15
M’orld Derelopmenr, Vol. 18, X:0. 7. pp. IOlj_1029, 1990. 030~750x/90S3.00+0.00 Printed in Great Britain. 0 1990 Pergamon Press plc Collaborative Sector Analysis: A Foreign Assistance Technique for Improving LDC Sector Management BRANDON ROBINSON* United States-Saudi Arabian Joint Commission on Economic Cooperation, Riyadh Summary. - Poor sector management holds back social and economic progress in many developing countries. Collaborative sector analysis is a technique for improving sector management. It provides findings of immediate utility for policy making, and strengthens less developed countries’capabilities in data collection, processing, analysis, and the formulation of policy. To be useful, the analysis should be limited to a few selected objectives, with aggregate and disaggregate measurements providing a baseline from which to monitor progress. The possible complementarities and the inescapable tradeoffs among objectives can be estimated periodically, and used to design sector strate$es that are technically, financially, politically and administratively feasible. If the sector analysis and design process is given a central, coordinating role, the less developed country will gradually reduce its intellectual and financial dependence on donors. 1. INTRODUCTION Improvements in the social and economic poli- cies of developing countries are not easily intro- duced and directed by foreign assistance agencies or the international lending institutions. Time and again “conditionality” turns out to be less effective than originally hoped and supposed. Loan conditions imposed by a multilateral or bilateral organization almost always carry the stigma of an infringement of national sovereignty. Besides, the “policy reforms” that are demanded by lenders are depressingly often based on an in- adequate diagnosis of problems - particularly at the decision-making level of sectors. Agriculture is an example. Donor recommen- dations for changes in a less developed country’s (LDC) policies seldom lead to a coherent and effective national agricultural sector strategy. The recommended levels of public expenditures fre- quently turn out to be excessive. Even when these costs can be met, an agricultural sector strategy that has been designed by donors may prove impractical because of unforeseen political oppo- sition or administrative obstacles. Donors very seldom possess the range of infor- mation needed to formulate a sector strategy that takes the LDC’s political, social, administrative and economic problems into proper account. This is not surprising. The main function of foreign assistance agencies is not to inquire, but to transfer resources. These organizations are usually eager to obligate funds and reluctant to accept delays by acknowledging information gaps. Each donor is strongly tempted, therefore, to find a standard way to behave, and to promote the same basic sector strategy in every developing country. The more recent foreign assistance strategies have usually been designed to bring about reductions in government activities and expen- ditures. This is a sound objective. IMany LDC governments are producing goods and providing services that could be produced and provided more efficiently by the private sector. Many LDCs are excessively centralized, many have bloated bureaucracies, and many LDC ministries over- regulate and have excessive workloads. The reduction of bureaucracy, government work- load, and public regulation cannot be effectively carried out under a single, standard approach *The support of John A. Patterson of USAID and Robert Bartram of the US Bureau of Census were essential for the design and execution of the collaborative sector analysis projects in Rwanda, Zambia, and El Salvador. The tech- nical expertise of Daniel Clay and Alan Saalfeld of the Bureau of Census made it possible to successfully com- plete the projects in El Salvador and Rwanda. Anne Gordon Drabek made valuable editorial suggestions. The views expressed here are not necessarily the views of USAID or the Bureau of Census. 1015

Collaborative sector analysis: A foreign assistance technique for improving LDC sector management

Embed Size (px)

Citation preview

M’orld Derelopmenr, Vol. 18, X:0. 7. pp. IOlj_1029, 1990. 030~750x/90S3.00+0.00 Printed in Great Britain. 0 1990 Pergamon Press plc

Collaborative Sector Analysis:

A Foreign Assistance Technique for Improving

LDC Sector Management

BRANDON ROBINSON* United States-Saudi Arabian Joint Commission on Economic Cooperation, Riyadh

Summary. - Poor sector management holds back social and economic progress in many developing countries. Collaborative sector analysis is a technique for improving sector management. It provides findings of immediate utility for policy making, and strengthens less developed countries’capabilities in data collection, processing, analysis, and the formulation of policy. To be useful, the analysis should be limited to a few selected objectives, with aggregate and disaggregate measurements providing a baseline from which to monitor progress. The possible complementarities and the inescapable tradeoffs among objectives can be estimated periodically, and used to design sector strate$es that are technically, financially, politically and administratively feasible. If the sector analysis and design process is given a central, coordinating role, the less developed country will gradually reduce its intellectual and financial dependence on donors.

1. INTRODUCTION

Improvements in the social and economic poli- cies of developing countries are not easily intro- duced and directed by foreign assistance agencies or the international lending institutions. Time and again “conditionality” turns out to be less effective than originally hoped and supposed.

Loan conditions imposed by a multilateral or bilateral organization almost always carry the stigma of an infringement of national sovereignty. Besides, the “policy reforms” that are demanded by lenders are depressingly often based on an in- adequate diagnosis of problems - particularly at the decision-making level of sectors.

Agriculture is an example. Donor recommen- dations for changes in a less developed country’s (LDC) policies seldom lead to a coherent and effective national agricultural sector strategy. The recommended levels of public expenditures fre- quently turn out to be excessive. Even when these costs can be met, an agricultural sector strategy that has been designed by donors may prove impractical because of unforeseen political oppo- sition or administrative obstacles.

Donors very seldom possess the range of infor- mation needed to formulate a sector strategy that takes the LDC’s political, social, administrative and economic problems into proper account. This is not surprising. The main function of foreign assistance agencies is not to inquire, but to transfer

resources. These organizations are usually eager to obligate funds and reluctant to accept delays by acknowledging information gaps. Each donor is strongly tempted, therefore, to find a standard way to behave, and to promote the same basic sector strategy in every developing country.

The more recent foreign assistance strategies have usually been designed to bring about reductions in government activities and expen- ditures. This is a sound objective. IMany LDC governments are producing goods and providing services that could be produced and provided more efficiently by the private sector. Many LDCs are excessively centralized, many have bloated bureaucracies, and many LDC ministries over- regulate and have excessive workloads.

The reduction of bureaucracy, government work- load, and public regulation cannot be effectively carried out under a single, standard approach

*The support of John A. Patterson of USAID and Robert Bartram of the US Bureau of Census were essential for the design and execution of the collaborative sector analysis projects in Rwanda, Zambia, and El Salvador. The tech- nical expertise of Daniel Clay and Alan Saalfeld of the Bureau of Census made it possible to successfully com- plete the projects in El Salvador and Rwanda. Anne Gordon Drabek made valuable editorial suggestions. The views expressed here are not necessarily the views of USAID or the Bureau of Census.

1015

1016 WORLD DEVELOPMENT

throughout the entire Third World. The specific policies that contribute most to inefficiency, the regulations that function most actively as dis- incentives to entrepreneurship, and the measures that are most needed for the removal of the obstacles vary greatly from country to country. Consequently, a country inquiry into the relevant socioeconomic, political and administrative con- ditions should lay the groundwork for the design of corrective measures. Each LDC needs to know what kinds of obstacles and constraints prevail in that country, and what kinds of policy changes are therefore needed. Each LDC needs to know what kinds of functions and activities need to be elim- inated or reduced in its ministries.

