation of the returns is an important measure of the riskiness of a strategy. Here we can see that the standard deviation of the procurement strategies using variable price, fixed price, maximum price and collar price are €1.17, €0.61, €0.13, €0.32 respectively. Table 2 Natural gas cost (€/m2) and standard deviation per price strategy in the period 2000-2005 for a flower greenhouse in the Netherlands

Year Contract type

variable price fixed price maximum price collar price 2002 7.45 8.17 7.62 7.61 2003 8.60 7.42 7.78 8.26 2004 7.50 8.65 7.87 7.92 2005 9.95 7.41 7.60 8.24

Average Cost 8.38 7.91 7.72 8.02 Standard Deviation 1.17 0.61 0.13 0.32

There are two ways to evaluate the four strategies:

Stochastic dominance framework (Hadar & Russell, 1969); Entrepreneur risk preferences.

Stochastic Dominance Framework In this case study we have analysed four years of procurement strategies and derived a mean and a variance for each strategy. If we assume that the distribution of returns for each strategy are normal and follow the mean and variance observed in our case study analysis, we can plot the cumulative distribution function of each strategy (see figure 3). Stochastic dominance theory can be a useful tool to analyse an entrepreneur's preference among several strategy choices. First order stochastic dominance theory suggests that if all points from a strategy's distribution are to the left of the distribution plot of a second strategy, the first strategy is said to dominate the second. If we have two return distributions with cumulative density functions F(x) and G(x) respectively, then F(x) first order stochastically dominates G(x) if and only if G(x) ≥ F(x) for all values of x. Second order Stochastic Dominance (SSD) can be used to analyse a dominant strategy in the case where the cumulative density function plots of the distributions cross. F(x) is said to stochastically dominate G(x) if its area to the left of G(x) is greater. In this case F(x) is more likely to yield a more favourable result (i.e. a lower procurement price in this case). Here we see that the maximum price strategy is most likely to yield the best natural gas purchase price, thus it can be said that the maximum price strategy is second order stochastically dominant.


272

Figure 3: Strategy cumulative distribution functions

Entrepreneur Risk Preferences A risk neutral entrepreneur is one who will select a strategy based on the optimal procurement price (lowest expected costs). In this case the maximum price contract offers the lowest expected procurement price. Therefore, an entrepreneur who is risk neutral will rationally select the maximum price contract strategy in the given case. A risk averse entrepreneur is one who weights his selection between risk and the likely outcome. Such an entrepreneur's utility function can be viewed as a convex curve where he trades return for risk in a decreasing way. The risk averse entrepreneur would seek the strategy or set of strategies which maximise his utility given his unique risk and return trade-off function. In this case a rational risk averse entrepreneur would also select the maximum price contract. Possibility For Derivatives And Need For Further Research Customised derivatives are just beginning to be offered for use in agriculture in the Netherlands. This research shows that there is promise for their use in the horticulture sector for natural gas procurement. It is likely that there are many more applications in other sectors where customised derivatives could be useful in managing commodity price risk inherent in a business. The development of these customised derivative tools will likely require a combined effort by skilled researchers, innovative entrepreneurs and a responsive and committed financial derivative sector. To develop customised derivatives a number of conditions ought to be present:

Natural Gas Cost (Euro per m2 and year)

Cumulative Distribution

6 7 8 9 10 11 12

0.2

0.4

0.6

0.8

1 A

B D C

A-Variable Price Contract B-Fixed Price Contract C-Maximum Price Contract D-Collar Price Contract


273

Commodity price availability and tradability. The creator of a derivative must have a direct or indirect way to manage the risk inherent with such offering. For example, if there is no trade in a hog futures market but there is very liquid trade in grain and protein products, a company could construct a model where the input price is a combination of grain and soy protein plus a production factor and plus a margin factor to arrive at an approximate live hog value. Using such a model, one may offer live hog derivatives which could be a useful tool for pork producers and processors. This technique is being applied in the agriculture and food industries in other countries. For example, high fructose corn syrup can be purchased based on corn prices plus processing costs plus a margin factor; - Price volatility. There should be a significant price risk present; - Price variation should impact farm or firm income. The commodity price volatility should have a noticeable impact on farm or firm income which will incentivise producers and firms to manage the price risk by using customised derivatives; - The market size should be considered. It is important for derivative creators and sellers to identify markets which meet a minimum size that provides an acceptable possibility to reward them for their efforts. Thus, one must consider the aggregate market size in determining which derivatives to offer; - There should be willingness by derivative firms and entrepreneurs to experiment and innovate in the realm of financial tools offered and used for commodity purchases and sales. Suggestions are listed for further research on this topic which may be helpful for the continued commercial development and application of customised derivatives. - Research could be undertaken to create a modelling tool for entrepreneurs to analyse the value and impact of incorporating the use of customised derivatives into their businesses; - Case study research utilising actual cases of the use of customised derivatives may be useful for entrepreneurs; - Further analysis of more exotic derivatives structures may offer additional benefits; - Research into the business implications of the use of customised derivatives could be helpful; - Research should be completed on decision tools for determining optimal derivative selection or optimal combinations of derivatives, including considering more detailed analysis of derivative selection under varying entrepreneur risk preferences. Also simulation with different strike prices or collar prices gives insight in the relation between premiums for lowering the risks and the change in volatility of costs or income; - Research into other implications of using customised derivatives such as cash flow considerations would be insightful; - Research into the further development of customised derivative tools in other commodities areas would help expand the knowledge and potential development of derivatives in other sectors; - Research other business models for introduction of derivatives in the market. For example the derivates studied in this paper could be offered by energy companies instead of a bank (the developer of the derivatives). Conclusions The horticulture sector in the Netherlands could find significant value by utilising natural gas price derivatives to manage the volatility risk and price risk of natural gas in the period 2000-2005. The cost of natural gas can be reduced through the prudent use of derivatives when compared with a variable price procurement strategy. Furthermore, all natural gas derivative strategies considered here offer less procurement cost volatility in the case study analysis. Using derivatives will reduce volatility of prices and volatility of procurement costs. The elimination of part of these risks has value for some farmers for example in case they need loans for extending the business.


274

Price Performance In the studied case, the variable price strategy resulted in a natural gas cost of €8.38 per square meter of production. The fixed price, maximum price and collar price contracts resulted in a cost of €7.91/m2, €7.72/m2 and €8.02/m2, respectively. This research shows that the best natural gas price for the last 4 years would have been achieved using a maximum price contract strategy. Of course these results on historical data are no guarantee that the use of derivatives will realise similar savings for further periods. Normally the lowest mean prices and highest volatility will be expected in the case of a variable price strategy. Risk Performance The case study analysis showed that the standard deviation of natural gas costs was €1.17/m2 for the variable price strategy. The fixed price, maximum price and collar price strategies yielded standard deviations of costs of €0.61/m2, €0.13/m2, and €0.32/m2, respectively. In the case study all derivative contracts were less risky than the variable price contract strategy, with the least risky strategy being the maximum price strategy. Business Benefits The savings in natural gas costs observed in the case study in the period 2000-2005, comparing derivative procurement strategies with the variable price procurement strategy, ranged from 5.5% to 8.5% savings in the cost of natural gas. The business impact of incorporating customised derivatives into the procurement and risk management plan of one's business appears to be a substantial reduction in the input cost volatility and approximately a 1% cost savings for the business as a whole. The use of derivatives leads to a strong decrease of volatility which means that the farm income is stabilised. This can lead to lower interest rates for loans or could increase the possibility for extra loans. References Black, F., 1976; 'The Pricing of Commodity Contracts'. In: Journal of Financial Economics 3; p 167-79. Hardarker, J.B., R.B.M. Huirne, J.R. Anderson and G. Lien, 2004; Coping with riskin Agriculture.

Second edition, CAB International, Wallingford. Horsager, K., W.H.M. Baltussen and G.B.C. Backus, 2006; Customised commodity derivatives; the case

of natural gas in the Dutch horticulture sector. Report 2.06.05, LEI, The Hague, The Netherlands. Hull, J., 2002; Options, Futures and Other Derivatives. 5th Edition, Prentice Hall, New Jersey. http://en.wikipedia.org/wiki/Volatility


275

Appendix 1 Black's option pricing model for futures and forwards ('BOPM') C = e-rt [ F N(d1) – X N(d2) ] where: d1 = ( ln (F/X) + (σ2 / 2t) / σ√t d2 = d1 – σ√t and: C is the price of the call option, F is the futures price, X is the strike price, t is the time remaining, σ is the standard deviation of returns or volatility, r is the risk free rate, ln denotes the natural logarithm, and N is the standard normal distribution function. Assumptions of the BOPM: There are no transaction costs; The interest rate remains known and constant; Prices are log normally distributed; Volatility is constant over the life of the option. One of the requirements of the Black’s Option Pricing model is a estimation of the future

volatility. One method is to look at the current market conditions. Because there is a lack of traded derivatives in the Netherlands another method is used which is based on historical volatility.

Historical volatility can be calculated as:

σh = Standard Deviation of ln(Pt/P(t-1) ) * √t where: σh is the historical volatility Pt is the price in period t, P(t-1) is the previous period's price and, t is the number of periods in a year (see http://en.wikipedia.org/wiki/Volatility) After calculating the historical volatility, we adjust the historical volatility in the following manner

to obtain the estimated future volatility: σe = 1.25 σh where: σe is the estimated volatility and, 1.25 is a constant based upon the observation that there is generally a premium for future

volatility relative to calculated historically volatility. The 1.25 is an arbitrary value. For commodities the range is 1.1 to 1.3. In the USA’s Natural Gas market the future volatility is at the moment about 1.3 times the historical volatility.


276

WHAT IS THE POTENTIAL FOR PRECISION AGRICULTURE BASED ON PLANT SENSING?

Jon T. Biermacher

The Sam Roberts Noble Foundation, Inc. 2510 Sam Noble Parkway

Ardmore, OK 73401 e-mail: [email protected]

Francis M. Epplin, B. Wade Brorsen, John B. Solie and William R. Raun

Oklahoma State University, US Abstract Plant-based precision nitrogen fertilizer application technologies have been developed as a way to predict nitrogen needs. This paper determines the expected profit from using plant sensing to determine nitrogen needs in winter wheat. The equipment necessary for precision application of nitrogen based on plant sensing is available commercially, but adoption has been slow. We find that plant sensing systems are roughly breakeven with current systems, which likely explains the slow rate of adoption. Keywords: nitrogen, precision agriculture, stochastic plateau, wheat What is the Potential for Precision Agriculture Based on Plant Sensing? Past research suggests that most agricultural producers apply more nitrogen than is needed in most years. Precision application of nitrogen based on soil sampling and yield monitors has been developed to help producers decide how much nitrogen to apply. However, costs and measurement errors have limited usefulness of nitrogen recommendations based on yield monitors and soil sampling of small grids (Babcock, Carriquiry, and Stern, 1996; Arslan and Colvin, 2002). Soil sampling for nitrogen has always been marginal due to low levels of available nitrogen in the soil. Use of yield monitors has also been limited because while yields vary substantially across the field they do not vary in the same way every year. These limitations associated with use of soil sampling and yield monitors might explain why few producers use these technologies to determine how much nitrogen to apply (Daberkow and McBride, 2000). More recently, plant sensing technologies have been introduced to agricultural crop producers. Plant sensing is promising since it is more precise than soil tests and yield monitors in predicting nitrogen response. However, adoption of such technologies has also been slow. Plant sensing is clearly an outstanding technical achievement, but it apparently faces some economic hurdles. One economic challenge to plant sensing is that it requires nitrogen be applied in liquid form whereas preplant nitrogen applications can use lower-priced anhydrous ammonia. Like soil tests and yield monitors, plant sensing technology is expensive. In addition, investment in plant sensing is irreversible in the sense that once the machines are bought the farmer cannot easily resell them to recoup the investment. When investment is irreversible there is an option value in postponing the decision to invest and search for new information. However, accurate information about producer benefits from using plant sensing is lacking. This lack of information may explain why adoption has been slow. Information about economic performance of plant sensing technologies is also valuable to agricultural manufacturers since it would provide them with a target cost needed to get producers to adopt these technologies.


277

The objective of this research is to determine the potential profitability of nitrogen recommendations based on whole field and variable rate wheat plant sensing relative to conventional practices. We develop a yield response to nitrogen function that is conditional on plant sensing. Using nine years of data containing wheat yield, optical reflectance, and levels of pre-plant nitrogen information, we find that plant sensing systems are roughly breakeven with current systems, which likely explains the slow rate of adoption. Theory Current plant sensing systems essentially require that a producer conduct a nitrogen response experiment in each field. The experiment consists of a single nitrogen-rich strip where enough nitrogen is applied so that nitrogen will not be the constraining input. Current plant sensing measures typically measure the normalized difference vegetative index, but the theory does not depend on what measure is used. With a nitrogen-rich strip, sensing is used to compare the fertilized and unfertilized plants and a formula is used to determine nitrogen needs. Nitrogen needs can vary across the field and systems have been developed to sense grids smaller than a square meter in an attempt to apply just the right amount of nitrogen to each grid. We assume here that the nitrogen application system chosen does not affect the optimal quantity of other inputs. Nitrogen can either be applied preplant, in which case anhydrous ammonia can be used or nitrogen can be applied as a topdress application to growing plants. Assuming that price and yield are uncorrelated, the producer’s optimization problem can be represented as

(2)

( )

.0,

and,}1,0{

)( then0 If

0 then0 If

0 then0 If

,

),(

s.t.

),1(

)E()E(max

P

2

3

1

)(223

21 ,,

≥

∀∈

=>

>>

>>

+=

=

+−−

−−−−=

T

i

T

T

P

TP

ORINT

NRSPTTPP

NN

NN

i

ORINN

N

N

NNN

Nyy

bb

bbNrNrypRTP

λ

λ

λ

λ

γ

λλλ

λλλλλλ

where R is net return above nitrogen fertilizer application costs; y is yield; N is the sum of preplant nitrogen (NP) and topdress nitrogen (NT); 1>γ is the relative efficiency of topdress nitrogen relative to anhydrous; p represents the expected price of wheat; λ = (λ1, λ2, λ3) is a vector of binary choice variables; rP and rT represent the prices of preplant nitrogen and topdress nitrogen, respectively; bP, bNRS, bN(ORI), and bT represent preplant nitrogen application costs, cost of the nitrogen-rich strip, cost of topdressing with optical sensing, and conventional topdress nitrogen application costs, respectively; and the function N(ORI) is the application rate algorithm based on precision sensing information (NRS). Note that λ3 is selected conditional on NRS being known. Increased yields with precision plant sensing could come about from conventional systems applying either too much or too little nitrogen. The evidence regarding whether excess nitrogen causes yields to decline is mixed (Biermacher et al.) but tends to suggest little or no yield decrease from applying excess nitrogen. A conventional system that applied too little nitrogen would clearly lead to lower yields than a


278

precision sensing system that applied exactly the amount of nitrogen needed. In practice however, producers apply more nitrogen than is needed in most years. As a result, most of the advantage of precision sensing is expected to be due to reduced cost of nitrogen fertilizer rather than increased yield. While the optical sensing system clearly uses less total nitrogen, it faces a major economic challenge because plant sensing uses nitrogen in liquid form, which is more expensive than anhydrous ammonia used with conventional technologies (i.e. rP < rT). Procedures Past research and characteristics underlying plant sensing technology suggest that stochastic plateau functions are more appropriate to represent yield response to nitrogen than polynomial and switching regressions (Tembo, Brorsen, and Epplin, 2003; Katibie et al., 2003; Katibie et al., 2007). Thus, we use a linear response stochastic plateau function to represent wheat response to nitrogen. The linear response function with a stochastic plateau can be written as (3) ,],)(min[ 210 itttM

T

it

P

it

S

itit uvNNORIy εµβββ +++++=

where ity is wheat yield in bushels per acre on grid i in year t; NP is the level of preplant nitrogen; NT is

the level topdress nitrogen; )( P

it

S

it NORI represents optical reflectance information taken in the spring on

grid i in year t; mµ is the average plateau yield, 210 and ,, βββ , are parameters to be estimated;

tv represents the plateau year random effect; tu is a year random effect that shifts the intercept, and itε is

an i.i.d. normal error term. Our data include preplant nitrogen and ORI readings for preplant nitrogen, but no topdress nitrogen. Therefore, equation (3) cannot be estimated. To circumvent this limitation, we assume that the marginal productivity of topdress nitrogen is the same (or at least proportional to) the marginal productivity of preplant nitrogen. Next, we estimate two separate regressions: wheat yield is regressed on optical reflectance information, and optical reflectance information is regressed on preplant nitrogen. The estimates from these regressions are then used to construct equation (3). Let the relationship between wheat yield and optical reflectance information be written as (4) ,)( it

P

it

S

itit NbORIay θ++=

where ity is wheat yield in bushels per acre on grid i in year t , a and b are the intercept and slope

coefficients to be estimated, and the error term itθ is partitioned into an independently and identically

distributed random error term *itθ that has mean zero and variance *

2 ,θ

σ and year random effect tω that has

mean zero and variance 2 .ωσ

Independence is assumed between the two variance components, and therefore the variance of the overall error term is *

2 2 2 .θ ω θσ σ σ= + The symbol )( P

it

S

it NORI is defined as the normalized difference vegetation

index (NDVI) sensor reading taken in the spring on grid i in year t and is adjusted by the number of growing degree days. The optical reflectance index (ORI) measures the amount of nitrogen available to the plants at the time of sensing, which in turn helps in quantifying the amount of additional nitrogen needed to reach plateau yields. The second regression used to construct equation (3) is the regression of optical reflectance information on preplant nitrogen. This relationship is defined as (5) ittt

MP

it

P

it

S

it uvORINNORI εβα ++++= ),min()(


279

where )( P

it

S

it NORI is an optical reflectance index reading taken in the spring on grid i in year t; α and

β are intercept and slope parameters to be estimated; P

itN is any nitrogen in grid i at the time of planting

in year t; MORI is the average sensor reading taken from the nitrogen rich strip; )(0,~ 2vt Nv σ represents

year random effects on the plateau ; )(0,~ 2ut Nu σ represents year random effects; and )(0,~ 2

ησε Nit is

the traditional random error term. The estimates from equations (4) and (5) are used to construct equation (3). Again, the key assumption is that the marginal productivity of topdress nitrogen is the same as (or at least proportional to) the marginal productivity of preplant nitrogen; that is, ,1 b=β and .2 ββ b= So, with this assumption we

set a=0β from equation (4), equation (3) can be re-written as

(6) ,],)(min[ itittt

MT

it

P

itit bbubvbORIaNbNbay θεββα ++++++++=

which imposes .// βbNyNy P

itit

T

itit =∂∂=∂∂

Determining the optimum preplant level of nitrogen analytically using the stochastic plateau model (6) is not straightforward because year and spatial random effects enter equation (6) nonlinearly. The optimal level of nitrogen to apply with this functional form has been developed by Tembo et al. (2007). The

optimum input level ( *PitN ) can be determined as

(7) ( )),1,0min(* ασ

βδ −+= v

Mit ZORIN P

where δZ is the critical Z-value where )/(1 βδ pbr=Φ−= is the observed probability in the right-hand

tail of the N(0, 1) distribution, r is the price of nitrogen, and p is the price of wheat. Parenthetically, if the variable rate plant sensing technology is applied, and we assume information from the NRS and each grid is sensed perfectly, then we can re-write equation (3) as (8) ,],)(min[ it

NRS

t

T

it

P

it

S

itit bORIaNbNbORIay θβ ++++=

where NRS

tORI is the in-field experimental measure from a nitrogen-rich strip or some other measure.

The model in (8) is a linear plateau model and the optimum is the level of nitrogen needed to reach the plateau on each grid, which is

(9) .*

β

S

it

NRS

tT

it

ORIORIN

−=

Note that we are implicitly assuming that none of the error in equation (5) represents measurement error. If we were to add measurement error, we would end up with the model developed by Berck and Helfand (1990) and Paris (1992). Adding measurement error would further reduce the value of sensing. In the case of a whole field application, the ORI on each grid is no longer known. In the case of uniform application using sensing, only an average measure of ORI is obtained from the response, which implies that spatial variation on each grid is expected to be present. However, since sensor measurements are taken from the NRS, which covers such a large area in the field, no error in the plateau is assumed. This implicitly assumes that all variation across grids is due to differences in available nitrogen and thus the variation across grids in the nitrogen-rich strip should be zero (which is not entirely true and is yet another assumption that causes our results to favor sensing). The response portion of the plateau thus has an additional error and the production function becomes:

(10) .],min[ itNRStit

Tit

S

itit bORIabNbORIbay θεβ +++++=


280

where S

itORI is an average ORI reading across an unfertilized portion of the field near the nitrogen-rich strip. The solution to the optimal level of nitrogen in (10) is analogous to (6) except that the upper rather

than the lower tail of the distribution is needed. The optimal whole field ( *WitN ) can be determined as

(11) ( )).1,0min(* S

itNRStit ORIZORIN

W

W −+= εδ σβ

where W

Zδ is the critical Z-value, )/( βδ pbrW =Φ= is the observed probability in the right-hand tail of

the N(0, 1) distribution, r is the price of nitrogen, and p is the price of wheat. Note that in an actual field the plateau might also vary across grids and so again this is a simplification that could cause the value of sensing to be overstated, unless the sensing could also identify the grids with less yield potential. Data and Empirical Procedures Parameters of equations (4) and (5) are estimated using data from nine years of on-farm winter wheat experiments conducted at seven locations located on or near agronomic research stations throughout the state of Oklahoma from 1998-2006. The data include observations for wheat yield, optical reflectance information, and level of preplant nitrogen. Data were collected at locations near Stillwater every year. Data were collected at Haskell from 1999-2002. At Hennessey, data were collected for 2000 and 2002. At Lahoma data were collected in all years except 1998 and 2001. At Perkins, data from two experiments were used; one included data collected in 1998 only, and the other utilized data from 1998-2006. At the Tipton site, data were only collected in 1998. Winter wheat was planted for grain only at a 78 kg ha-1 seeding rate using a 0.19 meter row spacing at all locations, excluding one of the two experiments at Perkins where spacing ranged from 0.15 to 0.30 meters. Nitrogen rich strips were placed in each experimental plot prior to planting wheat in late September or early October. All optical reflectance readings were taken during Feekes growth stages 4 (leaf sheaths beginning to lengthen) and 5 (pseudo-stem, formed by sheaths of leaves strongly erect) (Large, 1954). All reflectance readings from wheat collected from a 4.0 square-meter area between 10 a.m. and 4 p.m. were taken under natural lighting between January and March. Grain yield was measured from the same area where spectral reflectance data were collected. Additional information regarding the experiments can be found in Mullen, 2003. Parameters in equation (4) are estimated using a linear mixed effects model (PROC MIXED in SAS). Year random effects are tested using a likelihood ratio test. The parameters of the stochastic plateau model represented by equation (5) are estimated using SAS NLMIXED (2002-2003). Then, the estimates from equation (4) and (5) are used to construct equation (6), which is then used to simulate expected net returns from each of the seven nitrogen application systems considered.


281

Nitrogen Levels Equation (6) is used to compute the application levels of nitrogen fertilizer for each of several systems, including (1) an all-before-planting; (2) a whole field precision system; (3) a variable rate precision system; (4) the NFOA system developed by Raun et al., 2002; (5) the extension recommendation of 80 pounds per acre preplant system; and (6) an all-before-planting system that represents the average of what producers were actually found to be applying in the southern Plains (i.e., 63 pounds per acre) in a survey conducted in 2004 (Hossain et al., 2004). In addition, a check (system (7)) that has no nitrogen applied is included. Optimal application levels of nitrogen for systems 1, 2, and 3 are derived using the response function outlined in equation (3), and the optimal application level of nitrogen for system 4 is derived using the algorithm provided in Raun et al., 2002. Derivations of optimal levels of nitrogen for systems 1-4 are explained in detail in Biermacher, 2006. Simulation of Expected Net Returns Equation (6) is simulated to determine the expected net return from each of the alternative systems. Net returns on 250 sample grids within each of 250 sample years were simulated using the following steps. First, sample values for the error components in equation (6) are simulated using a random number generator. Errors are assumed normally distributed with mean zero and estimated variances provided from the regression procedures used to estimate equations (4) and (5). Intercepts, slopes, and expected value of optical reflectance information at the plateau are also provided from these regression procedures. In addition to the error components, values of NRS

t

S

it ORIORI and are simulated for each grid and year of

the sample. Moreover, application costs, and prices for NH3 and 28% UAN are included. A zero level of N is assumed when expected net returns from application are negative. The process for calculating sample values of optical reflectance information taken from the nitrogen rich strip is (12) ,tt

MNRS

t uvORIORI ++=

and the process for calculating sample values for the optical reflectance information on an individual grid and year is described by equation (5). Again, we note that since the NRS covers such a large area of the field, the plateau spatial variability is assumed to average to zero given that a substantial number of readings are taken from it. Once sample values for the errors and the optical reflectance information are simulated for each grid and year, then formulas for equations (7), (9), (11), and equation (17) in Biermacher, 2006 are used to generate samples of optimal nitrogen rates for each grid in each year for each system. The yield response function defined in equation (6) is then used to calculate sample values for wheat yield for each system, grid, and year in the sample. Net returns are then calculated as the difference between wheat revenue and cost of nitrogen and nitrogen application expenses for each grid in the year. The Monte Carlo integration is then completed by averaging net returns across the sample of years for each system. For each system, a long run average price of $3 per bushel was used for the expected price of wheat grain and market prices of $0.15 and $0.25 per pound are used for anhydrous ammonia and 28% UAN, respectively (Oklahoma Department of Agriculture). Gains in Efficiency It is believed that some gain in efficiency will be obtained when the plant-based sensing technology is used instead of the traditional preplant systems. However, it is not assumed as in Raun et al., 2002 that a seventy percent gain (i.e., 70.0=γ ) is achievable. For this study, we are assigning a twenty percent gain


282

in efficiency to the marginal product of nitrogen, such that the slope parameter β is effectively multiplied by an efficiency parameter γ that is set equal to 1.2. Results and Discussion Regression estimates of equation (4) are presented in table 1. Rejection of the null hypothesis that no random effects exist were based on the likelihood ratio test. The slope parameter (b) is significant at the .05 level. Table 1. Regression of Wheat Yield Response on Optical Reflectance Information

Statistic Symbol Estimatesa

Intercept a -5.2268 (3.52) Optical reflectance b 6.7291 (.42)

Year random effect 2ωσ 103.81

(25.83)

Error variance 2

*ϑσ 105.31

(5.33) a Asymptotic standard errors are in parentheses. Note, that the parameter estimates for equation (2) were estimated using PROC MIXED in SAS.

The intercept parameter (a) was not significant at the .05 level; however, it was significant at the .10 level. Estimates of equation (5) are presented in table 2. The marginal product of nitrogen ( )20.0)0297.07291.6( =×=βb suggests that approximately 2.27 kg of nitrogen should be applied to gain an additional bushel of wheat rather than the 0.65 kg suggested by the NFOA model.


283

Table 2. Stochastic Linear Plateau Model of Optical Reflectance Information as a Function of Nitrogen Statistic Symbol Estimatesa

Intercept α 5.6882 (.0640) Level of nitrogen β .0297 (.2022)

Average plateau ORI NRSORI 6.8879

(.0599)

Nitrogen at expected plateau NRS

tN 57.8045 (.1958)

Variance of plateau yield 2vσ 0.5861

(.0936)

Variance of year random effect 2uσ 0.7563

(.0737)

Variance of error term 2ησ 0.5097

(.0263) a Asymptotic standard errors are in parentheses. Note, the parameter estimates for equation (3) were estimated using NLMIXED procedure in SAS. Expected yield, optimal levels of nitrogen, and expected profits for each system are reported in table 3. The perfect (unachievable) information system had the largest expected profit of approximately $271 ha-1. Net return to nitrogen application for this system was approximately five percent greater than the average net return for the optimal preplant system determined using the TBE model, and was only approximately seven percent greater than the net return from the state recommendation of applying 90 kg ha-1 prior to planting in the fall, a value of $15 ha-1 over the state recommended system.


284

Table 3. Average Yield, Nitrogen, and Expected Profits from Alternative Nitrogen Management Systems without Plateau Spatial Variability

System

Estimate 0/0a 80/0b 63/0c 0/HHd TBE/0e 0/GSf 0/NFOAg

Average Yield (kg ha-1)

2196 2723 2696 2687 2693 2740 2476

Average Nitrogen (kg ha-1)

0.00 90 71 52 65 37 18

Average profit ($ ha-1)

242 256 259 265 260 271 255 a the check system with no nitrogen added. b the system that represents the state extension recommendation of 90 kg ha-1. c the system that represents the average level of nitrogen applied in the state of Oklahoma that was reported by producers via a survey conducted in 2004. d the system that represents the portable, handheld precision system where no nitrogen was applied prior to planting. e the system that represents the analytical approach developed by Tembo, Brorsen, and Epplin (2007) to determine the optimal level of nitrogen to apply in the fall prior to planting. f the system that represent the plant-based variable rate precision system that assumes perfect knowledge about the random processes. g the system that represents the NFOA developed by Raun et al. (2002). The portable handheld system had an average net return that was only $9 ha-1 greater than that obtained from the state extension system. Although, the TBE system realized a slightly higher yield, the gain from the reduction in fertilizer cost was the primary factor accounting for the difference. Note that using portable sensing provides the chance that some areas of the field could receive less nitrogen than actually needed, which will likely keep some yield in the field from reaching its potential plateau. A noteworthy comparison is the $16 ha-1 difference in net return between the perfect information system and the system that utilized the NFOA. This could be viewed as an indication that further improvements could be made to the NFOA. However, it is unlikely that the NFOA could ever perform as well as the perfect information system described in this paper. Note that the marginal product of nitrogen for the NFOA is too high and, adjusting it down to the size of that found using the data, the NFOA outcome would be similar to that given by the profits for the 90 kg ha-1 system. Sensitivity values for independent relative changes in the price of wheat, price of anhydrous ammonia, and the price of 28% Urea-ammonium nitrate are reported in table 4. The expected value of the perfect information system is not very sensitive to either below average or above average prices of wheat. In the extreme case where wheat price increases to $0.074 kg-1, the additional value of the perfect information system above that of the state recommendation is only about $3.75 ha-1.


285

Table 4. Sensitivity Values for Independent Relative Changes in Price of Wheat, Price of Anhydrous Ammonia, and Price of 28% Urea-Ammonium Nitrate

System

Parameter Price 0/0a 80/0b 63/0c 0/HHd TBE/0e 0/GSf 0/NFOAg

Price of Wheat ($/kg)

.030 161 155 160 168 162 171 165

.045 242 256 259 265 260 271 255

.060 323 356 358 363 360 372 346

.074 404 456 457 461 459 473 437

Price of NH3 ($/kg)

0.07 242 256 259 265 260 271 255

0.11 -------- 236

243 -------- 248 -------- --------

0.18 -------- 206 220 -------- 234 -------- --------

0.23 -------- 186 204 -------- 229 -------- --------

Price of UAN ($/kg)

0.11 242 256 259 265 260 271 255

0.16 -------- -------- -------- 257 -------- 263 251

0.20 -------- -------- -------- 248 -------- 255 248

0.23 -------- -------- -------- 244 -------- 251 247

a the check system with no nitrogen added. b the system that represents the state extension recommendation of 80 pounds per acre. . c the system that represents the average level of nitrogen applied in the state of Oklahoma that was reported by producers via a survey conducted in 2004. d the system that represents the portable, handheld precision system where no nitrogen was applied prior to planting. e the system that represents the analytical approach developed by Tembo, Brorsen, and Epplin to determine the optimal level of nitrogen to apply in the fall prior to planting. f the system that represent the plant-based variable rate precision system that assumes perfect knowledge about the random processes. g the system that represents the NFOA developed by Raun et al. (2002). As expected, the value of perfect sensing technology increases relative to the state system as the price of NH3 increases relative to the price of UAN. When the price of NH3 is increased to the point where it is equal to the price of UAN, the value of the variable rate system increased to approximately $40 ha-1 over that of the state-recommended system. The opposite relationship exists when the price of UAN increases relative to the price of NH3. If the price of UAN increases to $0.27 kg-1 , holding the price of NH3 constant at $0.07 kg-1, then the value of the state recommended system is approximately $10 ha-1 more profitable than the perfect variable rate system. In this situation, a typical producer would not be interested in adopting the plant-based precision system. Currently, this plant-based precision sensing technology is available on a commercial basis, and is being promoted to increase net returns to nitrogen fertilization by $25-$75 ha-1. However, the findings of this study do appear to explain why adoption has been slow. These findings also indicate that the optical sensing technology, including the nitrogen fertilizer optimization algorithm (NFOA), in many cases, does not apply enough nitrogen fertilizer, and therefore could be improved upon.


286

References Arslan S., and T.S. Colvin. 2002. “Grain Yield Mapping: Yield Sensing, Yield Reconstruction, and

Errors.” Precision Sensing. 3:135-154. Babcock, B.A., A.L. Carriquiry, and H.S. Stern. 1996. “Evaluation of Soil Test Information in

Agricultural Decision-Making.” Applied Statistics 45:447-461. Berck, P., and G. Helfand. 1992. “Reconciling the von Liebig and Differentiable Crop Production

Functions.” American Journal of Agricultural Economics 72,4:373-386. Biermacher, J.T. 2006. “Economic Feasibility of Site Specific Optical Reflectance Technology as an

Alternative Strategy for Managing Nitrogen Applications to Winter Wheat.” Unpublished Dissertation, Oklahoma State University.

Biermacher, J.T., F.M. Epplin, B.W. Brorsen, J.B. Solie, and W.R. Raun. 2006. “Maximum Benefit of a

Precise Nitrogen Application System for Wheat.” Precision Agriculture 7:193-204. Daberkow, S.G, and W.D. McBride. 2000. “Adoption of Precision Agriculture Technologies by U.S.

Farmers.” Proceedings of the 5th International Conference on Precision Agriculture, Minneapolis, MN, ASA/CSSA/SSSA, Madison, WI, July 16-19.

Greene, W. 2000. Econometric Analysis, 4th ed. Upper Saddle River NJ: Prentice-Hall. Hossian, I., F.M. Epplin, G.W. Horn, and E.G. Krenzer, Jr. 2004. “Wheat Production and Management

Practices Used by Oklahoma Grain and Livestock Producers.” Oklahoma Agricultural Experiment Station Bulletin 818.

Katibie, S., W.E. Nganje, B.W. Brorsen, and F.M. Epplin. 2007. “A Cox Parametric Bootstrap Test of the

von Liebig Hypothesis.” Canadian Journal of Agricultural Economics 55:15-25. Kaitibie, S., F.M. Epplin, B.W. Brorsen, G. W. Horn, E. G. Krenzer, Jr., and S.I. Paisley. 2003. “Optimal

Stocking Density for Dual-Purpose Winter Wheat Production.” Journal of Agricultural and Applied Economics 35:29-38.

Large, E.C. 1954. “Growth Stages in Cereals.” Plant Pathology 3:128-129. Mullen, R.W., K.W. Freeman, W.R. Raun, G.V. Johnson, M.L. Stone, and J.B. Solie. 2003. “Identifying

an In-Season Response Index and the Potential to Increase Whet Yield with Nitrogen.” Agronomy Journal 95:347-351.

Oklahoma Market Report, Oklahoma Department of Agriculture, Oklahoma City, OK. Various Issues. Paris, Q. 1992. “von Liebig Hypothesis.” American Journal of Agricultural Economics 74,4:1019-1028. Raun, W.R., J.B. Solie, M.L. Stone, K.L. Martin, K.W. Freeman, R.W, Mullen, H. Zhang, J.S. Schepers,

and G.V. Johnston. 2005. “Optimal Sensor Based Algorithm for Crop Nitrogen Fertilization.” Community Soil Science Plant Analysis 26:2759-2781.

Raun, W.R., J.B. Solie, G.V. Johnson, M.L. Stone, R.W. Mullen, K. W. Freeman, W.E. Thomason, and

E.V. Lukina. 2002. “Improving Nitrogen Use Efficiency in Cereal Grain Production with Optical Sensing and Variable Rate Application.” Agronomy Journal. 94:815-820.


287

Tembo, G., B.W. Brorsen, and F.M. Epplin. “Linear Response Stochastic Plateau Functions.” 2003.

Journal of Agricultural and Applied Economics 35:445. Tembo, G., B.W. Brorsen, F.M. Epplin, and E. Tostao. 2007. ”Crop Input Response Functions with

Stochastic Plateaus.” Unpublished working paper. Oklahoma State University.


288

A FEASIBILITY STUDY OF CONTRACT FINISHING OF HOGS

Bill Brown

Department of Agricultural Economics, College of Agriculture and Bioresoucres,

51 Campus Drive, University of Saskatchewan, S7N 5A8 Email [email protected]

Marv Painter,

Department of Management and Marketing College of Commerce,

University of Saskatchewan

Mark Ferguson, Industry and Policy Analysis,

Sask Pork.

Abstract A multi year financial model was used to evaluate the economics of contract finishing of hogs. The model includes projected income statements, balance sheets, and cash flow statements, as well as the calculation of the Internal Rate of Return (IRR) using discounted after tax cash flows. The model uses contract fees and incentive schemes, and the evaluation of manure and how much of it can be captured as income. It uses depreciation on capital investment (which affects the amount of income taxes paid), interest costs on borrowed capital, labor, utilities, insurance, maintenance, property taxes, and the cost of spreading the manure as expenses. The effect of an injection of patient capital was also calculated. The calculations were done for a 20 year period, from the time the facility is built and stocked with hogs to the end of the serviceable life of the barn. The results indicate that there are at least three conditions within the barn enterprise that have to be met in order for it to become economically viable. The first condition is the life of the contract. If the barn has a serviceable life of 20 years, it must be full of pigs for all or most of that time to generate competitive rates of return. The second condition is the capture of the nutrient value of the manure. The hog owners have not been able to capture the nutrient value of the manure because they don’t usually own the land surrounding their enterprises, that is to say stand alone barns have not been able to sell the manure at its full nutrient value. However, a contract feeder of hogs can locate the barn in the middle of their own land and take full advantage of the nutrient value of the manure through their cropping enterprises. The third condition is financial leverage. If the first two conditions are met and interest rates are below 8%, contract finishers of hogs can use financial leverage to their advantage. The addition of a 10% patient capital also helps the economic viability of the enterprise. Keywords: financial model, contract finishing of hogs Introduction A multi year financial model was used to evaluate the economics of contract feeding of hogs. The model includes projected income statements, balance sheets, and cash flow statements, as well as the calculation of the Internal Rate of Return (IRR) using discounted after tax cash flows. The IRR is the annual percentage rate of return on the original investment of equity capital generated by the after tax cash flows from the investment over its life. The model uses contract fees and incentive schemes and the evaluation of manure and how much of it can be captured as income. It uses depreciation on capital investment


289

(which affects the amount of income taxes paid), interest costs on borrowed capital, labor, utilities, insurance, maintenance, property taxes, and the cost of spreading the manure as expenses. The effect of a patient capital injection (no interest and repayment over 5 years starting in year 5) is also calculated. The calculations were done for a 20 year period, the serviceable life of the barn, from the time the facility is built and stocked with hogs. Most hog finishing contracts are for 5 years with an option to renew for another 5 years. In addition barn financing is generally done with a 10-year repayment period. Contract Hog Finishing Income Contract finishing of hogs usually requires the barn owner to build the barn to the general specifications of the hog owners. The hog owner supplies the hogs, the feed, veterinary expertise, treatments, and sometimes the labor as well. The barn owner is essentially renting the facility to the hog owner. The rental income or contract fee is based on the space required by a hog during the feeding period. The standard contract fee is approximately $54.00Cdn per pig space per year. This fee is based a 25 kilogram pig coming into the barn and taking 16 weeks to finish thereby resulting in 3 batches of pigs being finished per year. The capacity of the barn is usually based on 0.69 square meters of pen space required per pig. There are also incentives for good feed conversion that can add up to another $1.50Cdn per pig space per year. Some hog owners prefer to supply their own labor when contracting finishing. Others allow the barn owner to supply the labor. Contracts including barn owner supplied labor are usually a little lower but include training, supervision, and a larger incentive program. These contracts usually result in almost equivalent rates of return on investment to the barn owner plus the income generated by the labor. However, these contracts also include a management clause that will allow the hog owner to take over management if animal performance is compromised. Contract Finishing Capital Cost The two types of hog feeding systems that will be analyzed are a 2,400 head finishing facility with an earthen manure storage (EMS) system. The unit size was chosen because this fits nicely with the weekly supply of 25 kilogram pigs being produced from 3,000 or 6,000 sow unit operations weaning about 25 pigs per year per sow. The facility is equipped with fully slatted floors and small pens. The capital cost of the feeding facility is summarized in Table 1. It is assumed that most farm sites will have an existing water supply and exiting phone, natural gas, and power services. Given these assumptions, the total capital cost of this enterprise is $290.47Cdn per pig space. Table 1: Capital Cost ($Cdn) of 2,400 Head Barn with Earthen Manure Storage

(25 m x 75 m .6 m pits, EMS) 2.5m Ceiling Material Building Materials & Concrete

$ 353,593

Total Labor $ 318,535 Development Costs $ 25,000 Totals $ 697,127 Total Cost/Pig Space $ 290.47


290

Economic Evaluation of Hog Manure The economic evaluation of the hog manure will be based on a number of calculations. The first calculation is the amount of manure produced and the amount of land needed upon which to spread the manure. The second calculation pertains to the value of the manure. Hog operations that do not own the surrounding land cannot capture the full value of the manure because neighboring land owners are not willing to pay for the full nutrient value. However, hog operations that are associated with the owners of the surrounding land may be able to capture the value of the nutrients available to the crop or crops to be grown on the land. A final consideration is the cost of applying the manure to the land and the distance within which it should be transported. The Amount of Manure Produced A hog being fed from 25 kilograms to slaughter weight produces 8.5 liters of manure a day (Saskatchewan Agriculture and Food, 2006). It follows that a 2,400 head capacity hog operation will produce; 8.5 liters/day x 2,400 head x 365 days = 7,446,000 liters per year. The normal rate of application is 67,373 liters per hectare. Therefore a 2,400 head hog finishing operation will need 7,446,000 liters/67,373 liters per hectare = 110.5 hectares per year upon which to spread the manure. Given that the manure is spread on the land once every 3 years, a total of 331.5 hectares are needed within approximately 3.3 kilometers from the barn. The Value of the Nutrients The value of the manure produced by the finishing enterprise needs to be measured carefully to make sure its true economic benefit is calculated correctly. The first step is to measure the value of the nutrients in a unit of manure, say one thousand liters, as if one were to buy them in the market place. The major nutrients would be nitrogen, phosphorus, potassium, and sulfur. Unfortunately not all the nitrogen and phosphorus are available to a crop in the first year. Some of the nutrients will leach out of the soils, evaporate, or may stay in an inaccessible form for so long that their value is very limited. Table 2 presents an analysis of the value of the nutrients in a typical sample of 1,000 liters of hog manure. Nutrients available to the crop in years 2 and 3 are discounted by 20% per year. The nutrient value per acre using the usual rate of 67,373 liters per hectare is $3.88Cdn/1,000 liters x 67.373 = $261.41Cdn per hectare. Given the analysis in Table 2, the 2,400 head enterprise will produce 7,446,000 liters x $3.88/1,000 liters = $28,890.48Cdn per year. Table 2: The Value ($Cdn) of 1,000 liters of Hog Manure (Nutrients Available in Years 2 and 3 are Discounted by 20% per year)

Usable Fertilizer (kgs) Total Total Nutrients kgs/1,000

liters $/kg Year #1 Year #2 Year

#3 $

Total Nitrogen 3.09 Ammonium N (NH4) 1.9 $0.99 1.90 0.00 0.00 $1.88 Organic N 1.2 $0..99 0.30 0.12 0.05 $0.47 Total Phosphorus (P) 1.0 $0.93 0.50 0.10 0.04 $0.60 Total Potassium (K) 1.4 $0.60 1.40 0.00 0.00 $0.84 Total Sulfur (S) 2.0 $0.93 0.10 0.00 0.00 $0.09 Total $3.88

Source: Tri-Provincial Manure Application and Use Guidelines, Saskatchewan Agriculture & Food Fact Sheets, 2006


291

If the hog finishing enterprise is not associated with a crop enterprise that can take advantage of the nutrients then it can only realize a return based on the willingness to pay for the manure by neighboring crop farmers. Past practice in the industry has indicated that neighboring crop farmers have been willing to pay between $37 - $62 Cdn per hectare. It should also be noted that not all land nor crops grown in Western Canada are able to table advantage of this level of nutrients so the value quoted above should be considered an optimistic number. There is still an additional value calculation of the manure; that being the added value of the crop yield response to the manure application over and above what an equivalent amount of commercial fertilizer can provide. This phenomena deals with the increased crop response to the organic nature of the manure, its supply of other micro nutrients, and that about 1/4 inch of water is also being supplied. This additional value is not included in this study, though research conducted by soil scientists has established this bonus crop response to be a reality (Nagy et al. 2000). The Cost of Applying Hog Manure Industry standards indicate that the cost of injecting hog manure into the soil ranges from $0.00198 - $0.00242Cdn per liter at a rate of 67,373 liters per hectare. This works out to $133.41 – $163.05Cdn per hectare. Lighter rates may also be more economical to the crop on all soil types due to the limited ability of the crop to absorb all the nutrients available and translate them into higher yields for the part of the plant that is desired (Nagy et.al. 2000). If lower rates of application are preferred the costs would likely be more per liter (.0002199 per liter) as the equipment would have to be run longer and more distance would have to be covered resulting in increased fuel costs and wear on equipment.. The Effect of Hauling Distance on Cost The proximity of the application fields to the hog finishing operation will make a difference on the cost of applying the manure. Most manure applicators interviewed said that transporting the manure more than 3.3 kilometers from the EMS site would add significant costs. The Effect of Distance from Weanling Facilities, Feed Mills and Packing Plants The distances that the feed for the hogs and the hogs themselves have to travel will affect the contract finishing enterprise. The hog owners have to absorb these costs and therefore will not want to contract with finishers that are isolated from their weanling facilities and feed mills and their preferred slaughter plants as these distances add to their costs. Most hog owners interviewed felt that potential contract finishers should ideally be within 50 kilometers of the feed mill and weanling facilities. Most also felt that in order for a new hub of hog production to start, the core would have to be at least 3,000 to 6,000 sows and an adequate number of finishing barns in the area (20 – 25 2,400 head finisher barns) along with a feed mill. The distance from the weanling facilities and to the slaughter plant, were not as important as the distance from the feed mill. The cost of hauling the feed and the pigs is paid by the hog owners and therefore is not included in the calculations of the costs and returns associated with the barn enterprise. The Economics of Contract Finishing of Hogs The assumptions used as input into the base simulation for the 2,400 head barn are summarized in Table 3. It should be noted that the manure is valued at $37.00Cdn per hectare or $4,089Cdn per year, labor is charged at $15,000Cdn per year plus benefits of $2,673Cdn and all the capital supplied is in the form of cash or equity and no money is borrowed. The labor is


292

charged as an expense to the barn enterprise. This gives the true rate of return on the investment without the complicating effects of debt financing and labor income The barn is assumed to have a serviceable life of 20 years. The hog feeding contracts are assumed to be 5 years in length and renewable for another five years. At this time no hog owners are willing to commit to longer contracts. The barn owner is therefore going to have to take on the risk that the contract may not be renewed after 5 years and especially after 10 years. In addition the barn owner has to renegotiate the contract amount at each renewal. The assumption here is that the contract amount stays constant at $54.00Cdn per pig space. If the contract is not renewed the barn owner could feed his/her own pigs in the facility, but this entails an entirely different set of risk variables and is not simulated here. The assumption here is that the barn sits empty for the remainder of its life if the contract is not renewed. However, hog owners have said that they are committed to contract finishing rather than owning their own barns because of the huge investment cost of owning all their facilities and their inability to capture the higher value of the manure. Table 3 Assumptions of Base Simulation for 2,400 Head Barn ($Cdn)

Long Term Debt Interest Rate 0.0% Rate of Inflation (expenses) 2.0% Barn Rental - First 5-year contract $/Pig Space

$54

Barn Rental - Second 5-year contract $54 Manure Sales $4,089 Manure Sales Rate of Growth 0.0% Wage Laborers 1 Hourly Wage $15.00 Hours per worker 1,000 Utilities $9,000 Manure Disposal $16,381 Office/Barn Supplies $2,000 Maintenance $5,000 Insurance $5,500 Property Taxes $600 Payment Period 0 Percentage debt 0% Percentage Patient Debt 0% Long Term Debt $0 Patient Debt $0 Owner Equity $672,127

Table 4 presents the IRR for various contract lengths, bonuses, and valuation of manure scenarios assuming no debt. Obviously the barn owner needs to get the contract renewed for at least a second 5 year period for the investment to be viable. In addition, the barn owner needs to strive for the bonus, but more importantly needs to take advantage of the higher valuation of the manure which is based on its nutrient value is $28,891Cdn or $264.41Cdn per hectare per year. It can be seen that valuing the manure at its nutrient value has a significant effect on the


293

results. It is also important to note that only barn owners that also own the surrounding land upon which the manure is spread can capture this benefit.. It should also be noted that the value of the manure ($28,891Cdn) is assumed to be paid in cash from the cropping enterprise to the barn enterprise. However, this may not be necessary as long as the barn does not have any debt. The level of the IRRs presented in Table 4 are lower than the 15 – 20% rates of return usually required on business investments with similar risks. Given the fact that the contract may not be renewed or may be renewed at a lower level contributes to the risk of the barn investment. Taking advantage of the higher manure value is also not guaranteed. Given the current low prices for cereals and oilseeds, many landowners are cutting back on fertilizer rather than increasing. In addition some land may not be suitable for large amounts of hog manure applications. Investments of similar risks should return at least 15% if not higher. It would appear that the barn enterprise will have to rely on financial leverage to realize competitive levels of IRR. Table 4: Percentage IRR by Contract Length, Barn Rental Rate, Bonus, and High Manure Valuation, 0 Debt

5 years 10years 15 years 20 years Barn Rental ($54/pig space/year) -6.4 1.4 5.6 7.3 + Bonus ($1.50/pig space/year) -6.1 2.1 6.3 8.0 + High Nutrient Value of Manure -4.1 6.2 10.2 11.6 + Bonus and High Nutrient Value -3.8 6.9 10.8 12.2

Table 5 presents the IRR resulting from various combinations of interest rates and percentage of debt capital with a 10 year repayment period on the debt. The other assumptions include the standard contract of $54.00Cdn per pig space per year over the 20- year life of the barn, no bonuses, and selling the manure for $37.00Cdn per hectare rather than realizing the nutrient value of the manure. Combinations of interest rates and percentage of debt resulting in at least 1 year of negative cash flows are highlighted in bold. Given current interest rates in the 6% to 8% range, the barn cannot be more that 50% debt financed. Higher percentage debt financing will result in negative cash flows in at least 1 year.

Table 5: Percentage IRR by Interest Rate and Percent Debt, 10 Year Repayment (Bold Indicates Negative Cash Flow)

Interest Rate / % Debt

10% 20% 30% 40% 50% 60% 70% 80% 90%

0% 7.8 8.3 8.9 9.6 10.5 11.7 13.2 15.6 20.1 1% 7.7 8.2 8.7 9.3 10.1 11.1 12.5 14.5 18.1

2% 7.7 8.1 8.5 9.1 9.7 10.6 11.7 13.4 16.2

3% 7.6 8.0 8.3 8.8 9.4 10.1 11.0 12.3 14.4 4% 7.6 7.8 8.2 8.5 9.0 9.5 10.2 11.2 12.7 5% 7.5 7.7 8.0 8.2 8.6 9.0 9.5 10.2 11.2 6% 7.5 7.6 7.8 7.9 8.2 8.4 8.8 9.2 9.8 7% 7.4 7.5 7.6 7.7 7.8 7.9 8.0 8.2 8.5 8% 7.3 7.3 7.4 7.4 7.4 7.3 7.3 7.3 7.3 9% 7.3 7.2 7.1 7.1 6.9 6.8 6.6 6.4 6.2

10% 7.2 7.1 6.9 6.8 6.5 6.3 6.0 5.6 5.2 11% 7.2 7.0 6.7 6.4 6.1 5.8 5.3 4.8 4.3 12% 7.1 6.8 6.5 6.1 5.7 5.2 4.7 4.1 3.4


294

The effect of financial leverage can also be seen in Table 5. The IRR for the 0 debt scenario is 7.3% (Table 4). When debt capital can be secured for less than 7.3% the resulting IRR is higher than 7.3%. When debt capital has to be secured for more than 7.3% the IRR eventually is lower than 7.3%. Even though high IRRs can be attained by higher percentage of debt financing at low interest rates the barn enterprise itself would not be able to cash flow these payments so other sources of cash would have to be used. Table 6 presents the IRR resulting from various combinations of interest rates and percentage of debt capital with a 20 year repayment period on the debt. The other assumptions include the standard contract of $54.00Cdn per pig space per year over the 20 year life of the barn, no bonuses, and selling the manure for $37.00Cdn per hectare rather than realizing the nutrient value of the manure. Combinations of interest rates and percentage of debt resulting in at least 1 year of negative cash flows are highlighted in bold. Given current interest rates in the 6% to 8% range, the barn cannot be more that 70% to 80% debt financed. Higher percentage debt financing will result in negative cash flows in at least 1 year. Table 6: Percentage IRR by Interest Rate and Percent Debt, 20 Year Repayment (Bold Indicates Negative Cash Flow)


10% 20% 30% 40% 50% 60% 70% 80% 90%

0% 8.0 8.9 10.0 11.4 13.3 16.1 20.6 29.1 52.3 1% 8.0 8.8 9.7 11.0 12.8 15.3 19.4 27.2 48.5 2% 7.9 8.6 9.5 10.6 12.2 14.5 18.1 25.1 44.3 3% 7.8 8.4 9.2 10.2 11.5 13.5 16.7 22.8 39.4 4% 7.7 8.2 8.8 9.7 10.8 12.5 15.1 20.3 34.2 5% 7.6 8.0 8.5 9.2 10.0 11.3 13.4 17.5 28.6 6% 7.5 7.8 8.2 8.6 9.2 10.1 11.6 14.4 22.4 7% 7.4 7.6 7.8 8.0 8.3 8.8 9.6 11.1 15.3 8% 7.3 7.4 7.4 7.4 7.4 7.4 7.3 7.2 6.8 9% 7.2 7.1 7.0 6.7 6.4 5.8 4.9 2.8 - 100

10% 7.1 6.9 6.5 6.0 5.3 4.2 2.1 - 2.8 - 100 11% 7.0 6.6 6.1 5.3 4.2 2.3 - 1.2 - 100 - 100 12% 6.9 6.4 5.6 4.5 2.9 0.3 - 100 - 100 - 100

The effect of financial leverage can also be seen in Table 6. The IRR for the 0 debt scenario is 7.3% (Table 4). When debt capital can be secured for less than 7.3% the resulting IRR is higher than 7.3%, even to the point of positive infinity at 100% debt financing. However, when debt capital has to be secured for more than 7.3% the IRR eventually drops below 7.3%, even to the point of negative infinity (speedy bankruptcy) with as little at 70% debt financing at 12% interest rates. This increased financial risk must be considered by potential barn owners when contemplating highly leveraged (debt financed) scenarios. Table 7 presents the IRR resulting from various combinations of interest rates and percentage of debt capital with a 10 year repayment period on the debt, which is comparable to Table 5. However, in the case of Table 7 a 10% patient capital investment is added. The patient capital represents 10% of the barn investment value and is interest free and paid back in equal annual installments in years 5 to 10 of the 20 year simulation. The patient capital repayment needs to be delayed at least 3 years in order for the barn enterprise to establish itself. The other assumptions include the standard contract of $54.00Cdn per pig space per year over the 20 year life of the barn, no bonuses, and selling the manure for $37.00Cdn per hectare rather than realizing the nutrient value of the manure.


