Jour nal of Economic C ooperation C · PDF fileJour nal of Economic C ooperation ... The Real Effect of Government Debt: ... Bangladeshi producers and suppliers ought to

J o u r n a l o f E c o n o m i c C o o p e r a t i o n

& Development

J o u r n a l o f E c o n o m i c C o o p e r a t i o n

& Development

Stat i s t ica l Economic and Socia l Research

and Training Centre for Islamic Countries

(SESRIC)

Journ

al of Econ

omic C

ooperation

and

Develop

men

t

Se

pte

mb

er 2016

VO

LU

ME

37, N

o.3

ISSN 1308 - 7800 Volume 37, No. 3 September 2016ISSN 1308 - 7800 Volume 37, No. 3 September 2016

Oil Price Effects on Exchange Rate, Output and Consumer Price: A Case Study of Small Open Economy of Oman

The Impacts of Foreign Labour Entry on the Labour Productivity in the Malaysian Manufacturing Sector

The Real Effect of Government Debt: Evidence from the Malaysian Economy

Environmental Kuznets Curve for Deforestation in Indonesia: An ARDL Bounds Testing Approach

Determination of the Degree of Development and the Impact of the Information Environment on the Formation of A System of Social Control in Procurement Under the Russian Contract System (Method of Content Analysis of Information Resources on the Internet)

Bilateral Trade through Official Channel between India and Bangladesh: An Analysis with the Use of Time Series Forecasting Models

Ahmed Nawaz Hakro and Abdallah Mohammed Omezzine

Nur Sabrina Mohd Palel, Rahmah Ismail and Abdul Hair Awang

Siti Nurazira Mohd Daud

Efendi Agus Waluyo and Taku Terawaki

N.A. Mamedova and A.N. Baykova

Muhammad Mahboob Ali and Anita Medhekar

CMYK

CMYK

STATISTICAL, ECONOMIC AND SOCIAL RESEARCHAND TRAINING CENTRE FOR ISLAMIC COUNTRIES

Kudüs Cad. No:9 Diplomatik Site 06450 ORAN-Ankara, TurkeyTel: (90-312) 468 61 72-76 Fax: (90-312) 468 57 26

Email: [email protected] Web: www.sesric.org

STATISTICAL, ECONOMIC AND SOCIAL RESEARCHAND TRAINING CENTRE FOR ISLAMIC COUNTRIES

Kudüs Cad. No:9 Diplomatik Site 06450 ORAN-Ankara, TurkeyTel: (90-312) 468 61 72-76 Fax: (90-312) 468 57 26

Email: [email protected] Web: www.sesric.org

ISSN 1308 – 7800 Volume 37, No.3, September 2016

J o u r n a l o f

E c o n o m i c C o o p e r a t i o n

& D e v e l o p m e n t

Statistical Economic and Social Research

and Training Centre for Islamic Countries

(SESRIC)

Contents

Ahmed Nawaz Hakro and Abdallah Mohammed Omezzine

Oil Price Effects on Exchange Rate, Output and Consumer Price:

A Case Study of Small Open Economy of Oman .................................... 1

Nur Sabrina Mohd Palel, Rahmah Ismail and Abdul Hair Awang

The Impacts of Foreign Labour Entry on the Labour Productivity in the

Malaysian Manufacturing Sector ........................................................... 29


The Real Effect of Government Debt: Evidence from the Malaysian

Economy ................................................................................................. 57

Efendi Agus Waluyo and Taku Terawaki

Environmental Kuznets Curve for Deforestation in Indonesia:

An ARDL Bounds Testing Approach .................................................... 87

N.A. Mamedova and A.N. Baykova

Determination of the Degree of Development and the Impact of the

Information Environment on the Formation of A System of Social

Control in Procurement Under the Russian Contract System (Method of

Content Analysis of Information Resources on the Internet)............... 109

Muhammad Mahboob Ali and Anita Medhekar

Bilateral Trade through Official Channel between India and Bangladesh:

An Analysis with the Use of Time Series Forecasting Models ........... 135

EDITORIAL NOTE

While many developed and developing countries are still suffering the

negative impact of the global economic and financial crisis in terms of

continuous slowdown of economic growth and high unemployment rates,

the development of various sectors at international, regional and national

levels seems to be still struggling.