The multiplicity of donors and the plethora of foreign assistance activities often contribute to confusion concerning sector priorities and needed changes in policy. This is particularly true of African countries. The sector policy or strategy of an African country can usually be described as the collection of projects designed by foreign assist- ance agencies (FAAs). Generally speaking, the FAAs have done very little to coordinate their activities. The recurring costs of all these projects very often exceed the LDC’s current and future financial availabilities. The more able and better- trained LDC officials are pulled away from analy- sis and planning to deal with demanding foreign assistance agencies that are usually in a hurry to get their new projects approved and the pipelines of their old projects reduced. For example, Kenya’s foreign ajsistance projects increased in number from 506 m December 1981 to 689 in December 1983. In Zambia in 1981 there were 108 projects in the agricultural sector alone.

In many developing countries the large foreign assistance tail is so independent and so vigorous, it appears to be wagging the development dog. Instead of contributing to the development of a coherent and effective sector strategy, foreign assistance frequently has the effect of increasing sectoral incoherence. Besides, foreign assistance has usually failed to strengthen LDC capabilities for carrying out analyses on a periodic basis. These are the analyses that would help the LDC identify the constraints to, and the opportunities for, achieving its macroeconomic and sectoral goals.

The sector assessments or sector reports that are carried out by the foreign assistance agencies are often declared to be an alternative to the project- by-project approach. However, their main purpose is to identify foreign assistance projects. As sector strategies, or as a fund of information for the design of such strategies, these assessments are characterized by four basic flaws.

First, the short duration (three weeks to three months) of these studies precludes the collection

of new data and the consequent elimination of the critical data gaps. Second, these studies tend to be carried out by a group of specialists, each one of whom is expected to make action recommen- dations ‘in his or her area of expertise. Insuffi- cient consideration is given to the interactions across subsectors, to connections between the sector and the rest of the economy, and to the comparative importance or priority rankings of subsector problems and corrective measures. As a consequence, the sector program or strategy that arises from a sector assessment tends to include activities in each area for which the sector assess- ment team has a representative. The sector strategy is more likely to be a successful turf and budgetary compromise among the specialities than an effective solution of the more basic sectoral problems.

There is a third basic flaw in these sector assess- ments. These studies tend to concentrate on the action front (especially on the interventions that will be carried out by the foreign assistance agency), and to ignore the information front. They thus fail to contribute to an improvement of the LDC’s data base, and to its capabilities for analysis and the formulation of policy. Fourth, the sector assessments are not collaborative, institution- building endeavors. LDCs are usually reluctant to adopt policy recommendations emerging from studies that are carried out unilaterally by a foreign assistance agency.

2. COLLABORATIVE SECTOR ANALYSIS: PURPOSES AND CHARACTERISTICS

A collaborative sector analysis (CSA) has three purposes: (I) to provide findings of immediate utility for policy; (2) to strengthen LDC capa- bilities in data collection, processing, analysis, the formulation and implementation of policy; and (3) to help establish an LDC sector analysis process that will continue to function after termina- tion of foreign assistance. A CSA has both short- range and long-range objectives - to improve the LDC’s present performance, and its capabilities for future performance.

A successful collaborative sector analysis will be tailored to the LDC’s institutional interests and capabilities. It will involve the collection of data not previously obtained, the processing and ana- lytical interpretation of the data, and the use of the findings and conclusions for the formulation of a more coherent and effective sector strategy. The duration of a CSA will seldom be less than five years, and will in many cases be longer. For example, the Zambian Agriculture Training, Plan- ning and Institutional Development (ZATPID)

COLLABORATIVE SECTOR ANALYSIS 1017

project and the Rwandan Agricultural Survey and Analysis (RASA) project were both five-year col- laborative sector analyses that were extended to IO-year projects by the LDCs and the US Agency for International Development (USAID).

Institution building and technology transfer are two characteristics of a CSA. The project usually includes both on-the-job training, provided by resi- dent advisors and short-term consultants, and for- mal training in a university. Although the training and technical assistance expertise provided by the foreign assistance agency varies from sector to sector, all CSAs are concerned with improving LDC performance in data collection, data pro- cessing, the analytical interpretation of data, and the formulation of policy.

The role played by foreign advisors at the plan- ning or policy-making stage will vary among coun- tries. Some LDC institutions may seek advice at all stages of policy making; other LDC institutions may draw the line at the stage of identification of policy options or alternative strategies that are technically, financially and administratively feasible. Although up to now LDC institutions have welcomed the advice of the resident foreign advisors at the strategy design stage, there may be some institutions that will prefer to limit this technical assistance to the earlier stage of data collection, processing and analysis.

The particular ministry’s informational and analytical processes are improved in order to bring about better policy making and day-to-day management. Consequently, the collaborative design of a CSA should specify procedures for communicating findings, explaining analytical conclusions, and discussing recommendations for action. Ministry officials who are not dedicated on a full-time basis to the sector analysis process should participate in these discussions. The project agree- ment should call for periodic exchanges among analysts, policy makers, and managers. This matter is more fully considered as part of the last of the I I methodological guidelines.

3. ELEVEN METHODOLOGICAL GUIDELINES FOR COLLABORATIVE

SECTOR ANALYSIS AND DESIGN

There is no generally accepted set of rules for carrying out a collaborative sector analysis. Many of the concepts and disciplinary techniques depend on the particular sector and on the problems and issues that are specific to it. Nevertheless, there are some guiding principles that emerge from the USAID experience. These are specified below.

(a) Fulfill- the FAA responsibility for the adoption of a sound sector analysis methodology

Although a university, consulting firm, or another government agency, such as the US Bureau of Census or the US Department of Agri- culture, may provide the technical assistance needed to carry out the sector analysis, the foreign assist- ance agency that contracts for this work should take an interest in the methodology, assess the various proposals, and concur with the main fea- tures of the methodology that is finally approved. The FAA should make sure that the methodology employed is sound. A defective sector analysis serves to discredit the sector analysis function, the training that has been provided, and the new data collection, processing and analytical procedures that have been introduced.

(b) Specifv a limited set of socioeconomic objectives

Any inquiry needs to be structured in terms of a limited set of objectives. In the case of a sector analysis, these objectives should be sufficiently aggregate to qualify as ends for management of the sector. Increasing food production and con- sumption, farm family income, and agricultural exports qualify as sector-level objectives, and are therefore appropriate structural elements for an agricultural sector analysis.

An education sector analysis should also be structured by a limited number of sector-level goals. The most common are: (I) increase access to education and training; (2) increase the internal efficiency of education and training activities; and (3) improve the relevance or so-called external efficiency of the education system so that it will make a greater contribution to development.

As indicated by these examples, the goals or objectives of a sector analysis should not be quanti- fied. A purpose of the analysis is to estimate actual and future resources, identify tradeoffs and com- plementarities, and arrive thereby at attainable, quantitative targets. Resources are invariably scarce. The purpose of the analysis is to determine as precisely as possible the amount of goal attain- ment that is feasible, as well as desirable.

Quantitative targets are often set prematurely and, as a consequence, unrealistically. This com- mon occurrence is mainly due to the failure to adhere to the principle of logically prior analysis.

(c) Recognize the analytical chalenge posed by mukiple problem-objectives

It may be instructive to compare sector-level management with firm-level management in order

1019 WORLD DEVELOPblENT

to highlight a basic ditference and a basic similarity. Firm management usually involves the pursuit ofa single objective-profit maximization. As indicated in the examples of agriculture and education, sector management involves the pursuit of multiple objectives.