295

Table 7: Percentage IRR by Interest Rate and Percent Debt, 10 Year Repayment, With 10% Patient Capital Paid Back In Years 5 to 10, (Bold Indicates Negative Cash Flow)


0% 10% 20% 30% 40% 50% 60% 70% 80%

0% 8.0 8.5 9.1 9.9 10.9 12.2 13.9 16.8 22.6 1% 8.0 8.4 9.0 9.7 10.6 11.7 13.3 15.6 20.3

2% 8.0 8.4 8.9 9.5 10.3 11.2 12.6 14.6 18.3

3% 8.0 8.3 8.8 9.3 9.9 10.8 11.9 13.5 16.3 4% 8.0 8.3 8.6 9.1 9.6 10.3 11.2 12.4 14.5 5% 8.0 8.2 8.5 8.8 9.3 9.8 10.5 11.4 12.8 6% 8.0 8.1 8.4 8.6 8.9 9.3 9.8 10.4 11.3 7% 8.0 8.1 8.2 8.4 8.6 8.8 9.1 9.6 10.0 8% 8.0 8.0 8.1 8.1 8.2 8.3 8.4 8.5 8.7 9% 8.0 7.9 7.9 7.9 7.9 7.8 7.8 7.7 7.5

10% 8.0 7.9 7.8 7.7 7.5 7.3 7.1 6.8 6.5 11% 8.0 7.8 7.6 7.4 7.2 6.9 6.5 6.0 5.5 12% 8.0 7.7 7.5 7.2 6.8 6.4 5.9 5.3 4.6

Combinations of interest rates and percentage of debt resulting in at least 1 year of negative cash flows are highlighted in bold. The first thing to note is the patient capital contribution adds 0.7% (8.0% - 7.3% (Table 4)) to the IRR before any other money is borrowed. Given current interest rates in the 6% to 8% range, the barn still cannot be more that 50% debt financed. Higher percentage debt financing will result in negative cash flows in at least 1 year. However, the patient capital does result in a higher IRR to the barn owner. In the case of the 6% interest rate the advantage is +1.1% (9.3% - 8.2% (Table 4)). This difference is showing the effects of financial leverage. Conclusions There are at least three conditions within the barn enterprise that have to be met in order for it to become economically viable. The first condition is the life of the contract. If the barn has a serviceable life of 20 years, it must be full of pigs for all or most of that time to generate competitive rates of return (15%). Most industries are so risky that no company is going to sign a 20 year contract, but hog owners have indicated that they would rather rely on contract feeders than build the barns themselves. The reasons hog owners give for not wanting to own the barns themselves is because of the huge investment required. Other reasons not expressed as often are negative reaction from communities when many barns are being built and that the rates of return are not competitive. The second condition required to make contract finishing of hogs in Western Canada a viable enterprise is the capture of the nutrient value of the manure. The hog owners have not been able to capture the nutrient value of the manure because the surrounding landowners have not been willing to pay the nutrient value of the manure. However, a contract finisher of hogs can locate the barn in the middle of their own land and take full advantage of the nutrient value of the manure. Though they may not physically transfer funds from their cropping enterprise to their contract finishing enterprise, there is a net value that one of the enterprises does capture.


296

The third condition is financial leverage. If the first two conditions are met and interest rates stay below 8%, contract finishers of hogs can use financial leverage to their advantage. The addition of a 10% patient capital also helps the economic viability of the enterprise. References Nagy, C.N., J.J. Schoenau, and R.A. Schoney. Economic Returns and Hauling Distance of Hog and

Cattle Manure. 2000. Saskatchewan Agriculture and Food. Statistics Handbook, 2005. Saskatchewan Agriculture & Food. Tri-Provincial Manure Application and Use Guidelines, Fact Sheets,

2006 Sask Pork. Sask Pork Annual Report 2003-2004. Saskatoon. 2005


297

QUANTIFYING THE SOURCES OF DAIRY FARM BUSINESS RISK AND UNDERSTANDING THE IMPLICATIONS FOR RISK MANAGEMENT STRATEGIES1

Hung-Hao Chang Department of Agricultural Economics, National Taiwan University.

Richard N. Boisvert and Loren W. Tauer*

Applied Economics and Management, Cornell University, Ithaca, New York, 14851, USA Email: [email protected]

Abstract The major sources of variability of net farm income on individual New York dairy farms over the past 10 years are identified using methods in variance decomposition. The most important source of income variability is the fluctuation in the price of milk received by farmers, followed closely by year-to-year variation in the quantity of feed purchased. The degree of success in engaging in activities that increase diversification that lead to a reduction in the variance in farm income is higher for older farmers and for those that milk in a milking parlor, use recombinant bovine somatotropin, have greater assets per cow, and have engaged in activities to earn income from off-farm sources. Keywords: dairy farm income variability, risk, variance decomposition Introduction It is perceived that dairy farms currently experience greater risks than in past years. Support for this perception is found in data such as farm income per cow of a group of New York dairy farms (Figure 1). During the first half of this period, labor and management income ranged between a loss of $12 per cow to a profit of $240 per cow. In contrast, over the second half of this period, dairy farm income per cow ranged from a loss of $90 to a profit of $430. It is thought that this increased variability is due primarily to increased volatility in milk prices. According to the New York Agricultural Statistics, dairy farmers received an average of $32.47 per hectoliter of milk during the 10-year period ending in 2004; prices ranged from $38.17 to $29.09 per hectoliter, with a coefficient of variation of 0.10. For the prior 10-year period, New York dairy farmers received an average of $29.97 per hectoliter--slightly more than two dollars less. Prices varied over a narrower range—from $28.63 to $33.18 per hectoliter and the coefficient of variation was only 0.05. While recent fluctuations in milk prices explain some of the increased variability in farm income, there are other determinants. For example, over the 10-year period ending in 2004, dairy feed prices averaged $211 per metric ton in New York, but their relative variability (coefficient of variation of 0.10), was as large as that for milk prices (New York Agricultural Statistics). For the earlier period ending in 1994, average dairy feed prices were slightly lower, $189 per metric ton; the coefficient of variation of 0.06 was also much lower. Another factor that could affect income variability, milk production per cow, is much more stable across the State, but for individual farms, it can be substantially more variable.

1 This research was funded by Cornell University Hatch Project 121-7419, Integrated Risk Management Decision Strategies for Dairy Farmers.


298

Figure 1: Average Net Return to Labour and Management per Cow for all Farms Participating in the New York Dairy Farm Business Summary Program

-200

-100

0

100

200

300

400

500

1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005

Year

Ne

t In

co

me

pe

r C

ow

To address these concerns, there have been recent discussions about developing additional financial products and management strategies for reducing risk for agriculture (Dismukes and Durst). To identify the products needed to manage dairy risk effectively one must quantify the important sources of dairy farm income variability. Then farmers can begin to control fluctuating incomes through business and financial management strategies, including hedging or insurance. This paper reports research that quantifies the major sources of income variability on New York dairy farms. Using dairy farm record data, we decompose the variability in net farm income for the 10-year period ending in 2002 into the several components of revenue and cost, accounted for the variability in both the quantity and price associated with each component. We extend this decomposition analysis, by constructing a variable that is the ratio of the variance in farm income divided by the sum of the direct contributions of all components of farm income to variance (e.g. as though these sources are uncorrelated). We regress this variable on characteristics of the dairy operation. Those characteristics that reduce this ratio contribute to a farmer’s success in undertaking production activities to reduce risk by diversifying into activities that are negatively correlated with one another. Method for Decomposing the Variation in Net Farm Returns According to the theory of risk aversion, and mean-variance analysis, risky prospects are evaluated by examining the mean and variance in returns. In comparing alternatives with the same mean, the one with the lowest variance is considered the least risky, and is preferred (e.g. Boisvert and McCarl, 1990). Thus, the variance in returns serves as a measure of risk. The variance in net returns depends on the variability in the quantities of individual inputs and outputs, and on the variability in input and output prices. To isolate the effects of these prices and quantities on this measure of risk, we define net farm income (NFI) as the revenue from selling M commodities, qi (i = 1,…,M), at per unit prices, pi, less the cost of buying N inputs, xj (j = 1,…,N) at unit costs of, cj. Algebraically, NFI is:


299

.),,,(NFI)1(11∑∑

==

−=N

j

jj

M

i

ii xcqpxcqp

Mean NFI is:

)(E)(E][E)2(11

j

N

j

ji

M

i

i xcqpNFI ∑∑==

−= ;

E is the expectations operator. There are two ways to decompose variability in NFI. One way is treat each separate component of revenue and cost as the product of price times quantity (e.g. zi = piqi, and yj = cjxj). Thus, NFI is the sum of M random variables minus the sum of another N random variables, and its variance can be decomposed into:

,222)NFI(Var)3(1 1

,

1

1 1,

1

1 1,

1

2

1

22NFI ∑∑∑∑∑∑∑∑

= =

−

= =

−

= ===

−+++==M

i

N

j

yz

N

j

N

l

yy

M

i

M

k

zz

N

j

y

M

i

z jiljkijiσσσσσσ

where i < k, j < l, and 22222 ,,,,

jiijikiji yzyyzzyzσσσσσ are variances of revenue and cost components, and

covariances between pairs of revenue components, pairs of the cost component, and pairs of revenue and cost components, respectively. This expression isolates the contribution to the variance in NFI of each individual revenue and cost component. Because each component is the product of two variables (only some of which can be controlled by the farmer), equation (3) fails to isolate the contribution of individual quantities or prices to the variance in NFI. To circumvent this problem, Bohrnstedt and Goldberger (1969) derived a linear approximation to the variance of the product of random variables, which has been used to decompose the variance in such things as farm returns and returns to agricultural land (e.g. Burt and Finley, 1968, Boisvert and Bills,1984, and Schmit, et al., 2001).Using this approximation, the variance in NFI is:

( ) ( )[ ] ( ) ( )[ ]

( ) ( ) ( ) ( )

( ) ( ) ( ) ( ) ( ) ( ) ( ) ( )( )

( ) ( ) ( ) ( ) ( ) ( ) ( ) ( )( )

( ) ( ) ( ) ( ) ( ) ( ) ( ) ( )( )

,222

EEEEEEEE2

EEEEEEEE2

EEEEEEEE2

EEEE2

EEEE)4(

1 2,

1

1 2,

1

1 2,

11

1 2,,,,

1

1 2,,,,

1

1 2,,,,

1 1,,

1

2222

1

22222NFI

∑∑∑∑∑∑∑∑

∑∑

∑∑

∑∑

∑ ∑

∑∑

= =

−

= =

−

= ===

= =

−

= =

−

= =

= =

==

−++++

+++−

++++

++++

++

+++=

M

i

N

j

yz

N

j

N

l

yy

M

i

M

k

zz

N

j

y

M

i

z

M

i

N

l

xqjicqjixpjicpji

N

j

N

l

xxljcxljxcljcclj

M

i

M

k

qqkipqkiqpkippki

M

i

N

j

xcjjqpii

N

j

xjcj

M

i

qipi

jiljkiji

jijijiji

ljljljlj

kikikiki

jjii

jjii

RMRMRMRMRM

cpxpcqxq

ccxccxxx

ppqppqqq

xcqp

cxpq

σσσσ

σσσσ

σσσσ

σσ

σσσσσ

where i<k, j<l, E is the expectations operator, and the σ's are respective variances and covariances among components. The first line of (4) gives the direct contributions of qi, pi, xj, and cj, to the variance in NFI; their size depends on the square of the component’s expected value and its variance. The next four lines contain first-order interaction effects between pairs of components-- products of the component’s expected value and the covariance between them. Since the expected values of input and output prices and quantities are positive, each term has the sign of the covariance. If two components move in opposite directions over time, the covariance is negative; if they move in the same direction, the covariance is positive. Where both terms are revenue components (cost components), the variation in revenue (cost) increases if the covariance between the terms is positive, and ceteris paribus the variance in net return increases.


300

The situation is different for terms involving revenue and cost components (terms with a negative 2 in front of the brackets). If the covariance is negative, ceteris paribus cost and revenue move in opposite directions; the variance in NFI increases. Similarly, if the covariance between revenue and cost components is positive, cost and revenue move in the same direction; the variance in NFI is reduced. The last line of (4) contains a set of remainders (RM) that include the interaction effects of higher-order moments of the distribution that cannot be decomposed. If these remainders are small, other terms effectively isolate the individual contributions of prices and quantities to the variance in NFI. To identify the proportional effects of each price and quantity component, we normalize the direct and first-order interaction effects by dividing each term by total variance (Burt and Finley, 1968). The Data

For the analysis it is necessary to have annual data on a number of dairy farms over some period of time. We focus on the dairy farms that participated in New York’s Dairy Farm Business Summary Program each year from 1993 through 2002 (Knoblauch, et al., 2005). These 57 farms represent about one-fourth of the total farms participating in the program during any of these 10 years. These farms are located throughout New York. The ages of the farm operators vary significantly, as does the level of education. Farm operators utilize different milking systems. The average herd size is 270 cows, ranging from 40 to 1,160. Milk production per cow averaged over 8,618 kilograms, and ranged from about 3,629 kilograms to over 12,247 kilograms. For the decomposition, NFI is defined as total receipts minus operating expenses. The sources of income and expenses are: milk sales, cull cow sales, off-farm income, paid labor expenses, and purchased and grown feed expenditures. Fixed costs are not deducted from expenses, but in general year-to-year variations in fixed costs on these farms are small, and typically reflect changes in long term investments rather than annual changes in input and output prices or quantities. Because of its increasing importance, we add income from non-farm sources to our measure of net farm income to identify the extent to which non-farm income reduces variability of income to farm households. Measures of revenue and expenditures are calculated on an accrual basis. To put them on a comparable basis they are converted into constant (1993) dollars. Farm revenues are deflated by the U.S. Index of Farm Prices Received, while farm expenses are deflated by the U.S. Index of Farm Prices Paid. Off-farm income is deflated by the U.S. Consumer Price Index. To abstract from the effects of farm size, data are converted to a per cow basis. After converting to constant 1993 dollars, the NFI across these 57 farms averaged about $1,550 per cow and ranged from $609 to over $3,800. In most farm record systems, data on input quantities and expenses are often reported, but prices are not. To circumvent this problem, unique, implicit output and input prices are estimated for each farm for each year by dividing the deflated receipt or expenditure item by the physical quantity of input used or output sold by that farm. These implicit prices vary significantly across farms (Table 1).2 As an example, the average price paid for purchased feed is just over $97 per metric ton, but the range is from about $78 to $143 per metric ton. Some of this variation in prices may reflect local market conditions, but also the heterogeneity in the quality of labor or other inputs. Since the decomposition of net farm income is conducted separately by farm, problems in not controlling for differences in input quality are likely 2 Since the farm records contain data on the payment for off-farm work but not hours worked, we cannot calculate an implicit price. Thus, the quantity of off-farm work is measured in dollar units so the implicit price is a constant one dollar over all years. Similarly, since only the value of grown feed is reported, its implicit price is unity in all years as well. While these minor limitations in the data don’t allow us to decompose these revenue and expenditure items into their price and quantity components, they do not affect our ability to decompose the other revenue and expenditure components into the price and quantity effects.


301

minimized. The quality of labor or other inputs is likely to be relatively consistent across years for the same farm. Table 1: Major Compnents of Net Farm Income for the Sample of 57 Dairy Farms

Variable

Meana

Standard Deviation

Minimum

Maximum

Receipts ($ per cow)b

Milk Sales 2712.68 487.78 1228.21 4145.36

Cull Cow Sales 132.84 113.57 0.00 2445.49

Off-Farm Income 43.44 111.27 0.00 1073.48

Expenditures ($ per cow)b

Hired Labor 329.89 203.42 0.00 824.84

Purchased Feed 764.76 218.25 87.44 1542.60

Net Return ($ per cow)b 1551.63 382.46 608.95 3821.31

Pricesb

Milk ($ per hectoliter.) 32.27 2.98 24.65 42.45

Cull Cows ($ per kilo.) 0.32 0.11 0.59 0.99

Hired Labor ($ per month) 1781.65 706.89 0.00 8734.01

Purchased Feed ($ per metric ton)

90.44 18.34 78.52 143.00

Quantities

Milk ( kilograms per cow) 8,677 1,403 3,914 12,353

Cull Cows (kilograms per cow) 167.37 136.98 0.00 2917.41

Hired Labor (months per cow) 0.18 0.09 0.00 0.44

Purchased Feed (metric tons per cow)

9.86 3.31 1.28 22.29

Feed Grown ($ per cow) 242.68 104.80 31.90 663.58 a These are the 10-year averages for the 57 farms over the years 1993-2002. b These monetary values are deflated into 1993 constant dollars using the appropriate indices of prices received, prices paid, and the CPI as described in the text.

The Results of the Variance Decomposition The variance in NFI across the 10 years for each of the 57 farms is decomposed according to equation (4). The results are unique by farm. They are summarized in Table 2. Since the component effects are normalized, they sum to unity. Because some of the first-order correlations between components are negative, some direct contributions can be greater than unity. To draw inferences from these results, the linear approximation of the variance, as estimated by the direct contributions of prices and quantities and


302

the first-order interaction effects associated with the decomposition, must be a good approximation of the actual variance. The combined size of the RM terms in equation (4) must be small. Although not reported in Table 2, this is the case. The average absolute error is less than 8 percent. For just over 70 percent of the farms, absolute errors are no greater than 10 percent, and for nearly 90 percent absolute errors are less than 25 percent. Table 2: Normalised Decomposition of the Variance in Dairy Net Farm Income

Item Mean St. Dev. Minimum Maximum

Direct contribution of revenue component

Prices

Milk 1.19 0.73 0.11 4.30

Cull cows 0.01 0.01 0.00 0.03

Quantities

Milk 0.63 0.63 0.07 3.34

Cull cows 0.06 0.18 0.00 1.29

Off-farm work 0.06 0.13 0.00 0.77

Direct contribution of expenditure component

Prices

Labor 0.10 0.10 0.00 0.43

Purchased feed 0.53 0.38 0.06 2.11

Quantitities

Labor 0.09 0.10 0.00 0.57

Purchased feed 0.98 0.79 0.05 4.76

Grown feed 0.11 0.10 0.01 0.47

Indirect contribution of covariance termsa

Two Revenue Components

Milk-P & milk-Q -0.32 0.84 -3.06 0.87

Two Cost Components

Purchased feed-Q & grown feed-Q 0.18 0.24 -0.22 0.82

Purchased feed-P & purchased feed-Q

-1.00 0.79 -4.21 0.08

Purchased feed-P & grown feed-Q -0.16 0.18 -0.56 0.45

Revenue & Cost Componentsb

Milk-P & purchased feed-P 1.01 0.62 0.11 3.03

Milk-P & purchased feed-Q -1.27 1.00 -4.30 0.51

Milk-P & grown feed-Q -0.35 0.32 -1.13 0.04

Milk-Q & purchased feed-Q -0.34 0.61 -2.41 0.86

Milk-P & labor-P -0.18 0.27 -0.91 0.33 a While we report all the direct effects, only the first-order covariance terms greater than 0.15 in absolute value are reported. Thus, the components do not add to unity. b The signs on first-order covariance terms that involve a revenue and cost component implicitly include the (-2) from the 4th line of equation (4).


303

Factors Affecting Variance Directly

Examining the results in Table 2, it is evident that the price of milk, with an average contribution of 1.19, is the revenue component with the largest direct contributor to the variance in net return on these farms. If an effect of this magnitude persists into the future, farmers will likely find strategies to reduce risk such as the forward pricing of milk increasingly desirable and useful. It is also true that the variability in milk output is another component of revenue that contributes directly to the variability in net return. However, its average relative contribution of 0.63 is only about half the size of the contribution of milk prices. The other revenue components (price and quantity of cull cows and off-farm income) make only minor direct contributions to the variance in NFI; on average, the relative contribution to the variance in NFI is 0.06 for both components. This is hardly surprising since the sale of cull cows is primarily a by-product of milk production, and dairy farmers or their spouses typically work less off the farm than on other types of farms. On average, these activities constitute only about 4.6 and 1.5 percent of NFI, respectively (Table 1). Yet, for several farms, the effect of these components on variance in NFI is quite large, especially for cull cows, where the range is from 0 to 129 percent. The likely explanation is that for several farms, production or disease problems necessitated large cattle sales. While these problems are low probability events, there may be an opportunity to deal with them through an insurance product. For expenditures, the average direct contributions to NFI variability of the quantity of feed purchased and the price of purchased feed of 0.98 and 0.53, respectively, dwarf the direct contributions of other components (Table 2). The quantity and price of purchased feed are the third and fourth largest direct contributors to NFI variability, suggesting that forward pricing of purchased feed may be a useful strategy on dairy farms. However, based on the relatively small contribution of grown feed expense to variability in NFI (0.11), there may continue to be little interest among New York dairy farmers in crop insurance; this value, however, reflects grown feed expenditures and not grown feed production. Indirect Contributions to Variability in NFI The previous discussion underscores the importance of revenue and cost components that contribute directly to increased variability in NFI. However, there are important first-order covariance effects whereby the revenue and cost components interact to affect the variance in NFI. If these first-order correlation effects are positive, then the two components vary over time to increase the variance over and above the two separate direct effects. Alternatively, direct effects are tempered through negative first-order correlation effects. It is this type of negative relationship that makes diversification in a financial portfolio or diversification in economic production, sales, or purchase activities such an effective strategy to manage risk. To manage risk in this way, it is often necessary to accept somewhat smaller average return over time. To begin the discussion, the negative covariance effect (-0.32) between milk price and quantity in Table 2 does not reflect a normal production response to price changes, where output price and output quantity should be positively related, but such a response does leads to less variability in NFI. Farmers expand or contract through adjustments to both purchased and grown feed, but through the covariance effect (0.18) ceteris paribus, this leads to an increase in NFI variability. However, the natural opposite movements in the price and quantity of purchased feed (covariance effect of -1.00) tend to reduce the NFI variability, as does the price of purchased feed and grown feed quantities (covariance effect of -0.16). Since milk price is a revenue component and the feed price is a cost component, the positive effect of (1.01) means the components move in opposite directions. Increases in purchased feed prices appear to be accompanied by lower milk prices, leading to increased costs, decreased revenue, and increased variance in NFI. This inverse relationship is unfortunate from a risk management strategy since a natural hedge would exist if higher milk price were accompanied by higher purchased feed prices.


304

Analogously, the negative covariance effect (-1.27) squares with management decisions to purchase more feed when the price of milk is high—presumably to increase milk production; the combined result is a reduction in the NFI variability. The same logic explains the negative covariance effect (-0.35) between milk price and the quantity of feed grown. The negative covariance effect (-0.34) between milk production and purchased feed suggests that these quantities also tend to move in the same direction, serving to reinforce the variance-reduction effects on NFI due to positive correlation between feed use and milk prices. Variance Reduction Estimates A management action on a farm hopefully increases NFI but in the process may also increase its variability. In contrast, when the action is negatively correlated with other net income increasing actions, the variability in NFI falls, even though each individual activity adds to NFI variability in net farm income. This is the essence of diversification in selecting an appropriate portfolio of financial assets, or in selecting a combination of agricultural production decisions, where the negative correlation comes about through the complex interactions between components of revenue and cost. In dairy production, these interactions are captured by first-order covariance terms in Table 2. The effective diversification of dairy operations will differ from farm to farm. However, an implicit measure of this effectiveness is constructed by dividing our estimate of a farm’s variance in NFI by the sum of the direct contributions to income variability—those revenue and cost components in the first two sections of Table 2. This divisor estimates the variance in NFI assuming all components of cost and revenue are uncorrelated. This variable (DIVER) must be non-negative, and is likely to range between zero and unity, but it could exceed unity if there are no negative, but some positive first-order covariance effects. A low value for DIVER reflects successful diversification. One should expect that successful diversification depends on characteristics of the farm and farmer, and on management choices. To identify factors contributing to successful diversification the variable DIVER is regressed on various characteristics of the farm operations (Table 3). These variables are the farm average over the 10 years, except as indicated, such as off-farm income. Some of these factors, such as age and education of the farmer, reflect experience and the potential ability to make decisions. Other variables reflect the characteristic of the farm, such as the type of milking system (parlor or no parlor milking), size of the farm, or location within the state. Some may be proxies for unobserved factors that identify successful managers in terms of tactical and strategic decision making. Since a low value of DIVER reflects successful diversification, the effects of factors associated with good management are expected to be negative.


305

Table 3: Sample Statistics for 57 Farmers (Data used for OLS Model)

Variable Mean Std. Dev. Minimum Maximum

DIVERa 0.35 0.20 0.08 1.07 Operator age (years) 49.01 8.41 32.30 71.50 Operator education (years) 13.52 1.74 10.80 18.00 If parlor milking (any type) is used as milking system (=1)

0.82 0.38 0.00 1.00

Ratio of grown feed expenses to total feed expenses

0.25 0.08 0.06 0.45

If rBST adoption (=1) used on farm; 0 otherwise

0.79 0.41 0.00 1.00

Numbers of cows (1,000) 0.27 0.25 0.04 1.16 Located in western New York 0.39 0.49 0.00 1.00 Asset value per cow ($10,000) 0.69 0.20 0.34 1.63 If the farm household ever received off-farm income (=1)

0.91 0.29 0.00 1.00

a Sum of direct variances terms plus the sum of indirect covariance effects, all divided by the sum of direct variances. Direct variance consists of the components in the first two sections of Table 2. Indirect variance consists of the components in the last sections of Table 2.

From Table 4, the negative sign on a farmer’s age suggests that older farmers are more successful at diversification; for each year of age the DIVER variable decreases by 0.007. Farmers with more of education also appear to be more successful at diversification, although the effect is not statistically significant. This may be in part explained by the fact that years of education is an imperfect measure of educational attainment, or there may be too little variation in the variable to obtain a precise measure of its effect. Although these dairy farmers receive a small fraction of their income from off-farm jobs, those that do appear somewhat more effectively diversified. In contrast, increased farm size, as measured by the number of milk cows, seems to be associated with less effective diversification, but the effect is small, and it is not statistically significant. However, the level of capitalization of the farm, as measured by assets per cow, is also associated with less effective diversification, and this effect is statistically significant.


306

Table 4: OLS Estimation Results

Dependent Variable: DIVERa

Variable Estimate Std. Dev. t-value p-value

Intercept 1.127 0.316 3.570 0.001

Operator age (years) -0.007 0.003 -2.210 0.032

Operator education (years) -0.013 0.017 -0.790 0.435

If parlor milking (any type) is used as milking system (=1)

-0.151 0.068 -2.230 0.031

Ratio of grown feed expenses to total feed expenses

-0.231 0.324 -0.710 0.480

If rBST adoption (=1) is used on farm; 0 otherwise

-0.162 0.068 -2.380 0.021

Numbers of cows (1,000) 0.112 0.130 0.860 0.394

Located in western New York -0.085 0.061 -1.400 0.169

Asset value per cow ($10,000) 0.263 0.128 2.060 0.045

If the farm household ever received off-farm income (=1)

-0.161 0.081 -1.980 0.053

R-Square 0.478

Adj R-Sq 0.379

* 57 observations a Sum of direct variances terms plus the sum of indirect covariance effects, all divided by the sum of direct variances. Direct variance consists of the components in the first two sections of Table 2. Indirect variance consists of the components in the last sections of Table 2.

There are three rather specific management decisions that lead to effective diversification. Farms that milk using a parlor are more diversified, lowering the diversification index by 0.151. The use of recombinant bovine somatotropin is clearly associated with more effective diversification; the estimated parameter is -0.162. By increasing the proportion of feed grown on the farm, a farmer may be somewhat more insulated from fluctuating feed prices. The negative sign on this coefficient appears consistent with this expectation, but the effect is not statistically significant. Conclusions Net farm income varies from year to year, and the sources of that variation over a ten year period for a sample of 57 New York dairy farms are identified using a variance decomposition technique. The single largest source of net farm income variability is the variation in the price of milk, followed closely by the price of purchased feed. However, there was a positive covariance effect between the price of milk and the price of purchased feed, suggesting that if purchased feed prices increase, then milk price decrease that year. None-the-less, there may be opportunities to use insurance or forward pricing tactics to reduce income variability. On the price side, milk price and purchase feed prices are the prices that should be targeted. On the quantity side, milk output and grown feed might be insured. Interesting, although dairy farmers have had crop insurance products for a number of years, including insurance for grown corn silage, corn grain, and hay, the impact of variation in milk output on net income variability is much larger than for grown feed expenditures. Off-farm income is often considered to have a stabilizing effect on


307

income. Although more off-farm income would obviously increase net income, it represents such a small fraction of income for farms in this sample that it has almost no effect on income variability. Diversification on individual farms is measured as the sum of variances terms plus the sum of covariance effects, many of which are negative, all divided by the sum of variances—a measure of variance if all factors are uncorrelated. Regression of this variable, which differs by farm, on farm and farmer characteristics, suggests that age, use of a milking parlor and rBST, and reliance on off- farm income lead to more effective diversification. An older farmer may have a more stable farm operation with less variable income, and off-farm income should reduce income variability. The significance of the use of a milking parlor and rBST in leading to a more effectively diversified dairy operation indicates that the adoption of selected technologies may be effective risk reduction management decisions. References Bohrnstedt, G. W. and A. S. Goldberger. 1969. On the Exact Covariance of Products of Random

Variables. Journal of the American Statistical Association. 64: 1439-1442. Boisvert, R. N. and N. L. Bills. 1984. Variability of New York’s Agricultural Use Values and Its

Implications for Policy. Northeast Journal of Agricultural and Resource Economics. 13: 254-263. Boisvert, R. N. and B. McCarl. 1990. Agricultural Risk Modeling using Mathematical Programming.

Southern Cooperative Series, Bulletin No. 356. A. E. Research Bulletin 90-9, Department of Agricultural Economics, Cornell University.

Burt, O. R. and R. M. Finley. 1968. Statistical Analysis of Identities in Random Variables. American

Journal of Agricultural Economics. 50: 734-744. Dismukes, R. and R. L. Durst. 2006. Whole-Farm Approaches to a Safety Net. Economic Information

Bulletin No. (EIB-15). Economic Research Service, USDA. Knoblauch, W. A., L. D. Putnam and J. Karszes. 2005. Business Summary New York State, 2004.

Research Bulletin 2005-03, Department of Applied Economics and Management, Cornell University.

New York Agricultural Statistics (Annual Bulletins). New York Department of Agriculture and Markets.

10B Airline Drive, Albany, NY 12235. Schmit, T. M., R. N. Boisvert, L. W. Tauer. 2001. Measuring the Financial Risks on New York Dairy

Producers. Journal of Dairy Science. 84:411-420.


308

MILK COMPONENTS AND FARM BUSINESS CHARACTERISTICS: ESTIMATION OF PRODUCTION FUNCTIONS VERSUS A MULTIPLE OUTPUT DISTANCE FUNCTION1

Jaesung Cho and Loren W. Tauer*

Applied Economics and Management, Cornell University, Ithaca, New York, 14851, USA Email: [email protected]

Abstract: The effects of inputs and business factors on the four milk outputs of aggregate milk, butterfat, protein, and other solids were estimated using four individual production functions, and a separate stochastic output distance function, with New York dairy farm data. Results show that 13 independent variables out of 22 display statistically significant effects on the production of at least one of the four milk components. Differential impacts of some inputs on component production indicate that milk component composition can be modified given component prices. Profit increase potentials were computed for inputs. Keywords: farm business characteristics, milk component production, seemingly unrelated regression, multiple output distance function, technical efficiency. Introduction Dairy farmers in New York now receive milk payments under the Federal Milk Marketing Order multiple-component pricing system. Payments are based on the quantities of the three main milk components: butterfat, protein, and other solids. Because the price of each component is determined by the value of that milk component in processing dairy products, and ultimately the prices of final dairy products, component prices vary over time. This provides an opportunity for dairy farmers to increase profits by altering individual milk components in response to component prices. Buccola and Iizuka (1997) estimated hedonic cost models to determine how farmers might respond to component price changes and found little opportunity to adjust components. Bailey et al. (2005) and Smith and Snyder (1978), investigated the economics of milk components by dairy breed, the factor thought most responsible for component composition. However, an important aspect that has been overlooked is the relationship between milk component production and business factors such as farm ownership type, economic scale of the farm, operator labor quality, and intensity of machinery use. Thus, the objective of this paper is to examine the effects of inputs and business factors on the four decomposed milk outputs: aggregate milk, butterfat, protein, and other solids. This is accomplished by estimating individual production functions for the four milk components, and separately a stochastic output distance function using New York dairy farm data. Single-Output Production Functions Estimating separate production functions for a multi-output technology requires the imposition of separability conditions. However, since all of the production factors in milk production are non-allocable, no separability assumption is imposed on outputs and inputs in estimating the four single-output

1 This research was funded by Cornell University Hatch Project 121-7419, Integrated Risk Management Decision Strategies for Dairy Farmers. The authors thank Thomas Overton for his comments.


309

production functions; even though farmers might want to produce only one particular component using a specific production factor, the other milk components will also be produced. The log-linear form of the Cobb-Douglas2 production function used in this study can be expressed as:

(1) ∑ ∑++=k j jijmkikmmomi dxy ,, lnlnln δβα

where miy = the thm output production for farm i, with m = 1 for annual milk yield per cow, m = 2 for

annual butterfat yield per cow, m = 3 for annual protein yield per cow, and m = 4 for annual other solid yield per cow, kix = the thk input, jid = the thj categorical variable, and moα , km,β , and jm,δ are

coefficients. In this model, the coefficient estimates kimikm xy ln/ln, ∂∂=β represent the (partial) production elasticity

of the thk input for the thm output. Thus, if production elasticities of the thk input for the four outputs are identical, each individual component production will change proportionally according to a change in aggregate milk production, so that individual component productions, as percentages of aggregate milk, would be the same regardless of the amount of milk produced. On the other hand, if the production elasticity of each input is different for the four output productions, farmers can alter individual component productions by adjusting inputs. Output Distance Functions Stochastic production frontier models can be utilized in production technology with multiple outputs and inputs by incorporating a distance function (Shephard 1970; Brummer et al. 2002). If dairy farmers maximize outputs due to the difficulties of allocating inputs, the stochastic output distance function is an appropriate specification. The log-linear form of the Cobb-Douglas output distance function for farm i can be expressed as (2) ∑∑∑ +++=

j jijk kikm mimooi dxyD δβαα lnlnln

Because an output distance function is homogeneous of degree one in outputs, the imposition of homogeneity is accomplished by normalizing the outputs by one of the outputs. Hence, butterfat, protein, and other solids are normalized by milk production ( imimi yyy 1

* /= ), so that milk production becomes the

dependent variable and the independent output variables are represented as percentages of each component in milk, resulting in: (3) ∑∑∑ +++=−

j jijk kikm mimoioi dxyyD δβαα lnlnlnln *1

To take into account unobserved random variations, a random error term (vi ) is added to the output distance function. ln Doi is also moved to the right hand side, and replaced with ui where ui > 0. Then, for the purposes of easier comparison between the estimated results of the stochastic output distance function and the previous single-output production functions, the dependent variable in this equation is transformed to a positive iy1ln so that the signs of the estimated coefficients will be reversed,

corresponding to those in a general production function.

2 Since 22 defined independent variables will be used, it precludes the use of more flexible functional forms that require many more estimated coefficients. However, the Cobb-Douglas is satisfactory at estimating production slopes over small data ranges, which is the case here.


310

(4) ∑∑∑ −++++=j iijijk kikm mimoi uvdxyy δβαα lnlnln *

1

This stochastic output distance function has two separate error terms: the symmetric random error term ( iv ), and the one-sided efficiency error term ( iu ). Here iv is assumed to be independently and identically

distributed, ),0( 2vN σ , and iu , representing the technical inefficiency, is assumed to be independently half-

normally distributed, ),0( 2uN σ+ .

The coefficient estimates kk xy ln/ln 1 ∂∂=β from the output distance function measures the increase in

the primary output 1y , holding the output ratios of *my constant. Thus, this estimated elasticity kβ includes

the impact on the other outputs which keep their output ratios constant. However, by using the coefficient estimates kβ and the estimable form of the deterministic output distance function, the (partial) production

elasticity of an input for each individual output can be derived. To obtain this production elasticity, first, rearrange the estimable form3 of the deterministic output distance function as (5) ∑∑∑∑ ++++−=

j jijk kikm mimm imo dxyy )()ln()ln(ln)1(0 1 δβααα

Taking the anti-log of this function generates the equation:

(6) ∑∑= ∏∏+−

j jijkmm m

d

k kim miio exyyδβαα

α )(1)1(

1' , where oeαα ='

0

This multidimensional relationship can be represented by the general transformation function G. (7) 0),,( =ydxG Then, kxy ∂∂ /1 and km xy ∂∂ / can be computed by applying the implicit function theorem.

(8) )//()/(/ 11 yGxGxy kk ∂∂∂∂−=∂∂ ∑ ×+=m kiimk xy )/())1/(( 1αβ

(9) )//()/(/ mkkm yGxGxy ∂∂∂∂−=∂∂ )/()/( kimimk xy×−= αβ

Multiplying Equation (8) by )/( 1iki yx , and Equation (9) by )/( miki yx generates

(10) )1/(ln/ln 1,1 ∑+=∂∂=m mkk

ok xy αββ

(11) mkkm

o

km xy αββ /ln/ln, −=∂∂= , where m ≠ 1, and for mα < 0

These computed values represent partial production elasticities. Equation (10) measures the production elasticity of the thk input kx for aggregate milk, and Equation (11) measures the production elasticity of

the thk input kx for each individual component. However, in a case where a coefficient ( mα ) for an

output ratio ( 1* / yyy mm = ) is positive, the sign of the Equation (11) should be revised as

(12) mkkm

o

km xy αββ /ln/ln, =∂∂= , where m ≠ 1, and for mα > 0.

The reason is that a positive coefficient ( mα ) for an output ratio ( 1

* / yyy mm = ) implies that milk

component levels, displayed as percentages of aggregate milk, increase with aggregate milk production; this is especially true when an increase in an input causes one particular milk component to increase faster than the aggregate milk. In that case, the signs need to be the same (positive) for both the production elasticity of an input for aggregate milk ( kβ ) and the milk component production ( o

km,β ). However,

Equation (11) reveals that o

km,β and kβ have opposite signs because kβ > 0, mα < 0, and o

km,β <0. Hence,

Equation (12) should be used for computing the production elasticity of an input for a milk component production ( o

km,β ) when mα < 0.

3 Equation 4 with the reversed signs of the estimated coefficients (Table 4), corresponding to those in a general production function.


311

The conventional way to examine the relationships between outputs is to simply look at the production possibilities curve in )( / nmyy nm ≠ space. Yet, unlike a business firm where the manager can alter the

inputs used among various outputs, the dairy cow cannot be asked to produce different milk components from a fixed input bundle. So, in this application, a PPC in )( / nmyy nm ≠ space degenerates to a single

point. In other words, neither the movement along the PPC in the )( / nmyy nm ≠ space, nor the elasticity

between the outputs from the PPC in the )( / nmyy nm ≠ space are relevant concepts in this study.

However, by increasing, decreasing, or altering that input bundle, a dairy cow might respond by producing a different output composition. Consequently, the elasticity between the outputs can be obtained by using two output combination points in multidimensional spaces rather than from the PPC. Because my∆ , ny∆ , ny , and my are easily obtained from an old output combination point and a new output

combination point that is generated by a change in inputs, the elasticity between my and ny ( nm ≠ ) can

be calculated by nmyy yynm

ln/ln, ∂∂=ε = )/( nm yy ∆∆ )/( mn yy× .

If the effects of input x on both outputs y1 and y2 are always the same, output y2 would increase proportionally according to an increase in output y1, so that the output ratio (y2/y1) is always constant regardless of the amount of outputs produced. In this case, a new output combination point will be plotted on the ray-line OD that extends out from the origin O through the old output combination point A, as shown in Figure 1. Point B is, thus, the new output combination point. On the other hand, if the effects of input x on outputs y1 and y2 are different, the new production point does not appear on line OD, and, in this case, point C is the new output combination point. This new output combination point C implies that by increasing the input level from xo to x1, farmers are able to alter individual component productions by altering inputs. Figure 1. Two output combination points in y1/y2 space

To compute the elasticity between the four outputs, first rearrange Equation (5) as: (13) ∑∑ −−−++−=

k kkllnnm mom xyyyy )ln(lnlnln)1([ln 1 βαααα

∑−j mjjd αδ /)]( , where lnm ≠≠

The elasticity between aggregate milk and each individual component is simply computed by taking the derivative of Equation (13) with respect to 1ln y .

y2b

y2c

y1c

O

y2a

y1a

C

∆y2ac B

∆y1ac

A

∆y1ab

∆y2ab

y1

y2

D

y1b


312

(14) ∑+=∂∂=m mmmyy yy

mααε /)1(ln/ln 1, 1

On the other hand, taking the derivative of Equation (13) with respect to ln yn generates the elasticity between each individual component as Equation (15). (15) mnnmyy yy

nmααε /ln/ln, −=∂∂=

Data Data were obtained from the New York Dairy Farm Business Summary (DFBS) with 105 participating farms in 2003, and 107 participating farms in 2004. Although 94 of the farms submitted data for both years, estimating production functions with panel characteristics was precluded given only two years of data. Since these data were submitted on a voluntary basis, and the DFBS program is designed to assist farmers in improving their business management skills and accounting and financial analysis techniques, most of the participants are specialized commercial dairy farms. Thus, the average farm size and productivity represented in the data are larger and more productive than the average New York dairy farm. The average farm size at 448 cows is larger than the year 2004 average 95 cows New York dairy farm. Cow productivity of the DFBS farms is also higher at 21,059 lbs4 (= 9,540 kg) of milk per cow compared to the 2004 New York state average of 17,786 (= 8,057 kg). All outputs and inputs are measured on an accrual basis, reflecting what was actually produced and used during the year, rather than what was sold and produced. Aggregate milk and the three individual components – butterfat, protein, and other solids – are included in the model as output variables and are expressed as pounds (= 0.453 kg) per cow during a year. Outliers in the data distributions of butterfat, protein, and other solids may indicate data error, or data recorded under unusual circumstances such as diseases. Therefore, outliers were treated as missing data to prevent the possible distortion of regression results5. Input variables used in this study can be classified into five groups: feed, breed, labor, capital, and other managerial and environmental inputs. Among these inputs, the average herd size, bedding expense per cow per year, machinery cost per cow per year6, BST expense per cow per year, culling rate, bred heifer rate, daily milking frequency, the operator labor quality (age, wage, and education levels), and farm ownership type are considered to be business factors in milk production. Some of these like age cannot be controlled by the farmer, others like education are mostly pre-determined, others like herd size are long-run adjustments, while many expenditures can be quickly adjusted. A year dummy variable is also included to allow for unobserved technical change and environmental aspects such as temperature and sunlight variation between the years. A summary of the variable codes used in estimation of milk component production function is provided in Table 1. Estimation of the Four Production Functions by ISUR The four outputs may be simultaneously affected by non-observable variables which reflect the condition of the cow. Thus, iterative seemingly unrelated regression (ISUR), which takes into account the correlations between the error terms of each equation, was used to estimate the coefficients of the four single-output production functions. The residuals from the system of equations are highly correlated: 0.82 4 0.453 kg 5 Detailed criteria used to define the outliers are butterfat percentage less than 2 percent or greater than 5 percent, protein percentage less than 1 percent or greater than 5 percent, or other solids percentage less than 4 percent or greater than 8 percent. As a result of data correction, six missing data were deleted in each year. 6 This is the sum of expenses per cow for fuel, oil, grease, machinery repairs, vehicle expense, machine hire, machine rent, machine lease, interest (5%), and depreciation.


313

for aggregate milk and butterfat, 0.87 for aggregate milk, and protein, and 0.92 for aggregate milk and other solids, 0.86 butterfat and protein, 0.82 for butterfat and other solids, and 0.88 for protein and other solids Table 2 reports the different (partial) production elasticity of the thk input for the thm output production. This implies that the effect of this input on each output is different, so that farmers are able to alter individual component productions by adjusting this input. Table 1: Description of Variable Names and Sample Statistics *l bs = 0.453 kg **DM = Dry matter ***1 acre = 4046.8m2

Variable names Description Mean Std. Dev.

MILK_COW Milk production per cow per year (in pounds*) 21306.79 (9652kg)

3366.69 (1525kg)

BUTTERFAT_COW Butterfat production per cow per year (in pounds*) 775.25 (351kg)

105.96 (48kg)

PROTEIN_COW Protein production per cow per year (in pounds*) 638.85 (289kg)

94.19 (43kg)

OTHERSOLIDS_COW Othersolid production per cow per year (in pounds*) 1208.12 (547kg)

194.93 (88kg)

YEAR Dummy for year (2003=0 and 2004=1) 105 farms in 2003 107 farms in 2004

COWS Average number of cows on the farm 417.68 447.39

COWS_WKR Average number of cows per worker 38.48 12.61

OPER_AGE Average operator age 49.08 7.61

OPER_EDU Average operator education level 14.05 1.67

OPER_LABOR Average operator labor contribution per cow (in months) 13.28 2.68

WAGE_MONTH Average monthly wage for hired labor 2440.79 883.51

FORAGE_COW Tons of home-grown forage (DM**) per cow per year 8.08 2.58

FORAGE_ACRE Tons of home-grown forage (DM**) per acre** per year 4.07

1.24

CONCENTRATE_COW Expense for purchased concentrate per cow per year 909.17 212.12

ROUGHAGE_COW Expense for purchased roughage per cow per year 43.63 87.13

GENETICS_COW Expense for genetic improvement per cow per year 46.52 26.17

COW_VALUE Average annual cow value (in dollars) 1238.62 168.67

NON-HOLSTEIN The percentages of Non-Holstein herds on the farm 7.98 21.76

CULL_RATE Culling rate 32.07 7.94

HEIFER_RATE Bred heifer rate 22.07 5.97

BEDDING_COW Bedding expense per cow per year 49.74 39.66

MACHINERY_COW Machinery cost per cow per year 590.56 166.86

BST_COW BST expense per cow per year 37.87 37.23

PARLOR Dummy for milking system type (parlor system=1) 165 farms

SOLEOWNER Dummy for farm ownership type (sole owner=1) 89 farms

3ⅹMILKING Dummy for milking frequency (more than two times=1) 96 farms


314

FORAGE_ACRE is positive in each of the four equations. This variable may be a proxy for forage quality provided to a cow. In New York higher yield per acre may represent higher quality alfalfa rather the grass. Dairy concentrate per cow (CONCENTRATE_COW) represents annual dairy concentrates, measured as an expense in dollars, and the coefficient of this variable confirms that increased expenditures for purchased concentrate per cow has major impacts on the four outputs. An increase in dairy concentrate expense leads to a relatively large increase in milk production, especially for butterfat and protein production. The input variables related with genetics and breed, GENETICS_COW and NON-HOLSTEIN have significant effects on the four outputs. The coefficient for genetics per cow (GENETICS_COW) shows a significant impact on all four outputs. The negative coefficient for Non-Holstein breeds (NON-HOLSTEIN) shows that the Non-Holstein breed proportion on a farm results in a decrease in the quantity of aggregate milk and individual components. However, since Non-Holstein breeds produce milk containing higher butterfat and protein content as percentages of aggregate milk, the rates of decrease for butterfat and protein from a one percent increase in the percentage of Non-Holstein breeds are smaller than that of aggregate milk. Among the variables related with the human capital of an operator such as operator age (OPER_AGE), operator education level (OPER_EDU), and operator labor contribution per cow (OPER_LABOR), only average operator age (OPER_AGE) has statistically significant effects on aggregate milk, butterfat, and other solid production, implying that the productivity of a farmer and farmer age has an inverse relationship.


315

Table 2: Estimation Results by ISUR

Single-output production functions

ln y1

(aggregate milk)

ln y2

(butterfat)

ln y3

(protein)

ln y4

(other solids) Est St.Err Est. St.Err Est St.Err Est St.Err

DYEAR -0.0389 0.0118 -0.0574 0.0117 -0.0353 0.0114 -0.0405 0.0132 (0.00) (0.00) (0.00) (0.00)

ln (COWS) -0.0138 0.0153 -0.0046 0.0143 -0.0147 0.0151 -0.0120 0.0163 (0.37) (0.75) (0.33) (0.46)

ln (COWS_WKR) -0.0224 0.0322 -0.0094 0.0303 0.0203 0.0349 -0.0106 0.0352 (0.49) (0.76) (0.56) (0.76)

ln (OPER_AGE) -0.1278 0.0359 -0.0705 0.0377 -0.0543 0.0397 -0.1114 0.0449 (0.00) (0.06) (0.17) (0.01)

ln (OPER_EDU) 0.0462 0.0468 -0.0076 0.0442 0.0526 0.0475 0.0186 0.0534 (0.32) (0.86) (0.27) (0.73)

ln (OPER_LABOR) -0.0061 0.0284 0.0162 0.0315 0.0235 0.0279 0.0090 0.0311 (0.83) (0.61) (0.40) (0.77)

ln (WAGE_MONTH) 0.0567 0.0176 0.0386 0.0173 0.0581 0.0169 0.0542 0.0196 (0.00) (0.03) (0.00) (0.01)

ln (FORAGE_COW) -0.0579 0.0197 -0.0668 0.0187 -0.0590 0.0193 -0.0633 0.0215 (0.00) (0.00) (0.00) (0.00)

ln (FORAGE_ACRE) 0.0879 0.0343 0.0893 0.0293 0.0752 0.0327 0.0858 0.0367 (0.01) (0.00) (0.02) (0.02)

ln (CONCENTRATE_COW) 0.1056 0.0321 0.1063 0.0324 0.1140 0.0345 0.0980 0.0326 (0.00) (0.00) (0.00) (0.00)

ln (ROUGHAGE_COW) -0.0031 0.0038 -0.0062 0.0034 -0.0050 0.0037 -0.0017 0.0041 (0.41) (0.07) (0.17) (0.68)

ln (GENETICS_COW) 0.0552 0.0100 0.0606 0.0079 0.0574 0.0088 0.0608 0.0108 (0.00) (0.00) (0.00) (0.00)

ln (COW_VALUE) 0.0539 0.0449 0.0307 0.0405 0.0152 0.0484 0.0612 0.0491 (0.23) (0.45) (0.75) (0.21)

ln (NON-HOLSTEIN) -0.0433 0.0056 -0.0133 0.0045 -0.0238 0.0048 -0.0395 0.0063 (0.00) (0.00) (0.00) (0.00)

ln (CULL_RATE) 0.0047 0.0294 -0.0046 0.0293 0.0094 0.0286 0.0100 0.0331 (0.87) (0.88) (0.74) (0.76)

ln (HEIF_RATE) 0.0371 0.0201 0.0158 0.0183 0.0353 0.0195 0.0379 0.0224 (0.06) (0.39) (0.07) (0.09)

ln (BEDDING_COW) 0.0102 0.0061 0.0189 0.0056 0.0126 0.0064 0.0126 0.0069 (0.09) (0.00) (0.05) (0.07)

ln (MACHINERY_COW) 0.0816 0.0282 0.0947 0.0270 0.0982 0.0261 0.0924 0.0304 (0.00) (0.00) (0.00) (0.00)

ln (BST_COW) 0.0153 0.0042 0.0129 0.0040 0.0175 0.0040 0.0149 0.0046 (0.00) (0.00) (0.00) (0.00)

DPARLOR -0.0188 0.0245 -0.0467 0.0211 -0.0486 0.0234 -0.0255 0.0270 (0.44) (0.03) (0.04) (0.34)

DSOLEOWNER -0.0125 0.0162 -0.0093 0.0162 -0.0116 0.0158 -0.0176 0.0173 (0.44) (0.57) (0.46) (0.31)

D3ⅹMILKING 0.0941 0.0214 0.0841 0.0200 0.0958 0.0203 0.0971 0.0223 (0.00) (0.00) (0.00) (0.00)

Intercept 8.0622 0.6265 4.8246 0.5613 4.1270 0.5498 5.0272 0.6683 (0.00) (0.00) (0.00) (0.00) The log pseudolikelihood value of fitting constant-only model: 1217.3882

The log pseudolikelihood value of fitting full model: 1463.4251

(P>z)

The negative coefficient for parlor milking system7 (PARLOR) indicates that the parlor milking system has negative effects on butterfat and protein production. However, it is somewhat difficult to conclude that parlor milking system itself negatively affects milk production because milking system type is highly correlated with size and housing type of a farm. In the DFBS data, the correlation between parlor milking

7 Parlor type milking systems include the following: Herringbone which conventional exit (46% of all parlor types), Herringbone which rapid exit (9.9%), Parallel (30.1%), Parabone (4.4%), Rotary (1.1%), and other types of parlor (8.5%).


316

system and freestall barn is 0.89, and those farms have an average of 526 cows. On the other hand, the correlation between stanchion milking system and tiestall barn is 0.86, and those farms have an average of 61 cows. The average monthly wage for hired labor (WAGE_MONTH) measures the effect of hired labor quality, assuming a higher monthly wage reflects higher productivity. This was found to be true because this variable has positive effects on the four outputs. The total machinery cost per cow (MACHINERY_COW) is used as a measurement of equipment quality and non-obsolescence of that equipment, as well as the capital intensity of a farm. This variable was found to have the second most positive impact on milk production. Since bedding increases the comfort level and decreases the stress level of a cow, the bedding expense per cow (BEDDING_COW) coefficient estimates are positive for the four outputs. Superior bedding also provides a clean, dry rest area that helps prevent the spread of infectious diseases. As expected, the coefficient estimates for BST_COW and milking frequency (3×MILKING) are positive and statistically significant in the production of all four outputs. Since some inputs are measured as expenditures in dollars and output prices are available, the effects of inputs on the milk components can be computed as the additional profit (or loss) from a one percent change in the input expenditures. This calculated additional profit represents the profit change from changes in butterfat, protein, and other solids. For instance, the total additional revenue generated using year 2005 average component prices from a one percent increase in expense for genetic expenditure per cow (GENETICS_COW) is $1.81. This is the sum of the revenues from an increase in butterfat, protein, and other solids; each revenue resulting from an increase in an individual component is equal to the component price times the amount of additional output (∆ym), which is computed by taking the average production of each output times the production elasticity (β

m,k) of the input (GENETICS_COW) for that

output. On the other hand, the additional cost for a one percent increase in the input (GENETICS_COW) is $0.47, which is equal to a one percent increase in the average expense for genetic expenditure per cow ($0.47 = $46.52 × 1%). Thus, $1.35 in additional profit is generated by a one percent increase in the expense for genetic expenditure per cow (GENETICS_COW). Since the profit maximization production point is where the total additional revenue equals the additional cost, farmers should spend more until the maximization condition is satisfied. Table 3 shows that a one percent increase in GENETICS_COW, BEDDING_COW, MACHINERY_COW, and BST_COW will generate more profit for farmers; farmers can increase their profits by spending more money on bedding materials for cows, using more BST, buying more or better farm equipment related to milk production, and improving genetic traits of cows.


317

Table 3: Additional Profit from a One Percent Increase in Inputs

Input (xk): GENETICS_COW, Mean xk : $46.52

Output (ym) Mean

(lbs*) βm,k ∆ym (lbs*)

Ave. Price (2005)

Add. Revenue per cow

Total Add. Revenue

per cow

Add. Cost per cow

Add. Profit per cow

Butterfat 750.58

(340kg) 0.0606

0.4548 (0.206kg)

$2.46 $1.12 $1.81 $0.47 $1.35

Protein 618.33

(280kg) 0.0574

0.3549 (0.161kg)

$1.71 $0.61

Other solids 1170.92 (530kg)

0.0608 0.7119

(0.322kg) $0.12 $0.09

Input (xk): BEDDING_COW, Mean xk : $49.74

Output (ym) Mean (lbs*)

βm,k ∆ym (lbs*) Ave. Price

(2005) Add. Revenue

per cow

Total Add. Revenue

per cow

Add. Cost per cow

Add. Profit per cow

Butterfat 750.58

(340kg) 0.0189

0.1419 (0.064kg)

$2.46 $0.35 $0.50 $0.49 $0.01

Protein 618.33

(280kg) 0.0126

0.0779 (0.035kg)

$1.71 $0.13


0.0126 0.1475

(0.067kg) $0.12 $0.02

Input (xk): MACHINERY_COW, Mean xk : $590.56



(2005) Add. Revenue

per cow

Total Add. Revenue

per cow

Add. Cost per cow

Add. Profit per cow

Butterfat 750.58

(340kg) 0.0947

0.7108 (0.322kg)

$2.46 $1.75 $2.92 $0.02 $2.90

Protein 618.33

(280kg) 0.0986

0.6097 (0.276kg)

$1.71 $1.04


0.0924 1.0819

(0.490kg) $0.12 $0.13

Input (xk): BST_COW, Mean xk : $37.87



(2005) Add. Revenue

per cow

Total Add. Revenue

per cow

Add. Cost per cow

Add. Profit per cow

Butterfat 750.58

(340kg) 0.0129

0.0968 (0.044kg)

$2.46 $0.24 $0.44 $0.38 $0.06

Protein 618.33

(280kg) 0.0175

0.1082 (0.049kg)

$1.71 $0.18


0.0149 0.1745

(0.079kg) $0.12 $0.02

*0.453 kg Stochastic Output Distance Function Results The stochastic output distance function (4) is estimated from the maximum likelihood technique, and is reported in Table 4. The results of estimating the stochastic output distance function indicate that 12 out of 22 production factors have statistically significant effects on milk production at the 0.05 level. Since the specification of the stochastic output distance function (4) is different from the four single-output production functions, the interpretations of the coefficient estimates kβ in Table 4 are also somewhat

different from km ,β 8 from the four single-output production functions. The coefficient estimates kβ from

the stochastic output distance function represent the production elasticity of the thk input for the overall output, holding the other inputs and percentages of each milk component *

my constant. Thus, this

estimated elasticity kβ includes the impact on the other outputs which keep their output ratios constant.