In this current issue of the Journal of Economic Cooperation and

Development – September 2016, six valuable articles have been selected

that analyse global oil prices, the importance of labour productivity in the

manufacturing sector, the relationship between government debt and

economic growth, deforestation rates, social control in procurement, and

bilateral trade and focus on trends in some Asian countries such as

Bangladesh, India, Indonesia, Malaysia, Oman and Russia.

The first article examines and explores the sharp fluctuations in global oil

prices and their associated impact on global economic imbalances which

have contributed to the renewed debate among the policy makers regarding

the nature and extent of these fluctuations. It also investigates the impact of

oil prices on the small open economy of Oman.

The second article emphasizes that the improvement and strengthening of

labour productivity has become an important approach to accelerate the

growth of the manufacturing sector in Malaysia. It also attempts to analyse

the impacts of the entry of foreign workers on the labour productivity of the

manufacturing sector in Malaysia stressing the fact that the contribution of

foreign labour on labour productivity is smaller compared to the local

labour.

The third article investigates the real effect of government debt on

Malaysia’s economy stating the fact that there is a long-run relationship

between federal government debt and economic growth in Malaysia and

that there is also an evidence of a non-linear relationship between the

federal government debt and economic growth, which suggests the optimal

level of debt that the government should hold.

The fourth article is meant to empirically demonstrate the inverse U-shaped

relationship, which is generally called the environmental Kuznets curve

(EKC), between economic development and deforestation rate in Indonesia.

Results support the long-run inverted-U relationship, which implies that,

while the deforestation rate increases at the initial stage of economic

growth, it declines after a threshold point.

The fifth article is a result of research on the effect of environment on the

development of information system of social control in procurement.

Informatization process of social relations largely determines how certain

trends of social activity will be popular and durable. To determine the

degree of maturity of information content on the theme of social control in

procurement in Russia, the study also used the method of content analysis

of information resources.

The sixth and last article sheds light on bilateral trade between India and

Bangladesh which will be mutually beneficial to both countries and

improve welfare as per trade theory. It also tries to forecast impact of trade

between two countries considering the time period 1991-2014. By engaging

in bilateral trade with India, Bangladeshi producers and suppliers ought to

be concerned about attaining long term sustainability in their business, by

improving quality of the products so that export can be raised in a

competitive manner. This will help to promote and nurture bilateral trade

relations, ensure sustainability of business and mutually benefit both the

countries through free trade agreement.

Amb. Musa KULAKLIKAYA

Editor-in-chief

Journal of Economic Cooperation and Development, 37, 3 (2016), 1-28

Oil Price Effects on Exchange Rate, Output and Consumer Price:

A Case Study of Small Open Economy of Oman

Ahmed Nawaz Hakro1 and Abdallah Mohammed Omezzine

2

Sharp fluctuations in global oil prices and their associated impact on global

economic imbalances have contributed to the renewed debate among the policy

makers regarding the nature and the extent of these fluctuations. This study is

designed to investigate the impact of oil prices on the small open economy of

Oman. Structural Vector Auto Regressive (SVAR) model has been adopted to

trace the dynamic inter-relationships among the key macroeconomic variables.

Evidence suggests that changes in crude oil prices significantly affect output,

external balances and the monetary and fiscal variables. The external shocks

induced by positive changes in global oil prices likely affect the demand

management policies in the short and long-run. In the long-run, changes in oil

prices determine the output and subsequent fiscal and monetary policy

changes, while in the short-run, fluctuations are contained well through

demand management policies. Continuation of expansionary fiscal and

monetary policies may likely contain the effects of imported inflation.

However, in the long run, over reliance on expansionary policies may less

likely to be a feasible option.