The single objective pursued by the managers of the firm, and the multiple objectives pursued by the managers of the sector, are problematic objectives. This is the shared characteristic. In both cases, problems need to be solved and difficulties over- come in order to attain objectives. Indeed, increas- ing the firm’s profits and increasing national food production by specified amounts within specified periods can be correctly referred to as “problems.” Since we are talking about changes that constitute problems as well as objectives, it is useful to employ the term “problem-objectives.”

One of the main functions of the analysis is to uncover both the actual and the potential ways in which the problem-objectives are interrelated. Determining their possible complementarities and their inescapable tradeoffs provides the basis from which a more coherent and effective strategy can be designed. These problem-objectives must be kept constantly at the forefront. To lose sight of them is to lose control over the structure and the direc- tion of the collaborative inquiry.

(d) Provide s@cient time for a careful, collaboratice design of the sector analysis

metho$Aogy, and select sector ends, and not sector means, as the problem-objectices

Correctly distinguishing between means and ends is essential to good management. This dis- tinction varies with the decision-making and acting entity, and with the management stages of analysis, design and implementation.

Let us suppose that one of the conclusions of an agricultural sector analysis were that the improve- ment of the agricultural extension service had high priority for increasing production and farm family income. At the sector management level, this im- provement would constitute a means, not an end.

On the other hand, this improvement would become an end for the lower-level authority - the agricultural extension service. It would be the purpose or end-in-view of a project - the set of activities put into motion to bring about a specified improvement. The subsequent attainment of the specified improvement would constitute a future completion (or end) and termination of this effort by the agricultural extension service.

To prevent a potential subsector means from being taken for a sector problem-objective or end, it is necessary to provide sufficient time for dis-

cussions between the FAA and the LDC insti- tution. For example. during the design of the El Salvador collaborative education sector analysis. it was proposed that improvement of the iMinistry of Education be selected as one of the problem- objectives of the sector analysis.

The proposal was rejected after considerable dis- cussion. A comprehensive study of the ministry’s organization and operations would have constituted the substitution of a means for an end. Adoption of this problem-objective as a basic analytical com- ponent would have concentrated a large part of the professional time on internal ministry affairs. Such a study would probably have turned out to be a misallocation of analytical energies, since many of the ministry’s internal affairs were of little significance for: (I) increasing access to educa- tion and training; (2) increasing the internal effi- ciency of education and training activities; and (3) improving education’s external efficiency or relevance.

Observation and assessment of the ministry would be appropriate only when the analytical pur- suit of these objectives led the inquiry to a specific ministry department or operation. In such a case, the observational and analytical elTort would be beneficially directed by a clear and definite purpose. For example, were poor academic per- formance in a significant portion of the schools explained by a lack of training materials, and were the lack of materials traced to delivery failures on the part of the ministry, then it would be advisable to carry out observations of the ministry - not to assess the entire institution, but to examine the ministry’s ordering, warehousing and shipping procedures in order to identify flaws and to specify corrective measures.

To sum up, the fourth methodological guideline is to make sure that sector ends, and not sector means, are chosen as the problem-objectives for the sector analysis.

(e) Use the problem-objectives (the dependent cariables) as the criteria for selecting the factors of possible relevance (the independent variables)

The scope of a collaborative sector analysis should not be greater than the parties can carry out within the specified life of the project. The collaborative inquiry cannot cover every area pro- posed by every interested party. Any external party involved in a CSA must accept the fact that the specification of the research (which is likely to have considerable academic or theoretical interest) will be governed by the LDC’s practical, sector management concerns: (1) obtaining findings of immediate utility for policy: (2) developing the

COLLABORATIVE SECTOR ANALYSIS 1019

LDC’s capabilities in data collection, processing, analysis and the formulation of policy; and (3) establishing an ongoing sector analysis process in the LDC institution.

Among the first steps are the selection and measurement of the problem-objectives. Expressed symbolically, the problem-objectives are the depen- dent variables - a representation of those con- ditions that have been selected for increase or reduction. Food production can be used as an example.

If a time series of past food production levels is not available, the identification of variations in production and their causes may have to be limited to a cross-sectional analysis. This measurement of production (what is often referred to as a base line) is an essential part of a continuing sector analysis process.

Selection of the problem-objectives (the depen- dent variables) clears the way for selecting the fac- tors of possible relevance (the independent vari- ables). The latter are those factors which the inquiring parties believe have the greatest influence in raising or lowering the production levels in ques- tion. The independent variables represent hypoth- eses with implications for the formulation of policy and the allocation of resources.

Precise quantitative estimates of all the possible independent variables cannot be expected from the first sector analysis. This is but one of the reasons for institutionalizing the process in the LDC min- istry. A satisfactory explanation of the factors affecting the problem-objectives will often involve successive approximations. Increasingly accurate estimates, findings and conclusions are more likely to emerge from a sector analysis if it is a continuing activity.

(f) Identgy the relevant interconnections between the sector and the larger society to be captured in

the dependent and independent variables

The problem-objectives and the independent variables involve interconnections between the sec- tor and the larger society. There are some problem- objectives which, by definition, include an inter- sectoral objective. For example, for education the linkage is made by the relevance or external efficiency goal. To assess the external efficiency of a national education and training system is to assess the extent to which the system contributes to the attainment of desired conditions. These can be narrowly or broadly defined.

They can be limited to economic conditions and goals, such as growth in GDP and employment. Or the definition can be expanded so that it includes social and cultural conditions and goals,

such as the maintenance of national traditions, the improvement of cultural activities. an increase in political participation, etc. The LDC must decide how it wants this goal defined for purposes of analysis.

Similarly, when an increase in agricultural exports is selected as one of the problem-objectives of an agricultural sector analysis, this dependent variable links agriculture to the nation’s external financial accounts. Selection of this problem-objec- tive calls for yearly estimates of past and potential future contributions of the exported crop to increases in national foreign exchange reserves.

Some of the independent variables or hypo- thetical means will also represent intersectoral con- nections. For example, if insufficient effective demand is postulated as a major constraint to the desired increase in food production, there will be a need to gather evidence for and against this hypothesis. If inadequate nutrition is postulated as a major cause of poor learning performance in the schools, it will be necessary to move outside the sector to obtain measurement of food consump- tion or some proxy measure, such as household income.

(g) Obtain geographically and categorical!? disaggregate measures of the selected

problem-objectives

Many observers have been surprised by how much information of immediate utility for policy and day-to-day administration is provided by geo- graphically and categorically disaggregate measure- ments of sector goals or problem-objectives. To obtain the understanding of the sector that is required for the improvement of management, both aggregate and disaggregate measures are required. Aggregate measures, taken alone, can be highly misleading. National totals and averages lump together local and regional totals and aver- ages that are likely to vary greatly. National stat- istics, taken alone, often serve to mask realities of great importance for policy.

Under the El Salvador collaborative education sector analysis, increasing the internal efficiency of educational programs was a basic groblem-objec- tive. It was defined as a measurable objective in terms of four components: (1) reducing enroll- ment loss (dropout and pushout); (2) reducing repetition; (3) reducing per student cost; and (4) increasing learning and the acquisition of skills.

In order to measure the current learning levels, achievement tests were prepared in the ministry under the CSA. The achievement test results for basic education in the tested grades (3, 6 and 9) were disaggregated by the student’s income class

1030 WORLD DEVELOPMENT

and rural or urban residence, among various other categories. The average test scores for these two disaggregate groups indicated that, with the excep- tion of the highest of six income classes, there was no statistically significant difference in test scores. The rural students scored as well, and at times slightly better, than the urban students. Findings of this sort can lead to savings. For example, they can allow one to discard administratively bur- densome, costly proposals for “improving” rural education, e.g., developing a different and more “relevant” curriculum for the rural schools.