However, the resulting coefficients km ,β and kβ are almost identical; since milk components are only a

8 The (partial) production elasticity of the thk input for the particular output my


318

small portion of aggregate milk, the effects of an input on aggregate milk and on overall output are almost the same. Table 4: Stochastic Output Distance Function Estimate

Variable Estimate (βk) Std. Err. z

ln ( *2y ) -0.4904 0.1301 -3.77

ln ( *3y ) -0.2986 0.1620 -1.84

ln ( *4y ) 0.3727 0.1672 2.23

DYEAR -0.0470 0.0116 -4.05 ln (COWS) -0.0091 0.0139 -0.65

ln (COWS_WKR) -0.0078 0.0291 -0.27 ln (OPER_AGE) -0.0808 0.0380 -2.13 ln (OPER_EDU) 0.0422 0.0573 0.74

ln (OPER_LABOR) 0.0056 0.0297 0.19 ln (WAGE_MONTH) 0.0498 0.0170 2.93 ln (FORAGE_COW) -0.0579 0.0219 -2.65

ln (FORAGE_ACRE) 0.0853 0.0292 2.92 ln (CONCENTRATE_COW) 0.1110 0.0208 5.33

ln (ROUGHAGE_COW) -0.0051 0.0037 -1.37 ln (GENETICS_COW) 0.0553 0.0073 7.54

ln (COW_VALUE) 0.0325 0.0440 0.74 ln (NON-HOLSTEIN) -0.0236 0.0060 -3.95

ln (CULL_RATE) -0.0010 0.0191 -0.05 ln (HEIF_RATE) 0.0252 0.0155 1.62

ln (BEDDING_COW) 0.0143 0.0049 2.92 ln (MACHINERY_COW) 0.0934 0.0267 3.50

ln (BST_COW) 0.0143 0.0038 3.74 DPARLOR -0.0379 0.0199 -1.90

DSOLEOWNER -0.0064 0.0146 -0.44 D3ⅹMILKING 0.0866 0.0171 5.07

Intercept 6.3761 0.8388 7.60 σv 0.0640 0.0168 3.80 σu 0.0553 0.0539 1.03

σ²s = σ²v + σ²u 0.0072 0.0039 1.83 λ = σu/σv 0.8646 0.0704 12.28

There appears to be very little technical inefficiency among the New York dairy farms that participated in this DFBS project. The minimum value of estimated technical efficiency is 90% and the average is 96%. These data participants represent high performance farms, so production variation between farms may be relatively small, leading to a high minimum and average efficiency. In addition, the distance function included many business factors such as those representing the economic scale of the farm and the operator labor quality, which possibly affect the technical efficiency level of a farm. In this way, the effects of technical inefficiency are captured in the coefficient estimates, instead of in a one-sided error term.


319

The (partial) production elasticity o

km,β of each input for each of the four outputs is reported in Table 5.

These elasticities are computed by using Equations (10)-(12) and the coefficient estimates for the inputs that have significant effects on milk production. The interpretations of o

km,β are similar to the coefficient

estimates km ,β from the four single-output production functions. The absolute values of o

km,β and km ,β

are slightly different, but the signs and relative effects of the production factors that significantly affect the four output production are almost identical. Thus, the impact of the factors estimated from the distance function is not significantly different from the impacts estimated from the four separate production functions discussed previously. Table 5: Computed Production Elasticities of Inputs for the Four Outputs

ln y1

(aggregate milk)

ln y2

(butterfat)

ln y3

(protein)

ln y4

(other solids)

o

k,1β o

k,2β o

k,3β o

k,4β

DYEAR -0.0806 -0.0959 -0.1575 -0.1262

ln (COWS) -0.0156 -0.0186 -0.0305 -0.0245

ln (COWS_WKR) -0.0133 -0.0159 -0.0261 -0.0209

ln (OPER_AGE) -0.1384 -0.1647 -0.2706 -0.2168

ln (OPER_EDU) 0.0722 0.0860 0.1412 0.1131

ln (OPER_LABOR) 0.0096 0.0114 0.0187 0.0150

ln (WAGE_MONTH) 0.0853 0.1016 0.1668 0.1336

ln (FORAGE_COW) -0.0993 -0.1181 -0.1940 -0.1555

ln (FORAGE_ACRE) 0.1462 0.1740 0.2858 0.2290

ln (CONCENTRATE_COW) 0.1901 0.2263 0.3716 0.2977

ln (ROUGHAGE_COW) -0.0087 -0.0103 -0.0169 -0.0136

ln (GENETICS_COW) 0.0947 0.1127 0.1851 0.1483

ln (COW_VALUE) 0.0557 0.0663 0.1089 0.0873

ln (NON-HOLSTEIN) -0.0404 -0.0481 -0.0790 -0.0633

ln (CULL_RATE) -0.0017 -0.0020 -0.0033 -0.0026

ln (HEIF_RATE) 0.0432 0.0514 0.0844 0.0676

ln (BEDDING_COW) 0.0245 0.0292 0.0480 0.0384

ln (MACHINERY_COW) 0.1600 0.1904 0.3127 0.2505

ln (BST_COW) 0.0246 0.0292 0.0480 0.0385

DPARLOR -0.0650 -0.0773 -0.1270 -0.1018

DSOLEOWNER -0.0110 -0.0130 -0.0214 -0.0172

D3ⅹMILKING 0.1484 0.1766 0.2900 0.2323


320

Finally, elasticities between the outputs computed by using Equations (14) and (15) are presented in Table 6. Table 6: Elasticities between Outputs

y2

(butterfat) y3

(protein) y4

(other solids)

y1

(aggregate milk) -0.8402 -0.5116 0.6385

y2

(butterfat) -0.6089 0.7600

y3

(protein) 1.2482

Conclusions This study measured the responses of aggregate milk and individual milk component production to changes made in the dairy business. Four single-output production functions and a stochastic output distance function were estimated using New York Dairy Farm Business Summary (DFBS) data from 105 farms in 2003 and 107 farms in 2004. The empirical results demonstrate the possibility of altering individual component productions. However, since the differences between the effects of each input on each output are relatively small, the farmer’s ability to alter individual component productions may be limited. Yet, this is still important because, given the small profit margins that often occur in the dairy industry, this small ability provides the opportunity for farmers to increase profits by altering individual component production levels in response to each component price. References Bailey, K.W., Jones, C.M., and Heinrichs, A.J., 2005. Economic Returns to Holstein and Jersey Herds Under Multiple Component Pricing. Journal of Dairy Science, 88 (6), pp.2269-2280. Brummer, B., Glauben, T., and Thijssen, G., 2002. Decomposition of Productivity Growth Using

Distance Functions: The Case of Dairy Farms in Three European Countries. American Journal of Agricultural Economics, 84 (3), pp.628-644.

Buccola, S., and Iizuka, Y., 1997. Hedonic Cost Models and the Pricing of Milk Components. American

Journal of Agricultural Economics, 79 (2), pp.452-462. Shephard, R.W., 1970. Theory of Cost and Production Functions. Princeton: Princeton University Press. Smith, B.J., and Snyder, S.D., 1978. Effects of Protein and Fat Pricing on Farm Milk Prices for the Five Major U.S. Dairy Breeds. American Journal of Agricultural Economics, 60 (1), pp.126-131.


321

PRODUCTIVITY AND FARM SIZE

Paul Clark and Michael Langemeier* Dept. of Ag. Econ.

346 Waters Hall, Kansas State University, Manhattan, KS 66506 Email: [email protected]

Abstract The objective of this paper was to examine productivity differences among individual farms. The Malmquist index approach was used to estimate productivity for each farm and to decompose productivity change into technical change and efficiency change. The relationship between productivity and farm size was explored. In addition, the relationships between each productivity component and outputs, and each productivity component and inputs were explored. For the sample of Kansas farms used in this study, average annual productivity change over the 20-year period for the sample of farms was 2.16%. Technical change averaged 1.54% per year and efficiency change averaged 0.61% per year. Productivity was significant and positively related to farm size and feed grain production, and significant and negatively related to labor use. The largest farms, those with real gross farm income greater than $500,000, had the largest annual average productivity change at 3.20%. Keywords: productivity, competitive advantage, farm performance

Introduction Productivity measures the quantity of outputs of a production process relative to the level of inputs. The more output resulting from a given level of input, the more productive the process. Productivity growth has been a relatively constant feature of U.S. agriculture. Output increases relative to input use have allowed fewer farmers to produce increasing amounts of commodities on a relatively constant or declining land base. Annual output growth for U.S. agriculture was 1.76% from 1948 to 2002 (United States Department of Agriculture 2005). Rather than growth in inputs, almost all of this output growth was due to an increase in productivity. Productivity growth enabled farms to increase outputs in relation to inputs or improve the output/input ratio. By the end of the twenty-year period 1982 to 2002, 5% fewer U.S. farms were farming 5% more hectares. Kansas farm numbers over the same period show a similar trend, falling by 12% from 73,315 farms to 64,414 farms while hectares in farms remained relatively constant at around 19.1 million hectares (Kansas Department of Agriculture, 2004 Kansas Farm Facts). Clearly, U.S. and Kansas farms became more productive in general over this time period. Fewer people managing more total hectares is a continuation of a historical trend in U.S. agriculture. People fed per farm worker has increased from 15.3 people fed per farm worker in 1950 to 103 people fed per farm worker in 1998 (Hallberg 2001). Previous research has focused on the measurement of productivity at the state or country level (e.g., Ball et al. 1997; Arnade 1998; Ball et al. 2004). Research that examines productivity differences among farms is sparse. This research would be useful in understanding the structure of agriculture. Specifically, this research could be used to determine the competitive position of individual farms or groups of farms. This paper examines productivity differences among individual farms in Kansas. The Malmquist productivity index is computed for each farm. The Malmquist productivity index is then decomposed to


322

measure technical change and efficiency change for each farm. Differences in productivity, technical change, and efficiency change indices among farm size groups are presented and discussed.

Productivity Indices Productivity indices, technical change indices, and efficiency change indices are computed in this paper using the input based Malmquist index approach. As discussed by Färe et al. (1994) and Färe and Grosskopf (1996), with this approach distance functions and linear programming are used to estimate Malmquist indices for each pair of years. The advantage of using this approach to estimate productivity is that it does not impose a functional form on the underlying technologies. The Malmquist productivity index for each pair of years was decomposed into a technical change component and an efficiency change component. Technical change (TECHC) represents a shift in the production frontier and enables farms to produce more output with the same level of inputs or the same output with a lower level of inputs. Efficiency change (EFFC) involves a movement towards or away from the production frontier. If a farm exhibits positive efficiency change, they are said to be catching up. Positive efficiency change would enable a farm to have an output/input ratio that is similar to the most efficient farms or those on the production frontier. Improvements in productivity over time yield Malmquist indexes greater than one. Deterioration in productivity results in a Malmquist index that is less than one. Similarly, improvements in the TECHC and EFFC components of the Malmquist index are also associated with a value of one and deterioration less than one. While the product of the TECHC and EFFC must equal the Malmquist index, these components can be moving in different directions (Färe et al. 1994; Färe and Grosskopf 1996). Productivity indices were summarized by farm size group. The farms were categorized into groups using the following farm size categories: those with an average annual real gross farm income (rgfi) less than $100,000; those between $100,000 and $250,000; those between $250,000 and $500,000; and those farms with an average annual real gross farm income greater than $500,000. To further examine the relationship between productivity and farm size, the following regression was used: ln prodi = α + β(ln rgfi) (1) where ln is the natural logarithm, prodi is the Malmquist productivity index, and rgfi is real gross farm income. The β coefficient in this regression represents an elasticity which can be used to examine the sensitivity of productivity to changes in farm size. Regression analysis was also used to examine the relationship between each productivity component and outputs, and each productivity component and inputs. To ease interpretation of these regressions, outputs and inputs were normalized using real gross farm income.

Data Data collected and maintained by the Kansas Farm Management Association (KFMA) from 195 Kansas farms that had continuous data from 1984 to 2003 were utilized in this study. The KFMA database for 2003, from which the 20-year continuous member subset comes, contains 2,370 variables per farm for approximately 2,000 farms (Langemeier 2003). For this study, six outputs, small grain income (wheat),


323

feed grain income (afg), oilseed income (oil), hay and forage income (ahay), beef income (beefi), and other income (otheri) were used. Other income includes crop insurance proceeds, machine hire, farm program payments, and other miscellaneous income such as patronage dividends. Output quantities were derived by dividing production values by the appropriate price. Prices were collected from the Kansas Department of Agriculture (Agricultural Prices, various issues). Wheat prices were used for small grains, corn prices for feed grains, soybean prices for oilseeds, the price of all beef for beef, and the all hay price for hay and forage. For inputs, purchased inputs (pinputs), capital inputs (capital), and total labor (tlabor) were used. Purchased inputs include feed, seed, insurance, fertilizer, and chemicals. Capital includes interest, depreciation, repairs, fuel, and land. Total labor (i.e., workers per farm) includes hired labor and unpaid operator and family labor. Aggregate input prices were used to create implicit input quantity indices for capital and purchased inputs (United States Department of Agriculture, Agricultural Prices, various issues). Information pertaining to real gross farm income (rgfi) and total hectares (tacres) was collected. Rgfi includes all farm income deflated by the implicit price deflator for personal consumption expenditures (United States Department of Commerce). Total hectares include all hectares, cropland and pasture, owned and rented by the farm. Summary statistics for the 195 farms are displayed in Table 1. For the 195 farms, real gross farm income averaged $232,236 over the 20-year period 1984-2003 with a maximum of $850,337 and a minimum of $33,877. Total hectares averaged 668 hectares with a 20-year average minimum of 86 hectares and a maximum of 2,287 hectares. Output variables had 20-year averages for all farms of 319 metric tones (M/T) for wheat, 474 M/T for feed grains, 137 M/T for oilseeds, 69 M/T for all hay, and 25 M/T for all beef. Other income, which includes crop insurance proceeds, machine hire, farm, program payments, and other miscellaneous income, averaged $39,745. Purchased inputs, capital, and labor averaged 69,241, 100,892, and 1.46, respectively. Table 1: Summary Statistics for Sample of 195 Kansas Farms.a

Variable Unit Average Max Min Std. Dev. Real Gross Farm Income

$ 232,236 850,337 33,877 150,743

Total Hectares Hectares 668 2,287 86 365 Wheat M/T 319 1,624 5 270 Feed Grains M/T 474 2,575 1 507 Oilseeds M/T 137 1,123 -0- 185 Hay and Forage M/T 69 1,566 -0- 152 Beef M/T 25 279 -0- 37 Other Income $ 39,745 247,343 4,608 30,459 Purchased Inputs Index 69,241 435,609 8,224 57,494 Capital Inputs Index 100,892 35,8214 18,593 62,998 Total Labor Workers 1.46 6.56 0.40 0.07

a 20-year averages.


324

Table 2 summarizes farm characteristics for farms grouped by real gross farm income. Approximately 77% of the farms had a real gross farm income between $100,000 and $500,000. The average real gross farm income for the largest farm size group was $663,813 and ranged from $533,843 to $850,337. Table 2: Characteristics of Farms by Farm Size Category

Farms < $100k

Farms $100k to $250k


Farms > $500k

# of farms 34 96 55 10 Rgfi $76,370 $172,625 $354,169 $663,813 Max Rgfi $99,682 $245,231 $499,217 $850,337 Min Rgfi $33,877 $101,362 $251,327 $533,843 Hectares 282 626 898 1,121

Results Average productivity change for the 195 farms for the 20-year period was 2.16%. The largest farms, those with real gross farm income greater than $500,000, had an average annual productivity change of 3.20% (Table 3). The smallest farm group, those with real gross farm income less than $100,000, had an average annual productivity change of 2.13%. The two middle groups of farms, those with real gross farm income between $100,000 and $250,000 and those with real gross farm income between $250,000 and $500,000, had average annual productivity changes of 1.61% and 2.96%, respectively. Table 3: Productivity Measures Farm Size Category

Farms < $100k



Farms > $500k

Number 34 96 55 10 Rgfi $76,370 $172,625 $354,169 $663,813 Hectares 282 626 898 1,121 TECHC 1.0130 1.0115 1.0211 1.0305 EFFC 1.0082 1.0045 1.0083 1.0015 PRODI 1.0213 1.0161 1.0296 1.0320

TECHC – technical change EFFC – efficiency change PRODI – Malmquist productivity change index The fact that the group with average real gross farm income between $100,000 and $250,000 had the lowest productivity change index is interesting. A farm at the upper end of the real gross farm income range for this group would generate, on average, enough income to support one farm family. However, if the farm operator of a farm in this group has lower productivity than farms in the largest two farm size groups, he or she has a competitive disadvantage and thus will need to make a decision on whether he or she should decrease farm size and become a part-time farmer, or increase farm size to augment productivity. If the productivity remains relatively low for this group, it will become increasingly difficult for these farms to cover family living expenditures. The Malmquist index was decomposed into technical change (TECHC) and efficiency change (EFFC). Technical change averaged 1.54% per year and efficiency change averaged 0.61% per year for the sample of farms. For the largest farms, TECHC averaged 3.05% and EFFC averaged 0.15%. The middle groups of farms had TECHC of 1.15% and 2.11%, and EFFC of 0.45% and 0.83%, respectively. The smallest


325

farms averaged 1.30% for TECHC and 0.82% for EFFC. These decompositions suggest that technical change was, on average, a larger contributor to productivity change for the sample of farms. This was particularly true for larger farms for which almost all of the productivity change was due to technical change. The regression examining the relationship between productivity and farm size for the entire sample of farms yielded an estimated β of 0.0072. This parameter estimate was statistically significant at the 5% level. This would suggest that for all 195 farms in the sample, over the 20-year period, a 1% increase in real gross farm income resulted in a 0.0072% increase in productivity. A doubling of farm size (increasing average farm size from $232,236 to $464,472) would result in a 0.72% increase in productivity. As illustrated in Table 4, regressions examining the relationship between productivity and farm size were run for each farm size category. The estimated coefficients remained relatively small in magnitude in all four cases. None of these parameter estimates were statistically significant at the 5% level. Table 4: Regressions Examining the Relationship Prodivtivity and Farm Size

Farms < $100k



Farms > $500k

β‡ 0.0219 0.0070 0.0046 -0.0041

t-stat 1.4191 0.6037 0.2979 -0.0733 Adj-R2 0.0298 -0.0067 -0.0172 -0.1242

‡ Regression coefficient on the natural logarithm of real gross farm income * Significant at the 5% level ** Significant at the 1% level Further analysis was done to assess the impact of outputs and inputs on the PRODI measure, and on TECHC and EFFC. These results are summarized in Table 5, Table 6, and Table 7. Table 5 presents the results for productivity. Table 6 and Table 7 present the results for technical change and efficiency change, respectively. Productivity was significantly related to feed grain production and labor use (Table 5). Farms that increased the proportion of feed grain income to gross farm income were relatively more productive. Farms that used relatively less labor in proportion to gross farm income were relatively more productive. This result reveals the importance of labor efficiency improvements to productivity growth.


326

Table 5: Regressions Examining Relationship between Productivity, Outputs, and Inputs

β‡ t-stat Adj-R2

Prodi-outputs† 0.122466

Wheat 0.003426 1.587169 Afg 0.006457

3.132454**

Oil 0.001277 1.099791 Ahay -0.001110 -0.691965 Beef -0.001580 -1.178818 Prodi-inputs† 0.041138 Tlabor -0.014010 -2.785491** Pinputs 0.012644 1.611806

Capital 0.016505 1.759575

† Outputs and inputs normalized by real gross farm income ‡ Prodi regressed on normalized outputs and inputs * Significant at the 5% level ** Significant at the 1% level Table 6 presents the results of the regression analysis examining the relationship between technical change, and output and input mixes. Farms with higher levels of feed grain and oilseed production in relation to all other outputs exhibited higher levels of technical change. These results suggest that technology (e.g., adoption of no-till practices) was biased towards feed grains and oilseeds. All three inputs were significantly related to technical change. The input results in Table 6 suggest, in general, that technology was biased towards capital and purchased input use. Table 6: Regressions Examining Relationship between Techinical Change, Outputs, and Inputs

β‡ t-stat Adj-R2

TECHC-outputs† 0.211211

Wheat -0.000310 -0.156184 Afg 0.001411 2.177326* Oil 0.004573 2.877649** Ahay -1.59E-03 -0.916828 Beef -0.000210 -0.128979 TECHC–inputs† 0.144869 Tlabor -0.013910 -4.132480** Pinputs 0.018695 2.638690** Capital 0.016103 2.620173**

† Outputs and inputs normalized by real gross farm income ‡ TECHC regressed on normalized outputs and inputs * Significant at the 5% level ** Significant at the 1% level

The results of the regression analysis examining the relationship between efficiency change, and output and input mixes are presented in Table 7. Efficiency change was positively related to wheat production, thus farms with higher levels of wheat production moved towards the production frontier. There was not a significant relationship between efficiency change and any of the inputs. This result means that farms


327

that were moving towards the frontier did not use relatively more or less of any input. This result also suggests that the results with respect to productivity and labor discussed above were due to technical change rather than efficiency change. Table 7: Regressions Examining Relationship between Efficiency Change, Outputs, and Inputs

β‡ t-stat Adj-R2

EFFC-outputs† 0.044100

Wheat 0.003734 2.733881** Afg 0.001884 1.358308 Oil -0.000130 -0.127460 Ahay 0.000483 0.513618 Beef -0.001360 -1.959407 EFFC –inputs† -0.013780 Tlabor -0.000100 -0.024562 Pinputs -0.003460 -0.624718 Capital -0.002190 -0.358905

† Outputs and inputs normalized by real gross farm income ‡ EFFC regressed on normalized outputs and inputs * Significant at the 5% level ** Significant at the 1% level Summary and Implications Productivity measures the quantity of outputs of a production process relative to the level of inputs. The more output resulting from a given level of input, the more productive the process. Productivity growth has been a relatively constant feature of U.S. agriculture. Output increases, relative to input use, have allowed fewer farmers to produce increasing amounts of commodities on a relatively constant or declining land base. Productivity measures for a sample of KFMA farms that had continuous data for the period 1984-2003 were computed in this study. Annual average productivity change over the 20-year period for this sample of farms was 2.16%. Productivity increased by 0.0072% for every 1% increase in real gross farm income. The largest farms, those with real gross farm income greater than $500,000, had the largest annual average productivity change at 3.20%. When regressed against outputs, feed grain production had a statistically significant and positive impact on productivity, while labor use was negatively related to productivity. These results suggest that productivity increased as farms added more feed grains and reduced labor relative to gross farm income. Productivity was decomposed into a technical change component and an efficiency change component. Technical change averaged 1.54% per year and efficiency change averaged 0.61% per year for the sample of farms implying that most of the gains in productivity came through technological improvements rather than through gains in efficiency. In contrast to the small farms for which 38% of productivity was attributed to efficiency change, only 5% of productivity change for the largest farm size group was attributed to efficiency change. Thus, technical change played a much larger role in overall productivity change for the large farms. This study has implications for the structure of agriculture. The productivity growth for the large farms was substantially higher than the productivity growth for the small farms. Also, technical change, a major


328

component of productivity growth, was substantially higher for large farms. These results suggest that large farms have a competitive advantage and that consolidation will continue to be a major force impacting Kansas agriculture.

References Arnade, C 1998, ‘Using a programming approach to measure international agricultural efficiency and

productivity’, Journal of Agricultural Economics, vol. 49, no. 1, pp. 67-84. Ball, VE, Bureau, JC, Nehring, R & Somwaru, A 1997, ‘Agricultural productivity revisited’, American

Journal of Agricultural Economics, vol. 79, no. 4, pp. 1045-1063. Ball, VE, Hallahan, C & Nehring, R 2004, ‘Convergence of productivity: an analysis of the catch-up

hypothesis within a panel of states’, American Journal of Agricultural Economics, vol. 86, no. 5, pp. 1315-1321.

Färe, R, Grosskopf, S, Norris, M & Zhang, Z 1994, ‘Productivity growth, technical progress, and

efficiency change in industrialized countries’, American Economic Review, vol. 84, no. 1, pp. 66-83.

Färe, R. & Grosskopf, S 1996, Intertemporal production frontiers, Kluwer Academic Publishers, Boston. Hallberg, MC 2001, Economic trends in U.S. agriculture and food systems since World War II, Iowa State University, Ames, IA. Kansas Agricultural Statistic Service 2004, 2004 Kansas farm facts, Kansas Department of Agriculture,

Topeka, KS. Kansas Agricultural Statistic Service various years, Agricultural prices, Kansas Department of

Agriculture, Topeka, KS. Langemeier, MR 2003, ‘Kansas farm management, SAS data bank documentation’, Contribution No. 03-

420-D, Kansas Agricultural Experiment Station, Kansas State University, Manhattan, KS. Economic Research Service 2005, Agricultural productivity in the United States, United States

Department of Agriculture, Washington, D.C. (www.ers.usda.gov/data/agproductivity). National Agriculture Statistics Service various years, Agricultural prices, United States Department of

Agriculture, Washington, D.C. Bureau of Economic Analysis various issues, Personal income and outlays, United States Department of

Commerce, Washington, D.C.


329

A MODEL TO EVALUATE THE FEASIBILITY OF GM AND NON-GM CO-EXISTENCE IN EUROPE AT FARM AND COLLECTION FIRM LEVEL FOR MAIZE.

François Coléno

INRA UMR 1048 SAD-APT Bâtiment EGER, Site de Grignon, BP 1 78850 Thiverval-Grignon France.

Email : [email protected]

Abstract GM and non-GM coexistence, as defined by the European commission, defines a product as non-GM if it contains less than 0.9% of GM material. To avoid the risk of mixing GM and non-GM, we made a model of supply chain management rules and strategies for crop collection planning for a small farming region. It simulates (i) the GM and non-GM proportions at the end of the supply chain and (ii) the transportation and processing costs. Three strategies were evaluated. One has no specific planning. Batches were taken as crops arrived at the silo. A second is on a spatial basis and allocates each silo to a single crop batch. A third is time based and allocates part of the collection time to one product. We show that the spatial strategy allows all the non GM production to be segregated, but at a high cost. On the contrary the time strategy leads to a lower cost but with lower segregation results. Keywords: GM, supply chain management rules, maize Introduction The introduction of genetically modified (GM) crops into Europe has lead to conflict between supporters and opponents of the use of this technology (Levidow et al. 2000). These positions lead to an informal moratorium (since abandoned) on GM use and to several regulations for co-existence between GM and conventional crops, labelling and traceability. This regulation aims to guarantee that conventional production will not become mixed with GM crops. At industrial level, these regulations ensure traceability of GM products at each step of the supply chain, from farmers to consumption. Labelling of GM presence is needed as soon as a product contains more than 0.9% of GM (Arvanitoyannis et al. 2006; Beckmann et al. 2006; Jank et al. 2005). Concerning agricultural production the co-existence generates several problems. On a farm, use of the same agricultural machinery, such as a seed drill or harvester, for both GM and conventional production, increases the risk of admixture (Jank et al. 2006). Moreover, a farmer using GM seed has to be sure that his fields will not pollute the conventional production of his neighbours. To do so, it is recommended to have a buffer distance between GM and non GM fields (Byrne and Fromherz 2003), and to have time lag between GM and non GM production in order to minimise the risk of pollution (Angevin et al. 2005). At the industry level, the problem is to guarantee the absence of GM using the PCR test (Lüthy 1999) and using risk management policies. In this case use of HACCP to identify the critical point in the supply chain is recommended (Scipioni et al. 2005). Concerning agricultural territory, i.e. a small region of several square kilometres, the problem is to propose governance practices that minimise contamination between fields (Byrne and Fromherz 2003), and that allow collection firms to collect the two types of product. These firms have to segregate these two products using the existing infrastructure. This constraint obliges them to combine, in their collection silos and maize dryers, the production of several dozens of fields. Co-existence between GM and non-GM production leads to questions about agricultural production and about transformation and transportation of


330

these products. To answer the first set of questions, several agronomic models of gene dissemination were developed. These models allow the pollution risk from GM fields to be evaluated, taking into account spatial aspects and production system configurations (Angevin et al. 2003). Concerning the second set of question works of Le Bail (2003) identified the critical points in the collection chain. These critical points were concerned with cropping plan management, storage of harvested products and, in the case of maize, drying, which is a bottleneck in maize collection. At collection level it is possible to consider two different logics in order to ensure chain segregation. The first consists of using optimal collection scheduling and planning at the different stages of collection (Entrup et al. 2005) to ensure GM and non-GM segregation without any central organisation of product delivery from farmers. A second is a centralisation of the collection decision by organising farmers delivery during the collection period or in the collection territory. This second solution was proposed to manage GM and non-GM segregation where there was a low proportion of GM production. (Le Bail and Valceschini 2004). For a higher proportion of non-GM, such as one third of the whole maize, studies made in collaboration with French collection firms have identified some collection organisation strategies (Coléno et al. 2005). These strategies are based on : The separation of the two products in space, giving one chain to each production. So each collection silo receives only one type of product. Dryers are allocated to one type of product. The separation of the two products by the timing of the collection period. In this case, each product is delivered to the nearest collection silo to the farm, but at a specific time. Thus, GM can be delivered in the beginning of the collection period and the non-GM at the end. In this paper we propose to evaluate these management logics of decentralisation and centralisation using a simulation model of flow in the supply chain of the collection firm for a large proportion of non-GM collected. Concerning the centralisation logic we will take into account the two strategies of segregation in space and time. After presenting the model we will evaluate the different strategies using two criteria: the collection cost and the proportion of non-GM that is stored as non-GM at the end of the collection process. Presentation of the Model Maize collection in Europe occurs in autumn - generally from September to December. During this period, farmers harvest their maize and deliver it to the collection silos of the firm purchasing their harvest. Each of these silos is made up of different cells, all of the same size. The cells are small compared to the quantity of maize collected. Very often, maize is transferred from collection silos to dryers. When maize is dried it is stored in uniform batches in storage silos in harbours or railway stations. These storage silos may contain 300 000 tons or more. To ensure a high quality of maize, and hence access to the best food markets, the maximum time from harvesting to drying should be less than 48 hours. To ensure GM and non-GM segregation in the collection chain, several factors have been shown to be important. (Coléno et al 2005; Le Bail 2003) : Mixing of products can occur in the collection silos. When all the cells contain maize the silo manager has to choose between (i) accepting farmers’ deliveries and so mixing the two products and (ii) refusing some deliveries to avoid mixing but with the risk that the farmer will sell his crop to another firm. The type of relationship between the firm and the farmer, and whether there is another collection firm in the territory will influence the silo manager’s decision.


331

Mixing may also occur in the dryers. To reduce drying costs, dryers are used at their full capacity. In so doing, mixing may occur if there is not enough of one product. Moreover, to avoid contamination between products in the dryer, the first batch of non-GM that follows a GM lot must be sold as GM. The model deals with these two critical points and takes into account transport between collection silos and dryers. It is therefore made up of three modules: collection silos, dryers and transport. In order to take into account the decentralised logic we will consider two schedulings of collections - silos and dryers. The first one, favouring segregation, consists of making uniform batches. Conversely, the second focuses on cost minimisation using the total storage and drying capacity. Collection Silos Each day, a collection silo receives a quantity of GM and non-GM maize. If there is one cell that already contains the product delivered, the delivery is put into this cell if there is room. Once it is full, the rest is put in another cell containing the same product, or in an empty one. If there is no such cell, the management of the rest will depend on the scheduling of the collection silo: In the case of a scheduling in favour of segregation (SS1) the rest will be refused and deferred to the next day. In the case of a scheduling in favour of cost minimisation (SS2) the rest will be put in the first cell with sufficient free space. The maize in this cell will then be considered as GM.

Dryers

Drying facilities consist of two structures: dryer waiting silos, where maize is stored before being dried, and the actual dryers. Each day, a dryer dries one batch of maize. The management of dryers depends on their scheduling: In the case of a scheduling in favour of segregation (SD1) drying batches are uniform, and so contain only one type of product, even if the dryer is not used at its full capacity. In the case of a scheduling in favour of cost minimisation (SD2), mixing of GM and non-GM takes place as soon as there is not enough of one product to use the dryer’s full capacity. In this case the batch dried is treated as GM. Moreover, the dryer is managed to minimize the change of products from one day to another in order to minimise the amount of non-GM to be treated as GM. Transport Each day, the collection silos can call for transport if their stock is above a certain threshold. These requests are treated using the First In First Out management rule, the older batch being given priority. To take into account the time constraint of 48 hours for the food market, the delivery stocked at t-1 has the higher priority level. If it is not possible to store the incoming batch in the waiting silos at the drying facility, the delivery is deferred to the next day. The Model The model works with a half day time step. Each half day, collection silo stocks are calculated, taking into account the GM and non-GM deliveries. GM and non-GM quantities dried are calculated, taking into


332

account the waiting stock at the drying facility. From these new values of stocks in collection and dryer waiting silos, transport of maize from collection silos to drying facilities is calculated. In order to run a simulation, we use the values shown in table 1. These values are the ones we found in the collection firms we worked with. The territory we simulated contains ten collection silos and two dryers. Table 1: Value of the different parameters

Size of collection silos 4*100 t Size of dryer waiting silos 2*250 t Drying capacity 1000 t/ day Number of trucks 30 Size of trucks 36 t GM collected 100000 t Non-GM collected 50000 t

We first simulated the collection with 150000 t of one product in order to compare the cost of a situation with segregation with the present situation. The deliveries per day for the whole collection period are shown in figure 1. This curve is the ideal situation for collection firms. It comes from the combination of an optimal management of grain maturity and the desire of farmers and collection firms to harvest maize when it is as dry as possible. Figure 1: Deliveries per day for a collection with one product

0

500

1000

1500

2000

2500

3000

1 5 9

13

17

21

25

29

33

37

41

45

49

53

57

61

65

69

73

77

81

85

89

days

ton

s

Then we simulated three situations: One in which farmers can deliver their maize when and where they want. A spatial strategy where farmers can deliver their maize when they want to, but to a specific collection silo depending on the product (GM or non-GM). One third of the collection silos and one dryer were allocated to non-GM products. The curve of delivery for these two situations is shown in figure 2. A time strategy where farmers can deliver their products where they want but non-GM crops are collected only in the first month and GM crops are collected only during the two last months, (figure 2b).


333

Figure 2: Deliveries per day of GM (▲) and non-GM (●) with no constraint or with spatial organisation (a) or time organisation (b)

For each of these three situations we calculated the ratio of the quantities of product at the end and at the beginning of the collection process. The ratio of GM can therefore be higher than 100% if there is non-GM crop mixed with GM. To consider the cost we compared (i) the increase in transport cost from the situation with one product and (ii) the rate of dryer use, which is a good indicator of drying cost, as this cost is nearly independent of the quantity dried. Results Influence of Scheduling Rules Figure 3 shows the ratio of GM and non-GM crop at the end of the process from GM and non-GM at the beginning of the process. It is so possible to compare the different scheduling rules in the case of a decentralised logic. It shows that the combination of rules which does not favour segregation (SS2 X SD2) allows 20% of the non-GM product to be segregated. This comes from the size of the non-GM collection which is large enough to allow the “natural” constitution of a non-GM homogenous batch. But, if scheduling rules in favour of segregation are used (SS1 X SD1), the ratio of non-GM increases to 50 % for a very small cost increase. The increase in transport cost is zero and dryer use rate is 93 % (table 2). Figure 3: ratio of GM and non-GM crop segregated at the end of the process without any collection strategy.

0

20

40

60

80

100

120

140

%

SD1 SD2 SD1 SD2

SS1 SS2

% GMO at the end of process % non GMO at the end of process

a b

0

200

400

600

800

1000

1200

1400

1600

1800

2000

1 5 9

13

17

21

25

29

33

37

41

45

49

53

57

61

65

69

73

77

81

85

89

days

ton

s

GMO delivered non GMO delivered

0

500

1000

1500

2000

2500

3000

3500

1 5 9

13

17

21

25

29

33

37

41

45

49

53

57

61

65

69

73

77

81

85

89

days

ton

s

GMO delivered non GMO delivered


334

Table 2: Transport cost increase and dryer use rate for each collection strategy and for the different scheduling rules. (1) : SS1 x SD1. (2) SS1 x SD2. (3) SS2 x SD1. (4) SS2 x SD2

No collection strategy

Collection with a spatial strategy

Collection with a time strategy

Transport cost increase

01 02 03 04 6111 6112 6113 6114 4001 3902 4003 3904

Dryer use rate 931 1002 883 964 881 882 883 884 901 902 903 904

Collection with a Spatial Strategy In this case two different supply chains are made. One third of the collection silos are allocated to non-GM products and the two other thirds are used for GM. GM and non-GM are dried into two different dryers. Figure 4 shows the ratio of GM to non-GM at the end of the collection process. The ratio of non-GM at the end of the process is nearly 100%. But the ratio of GM is low (82%). This is due to the small size of the GM supply chain compare to the size of the GM collection This come form the fact that there is only one dryer allocated to GM. It is therefore not possible to dry all the GM maize collected. Conversely the size of the non-GM supply chain is greater than the non-GM collection. So, the dryer is used below its capacity. This is confirmed by a low rate for dryer use (88%) and hence a high drying cost. Moreover, this strategy has a high transport cost which is increased by 610%. This is because it is impossible to minimise the distance between collection silos and dryers, as each dryer is allocated to one product. Figure 4: Rate of GM and non-GM segregated at the end of process with a spatial strategy collection.

0

20

40

60

80

100

120

140

SD1 SD2 SD1 SD2

SS1 SS2

% GMO at the end of process % non GMO at the end of process

Collection with a Time Strategy In this case, the non-GM crop is collected in the first month of the collection period and refused after that. GM crops are collected in the last two months of the collection period and refused before then. The ratio of non-GM at the end of the process is 87% and that of GM is 90% (figure 5). Moreover, production costs are lower than for the spatial strategy. The dryers are used at 90% of their capacity and the transport cost


335

has increased by 400% compared with a situation without segregation. These results show that segregation is done at the expense of overall collection efficacy, as not all the maize collected is dried. It is obvious in the case of a time strategy as it is impossible to dry both the whole GM and non-GM crop collected. So these strategies could lead to an increase in the duration of collection in order to dry all the maize. This will lead to a loss of quality and thus to a decrease in the prices paid to farmers. Figure 5: Ratio of GM and non-GM segregated at the end of the process with a time strategy collection.

0

20

40

60

80

100

120

140

SD1 SD2 SD1 SD2

SS1 SS2

% GMO at the end of process

% non GMO at the end of process

Conclusion / Discussion This modelling work shows that use of decentralised scheduling rules over the course of time is less efficient than centralised decisions based on forecasting if the efficiency is judged by the quality of production (Li and Liu 2006). But we show that changing the efficiency criterion, in this case taking the production cost into account, changes this judgement. There is a choice to be made between centralised and decentralised planning and scheduling on the basis of the choice of an efficiency criterion. This decision must be made taking into account the type of market sought. Thus there are difficulties for co-existence between GM and non-GM production at the collection level. To overcome these difficulties it is necessary to plan the collection before the collection period in order to specialize the infrastructure for one or the other product. Doing so leads to an increase in the collection cost and a decrease in the quantity of maize dried in the time required to produce the higher quality wanted by the market. A spatial specialisation of the infrastructure allows most of the collected crop to be dried, but at a higher cost. Conversely, a specialisation of the infrastructure on a time basis, with GM at the end of the collection period, minimises the cost but with a decrease in the quantities dried. It is therefore necessary to seek an optimal collection plan, taking into account the cost and the quantity dried and segregated. This optimisation could focus on the optimal collection duration of each product and on the period when each product should be collected. A combination of the two strategies with spatial segregation in a specific period could be explored.


336

Moreover, such collection strategies lead to the possibility of stricter collection territory governance, as considered by Byrne and Fromherz (2003). It would not be possible to introduce such collection strategies without consultation with the farmers; otherwise there is a risk that farmers will change their relationships with collection firms and sell their harvest to the firm with the fewest restrictions. These different types of governance should be evaluated taking into account the cost to the farmers and for the collection firms, together with the ratio of maize segregated. References

Angevin, F., Roturier, C., Meynard, J. M., Klein, E. K. 2003, "Co-existence of GM, non-GM and organic maize crops in European agricultural landscapes : using MAPOD model to design necessary adjustments of farming practices", 1° European Conference on the co-existence of Genetically Modified Crops with Conventional and Organic Crops, Borupsgaard (DNK), 13-14/11/2003 pp. 166-168.

Angevin, F., Sester, M., Choimet C., Messean, A., Gomez-Barbero, M., Rodriguez-Cerzo, E. 2005,

"Using the GeneSys-beet model to evaluate and manage populations of Herbicid-Tolerant weed beet, and implications for coexixtence of Herbicid Tolerant and conventional sugar beets", Second International Conference on Co-existence between GM and non GM based agricultural supply chain edn, A. Messean, ed., Agropolis Production, Montpellier (FRA)14-15/11/2005, pp. 101-104.

Arvanitoyannis, I. S., Choreftaki, S., Tserkezou, P. 2006, "Presentation and comments on EU legislation

related to food industries-environment interactions: sustainable development, and protection of nature biodiversity- genitcally modified organisms", International journal of food science and technology, vol. 41, pp. 813-832.

Beckmann, V., Soregaroli, C., Wesseler, J. 2006, "Coexistence rules and regulations in the european

union", American Journal of Agricultural Economics, vol. 88, no. 5, pp. 1193-1199. Byrne, P. F. Fromherz, S. 2003, "Can GM and Non-GM Crops Coexist? Setting a Precedent in Boulder

County, Colorado, USA", Journal of Food, Agriculture & Environment, vol. 1, no. 2, pp. 258-261. Coléno, F. C., le bail, M., Raveneau, A. 2005, "Segregation of GM and non GM production at the primary

production level", Second International Conference on Co-existence between GM and non GM based agricultural supply chain edn, A. Messean, ed., Agropolis Production, Montpellier (FRA), 14-15/11/2005, pp. 169-172.

Entrup, M. L., Gunther, H. O., Van Beek, P., Grunow, M., Seiler, T. 2005, "Mixed-Integer Linear

Programming approaches to shelf-life-integrated planning and scheduling in yoghurt production", International Journal of Production Research, vol. 43, no. 23, pp. 5071-5100.

Jank, B., Rath, J., Gaugitsch, H. 2006, "Co-existence of agricultural production systems", Trends in

Biotechnology, vol. 24, no. 5, pp. 198-200. Jank, B., Rath, J., Spok, A. 2005, "Genetically modified organisms and the EU", Trends in

Biotechnology, vol. 23, no. 5, pp. 222-224. Le Bail, M. 2003, "GMO/non GMO segregation in the supply zone of country elevators.", 1° European

Conference on the co-existence of Genetically Modified Crops with Conventional and Organic Crops, Borupsgarrd(DNK), 13-14/11/2003 pp. 125-127.


337

Le Bail, M. and Valceschini, E. 2004, "Efficacité et organisation de la séparation OGM/non OGM.",

Economie et Société. Série «systèmes agroalimentaires», vol. 12, no. 4, pp. 18-29. Levidow, L., Carr, S., Wield, D. 2000, "Gentically modified crops in the European Union: regulatory

conflicts as precautionary opportunities", Journal of Risk Research, vol. 3, no. 3, pp. 189-208. Li, J. L. and Liu, L. W. 2006, "Supply chain coordination with quantity discount policy", International

Journal of Production Economics, vol. 101, no. 1, pp. 89-98. Lüthy, J. 1999, "Detection strategies for food authenticity and genetically modified foods", Food control,

vol. 10, pp. 259-361. Scipioni, A., Saccarola, G., Arena, F., Alberto, S. 2005, "Strategies to assure the absence of GMO in food

products application process in a confectionery firm", Food control, vol. 16, pp. 569-578.


338

THE VALUE OF PREGNANCY TESTING SPRING-CALVING BEEF COWS

Billy Cook, Jon T. Biermacher*, and Dan Childs The Sam Roberts Noble Foundation, Inc.

2510 Sam Noble Parkway Ardmore, OK 73401

E-mail: [email protected] Abstract Implementation of best management practices into the beef cow enterprise is critical for long-term success. Previous literature suggests that pregnancy testing is valuable to the beef cow operation; however, less than half of producers in the southern Plains region of the United States utilize pregnancy testing. The objective of this research is to determine the expected value of pregnancy testing and the subsequent adoption of an effective culling practice on first-time non-pregnant beef cows relative to a system that does not use pregnancy testing or a culling strategy. Results show that the value of adopting pregnancy testing and an effective culling practice for first-time non-pregnant cows ranged between $54 and $76 head-1, depending upon the year. With the cost of pregnancy testing ranging between $2 and $5 head-1, the value of the risk-reducing information gleaned from pregnancy testing tends to warrant adoption. Keywords: adoption, beef cows, culling, pregnancy testing The Value of Pregnancy Testing Spring-Calving Beef Cows There are approximately 37 thousand cow-calf producers operating in the south-central Oklahoma/north-central Texas region of the United States with herd sizes ranging between 10 and 4500 head (average of 35), accounting for approximately 1.3 million beef cows (USDA-Oklahoma, USDA-Texas, 2006)1. According to a recent survey, producers who manage herds larger than 100 head glean over 40 percent of their income from their cattle operations, and producers with herds smaller than 100 head received less than 40 percent of their income from cattle (Vestal et al. 2006). Regardless of the size, how well a herd is managed is critical for long-term profitability of the cow/calf business. There are several components and techniques to a successful cow-calf management strategy—a management plan for beef cow replacement decisions should be one of them. Previous literature that focused on beef cow replacement decisions suggests that utilizing a strategic culling practice on unproductive cows is an essential management practice for herd profitability (Jarvis, 1974; Yager, Greer, and Burt, 1980; Melton, 1980; Blake and Gray, 1981; Bentley and Shumway, 1981; Rucker, Burt, and LaFrance, 1984; Trapp, 1986; Bourdon and Brinks, 1987; Foster and Burt, 1992, Frasier and Pfeiffer, 1994; Marsh 1999; Mathews and Short, 2001; and Ibendahl, Anderson, and Anderson, 2004). The literature makes note that a cow is not likely to recover the lost revenue from being open just once; however, some authors discuss situations when culling a younger open cow is not the best decision (e.g., when biannual calving seasons are considered, or when the cost of a replacement heifer is high relative to cow production expenses) (Tronstad and Gum, 1994; Ibendahl, Anderson, and Anderson, 2004, and others).

1 This statistic does not include the number of dairy cows, which accounts for an additional 60 thousand head (USDA-NASS, Oklahoma and Texas Quick Stats).


339

Tronstad and Gum (1996) concluded that producers should utilize pregnancy testing as part of a comprehensive management strategy, even though there are circumstances when it may not be profitable in the short run to do so. Contrary to these conclusions, the survey by Vestal et al, 2007 reported that only approximately 14 percent of producers who manage less than 100 head of beef cows utilize pregnancy testing for the cows they own, and only about 25 percent of them utilize pregnancy testing on their raised heifer cows. More surprising, the survey reports that only about 30 percent of producers with herds larger than 100 head utilize pregnancy testing, and only about half use pregnancy testing on the heifers they raise. In addition, nearly half of the producers in the region do not adhere to a defined calving season; however, for the other half that does, approximately 75% utilize a spring-calving season that begins in late January and runs through the end of March (USDA, 2006 and 2007). Findings from the survey and recommendations based on results from the literature do not match up well with what is observed in the region regarding the producer rate of utilization of pregnancy testing and culling strategies. That is, the practice of pregnancy testing and culling management is promoted, in general, as economical, but adoption has been limited. Ibendahl, Anderson and Anderson (2004) argue that the usability of dynamic programming models by farm producers is limited. We feel that this argument can be made for other types of simulation models such as Markovian simulations and net present value simulations, and may help to explain why the rate of adoption of recommendations from the literature that uses such modeling techniques has been limited. This observation along with the number of producers in the region that neglect to utilize a defined calving season provides the impetus for demonstrating to producers via an on-farm demonstration experiment the economic value associated with utilizing pregnancy testing and an effective culling protocol for first-time open beef cows. The objectives of this research are to determine the expected maximum value of pregnancy testing and the subsequent adoption of an effective culling practice on first-time non-pregnant beef cows relative to a system that does not use pregnancy testing nor a culling strategy, and to communicate to producers in the region how this value affects the net profitability of the spring-calving cow/calf enterprise. One contribution of our research to the current way of thinking about beef cow replacement decisions is that our experiment was utilized in order to provide an actual demonstration to producers in the region, allowing them the opportunity to see first hand what is required to carry out the operations associated with pregnancy testing and culling protocol. Also noteworthy, our research does not make assumptions about calf weights, market prices, or input costs in our analysis. As a result, we feel that the findings from this research will likely have a substantial impact on the rate of adoption of pregnancy testing and an effective culling protocol by producers in our region. Moreover, we believe our research will have a sizeable effect on the rate at which producers adopt a defined calving season and a subsequent management protocol for that system, regardless of how it is defined (i.e., fall-calving, spring-calving, biannual-calving, etc.). Conceptual Framework Economic theory suggests that a producer operating in a competitive market will adopt a new technology or production practice if the expected profitability from the technology is unambiguously larger than their current method of production (Grilliches 1957, 1958; Feder, Just, and Zilberman, 1985). Conceptually, the profit-maximizing producer faces the following decision rule for whether or not he should adopt pregnancy testing and an effective culling practice into his cow/calf enterprise

(1) >−

=otherwise, no,

,))E(max E()) E(max E(if yes,Adoption

λRRP


340

where 0>λ is the cost of change, )E( PR is the expected net return per cow when pregnancy testing is used, and )E(R is the expected per cow net return when pregnancy testing is not used. More formally, we define the value of information gleaned from pregnancy testing as the difference between the expected net return per cow when pregnancy testing is administered and culling first-time open cows implemented and the expected per cow net return when pregnancy testing is forgone and culling of first-time open cows not implemented. We assume that adoption of pregnancy testing and a culling protocol on first-time open cows is a risk-reducing technology for producers, and so we utilize an expected net return framework as apposed to using the expected utility framework. Mathematically, the value of pregnancy testing can be written as

(2)

)]},/()(E)/(E)/(E)E([

)]/()/(

)/(E)/(E)/E()E({[

),E()E(

kt NbvcBpNCpNFpwp

NbNT

vcNBpNCpNFpwp

RRV

ttkt

B

ktkt

C

ktkt

F

ktkt

Ptt

P

tkt

B

ktkt

C

ktkt

F

ktktkt

P

−−+++−

−−

−+++=

−=

where V is the average value of pregnancy testing per cow; E(.) is the expectations operator; the superscript P in equation 2 denotes the system (herd) that uses pregnancy testing; p is the price paid to the cow/calf producer for a calf of herd average weight w sold in marketing period k in year t; pF is the price paid to producers for first-time open cows; pC is the price paid to producers for non-productive, older cull cows; pB is the price paid for bulls culled in marketing period k in year t; F, C, B are the total number of first-time open cows, sick or nonproductive older cows, and bulls culled from the herd in marketing period k in year t, respectively; N denotes the total number of cows in the herd; vct represent the average per cow production costs in year t; T is the per cow cost of pregnancy testing in year t; b denotes fixed production costs in year t associated with ownership of capital (cows, equipment, buildings, fences, etc.) used in the production process. Note that management of cows is not expected to differ when pregnancy testing is adopted except for administering the pregnancy test itself, which is conducted by a certified technician at the same time spring-born calves are sorted and separated from their dams in the fall. Average net return between systems is expected to differ by the cow production expenses. The system (herd) that utilizes pregnancy testing would over time be expected to have a reduction in cow production expenses. A positive expected value represents the average additional profit per cow that a producer would expect to earn from adopting pregnancy testing and a strict culling regiment of first-time open cows into his cow herd management practices. Herd Description A culling management strategy was initiated on a group of 30 head of spring calving, 3-6 year bred cows of Angus, Brahman and Simmental inheritance in 1998. Any cow that did not wean a calf or that was not palpated pregnant in the fall of each year was removed from the herd. Additional bred cows of similar breeding were added back to the herd in the fall of each year to maintain a 30 head herd. Prior to project implementation in the fall of 2000 a comparison group of 35 head of Angus, Hereford or Angus/Hereford cross bred heifers were purchased directly from a local producer. These cattle were selected to represent a typical set of English influenced heifers for the region. The cow herd composition used in the study, then, consisted of 27 mature cows with an average age of seven years, and 35 two-year old cows for a total of 62 cows. The herd was located at a research farm in


341

south-eastern Oklahoma, near the town of Allen. During the three-year study (2001-2004), no cows from either group were culled unless they died or displayed chronic unacceptable infirmities (e.g., broken leg). All 62 cows were exposed to 3 full-sib Angus bulls for 60 days from June 1 to August 1 of each year and similar management practices for all three years of the study.

Methodology The data provided the opportunity to determine the net return of keeping open cows in the herd for each of the three years of the study. Enterprise budgets were developed for each cow in each group (i.e., the mature group and the young group) for each year, including the non-pregnant cows. Cow costs for each group have been separated into variable expenses and fixed expenses. Variable expenses included the average costs for mineral, supplemental feed, hay for cows and bulls, pregnancy testing services, veterinary products for cows and bulls, machine hire/lease, pasture rent, pasture maintenance expenses (i.e., seed, custom hire, and fertilizer), labor, and miscellaneous expenses. Fixed costs include depreciation and interest for mature cows, young cows, bulls (sires), calf scales, and computer software used to keep track of the data and analysis. It is important to note here that the cost of an open cow was the same as the cost of a bred cow, except for any costs associated with the preconditioning program or any related feed yard expenses from the retained ownership program. There are alternative strategies in the region regarding how producers market their calves. Some producers elect to background their calves using a preset preconditioning program where value is added to them for a defined period of time, and then retain ownership of them via a retained ownership program with a feed yard. Alternatively, some producers make arrangements with their neighbors to share ownership and profit margins associated with placing calves on winter rye or wheat pastures, which can be a relatively cheap source of gain over the winter months in the region. However, the large majority of cattle producers operating a spring-calving operation in the region sell their spring-born calf crop at the time of weaning in early October. Under this system, producers will typically wean calves from their dams and immediately transport calves to a sale barn for quick sale so as to minimize shrink that is associated with stress due to transportation and handling. In an attempt to collect other useful information associated with the calves produced in this project, we elected to retain ownership of them with a feed yard. As a result, we did not actually sell calves from this study at the time of weaning. This required us to use an alternative approach to place value on the calves produced in our study. We calculated calf value as the average calf weight by gender (which we recorded at the time of weaning) in pounds times the average price paid per pound to producers who sold calves of similar weight at the Oklahoma City National Stockyards sale in early October. Weaning weights were adjusted by a shrink factor of three percent, which is common for this system in the region. Transportation and commission fees have been excluded for analytical convenience. Results and Discussion Descriptive statistics for the cows for each year are reported in Table 1. After the calving season in 2003, three cows from the mature herd and one cow from were sold due to chronic illness, reducing the total herd size to 58 for the 2004 production season. In 2002 it was determined that a total of 12 cows were open based on the pregnancy testing results. Of the 12 cows, three were from the mature group and nine from the young group. In 2003, there were 15 open cows, 13 of which from the young group. In percentage terms, approximately 37 percent of the cows in the young group were open relative to only 13


342

percent of the mature cows. By 2004, the results were better with only 10 open cows between both groups. Table 1: Descriptive Statistics for Cows and Calves by Cow Group and Year

Year

Number of Cows in Total

Herd

Number of Cows

in Mature Herd

Number of Cows in Young

Herd

Number of Open Cows

in Total Herd

Number of Open Cows in Mature

Herd

Number of Open Cows in Young

Herd

2002 62 27 35 12 3 9 2003 62 27 35 15 2 13 2004 58 24 34 10 3 7

Descriptive statistics for the calves for each year and group is reported in Table 2. The data show that there was a substantial difference between calving rates between the two groups in all three years. Over the three years of the study, the mature group of cows realized an average calving rate 17 percent greater than that of the younger group. The calving rate for the younger group was the lowest in 2003, which is not surprising given that almost 40 percent of the cows in that group were open. Table 2. Descriptive Statistics for Calf Crop by Cow Group and Year

Number Number Calving Calving Calving Number of Calves of Calves Rate Rate Rate of Cows In Mature in Young Total Mature Young

Year in Herd Group Group Herd Group Group 2002 50 24 26 81% 89% 74% 2003 47 24 23 76% 89% 66% 2004 47 21 26 81% 88% 76%

A count of non-pregnant cows for both groups by cow identification ear tag number is reported for each year in Table 3. Although several cows were identified as open over the three year period of the study, only two cows were identified as open in each of the three years of the study (i.e., cow number 1 in the mature group and cow number 72 in the young group). Table 3 also shows that cow number 41 from the younger group was found to be open in the first two years of the study (i.e., 2002 and 2003), but pregnant in the last year (2004). Moreover, we found that cows number 60, 61, and 65 from the younger group were open in the first year of the study (2002), pregnant in the second year (2003), but were found to be open again in the last year (2004).