1. Introduction

Oil price shocks have significantly shifted the wealth of nations; induce

huge windfalls and external imbalances for both oil importing and

exporting countries (see, for example, Coudert et. al., 2008). The impact

of shocks and their associated effects on output, consumer prices and on

external balances have been recognized by a number of scholars

(Schneider, 2004; Setser, 2007; Roubini and Setser, 2004; Allsopp,

2006, among others).

1 Associate Professor, University of Nizwa, P.O Box 33, Post Code, 616, Nizwa, Sultanate of

Oman. E-Mail: [email protected] 2 Professor, University of Nizwa, Sultanate of Oman

2 Oil Price Effects on Exchange Rate, Output and Consumer Price:


Theoretically, in fixed exchange rate economies, oil price shocks are

transmitted to exchange rates through terms of trade channel. A typical

positive oil price shock induces the consumer prices of imported and

non-traded goods in domestic economy and appreciates the real

exchange rates. Governments in these situations usually anticipate wage-

price spiral cycles. The inflationary expectations are being countered by

resorting towards expansionary fiscal policy measures, such as, price

subsidies and wage adjustments. Increasing inflationary pressure and

appreciation in real exchange rate are usually the compelling conditions

for turning the real interest rates into negative zone. This complicates

the conduct of fiscal and monetary policies. The use of expansionary

fiscal or monetary policies in these situations turns to be a riskier option

(expansionary fiscal policy at the times when it requires containing the

inflationary expectations, expansionary monetary policy may aggravate

the prices). A fall in oil prices may have a reverse effect such as loss in

government revenues, lower government spending or a situation of

disinflation and a rise in real interest rates. A restrictive monetary policy

could put the growth objective in danger.

The small oil-based open economy of Oman is an interesting case study

in this context. Oman is known as one of the impressive success stories

in the Gulf and in the Arab world, despite possessing relatively smaller

resources as compared to its neighbours. With a consistent high growth,

lower level of inflation and stable external account surpluses, Oman has

achieved a significant progress on the economic front. The economic

growth primarily is driven by its hydrocarbon sector. Nominal GDP is

roughly 80.5 billion of US dollars in 2014. The current account balance

(percentage of GDP) is 10.6 percent with a global rank of 15. Table 1

refers to the average trends in the major macroeconomic variables from

2011-2014. Most economic indicators show impressive trends in last

few years. Real GDP growth is 4.4 percent on average for last four

years. Consumer price index is around 2.25 percent on average. Fiscal

balance is 6.2 percent of GDP and current account balance is around

12.7 percent of GDP on average.

Journal of Economic Cooperation and Development 3

Table 1: Oman’s Economic Performance since 2011-2014

Variable Average

2011-14 2011 2012 2013 2014

Real GDP (annual change, percent)

4.4 4.0 5.7 4.8 3.4

Nominal GDP (billions of US dollars)

57.1 67.7 75.4 77.1 80.5

CPI (year average; percent) 4.0 3.2 2.8 0.3 2.7 Broad Money Growth (annual change; percent)

39.3 36.6 37.2 39.7 43.8

Fiscal Balance (percent of GDP)

6.2 9.4 4.6 8.1 3.0

Government Debt (percent of GDP)

6.8 5.6 6.2 7.3 8.1

Current Account Balance (percent of GDP)

12.7 15.8 13.3 11.9 9.9

Nominal effective exchange rate index (end of the period average)*

101.3 96.5 99.5 103.9 105.3

Sources: International Financial Statistics (IFS), World Development Indicators (WDI)

and Ministry of National Economy (MONE) Sultanate of Oman. Central Bank of

Oman (CBO) Sultanate of Oman

*https://www.quandl.com/#/data/WORLDBANK/OMN_NEER-Oman-Nominal-

Effecive-Exchange-Rate http://mecometer.com/whats/oman/gdp-per-capita-ppp/

Figure 1 indicates that trend in real GDP growth is steep during the last

three decades. The GDP per capita (PPP) of Oman is US$29,800 with a

global rank of 43.