The original selection of problem-objectives rep- resents the conviction of the policy makers that each selected goal (for example, increasing food production, increasing farm family income, reduc- ing rural-to-urban migration) has great economic and social importance as a problem and, therefore, as an objective. However strong these convictions may be, they are, logically speaking, hypotheses. Consequently, as well as helping locate the “prob- lem,” disaggregate measurement serves to con- firm or reject these hypotheses.

Some problem-objectives will have a more mark- edly tentative character than others. Let us con- sider the task of conserving natural resources in a developing country in which the erosion of topsoil is apparently widespread. How should the LDC go about obtaining the information required for: (I) determination of the extent and location of the problem; and (2) the design of a cost-effective strategy for combating it?

Inquiry’into the soil erosion problem in the LDC could be carried out in at least two ways. One way would be to make it a problem-objective in an analysis of the national resources sector. Although not completed, such an approach was initiated in the Sudan. As well as erosion, the selected problem-objectives included desertification, defor- estation, and the inadequate supply and dis- tribution of water. This sector analysis scope had the advantage of focusing the inquiry on the very close interrelations among these four problem areas in the Sudan, and on the agricultural and pastoral activities involved.

Another way to treat erosion is to make it a problem-objective in an agriculture sector analysis. In this case, geographically disaggregate measure- ment may serve more directly to confirm or reject the hypothesis that erosion is a widespread phenomenon with serious implications for sus- tained production.

The disaggregate measurement of the problem- objective can serve to eliminate programs and pro- jects that would waste resources. Included in this category are national, standardized programs that introduce corrective measures in regions that do not require them.

In sub-Saharan Africa, for example, a more selective distribution of food could bring about enormous savings. But a more selective dis- tribution of food requires much better measure- ment of production than is currently obtained in most African countries.

Under the Rwanda Agriculture Survey and Analysis project (RASA), procedures for accurate measurements of crops and plots were devised and successfully carried out. Each farm in the sample was given baskets of a standard size for each crop, along with a pamphlet or log. Each crop was iden- tified on different pages of the pamphlet or log through a picture of the crop at the top of the page and scores of pictured baskets printed below the crop picture. As baskets were filled with a crop the farm head (perhaps illiterate) would pencil a circle around another basket on the appropriate crop page. The plots on which this food was grown were also accurately measured, in this case by extension agents who received special training.

Detailed information concerning the household (age, sex, school attendance, time spent on the different kinds of farm work, including marketing, etc.), and production data were collected for a sample of farms. This sample was stratified so that reliable estimates could be made for prefectures (provinces) and ecological regions. Consequently, once the data processing began, it was possible to quickly identify regions of food surplus and deficit.

The importance of this data can be underscored by a well-known example - the inadequacy of the information used to guide food distribution under the Food for Peace program. In the absence of reliable, up-to-date food production data, areas in need can be overlooked, and the provision of cheap, imported food in areas in which the locally grown crops are sufficient or in surplus can act as a disincentive for farmers.

Better food distribution can be the first payoff of an agricultural sector analysis. Success in the identification and quantification of factors that impact on production can later provide a basis for designing a more coherent and effective agriculture sector strategy.

(h) Estimate the effect that changes in the sector means (the independent variables) have on the sector

ends (the dependent variables)

As .has been pointed out, finding difierences in the performance of a sector objective, such as learning or agricultural production, when mea- sured in disaggregate terms, is a first step in the process of identifying and explaining variations in the problem-objective - the sector performance indicators that represent one of the selected goals.

COLLABORATIVE SECTOR ANALYSIS IO21

It is noteworthy how often these disaggregate measurements constitute findings of significance for policy. For example, the tabulation of farm production by farm size might indicate a minimum size below which production is not economic or is insufficient to support a farm family. A correlation of monetized farm production per acre or per farm family member that compares farms that have taken anti-erosion measures and farms that have not taken such measures (this information was cap- tured in the questionnaire used in the Rwanda Agriculture Survey and Analysis) can provide a way to estimate the short-range and private cost- benefit ratio of these anti-erosion measures.

Some of the statistical results of the analysis may serve to discredit the means favored by the LDC and the FAAs - means which may have been engaging many officials in the public sector, and consuming large amounts of public sector revenue.

A sector analysis can identify changes in the national agricultural strategy which can make it more effective and efficient. Nevertheless, the scope of the analysis is necessarily more limited than the scope of considerations and issues which the strategy should address. The analytical or diag- nostic phase is concentrated on explanation, and is obliged, therefore, to look back at the past. The design, planning and programming phase looks ahead. It concentrates on future actions to be taken over a specified period of time.

A successful analysis will identify past and present constraints to, and missed opportunities for, greater progress in the attainment of each selected sector end. Shifting from the identification of defects and deficiencies to a specification of the conceptually related corrective measures is simply the first step in the design of a strategy. These selected corrective measures may have limited tech- nical validity. They may not be attainable in actual practice. In other words, they may confront impediments hitherto ignored. In order to assess the practical feasibility of the selected corrective measures, one must appraise them in the light of financial, administrative and political factors.

(i) Identgy and quantify possible complementarities and inescapable tradeoffs among ends and means.

Use these to design a sector strategy that is feasible

(i) Technical feasibility The multiplicity of development goals at macro-

economic and at sector levels makes it more diffi- cult to formulate policy and design programs. Too often the difficulty is compounded by unwar- ranted assumptions concerning the nature and interrelations of these goals. Policy makers are

inclined to view these goals in a disjunctive manner, as either wholly conflicting, in a position of trade- off, or as complementary and mutually supporting. This way of viewing social goals involves a priori assumptions that can be very misleading. It repre- sents the oversimplified approach of policy makers who are frequently inclined to ignore or dismiss the facts in order to move directly from theory to policy, from the textbook to a design. Such an approach neglects the necessary intermediate step of an empirical country analysis.

The collection and analysis of relevant national data places the relation among goals in a new light, and may alter previously held assumptions. An empirical analysis may reveal hitherto unknown conflicts among means. Potential complemen- tarities that were not previously perceived may also emerge from the empirical analysis. The iden- tification and quantification of present or actual ends and means, and potential, future ends and means, make the discussion of complementari- ties and conflicts more specific and useful.

Let us assume that the goals or problem-objec- tives of an agricultural sector analysis were to increase food production, farm family income, and the value of agricultural exports, and to reduce rural-to-urban migration. If the means required to achieve the selected food production target over a IO-year period were the same means required to achieve the targeted increases in family income for that period, these two goals would be comple- mentary. They would not compete for resources. If one of the findings of a survey were that members of those farm families that have attained the tar- geted levels of income were not migrating, then the goal of reducing rural-to-urban migration would also be complementary.

On the other hand, if the country’s main agri- cultural export earner were a crop grown on estates and not on family farms, then the financial resources that might be required for attaining this goal would be in competition with the financial resources required for the food production, family income and migration goals. But if the exported crop were grown or could be. grown efficiently on family farms, there might be very slight additional requirements for resources, such as credit, ferti- lizer, and the technical assistance needed to im- prove cultural practices and to alter crop mix.

One might assert a priori that food production and the export crop would necessarily compete for land, so commonly a scarce resource. But if the land were equitably distributed and inclusion of the export crop on family farms were both finan- cially and ecologically advisable, then even this resource would not necessarily represent a conflict among the goals.