343

Table 3: Open Cow Identification by Group, Year and Year-by-Year Interaction

Year

ID # Mature Group

ID # Young Group

2002 1,13,22 41,52,53,56,60,61,65,71,72 2003 1,10 38,41,51,54,55,57,59,63,66,68,69,72,75 2004 1,6,25 51,56,60,61,64,65,72 2002, 2003 1 41,72 2002, 2004 1 60,61,65,72 2003, 2004 1 51,72 2002, 2003, 2004 1 72

Interestingly, we see from Table 3 that cows 41, 60, 61 65, and 72 turned out to be open at least twice over the three years of the project while cows 52, 53, 56, and 71 were open only once over the three years of the project and appear to have become productive after just one year of being open. We can not say anything about cow number 64 in the final year of the study (2004), except to say that she was in fact open; we do not know whether or not she would have been more productive in time. Weaned pay weights for each cow group and year are reported in table 4. As expected, calves in the younger group realized a lower average weaned pay weight than did the cows in the mature group. Calves from the younger group, on average, weighted 18 kg less than the average pay weight of calves in the mature group. In addition, weaning weights of calves from both cows groups increased steadily each year of the project, reflecting heavier calves as cow age increases. As one would expect due to differences in age, weaning weights per cow exposed was greater for the mature cows compared to the younger cows. This result was consistent with findings reported by the Beef Improvement Federation (BIC). Table 4: Weaned Pay Weight by Group and Year (kg)

Variable 2002 2003 2004 Average Mature Group 222 224 227 224 Young Group 202 196 221 206

Pay weights at weaning provide useful information, but that information can be misleading as it relates to open cows. A better measure of animal productivity is weaning weight of calves per cow exposed, reported in table 5. Significantly less weight of weaned calves are available from the younger group than from the mature group as a result of open cows. When accounting for open cows, we see that the average weight at weaning of a calf in the young group is, on average, is 54 kg less than the average weaned weight when open cows are not considered (i.e., 206-152). Although a difference in weight of 16 kg exists in the mature group when open cows are considered, the difference is not as significant as that found in the young group. This is because the mature group was substantially more productive in terms of producing calves than the younger group.


344

Table 5: Weaning Pay Weights per Exposed Cow by Group and Year

Variable 2002 2003 2004 Average Total Herd 178 165 187 177 Mature Group 213 207 205 208 Young Group 151 131 174 152

The total cost of non-pregnant cows for each of the two groups and years are reported in table 6. As one can see, the total cost for all open cows in the herd (i.e., young and mature groups) over the duration of the study was approximately $18,600. Without much surprise we can see that there was an $11,250 difference between the total costs associated with the open cows in the mature herd versus that of the young herd. Over the span of the study, the average total cost of the open cows in the young herd was approximately $3,750 more than that of the mature group of cows. Table 6. Total Cost of Open Cows by Group and Year ($)

Total Mature Young

Year Herd Group Group Difference 2002 5,940 1,423 4,518 3,095 2003 7,772 905 6,867 5,963 2004 4,969 1,388 3,582 2,194 Total 18,681 3,716 14,967 11,250

Average 6,227 1,239 4,989 3,750 Prices for weaned calves by cow group and year are reported in table 7. The main point of this table is to highlight the fact that the lightweight calves from the younger group did receive as expected, in each of the three years, a higher price than did the heavier calves of the mature group. The significance of this is that we believe that the effect of the additional weight of the calves from the mature group relative to the young group on net return is not larger than the effect on net return of the additional weight in the mature group from having more calves relative to the younger group. This effect does not offset the losses from having fewer of the lighter weight animals, even if they do bring a higher price per kilogram. Table 7. Prices for Weaned Calves by Group and Year ($ kg-1)

Variable 2002 2003 2004 Average Mature Group 1.8771 2.3395 2.5887 2.2689 Young Group 1.9735 2.5093 2.6262 2.3704

Net return to all unpaid resources per cow for both the mature group and the young group of cows for each year is reported in Table 8. The older, more mature group of cows (those that received pregnancy testing and a strict culling protocol prior to the project implementation) outperformed the younger group (the group that did not receive testing or culling) on average and in all three years of the study. The average threshold value of information from pregnancy testing and implementation of a strict culling protocol on first-time open cows was equal to $64 cow-1 (i.e., the difference between the net return of the mature group and the net return of the young group). Net return, and hence the value of information


345

varied substantially across years. However, the value appears to be substantial enough to cover the $2 cost head-1 of administering the pregnancy testing. Table 8: Net Return for Weaned Calves by Group and Year ($)

Variable 2002 2003 2004 Average Mature Group -39 46 139 49 Young Group -101 -30 85 -15

In addition to the analysis of the data generating in this project, it is important to report that demonstration field days were advertised and made available to producers in the region each of the four years pregnancy testing was administered to the mature group of cows prior to project implementation in 2001. These demonstrations provided cow/calf producers the opportunity to see first-hand the process of administering the pregnancy testing procedures used for this project, and allowed them the opportunity to ask question of production animal scientists and trained technicians regarding pregnancy testing and culling options. A substantial turnout each year by producers is worth noting, as where the level of questions fielded during these demonstrations. An obvious limitation of this research is the total number of years the experiment was conducted. We would expect the average value of pregnancy testing to vary somewhat through the peaks and troughs of the cattle cycle. However, as the costs of testing remains low, and additional information useful for making better decisions is seen as a risk reducing technology, the net benefits from the information associated with the testing is believed to be beneficial to cattle producers. Conclusions A three year demonstration experiment was conducted in south-central Oklahoma to determine the expected value of adopting pregnancy testing and a strict culling protocol for first time open cows in spring-calving beef cow herds. The study yielded several useful pieces of information. First, it was discovered that the group of cows that did not utilize pregnancy testing and culling of first-time open cows realized a cost of $11,250 more than the group of cows that utilized pregnancy testing and culling of first time open cows. Second, the study was useful in that it demonstrated first-hand to producers in the region the technical aspects associated with pregnancy testing, and allowed them the opportunity to ask questions of certified technicians and animal scientists. Lastly, we found that the average value of information gained from pregnancy testing and culling first time open cows ranged from $54 and $76 head-1, providing ample justification for paying the $2 to $5 head-1 cost for testing. References Beef Improvement Federation (BIF). 2002. “Guidelines for uniform beef improvement programs.” 8th ed,

Animal and Dairy Science Department Publication, University of Georgia, Athens, GA. Found at: http://www.beefimprovement.org/library/06guidelines.pdf [Accessed, December 2006].

Bently, E., and C.R. Shumway. 1981. “Adaptive planning over the cattle price cycle.” Southern Journal

of Agricultural Economics 13, 139-148.


346

Blake, M., and J.R. Gray. 1981. “A method for determining ranch profit probabilities when livestock yields are normally distributed.” Western Journal of Agricultural Economics 6, 103-112.

Bourdon, R.M., and J.S. Brinks. 1987. “Simulated efficiency of range beef production: III. Culling

strategies and nontraditional management systems.” Journal of Animal Science 65, 963-969. Bulut, H., J.D. Lawrence, and R.E. Martin. 2006. “The value of third-party certification claims at Iowa’s

feeder cattle auctions.” Iowa Beef Center, Iowa State University Extension Publication, IBC 30. Feder, G., R.E. Just, and D. Zilberman. 1985. “Adoption of agricultural innovations in developing

countries: A survey.” Economic Development and Cultural Change 33, 2, 255-298. Foster, K.A., and O.R. Burt. 1992. “A dynamic model of investment in the U.S. beef cattle industry.”

Journal of Business and Economic Statistics 10, 419-426. Frasier W.M., and G.H. Pfeiffer. 1994. “Optimal replacement and management policies for beef cows.”

American Journal of Agricultural Economics 76, 847-858. Grilliches, Z. 1957. “Hybrid corn: an exploration in the economics of technological change.”

Econometrica 25, 4, 501-522. Grilliches, Z. 1958. “Research costs and social returns: hybrid corn and related innovations.” Journal of

Political Economy 66, 5, 419-431. Ibendahl, G.A., J.D. Anderson, and L.H. Anderson. Spring1994. “Deciding when to replace an open beef

cow.” Agricultural Finance Review, 61-74. Jarvis, L.S. 1974. “Cattle as capital goods and ranchers as portfolio managers: an application to the

Argentine cattle sector.” Journal of Political Economy 82, 489-520. Marsh, J.M. 1999. “The effects of breeding stock productivity on the U.S. beef cattle cycle.” American

Journal of Agricultural Economics 81, 335-346. Mathews, K.H., and S.D. Short. 2001. “The beef cow replacement decision.” Journal of Agribusiness 19,

2, 191-211. Melton, B.E. 1980. “Economics of beef cow culling and replacement decisions under genetic progress.”

Western Journal Agricultural Economics 5, 137-147. Rucker, R.R., O.R. Burt, and J.T. LaFrance. 1984. “An econometric model of cattle inventories.”

American Journal of Agricultural Economics 66, 131-144. Trapp, J.N. 1986. “Investment and disinvestment principles with nonconstant prices and varying firm size

applied to beef-breeding herds.” American Journal of Agricultural Economics 68, 692-703. Tronstad, R., and R. Gum. 1994. “Cow culling decision adapted for management with CART.” American

Journal of Agricultural Economics 76, 237-249. Tronstad, R., and R. Gum. 1996. “Value of pregnancy testing.” Arizona Ranchers’ Management Guide,

Arizona Cooperative Extension, 123-136.


347

U.S. Department of Agriculture, National Agricultural Statistics Service. 2007. “Fact Finders for Agriculture.” USDA, Washington, Published February 2, pg. 6.

U.S. Department of Agriculture, National Agricultural Statistics Service. 2007. “Fact Finders for

Agriculture.” USDA, Washington, Published July 2, pg. 6. U.S. Department of Agriculture, National Agricultural Statistics Service. 2007. “Quick Stats, Oklahoma,

County Livestock Data.” Online. Available at http://www.nass.usda.gov/Statistics_by_State/Oklahoma/index.asp#.html. [Accessed, January 2007].

U.S. Department of Agriculture, National Agricultural Statistics Service. 2007. “Quick Stats, Texas,

County Livestock Data.” Online. Available at http://www.nass.usda.gov/Statistics_by_State/Texas/index.asp#.html. [Accessed, January 2007].

Vestal, M.K., C.E. Ward, D.G. Doye, and D.L. Lalman. 2006. “Beef production and management

practices and implications for educators.” Selected paper presented at the American Agricultural Economics Association meeting, Long Beach, California, July 2006.

Yager, W.A., R.C. Greer, and O.R. Burt. 1980. “Optimal policies for marketing cull beef cows.”

American Journal of Agricultural Economics 62, 456-467.


348

THE DEVELOPMENT AND ROLE OF NOVEL FARM MANAGEMENT METHODS FOR USE BY SMALL-SCALE FARMERS IN DEVELOPING COUNTRIES

Peter Dorward* and Derek Shepherd School of Agriculture Policy and Development,

University of Reading, Box 236,Reading, RG6 6AT, UK


Mark Galpin Formerly of the School of Agriculture, Policy and Development, University of Reading. Now with

International Nepal Fellowship,Box 1230, Kathmandu, Nepal

Abstract A key aspect of farm management is decision making and a variety of methods to assist with decision making are widely used in commercial agriculture, ranging from simple budgets through to complex computer models. The vast majority of the worlds’ farmers however have relatively small units of land and are in developing countries. It is widely accepted that these farmers make rational decisions within the challenging, complex, and risky environments that they operate. Despite major training and dissemination initiatives over many years, supported by national governments and international organisations to encourage farmers and extension staff to use farm management budgeting methods, uptake has remained extremely low. This paper reports on a research project funded by the UK Department for International Development, that firstly identified what small-scale farmers wanted from decision making methods and then worked with farmers and advisors in Ghana and Zimbabwe to develop and evaluate new and appropriate methods. These methods, in particular participatory budgets, have subsequently been successfully used in a range of developing countries. Results from activities conducted to evaluate participatory budgets are presented together with experience of their use for a variety of farm management functions, including planning and decision making. The findings demonstrate that they provide a useful method for small-scale and for non and semi-literate farmers operating in challenging environments, as well as for extension and research staff working with them. Keywords: participatory budget, participatory farm management method, PRA, decision making, planning, response farming

The Need for New Farm Management Methods Farm management is essentially about decision making and farm management methods are widely used in commercial agriculture. These include various budgets such as gross margins, net margins, profit and loss accounts, balance sheets and more complex techniques including the use of linear programming models. However the majority of the world’s farmers do not use such conventional farm management methods and neither do most of the government and non government agencies who work with them. This is despite several initiatives to facilitate their use, particularly in the 1980s as part of extension approaches funded by international organisations. The majority of world’s farmers is small-scale, operate in challenging, complex and risky environments, and are widely regarded to make rational decisions about their farms and other activities they are engaged in. Many of these farmers are non or semi- literate. Despite several authors identifying the need for more


349

appropriate farm management methods for small-scale farmers in developing countries in the 1980s (Harding, 1982; Rehman and Dorward, 1984) an extensive review of the literature and consultation of experts in the field indicated that no work had been conducted to address this need by 1996 (Dorward et al., 1997). This paper provides an overview of research subsequently conducted to develop and evaluate novel farm management methods for small-scale farmers in developing countries. A research project was funded by the UK Department for International Development (DFID) and was mainly conducted in Zimbabwe and Ghana. The paper also draws on experience of the use of the methods in a wide range of countries. The work involved: Identifying the decision making requirements of small-scale farmers; Developing and modifying new farm management methods with farmers and extension staff; Evaluating the new methods through a variety of approaches and for different uses; Disseminating the new methods through training, extension, and publication of training materials. Developing New Farm Management Methods Figure 1 summarises the main processes used in developing new methods. Small-scale farmers, extension staff and relevant experts were consulted to consider what types of decisions small-scale farmers make about allocating resources, how they do so and therefore what types of methods are needed. This included widespread informal consultation in several countries and formal and informal survey work in Zimbabwe. A comprehensive review of the international literature was also conducted. These activities were also used to establish why existing conventional farm management methods were not widely used and had failed to meet small-scale farmers’ needs. The understanding gained was then used in the design of novel methods. Four main limitations of conventional farm management methods were identified and that novel methods would need to address:

1. Conventional farm management methods focus on financial measures e.g. profit, cash, or worth. They generally work on the premise that profit maximisation or increasing worth are the main objectives of users. In reality small-scale farmers operating in harsh and unpredictable environments frequently have other objectives such as to reduce risk through improving food security. Resources other than cash are therefore important in decision making. 2. Conventional farm management methods focus on the final output (e.g. profit at the end of the production period) and do not take into account changes with time during a production period or season. Changes during the production period may be crucial to small-scale farmers (e.g. availability of food, livestock forage, labour, cash) and unpredictable natural and economic conditions may have major impacts. Therefore farmers often take important decisions during a season, depending on conditions at that particular point (e.g when it is evident how good the rains have been, how much labour they have or how healthy a crop looks). Stewart (1986) has described this as ‘response farming’. 3. Conventional farm methods are relatively complex and difficult to use, particularly for non or semi-literate farmers. They are therefore not easily used by the majority of the world’s farmers, either on their own or with advisers.


350

4. Conventional methods often require a limited amount of equipment to use eg personal calculators or even computers. Even the use of pen and paper can be inappropriate with non or semi-literate farmers. Figure 1: The main activities in the development of new farm management methods

Participatory approaches, referred to as Participatory Learning and Action (PLA), have become widely used and furthermore demonstrate rural peoples’ abilities to diagram, map and score. The emphasis on visualisation, together with analysis and ownership of information by farmers, provided valuable lessons. Participatory farm management methods were therefore developed that drew on both PLA and on understanding gained of small-scale farmers requirements for farm management decision making methods noted above. Initial ideas were brainstormed and then tried and discussed with Ugandan farmers and subsequently further developed and refined working with Zimbabwean small scale farmers. Participatory Budgets Of the methods that were developed, participatory budgets received the most attention and have subsequently been used the most widely. The remainder of the paper therefore focuses on participatory budgets. (Descriptions of all the methods and examples of their use are presented in a training manual, Galpin et al., 2000.) Participatory budgets examine the use and production of resources over time for an enterprise. Normally they are constructed for one production period e.g. a whole season for a crop. They are prepared by farmers with counters and symbols on a board or grid. Figure 2 illustrates the basic layout. The columns represent periods of time e.g. months or weeks. The top row is used to show (with symbols) the activities for the enterprise in each time period (e.g. ploughing and planting in month one, weeding in month three). The second row is used for all resources required for each activity. Types of resources are indicated by different counters and amounts of resources are quantified by the numbers of counters. The final row (or rows) is used for all resources produced (the outputs or products) and different counters represent types of resources and the amounts of counters represent the quantities produced. Balances of resources can be calculated and if farmers want to, enterprise ‘profit’ can be calculated by giving all resources cash values. Drawings and symbols can be used instead of counters. Copies of participatory budgets are normally made on flip chart paper for farmers and other participants to refer back to.

Identifying small-scale farmers’ requirements for farm management methods Consulting farmers, extension staff, experts and literature

Identifying the limitations of existing farm management methods

Consulting farmers, extension staff, experts and literature

Developing concepts of new methods Brainstorming and discussions with farmers

Trying out and improving new methods

Using methods with farmers, reflecting on strengths and weaknesses, making improvements to methods, trying methods with farmers again


351

Figure 2. Example layout of a participatory budget

Participatory budgets build on lessons from a traditional board game played throughout much of Africa and in areas of south east Asia, known as ‘mancala’ (‘tsoro’ in Zimbabwe, ‘oware’ in Ghana). The game involves calculating and moving numbers of counters and often is played at high speed. It is played widely by non and semi-literate people thereby showing that they are often highly numerate. Participatory budgets can be used for several purposes including: Analysing farmers’ existing activities and use and production of resources; Exploring the implications of making a change to an enterprise (e.g. using organic rather than inorganic fertiliser, introducing an intercrop); Comparing different enterprises; Planning a new enterprise. These planning and decision-making functions of farm management can be conducted by individual or groups of farmers, often facilitated by an adviser. ‘What if’ questions can be identified by farmers and their possible outcomes (scenarios) explored e.g. what would happen if the rains failed or prices dropped in a particular month (rather like the use of a computer spreadsheet). Participatory budgets can also be used during production to help predict the effects of conditions once they are known (labour availability etc) and to make decisions on actions and allocation of resources for the remainder of the period. Finally, participatory budgets can be used by and with farmers to plan, conduct, and analyse the results from on-farm research. Although figure 2 illustrates their simple structure, participatory budgets normally contain much more information than is shown here. Also not reflected here is the learning that takes place during the creation of a participatory budget due to farmers and facilitators sharing information and discussing experience with each other.

Activities

Inputs

Outputs

Cash, Balance, « Profit »

Months


352

Evaluation of Participatory Budgets A range of both formal and informal activities were used in the research project to test and evaluate participatory budgets and for a variety of uses. These are summarised in table 1 and involved conducting specific exercises in order to establish how well participatory budgets met specific criteria when used for different purposes and in different contexts. Some of the exercises were conducted in one session (eg using participatory budgets with farmers to plan a possible new enterprise) and others involved working over a longer period (e.g. using participatory budgets with farmers throughout a whole growing season). The activities were carried out over a three year period, and with farmers, extension staff and researchers and in different farming systems in Ghana and Zimbabwe and therefore enabled relatively comprehensive evaluation. Table 1: Exercises and activities used to evaluate participatory budgets

Exercise or activity Main information and observations

1. Field testing of participatory budgets by extension and farmers, to describe farmers existing enterprises and explore potential improvements and new enterprises. Series of short exercises with 23 extension staff in Zimbabwe.

Farmers, researchers and extension staffs’ observations and scores e.g. for ease of use, usefulness

2. Long term evaluation with 22 extension staff incorporating use of participatory budgets into their work over seven months, in Zimbabwe.

Observations and uses recorded by extension staff. Evaluation workshop with extension staff at end of period. Experience of use of methods and strengths and weaknesses identified.

3. Exercises with farmers in two communities (one week each) to investigate the suitability of green manuring for vegetable farmers systems and resources. Ghana.

Documented report of results and of findings from the exercises. Farmers’, researchers’ and extension staffs’ observations.

4. Long term needs assessment over one season. Compared farmers planned participatory budgets and actual practice during a season, and explored reasons for this, to better understand farmers systems and constraints faced. Ghana and Zimbabwe.

Analysis of data from planned and actual participatory budgets. Observations from farmers and facilitators.

5. Short controlled exercise with 10 extension staff and farmers to compare participatory budgets with approaches currently used by extension staff for exploring the feasibility of starting a poultry production enterprise, in Zimbabwe.

Extent and relevance of information covered using different approaches was compared. Observers’ scores for extent to which feasibility of starting new enterprise (broken down into criteria) had been explored.

A variety of types of information and observations were identified prior to each activity. These included farmers’, researchers’, extension staff and trained observers’ observations and scores for specific criteria ie for ease of use, usefulness, level of participation, extent to which methods used achieved the objectives


353

of an exercise (eg exploring the feasibility of a new enterprise), strengths and weaknesses, and the extent and relevance of information included in budgets and in exercises. Information and observations were recorded during or immediately after each exercise. Further, a feedback and evaluation workshop was held for the 22 staff that had used participatory budgets as part of their extension work during a season. Five of these were then selected as case studies and visited in their own areas where the extent to which participatory budgets had been used and their usefulness were explored with the extension workers and with farmers. Data from the long term needs assessments of farmers was analysed to compare farmers’ planned and actual budgets and feedback sessions were held with the farmers. The overall findings are summarised here and detailed results from each exercise are presented in Galpin et al. (2000); Dorward (1999) and Galpin, (2000). The following numbered sections relate to the activities in table 1. Field Testing of Participatory Budgets by Extension and Farmers Farmers, extension and research staff found participatory budgets highly useful. The mean scores from eleven groups of farmers opinions on the ‘usefulness’ and ‘ease of use’ of participatory budgets, after they had first used them were almost the maximum possible; 9.5 and 9.0 respectively (possible scores between 1 and 10, 1 represents the most negative score possible ie ‘of no use at all’ and 10 represents the most positive score possible). Following the first exercise using participatory budgets with farmers, extension workers (19) gave a mean score of 8.4 for usefulness of participatory budgets to farmers (using the same scale noted above). Informal observations of farmers and extension using participatory budgeting throughout the research in both countries supported these high scores.


354

Table 2. Summary of options identified by farmers and their potential impacts

Producer Group

Green manure

crop planting

date

Tomato

harvest date

Impact

on

timing

Benefits /

Advantages

Costs / Risks

Early irrigators

October

April/May - June/July

No change

• reduced cash expenditure on fertiliser

• improved soil quality

• overall increase in labour required

• incorporation very intensive due to hard ground

• may lose green manure crop if drought - or very high labour costs for watering

• risk of fire damage to green manure crop

Mound transplanters

September

July

No change

• lower input costs ( no fertiliser required)

• reduced labour peaks in January / February

• other crops also benefit (including quality)

• reduced weed growth

• increased labour required in August / September (planting green manure) and October / November (incorporation)

• overall increase in labour required

Mound direct seeders

July

July

(Onions: November)

No

change

• green manure crop benefits other crops

• increased cash from minor season crop (if grown)

• higher input costs (for farmers who do not currently use fertiliser)

• requires cash outlay for minor crop (if grown)

Flat planters

March

September

Delay in tomato harvest

by 2 months (price

affected)

• no fertiliser costs • possible reduced

rates of abortion and flower drop as temperatures are lower

• possible benefit to subsequent crops

• increased labour particularly for incorporation

• production costs 6 -10 % higher

• higher disease incidence at harvest due to heavy rains

Adapted From Dorward et al. (2003) Long term evaluation with 22 extension staff incorporating use of participatory budgets into their work

over seven months


355

Extension staff that had been trained in the use of participatory budgets were asked to use them in their own work, where they saw suitable opportunities to, in the seven months following the training. They used them with farmers mainly for comparing various enterprises (i.e. to investigate which was more suitable and feasible) and for planning new enterprises. Enterprises commonly considered with farmers included poultry broiler production, vegetable gardening, maize and beans. Other uses included exploring the marketing and timing of some operations, replacing artificial fertiliser with compost, and investigating labour use. Extension workers’ opinions on participatory budgets were obtained during a feedback workshop and in addition several were visited in their own working locations. Feedback was obtained through use of questionnaires, interviews and a participatory activity where strengths and weaknesses identified by staff were scored and discussed. Extension staff considered participatory budgets to be useful for enabling farmers to identify and select the best enterprises for them, plan enterprises, and work out whether they have made a profit or not. They considered them suitable for non-literate and literate farmers and several reported that using participatory budgets had improved relationships with farmers. Feedback was positive although other wider factors influencing the extent to which staff could use participatory budgets were also noted including a general lack of opportunities (i.e. new enterprises) for farmers to improve their livelihoods. Exercises with Farmers in Two Communities (One Week Each) to Investigate the Suitability of Green Manuring for Vegetable Farmers Systems and Resources. This exercise was conducted to test the use of participatory budgets for assessing the suitability of potential innovations with farmers prior to implementing on-farm trials i.e. at a relatively early stage of technology development and adaptation. Green manuring had been identified by research staff in the Brong Ahafo region of Ghana as a possible means of addressing poor soil fertility for farmers producing tomatoes in the wet season. Before deciding whether on-farm trials should be established, a one week participatory analysis was conducted in each of two communities. Working with farmers representing five different types of tomato farmers in each community, participatory budgets were used to: a) describe existing tomato production; b) explain the green manuring technology; c) jointly explore the timing and resource implications of introducing green manuring into the cropping system; d) develop alternative timings and activities for fitting a green manure crop into the tomato production system; e) identify the likely resource use and production implications of options identified in d). Table 2 gives a summary of the options and their potential benefits and costs identified by different types of tomato farmers. For some types of tomato producers in each community it emerged that green manuring was not a sensible option and for others it may be with the adaptations suggested. These could then be explored by farmers and research staff in on-farm trials. The use of the participatory budgets had not only identified who the technology is likely, and importantly who it is not likely, to be suitable for, and how it can be adapted, but also what particular features need to be examined and focused on in trials and practical management. This case study illustrates that participatory budgets used in this way could improve the relevance and quality of subsequent on-farm research. Without this approach, several seasons of trials work could be conducted and associated resources used, before reaching the same findings (see Dorward et al., 2003). Furthermore the participatory budgets can be used by farmers and research staff for recording on-farm trial results (including resources) and for analysing and comparing findings. Long Term Needs Assessment Over One Season.

In order to investigate the potential to use participatory budgets during a production period and as a means of better understanding the constraints farmers operate in and their farming systems, exercises were conducted in two different farming systems in Ghana and Zimbabwe. At the start of the season individual farmers created participatory budgets for the season ahead. They were then visited each month during the season and revised their participatory budgets to reflect what had actually happened. At each visit and at the end of the season, each farmer and the facilitator working with them compared what had


356

been planned with the actual practice, and discussed reasons for any differences. In Zimbabwe six small-scale resource poor farmers who relied on maize as their main staple crop were worked with and in Ghana 22 small-scale tomato growers. The effects of the unpredictable natural and economic environments on farmers’ enterprises and decision making during seasons were very evident in both locations. Farmers’ practices were very different to plans expressed and consistent reasons for the differences were evident. Findings in Ghana included: The major cause of disruption to planned activities was time spent at funerals (an obligation) Early arrival of rains reduced labour demand for watering Actual inputs were different to those planned mainly due to farmers responding to the crop condition and input availability. Prices for inputs predicted were broadly accurate A glut of tomatoes on the market led to poor prices, reduced income and reduced expenditure on labour for picking (much of the crop remained unharvested). Despite the small sample size in Zimbabwe some consistent observations were evident, including: Unavailability of seed leading to inappropriate varieties being planted and low yields Late arrival of rains and low rainfall had major effects including delayed land preparation and planting. This contributed to labour competition later in the season and to low yields Again a major cause of labour shortage at important points in the season was the need to attend funerals Lack of cash, or alternative requirements for cash at particular times eg for school fees, resulted in no fertiliser being applied and no weeding being conducted by some farmers Illness and pregnancy in families resulted in reduced labour availability and delayed activities Farmers at both locations were clearly practicing ‘response farming’ as described earlier in this paper and having to make major decisions on resource use and activities during the season in response to unpredictable changes in the natural, social end economic conditions. Social factors were clearly important in influencing farmers’ responses. The use of the participatory budgets with farmers improved extension and research staffs’ understanding of the constraints faced by farmers and the nature of their decision making. Farmers in Zimbabwe observed that the process of using the participatory budgets as described here was helpful and in particular with planning and allocation of resources. The participatory budgets enabled farmers to visualise the impact of unpredicted events as well as of alternative management responses, and to allocate their resources in the light of this. Tomato farmers in Ghana were also positive about the use of the participatory budgets in the exercise. All of them noted that it helped improved their timeliness of their activities and most (20) noted how it enabled them to determine profitability of their enterprise which was not normally calculated. It had also helped with establishing the contribution of components to the success or failure of the enterprise e.g. labour costs. Despite the very different farming systems, cultures and environments of the locations in Ghana and Zimbabwe, participatory budgets had been useful to farmers and facilitators working with them, to understand their systems and constraints and to carry out farm management functions of decision making, planning and control.


357

Short Controlled Exercise with Extension Staff And Farmers to Compare Participatory Budgets with

Approaches Currently Used (To Explore The Feasibility Of Starting A Poultry Production Enterprise) In addition to the approaches to evaluating participatory budgets described above, more controlled exercises were designed and conducted. Five extension staff that had been trained in use of participatory budgets and five untrained staff (in Zimbabwe) conducted the same exercise. Each member of extension staff worked with a separate small group of farmers. Trained observers recorded information on the extent to which aspects of the exercise were completed. The exercise was divided into tasks to: a) explore the viability of them starting a broiler enterprise 2) explore how the new enterprise would fit with existing labour availability; 3) examine the possibility of taking out a loan; 4) consider what may go wrong with such an enterprise; 5) consider the impact of half the birds dying one week before the first sale. The extension staff that had not been trained in use of participatory budgets were asked to use whatever methods they normally would for such an activity. Scores for the extent to which, and how well, each task had been conducted were given based on observers’ scores and on analysis of detailed records of the information used and generated. Mean scores for staff and farmers using participatory budgets were higher than for those using conventional methods by between 225% and 600% for task 1 and between 11% and 250% for tasks 2 to 5. Conclusions In addition to the activities reported above, participatory budgets have since been used for a range of purposes in a variety of other developing countries. Examples include farmers working with research or extension staff analysing dairy systems in Mexico, exploring the potential of IPM options in maize-dairy systems in Kenya and investigating rice production in Bangladesh. Feedback from staff involved in these and other activities have generally supported the findings above. Participatory budgets provide a method that is appropriate to many of the resource-use decisions that small-scale farmers face, the factors that influence subsistence and near subsistence farmers decision making, and the unpredictable and changing environments they operate in. Furthermore they provide a way for research and extension staff together with farmers to explore the suitability of innovations and to take into account important differences in farmers’ access to resources and levels of poverty. Acknowledgements This work was mainly funded by the UK Department for International Development. The authors would also like to thank the Ministry of Food and Agriculture staff in Ghana and AGRITEX staff in Zimbabwe with whom much of the field work was conducted. In particular we would like to thank the many farmers in Ghana and Zimbabwe that we worked with.. References Dorward, P. (1999). Participatory farm management methods for improved agricultural extension with

smallholder farmers in Zimbabwe. PhD thesis. The University of Reading. Dorward, P., Galpin, M. and Shepherd, D.D. (2003). A case study on green manuring options for tomato

producers in Ghana. Agricultural Systems, 75, 97-117. Dorward, P., Shepherd, D.D. and Wolmer, W. (1997). Developing farm management type methods for

participatory needs assessment. Agricultural Systems 55, (2), 239-256.


358

Galpin, M., Dorward, P. and Shepherd D.D. (2000). Participatory farm management methods for

agricultural research and extension: A Training Manual. Department For International Development and the University of Reading.

Harding, T.J. (1982). Farm management advice to peasant agriculture: the transfer of technology. Journal

of Agricultural Economics, 33, 47-56. Rehman, T. and Dorward, A. (1984). Farm management techniques and their relevance to administration,

research and extension in agricultural development: Part 1 - Their evolution and use in developed countries; Part 2 - An appraisal of their potential in LDCs. Agricultural Administration 15, 177-190, 239-254.

Stewart, I. (1986). Response farming: A scientific approach to ending starvation and alleviating poverty

in drought zones of Africa. In Proceedings, African Agricultural Development conference: Technology, Ecology and Society. (Eds Yolander, and T. Moses). Pomona, CA: California State Polytechnic University.


359

A COMPARATIVE STUDY OF VARIABILITY IN AGRICULTURAL ENTERPRISES AND FISH FARMING

Ola Flaten

1Norwegian Agricultural Economics Research Institute (NILF), Box 8024 Dep., 0030 Oslo, Norway E-mail address: [email protected]

Gudbrand Lien

Hedmark University College, Rena, Norway

Ragnar Tveterås

Department of Industrial Economics, Risk Management and Planning, University of Stavanger, Norway Abstract Agri- and aquaculture have common features associated with their biological nature affecting risk exposure of the businesses. The aim of this paper is to compare risk exposure in salmon farming and agricultural enterprises in Norway by using an implicit error component model to examine the risk structure of yields, prices and economic returns. Panel data originated from the Norwegian Farm Accountancy Survey and the Norwegian Directorate of Fisheries. Results indicate a higher farm-level year-to-year variability in yields, prices and economic returns in salmon farming than in agricultural enterprises. Return on assets was highest in salmon farming with an average return of 9.2%. All of the agricultural farm types exhibited a negative average return on assets. Stochastic dominance tests of the distribution of economic returns from fish farming and agricultural enterprises showed salmon farming to the most economic viable alternative. Keywords: risk analysis; salmon farming; livestock production; crop farming; stochastic dominance; Norway Introduction Agri- and aquaculture1 are both biological production sectors and are exposed to widely varying and unpredictable elements of nature, like uncertainty in biological processes related to weather, diseases, pests, infertility, etc., which cause yield variability (production risk). In addition, activities are dispersed on heterogonous soils - or water conditions. Weather and spatial dispersion in agriculture particularly affects crops and grazing livestock. In contrast, confinement production of livestock partially controls production risk. Modern fish farming is essentially a batch production system, as in chicken or feeder-pig-to-finish operations, but fish is produced in open cages leading to less control of the biological processes than indoors. The biological uncertainty is a fundamental cause of price uncertainty. Consequently, the two sectors face many similar economic risks. However, there are also notable differences. Agriculture has existed for more than 10,000 years, and core agricultural techniques were developed early. Although the practice of aquaculture is ancient, fish farming only recently has become a specialised business. Open-net cage salmon farming in marine waters was pioneered in Norway in the late 1960s; this particular type of aquaculture technology is still a minor part of the production of farmed fish worldwide. It however dominates in salmon farming.

1 We refer to agriculture as the process of producing food, feed, fibre, fuel and other goods by the systematic raising of plants and animals on land. Aquaculture is the rearing of aquatic organisms under controlled or semi controlled conditions.


360

The two industries operate in different institutional environments. A large number of government interventions in agriculture are a common feature in many countries. The agricultural sector has built institutions and farmer cooperatives that, among other tasks, mitigate risk. In Norway, which we focus on in this study, agriculture mainly produces goods for the domestic market and receives substantial producer support, chiefly through import tariffs and government payments. The export-oriented aquaculture industry operates with more liberal market and trade regimes, and collective institutions to mitigate risk have not been developed. The current production level in Norway is around 600,000 metric tons of salmon, close to half of total world production, and more than 90% of its salmon production is exported. Finally, small, family-based firms dominate in agriculture, while aquaculture business structures have converted into a mix of medium-sized and large firms. Since the biological nature affects risk exposure in both sectors, we believe a better understanding of risk exposure can be achieved through comparative analysis of agricultural versus aquaculture businesses. Agri-and aquaculture products may be competitors in food markets and differences in risk exposure may be one important factor for predicting the success of the industries. Good risk estimates may be important for the industries, potential investors and policy makers. Also, the analysis can help to identify sector differences in the need for price and income stabilisation tools and risk management strategies. The aim of this paper is to compare risk exposure in salmon aquaculture and agricultural enterprises in Norway. First, we compute and compare the variability of yields, prices and economic returns at the farm level. Measures of variability in themselves may not indicate much about riskiness, except under specific probability distribution assumptions, such as normality. Second, we will employ a more general framework for addressing risk exposure, the stochastic efficiency methods, using measures of the farms’ economic returns. To the best of our knowledge, a cross-industry risk comparison like the one provided in this paper has not been done before. Methods Detrending procedures Improving technology and management influence the yield of most biological enterprises, and estimates of yield variability are conditional on having an appropriate model of the changes in the mean of output. Atwood et al. (2003) refers to three statistical procedures for detrending of yields: No time trend adjustment, which is likely to overestimate the variability since no trend is removed. Estimation of individual farm-level trends. If most or all farms in an area actually have similar underlying trends (which can be reasonable), estimating individual trends for each farm may result in non-robust trend parameters. An error component procedure that implicitly removes any common regional trend from the farm yields series (Atwood et al., 2003). This procedure, error components implicit detrending (ECID), was shown to better describe the reality in most cases than individually detrending farm-level data. Atwood et al. (2003) describes the ECID procedure as follows: Calculate the “regional” yield in year t, y

Rt, i.e., the area weighted average of farm yields in the region.

Compute each farm’s “yield” deviation from the regional yield, as: yy Rtitit -=∆ , where y

itis the yield of farm i in year t.


361

3) Compute farm i’s residuals as: )(-)-( yyyy RiRtitit −=ε (1)

where yi is farm i’s average yield, and y

R is the average regional yield in region R for the ti years of

farm output reported by producer i. The first term shows the farms deviation from the regional yield in year t and the second term the farms average deviation from the regional yield.

It can be shown that the resulting farm residual values have been implicitly detrended to the degree that farm yields follows a common regional yield trend. If there are reasons to believe that producers’ underlying yield trends could vary widely within a region, it is likely that the ECID procedure might generate biased residuals. However with short-term panel data, it will always be extremely difficult to identify whether a difference in an individually estimated trend occurred because of differences in actual trend or resulted from sampling anomalies. In this study we will use a modified version of the ECID procedure, where we also have included the relationship between the national and the regional yield level. In our ECID approach the decomposition of yield y

it at farm i in year t is expressed as:

εyyyyyy itRtRRiit ++)-(+)-(= (2) where y is average national output (average yield for all farms over all years) and y

R is average output

in region R (average yield for all farms in region R over all years). The four variability components in Eq. (2) can be expressed as: Time-invariant, farm-specific deviations, )-( yy

Ri , a farms average deviation from the regional yield

level. In other words, variability that arises from time-consistent, farm-related factors (soil/water properties, farmer skills, topographic position, permanent weather conditions, etc.) showing the variation between farms within a region. Time-invariant, region-specific deviations, )-( yy R , a regions average deviation from the national yield level, i.e., variation in yields between regions. Time-variant, region-specific deviations, y

Rt, average output in region R in year t expressing the variation

in yields between years in a region. Time-variant, farm-specific deviations, εit , the farm residuals, showing variation in yields between years on a farm caused by time-inconsistent factors such as weather variability and variable annual management decisions. We are particularly concerned with, from the farmers’ point of view, how yields vary between years on farms. We examined variability in yields between years within a farm, since this best describes variability in yields at the individual farm level. As a statistical measure of variability we used the coefficient of variation (CV), which expresses the standard deviation (SD) relative to the mean. The SD in yields within farms was estimated by taking the SD of the sum of variability components 3 and 4 in Eq. (2). Variance components were calculated by dividing the variance of a specific component by the sum of the variance of the four components in Eq. (2). A variance component represents the variance of a specific component as a fraction of total yield variance in an enterprise. The procedure does not take into account correlations between the components. To take notice of the correlations, within-farm correlations of yields were calculated and reported.


362

Estimation of price variability was based on the same error component procedure (ECID). The aim was to decompose variation in annual farm-level prices. All prices were converted to 2004 real Norwegian kroner (NOK, €1≈NOK 8.10), using the Norwegian consumer price index as price deflator. Also, for the examination of financial variability the ECID procedure was used.

Measures of financial performance Several measures of financial performance can be computed. When making comparisons across farms, it is useful to control for differences in their resource base. Also, since we are dealing with family farms using unpaid family labour as well as farms organised as corporations with paid labour we have to use a measure which can compare financial variability no matter how the farm business is organised. In the comparison of financial performance we first employed the rate of return on assets (ROA):

,(average) esasset valu farm total

assets return to=(%) assetson Return

where Return to assets = Net farm income from operations – opportunity cost of unpaid labour. ROA is the return on both debt and equity capital, since interest on debt capital is not included in net farm income from operations. To find the return to assets imputed charges for unpaid operator and family labour is deducted. Many agricultural businesses do have a negative ROA. Comparing two farms with the same negative return to assets, the one with the lowest asset values will have the most negative ROA. This may be confusing, since is it better to achieve a certain return to assets with the least use of assets. Therefore, we did also calculate a second financial performance measure for agricultural businesses, the profitability quotient (PQ), defined as:

100×labour ofcost y opportunit +assets farm all ofcost y Opportunit

operations from income farmNet =PQ

If PQ equals 100 (or higher) net farm income is sufficient to provide a return to capital and labour equal to (or higher) than their opportunity costs. Since the measures of economic returns already are relative terms, we will report the SD for the analysis of economic variability.

Stochastic efficiency analysis Risk can be measured as the variation in distribution of possible outcomes, and a risky choice is then one with a wide range of possible outcomes. Most managers do however associate risk with the negative outcomes, and a risky choice is one that contains a threat of a negative outcome (March and Shapira, 1987). Hence, from a managerial perspective, associating lower variability in economic returns with less risk can be misleading. Hardaker et al. (2004:266) have pointed out that the best route to risk efficiency is by finding strategies that improve the expected values of returns, rather than those that reduce dispersion as measured, for example, by the variance of returns. We isolated risk efficient solutions using stochastic dominance techniques. There are several stochastic dominance criteria, where this study used first (FSD) and second (SSD) degree stochastic dominance criteria. Stochastic dominance analysis requires comparison of cumulative distribution functions. Variability in economic returns within farms, estimated as the sum of component 3 and 4 in Eq. (2), was for each of the farm types used to generate empirical distributions of financial outcomes. An empirical distribution was


363

chosen because it avoids forcing a specific parametric distribution (such as the normal) on the economic returns. Data The data source for agriculture was the Norwegian Farm Accountancy Survey (NFAS) collected by the Norwegian Agricultural Economics Research Institute. The unbalanced panel data set includes farm production and financial data collected annually from about 1,000 farms. These farms are located throughout the country (divided into eight regions) and represent a wide range of farm sizes and types of farms. The total data set used in the analysis included 13,000 observations on 1970 farms from 1992 through 2004. Financial performance measures were only available at the whole-farm level. Many farms are mixtures of several enterprises. Farms in NFAS are classified according to their main categories of farming. To perform analysis of economic returns at the whole-farm level we included the most common farm types from the survey. Aquaculture was analysed with data from the Norwegian Directorate of Fisheries, which annually compiles data of salmon farm production data for their profitability survey of Norwegian fish farms. Firm level data for the years 1985-1998 were included. Later data were excluded, as region was only specified until 19982. In aquaculture, region is specified in terms of which county the farm belongs to. Ten of Norway’s counties have fish-farms represented in the sample. The sample annually includes 200-300 firms, typically representing over 50% of the total salmon production in Norway. In total the data set included 3600 observations. The accounting methods used in the data sets chiefly follow the rationale of conventional accounting, with its use of historical cost for the valuation of long-term assets3. Following the procedures of the NFAS, a flat labour charge per worked family hour equal to the wage rate for skilled farm workers were used to compute costs of unpaid labour. Opportunity costs of farm assets were set equal to the interest rate used in NFAS. Results and discussion Yield Variability and Correlation Yield variability and the variance components are reported in Table 1. Salmon farming showed the highest yield variability with a CV within farms of 58%. The high yield variability in salmon farming was expected, since the industry is rather young and has experienced rapid growth. The industry has been through periods where diseases and pests significantly have reduced the output and salmon farmed in sea cages are exposed to rough and variable weather conditions. Of the agricultural enterprises, only potatoes reached a CV of more than 50%. Forage followed with a CV of 38%. For grain crops the CV’s were within the range of 25 to 30%. Rasmussen (1997) found a CV

2 The two sectors are being compared across different time periods and for different length of times. Measures at the national aggregated level show the same mean ROA in aquaculture at 10% for the two periods 1985-1998 and 1992-2004. The CV was 6.3% in the former and 9.0% in the latter period. Even though the essence of problems concerning variability faced by individual producers is lost in the process of aggregation, these numbers suggests only modest impacts on the variability measures of the different time periods. 3 Market values for assets were not available. Non-farm assets (included the value of the farm dwelling) were not included in the asset values.


364

close to 20% for grain yields on Danish farms while the CV’s was around 10% among cereal growers in England (Webster and Williams, 1988). Among the livestock enterprises, sheep and hog-farrowing operations achieved the highest CV’s. Milk, goat milk and porkers were rather yield stable. Low CV’s have been found for milk and pork in Denmark (Rasmussen, 1997). It is reasonable that extensive grazing production like sheep is likely to have more variable yields than intensive livestock production, since the former is more severely exposed to the effects of variable weather conditions. Diseases and infertility may cause variations in herd productivity in hog-farrowing operations. Table 1 Estimated yield variability

Variance components

Enterprise Average

yield1

CV within farms2

Time-invariant

farm-specific

Time-invariant

region-specific

Time-variant region-specific

Time-variant

farm- specific

Barley, kg/ha 3859 0.27 0.33 0.13 0.24 0.30 Oats, kg/ha 4083 0.28 0.33 0.14 0.22 0.30 Wheat, kg/ha 4569 0.25 0.29 0.15 0.28 0.28 Potato, kg/ha 18572 0.51 0.33 0.13 0.17 0.37 Forage, feed units/ha 3720 0.38 0.27 0.33 0.19 0.21 Milk, litre sold/cow 5686 0.09 0.66 0.02 0.05 0.26 Sheep meat, kg/winter fed sheep 26.4 0.27 0.46 0.04 0.11 0.38 Goat milk, litre sold/goat 499 0.14 0.66 0.07 0.11 0.17 Weaners per sow per year 17.4 0.25 0.46 0.07 0.18 0.29 Finisher-hog, kg slaughter weight 75.9 0.08 0.41 0.02 0.36 0.21 Salmon, kg/m3 cage volume 27.6 0.58 0.28 0.03 0.27 0.43

1 Paid or sold crop yields 2 Mean of the farms

The variance components show that the variability in yield level between the farms within a region was relatively more important for livestock enterprises than for crop enterprises and salmon (column 4 in Table 1). The time-invariant region-specific component was small for salmon and livestock enterprises, i.e., small variations in yields between regions. The larger variation between regions in crop yields is primarily caused by time-consistent differences in climatic conditions for crop production. The time-variant region-specific component was generally lowest for the livestock enterprises. The higher region-wide variation for crops and salmon may be associated with their heavier influence of widespread weather and pest conditions. The farm-specific shocks were relatively highest for salmon but also for sheep and potatoes the proportion was considerable. Generally, farm-specific randomness may be caused by fluctuating farm-specific weather and disease conditions and variable management decisions. Table 2 shows estimates of within-farm yield correlations. Weather is the primary factor influencing crop yields and crops within the same growing season experience the same weather. Since crops are susceptible to different insects and diseases and do not react exactly in the same way on the weather crop yield correlations often tend to be moderately positive as found in Table 2. Crop yields and livestock performance were less closely correlated. Production rates among different types of livestock were little correlated. The main findings are in concurrence with a similar study in Denmark (Rasmussen, 1997).


365

Table 2 Within-farm yield correlations between agricultural enterprises1

Oats Wheat Potato Forage Milk Sheep Goat Piglets Barley 0.53 0.43 0.39 0.33 0.18 0.26 0.20 0.17 Oats 0.44 0.30 0.22 0.12 0.23 0.02 Wheat 0.31 0.20 0.14 -0.06 -0.04 Potato 0.69 -0.18 0.09 -0.31 0.06 Forage 0.00 0.08 -0.53 0.35 Milk 0.04 -0.13 0.12 Sheep 0.02 0.10 Goat 0.51

1 Bold numbers significantly different from 0 at 5% level

Price variability and correlation Table 3 shows the price variability results. Potato prices exhibited the largest relative price variability within farms (CV=68%), followed by salmon (CV=40%). The prices of the other agricultural commodities were fairly stable with CV’s around 10 to 20%. Farmer cooperatives’ market regulations within the maximum prices set by the government and supply control in milk production have tempered price fluctuations. Why was the potato price more variable than prices for salmon determined in fluctuating world markets? Potato growers face a greater exposure to market prices than other farmers as there are fewer market regulations. Prices are volatile due to the inelastic nature of the demand for potatoes and variations in supply between seasons. A much higher price variability for potatoes than for other agricultural commodities has also been found in Denmark (Rasmussen, 1997). Table 3 Estimated product price variability

Variance components

Mean prices

CV within farms1

Time-invariant,

farm-specific

Time-invariant,

region-specific

Time-variant, region-specific

Time-variant,

farm specific

Barley, NOK/kg 2.27 0.16 0.38 0.09 0.34 0.18 Oats, NOK/kg 2.08 0.16 0.39 0.11 0.29 0.20 Wheat, NOK/kg 2.77 0.18 0.30 0.22 0.31 0.17 Potato, NOK/kg 2.36 0.68 0.22 0.23 0.27 0.28 Milk, NOK/L 4.24 0.15 0.29 0.11 0.37 0.23 Beef, NOK/kg 42.47 0.18 0.34 0.03 0.40 0.22 Lamb, NOK/kg 47.96 0.17 0.29 0.10 0.25 0.36 Goat milk, NOK/L 6.77 0.11 0.30 0.09 0.32 0.29 Piglets, NOK 867 0.15 0.42 0.06 0.29 0.22 Pork, NOK/kg 25.36 0.15 0.35 0.06 0.39 0.21 Salmon, NOK/kg 37.93 0.40 0.22 0.05 0.53 0.21

1 Mean of the farms

For most products more than 50% of the price variability was attributable to factors that are not consistent in time (Table 3). The time-variant, region-specific component was usually larger than the time-variant, farm-specific component, indicating that variation in prices between years in a region is more important


366

than specific farm shocks. The time-invariant, region-specific component was particularly large for salmon while the farm-specific shocks were highest for lamb. Table 4 reports the magnitude of price-yield correlations at the farm level. Potatoes exhibited the strongest negative price-yield correlation followed by salmon; cf. their higher exposure to competitive markets for outputs. The price-yield correlations for the other commodities were moderate. The causes of the positive correlations for milk and goat milk may be associated with animal health performance. Animal diseases can result in lower yields as well as deteriorated milk quality implying a reduced milk price.

Table 4 Farm-level price-yield correlations1

Correlation

(price - yield) Barley -0,02 Oats -0,18 Wheat -0,10 Potato -0,58 Milk 0,26 Lamb -0,06 Goat milk 0,31 Piglets -0,11 Pork -0,03 Salmon -0,50

1 Bold numbers significantly different from 0 at 5% level

Variability in Economic Returns

The variability in return on assets (ROA) is reported in Table 5. ROA was highest in salmon farming with an average return of 9.2%. All of the agricultural farm types showed a negative average ROA. There were larger within-farm variations between years on salmon farms (SD of 19.1%) than in agricultural farm types4. It appears that farm-specific factors drive the majority of variations in financial performance of salmon farms. Approximately half of the financial variability was explained by time-variant, farm-specific factors while 30% could be explained by consistent farm-factors. These findings suggest that uncertain factors like variable weather conditions and time-inconsistent management are more important for determining average financial performance in salmon farming than differences between farms in managerial skills, biophysical resources, etc. The average profitability quotient (PQ) for agriculture in total was 58, i.e., a return to capital and labour far below their opportunity costs (Table 6). PQ was lowest for sheep with 35, and highest for grain/potatoes (70) and grain/hog (76). Hence, substantial variability existed among farm types in terms of PQ’s. The farm types grain/potatoes, grain and grain/hog showed the greatest economic return variability at SD within-farms around 40%. Dairying is often believed to have relatively low income variability over time, and the variability was actually lowest for dairy and milk goat farms.

4 Due to the potential problems with negative ROA’s, financial performance within agriculture will be further discussed using the profitability quotient.


367

Table 5 Variability in economic returns (return on assets, %)

Variance components

Farm type

Mean values

SD within farms1

Time-invariant,

farm-specific

Time-invariant,

region-specific


Time-variant,

farm specific

Dairy -9.14 9.96 0.59 0.04 0.11 0.25 Sheep -25.20 14.30 0.33 0.28 0.21 0.18 Goat -12.80 14.18 0.53 0.10 0.19 0.18 Grain -5.14 11.70 0.46 0.10 0.22 0.22 Grain and hog -0.64 13.11 0.13 0.65 0.14 0.08 Grain and potato -2.76 18.23 0.27 0.29 0.31 0.13 Aquaculture 9.19 19.11 0.30 0.01 0.18 0.51

1 Mean of the farms

The farm-specific factors did also drive the majority of variations in financial performance of livestock enterprises (Table 6). The regional components was more important for crop farms than for livestock operations, maybe related to the greater dependency of favourable region-wide weather conditions for financial success in cropping. A larger fraction of variability in financial performance for agricultural businesses than for salmon farms was attributable to temporally consistent factors. The difference in financial performance between farms (within a farm type) was thus strongly related to factors associated with permanent farmer skills and location (soil properties, topography, etc.). Table 6 Variability in economic returns in agriculture (profitability quotients)

Variance components

Farm type

Mean values

SD within farms1

Time-invariant,

farm-specific

Time-invariant,

region-specific


Time-variant,

farm specific

Agriculture 58 29.4 0.45 0.14 0.07 0.34 Dairy 61 19.2 0.54 0.05 0.12 0.29 Sheep 35 25.1 0.49 0.09 0.17 0.25 Goat 67 21.8 0.44 0.12 0.19 0.25 Grain 53 43.3 0.37 0.28 0.11 0.24 Grain and hog 76 39.0 0.25 0.33 0.19 0.23 Grain and potato 70 43.4 0.31 0.25 0.23 0.21

1 Mean of the farms

Is Salmon Farming More Risky?

We found in general higher variability in yields, prices and economic returns in salmon farming than in agricultural businesses. But is salmon farming more risky? Fig. 1 shows the empirical cumulative distribution functions (CDF) for ROA’s in the businesses.


368

Fig. 1 Cumulative distribution functions for return on assets in salmon farming and agricultural businesses

0,0

0,2

0,4

0,6

0,8

1,0

-120 -80 -40 0 40 80

Return on assets, %

Cu

mu

lati

ve p

rob

ab

ilit

y

Dairy Sheep Goat

Grain Grain and hog Grain and potato

Salmon

ROA from salmon farming was the most variable, since the CDF for salmon is less steep than the agricultural enterprises. However, we should not equate higher variability of returns with more risk. The CDF’s show that salmon farming first degree stochastic dominates the sheep, goat and grain/potatoes enterprises, since at any given probability level the value of returns from salmon farming is greater than that from these agricultural enterprises. Salmon farming was preferred to grain/hog by second degree stochastic dominance (SSD), since at any given probability level the accumulated returns from salmon was greater than the accumulated returns from grain/hog. The minimum ROA for the dairy and grain enterprises were higher than the minimum for salmon farming. Then salmon farming cannot dominate dairy and grain in the sense of SSD. However, by inspection of the CDF’s, a decision-maker should be extremely risk averse (i.e. give extremely weight to the lower left-tails of the CDF) to rank dairy and grain equally efficient as salmon farming. Out of, e.g., 100 outcomes salmon farming will have highest ROA in more than 96 of them, and the upside gains of salmon farming are substantial. Conclusions Results indicate that the year-to-year variability in yields, prices and economic returns at the farm level was larger in salmon farming than in agricultural enterprises. The only exception was higher price variability for potatoes. The variability in livestock enterprises was generally lower than for crop enterprises. Even though salmon farming offered more volatile economic returns than agricultural enterprises, stochastic dominance tests of the distribution of economic returns from the businesses showed salmon farming to be dominant over all agricultural businesses except dairy and grain. The substantial upside gains of salmon farming should also make it more economically attractive than dairy or grain for all except the extremely risk-averse decision-makers. In summary, it appears that the distribution of economic returns in salmon farming was to be preferred to that of agricultural businesses. The findings do not imply that agriculturists should switch to aquaculture. However, since only salmon farming has


369

been attractive from an investor’s perspective, it may help to explain why salmon farming has converted from family firms into large corporate ownership, while agriculture has remained in small, family-based firms. References Atwood, J., Shaik, S., Watts, M., 2003. Are crop yields normally distributed? A re-examination.