Figure 1: Real Gross Domestic Product of Oman (1980-2010



Figure 2 indicates the growth in GDP per capita since 1980. The trend

shows that GDP per capita is increasing on average at around 9-10

percent. Oman’s exchange rate is pegged with the US dollar and has

long been maintained at 0.35 Omani riyal to 1 US $ from 1975 till 1985

and thereafter, at 0.38 riyals to 1 US $ from 1986, and it remains stable

since long. The officially declared purpose of the peg is to maintain the

price stability in the country, apparently fixed exchange rate regime

which is linked to US interest rates3.

Figure 2: Growth Trend of GDP per Capita since 1980

Year (1980-2010- 31 observations)

However, the global economic trends are changing. In particular, the

changes are frequently occurring in the real value of US dollar and real

oil prices, which have continuously been affecting the business cycles of

both the oil exporting and importing countries as well. In these

circumstances, continuation of fixed or pegged exchange rate policy or

dollar pegging of Omani riyal has widely been questioned. The

continuation of pegging of Omani riyal with dollar may be a suitable

policy option to anchor the exchange rate fluctuations in short run, but at

least it may be a less feasible option in long run. Since the commodity

prices are traded in dollar, it is often noticed when real oil price rises the

real value of dollar declines. The oil exporting economies and their

dollar pegged currencies are usually appreciated in these situations.

3 Central Bank of Oman (CBO) has indicated that there are no plans to drop the peg to

the US dollar; fiscal policy will remain the main tool to curb inflation (Times of Oman,

16 March 2008).


The divergence and deviations in oil and dollar prices are likely to be

persistent in global economic trends in the medium and long run.

Therefore, countries with pegged exchange rates, may likely observe

fluctuations in their currencies in real terms.

It is, therefore, perceived that short run changes in global real oil prices

are likely affecting the domestic economy. These changes have usually

been perceived through terms of trade channel. Changes in tradable

prices are apparently channelled through real exchange rate

appreciation or depreciation, which in turn, affects the real interest rates,

government consumption expenditures and the domestic non-tradable

consumer prices. It indirectly affects the aggregate demand. If the price

shocks remain persistent, Oman economy may likely experience

positive terms of trade with exchange rate appreciation. Generally, the

oil price shock is assumed to pass-through the channels of real exchange

rate, terms trade, commodity prices to fiscal and monetary variables in

second round effects.

Consequently, it is very important to investigate the continuation of

existing policy options and to understand the impact of oil price shocks

on the real exchange rate, output and prices of small open economy of

Oman. This investigation aims to examine the relationship of changes in

oil prices with domestic economic dynamics of Oman. The study is an

attempt to establish the extent of inter-linkages of external and domestic

structural variants and their inter relationships in dynamic model.

Rest of the study is organized as follows, section two reviews the

relevant literature, section three discusses the methodology, while

section four focuses on the model estimation and presentation of results

and section five consists of conclusion and policy recommendations.

2. Literature Review

Coudert, Couharde and Mignon (2008) by compiling recent evidences

on the link between real effective exchange rate (REER) and commodity

terms of trade, establish the long run elasticity between the two, which is



around 0.5 on average4. Korhonen and Jurikkala (2007) reveal that the

price of oil has a significant and positive effect on real exchange rates

for OPEC and three oil producing commonwealth of independent states.

Habib and Kalamova (2007) in a study on Russia, Norway, and Saudi

Arabia indicate a long run relationship between real oil price and real

exchange rate but only for Russia.