Of course, the examples given are not typical.

1021 WORLD DEVELOPMENT

They are too simple and too aggregate. In almost any developing country the relation among these factors will differ from region to region. The geo- graphical disaggregation of data and the quan- titative analysis of means-end connections on a regional basis, under the sector analysis, provide the information basis for a regionally differentiated strategy. Such a strategy is more likely to bring about the attainment of goals.

If present and future resources are scarce, this regionalization of goals may well serve to highlight resource conflicts among regions, and the need to establish different regional priorities. The geo- graphic disaggregation of the data might demon- strate major weaknesses in a standard, national program, and lead to arrangements for the decen- tralization of policy making and programs.

A sector analysis that identifies the defects in national policies, public interventions, and regu- lations can be used to make the national sector strategy more coherent and effective. But the identification of defects and the selection of the corresponding corrective measures is simply the first step. Although each proposed corrective mea- sure may be technically sound as a separate or isolated reform, a sector strategy that incorporates all of the proposed corrective measures may not be financially, politically or administratively feasible.

(ii) Financial feasibility A financially feasible sector strategy calls for

studies or inquiries of three kinds: (I) a detailed analysis pf past programs to identify ineffective and expensive procedures, areas of waste, and potential savings; (2) estimates of future funding availabilities and the use of these estimates as financial constraints on the planning of sector pro- grams; and (3) predictions of the consequences of proposed policy changes, with particular attention to the financial implications of the expansion of public and private services.

The impact of reforms and corrective measures on the demand for private and public services, and the subsequent distribution of public revenues, can be illustrated by the education sector, in which the graduates of primary school are the candidates for secondary school, and secondary school graduates are the candidates for higher education. The expan- sion of public education at any one of these levels will serve to increase that level’s recurrent costs and to’deny these financial resources to the other two levels.

One of the problem-objectives of the El Salvador collaborative education sector analysis was to increase the internal efficiency of the formal edu- cation system. Studies of primary education identi- fied the major causes of enrollment loss and repe- tition in the primary schools, as well as ways to

reduce per student cost without lowering satis- factory learning achievement levels. This made it pos- sible in the design or strategy phase to specify the measures needed for increasing internal efficiency.

Another of the basic problem-objectives of the collaborative sector analysis was to increase access or enrollment at the various levels of the education system. It soon became clear that the objective of increasing access and the objective of increasing efficiency were not in a simple and uniform relation of either complementarity or tradeoff. The proper treatment of both objectives and their inter- connections required considerable quantification of enrollment targets and costs.

Increasing the internal efficiency of primary edu- cation has the effect of increasing access to sec- ondary and higher education. The reduction of repetition and enrollment loss accelerates the upward flow of students through the primary grades. It also increases the rate of entry into pri- mary school, particularly when there is a high population growth rate and pent-up demand for schooling. Both of these changes lead to a higher rate of graduation, and to greater demand for secondary education. More secondary graduates increase the demand for higher education.

The per student costs in secondary and higher education can be as much as 15 and 30 times greater than the per student cost in primary educa- tion. Consequently, if there is no forward planning for expanded access to each of these three levels, the increase in public recurrent costs at the higher levels is likely to absorb larger and larger portions of the public education budget, and preempt the expansion of primary education that is needed in countries with fast-growing school-age popu- lations. In effect, a careful review of actual expen- ditures (not budgets) during the IO years that preceded the El Salvador CSA planning period showed that the upward drift in expenditure had taken place, and had served to hold back the ex- pansion of primary education, which was already seriously undersupplied.

An excessively rapid upward drift in the shares of education expenditures reduces the possibility of maintaining the rate of primary education expansion required by a growing school-age popu-

lation. This is the trend in many nations. LDCs and bilateral and multilateral donors have not given this widespread phenomenon the attention it so clearly deserves.

In the case of El Salvador, it was clear from the analysis that the upward shift in education expenditure shares had been excessively rapid. How could this trend be reversed?

In El Salvador, the 1977-2002 planning period was used to project the growth in school-age popu- lation and to estimate the financial resources that

COLLABORATIVE SECTOR ANALYSIS 1023

would be availabie for the entire IMinistrv of Edu- cation budget, including the public administration of education, the special training programs, and the three closely linked activities of primary, secon- dary and higher education. Estimates of expanded enrollment in private education during the plan- ning period entered into the calculations. The esti- mate of future fiscal availabilities was deliber- ately made conservative, and a high. sustained school-age population growth rate was projected for the entire 25-year period.

The computer programs used the estimate of the fiscal resources available each year of the 25-year period as a constraint on the calculation of the different combinations of financially feasible en- rollment targets for each of the three levels. This 25-year period was also used for the scheduling of the efficiency improvements. For example, the incidence of first grade repeaters from the 1977 base-year level of 40% (due mainly to a shortage of space in the second grade) was scheduled for very gradual reduction to 10% by 2002.

The results of these estimates of financially feas- ible expanded enrollment were highly encouraging. They indicated that by using its own financial resources wisely, El Salvador could bring about enormous improvements in educational attain- ments at all educational levels, For example, in the rural areas the new entrants into grade 6, as a percentage of the age cohort (I 3-year-olds) would increase from 21% in 1977 to 76% in 2002. In the urban areas there would be an increase from 59% to 90%. These estimates, and the estimates for entrants into grade 9, the last year of high school, and for the graduates of higher education, appear in Table I.

Can these very large increases in the country’s level of educational attainment, or in the nation’s human capital stock, be achieved in other LDCs? What are the methodological lessons for analysis and design?

For one thing, it is clear that the LDC should explore the relations between two of the main prob- lem-objectives - access and internal efficiency -

Table I, Financially feasible enrollmenl targels for El Salvador

Year Matric. New

entrants Repeaters

New entrants and higher ed. graduates as % of age cohort

Rural zone

1977 17,956 2002 91,433

1977 3,219 2002 59,025

1977 0 2002 12,729

1977 0 2002 0

Urban zone

1977 40,610 2002 98,956

1977 35,741 2002 91,794

1977 16,242 2002 45,880

1977 34,795 2002 127,689

I2,28 1 87,024

3,143 57,935

Grade 6 5,675 4,409

Grade 9 77

I.090

Grade 12

21.06 75.88

5.77 54.37

0 12.94

Higher education 0 0

34,587 94,136

35,463 90,008

Grade 6 6,024 4,820

Grade 9 279

1,786

Grade 12

Higher education Graduates ___ -

10,756 1,132 34,286 6,547

58.83 90.33

64.39 91.81

31.48 49.44

Graduates as % cohort 2.45 8.00 (Urban cohort) 3.94 (Nat’1 cohort)

102-l WORLD DEVELOPMENT

to determine possible complementarities and in- escapable tradeoffs, and to prevent an excessively accelerated upward drift of expenditure, with a robbing of Peter (primary education) to pay Paul (secondary and higher). This was occurring during the 10 years prior to the analysis. This is what is occurring right now in most Latin American countries.

This third, overall appraisal is an assessment of the administrative and political feasibility of the entire proposed strategy. It should take place after the actions to be implemented have been identitied and scheduled. and after thejnancial viability of the entire plan has been verified.

If one of the LDC’s goals is to have an increas- ingly larger percentage of the school-age popu- lation completing primary education, the LDC will have to control the expansion of, and entry into, publiclv jinanced secondary schools and higher education institutions. lfmanagement of the sector is shaped entirely by responses to immediate pres- sures - if analysis and planning are absent and crisis management prevails - the LDC with scarce resources and a fast-growing school-age popu- lation will not achieve the progress it seeks in the desired movement to universal primary education.