American Journal of Agricultural Economics 85, 888-901. Hardaker, J.B., Huirne, R.B.M., Anderson, J.R., Lien, G., 2004. Coping with Risk in Agriculture, 2nd ed.

CABI Publishing, Wallingford. March, J.G., Shapira, Z., 1987. Managerial perspectives on risk and risk taking. Management Science 33,

1404-1418. Rasmussen, S., 1997. Yield and price variability in Danish agriculture: an empirical analysis. In: Huirne,

R.B.M., Hardaker, J.B., Dijkhuizen, A.A. (eds.), Risk Management Strategies in Agriculture – State of the Art and Future Perspectives. Backhuys Publishers, Leden, the Netherlands, pp. 37-44.

Webster, J.P.G., Williams, N.T., 1988. Changes in cereal production and yield variability on farms in

South East England. Journal of Agricultural Economics 39, 255-262.


370

THE INTRODUCTION OF A SUPPLY CHAIN/CONSUMER FOCUS IN FARMER CONTROLLED BUSINESSES IN THE UK

Francisco Gonzalez-Diaz,

Royal Agricultural College Email: [email protected]

David Newton Royal Agricultural College


John C Alliston Royal Agricultural College


Research Sponsors:

Abstract This paper assesses the effectiveness of the traditional models of cooperation and analyses best practice in collaboration between farmers seeking to gain significantly greater scale and flexibility in an increasingly global food chain. The UK grocery retail sector is one of the most concentrated in Europe and the total income from farming is estimated to have fallen 11% in real terms in comparison with 2004. Drawing upon broader strategic management thinking and relevant international practice, the study seeks to identify new feasible models for farmer collaboration. Existing forms and attitudes towards traditional business cooperation in the UK farming sector will continue to be insufficient to gain adequate market power and profitability within an increasingly competitive global food and farming industry. UK farmer collaboration needs to address the global scale of the food supply chain and to do so is likely to require a radical re-think of the most appropriate business structures and alliances. Keywords: supply chain, UK, model, farmer collaboration Introduction The Foot and Mouth Disease caused a huge trauma in the UK farming industry. In August 2001 the Government appointed the Policy Commission on the Future of Farming and Food, chaired by Sir Donald Curry CBE, to advise the Government on how to create a sustainable and competitive farming and food sector. The starting point for the commission was that the situation in the English farm and food industry was completely unsustainable. The diagnosis indicated that farming was detached from the other sectors of the economy and was “serving nobody well” (Curry, 2002). Among the recommendations of the commission, there was an important emphasis on the need to increase collaboration because it was seen as the best way for small farm businesses to obtain the benefits of being a large farm business (Curry, 2002). The need for strengthening collaboration within the food supply


371

chain was identified as so important, that one of the immediate consequences of the recommendations was the creation of English Farming and Food Partnership (EFFP) in 2003. The need for a radical change in the UK farming industry has been quite clear since the Mid Term Review (MTR) of the Common Agricultural Policy. Nowadays, food and farming is a global business, and the people involved in the sector have to recognise the urgent need of adapting to the new economic environment. Experts in farmer collaboration such as Parnell (1999) and The Plunkett Foundation (1992) had previously made clear the need for bigger, better, more effective and efficient Farmer Controlled Businesses (FCBs), and also set the challenge to explore and evaluate new approaches to develop farmer controlled enterprises more imaginatively. Aim and Objectives The main aim is to identify new forms of collaboration between farmers with the need to gain significantly greater scale and flexibility in an increasingly global food chain. The research objectives are: To assess the effectiveness of the traditional models of cooperation. To develop a new model of cooperation within the food chain from which UK farmers can achieve greater competitiveness

Procedures Primary research has been limited to EU member countries (primarily Spain and France) as the legislative framework of the Common Agricultural Policy (CAP) significantly limits opportunities for the transfer of operating models from a non-EU business environment. Using an inductive grounded theory approach comprising a series of Delphi iterative face to face interviews, two rounds of guided interviews were completed. These comprise, 35 experts in the field of business collaboration, selected using a purposive sampling approach. Interviewees include leading academics, government officials and advisors, and managers of the most profitable and/or innovative EU-based collaborative ventures. The objective of the first round of interviews was to identify the parameters of best practice and develop a working hypothesis of how current cooperative models might be supported or challenged effectively. The second stage was completed in order to refine these frameworks using expert opinion, particularly those who are dealing with farmers on a daily basis, in order to gather a closer and more practical view. The Food Industry The economic reality is clear; almost every sector of the UK’s food industry is suffering from over capacity and lack of investment. In addition to this, the supermarkets are cutting the number of suppliers, and in this reduction process they only deal with the companies that are able to deliver both the required scale and quality standards (Key Note, 2004). According to the Institute of Grocery Distribution (IGD) (2005) the UK grocery retail sector is one of the most concentrated in Europe. A measure of the market share of the 3 or 5 largest firms shows that the top five grocery retailers have 63%, and the top 3 firms 48%.


372

At the same time, the total income from farming is estimated to have fallen by 8.9% in current prices, or by 11% in real terms in comparison with 2004, reaching £2.5 billion in 2005. In real terms the total income has slipped below the levels of the late eighties, but is still 40% above the lowest point of 2000 (Defra, 2005). However, the London Economics (2004,) investigated the main determinants of the farm –retail price spread and concludes that: UK farm gate-retail price spreads are generally not among the highest in the EU Member States The results suggest that, overall and for the period covered by the analysis, concentration in the retail domestic market does not seem to have a significant impact in the evolution of spreads The sterling/euro exchange rate and costs in the supply chain, appear to increase retail farm spreads in most commodities groups, therefore, this would suggest that UK farm products are subject to significant competition from countries inside the Euro area. Farmer Controlled Business (FCBs) Regarding collaborative ventures and in order to compare the relative importance of FCBs in different countries, table1 shows their market share per sector for a selection of EU countries. Table 1. Market share of UK & non-UK FCBs 2001(%)

Dairy Fruit & Veg Meat Farm inputs Grain France 49 35-50 27-88 50-60 75 Germany 55-60 60 30 50-60 Denmark 93 20 – 25 66 - 93 59 - 64 87 Netherlands 82 70 – 96 35 40 – 50 Sweden 99 60 79 - 81 75 75 UK 50 35 – 40 20 20 - 25 20 Spain 45 35 35 50 35

Source: EFFP (2004) and MAPA (2003)

It is clear from table 1 that the UK FCBs hold a quite small proportion of the market if compared with other EU countries, particularly northern European countries. Even in sectors like meat, where the market share of UK FCBs has been traditionally very small, other European countries have FCBs with significant market power. Another way to evaluate the importance of the FCBs sector is to compare the cooperative turnover with the total agricultural output. Graph 1 shows that doing such a comparison the English FCB sector is quite small in relation to a selection of European and North American countries.


373

Graph 1 Co – operative turnover compared to agricultural output in Europe.

Source: EFFP (2004) It is also interesting to highlight the fact that in many countries the Cooperative turnover is higher, and in some cases more than double the total agricultural output, representing the significant importance of the FCB sector for the farm industry. The comparison of the top 30 FCBs in England with the top 30 FCBs in Europe is shown in Graph 2. The difference in turnover between both groups is very significant. Additionally, it can be added that all of the top 30 EU FCBs had in 2005 turnovers above £1 billion. Graph 2 Turnover of the top 30 FCBs in England Compared with EU

Source: EFFP (2005) Globalisation Fulton (2000) said that globalisation has increased both the rate and the nature of social and economic change, through rapid advances in technologies, the declining of the nation-state and its consequent borders, the fluid movement of goods and people and the blending of cultures. Thomson (2001) predicts that there are very good opportunities for well organised, market orientated and adequate size producers that can satisfy the needs for price, quality, marketing support and volume.

Cooperative turnover compared to agricultural output

0

50

100

150

200

250

300

Swed

en

Den

mar

k

Net

herla

nds

Franc

e

Ger

man

y

Can

ada

Unite

d Sta

tes

England

%


374

In the UK, English Food and Farming Partnerships (EFFP) (2006) emphasises that it is quite clear that in the future the competition will be between global supply chain and not within the supply chain, opening opportunities for world class FCBs. For that reason, EFFP suggests that English FCBs should change to a more global market driven mindset in order to exploit international market opportunities. Fulton and Gibbings (2000) said that globalisation and an industrialised agriculture are very related. Globalisation has increased the social economic changes, affecting the role of governments regarding agriculture by increasing their regulations on environment, health and food safety. Simultaneously, the consumers are widening their demands for different products, and changes in technology are affecting the way food is produced, processed and distributed Thinking about the competitiveness of the UK farming sector in this new era, Hampson (2006) reports that even the most cost effective UK farmers might not be able to compete with the low-cost imports on price alone. This suggests that market differentiation will be the way forward to keep farmers’ viability in a globalised market. Torgerson (2004) argues that direct payments to farmers have acted as incentives to increase farm size, but also as a barrier to farmers organizing together in order to increase their income from the market. In other words, farmers traditionally have relied on government support as their major source of income isolating them from the market forces. The New Way In her research about cooperation in the UK red meat sector, Bowles (2004) said that the different social economic environment that is affecting the UK faming, in comparison with the economic environment at the times when agriculture cooperation started and developed in mainland Europe, requires different cooperative models able to achieve significant scale of operation in the short term. To gain more value for farmers, FCBs should change from a defensive cooperative model to a more offensive and risk-orientated model, which has to be able to attract investments. The traditional model of cooperation has its limitations in being able to raise capital from non-members; therefore there is a need for new models that are more attractive to external investors (EFFP, 2006). “Inevitably, the traditional organization of ownership, control, and business conduct of cooperatives restricts vertical expansion into value-added activities, exploitation of market opportunities at both farm and processing level, and creation of superior customer value. As a matter of fact, the organizational arrangements of traditional cooperatives hinders them from making their escape from production-orientated to market-orientated business.” Kyriakopolus and van Bekkum (1999) Thelwell (2004) argues that the fragmentation at production level is the main barrier for UK farmers to gain more economic and political power. Therefore, collaboration is the only way to gain the required scale to influence the market, and to do so, farmers will have to compromise some of their individual freedom and start to invest in their market beyond the farm gate in order to change from production orientated operations to market led ones. Traditional cooperatives need to be adapted to the post-modern business environment (Goldsmith, 2004) The message is clear: to achieve real benefits there is a need for a change in the mindset of all businesses involve in the UK supply chain (English Food & Farming Partnerships, 2006). British farming has to approach the new political and market developments with an alternative “thinking” (Askew, 2006).


375

Results: The Models The outcome of the research has been to develop three discreet but combinable models of collaboration. Each model requires different levels of commitment from its members and would suit different business situations. It is assumed that the prime consideration for members of any organization of primary producers is a desire to receive enhanced financial benefit from their participation. Each model reflects the detailed comments of interviewees. Model 1: NETASSOC This model allows a group of farmers (also industry-related non-farmer businesses) to collaborate in a formal business relationship. The volume, price and quality of the products are agreed in advance, (there must be at least a clear description of the products to be traded). In some cases it will be possible to have standardised contracts between the members and guidelines about the operational requirements to participate in the NetAssoc. For example, a beef farmer and a finisher might be members of a NetAssoc. The farmer might agree to sell his calves to a finisher. The number of calves, the breed, the weights, delivery dates, etc, would have been agreed in advance. Members would be registered with the NetAssoc but would be flexible in agreeing to contracts in any one trading period. This would allow a better coordination of the chain; increasing the efficiency and the quality of the final product. Here supply and demand are matched within a flexible and yet agreed framework where the parties would have redress to law. Interaction among members would need little additional supervision or control. The level of organisation can vary, being loose or tight at varying times. There are no entry barriers beyond acceptance by existing members of the basic rules of the NetAssoc and no exit barriers other than the restrictions of an individual contract. Co-ordination could be shared by the members to reduce overheads. Collectively members of the NetAssoc might decide to bid for contracts – if successful they may decide to appoint professional co-ordinators. However a simple database of contracts would also suffice. The NetAssoc thus promotes the matching of supply and demand. Production quality is determined by individual contracts. Increased communication, improved flows of information and increased mutual trust and dependency should result. It is extremely flexible for the participants and no initial investment is required. Basic governance would develop only on the basis of success and mutual agreement. As scale increased the members of the NetAssoc might conceivably vote to become a group member of a NetCoop.

Model 2: NETCOOP This model is an adaptation of the traditional model of cooperatives. Members have to acquire “rights” to participate in the coop (buying or delivering products). The number of rights purchased for each member will be in relation to the amount of products allowed to be traded, and will relate to the voting power of the member, increasing the commitment and the sense of ownership towards the NetCoop. Members will receive market price for their products, and a further “bonus” which represents the ability of the NetCoop to add value to the inputs, allowing a clear differentiation between the product delivered by the member and the performance of the NetCoop as a business. This kind of procedure drives the producers to increase the quality of their production and it is a good way to evaluate the performance of the NetCoop’s management team. Where the NetCoop operates in more than one sector (business or products), the member’s “rights” will determine the number of “participations” acquired by a particular member. These “participations”


376

represent the share of the whole NetCoop that belongs to each member and therefore relate to the voting power i.e. a member will receive market price for the products traded (x amount of potatoes), plus a later “bonus” based upon the added value of the product sector (performance of the potatoes business within the NetCoop), plus a “participation dividend” related to the profitability of the NetCoop as a whole. This increases the global vision of the business and spreads the risk of the membership. The valuation of the members “rights” will vary with the overall performance of the Netcoop and these rights will be tradable and may therefore offer the opportunity for a capital gain. The model offers many options and could be the basis for a federal model, where individual producers own the ‘rights’ and a first tier coop or Net Assoc owns the ‘participations’ in a larger NetCoop (related to the number of rights of its members), thus giving them the voting power and participation in the overall performance of the federated coop. The performance of the NetCoop and a real sense of ownership are at the centre of this model, compelling forces to drive a focus upon the needs for consumer and a broader vision of the requirements of the supply chain. It is also a flexible structure more suited to a complex and changing business environment.

Model 3: NETBUS The NetBus is a traditional registered company, with the key difference that its shareholders are other business/companies who are participants of a specific supply chain. It is a network of businesses that form a company, bringing integration, coordination and flexibility to the supply chain. The members form a company in order to increase collaboration and commitment with the common objective of long-term sustainability and competitiveness. Because every stage (every individual business member) of the chain owns shares in the company, there will be clear benefits from the sharing of information and the seeking of maximum efficiency at every stage. The structure could be a horizontal or a vertical organization, so the possible shareholders are: primary producers, processors, input companies, traders, financial institutions, service companies, universities, NetCoops, FCB, and so on. Everybody buys shares, participates in the profits, and the company is run as a normal profitable business. Therefore, each member has to deliver (products or services) to meet the company expectations, otherwise it should be provided by someone else. These kinds of requirements will pressure each member to be the best in their particular area, and the return will come as dividends and as an increase in the share price. Some restrictions are necessary to ensure the primary producers maintain majority ownership, and the NetBus may decide to limit membership. This could be done by the introduction of different type of shares with different rights over profits and over voting power. There is a huge potential for the synergy coming from the participation and commitment of businesses from different elements of a supply chain. Without doubt it will increase the coordination, efficiency and long-term competitiveness of the participating members. Bargaining power increases exponentially allied to a significantly improved flow of commercial information between its participants. Because of the integrative nature of such an organisation, there will be a need to demonstrate compliance with anti-trust or Office of Fair Trading regulations, however it is an ideal model to compete against other supply chains and to develop new products or enter competitive markets. Conclusions The proposed models seek to overcome the perceived limitations of traditional models of UK Farmer Controlled Businesses in the new economic environment. The second round of interviews presented the


377

business models to experts in the field of collaboration in order to refine them and adjust them. The overall conclusions are: The models offer a different framework that will increase the consumer/supply chain focus and the flexibility required by UK FCBs to increase their competitiveness. The payments of dividends or bonuses act as a more visible benefit to being part of the organization. There needs to be established the right to trade ownership and a financial framework attractive to external capital. The proposed models offer mechanisms that increase the motivation for participation and develop a stronger sense of ownership that will be reflected in higher commitment. The models 2 and 3 provide a better environment for the development of the management, increasing the possibility of better performance, control and career options. The models offer new opportunities for proactive new members with a continual interest in increasing their level of participations. Such participating members will support the recruitment of well-educated and/or experienced leaders. References Askew, M. F. (2006) The challenge of alternative crops. The national farm management conference, 15th-

16th November 2006, Marston Hampshire Centrecourt Hotel, Basingstoke, England. Bowles, L. (2004). Cooperation and collaboration in the red meat sector. Nuffield Farming Scholarship

Trust Custer, R.L; Scarcella, J.A. and Stewart, B.R (1999). The modified Delphi Technique: a rotational

Modification. Journal of Vocational and Technical Education. Vol 15, No2, pp50-58 Defra (2005). Agriculture in the United Kingdom 2005. Defra English Farming and Food Partnerships et al. (2004). Farming and food: collaboration for profit. EFFP English Farming and Food Partnerships (2005). Performance indicators for the English FCB sector 2005.

EFFP English Food and Farming Partnerships (2006). FCBs key to growth. News issue 6, November. English

Food and Farming Partnerships. Fulton, M. and Gibbings, J. (2000). Response and adaptation: Canadian agricultural cooperatives in the

21st century. In, Canadian Agricultural Cooperatives: critical success factors in the 21st century. Goldsmith, P. (2004). Creating value in a Knowledge-based agriculture: a theory of new generation

cooperatives. In Cooperatives and local development: theory and applications for the 21st century. Ed Merrett, C.D. and Walzer, N. M.E.Sharpe.

Hampson, S (2006). Differentiation: a sustainable future for UK agriculture. Royal Agricultural Society

of England. IGD (2005). Grocery retailing 2005. IGD


378

Key Note; Smith, P. (ed) (2004). Key note: market forecast. Key Note Kyriakopoulus, K. and van Bekkum, O. F. (1999) Market Orientation of European Agricultural

Cooperatives: strategic and structural issues. In, IXth Congress of the European Association of Agricultural Economists, August 1999, Warsaw, Poland

London Economics (2004). Investigation of the determinants of farm-retail price spreads, p.93-94.

London Economics and Defra Parnell, E. (1999) Making cooperation work: rights, duties and internal contracts. In, The world of

cooperative enterprise (1998) Plunkett Foundation. Plunkett Foundation (1992). Farmer controlled business: bringing balance to the market place. Plunkett

Foundation. Policy Commission on the Future of Farming and Food (2002). Farming and food: a sustainable future:

report. Policy Commission on the Future of Farming and Food [chairman Donald Curry] Saunders, M; Lewis, P; Thornhill, A. (2003) Research methods for business students (3rd ed). Pearson

Education Thelwall, D. (2004). Raising the game. Royal Agricultural Society of England. Prospect Management

Services. Thompson, G. (2001) Supply chain management: building partnerships and alliances in international food

and agribusiness. Rural Industries Research and Development Corporation. RIRDC Publications. (Publication No 01/31).

Torgeston, R. E. (2004). Producer marketing through cooperatives. In Cooperatives and local

development: theory and applications for the 21st century. Edited by Merrett, C D and Walzer, N.; M.E. Sharpe: 224-246.


379

ACHIEVING SUSTAINED IMPROVEMENTS IN PROFITABILITY IN BEEF ENTERPRISES AND REGIONS IN SOUTH AFRICA AND AUSTRALIA

.

Richard Clarkad, Garry Griffithbd, Percy Madzivhandilac, Baldwin Nengovhelac, Peter Parnellbd and Janice Timmsad

a Queensland Department of Primary Industries, 80 Ann Street, Brisbane, QLD 4000, Australia.

b NSW Department of Primary Industries, Beef Industry Centre, University of New England, Armidale, NSW 2351, Australia. c Agricultural Research Council, Animal Production Institute, Irene, South Africa.

d CRC for Beef Genetic Technologies, University of New England, Armidale, NSW 2351, Australia.

Abstract Contemporary Research and Development (R&D) projects are increasingly concerned with setting outcome targets such as measurable improvements in enterprise profit, industry growth, environmental health or resource use efficiency. This paper describes the application and ongoing development of a ‘Sustainable Improvement and Innovation’ (SI&I) model for designing and managing medium to longer term R&D projects to achieve and sustain outcomes, improvements and innovations in agricultural enterprises, industries and regions. The model is applied in two Beef Profit Partnerships (BPP) projects, one by emerging farmers in two provinces in South Africa, and the other by commercial beef producers in Australia. The South African project was funded by the Australian Centre for International Agricultural Research over five years; the Australian project is funded by the Cooperative Research Centre for Beef Genetic Technologies and is in the second of seven years. The primary objective of both projects is to accelerate the rate of adoption of new technologies and to measure and monitor the productivity and profitability outcomes in farming systems and in the broader region. Another objective is to undertake adoption science research on this process and to measure and monitor partnership and capacity building outcomes. Through the use of the model in the South African BPP project, measurable improvements in profit per beef enterprise, each year, in the participating communities and regions have been achieved, and these improvements have been sustained across an increasing number of cattle enterprises and communities. Further, the SI&I process implemented as an integral part of the BPP project has been demonstrated to lead to measurable, positive economic outcomes, even over a relatively short period. The SI&I model is presented in detail, the results from the South African application are reported, and the value of the model and its methods for extending the approach to Australia and elsewhere are discussed. Conclusions are made about the application of the SI&I model in R&D projects focused on high rates and scales of impact. Keywords: R&D projects; outcomes; continuous improvement; continuous innovation; partnerships; systems


380

Introduction Rural Research and Development (R&D) projects are being criticised for poor achievement of outcomes, rates and scales of impacts, and ongoing improvement and innovation during and after the end of projects1. R&D investors often settle for outputs like the development of new knowledge, technical products, information packages, publications, or a new or improved practice that, if used or adopted, could provide outcomes/benefits. They fall short of setting outcome targets such as measurable improvements in regional enterprise and industry profit, environmental health or resource use efficiency. This paper reports on the application and development of a Sustainable Improvement and Innovation (SI&I) model to achieve and sustain outcomes, improvements and innovations (at a regional scale) in 3-5 year project timeframes. The SI&I model is an integral part of the Beef Profit Partnerships (BPP) project that has been implemented in South Africa and Australia. In this paper, the SI&I model is presented in detail, the results from the South African application are reported, and the value of the model and its methods for extending the approach to Australia and elsewhere are discussed. Conclusions are made about the application of the SI&I model in R&D projects focused on high rates and scales of impact. The Sustainable Improvement and Innovation Model There is a growing literature that advocates the design and development of a human/social system to achieve and sustain improvements and innovations at a regional and/or national scale. System-based approaches to project management have been advocated for a long time. Such approaches: help to identify, understand and work with those elements essential to achieving target outcomes, and the key relationships and interdependencies between these components; help to better visualise, understand and develop a shared mental-model of the whole system; and provide a practical, easy to use project management framework for thinking, implementation, regular measurement/assessment, and continuous improvement of project performance. The application of an outcome-focused, whole-system model could overcome constraints often experienced in R&D projects which produce outputs but fail to achieve outcomes within project timeframes, and could increase the ease and efficiency with which target outcomes are achieved. Figure 1 shows the six interconnected elements of the SI&I model that have been identified using systems thinking tools. The level of input required in each interconnected element (in each timeframe) is also shown, based on our experience.

1 There is an extensive reference list that accompanies this material on the background and implementation of the SI&I model. For space reasons references are omitted from this version, but a full version of the paper may be obtained from the corresponding author.


381

Figure 1: The interconnected elements of the Sustainable Improvement and Innovation (SI&I) model In the following sections the function and purpose of each element of the SI&I model is described and the key methods and tools used to make each element of the model functional are highlighted. Element 1 – Shared Outcome Focus, Targets and Performance Measures A number of authors have emphasised the value of Focus in achieving improvements and innovations. To achieve satisfying results it is important that people set outcome-based targets rather than activity-based goals. When working in partnerships it is crucial that partners have a shared understanding of target outcomes and the key concepts associated with these outcomes. To sustain improvement and innovation it is essential to make success measurable so that people can see tangible results and be rewarded and motivated from their efforts. Performance measurement drives behaviour and behaviour change, supports the prioritisation of actions and enables comparing and tracking of performance changes and differences. The use of the Critical Success Factors (CSFs) method enables people to identify, action and measure those factors critical to success. The measures of performance must align with the purpose of the measurement, thus the identification of Key Performance Indicators (KPIs) with clear links to CSFs and target outcomes is crucial. The purpose of Element 1 is to enable project teams, partners and individuals to develop clear target outcomes, CSFs and timely KPIs to focus their thinking and action on achieving and recording results linked to their target outcomes. A key tool that has been used to make Element 1 functional in agricultural R&D projects is Focusing Frameworks. Focusing Frameworks like the one in Figure 2 enable individuals and partners in projects to: develop SMARTT focuses for action; develop a shared understanding of key concepts and terms like profit; develop target outcomes, CSFs and timely KPIs; benefit from the thinking and action of individuals and teams using a shared framework tool; and

LL LLee eevv vvee eell ll oo ooff ff ii ii nn nnpp ppuu uutt tt bb bbyy yy mm mmaa aann nnaa aagg ggee eerr rr ss ss,, ,, ll ll ee eeaa aadd ddee eerr rr ss ss && && aa aacc cchh hhii ii ee eevv vvee eerr rr ss ss

EarlyEarlyEarlyEarly MiddleMiddleMiddleMiddle LateLateLateLate

1 1 1 1 Shared FocusShared FocusShared FocusShared Focus,,,, Targets Targets Targets Targets &&&& Performance MeasuresPerformance MeasuresPerformance MeasuresPerformance Measures

2 2 2 2 NetworkNetworkNetworkNetwork Design Design Design Design & & & & ManagementManagementManagementManagement3 3 3 3 Technology Technology Technology Technology & & & & Information Development Information Development Information Development Information Development & & & & IntegrationIntegrationIntegrationIntegration

4 4 4 4 Continuous Improvement Continuous Improvement Continuous Improvement Continuous Improvement & & & & InnovationInnovationInnovationInnovation5 5 5 5 Capacity BuildingCapacity BuildingCapacity BuildingCapacity Building6666 Growth Growth Growth Growth & & & & Momentum ManagementMomentum ManagementMomentum ManagementMomentum Management

6666

2222

3333

4444

5555

1111

TimingTimingTimingTiming


382

give focused reports on, and mutual support for, impact on targets. The power of Focusing Frameworks can be enhanced by developing them into Performance Management Frameworks and strategies for project teams and individuals. When Focusing Frameworks become a shared conceptual model they enhance collaborative efforts to achieve targets. Figure 1 shows from our experience that the input required to achieve a shared outcome focus, targets, CSFs and KPIs is relatively high in the early phase of projects but requires less input once the majority of project partners develop a shared understanding. The use of a project glossary has enhanced the achievement of a shared understanding of key concepts and terms like targets, profit, productivity, CSFs and KPIs. Figure 2: “The Beef Profit Driver Tree”

Element 2 – Network Design and Management To enable individuals to achieve improvements and innovations collectively requires the establishment of an effective and sustainable social organisation/infrastructure. The concept and principles of partnership help achieve productive collaborations because they promote mutual responsibility-taking and mutual proactive support. Effective partnerships require necessary functions, roles and responsibilities to be fulfilled through the active involvement of the right proportion of partners in the most appropriate infrastructure. The partnership infrastructure considered most appropriate for SI&I are networks of individuals and teams at local and regional levels. It is estimated that an optimum size for a regional network is about 100 members. But effective regional networks don’t just happen! They need design and management. Figure 3 shows a typical regional network design and management concept. Three key groups are: Achievers i.e. all members of the network; Leaders i.e. about 15% of network members; and Managers i.e. about 5% of network members. The purpose of Element 2 is to enable people interested in achieving target outcomes (Element 1) to build a viable partnership and to work productively as individuals and in teams.

ProfitProfitProfitProfit////EnterpriseEnterpriseEnterpriseEnterpriseR/kg/ha/year

ProfitProfitProfitProfit////RegionRegionRegionRegionR/Region

ProfitProfitProfitProfit////CommunityCommunityCommunityCommunityR/Community

GrowthGrowthGrowthGrowth

Kg/hd/yr

ReprodReprodReprodReprod

%

DeathsDeathsDeathsDeaths

%

ThroughputThroughputThroughputThroughput

Kg/ha/year

MarketingMarketingMarketingMarketing

HealthHealthHealthHealth

CostsCostsCostsCosts

R/kgManagementManagementManagementManagement

PricePricePricePrice

R/kg

ProfitProfitProfitProfit IncomeIncomeIncomeIncome CostsCostsCostsCosts==== ----

==== xxxx ---- CostsCostsCostsCostsThroughputThroughputThroughputThroughputPricePricePricePriceProfitProfitProfitProfit

GeneticsGeneticsGeneticsGenetics

NutritionNutritionNutritionNutrition


383

Key methods used to make this element functional in a project are: explaining and negotiating the benefits, values, key functions and structure required for productive regional improvement and innovation networks; negotiating criteria for involving the right proportion of partners in local teams and regional networks; ensuring the right proportions of Achievers, Leaders and Managers in regional networks; and negotiating shared focuses, CSFs, KPIs and methods with teams and individuals in regional networks. Attrition of vital role-players (and teams) in networks is to be expected and succession should be planned for. The role of local, provincial, national industry, government and academic agencies is crucial for network vitality. It is best if local teams and regional networks are interdependent of, not dependent on, one another. Figure 3: The functions of management, leadership and achievement required for an effective regional improvement and innovation network

Element 3 – Technology and Information Development and Use In a well-planned and sustainable society, it is not simply the availability of new technologies that fuels economic growth and sustained productivity, but more the wise adoption, adaptation and application of those technologies. To achieve sustainable improvement and innovation, the on-going generation and use of new knowledge, information and technology is required. The purpose of Element 3 is to enhance the research, development and use of technology and information to achieve target outcomes year by year. Methods need to be used to ensure that specific items of information and technology are identified and linked to end-user needs and feedback i.e. using the principles of “market in” rather than “product out”. Focusing Frameworks (Figure 2), and benefit/cost analysis tools like gross margins or whole-farm financial analyses enable the potential impact of information and technologies on targets, CSFs and KPIs to be identified, assessed and prioritised for development and use. The development of information and technology needs to be a dynamic process closely integrated with the end-user needs and input. Timely feedback is vital to overcome constraints to the adoption of technologies - this is enabled by strong integration with Element 4.

FOCUS FOCUS FOCUS FOCUS

FOCUS

FOCUSFOCUS

ManagingOrganizing a supportive system

LeadingInspiring thinking & action

AchievingAchieving results

Regional Improvement & Innovation Network

FOCUSFOCUS


384

Element 4 – Continuous Improvement and Innovation Through experience of useful activity and memories, individuals form a store of useful knowledge. In this way organisms improve their methods. This continual refinement and improvement is called growth: “a self renewing process through action upon the environment”. To achieve impact and accumulated (enterprise and industry) growth on a large scale it is necessary that individuals, and teams and networks use effective and efficient shared processes. The purpose of Element 4 is to enable all project partners to develop and use a shared process to focus on targets and CSFs, benchmark and measure KPIs, take focused action, seek feedback and support, and achieve high rates of improvement and growth. A key methodology is the Continuous Improvement and Innovation (CI&I) process. To be of value in a partnership targets, CSFs and KPIs need to be meaningful and easily shared so that they can be used to identify and promote methods that achieve success i.e. Evidence-based Practice. KPIs need to give early and meaningful indication if actions are achieving impact, or not. The timing of CI&I activities at 30, 90, 180 and 360-days has been found to be useful to achieve motivating progress, supportive feedback, the creation of new ideas and opportunities, high rates of improvements per year, and to manage the dynamics of enterprises and networks. The process of CI&I is not taught at school and it is not currently commonly used in many rural communities and industries. Building the capacity of leaders of regional teams and networks to achieve continuous improvements and innovations in beef enterprises and communities now and in the future is one of the more difficult tasks of implementing the SI&I model.

Element 5 – Capacity building All the partners in SI&I projects and networks need to be equipped with the necessary capacity to fulfil their functions and roles, and for sustainability, communities need to be equipped to design their own systems and processes – not have these done to or for them. The purpose of Element 5 is to equip BPP partners with the knowledge, tools, technologies, skills and support at appropriate times to achieve sustained improvement in profit per beef enterprise, per year, in a growing number of enterprises, communities and regions. Key methods are the use of skills training in appropriate tools, coupled to immediate practise and support, CI&I (Element 4), and Evidence-based Practice. R&D of improved methods for achieving rewarding results regularly also contributes to capacity. Skill’s training is often separately designed for Achieving sustained improvement of profit, Leading regional teams and networks, and Managing regional networks (Figure 3). Training needs to be planned to meet estimated rates of project/network participant/personnel attrition, due to resignations and progressions that create gaps in the capacity of partnerships. It also needs to be linked to the practice of CI&I and Evidence-based Practice, and to be timely and progressive, not repetitive. The level of investment in capacity building is often a potential weak point in a project.


385

Element 6 – Growth and Momentum Management Growth is the accumulated impact over a period of time i.e. the sum over 5 years of the product of the number of improvements and the impact of each of those improvements. Momentum is the level of impetus that sustains growth. Momentum and growth need to be achieved with efficiency to achieve Return on Investment. In dynamic environments it is reasonable to suggest that a never-ending series of initiatives aimed at a constant readjustment and realignment is necessary. The purpose of Element 6 is to achieve impact and accumulated impact per region though effective, efficient and sustainable systems. Key methods include: regular and frequent face-to face meetings of network partners (Element 4) to achieve results, and share results, give feedback and promote proven and successful methods, receive feedback and create opportunities; regular and frequent communications to achieve awareness, understanding, quality relationships, marketing of results and “proof-of-concept” and “proof-of-profit” to specific target audiences; and regular assessment and management of the whole-of-system model for vitality. From our experience the input required to achieve growth and momentum is relatively low in the early phase of projects but requires more input later on with the aim of achieving the targeted state of sustainability by the end of the project. The South African BPP Project The South African Beef Profit Partnerships (BPP) project was initiated in 2000 by the Australian Centre for International Agricultural Research in partnership with the Agricultural Research Council and the Limpopo and North West provincial governments in South Africa. The project was scheduled to end in June 2006, but has been extended one year. The BPP project is targeting improved profits for emerging farmers, who own 40% of the beef cattle breeding herds in South Africa but generate only 5% of cattle sector returns. The income from these enterprises is very low (Tapson 1990). The BPP project was designed to achieve target outcomes from the outset and to sustain outcomes post project. The specific target outcome of the BPP project was: “to achieve sustained improvement in profit per beef enterprise, per year, in a growing number of enterprises, communities and regions, in two provinces in northern and north western South Africa”. Fifteen farmer teams commenced in the BPP project in 2001 and the number had risen to 24 by 2005. These farmer teams routinely measured a number of price, cost and herd productivity KPIs based on the model set out in Figure 2. Following specialised training and capacity building workshops, a subset of the farmer teams also routinely calculated and recorded gross margins for their beef enterprises. For the analysis reported here, the relevant price, cost and herd productivity KPIs were averaged or summed as required across the number of farmer teams reporting each year. These data are given in Table 1. As well, a number of KPIs for each element of the SI&I model described above were developed and routinely assessed by the project management team. These are discussed in detail in Clark et al. (2005b), Timms et al. (2005) and Nengovhela et al. (2007).


386

Table 1: Impact on beef profit CSFs and KPIs year by year in the BPP project (R = Rand)

KPI 2001 2002 2003 2004 2005 Price – Ave R/kg 4.56 8.5 7.13 7.23 8.8 Growth – Ave weight (kg) of calves sold 188 210 205 194 Reproduction Rate - Ave % calves/100 cows mated

43 51 53 62.6 64

Health - Ave pre-weaning mortality % - - 8 3.7 9.32 Throughput – Number sold/year - 23 187 219 389 Based on the recorded data from the farmer teams, the BPP project increased revenue to the emerging farmers involved in the BPP farmer teams by more than 1.25 million Rand over the period 2001-2005 (Madzivhandila et al. 2007). These additional revenues represent between 216 R per farmer team in 2001 to 26,769 R per farmer team in 2005. The average across these five years is 14,358 R per farmer team. Tapson (1990) suggested that prior to the BPP project, an emerging farmer with 25 breeding cows would be able to generate a gross income of only 1,050 R per year from those cattle. From the data in Table 1 we can suggest that Tapson’s farmer would have received an annual income of around 20,000 R if he had been a participant in the BPP project. Based on the recorded gross margin data from the subset of farmer teams, the BPP project increased profits to these teams by 198,610 R over the period 2002-2005. This translates into an average improvement in gross margin due to the BPP project of 9,617 R per selected farmer team per year. Therefore, the BPP project has been able to achieve measurable improvements in profit per beef enterprise, each year, in the participating communities and regions. Overall, at least half and potentially up to 66% of the additional revenue estimated to be attributable to the BPP project would be expected to be retained as additional profits to the participating farmer teams. Thus each Rand spent on improvements in cattle production and marketing has resulted in at least a two Rand return to farmers. These improvements have been sustained across an increasing number of cattle enterprises and communities (Figure 4). The rising trend of additional revenue due to the BPP project is seen to be a function of the increasing number of heavier calves sold by an increasing number of participating farmers (the number of improvements), and the higher per kilogram price received for these calves (the impact per improvement).


387

Figure 4: Accumulated Income Finally, the SI&I process implemented as an integral part of the BPP project has been demonstrated to lead to measurable, positive economic outcomes, even over a relatively short period.

The Australian BPP Project The Sustainable Beef Profit Partnerships (also BPP) Project of the Cooperative Research Centre for Beef Genetic Technologies (Beef CRC) is designed to work in partnerships with beef businesses, value chains and the broader Australian beef industry to accelerate improvements, innovations and adoption and assist in meeting the overall Beef CRC target outcome of $179 million extra profit per year by 2012. The BPP Project has specified the following short-term focus, which all groups are encouraged to adopt: “To achieve an additional 5 per cent improvement in annual business growth among Beef Profit Partners within two years”, and the following target outcomes: Rapid and measurable improvements in productivity, profit and growth; Supportive network of rewarding partnerships, contributing to accelerated industry growth; and Partners equipped to achieve sustainable improvement and innovation. Some 50 BPP groups are being set up across the various beef production environments in Australia and New Zealand. Most of these will be commercial cattle producers, up to 5 will be full supply chains. Each group will have access to a trained facilitator and specialist economic and other technical support as required. Each facilitator and many producer partners have undertaken or will undertake CI&I training. Three particular aspects of this project are noteworthy. First, following the outstanding success of the South African project, the use of a clear shared process of CI&I is advocated to enhance the rate of improvements and innovations. Each partnership is encouraged to adopt CI&I principles and practices to achieve improvements, innovations and adoption, and so assist in meeting the project focus and outcomes, and to measure and report their successes and failures.

0

100

200

300

400

500

600

700

2001 2002 2003 2004 2005

Years

00

0 R

an

d Additional Income 000R

Throughput

Additional Price


388

Second, again based on the experiences in South Africa, a system-wide approach was developed to coordinate and manage the various CI&I partnerships, the linkages between them and their linkages with the broader beef industry, to assist in implementing efficient and effective mechanisms that will achieve the target outcomes. Third, as part of implementing this system approach, the BPP project has designed and is managing a number of formal strategies (Figure 5): Capacity, capability and competency - To ensure partners and industry are equipped and supported to achieve and accelerate improvements and innovations for sustainable impact on business profit and industry growth; Communication, promotion and marketing - To ensure all partners have a shared vision of the project, and that the partnership network and industry are adequately informed of the project achievements, and share and promote improvements and innovations; Research and development - To improve, discover and create more effective and efficient mechanisms to achieve accelerated improvement and innovation; Measuring, monitoring and evaluation - To ensure partners and industry are able to demonstrate achievements and obtain feedback and support to contribute to achieving further improvements and innovations; Partnership and industry support – To achieve momentum and institutionalisation of the CI&I process during and after the project; and System management and improvement – To ensure CI&I principles are applied to all elements, strategies, processes, methodology/mechanisms, human infrastructure and the project system as a whole. Figure 5: Six strategies to ensure effectiveness of CI&I partnerships and networks for beef business profit and growth

The strategy of most interest to this audience is the Measuring, Monitoring and evaluation (MM&E) strategy. Considerable effort has been put into designing and implementing effective and efficient monitoring and recording mechanisms that will assist in achieving the project’s three overall target


389

outcomes. Each of these outcomes has a set of KPIs that can be measured by the producer partners or the group facilitators (Table 2, Appendix 1) (ISNAR 2003). Further details on the Australian project are available in Griffith et al. (2007).

Discussion In this paper it has not been possible to fully describe either the SI&I process, its successful application in South Africa, or its potential application in Australia. The results so far suggest that the model is a valuable addition to the toolkit of extension practitioners and adoption scientists, and that its methods can be readily applied in either a developing or developed country context where the focus of the project is on rapid rates and/or broad scales of impact. References Bessant, J., Caffyn, S., Gilbert, J., Harding, R. and Webb, S. (1994), “Rediscovering continuous

improvement'', Technovation 14(1), 17-29. Chapman, R.L. and Hyland, P.W. (1997), “Continuous improvement strategies across selected Australian

manufacturing sectors'', Benchmarking for Quality Management and Technology 4(3), 175-88. Clark, R., Bacusmo, J., Bond, H., Gabunada, F., Madzivhandila, T.P., Matjuda, L.E., Motiang, D.M.,

Nengovhela, N.B., Taveros, A.A., Timms, J. and Toribio, J. (2005), “A model for achieving sustainable improvement and innovation in regions”, International Conference on Engaging Communities, Brisbane, August.

Griffith, G., Parnell, P., Clark, R. and Timms, J. (2007), "The Beef Profit Partnership approach to

adoption of new technologies", paper presented at the 51st Annual Conference of the Australian Agricultural and Resource Economics Society, Queenstown, New Zealand, 13-16 February.

Hyland, P., Mellor, R., O'Mara, E. and Kondepudi, R. (2000), “A comparison of Australian firms and

their use of continuous improvement tools”, The TQM Magazine 12(2), 117-124. ISNAR (2003), Monitoring, Brochure for the Regional Training Workshop on Evaluation and Impact

Assessment of R&D Investments in Agriculture, organised by the Post Graduate School of Agriculture and Rural Development, Faculty of Natural and Agricultural Science, University of Pretoria, and the International Service for National Agricultural Research, http://www.isnar.cgiar.org/learning/ImpactSept.htm.

Kaplan, R.S. and Norton, D.P. (1992), “The balanced scorecard - measures that drive performance”,

Harvard Business Review. 70 (1), 71-77. Madzivhandila, T.P., Nengovhela, N.B., Griffith, G.R. and Clark, R.E. (2007), “The South African Beef

Profit Partnerships Project: estimating the aggregate economic impacts to date”, paper to be presented at Living on the Margins – vulnerability, social exclusion and the state of the informal economy, Cape Town, South Africa, 26-28 March.

Nengovhela, N.B., Matjuda, L.E., Motiang, D.M., Madzivhandila, T.P., Banga, C., Masia, S., Clark,

R.A., and Timms, J. (2007), “Achieving sustained improvements in profitability in beef


390

enterprises and regions in South Africa”, Australian Journal of Experimental Agriculture (under revision).

Tapson, D. R. (1990), A socio-economic analysis of smallholder cattle producers in KwaZulu, PhD

Thesis, Vista University, Pretoria, Republic of South Africa. Timms, J., Clark, R., Espinosa, E., Gabunada, F., Madzivhandila, T.P., Maleza, Z., Matjuda, L.E.,

McCartney, A., Motiang, D.M., Nengovhela, N.B., Stewart, P. and Taveros, A.A. (2005), “Effective regional improvement and innovation networks”, International Conference on Engaging Communities, Brisbane, August.


391

Appendix 1: Table 2 Beef Profit Partnerships – Project Performance Measures

Target Outcome 1 - Rapid and measurable improvements in productivity, profit and growth

1 KPIs measured every 180 days 2 Results

1. Price - $ / kg 3

2. Throughput - kg / ha 4

3. Costs - $ / kg 5

4. Profit - $ / ha (per product, enterprise or business)

6

5. Relevant on-farm productivity KPIs (e.g. growth rate, reproduction %, death %)

7

6. Profit & productivity improvement in other enterprises

8

Target Outcome 2 - Supportive network of rewarding partnerships, contributing to accelerated industry growth


1. Number & type of BPP partners 11 Number of business managers, industry leaders/facilitators, specialists & researchers in the regional BPP network

2. Number & value of BPP focuses and activities

12 Number & type of BPP meetings. Scores of value (average & range out of 10). What liked & why; what not liked & why

3. Number & value of BPP communications, resources and specialist support

13 Number & score of value for kits, brochures, newsletters

4. Number and type of improvements & innovations shared

14 Number of improvements reported

5. Value of BPP groups/teams 15 % of meeting attendance. Feedback on BPP

6. Value of the BPP network 16 % of meeting attendance. Feedback on BPP

Target Outcome 3 – Partners equipped to achieve sustainable improvement and innovation


392


1. Number of partners who understand and value, the concepts and process of CI&I

19

2. Number and value of CI&I tools used 20

3. Number & description of improvements & innovations implemented

21 Reports on Action & Monitoring

4. Number of improvement opportunities assessed

22 Reports on Performance Analysis & Evaluation

5. Improved knowledge & skills of concepts, methods, tools & technologies

23 Reports on what individuals have learnt & changed that they did not know or do before

6. Number of concepts, methods, tools & technologies created, used &/or improved

24 Reports on new ways of assessing & managing the concepts like ‘throughput’, new products etc.


393

FACTORS INFLUENCING EAR INITIATION AND EAR EMERGENCE DEVELOPMENT OF PERENNIAL RYEGRASS CULTIVARS AT TWO DIFFERENT LATITUDES

G. Hurley*1, 2, M. O’Donovan1 and T. J. Gilliland2, 3

1Dairy Production Research Centre, Teagasc Moorepark, Fermoy, Co. Cork, Ireland

2Faculty of Science and Agriculture, Queens University Belfast, Northern Ireland 3Agri-food & Biosciences Institute, Crossnacreevy, Northern Ireland


Abstract The objective of this study was to define the relationship between ear initiation (EI) and ear emergence (EE) of perennial ryegrass cultivars at two latitudes. This investigation comprised three treatments (outdoor site at 54oN (CROSS’05), outdoor site at 50oN (MPK’06) and glasshouse at 54oN (CROSS’06)) on a common set of 40 cloned spaced plants of eight cultivars. EI date between MPK’06 and CROSS’05 was similar (+/- 1 day) while EI date at CROSS’06 was earlier (-16 days). Plants at MPK’06 had an EE date eight days earlier than plants at CROSS ’05 and CROSS’06. The interval between EI and EE was longer at CROSS’06 (+19 days) and CROSS’05 (+6 days) compared to MPK’06. Later heading cultivars had shorter period between EI and EE than earlier heading cultivars. A strong relationship between plant EE and EI was found at Moorepark (r2 =0.93) and Crossnacreevy (r2 = 0.94). Predicting cultivar EI date is a key indicator of the timing of sward quality deterioration. Keywords: perennial ryegrass, cultivars, ear initiation, ear emergence

Introduction In Irish dairy production systems increased emphasis is being placed on ensuring that the price paid for milk reflects the market returns that can be obtained from that milk in terms of processed products (Kennedy, 2005). It is essential therefore, that milk composition, in particular protein content is maximised. With the onset of lower product prices and the rising production costs, a low cost quality feed must obtain these objectives. During the mid-season period high herd performance can be achieved from an unsupplemented grass based diet (O'Donovan et al., 2004). During spring, swards change from vegetative to reproductive growth and by the inflorescent period (May/June), the sward is predominately made up of reproductive tillers resulting in sward quality deterioration, which has a negative impact on milk production and milk composition. One of the first morphological signs before the transition to reproductive growth, occurs at the shoot apex where bud primordia develop in the axils of the older leaf primordia to give the shoot apex a ‘double-ridge’ appearance (Jones & Lazenby, 1988) which occurs during spring. Laredo and Minson (1975) established that the leaves of Lolium perenne (perennial ryegrass) had a 20% higher VDMI (voluntary dry matter intake) than the stem fraction even though the DM digestibilities were only slightly higher (67.3 v. 64.8%, respectively). Significant differences for green leaf proportion exist between cultivars with leaf proportions as low as 63% in mid-season and late heading cultivars tending to have a higher leaf content (Gilliland et al., 2002). While it has been previously documented that reproductive initiation is day length dependant and ear emergence is largely energy dependent, the interaction between growing conditions, genotypes and locations would be expected to provide a better understanding of the relative importance of each factor


394

and could indicate potential strategies for controlling seed head development. The objective of this study was to investigate the effect of latitude and meteorological conditions on the timing of reproductive initiation and seed head development of eight perennial ryegrass cultivars. Materials and Methods Experimental Design and Management The experiment was undertaken at two different latitudes in Ireland over two consecutive years; Northern Ireland Plant Testing Station, Crossnacreevy, Belfast (latitude 54°32’N) during 2005 and 2006, and Moorepark Dairy Production Research Centre, Fermoy, Co. Cork (latitude 50°07’N) during 2006. In 2004, 40 plants of each cultivar (Table 1) were established at 0.75m spacing in an outdoor site at Crossnacreevy. These were vernalised over winter and examinations began in spring 2005 (CROSS’05). Table 1: Details of Lolium perenne cultivars assessed

25 Cultivar 26 Maturity level

27 Aberdart 28 Intermediate

29 Fennema 30 Intermediate

31 Corbet 32 Intermediate

33 Aberavon 34 Late

35 Foxtrot 36 Late

37 Mezquita 38 Late

39 Melle 40 Late

41 Twystar 42 Late

In the following autumn 2005, two clones of each plant were created by excising two tillers and transplanting them into multipot trays. One set of these plant clones were over wintered in a cool (frost free) glasshouse in multipots and then retained under controlled conditions in the glasshouse until all plants had initiated in spring (CROSS’06). These were then immediately transplanted to an outdoor site. The other set of tiller clones were transplanted at 0.75m spacing to an outdoor site in Moorepark in November 2005 and subsequently vernalised over winter. Examination of these plants began in parallel with the glasshouse experiment during the following spring of 2006 (MPK’06). In total therefore, this investigation comprised three treatments on a common set of 40 cloned plants for each cultivar. The treatments were ambient conditions at 54oN in 2005 (CROSS’05), ambient conditions at 50oN in 2006 (MPK’06) and glasshouse conditions at 54oN in 2006 (CROSS’06). Plant Measurements


395

Previous reproductive initiation data for three perennial ryegrass cultivars collected by Camlin (1977), were used to generate an ‘ear initiation’ (EI) versus ‘ear emergence’ (EE) regression coefficient. This was used as a guide to calculate an expected EI date for each cultivar in the present study, using their published ear emergence dates. Sampling for EI began in mid March prior to the expected EI date of the earliest cultivar. On alternate days one tiller was removed from each plant and examined under the microscope for reproductive budding as described by Sweet et al. (1991). Leaves were removed until the apex was visible under close examination. The presence of a double ridge on the apex indicated that the tiller had initiated or turned reproductive. Examinations continued until all plants had initiated. The mean EI date for each cultivar was then determined. Critical day length, which is the minimum length of daylight required to trigger the growth of the reproductive apex, was subsequently calculated for each plant. Photoperiodic data at the time of EI were obtained by interpolation of the corresponding latitude (Fig 1) from a table compiled by (Lam, 12 Nov. 2006) who calculated the year-round hours of daylight at five degree latitude intervals. Ear emergence (EE) date was recorded, as the date when three seed heads had visibly emerged on a spaced plant Cooper (1952). Plant energy requirements were calculated in terms mean daily temperature (>0oC) and photosynthetically active radiation (PAR) from EI to EE. Photosynthetically active radiation is the amount of useable light energy received by a plant and is dependent on light intensity and day length. PAR is a direct measurement of radiation in the wave band 400-700nm measured in MJ m2 -1 day-1. Statistical Analysis All statistical analyses were carried out using the statistical package SAS (SAS, 2002). Measurements were subjected to analysis of variance using the following model: Yij = µ +Si+Cj+Si X Cj+eij

where µ = mean; Si = site effect (i = 1-3); Cj = cultivar effect (j = 1-8); Si X Cj = interaction of site and cultivar; eijk = residual error term. A linear regression graph was drawn up for Crossnacreevy (54° 32’ N) to determine the relationship (R2) between EE and EI where the ten year average EE date was regressed on EI date (CROSS’05) for each cultivar. Results and Discussion Climatic Conditions Table 2 shows the total monthly rainfall (mm) and average mean daily temperature (°C) for the three sites during January to June 2005 and 2006. Overall, from January to June, the driest and warmest conditions were in 2006 at Moorepark and the wettest and the coolest conditions were at Crossnacreevy in 2005 and 2006, respectively. These records show, therefore, that the differences in these two key climatic parameters were of sufficient magnitude to induce differential timing of any climatically influenced physiological development.


396

Table 2: Total monthly rainfall (mm) and mean air temperatures (°C)

43 44 CROSS’05 45 CROSS’06 46 MPK’06

47 48 Total Rainfall

49 Mean air temp

50 Total Rainfall

51 Mean air temp

52 Total Rainfall

53 Mean air temp

54 Jan 55 96.3 56 6.1 57 32.0 58 5.3 59 55.2 60 5.0

61 Feb 62 47.8 63 4.8 64 44.2 65 4.9 66 26.3 67 5.4

68 Mar 69 72.8 70 7.1 71 153.0 72 4.8 73 108.1 74 6.1

75 Apr 76 84.0 77 7.5 78 59.9 79 7.5 80 29.3 81 8.5

82 May 83 88.6 84 9.7 85 106.5 86 10.4 87 115.2 88 10.8

89 June 90 40.8 91 14.1 92 40.6 93 14.3 94 13.7 95 15.0

96 97 98 99 100 101 102

103 Mean 104 78.1 105 8.2 106 72.7 107 7.8 108 58.0 109 8.5


397

In addition to climatic variation, differences in latitude between the two experimental sites provided an additional factor in this investigation. Fig 1 shows the magnitude of difference in day length between the two trial sites caused by their difference in latitude. The southern site (Moorepark) has longer day lengths in the winter period than the northern site (Crossnacreevy), but this is reversed in the summer period. Production of this day length correction graph for these two sites allowed the influences of differences in accumulated climatic conditions at the two sites to be separated from the influences of differences in day length, so facilitating an assessment of which parameter EI and EE were most influenced.

Figure 1: Hours of daylight (day length) at Crossnacreevy (54°N) ● and Moorepark (50°N) ■

6

8

10

12

14

16

18

01-

Jan

01-

Feb

01-

Mar

01-

Apr

01-

May

01-

Jun

01-

Jul

01-

Aug

01-

Sep

01-

Oct

01-

Nov

01-

Dec

Time of year

Dayle

ngth

(hours

)

Ear Initiation Table 3 shows the EI date and critical day length of each cultivar at each site. The average EI date between CROSS’05 and MPK’06 was similar (+/- 1 day) while the average EI date at CROSS’06 was much earlier; 16 and 15 days earlier than the EI date at CROSS’05 and MPK’06 respectively. There was a significant (P<0.001) site by cultivar interaction for EI date which was due to the large difference in the EI date (14 – 20 days) of each cultivar between the three sites. The average critical day length between CROSS’05 and MPK’06 was similar (+/- 0.4 hour) but was significantly less at CROSS’06 than CROSS’05 (-1.2 hours) and MPK’06 (-0.8 hours). Critical day length was significantly different (P<0.001) between sites however these differences were very small (≤ 1.2 hours) and so were in practical terms of little consequence. Data were sensitive to analysis due to the large number of replicate plants used. This can also be said of the significant (P<0.001) interaction between site and cultivar for EI critical day length, which was a result of a small yet significant range in day length hours (1.0 - 1.4) between cultivars at the three sites.