A nation’s education is a serious, long-range affair that should not be treated lightly. An edu- cation strategy requires an extended planning period. Without estimates of future financial avail- abilities, and future capital and recurrent costs, the consequences of present actions are unpredictable. An extended planning period is needed, as well, for the scheduling of corrective measures. Without such scheduling the probability of successful imple- mentation is considerably diminished.

The reliability of projections diminishes as they are extended into the future. This is one of the reasons that analysis, design and planning should become a,continuing process in the Ministry of Education. The actions inescapably affected by this year’s projections do not lie in a distant future. They are the actions that are taken this year. Consequently, the improvement of estimates and the updating of projections should take place erery year. This is not a difficult task. Once computer programs have been designed, tested, and put into place, comparatively little time and expense are required to update and correct the projections with new data and revised estimates.

Was the El Salvador 25-year education plan pol- itically viable? It did not call for a larger share of GDP or national government revenues for publicly financed education. Indeed, the plan possessed the unusual feature of calling for reductions in these two financial shares. In this respect it appeared to satisfy the requirement of political feasibility.

There was another dimension of the plan. how- ever, that might not be as popular: the need to have more stringent admission policies for publicly financed secondary and higher education. This need was not immediately apparent. After all, the sector plan called for steady and substantial increases in the percentages of the school-age population or grade cohort. By the year 2002, high school and higher education graduates would rep- resent a very much greater percentage of the IS- and 22-year-olds in the population. High school graduates would increase from 0% to 13% of the school-age cohort (I S-year-olds) in the rural zones, and from 3 I % to 49% in the urban zones. Higher education graduates would increase from 2.5% to 8% of the national 22-year-old cohort. Obviously. targets representing such large increases (increases in the proportions of a growing population that graduate from primary, secondary and higher edu- cation) are good news for the population at large, and can be used by the national government to muster initial support for the plan.

(iii) Political and administrative feasibility Thefinancial feasibility of a sector plan is obvi-

ously a basic issue. But, as has been suggested, the feasibility of a sector policy or strategy is also administrative and political.

Unfortunately, this is not the end of the matter. Success can generate new problems and new kinds of dissatisfaction. Graduation from primary school creates an appetite for secondary school. Similarly, increased graduation from secondary school increases the demand for education at the tertiary level. Now, each of the levels cannot and should not increase at the same rate. Such increases are not financially feasible, nor wise from a devel- opmental and labor force planning vievvpoint. The rate at which they should increase will often involve a reduction in the percentages of primary school graduates that enter secondary school and in the percentages of secondary school graduates that enter the higher education level.

Issues of political and administrative feasibility Let us suppose that before the expanston and should be considered in the following ways: (I) by improvement of primary education, eight of every making an inventory of administrative resources 10 primary graduates could find places in sec- and reviewing political positions and forces; (2) by ondary school. After the expansion and improve- assessing the political and administrative feasibility ment there will be many more primary school of individual corrective measures and expansion graduates, not only in absolute terms, but also as targets; and (3) by carrying out a comprehensive a proportion of the IS-year-old cohort. Never- appraisal after the synthesis or design. theless, at this later date there may be public sec-

COLLABOR.iTIVE SECTOR ANALYSIS lOl5

ondary school places for only five of every IO pri- mary school graduates. It is this future reduction in the admission rnfe that could lead to protests.

An expansion of private education, stimulated perhaps by appropriate government incentives and reformed policies, can help reduce social pressures for the expansion of publicly financed education. But in many of the poorer nations with fast-grow- ing school-age populations, expansion of the pri- vate education sector will be too slow, as well as too costly for lower-income families. It will fail to keep up with social demand. In this case, more stringent admission policies, based on academic merit and economic need, should be applied at publicly financed secondary and higher education institutions.

To make sure that more stringent admission policies and their political rationales are devel- oped in time, and communicated persuasively to the public, the problems created by continuous increases in the percentages of primary and sec- ondary school graduates need to be anticipated. Such problems are less likely to be solved in situ- ations of crisis - at least, less likely to be solved in ways that contribute to equity and to economic growth. One of the functions of analysis and for- ward planning is to help anticipate sector problems. Admittedly, there are many sector prob- lems that cannot be anticipated. But anticipat- ing future large-scale problems of the kind just discussed is an integral part of sound sector management.

Many important issues will not be identified in the absence of analytical and systematic pro- cedures. If the feasibility reviews do not take place on a scheduled basis, major problems will be overlooked. For example, a sector strategy or program may call for professional services not pres- ently available. If the need for these professionals is not anticipated, and candidates are not educated or trained, the professionals will not be available when future actions require them, and the plan will fail. To prevent this kind of failure, it is necessary to specify future staffing requirements and the cor- responding training programs.

In the El Salvador CSA, the secondary and higher education enrollment increases included normal schools for the preparation of the required numbers of future primary school teachers and university programs for the preparation of future high school teachers.

The expansion and corrective measures of an agricultural sector strategy also call for trained personnel. There should be a special review to quantify possible future shortfalls in agricultural professionals and technicians so that the needed training programs can be made part of the strategy.

If a sector strategy is not technically, financially,

administratively and politically feasible, it cannot be implemented. We restate this self-evident truth to underscore the importance of reviewing all planned public activities in the light of predicted future conditions and estimated availabilities. Applying these predictions and estimates as con- straints on planned activities will often lead to the recognition that the public sector workload is excessive. The inclusion of the various feasibility tests in the design process sets the stage for a second assessment ofcomplementarities and tradeoffs that may lead to a change in priorities, and to a reduction in the number of interventions and total amount of expenditures.

Admittedly, the ministry will often lack the needed expertise - a staff with substantial training and experience in data collection, processing, and analysis. Under such conditions, it will not be able to begin a rigorous application of each of the feasi- bility tests. Establishing a continuing sector analy- sis process in the LDC ministry makes it possible to gradually add each of the feasibility inquiries, and to limit the scope of the first CSA. In the first analysis and design, it is probably advisable to concentrate on cost control This is usually the most urgent of the feasibility issues. Besides, the resolution of this issue will often contribute to the resolution of administrative, technical and political issues.

(j) Institutionalize a continuing sector analysis process in the appropriate LDC organization

A firm LDC commitment to institutionalize a continuing sector analysis process in the policy- making agency or ministry is necessary for the initiation of a CSA project. In addition to staffing and training. the LDC and the FAA must be will- ing to pay salaries that are sufficiently attractive to retain competent statisticians, economists, survey managers, computer scientists, sociologists, etc. Some of these LDC officials will need prior formal training; others will benefit more from the intensive, on-the-job training that can be provided by resi- dent advisors. In many cases it will be advisable to alternate both kinds of training - with one specialized ministry official receiving on-the-job training while another official with the same spe- cialty is receiving formal training in an aca- demic institution, and then later having them ex- change roles. However, training is only one of the many requirements that need to be met. If the LDC and the FAA are not prepared to cooperate in establishing the long-term working conditions that are needed to attract and retain a cadre of com- petent analysts, it is probably best not to initiate a CSA.

1026 WORLD DEVELOPMENT

The needs will vary among countries and sec- tors. The Ministry of Agriculture in Rwanda had to start its institution building from scratch. Tq initiate the Rwanda Agricultural Survey and Analysis (RASA) project, the ministry had to create an entirely new statistical and analytical unit

’ and staff it with university graduates. Some were sent by the project to US universities as candidates for graduate degrees, and others received on-the- job training from resident advisors and short-term consultants.

A radically different governmental and insti- tutional situation was addressed under the Zam- bian Agricultural Training, Planning and Insti- tutional Development (ZATPID) project. This CSA project did not require the creation of a new unit. The main task in Zambia was to improve the analysis, design, implementation and evaluation performance of the many national entities with responsibilities in the agricultural sector, and to bring about better coordination of their activities. A large part of project resources was channeled to the Planning Division of the Ministry for Agri- culture and Water Development and to those units in the Central Statistics Office responsible for the collection, processing, tabulation and analysis of agricultural data. However, training and technical assistance were also provided to the National Com- mission for Development Planning, the National Marketing Board, the Ministry of Finance, and the Rura! Development Studies Board of the National University. To contribute successfully to the instikutionalization of a continuing sector analysis process, the CSA must be properly tailored to the LDC’s three “I? - institutional situation, status of information, and interventions.

Another key internalization and institution- building issue involves decisions as to what to do first. Whether a “general survey” or “special studies” are more important for improving sector performance is an argument that takes place in development circles with lamentable frequency. Such a dispute is misleading. Both kinds of inquir- ies are needed. The significant issues concern the sequence, timing, nature and scope of these studies. The project should not be overloaded. It should not be burdened with an excess of simultaneous activities. Under the ZATPID project, a special effort was made to stretch out the timing of special studies in areas such as credit, fertilizer costs and benefits, groundnut production and marketing, tractor rental, etc., into a multiyear schedule, so that at least one Zambian official would get on- the-job training by working closely with the foreign expert contracted on a short-term basis for the specialized study.

The amount of ministry activity and training in the areas of data collection, processing, and analy-

sis that can be carried out during a year is obviously limited. Consequently, it is necessary to select the studies to be postponed or delayed. In Rwanda. the Ministry of Agriculture and USAID agreed that the general survey of farms, agricultural pro- duction and rural households was a practical first step. The ethnic homogeneity and the small size of the country made this general survey a manageable task, and the best way to begin.

In Zambia, there were many factors that made such a beginning inadvisable. The country is much larger, and some agricultural areas are com- paratively inaccessible. The ethnic heterogeneity would call for the training of data collectors and interviewers with greater linguistic diversity; the national institutions regulating and intervening in the sector were more numerous and intrusive; data collection could not begin with the advantages of a fresh slate and would require the elimination or revision of unrepresentative samples and poorly constructed questionnaires.

Consequently, a successful sample survey of farms could not initiate the CSA. At the start, it would not be physically and administratively possible to obtain the reliable baseline data on agricultural production, land, labor and capital that are definitely needed for sound and sustained management of the agricultural sector. The design of ZATPID reflected the recognition that in Zam- bia the “special studies” (with emphasis on various existing policies, including those that governed pricing and credit) would have to precede the “general survey.” Since considerable prior reor- ganization, procedural change and training were needed to successfully carry out such a survey, these preparatory activities were made a part of the pioject.

(k) Help strengthen LDC management capabilities and help bring about a gradual transition

to LDC leadership by supporting periodic, in-country* sector meetings that are chaired

by the LDC and attended by donors

The 10 preceding guidelines for a CS.4 have been gleaned from experiences in El Salvador, Zambia and Rwanda. However, none of these LDCs can claim a markedly superior record in the man- agement of the particular sector involved. Why is this so? Is the CSA concept inherently flawed? If not, are there other conditions that need to be met if a CSA is to achieve its purposes?

It is important to underscore the fact that none of these three CSAs has enjoyed broad-based or coordinated support. All were bilateral USAID projects in countries with a proliferation of projects. In each of these countries, other foreign

COLLABORATIVE SECTOR ANALYSIS 1027

assistance projects competed with the CSA for the LDC’s highly limited sector resources in the areas of administration, logistical support, finances and trained personnel. None of the three CSAs was made the central sector project. None was sys- tematically monitored in periodic, in-country sec- tor meetings. chaired by the LDC and attended by donors.

The need for these meetings can be illustrated through consideration of a hypothetical worst-case situation. Let us assume that the LDC sector in question has the involvement of a very large num- ber of donors that do not coordinate activities. Let us suppose that the “sector strategy” is little more than the set of projects that have been designed and initiated by donors, that the proliferation of projects generates existing and future recurrent costs greater than current revenues and the esti- mate of future availabilities, and that admin- istrative and trained staffing demands of these pro- jects are greater than the LDC can provide. Let us assume, as well, that the LDC is seriously deficient in the analytical, design and implementation capa- bilities required for successful management of the sector. Unfortunately, these are usually valid assumptions for sub-Saharan LDCs.

The LDC needs to move on two fronts - the action front and the information front. It depends on the donors for analysis and design, and on donor financial and technical resources for much of the implementation. This great dependence on donors is likely to be self-perpetuating. The trans- fer of resources is governed by the obligation schedules of the foreign assistance agencies. This leads to the continuous initiation of projects; and the proliferation of projects makes it more difficult to identify priorities. Comprehensive sector analy- sis and design are not likely to take place when almost all the time the better-trained LDC pro- fessionals is dedicated to the negotiation and implementation of foreign assistance projects.

A successful CSA would bring about a thorough, although necessarily gradual, change in the worst-case scenario. It would begin to provide the information required for improving policies. It would lead to the elimination of low-priority projects and would reduce expenditures.

Unfortunately, a CSA is likely to be most urgently needed in the situation in which it is most likely to be starved for national resources. When donors are constantly starting projects, they are preempting a very large portion of the LDC’s financial, physical and trained human resources. In such a situation, to initiate a CSA without also reducing the amount of donor projects is to lay the groundwork for failure.

What should the foreign assistance community do to help the LDC move away from an incoherent

sector “strategy,” designed and executed in a piece- meal fashion by donors? How should the com- munity help the LDC move toward a more coher- ent and effective sector strategy that is increasingly the product of LDC analysis and design?

Improvements in communication and coor- dination in the LDC are required for a gradual and successful transition from a fragmented to an integrated strategy. During the transition period, the LDC would continue to receive donor guidance in design, as well as commodities, technical assist- ance and training. But this donor guidance would no longer be given on a piecemeal, project-by- project basis. The development of a comprehensive strategy for the sector would be a topic for periodic discussion; and there would be a much greater and more constant effort to coordinate the inter- ventions of donors in the sector. Donor coor- dination would cease to be a topic that is reserved for occasional and superficial treatment in a so- called Consultative Meeting held in Paris, Wash- ington or Rome. Donor coordination would become a systematic, periodic activity that takes place in the host country, with the host country playing the leading role.

How can this coordination be assured? The initiation of a CSA could require a prior agreement among the LDC authorities and the donors involved in the sector. Let us assume that this agreement provides for monthly, bimonthly or quarterly meetings of the LDC and donors. Accountability will be the guiding principle of these meetings. To strengthen LDC management capabilities, and gradually move the LDC to a position of leadership, the donors will require much better accounting of the sector’s resources. They will need better estimates of resource use and efficiency, and a more complete picture of financial flows, unit costs, and administrative operations.