398

Table 3: Ear initiation dates and EI critical day lengths of eight test cultivars at three sites

EI date EI Critical day length (hours) Cultivar CROSS

’05 CROSS

’06 MPK ’06

Range (days)

CROSS ’05

CROSS ’06

MPK ’06

Range (hours)

AD 30 Mar 15 Mar 4 Apr 20 12.7 11.5 12.8 1.3 FN 1 Apr 17 Mar 3 Apr 17 12.8 11.7 12.8 1.1 CB 13 Apr 29 Mar 17 Apr 19 13.6 12.6 13.5 1.0 AV 14 Apr 31 Mar 12 Apr 14 13.7 12.6 13.3 1.1 FX 20 Apr 4 Apr 17 Apr 16 14.2 13.0 13.6 1.2 MZ 22 Apr 6 Apr 18 Apr 16 14.3 13.1 13.7 1.2 ML 30 Apr 11 Apr 27 Apr 19 14.9 13.5 14.2 1.4 TR 2 May 12 Apr 28 Apr 20 15.0 13.6 14.3 1.4 Mean 16-Apr 31-Mar 15-Apr 18 13.9 12.7 13.5 1.2 SED 1.16 0.076 S *** *** C *** *** S*C *** ***

AD = Aberdart; FN = Fennema; CB = Corbet; AV = Aberavon; FX = Foxtrot; MZ = Mezquita; ML = Melle; TR = Twystar; EI = Ear initiation; S = site; C = Cultivar; SED = Standard Error of Difference, *** = P<0.001 While it is has been previously documented that photoperiod determines the initiation date (Evans, 1964), the effect of temperature during the early growth stages of the stem apex may have had a minor influence on EI. Temperature was not measured in the glasshouse however it can be safely assumed that temperatures in the glasshouse were higher than the outdoor conditions at Crossnacreevy in 2006. This may have accelerated EI at CROSS’06 compared to EI at the other two sites. According to Evans (1964) temperatures within the range of -6oC to about 14oC are required for vernalization, therefore it is safe to say that glasshouse temperatures were within this range or above the lowest threshold. According to Keatinge et al. (1979) increased temperature may stimulate initiation at an earlier stage by controlling the growth rate of the stem apex. Higher glasshouse temperatures may have therefore accelerated the growth of the stem apex, resulting in an earlier EI date at CROSS’06. There was also a significant (P<0.001) difference in the EI date between cultivars at each site however prior to the commencement of the experiment cultivars were paired with similar heading dates thus pairs generally initiated together, which suggests a strong relationship between the EI and EE dates. All cultivars had a significantly lower critical day length requirement at CROSS’06 than the other two sites. Grass species vary in their day length requirements for inflorescence initiation. Lolium perenne is an obligate long-day plant (Evans, 1964), flowering only when the photoperiod exceeds a critical day length (Cooper, 1952; Cooper, 1960), while other species respond to short day length exposure. Work by Gangi (1983) has concluded that different cultivars of perennial ryegrass have different requirements for vegetative and reproductive growth and development. Ear Emergence As stated, day length is the most decisive component affecting inflorescence initiation. Once floral initiation is attained the subsequent rate of elongation of the fertile shoot and inflorescence development is controlled by spring temperatures, however for each cultivar there is a minimum time from ear initiation to ear emergence, under optimum temperature and photoperiod (Cooper, 1952).


399

Ear emergence was six days later at CROSS’05 than MPK’06 with CROSS’06 intermediate between these two sites (Table 4). Latitude had a large effect on the EE date as the rate of ear development is strongly influenced by temperature. Cooper (1952) demonstrated a close relation between spring temperatures and the date of ear emergence in perennial ryegrass. A transition to long days and higher temperatures is usually needed for heading and anthesis (Heide, 1994). There were large differences between latitude for temperature and light intensity, which will be discussed later. There was no significant difference in the EE date of six cultivars between MPK’06 and CROSS’06 however Fennema and Foxtrot were significantly (P<0.001) different. Plants at CROSS’06 were transplanted to an outdoor site after EI, and as the rate of ear development is largely influenced by spring temperature (Cooper, 1952), higher indoor spring temperatures prior to EI would have triggered an earlier EE date. Table 4: Ear emergence (EE) dates and number of days between EI and EE of eight test cultivars

EE date EI – EE (days) Cultivar CROSS

’05 CROSS

’06 MPK ’06

Range (days)

CROSS ’05

CROSS ’06

MPK ’06

Range (days)

AD 31 May 27 May 24 May 7 61 73 50 23 FN 29 May 28 May 23 May 6 58 72 50 22 CB 6 June 3 June 31 May 6 55 65 45 20 AV 9 June 5 June 2 June 7 56 67 50 17 FX 7 June 5 June 31 May 7 47 62 44 18 MZ 10 June 6 June 3 June 7 49 61 46 15 ML 16 June 11 June 9 June 7 47 61 43 18 TR 17 June 12 June 9 June 8 46 61 42 19 Mean 7 Jun 4 Jun 1 Jun 6.9 52 65 46 19 SED 1.48 1.40 S *** *** C *** *** S*C NS ***

AD = Aberdart; FN = Fennema; CB = Corbet; AV = Aberavon; FX = Foxtrot; MZ = Mezquita; ML = Melle; TR = Twystar; EE = Ear emergence; EI = Ear initiation; S = site; C = Cultivar; SED = Standard Error of Difference, *** = P<0.001; NS = Non Significant There was a significant (P<0.001) site by cultivar interaction for the number of days between EI and EE, which is due to the large difference in days between EI and EE (15 - 23) of each cultivar between the three sites (Table 4). CROSS’06 had significantly more (+13) days between EI and EE than CROSS’05, both of which are at the same latitude which is a result of an earlier EI date at CROSS’06 (-16 days). The southern site (MPK’06) had the shortest period between EI and EE (19 days less than CROSS’06) for the production of reproductive material. The number of days between EI and EE of several perennial ryegrass cultivars measured by Keatinge (1979) was between 50 to 70 days with later heading cultivars having a shorter period between EI and EE. A large difference existed between cultivars with later heading cultivars tending to have a shorter time interval between EI and EE than earlier heading cultivars. To increase sward quality late heading cultivars would therefore be more advantageous as the sward is in a reproductive growth mode for a shorter period than earlier heading cultivars. Mean daily temperature (°C) and mean daily PAR were significantly (P<0.001) different between site and cultivar (Table 5). Mean daily temperature (°C) and PAR increased with cultivar maturity, that is, the later heading cultivars received more plant energy within a shorter period than earlier heading cultivars.


400

There was no significant difference in mean daily temperature and PAR between CROSS’05 and CROSS’06 during EI to EE; however both sites were significantly different to MPK’06, which had significantly greater values. Work by Hennessy (2005) has shown that greater mean daily temperatures increase the leaf appearance rate and leaf extension rate resulting in a greater leaf content and higher DM yields. Schapendont et al., (1998) also agreed with this where an increase in temperature had a positive effect on yield and was obvious at more southerly latitudes. Johnson and Thornley (1983) found that an increase in the incident light flux density causes an increase in photosynthesis and hence greater yield. In conclusion, the southern site, MPK’06 had a greater mean daily temperature and greater mean daily radiant exposure (PAR), both of which have a positive impact on leaf content and DM yield. Later heading cultivars also had higher values than earlier heading cultivars during a shorter period (EI – EE). Table 5: Daily mean energy requirements between EI and EE of eight test cultivars

Daily mean temperature (°C) Daily mean PAR (MJ m2 -1 day-1)

Cultivar CROSS

'05 CROSS

'06 MPK

'06 Range

CROSS '05

CROSS '06

MPK '06

Range

AD 8.9 8.4 10.1 1.7 6.7 6.1 8.9 2.8 FN 8.9 8.9 10.1 1.2 6.8 6.2 8.7 2.5 CB 9.8 9.5 11.0 1.5 7.1 7.0 9.4 2.4 AV 10.0 9.6 10.9 1.3 7.2 7.2 9.4 2.2 FX 10.4 9.8 11.0 1.2 7.5 7.3 9.4 2.1 MZ 10.4 10.1 11.5 1.4 7.4 7.4 9.8 2.4 ML 11.2 10.9 12.1 1.2 7.4 7.7 11.0 3.6 TR 11.2 11.0 12.5 1.5 7.6 7.7 11.0 3.4 Mean 10.1 9.8 11.2 1.4 7.2 7.1 9.7 2.6 SED 0.154 0.121 S *** *** C *** *** S*C NS ***

AD = Aberdart; FN = Fennema; CB = Corbet; AV = Aberavon; FX = Foxtrot; MZ = Mezquita; ML = Melle; TR = Twystar; EI = Ear initiation; EE = Ear emergence; PAR = Phototsynthetically active radiation; S = site; C = Cultivar; SED = Standard Error of Difference, *** = P<0.001; P<0.05 Relationship between EI and EE date A linear regression line was drawn up to demonstrate the EI to EE relationship at Crossnacreevy using outdoor plant EI date data (CROSS’05) and the ten year average EE date (personal communication) for each cultivar. A strong relationship between plant EE and EI was found at Crossnacreevy (R2 = 0.96; Fig. 2). Given such a strong relationship, the regression line and/or predicted equation of the EE/EI correlation can be used to determine the EI date for any perennial ryegrass cultivar using a known EE date at the corresponding latitude.


401

Figure 2: Relationship between ear emergence (EE) and ear initiation (EI) of eight cultivars at Crossnacreevy (54° 32’ N) based on ten year average EE date

y = 0.7733x + 39.57

R2 = 0.9569

145

150

155

160

165

170

175

140 145 150 155 160 165 170

Ear Emergence date

Ea

r In

itia

tio

n d

ate

Conclusion Maintaining high quality herbage is imperative in increasing animal output and performance mid-season. A good understanding of sward physiological changes throughout the growing season and the factors affecting these changes is vital in minimizing herbage deterioration. Large variation for plant initiation exists between cultivars which is largely influenced by latitude and day length. Predicting cultivar EI date will determine when swards change from vegetative to reproductive growth. Day length is the principal factor influencing plant initiation and therefore cannot be controlled; however by combining this information on the timing of plant initiation date with appropriate spring grazing management may help reduce sward quality deterioration mid season. In Ireland the Irish recommended cultivar list evaluates cultivars on DM yield, yet no information on sward quality is available, a factor which has a substantial impact on animal intake and production. Sward quality reduction is associated with increasing sward maturity; therefore cultivar heading date is the only indicator of cultivar sward quality. The date of ear initiation is the date when growth changes from vegetative to reproductive and there is a rapid increase of poorly digestible reproductive material in the sward. Predicting the EI date of various cultivars is a key indicator of the timing of sward quality deterioration, and this information can be incorporated into a sward management plan so as to improve herbage quality mid season. Cultivar choice may also have a positive effect on mid-season sward quality as later heading cultivars tend to have a shorter period (EI-EE) for the production of reproductive material. References Camlin, M.S. (1977) Growth and management of early, mid-season and late varieties of perennial

ryegrass. In Advisers and lecturers conference Loughry college. Cooper, J.P. (1952) Studies on growth and development in Lolium perenne III. Influence of season and

latitude on ear emergence. Journal of Ecology 40, 352-379. Cooper, J.P. (1960) Short-day and low-temperature induction in Lolium. Annals of Botany 24, 232-246.


402

Evans, L.T. (1964) Reproduction. In Grasses & GrasslandsEdited by C. Barnard. Canberra: Macmillan, London.

Gangi, A.S., Chilcote, D.O. & Frakes, R.V. (1983) Growth, floral induction and reproductive

development in selected perennial ryegrass Lolium perenne L. cultivars. Journal of Applied Seed Production 1, 34-38.

Gilliland, T.J., Barrett, P.D., Mann, R.L., Agnew, R.E. & Fearon, A.M. (2002) Canopy morphology and

nutritional quality traits as potential grazing value indicators for Lolium perenne varieties. Journal of Agricultural Science 139, 257-273.

Heide, O.M. (1994) Control of flowering and reproduction in temperate grasses. New Phytologist 128,

347-362. Hennessy, D. (2005) Manipulation of grass supply to meet feed demand of beef cattle and dairy cows. In

Faculty of Science and Agriculture Belfast: Queen's University Belfast. Johnson, I.R. & Thornley, J.H.M. (1983) Vegetative crop growth model incorporating leaf area expansion

and senescence, and applied to grass. Plant; Cell and Environment 6, 721-729. Jones, M.B. & Lazenby, A. (1988) The Grass Crop: The physiological basis of production. Chapman and

Hall Ltd. Keatinge, J.D.H., Camlin, M.S. & Stewart, R.H. (1979) The influence of climatic factors on physiological

development in perennial ryegrass. Grass and Forage Science 34, 55-59. Kennedy, J. (2005) Maximising Milk Price - Producer (Part 2). In National Dairy Conference 'Winning in

Changing Times' Waterford, Ireland: Teagasc. Lam, S. (12 Nov. 2006) Day Length for Various Latitudes.

http://www.orchidculture.com/COD/daylength.html. Laredo, M.A. & Minson, D.J. (1975) The voluntary intake and digestibility by sheep of leaf and stem

fractions of Lolium perenne. Journal of British Grassland Society 30, 73-77. O'Donovan, M., Delaby, L. & Peyraud, J.L. (2004) Effect of time of initial grazing date and subsequent

stocking rate on pasture production and dairy cow performance. Animal research 53, 1-11. Schapendonk, A.H.C.M., Stol, W., Kraalingen, D.W.G.v. & Bouman, B.A.M. (1998) LINGRA, a

sink/source model to simulate grassland productivity in Europe. European Journal of Agronomy 9, 87-100.

Sweet, N., Wiltshire, J.J.J. & Baker, C.K. (1991) A new descriptive scale for early reproductive

development in Lolium perenne L. Grass and Forage Science 46, 201-206


403

DETERMINING LABOUR EFFICIENCY OF U.S. ROW CROP PRODUCTION

Gregg Ibendahl and John Anderson Department of Agricultural Economics, Mississippi State University,

P.O. Box 5187, Mississippi State, MS 39762, USA. Email: [email protected]

Abstract Technology is continually improving the technical efficiency of agriculture. Advances in seeds, chemicals, machinery, and other inputs are allowing farmers to produce more than ever before and with fewer inputs. In addition, the available supply of agricultural labour has been shrinking. One problem facing producers is determining what practices lead to labour savings and where is additional labour savings likely to occur. As quality labour becomes more expensive and difficult to obtain, producers will want to know how best to allocate their resources in order to obtain maximum labour efficiency. This paper uses seven years of farmer data from cotton and soybean production to develop a model that shows the factors determining the hours of labor required to produce each of the crops. The model is based on a regression analysis of 900 farmer observations from the Mississippi delta. In addition, the model shows how effective each factor is for reducing labour and whether the factor is more important for cotton or soybeans. Results show that farm size, field size, percent rented land, percent of farm planted to soybeans or cotton, percent custom expenses, percent GMO seed varieties, and row spacing can all be important factors determining labour hour requirements. Bigger farms have economies of scale for both cotton and soybean farms. However, the coefficient for cotton is twice as large meaning that cotton farms see a bigger gain in labour reduction by expanding than do soybean farms. The use of GMO had a similar effect as it both reduced labour and was more effective for cotton than soybeans. The major difference between cotton and soybean farms was in the degree of specialisation. For cotton farms, adding more cotton reduced labour while for soybeans, adding more increased labour. These results indicate that cotton farms are likely to continue to expand and also be more specialised. Farms growing soybeans are likely to grow a mix of crops but will continue to expand as well. These results should be useful to producers looking for ways to save labour and also to policy makers considering minimum wage laws and payment programs that might limit farm size. Keywords: labour, efficiency, cotton, soybeans, production Background As might be expected, data about farm labour use has existed for a long time. For example, the number of agricultural workers in a state can be found back to at least the year 1800 (http://eh.net/databases/agriculture/). At a very aggregate level, the number of labour hours required to produce a crop can be calculated by dividing the labour hours available by the number of acres of a crop produced. Some published estimates of crop production labour go back to the 1800’s as well. According to Welch and Miley (1950), 601 hours were required to pick a single bale of cotton. By 1925 this had dropped to 269 hours and by 1945, labour per bale was down to 182 hours. More recent data from the USDA shows how dramatically labour hours to produce a crop acre have decreased.


404

Figure 1: Labour hours to produce an acre of cotton and soybeans

Figure 1 is based on the USDA’s Agricultural Resource Management Survey (ARMS). This survey estimates the costs to produce an acre of crop. The labour cost is converted to hours by dividing by a labour rate per hour. As Figure 1 shows, the time requirement is under five hours per acre for both crops. Sampling by the USDA changed in 2003 explaining the big jump in cotton for that year. Labour required to produce a given crop acre has declined for two reasons. First, out of necessity as the available supply of labour has been shrinking since the 1900’s. Mundlak (2005) shows that throughout the 1800’s land, labour, and capital all increased with growth rates ranging from one to three percent. From 1900 to 1940, the agricultural labour supply started to decrease by 0.5% while the growth rates of land and capital slowed to a positive 0.5%. Since 1940, labour has decreased at a 2% rate and land has decreased at a 0.5% rate. The second reason labour requirements for crop production have decreased is technical innovation. These technological changes can be divided into mechanical and chemical. For the mechanical changes, the most important factor has been the development of the cotton harvester. As shown in Peterson and Kislev (1986), the percent of cotton mechanically harvested went from 6% in 1949 to 96% in 1969. Mechanical changes occurred at a faster rate from 1950 to 1970 than they did from 1930 to 1950 (Kislev and Peterson, 1982). Other mechanical changes include better and larger planters as well as new tools and techniques to get the cotton from the field to the gin. The development of the boll buggy and module builder was nearly as revolutionary as the mechanized cotton picker. Chemical changes are many and include new herbicides and insecticides and better defoliating tools. Related to chemical changes in both cotton and soybeans, is the development of genetically modified (GMO) seed. For cotton, GMO seed has resulted in Bt varieties that reduce or eliminate the need to spray insecticides. Bt cotton has been shown to have economic benefits to farmers as well (Pray and Ma, 2001). For both cotton and soybeans, GMO seed has also resulted in glyphosate-tolerant or “Roundup Ready” varieties. These latter GMO products allow producers to use glyphosate as a weed control herbicide. Glyphosate (Roundup) can result is labour savings as it may reduce trips over a field and is less time sensitive to application.


405

Peterson and Kislev (1986) examine which is the most important factor in the reduction of crop production labour. Has it been the reduction in supply of agricultural labour or has it been technological change reducing the need for labour? In the former case, technology was developed to fill the gap as labour left. Peterson and Kislev conclude that the pull effect of labour leaving the farm is four times greater than technology pushing labour away. Model and Data Every year in Mississippi a survey is conducted of producers to determine the cost of production for all the major crops grown in the state. This is a phone survey of around 150 producers. The survey consists of some basic questions about acres owned, acres rented, acres of each crop grown, types of seed planted, irrigation, and rental information. The bulk of the survey though is about a specific field and the operations applied to that field. Every trip over that field is recorded as an operation and these questions ask about what was done during that trip and the type and size of equipment used. In addition, if the field operation involved planting, spraying, or fertilization, then the material and rate is also recorded. The survey does not specifically ask for labour hours or any costs associated with the field operations. These are brought in from a database of machinery and material cost items. For example, planting a specific seed variety with specific size planter and tractor will generate a labour time and costs for the tractor, planter, and seed that is taken from the cost database. Producers probably do not have a good time and cost estimate at the field level but do have good information about what was done to the field on each trip across the field. Taking the field operation information and applying the typical associated costs for the machines and material used gives a better estimate of costs and time than if farmers tried to provide this information directly. The econometric model is based on 910 observations for cotton and 947 observations for soybeans. Different producers are surveyed for each crop and the mix of producers used each year is different. For the analysis in this paper, separate econometric models were developed for each crop. The independent variable is the number of hours per acre required to produce either an acre of cotton or an acre of soybeans. The labour hours include operator labour, hand labour, and irrigation labour. Operator labour hours are those associated with operating a piece of equipment (i.e., cotton picker, tractor, etc). Hand labour hours are extra time needed to perform the various field operations which are not time directly on the machine. For example, planting requires extra time to fill the planter boxes. This hand labour is usually a lower skilled, lower paid source. Irrigation labour hours are associated with operating irrigation equipment. In addition to these three sources of direct labour there is unallocated labour that is difficult to pin down to a specific field operation. This labour amount is estimated as a function of the other costs. However, this unallocated labour is not included in the analysis as it is just based on other data. As a result, the labour totals per acre will likely appear low compared to some other estimates of labour. This does not affect the analysis of what drives labour use in crop production.


406

Figure 2: PDF approximation of the labour hours per acre – cotton and soybeans

Hours_Cotton Hours_Soybeans

Figure 2 shows the probability density functions (PDF) for the labour hours required to produce an acre of cotton and an acre of soybeans. As the graph indicates, the mean hours for soybeans are less than the mean hours for cotton. In addition, the variance for cotton labour hours is greater. Specifically, the mean soybean and cotton labour hours per acre are respectively: 0.70 and 1.85 hours. The respective variances of soybean and cotton labour hours per acre are: 0.08 and 0.30. As mentioned above, this is only the directly measured labour hours. The unallocated labour could add 50% or more to the labour totals. Table 1: Descriptive statistics for some cotton variables

Table 2: Descriptive statistics for some soybean variables

Variable: Crop Acres

Soybean

Acres

Soyean

acres Irr Non-GMO

Direct

exp

Fixed

exp Yield

Min.: 12 12 - - 41 2 2

Max.: 13,621 8,926 6,700 5,200 203 95 70

Mean: 1,402 759 203 183 95 27 30

Median: 925 456 - - 91 23 30

Variance: 2,498,270 684,602 283,100 237,997 572 259 163

Std. Dev.: 1,581 827 532 488 24 16 13

Std. Err.: 51.86 27.15 17.46 16.01 0.78 0.53 0.42

Skewness: 2.88 2.78 4.84 4.87 0.81 1.29 0.06 Sum: 1,302,860 705,089 188,167 169,700 88,111 #### ####

Variable: Crop Acres

Cotton

Acres

Cotton

acres Irr Non-GMO

Direct

exp

Fixed

exp Yield

Min.: 12 12 - - 185 9 44

Max.: 13,700 10,200 8,000 3,500 642 162 1,700

Mean: 1,641 990 363 88 354 72 792

Median: 1,195 700 - - 349 68 800

Variance: 2,416,270 989,229 532,104 88,381 4,975 598 59,860

Std. Dev.: 1,554 995 729 297 71 24 245

Std. Err.: 51.84 33.17 24.33 9.92 2.35 0.82 8.16

Skewness: 2.50 2.99 3.96 6.34 0.33 0.43 0.03 N: 899 899 899 899 899 899 899


407

Tables 1 and 2 list some of the descriptive statistics about the farms used in the analysis. This table has the crop acres, soybean/cotton acres, soybean/cotton irrigated acres, acres of non-GMO seed, direct expenses per acre, fixed expenses per acre, and yield per acre. As can be seen, most of the farms are fairly large with a mean size of 1,500 acres. This is certainly bigger than the typical U.S. farm but is fairly typical of full-time farms in the Mississippi delta.

It should also be pointed out that the number of acres planted to non-GMO seed is fairly small. While the data range for this dataset captures some of the transition to GMO seeds, the transition was well underway by the first year of this data. By the last year of the dataset nearly all the cotton seed and a majority of the soybean seed is GMO based. Results The following equation is the final form of the econometric model

Hr = β0 + β1 ⋅ Cropland + β2 ⋅ Fld_size + β3 ⋅ P_rent + β4 ⋅ P_AC_cotton

+ β5 ⋅ P_custom_exp + β6 ⋅ P_GMO_ac + β7 ⋅ Is_irr + β8 ⋅ Is_skiprow

+ β9 ⋅ Is_2004 + β10 ⋅ Is_2003 + β11 ⋅ Is_2002 + β12 ⋅ Is_2001+

+ β13 ⋅ Is_2004 + β14 ⋅ Is_1999 + β15 ⋅ ave _ yld

The variables in this equation are defined as follows: Cropland – The number of acres of cropland farmed. Fld_size – The size of the field used in the survey questionnaire. P_rent – The percentage of crop acres that are rented. P_AC_cotton – The percentage of crop acres planted to cotton. The soybean analysis has a similar variable. P_custom_exp – The dollar amount of custom operation expenses divided by the total expenses. This percentage is for either the cotton or soybean acres only. P_GMO_ac – The percentage of either cotton or soybean acres that were planted to GMO seed varieties. GMO seeds for cotton could either be Roundup Ready varieties, Bt varieties, or stacked genetics varieties (i.e., varieties containing both Roundup Ready and Bt genes). Is_irr – A dummy variable to represent if the crop acreage was irrigated. Is_skiprow – A dummy variable to represent skip row cotton. For soybeans, this is a continuous variable for row spacing. Is_2004 – A dummy variable signifying crops grown in 2004. There are dummy variables for years 1999 through 2003 as well. This makes 2005 the reference year. Ave_yld – the weighted average irrigated and dry land yield. Other variables were investigated in the econometric analysis but these were the only ones to have any statistical significance.


408

Table 3: Regression results for cotton model

Beta S.E. t-test Prob(t)

Intercept 2.29367 0.11114 20.63788 0.00000

Cropland -0.00006 0.00001 -5.64468 0.00000

Fld_size -0.00029 0.00021 -1.38194 0.16733

P_rent -0.05361 0.04291 -1.24921 0.21192

P_AC_cotton -0.15617 0.05895 -2.64932 0.00821

P_custom_exp -4.35934 0.28115 -15.50552 0.00000

P_GMO_ac -0.12577 0.06504 -1.93390 0.05344

is_irr 0.14858 0.03394 4.37751 0.00001

is_skiprow -0.19064 0.05767 -3.30601 0.00098

is_2004 0.01085 0.05752 0.18866 0.85040

is_2003 0.12314 0.05740 2.14530 0.03220

is_2002 -0.28730 0.05465 -5.25703 0.00000

is_2001 -0.33732 0.05676 -5.94290 0.00000

is_2000 -0.29116 0.05569 -5.22828 0.00000

is_1999 0.08241 0.05551 1.48458 0.13801

ave_yld 0.00042 0.00007 5.79017 0.00000

Table 4: Regression results for soybean model

Beta S.E. t-test Prob(t)

Intercept 0.65262 0.05096 12.80663 0.00000

Cropland -0.00003 0.00001 -4.68119 0.00000

Fld_size -0.00028 0.00010 -2.68882 0.00730

Row_space 0.00493 0.00079 6.21605 0.00000

P_rent -0.06675 0.02139 -3.12114 0.00186

P_AC_soybeans 0.07891 0.02891 2.72995 0.00645

P_custom_exp -1.15437 0.08371 -13.78994 0.00000

P_GMO_ac -0.07985 0.02170 -3.67931 0.00025

is_irr 0.13807 0.01956 7.05792 0.00000

is_2004 0.08622 0.02803 3.07631 0.00216

is_2003 0.07312 0.02924 2.50041 0.01258

is_2002 0.00382 0.02724 0.14034 0.88843

is_2001 0.07410 0.02895 2.55940 0.01064

is_2000 0.09208 0.03004 3.06515 0.00224

is_1999 0.15426 0.02965 5.20213 0.00000

ave_yld 0.00252 0.00075 3.34844 0.00085

Tables 3 and 4 present the regression results for the cotton and soybean analysis respectively. Every variable in the soybean analysis is significant while in the cotton analysis; field size and percent rent were not significant. However, these two variables were left in so that both crop models would match. The cotton model has an adjusted R-squared of 0.38 while the soybean model has an adjusted R-squared of 0.34 All the variables have the correct sign. As farm size gets bigger, the required labour gets smaller. The coefficient is very small but this only represents the fraction of a labour hour per acre reduction caused by farming one additional acre. However, cotton labour was reduced at twice the rate of soybean labour. The presence of a negative sign on the coefficient, even in the presence of some fairly large farms, would indicate that there are still some returns to scale occurring.


409

Field size is only significant for the soybean analysis but its negative sign shows that there are advantages to big fields. This might perhaps represent having fields closer together rather than farther apart. Saving time on the road is probably what is occurring here. The percentage of land rented tends to reduce labour requirements for soybean production. However, this variable is not significant for cotton. Row spacing is only a variable for soybean production as most of the cotton was only in two row widths and these were very close together. Surprisingly, wider rows led to more labour. Perhaps weed control was a bigger issue with wider rows. The percentage of custom operations affected both cotton and soybeans the most. This is totally expected as hiring custom work takes directly away from what the producer must provide. The increased use of GMO seed reduced labour as well. Again this is an expected result. However, the variable is barely significant for cotton. Given that the dataset does not cover the entire GMO transition, some of the significance may have been lost. The magnitude of the GMO seed coefficient for cotton is nearly twice that of soybeans. Because cotton has two types of GMO characteristics, this is expected. The one variable where the cotton and soybean analysis differ is for the percentage of farm acres planted to the crop in question. For cotton farms, increasing the percentage of cotton acreage decreased labour. For soybean farms, increasing the percentage of soybean acres increased labour. Figure 3: Scatter plot of labour hours for cotton for various farm sizes

Actual Hours Predicted Hours


410

Figure 4: Scatter plot of labour hour for soybeans for various field sizes

Actual Hours Predicted Hours

Figures 3 and 4 show how well the model predicts labour hours. These two figures match the hours by using the variable in the analysis that appears to show the biggest deficiency. For cotton farms, the econometric model does not do all that well at predicting labour hours for the smaller farms. For the soybean farms, the econometric model visually appears to do the poorest predicting with the smaller field sizes Conclusions and Discussions As shown here, labour hours to produce an acre of cotton and soybeans have been reduced considerably over the last century. Based on USDA data, cotton and soybean labor hours per acre are 4.5 and 2.5 hours, respectively. Based on data collected from Mississippi producers, the direct labor hours per acre for cotton and soybeans are 1.85 and 0.70 hours, respectively. Labour reduction is still occurring thanks to continuing improvements in mechanisation and improved seed and chemical technologies. However, the dramatic improvements in labour reduction are likely past as even in the highest cotton estimate by the USDA, cotton labour was already below five hours per acre. Thus, production changes that shaved hundreds of hours per acre are just not possible. Still, with labour becoming more scarce and expensive, any improvements in labor requirements should be welcome by producers. Returns to scale are still occurring and are possible even with some of the bigger than average farms in the Mississippi delta. Therefore, farm size is likely to increase even more. Given that cotton farms showed a rate of improvement twice that of soybean farms, cotton farms are more likely to expand than soybean farms. The specialisation variable in the results showed conflicting results. Cotton farms were able to reduce labour by growing more of the farm in cotton while the reverse was true of soybean farms. Thus, cotton farms are likely to become even more specialized in cotton while soybeans farms are likely to have a variety of crops grown. GMO seed varieties have certainly help reduce labour. However, this analysis probably missed some of the labor savings as the GMO transition was well underway before this variable was measured in the dataset.


411

Soybeans have the lower labor requirements than cotton and probably always will. However, cotton also has the most room for improvement and shows the most promise for size expansion and improvements in mechanization and seed/chemical technologies. This paper should be useful to U.S. policy makers because of potential changes to minimum wage laws as well as potential changes to farm policy relating to farm size. Cotton farms are some of the biggest farms and could be directly affected by a U.S. cap on government payments. If cotton farms are still showing benefits to increasing farm size then a payment cap could make U.S. agriculture less efficient. Finally, this paper should help farmers and farm managers as they make expansion decisions and allocate scare labor. References Carl, Pray et al. 2001, 'Impact of Bt Cotton in China', World Development, vol. 29, pp. 813-825. Kislev, Y. & Peterson, W. 1982, 'Prices, Technology, and Farm Size', The Journal of Political Economy,

vol. 90, pp. 578-595. Mundlak, Y. 2005, 'Economic Growth: Lessons from Two Centuries of American Agriculture', Journal of

Economic Literature, vol. 43, pp. 989-1024. Peterson, W. & Kislev, Y. 1986, 'The cotton harvester in retrospect: labor displacement or replacement',

Journal of economic history, vol. 46, pp. 199-216. United States Department of Agriculture (USDA) (2007) [Online]. Available at <http://www.usda.gov>

[Accessed 1 March 2007] Welch, F.J. & Miley, D.G. 1950, 'Cotton Labor Requirements', Journal of Farm Economics, vol. 32, no.

4, pp. 752-758.


412

THE USE OF RELEVANT COST ANALYSIS TO ASSESS PRODUCTION VIABILITY FOLLOWING THE DECOUPLING OF SUPPORT PAYMENTS IN ENGLAND1

Dr James V. H. Jones

Principal Lecturer and Head of Farm Management, School of Business, Royal Agricultural College, Cirencester, Gloucestershire,

GL7 6JS, United Kingdom E mail: [email protected]

Abstract Relevant cost analysis is a well recognised technique in management accounting used to decide whether production is viable in the short run. The removal of production-related subsidies following the reform of the CAP has left many farm enterprises with net margins that show a substantial loss. Relevant cost analysis is used to determine whether nevertheless it remains worthwhile continuing to produce in the short run. Relevant margins were calculated from industry costings for combinable crop, beef and sheep enterprises for 2004/5. These show that costed on this basis it is only the beef enterprises that look financially unviable. The paper argues that relevant cost analysis not only provides a very useful aid to farm level decision-making but also represents a very useful tool for guiding policy makers and industry analysts on the vulnerability of production and the potential for resultant structural changes. Keywords: relevant cost; viability; decoupling; subsidies; production What is Relevant Cost Analysis? Relevant cost analysis is a well established method in management accounting used to assess the viability of production decisions. Although it has not been widely used in agriculture as such, it uses virtually the same principles as those applied in partial budgeting. It is therefore a concept that will be both easy to appreciate and to apply for farm management economists. Partial budgeting isolates costs and revenues that are relevant to a change which is taking place. This generally involves a factor or product substitution. Relevant cost analysis typically looks at the decision to produce for a particular enterprise in isolation. Both techniques seek to isolate costs and revenues that are relevant to the production decision being examined ignoring those that are deemed to be irrelevant. Drury (2004, p.37) defines relevant costs and revenues as ‘those future costs and revenues that will be changed by a decision, whereas irrelevant costs and revenues are those that will be not be affected by a decision’. This generally restricts consideration to cash costs because these are the ones that alter. Drury concludes that ‘future cash flows’ tend to be ‘the relevant financial inputs for decision-making purposes’. However it is worth pointing out that if there is an opportunity cost this could put a cash value on the release of a resource priced on a non-cash notional basis. Irrelevant costs according to Jay (2004) include: Fixed overheads – because these will be incurred regardless of the decision. Notional costs - as these costs are only a book exercise and do not represent a real cash flow. Past or sunk costs – because these have already been incurred and they cannot be affected by a future decision. Book values – i.e. the price paid for stock in the past.

1 Some of the analysis in this paper has already been presented, with the permission of the 16th International Farm Management Congress organisers as an unrefereed paper (Jones, 2007) to the RICS Rural Research ROOTS 2007 Conference, 17th April 2007.


413

The Case for Considering the use of Relevant Margins as Opposed to Gross or Net Margins In a sense it may sound as though the definition of relevant costs is very similar to the definition of variable costs which are the basis of gross margins. However the identification of variable costs tends to be exclusive to those that can be easily allocated to enterprises. Costs which do vary, in the sense that they are affected by whether production takes place or not, which are nevertheless hard to allocate, tend to be treated as ‘fixed’. A good example of this is machinery costs. The elements that are directly affected by machinery use i.e. fuel and repairs, tend to be treated as part of fixed cost along with other elements, such as depreciation, insurance and road fund licences, which can to an extent be regarded as a fixed regardless of the use of the machine. In a relevant cost analysis all items that vary must be identified separately. This may involve a certain amount of estimation or apportionment. But no attempt needs to be made to allocate costs that are irrelevant to the decision to produce. There has been a long-running debate amongst farm management economists about whether net margins should not be more widely used as an alternative to gross margins in order to recognise the importance of the ‘fixed cost’ implications of enterprise choice (Warren, 1998). The case for this was put by Giles (1986 and 1987) and to an extent refuted by Kerr (1988). Nevertheless the net margin approach is now being more widely adopted and the arguments for doing so are set out in case for using net margins in the Farm Business Survey (FBS) as put by Wilson and Seabrook (2005). One of the objections to net margins as a measure of enterprise profitability is that there is no defined limit on which costs can and should be allocated. As a result the size of a net margin may be as much a reflection of the ability to allocate costs as to the amount of them. It is also a moot point as to whether the costs really can be allocated fairly to the enterprise. Relevant costs analysis provides a clear definition as to what costs should be allocated and why. This of course means that some of the costs that are deemed irrelevant are nevertheless real costs that have to paid somehow or other. But taking irrelevant costs out of the picture is helpful in identifying what could be a perfectly sound rationale for continuing to produce even when the net margin may indicate a substantial loss. A good illustration of why a distinction between relevant and irrelevant costs is a useful one is the cost of land and buildings. In the short run these costs will have to be met whether production takes place or not because of contracts with landlords, mortgage providers etc. Opportunities to buy or rent land are scarce and farmers will not give land up just because prices are unfavourable for production for the time being. To these practical considerations is added the need to retain land in order to claim the Single Farm Payment (SFP), which is area based. If the SFP is more than the costs of rent or finance associated with the land then getting rid of the land would be counter-productive. Yet net margins generally allocate the costs of land as though they were attributable to the enterprise and nothing else. Relevant cost analysis ignores these costs as fixed and would include consideration of any revenues that were lost as a result of getting rid of the land as ‘relevant’. It is quite likely that relevant costs will be different in the long run from the short run. This is because in the long run the farmer can change costs that it would be inadvisable or impractical to change in the short run i.e. staff can be made redundant, machinery sold, land can be sold or let etc. However long run relevant margins are likely to be much more difficult or tenuous to determine than short run. Why is it Particularly Pertinent to Consider the use of Relevant Margins at the Present Time? The importance of assessing the short run relevant margin is that if this not positive then the farmer has no financial justification for continuing to produce other than challenging the assumptions on which the calculations are based or taking into consideration potential tax advantages. There are of course plenty of non-financial criteria that could be applied to continuing to produce with a negative relevant margin but


414

they may not be sustainable in the long run. The relevant margin indicates the degree safety with which it can be assumed that production should continue. It has been argued by Jones (2005) that the reform of the Common Agricultural Policy (CAP) and sharp increases in some factor costs (notably those related to the price of crude oil) has now made it quite pertinent to assess whether production should continue whereas previously this might have been rather academic. The new context for production decision-making and the case for using relevant cost accounting have been explored more fully by Jack and Jones (2006). They argue that there are other factors which have created a case of need including, with the rapid increase in agri-environment funding, situations in which production decisions have to factor in the possibility of payments for cutting down production in order to benefit the environment and wildlife. There is real concern currently that some sectors of agriculture cannot produce adequate returns to support their continued existence in the longer term. The latest set of figures for beef and sheep costings from the English Beef and Lamb Executive (EBLEX, 2006) are the first to show performance without the benefit of subsidy. They show losses at the net margin level (particularly after adding in notional rent and unpaid labour) across all enterprises at all levels of performance. These losses are particularly large for beef enterprises. Gross margins however at average levels of performance were all positive. This situation is not unique to the beef and sheep sectors. Recent special studies of FBS cereal and oilseed rape net margins (Newcastle University, 2006 and Lang and Allin, 2006) also indicate that net margins would be negative without the Arable Area Payment (AAP) although gross margins would be positive. This creates a confusing picture with gross margins indicating that all these enterprises could be profitable and net margins indicating that none of them are. Relevant margins provide the opportunity to determine whether, between these two extremes, there is a good rationale to continue with production under these conditions and how generous the margin is. Combinable Crops The FBS studies provided a break down of the full net margin costs of oilseed rape (Newcastle University, 2006) and cereals (Lang and Allin, 2006) for the harvest year 2004. This was the last year before the Arable Area Payment (AAP) was replaced by the decoupled Single Farm Payment (SFP). It is therefore possible to see what the margins looked like with and without the subsidy. In order to establish an estimate of the relevant margin under these conditions certain assumptions had to be made about which costs were to be treated as relevant costs. These costs are identified in Table 2 under the cost headings used in the special studies (Newcastle University, 2006 and Lang and Allin, 2006). The variable costs are all obviously relevant costs because they are totally linked to production. The ‘overheads’ were not identified in any detail. They were thus treated as common costs and therefore most likely to be fixed and unaffected by a marginal change in production. The ‘fixed costs’ were largely treated as fixed and therefore irrelevant to any short run marginal change in the cropping unless it was felt from their description that they might have elements that would vary directly with production.


415

Table 2: The identification of relevant costs for growing combinable crops

Relevant costs Costs treated as 'irrelevant'

Variable costs Fixed costs

Seed Farmer's own labourFertiliser Unpaid labourSprays Paid labour - ordinary hoursCasual labour Machinery costs - depreciation, insurance & licencesContract Grain storage plant - depreciationFuel for drying Grain storage buildings repairs and depreciationMarketing costs RentMiscellaneous Drainage charges

Fixed costs Overheads

Paid labour - overtime LabourMachinery costs - fuel and repairs MachineryGrain storage plant - repairs Buildings

General The overtime element of paid labour was considered to be a relevant cost. This was estimated at 27.6% of the total on the basis of typical annual labour cost assumptions in Nix (2006, p.133). Machinery repairs, fuel and oil were estimated at 50% of total machinery costs as an approximation based on actual proportion that these costs represented on FBS cereal farms in 2004/5 (DEFRA, 2006, Table 1). The relevant costs also included a cost of interest on working capital (not mentioned in Table 2). The annual average working capital was estimated at 66.7% of total relevant costs for autumn sown crops and 50% for spring sown. The interest rate used was 7.5%. The crop gross margins and net margins from the FBS special studies and the relevant margins derived from them are shown in Figure 1. These are shown both with and without the AAP. It can be seen that with the AAP the crops all make a positive net margin, with the exception of spring oilseed rape which makes a small loss. There is therefore little need to look at the relevant margin to determine whether production is viable. However without the AAP the situation is much less clear. The crops all still make a positive gross margin but the net margins all show a loss of in excess of £100/hectare. Under these circumstances it is important to assess whether the relevant margin is sufficient to make it worthwhile continuing in production. The results show that it is viable, albeit with a very small margin for spring oilseed rape.


416

Figure 1: Gross margins, net margins and relevant margins of combinable crops in 2004/5 with and without the Arable Area Payment

-£300

-£200

-£100

£0

£100

£200

£300

£400

£500

£600

Gross marginincluding AAP

Net margin includingAAP

Relevant marginincluding AAP

Gross marginexcluding AAP

Net margin excludingAAP

Relevant marginexcluding AAP

£ pe

r ha

Winter OSR Spring OSR

Winter wheat Spring barley

Winter barley winter oats

Source: Based on data contained in Newcastle University (2006) Table 3.1 and Lang and Allin (2006) Tables 4.1 and 4.2 Beef and Sheep The data used to analyse production viability of beef and sheep enterprises was sourced from the EBLEX costings for 2004/5 (EBLEX, 2005). Although more recent costings are available (EBLEX, 2006) this was the last year before beef and sheep headage payments were abandoned in favour of the decoupled SFP. It therefore makes it possible to show what the returns were both with and without the subsidy. The EBLEX definition of net margin includes different costs from those allowed for in the FBS combinable crop special studies. They do not include any imputed rent or family labour costs, whereas Newcastle University (2006) and Lang and Allin (2006) do. However the arable studies do not include finance costs whereas EBLEX (2005) have included them. This illustrates the problems caused by a lack of a common basis for net margins referred to earlier. Relevant costs had to be extracted from the EBLEX figures by estimation. ‘Labour costs’ included regular wages, contract labour and casual labour. It was decided to take just the estimate of overtime cost on the whole at 27.6% (based on Nix, 2005, p.133) on the assumption that regular labour was likely to be the largest component of these costs. ‘Power and machinery’ consisted of machinery repairs, fuel, electricity, general contract, machinery hire, tax and insurance. It was assumed that 92.3% were relevant costs based on average tractor running cost assumptions in Nix (2005, p.165). ‘Administration costs’ comprised insurance, office costs and miscellaneous sundries. An estimate of 25% was placed on the likely proportion of relevant costs. ‘Property charges’ included water, council tax and farm and property repairs. The most important relevant cost within this would be metered water and it was assumed that this would comprise 20%. ‘Land resource costs’ were made up of actual rents and this is not a relevant cost so they were not included. ‘Machinery and fixtures’ comprised machinery depreciation, fixtures depreciation, machinery and equipment leasing. None of these are relevant costs. Finally the ‘finance costs’ were removed and replaced with a figure based on the relevant cost of interest (at 7.5%) on average working capital. The latter was calculated by taking costs incurred throughout the production cycle and adding half the costs incurred during the cycle adjusting the figures to an annualised basis.


417

The gross margins, net margins and relevant margins for various beef enterprises are shown in Figure 2. It can be seen that extensive beef finishing was not making a positive net margin even with the subsidy payments. However the other beef systems were making a positive net margin and costed on a relevant cost basis all the enterprises make a margin of over £100/head. However without the subsidies the picture is very different. Intensive beef finishing makes a tiny gross margin and all the systems show a loss on a net margin basis. It should be noted that the costs make no allowance for unpaid family labour (unlike the FBS special studies). If this was included the results would be even more dramatic. The relevant margins show a positive return, albeit a small one, for suckler cows and a negative return for both extensive and intensive beef finishing. This indicates a position which is not financially sustainable in the long term. Figure 2: Gross margins, net margins and relevant margins for beef enterprises in 2004/5 with and without subsidies

-£200

-£100

£0

£100

£200

£300

£400

Gross marginincluding subsidies

Net margin includingsubsidies

Relevant marginincluding subsidies

Gross marginexcluding subsidies

Net margin excludingsubsidies

Relevant marginexcluding subsidies

£/he

ad

Lowland sucklers

LFA sucklers

Intensive beef finishing

Extensive beef finishing

Data source: EBLEX (2005) pages 4 - 7

The gross margins, net margins and relevant margins for breeding sheep and store lamb finishing are shown in Figure 3. This shows that positive net margins were being made with the benefit of the Sheep Annual Premium (SAP). Removal of that subsidy has no effect on the store lamb finishing (ceteris paribus). However breeding sheep net margins become negative. This raises the question as to whether breeding sheep are financially sustainable without the subsidy? The answer however, as contained in the relevant margins, is positive with margins on average performance of over £10/ewe. This indicates that at least in the short run the producer should remain in production and contrasts with the beef relevant margins which did not provide this comfort.


418

Figure 3: Gross margins, net margins and relevant margins for sheep enterprises in 2004/5 with and without subsidy

-£20

-£10

£0

£10

£20

£30

£40

£50

Gross marginincluding subsidies

Net margin includingsubsidies

Relevant marginincluding subsidies

Gross marginexcluding subsidies

Net margin excludingsubsidies

Relevant marginexcluding subsidies

£/he

ad

Lowland breeding sheep

LFA breeding sheep

Store lamb finishing

Data source: EBLEX (2005) pages 8 - 10

Sensitivity of the Relevant Margin to Changes in Output It is useful to test the sensitivity of the relevant margins to changes in assumptions on price or physical performance. This is to show the degree of vulnerability or comfort enjoyed by the high and low performers and what an alteration in prices might do. This is particularly important as studies predict some price increases in response to subsidy decoupling (Moss et. al., 2002). It is the sensitivity to the change in the primary product that is felt to be the most significant and interesting. Therefore income from secondary products (wool with the sheep and straw with the cereal enterprises) have been assumed to be fixed. The percentage change required in the output of the primary product for the relevant margin (without subsidies) to break-even point is shown Figure 4.. This shows that most of the arable crops would require a substantial reduction in yield and/or price for the relevant margin to be at the break-even level. The sheep enterprises are closer to the break-even but the change required in lamb output (i.e. due to price, lamb numbers or weight) would still be quite substantial. But the beef suckers are close to beak-even and very exposed to any reduction in output (particularly in the lowlands) and the fattening systems require an increase in output (from the price or weight of the fat animal or a reduction in the store price) just to reach the break-even point.


419

Figure 4: Percentage change in primary output required for the relevant margin to break-even

-0.4%

-12.0%

7.8%5.6%

-20.4%

-25.1%

-9.1%

-28.2%

-11.6%

-33.9%-36.8%

-38.5%

-43.9%

-50%

-40%

-30%

-20%

-10%

0%

10%

20%

Lowlandsucklers

LFAsucklers

Intensivebeef

finishing

Extensivebeef

finishing

Lowlandbreeding

sheep

LFAbreeding

sheep

Storelamb

finishing

WinterOSR

SpringOSR

Winterwheat

Springbarley

Winterbarley

Winteroats

% c

hang

e in

out

put

Source: Based on data contained in EBLEX (2005), Newcastle University (2006) Table 3.1 and Lang and Allin (2006) Tables 4.1 and 4.2 Beef results for 2005/6 (EBLEX, 2006) show that intensive finishers have increased their margins as a result of a drop in the price they pay for stores. But in general terms the picture does not look any more favourable than in 2004/5. Output has dropped for suckler cows although there has been a small improvement in the gross margin for lowland sucklers. EBLEX (2006, p.7 – 8) now show a net margin for lowland sucklers after having deducted notional costs for unpaid family labour and land at a loss of £351/cow and a loss of £425/cow in the Less Favoured Area (LFA). Conclusions Relevant margins provide a very useful way of assessing whether enterprises that show a loss at the net margin level are nevertheless viable in the short term. The use of actual figures and assumptions on what are likely to be the relevant cost elements illustrate the use and the value of the technique. It shows that without production subsidies on average all combinable crop, beef and sheep enterprises in 2004/5 looked unprofitable at the net margin level. But by removing irrelevant costs it showed that only the beef enterprises looked to be financially unviable in the short run. This kind of analysis is clearly very valuable to farmers trying to make tactical production decisions and strategic plans for the future. It also could be a valuable tool in informing policy makers and industry analysts about the vulnerability of individual enterprises and production systems. The relevant margin will vary on each farm according to circumstances. Those that have the most flexible and responsive cost structures will have the lowest margins and those whose costs are mostly fixed and/or notional will have the highest for a given level of output. This does not indicate that their businesses are more profitable overall; in fact they may well be less profitable. However it does show that in terms of production decision-making they are likely to continue in production at lower prices and output levels than those whose costs are less fixed. This helps to explain the resilience of the self-sufficient family farmer remaining in production when on a total cost basis the enterprise looks to be financially unviable.


420

References Drury, C. (2004) Management and cost accounting, 6th Edition, Thomson, London DEFRA (2006) Farm accounts in England 2004/5: Results of the Farm Business Survey, May 2006,

Department for Environment Food and Rural Affairs EBLEX (2005) Business pointers for livestock enterprises: Cattle and sheep costings 2005/6, English

Beef and Lamb Executive, Farmers’ Weekly supplement, 25/11/2005 EBLEX (2006) Business pointers for livestock enterprises: Cattle and sheep costings, English Beef and

Lamb Executive, Farmers’ Weekly supplement, 24/11/2006 Giles, A.K. (1986) Net Margins and all that – an Essay in Management Accounting in Agriculture, Study

No.9, Farm Management Unit, Reading University Giles, A.K. (1987) ‘Net margins’, Farm Management, 6 (6), pp271-279 Jack, L. and Jones, J.V.H. (2006) ‘Facing up to new realities: The case for using relevant cost and target

cost approaches in agriculture’, Paper presented to the 80th Annual Conference of the Agricultural Economics Society held on 30th - 31st March 2006 at the INA P-G, Paris

Jay, B. (2004) ‘Relevant costs for decision-making’, 24th October 2004

http://www.accountancy.com.pk/articles_students.asp?id=142 accessed on 7/7/07 Jones, J.V.H. (2005) To produce or not to produce – is that the question?, Journal of Farm Management,

12 (5) pp 235-249 Jones, J.V.H. (2007) ‘The use of relevant cost analysis in identifying thresholds of production viability

post CAP reform’, paper delivered to the RICS Rural Research, ROOTS 2007 conference, 17th April 2007, RICS, Great George Street, London

Kerr, H.W.T. (1988) ‘Management accounting and the marginal concept’, Farm Management, 6 (9),

pp.349-351 Lang, B. and Allin, R. (2006) Special Study into the Economics of Cereal Production 2004, DEFRA

Special Studies in Agricultural Economics, Report No. 64, Rural Business Unit, Department of Land Economy, University of Cambridge

Moss, J., McErlean, S.,Kostov, P., Patton, M. , Westoff, P. and Binfield, J. (2002) Analysis of the Impact

of Decoupling on Agriculture in the UK, Agriculture and Food Science Centre, Queen’s University, Belfast

Newcastle University (2006) The Economics of Oilseed Rape in England, 2004, DEFRA Special Studies

in Agricultural Economics, Report No. 65, Farm Business Survey, School of Agriculture, University of Newcastle upon Tyne

Nix, J.S. (2005) Farm Management Pocketbook, 36th Edition, Wye Campus, Imperial College London Warren, M.F. (1998) ‘Banishing ‘fixed’ and ‘variable’ costs: Time to bring farm accounting into the real

world?’, Farm Management, 10 (2) , pp75-79


421

Wilson, P. and Seabrook, M. (2005) ‘Developments in the Farm Business Survey and Implications for Agricultural and Rural Economics Research’, Paper presented to the Agricultural Economics Society 79th Annual Conference, 4 - 6 April 2005, University of Nottingham.


422

FARM INCOME ON FULL AND PART-TIME FARMS – 2005

Anne Kinsella and Brian Moran Farm Management Department, Teagasc, RERC, Athenry Ireland

Email: [email protected] Abstract The Teagasc National Farm Survey (NFS) is undertaken annually to determine the financial situation on Irish farms. The principal measure of the income arising from farming activities is Family Farm Income (FFI) per farm. In addition to analysing farm income by system and size of farm, NFS data can also be analysed for Full-time and Part-time farms to determine the variation in income that occurs. In this paper the variation in FFI on Part-time and Full-time farms is analysed by system of farming. In the NFS Full-time farms are defined as those which require at least 0.75 Standard Labour Units to operate, as calculated on a Standard Man Day (SMD) basis. Farms are therefore divided into Full-time and Part-time on the basis of the estimated labour required to operate their farms as distinct from labour available, which is often in excess of that required. The total number of farms represented nationally is 111,115. Full-time farms represent the larger more commercial sector of farming and in 2005 accounted for 38% (or 42,300) of all farms represented by the NFS. Fifty five percent of Full-time farms were in the two dairying systems with 34% in the drystock systems with the remaining 11% in the Tillage systems. Of the 62% of farms which were Part-time, 88% were in the drystock systems. The average FFI on all Part-time farms in 2005 was €11,372, ranging from €16,933 on Dairy farms to €9,995 on Sheep farms. On 58% of Part-time farms either the farm holder or spouse had an off-farm job and on 94% of farms, there was another source of income – either from an off-farm job, pension or social assistance. Full-time farms are two and a half times the size (ha) of Part-time farms and represent the more commercially viable sector of farming. Keywords: family farm income, full-time farms, part-time farms, standard labour units Introduction The Teagasc National Farm Survey (NFS) is undertaken annually with its primary objective to determine the financial situation on Irish farms by measuring the level of gross output, costs, income, investment and indebtedness across the spectrum of farming systems and sizes,. The NFS is responsible for provision of data on Irish farms to the EU Commission. The principal measure of the income arising from farming activities is Family Farm Income (FFI) per farm, representing the financial reward to the family labour, management and capital investment in the farm business. This is calculated by deducting all the farm costs (direct and overhead) from the value of farm gross output. It does not include income from non-farming sources and thus may not be equated to household income. For 2005 year there are 1177 farms included in the analysis, representing 111,115 farms nationally. Figure 1 shows average Family Farm Income (FFI) per farm in current and real terms over the period 1995 to 2005. The data shows farm income in 2005 was 58% above that for 1995 in current terms and when inflation (CPI) is taken into account that FFI has increased from €14,236 in 1995 to €16,651 in 2005, an increase of 17% in real terms. The trend in FFI in current and real terms is shown in Fig 1. The main reason for the increase shown from 2004 to 2005 is the once-off carryover of arrears of direct payments from 2004. For all farms in 2005 FFI increased from €15,557 per farm in 2004 to €22,460 in 2005 – an increase of 44.4%. This phenomenal increase in 2005 farm incomes was due mainly to the change in EU policy from a coupled to a decoupled system, implemented in Ireland in the 2005 year. In 2005 Irish farmers received an average once-off payment of €5,266 per farm due to the carry-over of arrears from the 2004 coupled


423

direct payments. Thirty four percent of the increase in average farm income was due to this exceptional payment viz. Family Farm Income (FFI) would have increased by 10.4 per cent from 2004 to 2005 were it not for this direct payment arrears. Across the different farm systems, size groups and regionally there is variation in FFI per farm. Figure 1: Family Farm Income per Farm (€) 1995- 2005

Full-time and Part-time Farms In addition to analysing farm income by system and size of farm, NFS data can also be analysed for Full-time and Part-time farms to determine the variation in income that occurs. This paper focuses on the variation in FFI on Part-time and Full-time farms. Thirty eight per cent of the total population (or 42,300 farms) are classified as Full-time farms. During 2006 a supplementary survey on the NFS sample was also undertaken to determine farmers’ perception as to whether they were full-time or part-time farms. Results of this are also further analysed in this paper. In the NFS Full-time farms are defined as those which require at least 0.75 Standard Labour Units to operate, as calculated on a Standard Man Day (SMD) basis. , whilst Part-time farms require less than 0.75 labour units Farms are therefore divided into Full-time and Part-time on the basis of the estimated labour required to operate their farms as distinct from labour available, which is often in excess of that required. Standard labour requirements are measured in SMD for each farm enterprise and these are used to estimate overall labour required to operate the farm. A SMD is based on eight hours of work supplied by a person over eighteen years of age. The number of SMD required per hectare for different crops and per head for various categories of livestock is used to calculate the total number of SMD required to operate the farm. The presence of an off-farm job is not taken into consideration in the definition.