The four components of internal efficiency in the education sector can be used as an example. The indicators or measures of primary education efficiency - the geographically and categorically disaggregate estimates of enrollment loss, repe- tition, per student cost and learning - could be presented to the minister, the permanent sec- retary and, say, the heads of departments every month, quarter and semester, in the form of highly condensed reports or executive summaries. These reports would be used by these officials to monitor progress. They would be used for decision making, and lead to changes in various areas, such as aca- demic programs, construction plans, and teacher training. Holding ministry meetings of officials from the various departments with the sector analysis and design officials would open the way for a periodic exchange between the individuals responsible for administration and the individuals

1028 WORLD DEVELOPMENT

responsible for analysis and planning. Such an exchange is essential. It helps assure that the ana- lytical and planning functions continue to operate on an effective and practical basis.

Reports on access, flow, graduation, learning and costs that provide a clear picture of trends, updated targets and goals, would be prepared every year for administrative and policy-making purposes. For example, reports on secondary edu- cation that highlighted a large difference in the per student costs of technical and nontechnical schools could lead the authorities to reduce the number of technical schools and have significant cost-saving impacts.

Data summaries and interpretations in the form of brief reports and graphic presentations that serve to inform LDC policy makers, LDC admin- istrators, and donor representatives can also be used to keep the LDC public informed. This is, of course, a national matter. The LDC government would have to decide whether these reports should be widely disseminated.

Admittedly, donors who collaborate with LDCs in obtaining information will sometimes find them- selves in a dilemma. They will doubtlessly recog- nize that it is up to the LDC government to decide whether to disseminate the sector information of general interest and significance. But they may also recognize that a broad sharing of information through documentation and dissemination can contribute mightily to the undoing of government abuse and corruption. The LDC government that is hostile to the idea of providing a full account of the financing, performance and management of its sector should probably not accept the initiation of a donor-supported CSA.

In the worst-case situation, some years will elapse before the sector analysis process can pro- vide the LDC and FAA members of the sector meetings with all the information they need. Mean- while, it will be necessary to assess donor pro- posals - additional projects or new interventions carried out at the subsector level and intended to strengthen a sector component or means, such as agricultural research, the extension service, a seed production or marketing center, an irrigation scheme or the production of a new crop.

The expansion or improvement objective of the proposed project will be a subsector end (i.e., a sector means) believed to be important for attain- ing a sector goal or end. It will therefore be advis- able to have the sector meeting members assess the proposed project in terms of its likely impact in easing a major constraint to, or in seizing a sig- nificant opportunity for, attainment of the sector goal in question. It will also be useful for the mem- bers to evaluate the proposal in terms of estimated financial constraints, and an administratively

attainable workload. In short. the methodological approach taken under the CSA would also be taken at the sector meetings, albeit in a less rigorous and less quantitative fashion.

Adoption of this coordinated sector approach by the LDC and donors will oblige all the parties to meet the requirements of sound management. Under this kind of sector approach, the LDC and donors will make sector analysis the basis for strat- egy. They will have to estimate the present and future financial availabilities of the sector, and plan and initiate projects within these constraints. This will force the parties to make choices. In order to make choices they should identify constraints to, and opportunities for, the attainment of sector goals, and rank activities and expenditures accord- ingly. Establishing priorities and controlling costs is the first order of business.

This sector approach should make it more feas- ible to shift part of the foreign assistance provided each year from new or extended projects to sector support. For example, a shift might take place if the LDC and donors were to decide that a pro- posed project did not have high priority, and that the proposed resources should be used to provide technical and administrative training instead, or to ease the sector’s financial constraint. Review of the sector’s total public capital, operating costs and cost structure would provide the criteria needed to plan for a gradual decrease in LDC financial dependence on donors.

Such an approach to the sector provides better criteria for assessing the need for an increase in financial assistance from donors on a provisional basis. In certain circumstances, financing recurrent costs should be considered. Donor financing for a specified amount would be tied to a reduction and phase-out that would take place on a rigorously scheduled basis. The scheduled period would be used by all parties to ease the LDC’s transition to program sustainability, and to intellectual, as well as financial, LDC self-reliance.

The parties involved in the agreement would have to work out the details of their qrtidpro quo. Speaking generally, the precondition for a tem- porary provision of funds for recurrent costs, on a scheduled phase-out basis, would be a full and detailed accounting of revenues, budgets, expen- ditures, and progress in attaining the goals that were selected by the LDC for the sector-the very same accounting that is needed by the LDC to improve its sector management.

4. EXPANDED SECTOR ASSESSMENTS

A collaborative sector analysis (CSA) has three purposes: (1) to provide findings of immediate util-

COLLABORATIVE SECTOR ANALYSIS IO19

ity for policy; (2) to strengthen LDC capabilities in data collection, data processing, analysis and the formulation of policy; and (3) to help consolidate a continuing sector analysis and design process in the pertinent ministry.

The foreign assistance project that pursues these objectives will usually have a duration of 5-10 years and require a considerable amount of resources. The physical, financial and trained human resources that the LDC must provide are invariably scarce. A CSA is a major undertaking, and the decision whether to initiate such a project should be carefully considered.

Since resources are scarce, and there is com- petition between new and old projects for these limited resources in every country and sector, it is highly desirable to have sound criteria for project selection. Since a sector analysis can provide such criteria, and a CSA helps introduce and con- solidate a sector analysis process, initiating a CSA may appear to be invariably desirable. This is not necessarily so.

A CSA is also a project. It competes for resources with other projects. Consequently, there are always significant questions concerning the advisability of initiating a CSA. Is a CSA project in this sector needed? Is it likely to succeed? What priority does it have? How does it rank in import- ance when compared to existing activities and other proposed and potential projects? It may be relatively easy to compare the cost of a CSA with other projects. However, the prediction and quantification of benefits will be extremely diffi- cult. Is there any way of deciding what the short-, medium- and long-term benefits of a CSA in a specific country and sector are likely to be?

Since a CSA calls for the commitment of a con- siderable amount of resources by both the LDC and the foreign assistance agency, these parties should come to their own independent conclusions.

The current studies carried out by donors do not provide the answers to basic questions concerning a CSA’s desirability and feasibility. As has been

pointed out, sector assessments are almost always concentrated on the action front. They are pri- marily concerned with the identification and initiation of foreign assistance projects, and only secondarily concerned with LDC management and informational needs and capabilities. If improving LDC sector management becomes a central donor concern, it will be necessary to expand the scope of most sector assessments and reports.

An expanded sector assessment will cover the range of phenomena referred to earlier as the three “1”s - interventions, institutions and infor- mation. Such an assessment should be carried out before initiating a CSA. As well as reviewing the projects that are being carried out by the LDC and donors, an expanded sector assessment evaluates the LDC’s performance and capabilities in data collection, processing, analysis, policy formu- lation, and the design and implementation of the sector’s programs and projects. The workload in each of these areas currently carried out by the ministry, the adequacy of its organizational arrangements and administrative procedures, and personnel education and training should all be part of this assessment.

An expanded sector assessment can help answer the kinds ofquestions that need to be asked prior to deciding whether to initiate a collaborative sector analysis. What are the strengths and weaknesses in current LDC performance in the collection, pro- cessing, and analysis of data, and the formulation of policy? Is this performance adequate? If not, would a CSA bring about sufficient improvements to warrant its initiation? If so, what should be the scope of the first CSA? What are the institutional and informational defects and deficiencies that need to be taken into account? What sort of collaborative sector analysis will be most effec- tive in developing the LDC’s analytical and design capabilities?

Donors who wish to help LDCs improve their sector management could begin by preparing expanded sector assessments.