2,500

5,000

7,500

10,000

12,500

15,000

17,500

20,000

22,500

25,0001

995

19

96

19

97

19

98

19

99

20

00

20

01

20

02

20

03

20

04

20

05

Year

€

Current Real


424

Figure 2: Population of Full-time and Part-time Farms - 2005

Figure 2 shows the population of Full-time and Part-time farms, by system, in the NFS for 2005. The majority of part-time farms are in the three drystock systems, namely Cattle Rearing, Cattle Other and Sheep whilst the majority of full-time are in the specialist Dairying system.

0

5000

10000

15000

20000

25000

30000

35000

Dairying Dairying

& Other

Cattle

Rearing

Cattle

Other

Sheep Tillage

popula

tion

Full-time Farms Part-time Farms


425

Data in Table 1 show FFI, farm size measured in Utilised Agricultural Area (UAA) Direct Payments (DP) and off-farm employment for Full-time and Part-time farms by system of farming in 2005. The total number of farms represented nationally is 111,115 and population estimates are shown for each category of farms. Full-time farms represent the larger more commercial sector of farming and in 2005 accounted for 38% (or 42,300) of all farms represented by the NFS. Although Full-time farms account for only 38% of the population they contribute over two and a half times of gross output compared to the Part-time farms. Fifty five percent of Full-time farms were in the two dairying systems with 34% in the drystock systems and the remaining 11% in the Tillage systems. As highlighted also in Table 1, the average FFI on Full-time farms was €40,483, ranging from €49,102 on Dairying and Other System to €28,529 on the Sheep System. Direct payments contribution to FFI ranged from 48% on Dairy farms to 122% on Cattle Other farms and was 78% for all farms. Overall either the farm holder or spouse had an off-farm income on 49% of all Full-time farms. The incidence of off-farm jobs was higher on Full-time dairy farms than on Part-time dairy farms at 51%. However, on 18% of Full-time farms the farmer had an off-farm job, while on 38% of farms the spouse had an off-farm job. Of the 62% of farms which were Part-time, 88% were in the drystock systems. The average FFI on all Part-time farms in 2005 was €11,372, ranging from €16,933 on Dairy farms to €9,995 on Sheep farms. Figure 3 highlights the difference in FFI between Full-time and Part-time farms and compares this to FFI on “All” farms. Direct payments as a percentage of FFI ranged from 46% on Dairy farms to 137% on

Table 1: Full-time and Part-time Farms by System of Farming - 2005 System Dairying Dairying

Other Cattle

Rearing Cattle Other

Sheep Tillage All Systems

Full-time Farms

% of Population

15 6 3 5 5 4 38

UAA (ha) 45.8 67.3 53.7 64.9 69.0 88.3 59.6 Family Farm Income (FFI) €

41,357 49,102 29,240 42,132 28,529 44,709 40,483

Direct Payments €

19,712 35,632 35,565 49,372 34,472 43,503 31,724

DPs as a % of FFI

48 73 122 117 120 97 78

% of Farms Off-farm Jobs

51 41 50 45 57 47 49

Part-time Farms

% of Population

1 3 22 22 11 3 62

UAA (ha) 20.0 20.3 24.2 21.9 25.4 27.7 23.5 Family Farm Income (FFI) €

16,933 8,807 10,812 12,481 9,995 13,209 11,372

Direct Payments €

7,771 12,083 14,280 16,444 12,957 14,013 14,567

DPs as a % of FFI

46 137 132 132 130 106 128

% of Farms Off-farm Jobs

42 39 62 60 55 57 58

Source: Teagasc, National Farm Survey 2005


426

Dairying and Other farms. Direct payments (DP) include all subsidies paid to farmers and in 2005 include arrears from the 2004 accounting year plus the Single Farm Payment (SFP), Rural Environment Protection Scheme (REPS) and Disadvantaged Area Compensatory Allowance Scheme (DACAS). On Drystock farms DP account for more than 100% of FFI when market based output is not sufficient to cover total costs. On 58% of Part-time farms either the farm holder or spouse had an off- farm job and on 94% of farms, there was another source of income – either an off-farm job, pension or social assistance. Farmers on part-time farms were older (56 years) than those on Full-time farms (51 years) and 62% were married compared to 75% on Full-time farms. Figure 3: Family Farm Income (FFI) on Full-time and Part-time Farms – 2005 Data in Fig. 4 details FFI, direct payments and farm size for the full-time farms by farming system. In 2005 the normal pattern of income distribution between full-time farm systems changed due mainly to the unusual direct payment situation. The Dairying Other System and Tillage system had the highest FFI per farm at €49,102 and €44,709 respectively, followed by Cattle Other at €42,132. In previous years the Specialist Dairy system and Tillage always had the highest farm incomes when confined to full-time farms. Details of FFI, direct payments and farm size for Part-time farms are detailed graphically in Fig. 5

0

10000

20000

30000

40000

50000

60000

Dairying Dairying &

Other

Cattle Rearing Cattle Other Mainly Sheep Mainly Tillage All

0

10000

20000

30000

40000

50000

60000

Full-time Farms Part-time Farms All Farms


427

Figure 4: FFI, Direct Farm Payments/Subsidies for Full Time Farms by Farming System - 2005

Approximately 68,800 of farms were part-time with an average FFI of €11,372, ranging from €16,934 on the specialist Dairy Systems to €9,995 on the Sheep system. The average cash income on part-time farms was €13,583 in 2005 compared to €9,015 in 2004. Average direct payments and subsidies were €14,567 in 2005 i.e. 128% of FFI, reflecting the general situation on drystock farms (88% of part-time farmers in drystock systems) where output from the market place is insufficient to cover total production costs. Farmers in these drystock systems, and indeed in the Mainly Tillage system, will need to re-plan their enterprises and enterprise mix to take account of the decoupled policy introduced in 2005. On 58% of these Part-time farms either the farmer or spouse had off farm employment and on 94% of farms there was another source of income – either from off farm job, pension or social assistance. The farmers on part-time farms were older (56 years) than those on full-time farms (51 years) and 62% were married compared to 75% on full-time farms.

Summer Survey 2006

In Summer of 2006 an additional questionnaire was undertaken on the NFS sample. One of the questions included on this survey was to determine farmers’ perceptions as to whether they regarded their farm as a Full or Part-time farm based on the standard labour unit requirements. This survey represented a population of over 95,000 farmers, with over 44.,000 of those farmers regarding their farm as Full-time. Results by system of farming are detailed in Table 2. Overall 46% of farmers considered their farm to be Full-time, with the highest incidence occurring in the Dairying system, at 85%.

0

10000

20000

30000

40000

50000

60000

€/Farm

30

40

50

60

70

80

90

100

Ha

FFI 41357 49102 29240 42132 28529 44709 40483

Direct Payments 19712 35632 35565 49372 34472 43503 31724

Farm Size (Ha) 45.8 67.3 53.7 64.9 69.0 88.3 59.6

DairyingDairying/

Other

Cattle

Rearing

Cattle

Other

Mainly

Sheep

Mainly

TillageAll


428

Table 2: Do you consider this farm a full-time Farm?

System Dairying Dairying Other

Cattle Rearing

Cattle Other


% Yes 85 67 30 33 43 45 46 No 15 33 70 67 57 55 54

Source: Summer Survey – NFS 2006

When results of the survey are compared with farms which are actually calculated as Full-time in the NFS, it is interesting to note that only 35% of NFS farms are calculated as Full-time as compared to 46% that regarded their farms as Full-time. The highest incidence of Full-time farms occurs in the Dairying system with 93% of farms in this system calculated as Full-time, whilst only 85% of farmers in this category regarded their farm as Full-time. The drystock farms, namely Cattle Reraing, Cattle Other and Sheep farms, all overestimated their labour requirements (SMDs) with Tillage farms scoring exactly the same as that which was calculated for their farms. Table 3: Calculated on NFS as Full-time Farms

System Dairying Dairying Other

Cattle Rearing

Cattle Other


% 93 55 10 17 29 45 35

Source: Summer Survey – NFS 2006 Conclusions Full-time farms are two and a half times the size (UAA) of Part-time farms and represent the more commercially viable sector of farming. Over half of all Full-time farms are in Dairying Systems even though Dairy systems only account for 25% of all farms nationally. The income was higher on the Full-time farms, with the highest FFI in the Dairying Other system. Direct payments as a percentage of FFI were higher on the Part-time farms. Full-time farms were demographically more viable than Part-time farms, with a higher percentage of households having at least one member below 45 years of age. More farms consider themselves as Full-time farms than those defined as Full-time in the NFS. Overall 35% of NFS farms are calculated as Full-time as compared to 46% that regarded their farms as Full-time farms. The highest incidence of Full-time farms occurs in the Dairying system with 93% of farms in this system calculated as Full-time, while only 85% of farmers in this category regarded their farm as Full-time. Drystock farms overestimated their labour requirements (SMDs) with more farms in these systems considering themselves Full-time than what were actually calculated as Full-time farms. References Connolly, L., Kinsella, A., Quinlan, G. & Moran, B., National Farm Survey – 2005, Teagasc. Connolly, L., Kinsella, A., Moran, B. & Cushion, M., Summer Survey 2006, National Farm Survey,

Teagasc.


429

STRATEGIES OF POLISH FARMERS – AN ATTEMPT OF CLASSIFICATION

Edward Majewski and Piotr Sulewski Warsaw University of Life Sciences

Nowoursynowska 166, 02-787 Warsaw, Poland Email: [email protected]

Abstract In the paper results of the study on strategies of Polish farmers are presented. For a sample of 100 commercial family farms researched in the year 1996 the survey has been repeated in the year 2006. Ex-post analysis of farmers’ decisions in the key strategic areas revealed a distinct pattern of behavior which allowed to identify six basic types of strategies. Growth oriented strategies, in particular strategy characterized by expansive increase of farm area and scale of animal production resulted in a noticeably high increase of agricultural income. There is also a relatively large group of farmers implementing reduction strategies. This leads to the conclusion that ongoing structural changes in Polish agriculture may be accelerated in the near future. Keywords: Polish farmers, strategy Introduction Agriculture was one of the first sectors of the Polish economy to experience the effects of economic transformation, which began in 1989. Market liberalization resulted in the increase of prices of inputs, largely due to the removal of subsidies, as well as the imports of agricultural and food products competing successfully with domestic production. This led to a decrease of real agricultural income [Woś 1998, Józwiak 1998]. Since mid 1990s agricultural policy in Poland has undergone further changes due to preparations for accession to the EU, and since 2004 Polish agriculture has been included in the Common Agricultural Policy. The dynamic changes in Polish agriculture at the end of the previous century and the beginning of the 21st century brought about many threats, but also created opportunities for farmers. The vast majority of farmers from commercial farms took advantage of these opportunities, adjusting their farms to the new policy and market environment. Some, however, were not able to face the new challenges, which resulted in a deterioration of their financial situation. In the study, on which this paper is based, an attempt has been made to find out what strategic choices farmers had made before the Polish accession to the EU and how effective farmers’ strategies were. In general, it is rather unique that farmers, especially from small scale family farms, develop any formal, strategic plans. However, it does not mean that they do not apply any kind of long-term strategies. These are visible in ex-post research, when a course of actions and decisions made are analyzed. This survey is based on the assumption that every farmer implements a strategy to some extent, even without being aware of the theory of strategic planning and management. According to H. Mintzberg [1989, 1992], strategies are not only a result of formal planning process, but also a reflection of evolutionary character of organization’s functioning. As Mintzberg states, all decisions and actions within an organization create a certain pattern, which might be interpreted as a kind of strategy.


430

Methodology The panel survey was conducted in 2006 on a sample of 100 commercial farms from various regions of Poland. The same farms had been researched in a different project carried out in 199647. In both cases structured interviews were conducted to obtain the data. In 1995, the average size of the farm was 20,6 ha of agricultural land. At that time, the average farm size in Poland was about 6 ha. This is an indication that all the researched farms belonged at that time to the group of family commercial farms. Detailed interviews enabled a comprehensive analysis of the farms’ organization and performance, and provided information on changes that occurred in the surveyed farms between 1995 and 2005. The changes observed within the farms became the basis for identification of strategy types implemented by farmers. In order to identify the strategies, areas of farm management which can be considered strategic have been distinguished. Furthermore, methods of multi-dimensional analysis (cluster analysis and principal components analysis – PCA) have been applied. Identification of strategies realised by farmers Following some suggestions from the general theory of strategic decisions [e.g. Niedzielski and Fedejko 1995, Olson 2001] areas of strategic decisions (of strategic importance) possible to examine in farm businesses have been determined: Farm area – differences in the farm size were expressed by the percentage change of the agricultural land between year 1995 and 2005 and also by percentage share of land lease; Investment activities – described by a factor calculated as follows: value of the investment realised between 1995-2005 (in fixed prices) divided by the value of fixed assets in 1995; Type of investment financing – described by the indebtedness factor of the assets due to loans taken and by the factor determining the share of EU subsidies in financing farm investments Animal production scale and importance – described by the percentage change of livestock units number on a farm between 1995-2005 and the share of revenue from animal production in the overall farm revenue Crop production intensity – expressed by the percentage change of material costs calculated per hectare of crop production Production specialization degree – expressed by the factor describing the share of revenue from main activity in the overall farm revenue Income source diversification – expressed by the factor describing the share of non-agricultural income in the overall farm income Strength of farmer – customer links – expressed as the share of production delivered under long-term agreements in the overall sales value.

47 Majewski E., 2002. Economic and organizational conditions for dissemination of Integrated Farming System in Poland. Wyd. SGGW, pp. 190.


431

Thus established indicators were used for identification of farmers’ strategies. The first stage of the analysis showed farms, which in 2006, relied mainly on non-agricultural sources of income. Such farms accounted for 10% of all the farms and their rate of income from non-agricultural sources was more than 50%. They were excluded from statistical analysis because the majority of variables used for describing the strategy characterize agricultural functions of a farm, whilst in this group non-farming functions prevail over the agricultural ones. The strategy for this group of farms was named “reduction strategy with income diversification”. Since variables describing diversification of income sources in other farms did not noticeably differ, they were excluded from further analysis. In order to divide the rest of the farms into groups of similar strategic functioning areas, a two-phase cluster analysis was conducted. Classification into groups was based on the k-mean method, which due to optimisation enabled the formation of k-clusters. They are characterised by a maximum variability between each other and a minimum variability within each one [Internet Handbook Statsoft]. The use of this method means, however, that the researcher has to make an arbitrary decision on the number of clusters [Aldenderfer and Blashfield 1984]. In order to avoid this, initial analysis using hierarchical agglomeration method (Ward method, Euclid distances) was conducted in phase one, as suggested by Guidici [2003, after Harańczuk 2005]. This led to accepting 5 clusters as an optimum solution. In the second phase, the iterative k-mean method was used to group the farms into 5 clusters. Calculations have been done using Cluster analysis module of Statistica. The final classification of farms into groups and the characteristics of each cluster according to the strategic areas is shown in Table 1. Table 1: Average values of variables describing farm strategic function areas

Cluster Variables

No. No of farms

Far

m

size

ch

ange

[%

]

Lan

d le

ase

[%]

Inve

stm

ents

Inde

bted

ness

Use

of

EU

fund

s

for

in

vest

men

t [%

]

Cha

nge

in s

cale

of

an

imal

pr

oduc

tion

[%

]

Impo

rtan

ce

of

anim

al

prod

ucti

on [

%]

Cha

nge

in c

rop

prod

ucti

on

inte

nsit

y [

%]

Spe

cial

izat

ion

[%]

Cus

tom

er

link

s st

eren

gth[

%]

1 6 -40 3 0 1 0 -4 66 -66 65 0 2 19 -2 4 24 3 0 43 70 -13 75 33

3 29 108 35 126 20

7 222 81 3 82 58

4 29 65 16 17 13

0 94 71 -12 72 22

5 7 82 12 16 7 32 -100 0 71 100 33 6* 10 -14 3 16 1 0 -7 39 -7 83 5

*group with dominance of non-agricultural income, excluded from cluster analysis Source: Own research. The quality of this division was tested using multi-dimensional variance analysis, which confirmed the statistical significance of differences between the mean values of the characteristics of every group of farms. Thus determined clusters may be referred to as strategic groups which, according to the assumptions made, consist of farms implementing similar strategies. The method of classifying the strategies applied in strategic management is usually a two-dimensional matrix. It facilitates determination of the strategy or construction of the strategic groups’ map. In case of a multi-dimensional model, as in this situation,


432

where it takes into account different aspects of farm’s operations, the application of the classical approach proved to be impossible because of interpretation problems. That is why it was decided to simplify the model by applying factor analysis. The aim was to reduce the number of variables characterising a farm and achieve a better recognition of the data structure [Rummmel 2002, Trucker and MacCallum 1997]. This was supposed to facilitate drawing conclusions about the farmers’ strategies in different clusters. In this analysis the principal components analysis (PCA) was applied. Before conducting factor analysis, correlation between the primary variables was checked using the correlation matrix. Because the average correlation coefficients were bigger than 0.3, factor analysis was necessary [Sokołowski, Sagan 2005]. The number of main components was determined using the Keiser criterion (only factors whose value was above 1 were kept/left). Three separated, mutually independent, principal components jointly accounted for almost 60% variances of the original variables analyzed [Table 2]. Table 2: Values of selected principal components

Principal component no.

Eigenvalue % total variance

Cumulated Eigenvalue

Cumulated % total variance

1 2,9 29 2,9 29 2 1,9 19 4,8 48 3 1,1 11 5,9 59

Source: Own research. In order to determine relation between the selected principal components and original variables, factor loadings were analyzed [Table 3]. Table 3: Factor loadings

Selected principal components Initial variable Factor 1 Factor 2 Factor 3

AREA 0,82 0,05 0,06 LEASE 0,61 0,01 -0,27 INVEST 0,77 -0,15 0,27 DEBTS 0,64 0,31 -0,24 FUNDS 0,13 0,14 0,80 SCALE 0,73 -0,23 0,32 LIVEST 0,38 -0,68 -0,04 INTENS 0,08 0,81 -0,04 SPEC 0,22 0,14 0,45 SALE 0,01 0,73 0,28

Source: Own research. They might be interpreted as correlation coefficients between the processed variables and the primary variables [Internet Handbook Statsoft]. Factor loadings were subject to Varimax rotation in order to achieve more clarity. The character of variables correlated with different factors enables different interpretations of the situation [Rószkiewicz 2003]. The analysis showed that factor one (principal component) is loaded mainly by such variables as: area changes, rate of land lease, investment coefficient, indebtedness coefficient and changes in animal production scale. The variables grouped in this factor pointed to the development features. Factor two was related mainly to the production type and with the variable determining the strength of farmer – customer links. High value of this indicator suggests great importance of crop production and lower of animal production. The third factor, of the smallest weight in explaining the overall variability, was loaded mainly by the variable describing the rate


433

of EU subsidies and, to a smaller extent, by the one describing production specialization degree. The results of the factor analysis suggest that the variables grouped in the first factor were of key importance for explaining the overall variability. It means that factor one had the greatest influence on the strategies adopted by farmers. In the further part of the analysis, factor values were calculated in order to estimate the importance of each factor for the identified clusters of farms. Values of individual factors were calculated for every farm and then the mean values were derived for each cluster [Figure 1]. The received values of factors do not have the straightforward representation in figures but can only be interpreted in terms of their overall meaning. [Wójcik 2006], which points at the difference in their position in the 3D space. The analysis shows that clusters number one and two have reached the lowest values for factor one suggesting growth. At the same time, the farms from group one are characterised by a low weight of factor number two, which means decreasing intensity and rather low importance of crop production. The highest value of factor one (growth) was reached by farms from cluster 3, while farms from cluster 4 were in the middle of the range. Cluster number 5 reached the highest values of factor two, which suggests high importance of the intensity of crop production and low of animal production, and of factor three, which means large share of EU subsidies in investment financing and high degree of specialization (similar to cluster 3). Figure 1: Factor values for the determined groups of farms.

Source: Own research. The analysis, however, did not give the answer to the question of what strategies are implemented by farmers in each group. Still, it was helpful in interpreting the volumes of the variables characterising particular clusters [table 1]. The results of the factor analysis and the original parameters describing particular clusters helped determine the strategy types applied by farmers in the identified groups: Cluster one: Simple reduction strategy. The farms in which this strategy was implemented, were characterized by low values of all three factors selected in the factor analysis. Compared to the initial situation (original parameters), the average farm area in this group was reduced by 40 % and intensity decreased. There were no investments made between 1995 and 2005. Moreover, farms in this group are of the lowest degree of specialization and have no formal links with customers. Cluster two: Continuation strategy. These farms showed slightly higher values of all three factors. Comparison with the initial situation shows that the area of agricultural land has not changed. Similarly, the scale of animal breeding and production intensity have remained the same. The scale of investment


434

did not guarantee the replacement of fixed assets. The links between farmers and customers were of medium strength. Cluster 3: These farms showed the highest level of the growth related factor. The area of agricultural land has more than doubled and the number of animals has increased by 220%. Investments between 1995 and 2005 exceeded the value of the fixed assets in 1995. The farms in this group have the highest share of land lease and the highest indebtedness rate. Animal production is of the greatest importance within this group, same as links with customers. The strategy for this group is called Expansive growth strategy. Cluster 4: The value of factor one in this case was lower than in cluster three but higher than in all other clusters. Values of factor two and three were at the average level. The analysis of the initial variables showed that these farms are similar to farms in cluster three in terms of direction of changes but the scale of changes is much lower. Therefore, the strategy for this group is called Restricted growth strategy. Cluster 5: The farms in this cluster had similar value of factor one as farms in cluster two and the highest values of the two other factors. More detailed analysis showed that the relatively low value of factor one is caused by the abandonment of animal production, although all the other factors clearly indicate growth. In this case, animal production proves to be of no importance and crop production is intensified. That is why the strategy for these farms is called Growth with the focus on crop production strategy. The last group selected at the beginning of the analysis consists of farms relying mainly on non-agricultural sources of income. The strategy for these farms is called Reduction strategy with income diversification. For these farms agricultural production was constantly reduced (reduction in area, animal production scale and crop production intensity). The observed value of the investment ratio is the result of fixed asset purchases made for other than agricultural activities (e.g. for services). Figure 3 presents obtained strategy types.

Figure 3: Strategies accomplished in the researched group of farms.

Strategy and financial performance of farms

The characteristics of farms belonging to different strategic groups clearly indicates that all growth oriented strategies have resulted in significant improvement of economic performance of farms (table 4). Farms with reduction strategies noticeably decreased the production scale in the period 1995-2005 and

Sub-type:

STRATEGY

REDUCTION CONTINUATION (19 farms)

GROWTH General type:

- simple reduction (6 farms) - reduction with income diversification (10 farms)

- expansive growth (29 farms) - restricted growth (29 farms) -growth with focus on crop production (7 farms)


435

the level of net farm income in the year 2005 was close to zero. In diversified farms off-farm incomes were too low to compensate sufficiently the reduction of farm income and, consequently, the disposable household income decreased in relation to the year 1995. Table 4: Farms’ characteristics according to the strategy applied.

Strategy Simple reduction

Reduction – diversifi-cation

Continuation Expansive growth

Restricted growth

Growth with focus on crop production

Farmer’s age 55 46 48 41 42 45 1995 19,3 16,3 18,6 25,2 20,5 19,5 Area of

agricultural land [ha]

2005 11,6 13,1 18,3 52,4 32,3 35,5

1995 8,7 6,3 12,8 16,9 12,7 5,2 Number of animals [livestock unit] 1995

2005 7,4 5,8 18,3 54,4 24,7 0,0

1995 -3 13,3 7,2 30,5 10,1 -3,8 Per farm 2005 0,8 -1,6 28,1 95,0 44,0 24,3

Net farm income [thousand PLN ]

Per fully employed person 2005

0,4 -1,4 15,5 40,2 20,5 14,7

1995 29,9 38,1 47,8 75,5 48,3 42,4 Per farm 2005 19,8 29,8 54,4 138,1 72,0 42,9

Disposable income [thousand PLN

Per family member 2005

7,0 6,9 13,1 28,4 15,1 6,8

Farms which apply continuation strategy achieve relatively good results due to high productivity of land and efficiency of investments. Slightly lower incomes in 2005 were achieved by farms which pursued the strategy of “focusing on crop production”. It must be stressed however that production on these farms was dominated by cereals, which are less profitable than other crop production activities. Despite the limited growth of scale of production, farms belonging to this group registered the biggest growth of agricultural income in the period of 1995-2005. The highest income level was achieved by farms which applied „expansive growth” strategy which led to a significant increase in both the land area as well as the scale of animal production. It should be stressed that in 1995 farms of this group were characterized by only slightly bigger production potential while the level of agricultural income was considerably higher. The above leads to a conclusion, that thanks to the increased economic power, these farms were capable of expansive growth, which resulted in increased distance to the other farms, in terms of both, production scale and profitability. Conclusions The ex-post research showed that long-term activity of farmers revealed a distinct pattern of behaviour which may be called a strategy. The proposed method proved to be an effective way to conduct identification and classification of strategies implemented by farmers. In the examined group of commercial farms there were three basic strategy types applied – reduction, continuation and growth -


436

which comprised sub-strategies showing differences in directions of changes in farm organization. Growth strategies were applied by over 60% of the examined farms. Strategy types adopted by farms strongly correspond with the nature of structural changes occurring in Polish agriculture. Over the recent years considerable portion of agricultural land has been transferred to larger farms; similarly, the process of animal concentration has advanced. Appearance of a large number of farms applying growth strategies and improving their financial standing allows us to anticipate continuation of the present trend of farm structural changes in the years to come. The rate of changes may be considerably accelerated if we assume that the number of farmers retiring from farming (e.g. because of old age) will increase, which would release limitations on the land resources. References Aldenderfer M. S., Blashfield R.K.,1984 : Cluster analysis. SAGE Publications Józwiak W.,1998: Procesy dostosowawcze gospodarstw rolnych do zmiennej sytuacji rynkowej.

(Processes of adjustment to changing market situation in agricultural farms). In Polish agriculture in system transformation period (1989-1997)”, IERiGŻ, Warszawa

Mintzberg H. i Waters J.A.,1989: Of Strategies, Deliberate and Emergent. Readings in Strategic

Management, edit. D.Asch i C. Bowman. MacMilan; London Mintzberg H. Quinn J.B., 1992: The strategy Process. Concepts and Contexts. Prentice-Hall.Inc Niedzielski E., Fidejko B.,1995: Zarządzanie strategiczne przedsiębiorstwem rolniczym. (Strategic

management of agricultural enterprise) Wyd. ART.; Olsztyn Olson K.2001: A Strategic Management Primer For Farmers. Staff Paper Series, Department Of Applied

Economics, College Of Agricultural, Food, and Environmental Sciences, University of Minnesota, http://agecon.lib.umn.edu/, 2001

Rószkiewicz M.,2003: Zastosowanie narzędzi statystycznych w strategii pozycjonowania. (Statistical

tools in positioning strategy) Wydawnictwo Wydziału Zarządzania Uniwersytetu Warszawskiego; Warszawa

Sokołowski A., Sagan A.,2005: Przykłady stosowania analizy danych w marketingu i badaniu opinii

publicznej. (Examples of applying data analysis in marketing and public opinion surveys) StatSoft 2005

Trucker L.R., MacCallum R.C.1997: Exploratory Factor Analysis.The University of North Carolina,

http://www.unc.edu/~rcm/book/factor.pdf Wójcik P., 2006:,,Statystyczna analiza danych z pakietem SAS. Metody analizy wielowymiarowej –

analiza czynnikowa” (Statistical data analysis by means of SAS package. Multi-dimensional analysis method – factor analysis).

Woś A., 1998: Ustrojowe podstawy transformacji sektora żywnościowego (System foundations of food

processing sector transformation). In Polish agriculture in system transformation period (1989-1997)”. IERiGŻ; Warszawa


437

FARM INCOME RISK ASSESSMENT FOR SELECTED FARM TYPES IN POLAND - IMPLICATIONS OF FUTURE POLICY REFORMS1

Edward Majewski

Warsaw Agricultural University Faculty of Agricultural Economics

Nowoursynowska 166, 02-787 Warsaw, Poland Email: [email protected]

Waldemar Guba

Ministry of Agriculture, Poland Email: [email protected]

Adam Was

Warsaw Agricultural University, Poland Email: [email protected]

Abstract In the traditional understanding “risk arises when the stochastic elements of a decision problem can be characterized in terms of numerical objective probabilities, whereas uncertainty refers to decision settings with random outcomes that lack such objective probabilities” [Moschini, Hennessy 2001, p. 91]. This distinction, attributed to Knight (1921) is largely ignored in more recent publications. Moschini and Hennessy [op.cit.] state “we tend to use the word uncertainty mostly to describe the environment in which economic decisions are made, and the word risk to characterize the economically relevant implications of uncertainty”. Following this view any ex-ante considerations in the decision making process refer to uncertainty regarding the prediction of an un-known future, whilst risk relates more to the ex-post measuring or ex-ante assessment of economic impacts of the decisions made. As such “risk” can be defined as “uncertainty of outcomes” [EC Working Document 2001, after Hardaker, Huirne and Anderson 1997; M]. For some reason, farm-businesses, more than any other businesses, may be subject to a variety of risks, such as human or personal, asset, production or yield, price, institutional and financial [EC Working Document 2001]. Of those, production risk, mainly due to the nature of agricultural production exposed to weather conditions and dependent on the healthy growth of animals, as well as price risk, resulting from the volatility of agricultural markets, have a direct and probably the greatest impact on farm incomes. In addition the dependence of European farming on policy related transfers (market price support and direct income support) means that farm incomes are increasingly exposed to price and income risk related with the CAP reforms. For example, the consequences of the new WTO agreement may result in lower price support and greater exposure to world market price volatility. At the same time pending the policy debate on the EU budget for the next programming period will put both the forms and levels of direct farm support under public scrutiny. This paper deals with the assessment of risk for selected farm types in Poland in the perspective of the years 2013 and 2018 considering different EU farm policy scenarios creating an additional institutional risk, which through different market-related measures may affect, direct prices support and thus incomes. The typical Polish heterogeneity of farm structure creates a good basis for comparison of the income situation of different types of farms, with a focus on the probability of achieving low incomes threatening the farm’s existence. For this purpose a static simulation model using a Monte Carlo method was constructed. No changes in production structure and other possible adjustments, including investments, were considered. Keywords: risk assessment, Poland, EU farm policy, income

1 The research described in this paper is part of the 6th FP project “Income Stabilisation”, see www.incomestabilisation.org


438

Introduction In the traditional understanding “risk arises when the stochastic elements of a decision problem can be characterized in terms of numerical objective probabilities, whereas uncertainty refers to decision settings with random outcomes that lack such objective probabilities” [Moschini, Hennessy 2001, p. 91]. This distinction, attributed to Knight is largely ignored in more recent publications. Moschini and Hennessy [op.cit.] state “we tend to use the word uncertainty mostly to describe the environment in which economic decisions are made, and the word risk to characterize the economically relevant implications of uncertainty”. Following this view any ex-ante considerations in the decision making process refer to uncertainty regarding the prediction of an un-known future, whilst risk relates more to the ex-post measuring or ex-ante assessment of economic impacts of the decisions made. As such “risk” can be defined as “uncertainty of outcomes” [EC Working Document 2001, after Hardaker, Huirne and Anderson 1997]. For some reason, farm-businesses, more than any other businesses, may be subject to a variety of risks, such as human or personal, asset, production or yield, price, institutional and financial [EC Working Document 2001]. Of those, production risk, mainly due to the nature of agricultural production exposed to weather conditions and dependent on the healthy growth of animals and crops, as well as price risk, resulting from the volatility of agricultural markets, have a direct and probably the greatest impact on farm incomes. In addition the dependence of European farming on policy related transfers (market price support and direct income support) means that farm incomes are increasingly exposed to price and income risk related with the Common Agricultural Policy (CAP) reforms. For example, the consequences of the new WTO agreement may result in lower price support and greater exposure to world market price volatility. At the same time the policy debate on the EU budget for the next programming period will put both the forms and levels of direct farm support under public scrutiny. This paper deals with the assessment of risk for selected farm types in Poland in the perspective of the years 2013 and 2018 considering different EU farm policy scenarios creating an additional institutional risk, which through different market-related measures may affect, direct prices support and thus incomes. The typical Polish heterogeneity of farm structure creates a good basis for comparison of the income situation of different types of farms, with a focus on the probability of achieving low incomes threatening the farm’s existence. For this purpose a static simulation model using a Monte Carlo method was constructed. No changes in production structure and other possible adjustments, including investments, were considered. Methodology The level and volatility of farm incomes were estimated using a Monte Carlo simulation method in a farm model constructed for the @Risk package. For the simulation of farm incomes six of the most common production types in Poland (TF14), according to the FADN typology [FADN 2006a] were selected, each divided into 4 clusters by economic size (8-16, 16-40, 40-100, more than 100 ESU). The following EU agricultural policy scenarios were considered: Base 2004 Historic reference scenario. Current CAP 2013 Reflection of continuation of all existing policies, including implementation of the already agreed reforms (Luxembourg 2003) with minor assumed changes (10% mandatory modulation of direct payments).


439

Further Liberalisation 2018 Best guess for the future 2018 CAP, based on the assumption of new WTO deal and its consequences for EU price, trade and direct support policy in the sector. Full decoupling and mandatory modulation of 10% of direct payments is assumed (only the biggest farms). Full Liberalisation 2018 Withdrawal of all market and direct support measures. EU farm prices equal world market prices. The scenario parameters have been set in the following way: base scenario parameters were calculated from historical data; future price levels were assumed on the basis of the most recent FAPRI and OECD “baseline” projections. The projected prices were adjusted by assuming deviations from those projections driven by scenario specific changes in policy instruments. The assumed impact of liberalisation reflects the level of current market price support and projected world price level. volatility (standard deviation) of future prices was assumed based on analysis of variability of historic time-series for EU and world market prices. It was assumed that liberalisation enhances the volatility of future EU prices up to the level observed (historically) in case of world market prices. future yields and their volatility were estimated based on the extrapolation of the long-term historic trends in yields. The same levels of yields for all scenarios have been assumed; future inputs and costs were assumed on the basis of expert judgment. The assumption reflect changes in input prices and takes into account variety of factors driven each input market. Examples of basic assumptions regarding model parameters for future policy scenarios are presented in tables 1-4.

The key parameters of the base model which were calculated from historical data can be grouped as follows: Means of structural variables to describe the farm types (e.g. size of activities, yield, prices, inputs or costs) calculated from FADN data base for the years 2002-2004; Standard Deviation for selected variables; Cross correlations: farm related (input-output, input-input) from historical farm data; market related (price-price, price-yield; yield-yield) from national statistics data. Due to data limitations input-output correlations for crop production were not included in the model. Most of the farm activities in the model were described by the parameters of the distributions (standard deviation) of yields and prices. Similarly, the standard deviation was estimated for selected cost variables (energy, fertilizers, pesticides, seeds, purchased and farm produced feed for animals). Other variables of the model (e.g. fixed costs) were introduced as constant values specific for each farm type. For simplification a normal distribution for all variables was assumed. The distribution was truncated on

the left side at 0 for yields and for prices at the values, optionally, of σ2−x or 0 or the intervention price, depending on which was the highest.


440

The estimation of standard deviation in the base period, which is a basic measure of instability of yields and prices in the simulation model, created some difficulties related mainly to available sources of data. Data from two different sources for the period 2002-2004 and Farm Survey2 for the years 1997-2001, adjusted to FADN standards have been merged. For a given farm type (activity, size) all observations have been pooled across years (1997-2004) and standard deviations were estimated for the whole set of variables. Both data bases were merged for our estimations in the following way: all farms from the Farm Survey which represent farm types selected for simulations; randomly drawn 10% of FADN farm population. As a result the total number of farms in the “merged” data base varied, in consecutive years, between 377 in 2003 and 732 in 2004. Splitting the population of farms into selected farm types, and drawing data on single activities from smaller samples, which do not appear in all farms, reduces strongly the number of observations which can be used for estimation. As a consequence, a “within farms and across years” approach, which could be considered as the most appropriate, was not possible. That is why it was decided to pool all the observations within each farm type and estimate the standard deviation for the whole set of variables. Consequently, the analysis and its results are interpreted in relation to the experienced (ex-post for the base period) and envisaged (ex-ante for scenario analysis) situation in the population of farms, rather then in a single farm. Any estimates of means and standard deviations from the pooled data in the simulations produce a randomly chosen value that depends on all the combined sources of variation, including the hopefully small net sampling errors, between farm performance levels and year to year variations due to weather and market/policy conditions. As a result, the simulated income distributions for the represented farm population show the proportion of all farms likely to fall below some critical level, reflecting their economic viability. The statistics (mean and SD) capture all the variation even though it is not separated out into its respective components. Model parameters for the future policy scenarios, taken from available forecasts, estimated from extrapolation of existing trends or assumed according to experts’ judgment are presented in tables 1 – 4.

2 Farm Survey conducted by the Institute of Agricultural and Food Economics in Warsaw. Polish FADN, which have been established very recently, provides data for the years 2002-2004 only, but for a large sample of farms (12000 in the year 2004). The Farm Survey, which is not fully compatible with FADN, provides historical data for a long period, however for much smaller population of farms (about 1000 on average in the period considered).


441

Table 1: Price assumptions for selected commodities (Indices of EU nominal prices; Base 2004 = 100)

Commodity Current CAP

2013

Further Liberalisation

2018

Full Liberalization

2018

Wheat 105 100 98

Oil seed 105 105 105

Sugar Beets 56 56 39

Milk 83 86 69

Source: Own assumptions based on OECD-FAO 2005 and FAPRI 2006

Table 2: Estimated and assumed parameters for wheat prices and their volatility in the specialised cereal farms

Farm size (in

ESU)

Parameter Base 2004

Current CAP 2013


2018

Full Liberalization

2018

Price level* (PLN/dt)

45,7 48,3 45,7 44,6

Standard deviation

7,1 8,42 9,73 13,82 8-16 Volatility (coeff. of var. in %)

15,48 17,41 21,28 30,95


45,2 47,8 45,2 44,1

Standard deviation

6,7 8,0 9,2 13,1 16-40

Volatility (coeff. of var. in %)

14,81 16,7 20,4 29,6


47,6 50,4 47,6 46,5

Standard deviation

7,0 8,3 9,6 13,7 40-100


14,69 16,5 20,2 29,4


43,6 46,1 43,6 42,6

Standard deviation

5,2 6,1 7,1 10,1 100>


11,82 13,3 16,3 23,6

* price level for future scenarios were calculated by application of the indices for EU nominal prices from Table 1 Source: Own estimates based on FADN (Base) and OECD-FAO 2005 and FAPRI 2006 (future scenarios)


442

Table 3: Input price change assumptions (Base 2004 = 100)

Cost element Current CAP

2013


2018

Full Liberalization

2018

Fertilizers, pesticides 120% 130% 115%

Seeds 125% 140% 125%

Purchased concentrates 110% 110% 95%

Cash crops for feed 110% 110% 95%

Energy 120% 130% 130%

Land lease cost (per farm)

120% 115% 75%

Taxes (per farm) 150% 200% 150%

Other costs (per farm) 120% 130% 130%

Hired labour 150% 180% 180%

Off farm income 130% 150% 150%

Source: Own assumptions

Table 4: Assumed future yields Yields (dt/ha)

Crops

Annual rate of yield

increase (1992-2004)

Assumed future

annual rate of yield increase

Mean [2002 - 2004]

2013 2018

Wheat 0,93% 1,80% 38,4 45,1 49,3

Rye 0,85% 0,90% 24,5 26,6 27,8

Barley 1,34% 1,30% 31,7 35,6 38,0

Corn 4,13% 1,50% 57,1 65,3 70,4

Potatoes 1,84% 2,00% 189,3 226,3 249,8

Sugar beet 2,60% 2,00% 427,0 510,3 563,4

Oilseed rape 0,55% 0,50% 23,5 24,6 25,2

Milk 3,17% 2,50% 4127,3 5154,5 5831,8

Oats 1,66% 1,40% 24,8 28,1 30,2 Source: Own estimates based national statistics.

Characteristics of the analysed farm types

Basic characteristics of the FADN farm types selected for simulation is presented in table 5.


443

Table 5: Characteristics of selected farm types (mean values from the period 2002-2004)

Share in LU [%] Farm Type

[TF14]

Size class [ESU]

Agricultural area [ha]

Number of

livestock units

cows other cattle

pigs

Stocking rate

[LU/100 ha]

Specialized farm types 8-16 51,7 2,6 13% 9% 78% 5,2 16-40 112,3 3,6 11% 7% 82% 3,1 40-100 252,0 1,5 0% 98% 2% 0,6

13 cereal, proteins, oilseeds

> 100 511,7 38,8 1% 10% 89% 7,6

8-16 22,1 21,8 73% 21% 6% 98,9 41 dairy

16-40 38,5 39,5 73% 24% 3% 99,9

8-16 15,6 23,1 2% 2% 96% 148,5 16-40 27,6 48,9 1% 1% 97% 177,1 40-100 56,0 111,6 0% 1% 99% 199,1

50 pigs

> 100 128,5 442,9 0% 0% 100% 344,6 Mixed farm types

8-16 23,2 14,7 25% 16% 59% 63,4 16-40 51,6 31,8 19% 11% 70% 61,6 40-100 118,4 82,3 6% 7% 87% 69,5

81-82 mixed crop and livestock

> 100 482,8 245,7 30% 14% 56% 50,9

8-16 21,0 9,8 25% 21% 54% 46,7 16-40 41,8 18,8 16% 25% 59% 45,2

60 mixed crops 40-100 134,2 51,3 20% 15% 64% 38,3

8-16 19,6 17,9 44% 26% 31% 91,0 16-40 36,9 35,3 44% 27% 29% 95,7

71 mixed livestock 40-100 73,3 84,9 48% 20% 32% 115,8

Source : own calculations base on FADN and pre-FADN databases.

All the farms in the sample can be classified as commercial, mostly family farms. They are characterized by different production orientation and also varied level of specialization. Simulation results Simulation results allow the assessment of the impact of the policy scenarios on the average level and variability of farm incomes for the considered farm types. The risk effect was measured as a percentage of farms with negative income and income falling below the minimum wage level (for two persons). This minimum was set at the average wage level in Poland. It was adjusted for the future scenarios assuming an increase of wages in the economy (30% by 2013 and 50% by 2018). The average absolute farm incomes (table 6) show considerable stability over the analysed time horizon under CAP scenarios (Base, Most Likely, Likely). This reflects an impact of two major forces influencing


444

farm incomes in Poland in the near future, both being elements of EU farm policy, however acting in opposite directions:

• positively affecting farm incomes is the gradual implementation of the direct payments system (phasing-in), increasing payments from the initial 2004 level (55% of the full eligible rate) up to 100% in the year 2013;

• adversely affecting farm incomes is the gradual decline in farm prices due to the implementation of recent reforms (dairy) and the assumed consequences of the Doha round for the CAP market price support policy. After 2013 the only factor compensating for the decline in support of agricultural prices and increase in farm costs will be technical progress improving farm productivity.

The Liberal scenario results in a considerable decline in average farm incomes, which clearly reflects the ‘size’ of the current income support provided by the CAP. Another observation is that the mean values of farm incomes are slightly higher in specialized farms. This cannot be attributed to the fact of specialization only, because mixed farms in the sample are on average smaller in terms of the area and stocking density. The results of the simulation also show, what is quite obvious, that the greater the economic size of farms, the higher the farm incomes that are generated. Table 6: Farm income – mean (Euro per farm)

Farm Type [TF14]

Size class [ESU]

Base 2004 Current

CAP 2013

Further Liberalisati

on 2018

Full Liberalizat

ion 2018

Specialized farm types

8-16 10104 14691 13009 1902 16-40 16912 33143 30066 7663 40-100 62434 102284 92919 45058

13 cereal, proteins, oilseeds

> 100 191218 296866 282333 177758

8-16 9527 10809 13210 4351 41 dairy

16-40 21593 23429 28045 11583

8-16 7076 6970 6844 2266 16-40 13482 15338 15549 8053 40-100 32242 39207 43329 26715

50 pigs

> 100 85867 115896 131492 76043 Mixed farm types

8-16 6910 8232 8592 2164 16-40 17521 20741 21264 7428 40-100 38214 51865 52194 23134

81-82 mixed crop and livestock > 100 116556 155451 152739 26645

8-16 7012 6449 5794 501 16-40 12632 13802 13238 2247

60 mixed crops 40-100 65923 81198 76545 47524

8-16 5151 6437 7366 1173 16-40 12019 14839 17532 4975

71 mixed livestock 40-100 39354 41570 48088 19401

Source: Own calculations


445

A similar pattern emerges from the simulated distribution of farm incomes (table 7) . Even though farm types differ in terms of the percentage of holdings with negative income in any scenario, these percentages for most types remain below 10% in all CAP scenarios. However, the situation changes drastically in the liberal scenario under which the percentage of farms with negative incomes is significantly increased. Nevertheless, more than half of farms generate positive incomes in the liberal scenarios. The risk of farm incomes falling below the assumed minimum wage level increases in all farm types and policy scenarios (table 8). Under the liberal scenario, the majority of farms of small economic size are not able to reach an adequate income. The ability to generate a positive Farm Income allowing the covering of cost of own labour is one of the key factors determining economic sustainability of a farm. In our simulation the assumption was made that a farm should provide an income equal to at least the national minimum wages for 2 fully employed persons (in the base year 2400 Euro per person). Respectively, 2013 and 2018 minimum wages were assumed at the level of 3120 and 3600 Euro per person, taking into account the expected growth of the Polish economy and likely increase of wages. It is important to emphasize that the model applied is a static one and no adjustments such as changes in production structure or increases of the production scale were considered. The simulation results provide an insight into the proneness of different farm types to assumed policy changes. It turns out that the risk of making losses is considerably dependent on the production orientation and economic size of a farm. The least susceptible to income risk appear to be farms specializing in milk production (TF 41), which can be explained by a relative stability of milk prices and milk yields. Among specializing farms a somewhat higher exposure to risk was detected for cereal farms (TF 13), which partly results from greater yield and price volatility. Those most exposed to income risk appear to be farms specializing in pig production (TF 501), which can be explained by the high volatility of pig and feed prices. A wider portfolio of production activities on a farm should diminish the risk of negative income. This hypothesis is confirmed by the simulation results: non-specialized farms with a mixed production structure showed lower income risk compared with specialized pig farms, despite the fact that pigs accounted for more than half of the livestock. Among the mixed farms considered the lowest risk occurs in mixed livestock farms (mainly grazing - TF 71). These farms are characterized by about 30% share of pigs in the total number of livestock units, which enhances income instability. On the other hand, the domination of cattle has a stabilizing impact on farm income in that category of farms. Somewhat higher income risk occurs in mixed cropping farms (TF 60) and mixed farms (TF 81-82), though in the liberal scenario mixed farms (TF 81-82) are less exposed to risk than the farms where crop production dominates.


446

Table 7: Risk of getting farm income below zero

Farm Type [TF14]

Size class [ESU]

Base 2004 Current

CAP 2013


on 2018

Full Liberalizat

ion 2018

Specialized farm types 8-16 5,7% 1,8% 3,4% 40,1% 16-40 14,0% 3,5% 6,0% 34,6% 40-100 5,7% 1,4% 2,6% 19,3%

13 cereal, proteins, oilseeds > 100 2,1% 0,4% 0,9% 8,3%

8-16 0,4% 0,1% 0,1% 11,0% 41 dairy 16-40 0,1% 0,0% 0,0% 3,4%

8-16 13,5% 17,7% 20,3% 39,5% 16-40 17,0% 15,8% 18,5% 33,2% 40-100 16,8% 13,8% 15,1% 26,3%

50 pigs

> 100 25,8% 21,4% 21,4% 32,5% Mixed farm types

8-16 2,6% 1,5% 1,7% 31,4% 16-40 2,1% 1,2% 1,8% 24,2% 40-100 5,8% 2,2% 3,6% 22,2%

81-82 mixed crop and livestock > 100 7,2% 4,0% 5,3% 39,1%

8-16 1,1% 2,1% 5,3% 44,8% 16-40 1,4% 1,5% 2,3% 36,7%

60 mixed crops 40-100 0,2% 0,1% 0,5% 4,7%

8-16 3,2% 1,4% 1,0% 36,4% 16-40 1,3% 0,4% 0,3% 19,6%

71 mixed livestock 40-100 0,2% 0,2% 0,2% 8,5%



447

Table 8: Risk of getting farm income below minimum wage

Farm Type [TF14]

Size class [ESU]

Base 2004 Current

CAP 2013


on 2018

Full Liberalizat

ion 2018 Specialized farm types

8-16 19% 11,7% 22,5% 76,1% 16-40 22% 6,5% 12% 49,5% 40-100 7,6% 2,0% 3,8% 24,0%

13 cereal, proteins, oilseeds > 100 2,8% 0,5% 1,0% 9,9%

8-16 11,0% 10,6% 7,0% 80,5% 41 dairy 16-40 1% 1% 0% 27,3%

8-16 37% 46% 52% 73,0% 16-40 28% 28% 31% 51,2% 40-100 20% 18% 20% 33,9%

50 pigs

> 100 27% 22% 22% 33,4% Mixed farm types

8-16 27,6% 30,6% 38,9% 88,6% 16-40 7,5% 6,5% 8,7% 50,8% 40-100 8,7% 4,2% 6,6% 29,9%

81-82 mixed crop and livestock > 100 8,4% 4,4% 6,5% 43,1%

8-16 24,4% 47,0% 64,7% 96,6% 16-40 9,1% 11,0% 18,0% 78,3%

60 mixed crops 40-100 0,5% 0,3% 0,9% 8,5%

8-16 46,2% 48,7% 50,2% 97,3% 16-40 9,6% 5,9% 5,7% 65,8%

71 mixed livestock 40-100 0,8% 0,7% 0,4% 19,3%


Simulation results verify the hypothesis that farms from clusters of smaller economic size are to the greatest extent exposed to the risk of generating farm incomes below the set threshold level. This is visible especially under the liberal scenario, within the assumptions made. There is a significant difference in the percentage of farms with negative incomes if the CAP and the Liberal scenarios results are compared. It indicates the strong income stabilizing effects of the CAP for smaller scale farms. Conclusions Assessment of risk in farms of different types gives an insight into the phenomenon of strongly diversified farm structure in Poland and allows us to simulate the impact of policy changes on the risk of financial outcomes in the future. The simulation results show that likely reduction of the price protection for most agricultural commodities and reduction of direct payments may result in the increased risk of achieving low farm incomes, although the differences between the three CAP scenarios are not very marked. The gradual liberalization of the CAP which may be expected in the future will moderately affect Polish farms especially in the perspective of the year 2013 due to the increase of direct payments in line with the phasing-in schedule. The most exposed to the risk of low incomes are pig farms and mixed farms with a


448

high share of pigs in the total livestock numbers. Farm types which seem to be least sensitive to the assumed policy changes are dairy and farms from all production orientation types of greater economic size. Still, the smaller farms are relatively strongly protected by all CAP scenarios. More radical policy changes, as represented by the Liberal scenario, would dramatically worsen the financial situation of smaller farms, very likely driving a large number of farm holdings out of business. The most recent years in Poland are marked by the rapid concentration of production in commercial farms in almost all farm activities. A strong liberalization of the agricultural policies would speed up significantly such structural changes in the Polish farming sector. References: European Commission, Agriculture Directorate-General, 2001. Working Document “Risk Management

Tools for EU Agriculture”. FADN Poland, 2006a. Standard results obtained by the farms participating in Polish FADN in 2005.

http://www.fadn.pl/mediacatalog/documents/wyniki_stand_ogolne_2005r.pdf FADN Poland, 2006b. Standard results obtained by the individual farms having the accountancy system

in 2005. http://www.fadn.pl/mediacatalog/documents/wyniki_stand_indywidualne_2005r.pdf FAPRI 2006 U.S. and World Agricultural Outlook; Food and Agricultural Policy Research Institute; Iowa

USA Hardaker, J.B., Huirne, R.B.M. and Anderson, J.R. (1997), Coping with Risk in Agriculture. CAB

International, Oxon, United Kingdom. ISBN 0 85199 199 X. Moschini G., Hennessy D.A., 2001. Uncertainty, risk aversion, and risk management for Agricultural

producers. Handbook of Agricultural Economic, vol. 1A Agricultural Production, edit. Gardner B.L., Rausser G.C. Elsevier.

OECD – FAO Agricultural Outlook: 2005 – 2014 Statistical Yearbook Poland – Main Statistic Office publications 1979-2006


449

ESTIMATING INPUT-SPECIFIC RECOMMENDATIONS FOR TECHNICALLY INEFFICIENT CROP FARMERS

N Matthews*, B Grové and G Kundhlande

Address: Department of Agricultural Economics University of the Free State

PO Box 339, Bloemfontein, South Africa,9301 Email: [email protected]

Abstract The main objective of this paper is to demonstrate how data envelopment analysis (DEA) and production function analysis are combined to evaluate the input usage decisions of technically inefficient maize producers and to estimate potential yield gains from reallocating inputs. The data for this research originates from a production cost survey. Two distinct quadratic production functions were estimated for DEA-determined technically efficient and inefficient farmers using a dummy variable approach. Results indicate that the yield gains are large if farmers are able to produce in a technically efficient manner while maintaining current input levels. Results further indicate that even though a producer may be technically inefficient, yield gains and therefore increased profits are possible if such farmers adjust nitrogen application rates. The conclusion is that failure to recognise the fact that DEA-determined technically efficient and inefficient producers’ production processes are characterised by two distinct production functions renders uniform input recommendations inappropriate. In future, extension officers should aim to develop input recommendations taking these inefficiencies into account. Keywords: technical efficiency, input recommendations, production functions, data envelopment analysis Introduction According to Coelli (1996) modern efficiency measurement began with the work of Farrell (1957) who defines technical and allocative efficiency as two important concepts of economic efficiency. Technical efficiency (TE) reflects the ability of the farm to produce a certain level of outputs with minimal inputs, whereas allocative efficiency (AE) refers to the ability of the farm to choose optimal combinations of inputs given their respective prices. One needs to be technically efficient before one can be allocatively efficient, and attainment of both technical efficiency and allocative efficiency is required for economic efficiency. Central tendency measures have been used for many years to identify and estimate inefficiencies in the performance of decision-making units (DMUs). According to Bardhan et al. (1998), Feldstein (1974) was one of the first to use ordinary least square (OLS) regression to evaluate efficiency by estimating a production function. A DMU is classified as efficient if its observed output is higher than the estimated output from the regression, and inefficient if it is lower. Within a South African context Joubert and Viljoen (1974) used a Cobb-Douglas function to determine the relationship between production cost and technical efficiency. Viljoen and Groenewald (1977) also used a cross-sectional Cobb-Douglas production function to distinguish between different levels of efficiency within a homogeneous set of DMUs. However, advances in analytical approaches such as stochastic frontier analysis (SFA) and data envelopment analysis (DEA) have reoriented the evaluation of DMUs away from average or central tendency approaches toward best practice or frontier approaches. Recently Mushunje et al. (2005) explained the differences in technical efficiency between cotton farmers on communal land and those on resettled land in Zimbabwe using a Cobb-Douglas-type stochastic


450

production frontier model. Other researchers have used regression analyses to explain technical efficiency scores calculated with DEA (Galanopoulos et al., 2006). In most of the literature in which DEA is applied, researchers seek to measure the efficiency of DMUs and to explain the estimated technical efficiency (TE) scores using socio-economic and other production variables (e.g. Charnes et al., 1978; Galanopoulos et al., 2006). Some researchers also use TE scores to try and explain observed performance of DMUs (Zaibet et al., 2004). Most early studies on technical efficiency concentrated on explaining the inefficiencies through the use of regression analysis. Although these studies were able to explain the inefficiencies, they were unable to determine specific sources and amounts of inefficiency attributed to input use. In the literature there is evidence of frustration with the failure to provide quantitative information to guide decisions on reallocation of input use in most efficiency studies, for example that of Bowlin (1998). By implication, agricultural advisors are therefore unable to make input-use recommendations to a technically inefficient farmer or to estimate potential yield gains if such a farmer were to become technically efficient. The main objective of this paper is to demonstrate how DEA and production function analysis can be combined to evaluate the input use decisions of technically inefficient maize producers in South Africa and to estimate potential yield gains from improvement in technical efficiency. The above objective is achieved using a procedure proposed by Bardhan et al. (1998), which combines DEA and OLS in a two-stage manner to estimate sources and amounts of input-specific inefficiencies. The rest of this paper is structured as follows: In section 2 the data used in the study is discussed, while section 3 gives an outline of the linear programming and econometric models, as well as the procedure for estimating yield gains when a DMU increases its technical efficiency. Section 4 presents the results of the analyses, while section 5 provides a summary and conclusion. Data The data for this research originates from a production cost survey of white-maize producers in the Bothaville, Wesselsbron and Viljoenskroon areas in the Free State Province of South Africa (Le Clus et al., 2004). The farmers in this area are cash-crop producers but focus on maize production, as other crops are precluded by certain factors (e.g. product prices). The production cost survey covered 62 producers, but due to shortcomings in the data it was possible to extract a complete dataset for only 24 farmers in terms of crop yield, fertiliser applications, tractor size, and seeding rates on a per-hectare basis. However, seed (plant density) was not included in the analysis due to cultivar differences. Fertiliser applications were standardised on a mineral N basis, measured in kilograms per hectare (kg/ha), while tractor size, in kilowatts (kW), was measured as average tractor size per area of maize planted. Tractor size utilisation is a proxy for timeliness in carrying out critical operations such as planting, which indicates whether tractors are used effectively during the production process. Thus, by using tractors more intensively, farmers can increase the area of land cultivated, but have to work longer hours during the land preparation and planting stages of the production process so that planting is completed by the onset of the rainy season to ensure an increase in tractor productivity. Data on soil tests was not available, and as a result the study did not account for any nitrogen already in the soil. Weather, crop rotation, production method and labour were also not taken into consideration in the analyses, as the relevant data was not available. Table 1 shows the input use and maize yield for the total sample, as well as the DEA-determined technically efficient and inefficient sub-groups of farmers.


451

Table 1: Summary statistics on input use and maize yield for DEA-determined technically efficient and inefficient farmers in the Bothaville, Wesselsbron and Viljoenskroon areas

Average Minimum Maximum

Standard deviation

All farmers (n = 24) Area planted (ha) 831 120 2369 661 Yield (ton/ha) 4.00 1.50 5.70 0.92 Tractor size utilisation (kW/ha)

14.58 2.63 50.00 10.93

Nitrogen (kg/ha) 71.20 29.83 107.27 20.15 Efficient (n1 = 8 ) Area planted (ha) 1411* 312 2369 825 Yield (ton/ha) 4.08 2.90 5.70 1.00 Tractor size utilisation (kW/ha)

7.82* 2.63 17.95 5.41

Nitrogen (kg/ha) 65.84 29.83 99.45 23.15 Inefficient (n2 = 16 ) Area planted (ha) 540* 120 1100 290 Yield (ton/ha) 3.90 1.50 5.10 0.91 Tractor size utilisation (kW/ha)

17.96* 6.91 50.00 11.53

Nitrogen (kg/ha) 73.88 42.12 107.27 19.19 * H0: µ1=µ2, Rejected at a 5% level The average area cultivated by maize farmers can be seen in Table 1 as 831 ha. Efficient farmers cultivate 1141 ha, which is significantly (p<0.05) more than that cultivated by inefficient farmers, i.e. 540 ha. On average 71.2 kg of nitrogen and 14.58 kW per cultivated hectare are used to achieve a crop yield of 4 t/ha. DEA-determined technically efficient farmers achieve the same crop yields (4.08 t/ha) as the sample average. However, the DEA-efficient farmer uses 5.36 kg less nitrogen fertiliser, while the tractor size utilisation factor is 6.67 kW lower than the average of all the sample farmers. On the other hand, DEA-determined technically inefficient farmers achieve more or less the same crop yield (3.9 t/ha) as the average for all farmers, but utilise higher input amounts. On average, DEA-inefficient farmers use 2.68 kg more nitrogen, while the tractor size utilisation factor is 3.38 kW more. There is a significant difference (p<0.05) between DEA-determined inefficient and efficient farmers in respect of tractor utilisation. The lower tractor utilisation factor for the DEA-efficient farmers implies that these farmers cultivate larger areas per average tractor size than DEA-inefficient farmers. Procedures Classification of DEA-determined technically efficient and inefficient farmers The dual formulation of the mathematical programming model proposed by Banker et al. (1984) is used to derive the efficiency frontier under conditions of variable returns to scale. Following Kalvelagen (2002) the specification of the model is: j

zz Θ=

,min

λ, j = 1,2,…J (1)

subject to kj

J

j

kjj yy ≥∑=1

λ , k = 1, 2, ..K (2)


452

∑≥Θj

ijjijj xx λ , I = 1, 2, …I (3)

11

=∑=

J

j

jλ (4)

For the above model it is assumed that there are J DMUs each producing K outputs using I inputs, such that ykj represents the amount of the kth output produced by the jth DMU, and xij represents the amount of the ith input used by the jth DMU. z is the measure of technical efficiency and equals the ratio of the minimal feasible input usage to the current input usage and therefore Θj is the relative efficiency score for the jth DMU. λj are the weights to be used as multipliers for the input levels of the referent DMU to indicate the input use level for which an inefficient farm should aim in order to achieve efficiency. The model specification requires that optimisation be carried out DMU by DMU. The assumption of constant returns to scale means that farmers are able to linearly scale inputs and outputs without increasing or decreasing efficiency. However, this assumption only holds if the farm operates at an optimal size. Factors such as imperfect competition, constraints on finance, etc. can result in a farm not operating at optimal scale (Cinemre et al., 2005). This study assumes variable returns to scale, as the market under which a farmer operates is not perfect. GAMS (Brooke et al., 1998) was used to develop code to optimise each DMU’s efficiency score within a loop. The optimised TE scores indicate that eight farmers (33%) defined the efficiency frontier. The efficiency of the DEA-determined technically inefficient farmers ranged from 54%-89% with an average of 74%. In order to make input recommendations and to calculate yield gains, production functions were needed for the DEA-determined efficient and inefficient farmers. The following section describes the procedure used to estimate these functions. Estimation of production functions Both a Cobb-Douglas and a quadratic production function were explored as the functional forms to characterise the production process. However, the Cobb-Douglas form was dropped in favour of the mixed linear quadratic form due to statistical significance. Problems with degrees of freedom may occur, because only eight farmers were DEA efficient. To overcome the problem, production functions for both groups of DMUs were estimated simultaneously using a dummy variable approach (Bardhan et al., 1998). More specifically, the following function was estimated:

22322110

22322110 DxDxDxDxxxy δδδδββββ +++++++= (5)

where x1 is kilowatts, x2 is mineral nitrogen and D is a dummy variable indicating DEA-determined technically efficient DMUs (D=1). The symbols β0, β1, β2, and β3 are the estimated coefficients for DEA-determined technically inefficient DMUs, whereas the symbols δ0, δ1, δ2 and δ3 estimate changes to the β parameters due to efficiency differences between the two groups of farmers. One benefit of using interaction terms (dummy variable interacting with inputs) is that it allows one to test whether there is a significant difference in the estimated production functions for DEA-determined technically efficient and inefficient farmers. Calculation of gains to DMU Using the estimated production functions it is possible to calculate the expected yield gains available to DEA-inefficient farmers. These gains may stem from two sources: Firstly, yields may improve at current


453

input levels if DEA-determined technically inefficient DMUs are able to become technically efficient (movement to DEA-efficient production function). Secondly, DEA-inefficient DMUs may gain through changes in input levels while remaining on the inefficient production function (movement along DEA-inefficient production function). Gains from moving to DEA-efficient production function The following equation is used to estimate the potential yield gains of farmers moving towards the DEA-efficient production function at current input combination (Bardham et al., 1998):

22322110

1

2221

1

2221 0),,(1),,(

xxx

DxxxyEDxxxyEn

i

iii

n

i

iii

δδδδ +++=

=−

= ∑∑

== (6)

The δ1 term represents the average increase in yield associated with the first input, while the δ2 term represents the average increase associated with the second input, and the δ3 term represents the quadratic form of the second input. The δ0 term records the expected average gains from using both inputs in non-zero amounts. Note that δ refers to the estimated changes in the β parameters of the DEA-inefficient production functions. Thus, if δ parameters are not significant, only one production function is estimated and no yield gains are possible. Gains from changing input combinations on the DEA-inefficient production function Standard production economic theory is used to guide changes in input use when inefficient DMUs are unable to enhance their efficiency through a movement to the DEA-efficient production function. Figure 1 shows the average physical product (APP), marginal physical product (MPP), and the stages of production. Stage I of production is irrational, because APP is increasing. Furthermore, MPP is greater than APP and as a result the elasticity of production (EP) is greater than one. Production in stage III is also irrational because MPP<0, which implies that yield starts to decrease. Without any price information it is rational to increase input use to at least the beginning of stage II of production (xmin). On the other hand, if excessive input quantities are used, input use should be reduced to the end of stage II of production (xmax). Any level of input use between xmin and xmax is technically rationed. To determine the specific level of input within the range of xmin and xmax, price information is necessary. The calculated EP with respect to all inputs indicates that tractor utilisation for all DMUs occurs in stage II of production. For the DEA-efficient DMUs, nitrogen use also occurs in stage II, but for the DEA-inefficient DMUs, nitrogen use for production occurs in stages I, II and III. Fifty percent of the DEA-inefficient DMUs use nitrogen in accordance with stage I (EP <1), while 25% use nitrogen in accordance with stage II (0< EP <1) and the last 25% use nitrogen in accordance with stage III (EP <0). The gains to DMUs producing in stage I were calculated by increasing input use for each DMU to xmin. On the other hand the gains for DMUs producing in stage III were calculated by decreasing input use to xmax.


454

Figure 1: Graphical representation of the stages of production using average physical product (APP) and marginal physical product (MPP)

Results Estimated production functions for DEA-determined technically efficient and inefficient farmers The estimated production function for both the DEA-efficient and the DEA-inefficient farmers is shown in Table 2. The estimated coefficients show that there are significant differences between DEA-efficient and DEA-inefficient farmers. In essence two separate production functions exist – one for DEA-efficient farmers and another for DEA-inefficient farmers. Both sets of farmers need to use inputs in non-zero amounts to realise an output. All the coefficients for the inefficient production function are significant at a 1% level with the exception of tractor size utilisation, which is significant at a 10% level. Changes to the production function coefficients due to improvements in efficiency are all significant at a 5% level. Table 2: OLS estimates of the coefficients of production function

Coefficient Standard

error t-stat

Intercept (β0) -8.765 2.189 -4.004* Tractor size utilisation (β1) 0.023 0.012 1.866*** Nitrogen (β2) 0.303 0.056 5.378* Nitrogen2 (β3) -0.002 0.000 -4.878* Dummy (DEA-efficient farmers)

(δ0) 6.903 2.814 2.453**

D_Tractor size utilisation

(δ1) 0.112 0.043 2.612**

D_Nitrogen (δ2) -0.193 0.076 -2.541** D_Nitrogen2 (δ3) 0.001 0.001 2.429** R2 0.804 Adjusted R2 0.718 F-Statistics 9.379*

* Significant at a 1% level ** Significant at a 5% level *** Significant at a 10% level


455

The relationship between crop yield and the tractor utilisation factor is linear and positive. The estimated values indicate that the rate at which the tractor utilisation factor contributes to yield is 0.112 units higher for the technically efficient farmers. Thus, if a DEA-efficient farmer were to increase the tractor size or decrease the area over which a tractor is used, the increase in yield would be greater than if a DEA-inefficient farmer were to increase the tractor size or decrease the area planted in the same manner. What is interesting to note is that the rate at which nitrogen is transformed into crop yield is significantly lower (-0.193) for the DEA-efficient farmers. In addition, an estimated coefficient of 0.001 for the squared nitrogen term indicates that the production function has a lower curvature (i.e. the rate at which the production function turns as more nitrogen is applied is slower) than that for the DEA-inefficient farmers. Furthermore, the differences in the intercept term suggest that there might be other factors that could help explain the differences between DEA-efficient and DEA-inefficient farmers. The values of the coefficients β0 and δ0 suggest that the DEA-inefficient production function lies below the production function of the DEA-determined technically efficient farmers. In addition, the estimated coefficients also suggest that the production function associated with the DEA-efficient farmers is flatter than that of the DEA-inefficient farmers. Looking at the partial productivity of the inputs, the marginal productivity of nitrogen is lower for DEA-efficient farmers compared to DEA-inefficient farmers. One possible explanation for this might be that only a portion of the production function is estimated. For instance, if all the efficient farmers are producing in the second phase of production, the estimated coefficients will characterise the technical transformation of inputs in that phase. Estimated gains available to farmers Table 3 shows the yield gains from becoming DEA efficient (movement to efficient production function) and from changing the level of input use (movement along the inefficient production function). Farmers who are able to increase their efficiency by moving to the higher DEA-efficient production function stand to receive relatively large gains. The calculated average yield gains indicate that maize yield would increase by 1.92 t/ha if the DMU were to become DEA efficient. Some of the indicated gains in Table 3 are unrealistic (DMU 6 and DMU 8). This can be the result of a random shock that occurred during the year under consideration that was not taken into account during estimation of the production function. If the farmer is unable to move to the higher DEA-efficient production function due to some factor like incomplete information or inability to acquire better technology, the farmer should at least aim to move into the rational stage of production (0< EP <1). Among the DEA-inefficient farmers, eight operate in stage I of production and should increase nitrogen application. The average amount of nitrogen that should be added to current input use levels is 10.79 kg/ha, which would result in a yield increase of 0.95 t/ha. Four DMUs are using input up to stage III of production and these DMUs would therefore gain by reducing input use levels. On average, 15.8 kg/ha less nitrogen should be applied, which would result in average yield gains of 0.41 t/ha.


456

Table 3: Yield gains for inefficient farmers resulting from change in technical efficiency (movement to efficient production function) and from a changing level of input use (movement along the inefficient production function)

Movement along the production function

Technical efficiency change

(moving to higher

production function)

Changes in input amounts and crop yields

From stage I to II

From stage III to II

Respondent

Yield gains Nitrogen

Yield Nitrogen

Yield

t/ha kg/ha t/ha kg/ha t/ha 1 0.63 4.09 0.27 2 1.56 5.25 0.35 3 0.63 5.00 0.34 4 1.38 8.51 0.62 5 0.96 9.60 0.72 6 5.82 13.11 1.25 7 1.83 13.90 1.15 8 4.23 26.88 2.93 9 1.45 -21.27 0.75 10 1.26 -18.06 0.53 11 2.71 -13.67 0.30 12 0.59 -7.30 0.08 Average 1.92 10.79 0.95 -15.08 0.41 Standard deviation

1.60 7.46 0.88 6.05 0.29

Summary and Conclusions The research demonstrates how DEA and OLS regression can be combined to estimate gains to DMUs attributed to technical efficiency gains. During the first stage, DEA is used to group DMUs into two groups, namely technically efficient and technically inefficient farmers. The classification is carried forward in the form of a dummy variable to the next stage where production functions are estimated. These production functions are then used to determine benefits to DMUs attributed to technical efficiency gains. The results indicate that two distinct production functions exist for DEA-determined technically efficient and inefficient farmers. Results further indicate that yield gains are large if farmers are able to produce in a technically efficient manner while maintaining current input levels (movement from DEA-determined technically inefficient to DEA-determined technically efficient production function). However, care should be taken when calculating yield gains, as there are differences between the two production functions. Recall that the estimated production function of the DEA-efficient farmers has a lower curvature than that of the inefficient farmers. In essence these two production functions diverge from each other, which may have resulted in unrealistic yield gains. The procedure thus needs to be validated further with a larger dataset, which would improve the validity of the production functions.


457

Using the production function of the DEA-inefficient farmers to make input recommendations is a novel approach and should be explored further. Results indicate that even though a producer is technically inefficient, yield gains and therefore increased profits are possible if the farmer adjusts nitrogen application in accordance to tractor utilisation rate (movement on the DEA-determined technically inefficient production function). It can be concluded that failure to recognise the fact that DEA-determined technically efficient and inefficient producers’ production processes are characterised by two distinct production functions renders uniform input recommendations inappropriate. In future, extension officers should aim to develop input recommendations taking these inefficiencies into account. References BANKER, R.D., CHARNES, A., and COOPER, W.W. (1984). Some models for estimating technical and

scale inefficiencies in Data Envelopment Analysis. Management Science, 30(9): 1078-1092. BARDHAN, I.R., COOPER, W.W., & KUMHAKAR, S.C. (1998). A simulation study of joint uses of

data envelopment analysis and statistical regressions for production function estimation and efficiency evaluation. Journal of Productivity Analysis, 9:249-278.

BOWLIN, W.F. (1998). Measuring performance: An introduction to data envelopment analysis (DEA).

Technical Report, Department of Accounting, University of Northern Iowa, Cedar Falls, IA. BROOKE, A., KENDRICK, D., MEERAUS, A, & RAMAN, R. (1998). GAMS: A User’s Guide. New

York: GAMS Development Corporation. CHARNES, A., COOPER, W., & RHODES, E. (1978). Measuring the efficiency of decision making

units. European Journal of Operational Research 2: 429-444. CIMENRE, H.A., CEYHAN, V., BOZOGLU, M., DEMIRYUREK, K., & KILIC, O. (2005). The cost

efficiency of trout farms in the Black Sea Region, Turkey. Aquaculture, Article in press. COELLI, T. (1996). A guide to DEAP Version 2.1: A data envelopment analysis (computer) program.

Centre for Efficiency and Productivity Analysis (CEPA) Working paper 96/08. FARRELL, M.J. (1957). The measurement of productive efficiency. Journal of the Royal Statistical

Society A, 120(3):253-290. FELDSTEIN, M.S. (1974). Econometric studies of health economics. In M. Intriligator and D. Kendrick

(eds.), Frontiers of quantitative economics, Amsterdam: North-Holland Press GALANOPOULOS, K., AGGELOPOULOS, S., KAMENIDOU, I., & MATTA, K. (2006). Assessing

the effects of managerial and production practices on the efficiency of commercial pig farming. Agricultural Systems, 88(2/3):125-141.

JOUBERT, J.S.G. & VILJOEN, P. (1974). Production costs of crops in the North-Western Free State and

the factors which influence these costs. Agrekon, 13(3):10-19. KALVELAGEN, E. (2002). Efficiency solving DEA models with GAMS. GAMS CORPORATION.

Online <http://www.gams.com/~erwin/dea/dea.pdf>


458

LE CLUS, C.F., SPIES, D.C., TALJAARD, P.R., JOOSTE, A. & VAN SCHALKWYK, H.D. (2004). Report on a study to investigate methodology for the determination of the average production cost of maize. Grain SA, Bothaville.

MUSHUNJE, A., FRASER, G., & BELETE, A (2005). Measuring technical efficiency of cotton farmers

in Manicaland Province in Zimbabwe. Paper presented at the 43rd Annual Conference of the Agricultural Economics Association of South Africa (AEASA). Polokwane, Limpopo Province, 21-23 September 2005.

VILJOEN, P. & GROENEWALD, J.A. (1977). An approach to farming efficiency analysis as applied in

Rûens. Agrekon, 16(4):6-13. ZAIBET, L., DHARMAPALA, P.S., BOUGHANMI, H., MAHGOUB, O. & AL-MARSHUDI, A.

(2004). Social changes, economic performance and development: the case of goat production in Oman. Small Ruminant Research, 54:131-140.


459

ECONOMIC COMPARISON OF DIVERGENT STRAINS OF HOLSTEIN-FRIESIAN COWS IN VARIOUS PASTURE-BASED PRODUCTION SYSTEMS

S. McCarthy, B. Horan*, P. Dillon, P. O’Connor

Moorepark Dairy Production Research Centre, Fermoy, Co. Cork, Ireland


M. Rath

School of Agriculture, Food Science & Veterinary Medicine, UCD, Belfield, Dublin 4

Ireland

L. Shalloo

Moorepark Dairy Production Research Centre, Fermoy, Co. Cork, Ireland

Abstract The objective of this paper was to compare the economic efficiency of three divergent strains of Holstein-Friesian cows. Each strain was randomly allocated to one of three feeding systems: high milk output per cow from pasture (MP), high concentrate feeding system at pasture (HC), and high milk output per unit area from pasture (HS). Physical performance data was obtained from a 5-yr study conducted previously. A stochastic budgetary simulation model was used to simulate a model farm. The economic performance of each strain and feed system was derived for three production scenarios. Within all scenarios, profit was maximised where production was achieved at minimum cost as demonstrated by the comparably greater profitability of the low concentrate (MP and HS) systems. These results show that exclusive genetic selection for increased milk production results in reduced farm profitability as the productivity gains achieved are outweighed by associated increase in reproductive wastage costs in a pasture-based system. Keywords: strain of Holstein-Friesian, economic scenario, pasture-based system

Introduction The Irish dairy industry will experience considerable change in the years ahead (Hennessey and Thorne, 2006). Among the main agents of change, reform of EU agricultural policy, increased environmental regulation and economic prosperity will dramatically change the production landscape. The challenge for Irish dairy farmers is to increase the competitiveness of their business through innovation, productivity gain and increased operational scaas the industry evolves (Shalloo et al., 2004a). Genetic improvement of the dairy herd is one avenue to increased profitability on Irish dairy farms (Veerkamp et al., 2002) in a more competitive international dairy production environment. To evaluate the economic effects of animal performance variation arising from various alternate genetic selection strategies, a comprehensive multidisciplinary systems approach is required incorporating the effects on all major farm components including production revenues as well as variable and fixed costs. Agricultural policy has major implications for the evolution of production systems with reforms likely to result in a single world market focused dairy industry free from milk quota restrictions. Such reforms may result in reduced and more unstable farm gate prices (Dillon et al., 2005). The consequences of various genetic selection strategies must therefore be appraised with due consideration to future agricultural policy outcomes.


460

Until recently, milk yield has been the main objective criterion for selection in most temperate countries and the use of Holstein-Friesian genetics of North American ancestry has been ubiquitous. The popularity of the North American Holstein-Friesian was most likely because of its increased productivity over other dairy breeds. Overwhelming evidence now shows antagonistic associations between production and health traits (Horan et al., 2004; Evans et al., 2004; Rauw et al., 1998), continued selection for greater milk yield is anticipated to have deleterious consequences for health and fitness of the dairy herd (Pryce and Veerkamp, 2001). Reproductive performance affects the amount of milk produced per cow per day of herd life, breeding costs, rate of voluntary and involuntary culling and rate of genetic progress for traits of importance (Plaizier et al., 1997; Lopez-Villalobos et al., 2000), as well as having a significant effect on the overall profitability of a dairy herd (Britt 1985). The Economic Breeding Index (EBI) was introduced in Ireland in 2001 to identify genetically superior animals to increase profitability within Irish dairy herds (Veerkamp et al., 2002). The EBI is currently composed of five sub-indexes (relative emphasis in parenthesis): milk production (49%), fertility/survival (32%), calving performance (8%), beef performance (6%) and health (5%). The objective of the present paper was to investigate the profitability of three strains of Holstein-Friesian dairy cows differing in genetic potential for milk production and reproductive performance across three pasture-based production systems based on various alternate production scenarios arising from changes in EU Agricultural policy. Materials and Methods Production Study Details

The design of the 5-yr study and a subset of the production and reproduction data used in the analysis of the various strains and systems of production in the present evaluation have been reported by Horan et al. (2004, 2005). Briefly, three strains of Holstein-Friesian cows were compared: high production North American (HP), high durability North American (HD) and New Zealand (NZ). The HP strain was chosen on the basis of superior pedigree index for milk production, while the HD strain was selected on the basis of superior pedigree index for milk production, fertility and muscularity traits. The NZ strain was selected using the highest possible genetic merit expressed in the New Zealand genetic evaluation system (Breeding Worth). Primiparous animals entering the herd from spring 2003 onwards were bred from within each strain using sires concurrent to the different breeding objectives as outlined above relative to that strain. Each strain represented on average, thirteen sires over the five years of the study. The mean pedigree index (from the February 2004 international evaluations of the INTERBULL Animal Center, Uppsala, Sweden) for each strain is displayed in Table 1.


461

Table 1: The mean pedigree index for the three strains of Holstein-Friesian cows studied based on their predicted transmitting abilities (and SD) for milk production, survival and calving interval.

Strain High Production

High Durability New Zealand

Milk (kg) +194(90.8) +76(61.4) +52(56.0) Fat (kg) +9.0(2.96) +6.3(2.84) +8.6(2.66) Protein (kg) +8.8(2.39) +5.7(1.58) +4.2(1.33) Fat (g/kg) +0.3(0.53) +0.7(0.56) +1.3(0.58) Protein (g/kg) +0.4(0.23) +0.6(0.30) +0.5(0.21) Survival (%) -0.5(1.11) +0.4(0.51) +1.2(0.62) Calving interval (days) +0.44(1.57) -1.2(0.71) -1.6(0.86) Overall EBI1 (€) 51 58 75 Sub indices2: Milk (€) 46 32 41 Fertility (€) 2 25 38 Calving (€) 2 0 5 Health (€) -2 0 -5 Beef (€) 1 1 -9

All predicted differences were obtained from the February 2004 international evaluations of the INTERBULL Animal Centre (Uppsala, Sweden) using the MACE (multi-trait across-country evaluation). 1EBI = Economic Breeding Index. 2Subindices are derived from the economic values of individual traits: Milk (-€0.084/kg) fat (€1.55/kg), protein (€5.27/kg), survival (€10.80/%), calving interval (-€7.17/day), Health (- €55.48/unit logSCC & €1.13/standardised locomotion score), Beef (€2.94). Each strain was allocated to one of three feed systems (FS); high milk output per cow from pasture (MP), high concentrate feeding system at pasture (HC), and high milk output per unit area from pasture (HS). The MP system had an overall stocking rate of 2.47 cows/ha, N fertilizer input of 290 kg N/ha and received 325 kg concentrate /cow in early lactation with the remainder of the lactation diet comprised of grazed grass. The HC feed system had a similar overall stocking rate and N input as the MP feed system but 1,445 kg concentrate/cow was fed. The HS group had similar concentrate (327 kg per cow) and N inputs as the MP system but had an overall stocking rate of 2.74 cows/ha. Milk production, live weight and reproductive performance data over the five years used in the economic modeling are shown in Table 2 (Horan et al., 2004, 2005). The milk production data shown has been modified based on differences in reproductive performance observed over the study (Horan et al., 2004) to reflect the expected levels achievable in a stable herd where the strains will differ in maturity. Hence, reduced reproductive performance results in an increased proportion of younger cows in the herd of lower milk yields. As reproductive performance did not differ significantly between feed systems (Horan et al., 2004), in agreement with previously published research (Kennedy et al., 2003), all feed systems were assumed to have the same reproductive performance in this economic analysis.


462

Table 2: The effect of strain of Holstein Friesian on milk production, bodyweight and reproductive performance in three pasture-based feeding systems

Feed system MP HC HS

Strain HP HD NZ HP HD NZ HP HD NZ Number of lactations 65 65 65 65 65 65 65 65 65 Milk Production Milk (kg/cow) 6,748 6,656 6,335 7,724 7,588 6,597 6,531 6,527 6,255 Fat (g/kg) 40.6 40.9 43.9 40.0 40.1 44.5 41.0 41.1 45.6 Protein (g/kg) 34.5 35.6 36.5 35.4 35.8 37.2 34.8 35.5 36.1 Lactose (g/kg) 46.3 46.6 46.7 47.7 47.1 47.5 46.6 46.7 46.6 Average live-weight (kg) 558 590 552 564 594 541 551 580 542 Reproduction* Gestation length (days) 284 284 278 284 284 278 284 284 278 42-day in-calf rate (%) 54 65 74 54 65 74 54 65 74 Overall Pregnancy rate (%) 74 86 93 74 86 93 74 86 93 Total services per cow 2.07 1.79 1.61 2.07 1.79 1.61 2.07 1.79 1.61

*Breeding was initiated at on average 60 days in milk.

Economic Analysis

The Moorepark Dairy Systems Model (MDSM) (Shalloo et al., 2004), a stochastic budgetary simulation model was used to simulate a model farm integrating biological data for each strain in each feed system. The model integrates animal inventory and valuation, milk production, feed requirement, land and labour utilisation and economic analysis. The assumptions used in the model are outlined in Table 3 below. Land area was treated as an opportunity cost with additional land rented in when required and leased out when not required for on-farm feeding of animals. Variable costs (fertiliser, contractor charges, medical and veterinarian, artificial insemination, silage and reseeding), fixed costs (machinery maintenance and running costs, farm maintenance, car, telephone, electricity and insurance) and prices (calf, milk and cow) were based on current prices (Teagasc, 2004). A differential was placed between the strains in terms of male calf and cull cow value based on the variation in strain bodyweight. Three economic scenarios were investigated. In Scenario 1 (S1), it was assumed that farmers were constrained by the EU milk quota, i.e. quota applied at farm level. Farmers with cows producing greater yields would reduce cow numbers to exactly meet quota (evaluation based on a fixed output). Surplus land was leased out. In Scenario 2 (S2), it was assumed that EU milk quota applied at an industry level thereby allowing farms with high producing cows to maintain cow numbers and lease the additional quota required. Where quota leasing was an option the lease cost was taken at 4.79 c/kg of milk. In Scenario 3 (S3) it was assumed that farmers were constrained by land area but leasing milk quota was possible, therefore output could be increased through increased feed input.


463

Table 3: Assumptions used in the model farm for the years 2005 and 2013

Strain of Holstein Friesian HP HD NZ

Farm size (ha) 40.0 40.0 40.0 Quota (kg) 468,000 468,000 468,000 Reference fat (g/kg) 36.0 36.0 36.0 Price protein to fat 2 2 2 Quota lease price (c/kg) 4.8 4.8 4.8 Replacement Heifer price (€) 1,397 1,397 1,397 Labour costs (€/month) 1,905 1,905 1,905 Prices and Costs: 2013 Gross milk price (c/kg) 22.3 22.3 22.3 Reference cull cow price (€) 270 270 257 Reference male calf price (€) 102 102 64 Concentrate costs (€/tonne)

189 189 189

Opportunity cost of land (€/ha) 267 267 267

Results The benefits of the economic appraisal within a farmlet study such as this, is to quantify the effect of genetic change within a controlled management environment where observed differences can be attributed to genetics, feeding system or a combination of these factors. Current Milk Production Environment (S1 scenario) Table 4 shows the key herd output parameters from the model for the three strains in the MP, HC and HS feed systems in scenario 1. In this scenario all groups are restricted to a butterfat-corrected fixed quota of 468,000 kg thereby not requiring the entire 40ha of land for production. Within each FS the highest farm profit was realised with the NZ strain with the farm profit of the HP strain lowest and the HD strain intermediate. Within the HP strain, the highest profit was realized in the HC FS (€17,295), with the lowest profit within the HS FS (€14,232). Regarding the HD and NZ strains, maximum farm profit was realized in the MP FS (€24,925 and €27,869, respectively), the lowest farm profit realized in the HC FS (€21,385 and €21,712, respectively) while the HS FS was intermediate (€23,916 and €27,620, respectively). The increase in farm profit for the HP strain in going from the MP to the HC FS was associated with a large milk production response to increased supplementation in the HC FS thereby requiring 5.6 fewer cows calving to fill the quota, 5.9 fewer hectares of land, and also resulting in a reduction in both labour and replacement costs. In contrast the reduction in farm profit for the NZ and HD strains in going from the MP to the HC FS (€6,157 and €3,540, respectively) was a consequence of a smaller reduction in cow numbers and land requirements because of lesser milk production responses to concentrate and a reduction in milk returns associated with a disproportionate increase of fat to protein in the HC FS for the NZ strain resulting in a marginal reduction in profitability in the HC FS. Unlike the NZ strain which undergoes little change in farm profit going from the MP to the HS FS, the reduction in farm profit for the HD and HP strains was associated with an increase in the number of cows calving, replacement costs, and labour costs.


464

Table 4: Key herd parameters in a fixed quota scenario (S1) using anticipated future costs and prices for three strains of Holstein Friesian cows; High Production (HP), High Durability (HD) and New Zealand (NZ), within the Moorepark (MP), High Concentrate (HC) and High Stocking Rate (HS) feed systems.

Feed System MP HC HS Strain of Holstein

Friesian NZ HD HP NZ HD HP NZ HD HP

Milk price (c/kg) 26.2 24.7 24.5 26.7 24.4 24.5 26.3 24.7 24.5 Farm size Total hectares used (ha)

32.0 32.5 32.2 25.5 27.0 26.3 28.4 30.7 30.9

Quota lease (kg) - - - - - - - - - # Cows calving (no.)

65.8 65.1 64.5 61.1 60.8 58.9 64.9 68.3 68.8

Livestock units (LU)

74.1 73.1 71.8 68.9 68.3 65.6 73.2 76.7 76.6

Stocking rate (LU/ha)

2.32 2.25 2.23 2.70 2.53 2.49 2.58 2.50 2.48

Labour units (h) 2,311 2,304 2,296 2,254 2,250 2,227 2,301 2,343 2,350 Milk produced (kg) 419,9

40 440,7

07 438,3

63 402,5

45 450,8

13 450,4

26 406,6

92 444,4

20 445,4

00 Milk sales (kg) 407,8

44 428,7

25 426,5

00 391,3

04 439,6

29 439,5

95 394,7

47 431,8

51 432,7

45 Fat sales (kg) 17,59

5 17,43

3 17,45

5 17,65

0 17,30

7 17,30

8 17,64

4 17,40

0 17,39

0 Protein sales (kg) 14,76

6 14,97

4 14,67

6 14,16

6 15,36

7 15,47

8 14,05

4 15,12

8 14,99

4 Milk returns (€) 106,7

10 106,1

01 104,6

29 104,5

38 107,2

69 107,8

95 103,6

72 106,6

74 105,8

37 Livestock sales (€) 26,89

7 29,01

1 30,59

9 24,99

7 27,07

7 27,94

0 26,56

2 30,43

0 32,64

2 Total costs (€) 105,7

15 110,1

63 118,7

89 107,8

01 112,9

36 118,5

16 102,5

92 113,1

65 124,2

24 Margin per cow (€) 424 383 255 355 352 294 425 350 207 Margin per kg milk (cents)

6.64 5.66 3.74 5.39 4.74 3.84 6.79 5.38 3.20

Feed costs per kg milk (cents)

4.60 4.50 4.50 7.00 6.40 6.20 4.50 4.40 4.50

Replacement costs (€)

15,555

19,747

28,912

14,456

18,431

26,400

15,361

20,713

30,843

Labour costs (€) 31,844

31,473

31,005

29,594

29,375

28,310

31,446

33,014

33,075

Total profit per farm (€)

27,869

24,925

16,416

21,712

21,385

17,295

27,620

23,916

14,232


465

Quota Leasing Environment (S2 scenario)

The key herd output parameters from the model for the three strains in the three feed systems, within a quota leasing environment (S2) are shown in Table 5. In this scenario an equal number of cows (89.8) were calved for each strain within each feed system. The NZ strain again achieved the highest farm profit in all systems of production, with the HP strain lowest and the HD strain intermediate. The highest profit for the HP strain (€18,846) was achieved in the HC FS as in S1, with the lowest again realised in the HS FS (€14,291). Within this scenario the HD strain again achieved their greatest farm profit in the MP FS and suffered largest reductions in profit when moving to the HC FS. The HP strain increased margin per cow going from the MP to the HC FS unlike both other strains. Similar to the S1 scenario, the greatest farm profit for the NZ strain was realised in both the MP and HS FS. The HP and HD strains encountered reductions of €3,177 and €2,085 in farm profit in going from the MP to the HS FS. Relative to S1, the profitability in all cases in this scenario has increased. Limited Land within Quota Leasing Environment (S3 scenario) Table 6 shows the key herd output parameters from the model for the three strains in the MP, HC and HS feed systems in a fixed land area scenario. In this scenario an equal land base (40.0ha) was available to each farm. The highest farm profit was achieved by the NZ strain in the HS FS (€33,947) with 91.6 cows calving or 9.4 more than in the MP FS. This results in an overall stocking rate of 2.58LU/ha and an increase in margin per hectare of €58 and €169 compared with the MP and HC FS for this strain. As in scenarios 1 and 2 the HP strain achieved the lowest profit, achieving greatest profit in the HC FS (€18,835) or €471/ha. Within this environment (S3) cow numbers and feed costs were greatest for the NZ strain in the HC FS whereas margin per hectare and margin per kg milk were reduced by €111 and 1.72c compared with the MP FS. The HD strain achieved maximum profit in the MP FS (€27,475), achieving a margin of €687/ha with lowest profit for this strain realised in the HC FS as in both other scenarios. This was related to an increase of ten cows calving, a reduction in margin per cow of €68 and an increase in feed costs of 1.9c/kg milk.


466

Table 5: Key herd parameters in a quota leasing scenario (S2) using anticipated future costs and prices for three strains of Holstein Friesian cows; High Production (HP), High Durability (HD) and New Zealand (NZ), within the Moorepark (MP),High Concentrate (HC) and High Stocking Rate (HS) feed systems




43.7 44.8 44.8 37.5 39.8 40.1 39.2 40.3 40.3

Quota lease (kg) 148,972

162,146

167,246

183,546

209,551

230,621

150,997

135,562

131,981

# Cows calving (no.)

89.8 89.8 89.8 89.8 89.8 89.8 89.8 89.8 89.8


101.2 100.8 100.0 101.2 100.8 100.0 101.2 100.8 100.0


2.32 2.25 2.23 2.70 2.53 2.49 2.58 2.50 2.48


30 607,3

85 610,2

60 591,3

64 665,6

94 686,7

30 562,2

58 583,9

27 581,2


16 590,8

71 593,7

46 574,8

50 649,1

80 670,2

16 545,7

43 567,4

13 564,7


2 24,02

6 24,30

0 25,29

8 25,55

7 26,38

8 24,39

2 22,86

2 22,69


0 20,63

7 20,43

1 20,81

1 22,69

2 23,59

8 19,43

0 19,87

6 19,56


03 146,2

45 145,6

74 153,5

90 158,4

16 164,5

17 143,3

44 140,1

75 138,1


2 39,98

3 42,59

8 36,72

2 39,98

3 42,59

8 36,72

2 39,98

3 42,59


13 157,0

59 170,7

57 164,0

02 173,6

02 188,2

15 146,5

17 153,0

96 166,4


5.82 4.79 2.86 4.44 3.72 2.74 5.96 4.63 2.46


4.60 4.50 4.50 7.00 6.40 6.20 4.50 4.40 4.50


21,236

27,215

40,250

21,236

27,215

40,250

21,236

27,215

40,250


43,377

43,163

43,475

43,377

43,163

43,475

43,377

43,163


33,367

29,121

17,468

26,262

24,746

18,846

33,522

27,036

14,291

Influence of Genetic Strain on Farm Profit


467

Table 6: Key herd parameters in a limited land base scenario (S3) using anticipated future costs and prices for three strains of Holstein Friesian cows; High Production (HP), High Durability (HD) and New Zealand (NZ), within the Moorepark (MP),High Concentrate (HC) and High Stocking Rate (HS) feed systems




40.0 40.0 40.0 40.0 40.0 40.0 40.0 40.0 40.0

Quota lease (kg) 102,090

98,512

103,599

221,575

212,081

229,129

161,937

131,075

133,301

# Cows calving (no.)

82.2 80.1 80.2 95.7 90.1 89.6 91.6 89.1 90.0


92.7 89.9 89.3 107.9 101.2 99.8 103.2 100.0 100.2


2.32 2.25 2.23 2.70 2.53 2.49 2.58 2.50 2.48


58 541,9

72 544,8

43 630,4

86 668,2

89 685,2

00 573,5

29 5793

10 582,5


34 527,2

37 530,0

99 612,8

79 651,7

10 668,7

23 556,6

84 562,9

26 566,0


9 21,43

9 21,69

5 27,64

4 25,65

6 26,32

9 24,88

1 22,68

1 22,74


2 18,41

5 18,24

1 22,18

8 22,78

0 23,54

5 19,82

0 19,71

9 19,61


21 130,4

80 130,0

44 163,7

33 159,0

16 164,1

33 146,2

02 139,0

51 138,4


0 35,67

7 38,03

1 39,15

1 40,13

9 42,50

3 37,45

8 39,66

7 42,69


14 138,6

82 151,0

06 175,6

84 174,3

69 187,8

00 149,7

13 151,7

86 166,8


6.03 5.07 3.13 4.31 3.71 2.75 5.92 4.65 2.45


4.60 4.50 4.50 7.00 6.40 6.20 4.50 4.40 4.50


19,448

24,284

35,935

22,641

27,321

40,160

21,662

27,000

40,344


38,705

38,536

46,350

43,546

43,067

44,346

43,034

43,264


31,637

27,475

17,069

27,201

24,785

18,835

33,947

26,931

14,292


468

Discussion

The productivity and subsequent profitability of a dairy cow is determined by its environment (especially feeding) as well as its own inherent capabilities (genetic potential for production and health traits) (Holmes et al., 2002). Both animal and feed factors influencing farm profitability are numerous and differ greatly in significance depending on the economic characteristics of the production environment. For this reason, the findings of studies investigating the economic influence of alternate genetic selection strategies are often contradictory and the extrapolation of results to alternative production environments is erroneous. This study highlights the large influence of genetic strain and production system on farm profitability within Irish pasture-based production systems. Similar to the current study, previous studies (Shalloo et al., 2004a; Evans et al., 2006) have shown significant genetic influences on farm profitability in a variety of pasture-based feeding systems. These results reinforce the significance of reproductive capacity within pasture-based systems (Schmidt, 1989; Plaizier et al., 1997) with high profitability realised with animals combining high genetic potential for both production and fertility traits (HD and NZ strains) rather than with those selected purely for increased milk production potential (HP strain). Reductions in economic performance through reduced fertility arise through; reduced milk yield per cow per day of herd life, increased culling for reproductive reasons, fewer available replacement heifers, increased semen usage, and added costs of veterinarian interventions (Britt, 1985; Plaizier et al., 1997). Esslemont and Peeler (1993) reported desired annual total culling rates of 18% to maximise the benefit of age and genetic improvement while Esslemont et al. (2001) reported optimal financial performance to arise with a 365- to 370-d calving interval and a failure to conceive culling rate of about 7%.

Simm, (2000) postulated that the optimum method of selection on a number of traits was to use a selection index which places a weighted emphasis on traits based on their economic importance. In 2001 in Ireland, the Economic Breeding Index (EBI; Veerkamp et al., 2002), a profit based index, selecting dairy cows for the predominantly grass based, seasonal calving systems of milk production was developed to increase the profitability in dairy herds through genetic selection using the precepts of selection index theory (Hazel, 1943). Kahi et al. (1998) stated that the most profitable genotype was that which gives the highest profit per unit of the most limiting input. Within an Irish context, currently quota is the limiting factor (as reflected by the S1 scenario) whereas expected changes in the agricultural policy environment are likely to result in the S2 or S3 scenarios prevailing in future years. These results and those of Veerkamp et al. (2002) demonstrate that increased farm profitability for Irish dairy farmers, within probable future economic climates, can only be realised where productivity gains are achieved without detrimental impact on health and welfare traits. Selection index theory, as outlined by Hazel, (1943) is based on the premise that an animal may be poor on one trait and still achieve a high genetic evaluation by compensating based on their superiority on other traits within the index. In this study, the highest EBI (EBI = €80) NZ strain returned the highest profit, the lowest EBI (EBI = €52) HP strain returned the lowest profit whereas the HD strain were intermediate on EBI (€57) and farm profit in all scenarios investigated thereby validating EBI as an accurate genetic selection tool to predict the potential profitability of pasture-based dairy cows in the Irish economic climate. It can also be concluded from this result that the overall level of genetic potential of the HD and NZ strains for milk production and health traits (as measured by EBI) rather than their geographic origin is responsible for the profitability differential based on the similarity between observed and predicted economic performance. In contrast, the large differential in overall profitability between the HP and HD strains in these results is unexpected given the relatively small differential in genetic potential (EBI of €52 and €57, respectively). While the greater replacement costs incurred by the HP strain were expected, it was anticipated that their superior milk production potential (milk sub-index = €46) would compensate and therefore deliver a similar overall profitability to the HD strain. The increased


469

productivity of the HP strain was not realised in this study as impaired reproductive performance reduced milk productivity similar to the observations of Britt, 1985; Garcia and Holmes, 1999 and Stevens et al., 2000. This result implied that for Irish seasonal pasture-based systems, the economic significance of fertility traits is underestimated and must be considerably increased to reflect the significant influence of fertility on farm profitability. Alternatively, the milk production potential of sires with inferior genetic potential should be revised downward to provide a more accurate estimate of their true production potential within a seasonal production system. Influence of Feed System on Farm Profit Previous research has shown that increased concentrate supplementation at pasture does not influence the reproductive performance of animals when adequate amounts of high quality pasture are provided (Horan et al., 2004; Kennedy et al., 2003). Similarly, McCarthy et al. (2006a) reported no significant effect of feed system on udder health while Roche et al. (2006) found no effect of feed system on the rate of body condition score loss in early lactation. Consequently, where adequate nutrients are supplied in the basal diet, supplementary concentrate feeding can only influence overall farm profitability through its influence on animal production performance. The data collected here suggests that the revenue gains associated with genetic improvement considerably overshadow any influence of feeding system on farm profitability. The optimum system of milk production depends greatly on the prevailing economic environment (milk price, feed costs etc.) as well as the relative availability of the key factors of production (land, milk quota etc.). Within a milk quota scenario (S1), profit is maximised where production is achieved at minimum cost as demonstrated by the comparably greater profitability of the low concentrate (MP and HS) systems. This is similar to findings by Harris and Freeman (1993) where it was shown using a linear programming model that economic weight for herd life substantially increased in the restrictive quota situation. The limitation on output results in more emphasis being put on efficiency for each litre of milk produced. In a low milk price situation, pasture based systems are also more favourable, through their capability for low-cost milk production with the achievement of high output per hectare (Penno et al., 1996). Within an environment where milk quota is not a limiting factor (S2 and S3), land availability becomes the next limitation to the pasture-based systems under consideration. Similar to previous studies (Penno et al., 1996), this analysis shows that based on the anticipated reduction in milk price in future years, higher stocking rates (HS) systems will be most profitable. Such systems will be characterised by their capability for low-cost high milk productivity per hectare with lesser milk production per cow. Similar to previous studies (Lopes-Villalobos et al., 2000; Grainger and Goddard, 2004), the data show that under a scenario where land is limited and stocking rates increase (S3), the economic advantage of the smaller NZ strain will be increased due to the comparably lesser reduction in margin per cow in the HS system. Strain of Holstein Friesian by feed system interactions have been reported within regard to milk production (Horan et al., 2005) and DM intake (Horan et al., 2006). This suggests that the type of cow used may differ depending on the system of production. The results of this study show that the highest farm profit observed on the study was with animals of lesser milk production with good fertility on a MP or HS FS in all scenarios, whereas the highest farm profit for a HP strain animal was the HC system in all scenarios. While the HC system did improve the profitability of the HP animal, the increased profitability was still inferior to that of genetically superior animals across all systems of production.


470

Influence of Agricultural Policy change on Farm Profit

Comparisons of genetic groups or feeding systems must be made on the basis of that which gives the highest profit per unit of the most limiting input (Kahi et al., 1998). The economic principles applying to a no quota environment are substantially different to those that apply within a quota environment (Shalloo et al., 2004a). Where quota is not limiting, output from the farm is maximised through increasing milk sales until marginal revenue from additional milk sales is equal to the marginal cost of the additional milk. The system of milk production operated is therefore governed by the concentrate to milk price ratio (Clark and Kanneganti, 1999) and the milk production response to the concentrate supplementation. Where the milk price is high, systems adopted will maximise realised profitability through increased concentrate supplementation (Soder and Rotz, 2001). The Common Agricultural Policy is currently undergoing significant change with the most recent reform, the Luxembourg agreement anticipated to result in a reduction in milk price of 5c/L (from 27 to 22c/L) (Binfield et al., 2003) for EU milk producers with further reductions also likely. It is evident from this analysis that both within the current quota system and based on projected changes to their production environment, the future viability of Irish dairy farmers depends on the realisation of maximum efficiency in pasture based milk production systems, through the further development of low cost pasture-based production systems (similar to the MP system) focused on increased productivity. The aim within such a feed system must be to maximise the proportion of grazed grass in the diet, increase utilisation and maintain high intakes (Horan et al., 2006) throughout the grazing season. Complementary genetic selection must therefore deliver animals capable of high productivity from pasture. Based on the current analysis, it is apparent that these animals will be characterised by both high milk production and reproductive potential. Conclusions The purpose of this paper is to demonstrate the magnitude of variation in profitability between strains of Holstein-Friesian dairy cows, differing in genetic potential for milk production and reproductive performance, across different pasture-based production systems and within various production scenarios, and not to recommend any given existing strain of Holstein-Friesian for use in Irish pasture-based systems. Large variation in farm profit arises from various genetic selection strategies and production system choices with the optimum genetics and production system being the combination which results in the greatest farm profit within that production environment. This study demonstrates how genetic selection for increased milk production (HP strain) in conjunction with increased concentrate supplementation within Irish pasture-based systems will result in reduced profitability in future years relative to selection on a combination of production and reproductive traits (HD and NZ strains) within a greater reliance on high quality grazed pasture. These results validate the use of EBI as a valuable genetic selection tool but suggest that the weighting on fertility traits needs to be increased within the index to reflect the true value of fertility to farm profitability.

Acknowledgments This study is part of a joint project between Dexcel (New Zealand), Massey University (New Zealand), and Teagasc (Moorepark). We would like to acknowledge the support of Professor Colin Holmes (Massey University). We thank the staff of Curtins farm for their co-operation, care and management of the experimental animals.


471

References Binfield J., T. Donnellan, K. Hanahran, and P. Westhoff. 2003. The Luxembourg CAP reform agreement:

Implications for EU and Irish Agriculture. The Luxembourg CAP Reform Agreement: Analysis of the Impact on EU and Irish Agriculture, Teagasc, Rural Economy Research Centre, Dublin. Page 1-69.

Britt, J.H. 1985. Enhanced reproduction and economic implications. J. Dairy Sci. 68: 1585-1592. Clark, D. A., and V. R. Kanneganti. 1999. Grazing management systems for dairy cattle in Grass for dairy

cattle (eds J. H Cherney and D. J. R. Cherney)

Dillon, P., J.R. Roche, L. Shalloo, and B. Horan. 2005. Optimising financial return from grazing in temperate pastures. Utilisation of grazed grass in temporal animal systems. Page 131-147.

Esslemont, R.J., M.A. Kossaibati, and J. Allock. 2001. Economics of fertility in dairy cows. In: Diskin,

M.G. (Ed), Fertility in the High Producing Cows, British Soc. Of Anim. Sci. Occassional Publication No. 26, pp. 19-29.

Esslemont, R.J. and E.J. Peeler. 1993. The scope for raising margins in dairy herds by improving fertility

and health. Br. Vet. J. 149:537-547. Evans, R. D., M. Wallace, L. Shalloo, D.J. Garrick, and P. Dillon. 2006. Financial implications of recent

declines in reproduction and survival of Holstein-Friesian cows in spring-calving Irish dairy herds. Agricultural Systems 89:165-183.

Evans, R. D., P. Dillon, F. Buckley, M. Wallace, V. Ducrocq, and D.J. Garrick. 2004. Trends in milk

production, fertility and survival of Irish dairy cows as a result of the introgression of Holstein-Friesian genes. In: Proceedings of the Agricultural Research Forum, Tullamore, Ireland, p52 (abstr.).

Garcia, S.C., and C.W. Holmes. 1999. Effect of time of calving on the productivity of pasture based dairy

systems: a review. New Zealand society of agricultural research, 42:347-362. Grainger, C. and M. E. Goddard. 2004. A review of the effects of dairy breed on feed conversion

efficiency - an opportunity lost?. In: Animal Production in Australia, Proceedings of the 25th Biennial Conference of the Australian Society of Animal Production, University of Melbourne, Victoria 25: 77- 80.

Harris, B.L., and A.E. Freeman. 1993. Economic weights for milk yield traits and herd life under various

economic conditions and production quotas. J. Dairy Sci. 76:868-879. Hazel, L.N. 1943. The genetic basis for constructing selection indexes. Genetics 28:476. Hennessey, T., and F. Thorne. 2006. The impact of WTO Doha development round on farming in Ireland.

Teagasc Rural Economy Research Centre. Holmes, C.W., I.M. Brookes, D.J. Garrick, D.D.S. Mackenzie, T.J. Parkinson, and G.F. Wilson. 2002.

Milk production from pasture. Principles and practices. Massey-University, Palmerston North, New Zealand.


472

Horan, B., P. Faverdin, L. Delaby, M. Rath, and P. Dillon. 2006. The effect of strain of holstein-Friesian dairy cow on grass intake and milk production in various pasture-based systems. Anim. Sci: in press.

Horan, B., P. Dillon, P. Faverdin, L. Delaby, F. Buckley, and M. Rath. 2005. The interaction of strain of

Holstein Friesian cows and pasture based feed systems on milk yield, body weight and body condition score. J. Dairy Sci. 88:1231-1243.

Horan, B., J. F.Mee, M. Rath, P. O’Connor, and P. Dillon. 2004. The effect of strain of Holstein-Friesian

cow and feed system on reproductive performance in seasonal-calving milk production systems. Anim Sci. 79: 453-467.

Kahi, A.K., I.S. Kosgey, V.L. Cardoso, and J.A.M. Van Arendonk. 1998. Influence of production

circumstances and economic evaluation criteria on economic comparison of breeds and breed crosses. J. Dairy Sci. 81:2271-2279.

Kennedy, J., P. Dillon, K.O’ Sullivan, F. Buckley, and M. Rath. 2003. The effect of genetic merit and

concentrate feeding level on reproductive performance of Holstein-Friesian dairy cows in a grass based milk production system. Anim. Sci. 76: 297-308.

Lopez-Villalobos, N., D.J.Garrick, C.W. Holmes, H.T. Blair, and R.J. Spelman. 2000. Profitabilities of

some mating systems for dairy herds in New Zealand. J. Dairy Sci. 83:144-153. McCarthy, S., D.P. Berry, P. Dillon, M. Rath, and B. Horan. 2006a. Effect of strain of Holstein-Friesian

and feed system on udder health and milking characteristics. Livest. Sci. (In Press) Penno J. W., K.A. MacDonald, and A. M. Bryant. 1996. The economics of the No. 2 Dairy systems. In

Proceedings of the Ruakara Farmers Conference 48: Page 11-19. Plaizier, J.C.B., G.J.King, J.C.M. Dekkers, and K. Lissemore. 1997. Estimation of economic values of

indices for reproductive performance in dairy herds using computer simulation. J. Dairy Sci. 80:2775-2783.

Pryce, J. E., and R.F. Veerkamp. 2001. The incorporation of fertility indices in genetic improvement

programmes. BSAS occasional publication No. 26: 237-249. Rauw, W.M., E. Kanis, E.N. Noordhuizen-Stassen, and F.J. Grommers. 1998. Undesirable side effects of

selection for high production efficiency in farm animals: a review. Livest. Prod. Sci. 56:15-33. Roche, J.R., D.P. Berry, and E.S. Kolver. 2006. Holstein-Friesian strain and feed effects on milk

production, bodyweight and body condition score profiles in grazing dairy cows. J. Dairy Sci. 89: 3532-3543.

Schmidt, G.H. 1989. Effect of length of calving interval on income over feed and variable costs. J. Dairy

Sci. 72:1605-1611.

Shalloo L., P. Dillon, M. Rath, and E. M. Wallace. 2004. Description and validation of the Moorepark Dairy Systems Model. J. Dairy Sci. 87:1945-1959.


473

Shalloo, L., J. Kennedy, M. Wallace, M. Rath, and P. Dillon. 2004a. The economic impact of cow genetic potential for milk production and concentrate supplementation level on the profitability of pasture based systems under different EU milk quota scenarios. J. Ag. Sci. 142:357-369.

Simm, G. 2000. Genetic improvement of cattle and sheep. Miller Freeman UK Ltd. Soder, K.J., and C.A. Rotz. 2001. Economic and environmental impact of four levels of concentrate

supplementation in grazing dairy herds. J. Dairy Sci. 84:2560-2572. Stevens, J., L. Burton, and J. Rendel. 2000. Induced calving. In: Australian and New Zealand Dairy

Veterinarians Conference, May 2000. Editor T. Parkinson, Veterinary Continuing Education, Massey University. Page 63 – 78.

Teagasc 2004. Management data for farm planning, Teagasc, Dublin, Ireland. Veerkamp, R.F., P. Dillon, E. Kelly, A.R. Cromie, and A.F. Groen. 2002. Dairy cattle breeding objectives

combining yield, survival and calving interval for pasture-based systems in Ireland under different milk quota scenarios. Livest. Prod. Sci. 76:137-151.