Besides the oil price shocks, foreign output shocks also play an

important role in business cycle fluctuations in most developing

counties. In this context (Mendoza, 1995; Kose, 2002; and Kim et. al.,

2005) conclude that most of the business cycle fluctuations in aggregate

output are largely explained by external shocks. Hoffmaister and Roldos

(1997) and Ahmed and Loungani (1999) conclude that external output

and oil price shocks play an important role in business cyclical

fluctuations in developing countries. Hahn (2003) results suggest that

the size and the speed of the pass through in Euro zone area appear to be

robust over the time under different identification schemes. Similar

work of McCarthy (2000) indicates that exchange rates have modest

effects on domestic price inflation while import prices have stronger

effects. Pass-through is larger and has a prominent role in the inflation

process in countries with a larger import share and more persistent

exchange rates. Bems and Filho (2009) discover strong links between

real exchange rates and the terms of trade but with limited explanatory

power, while current account variable fits in the data well for oil

exporting countries. Authors use the price related methodologies

suggested by (Bayoumi, Tamim et. al., 1994; Williamson, 1994; Isard

and Faruqee, 1998; Abiad et. al., 2009). Bahamani-Oskooee and Kutan

(2008) examine the impact of exchange rate devaluation and

depreciation on output in context of nine emerging economies of the

Eastern Europe. They explain that economies which are relatively small

and heavily open depend on export revenues to promote their economic

growth, exchange rate devaluations affect their economic growth

negatively. Ito and Sato (2006), on East Asian countries after financial

crises, suggest that exchange rate depreciation results in higher rates of

inflation, especially in Indonesia. These studies establish the channels

4The evidences are based on the studies (Amano and Van Norden, 1995; Chen and

Rogoff, 2003; MacDonald and Ricci, 2001; Cashin et. al., 2004; Ricci et. al., 2008) .


and pass through links between the global oil price and external shocks

with the domestic macroeconomic dynamics of stated economies.

A number of other studies which determine the global oil price shocks

related to Middle East and North African (MENA) region countries are

also conducted. Hirata, Kim and Kose (2004) recognize a substantial

fraction of cyclical fluctuations for MENA region countries. They find

that 60 per cent of variation in aggregate output, domestic productivity

shocks explain close to 40 per cent of business cycle variation in

aggregate output for the region. Spending shocks and world interest rate

shocks are important in accounting for the volatility of business cycles

in certain macroeconomic variables. Makdisi, Fattah and Limon (2006)

suggest that MENA economies are quite vulnerable to exogenous shocks

associated with the terms of trade fluctuations, as these economes are

heavily dependent on export revenues of their primary products. Shahin

and El-Achkar (2010) study the impact of exchange rate policies on

price stability in eighteen MENA region countries from 1975-2005.

They find exchange rates along with monetary variables such as money

growth and lag inflation are the contributing variables to lower

inflation5. Bhattacharya (2003) study concludes that a lack of evidence

to account for the impact of exchange rates on real wage and relative

price flexibility and the difficulty in finding the substitute for exchange

rates as a nominal anchor. Jbili and Kramarenko (2003) in their analysis

clarify those different results that suggest the choice of exchange rates

for Lebanon and Jordan. Ghosh, Gulde, Ostry and Wolf (1997) state the

relationship between the nature of exchange rate regime inflation and

economic growth. The results indicate that inflation is lower and stable

under pegged exchange rate arrangements than in floating regimes.

Money and output growth are highly significant, whereas the interest

rate term is very insignificant. Clarida, Richard, Gali, Jordi, Gertler and

Mark (1999) investigate the terms of trade impact on exchange rate for

commodity and oil exporters’ case and reveal that real exchange rates

co-move with commodity prices in the long run and the response to oil

5 The authors find no significant link between exchange rate regime and inflation of

any peg periods for industrialized countries.



prices is somewhat lower than the commodity prices6. Coudert,

Couharde and Mignon (2008) estimate a long term relationship between

real effective exchange rate and economic fundamentals. Their results

demonstrate that real exchange rates co-move with commodity prices in

the long run and respond to oil prices somewhat less than the

commodity prices. Hakro and Omezzine (2010) also measure the global

food and oil price shocks and consequent macroeconomic implications

for Oman economy. They discover that external oil and food price

shocks significantly affect real exchange rate, consumer price index and

other macroeconomic variables of Oman economy.

Edward (1989) widely quoted study raises the crucial question about the

nature of link between the fluctuation in the exchange rate with the

output in short run or long run. By using 12 developing countries data,

the study regresses the real GDP on nominal and real exchange rate

along with other macroeconomic variables. He finds mixed evidence

which indicates that initial contractionary effects could be reversed after

some time. Agenor (1991) in a similar study on 24 developing countries

reveals surprisingly that real exchange rate depreciation actually boosts

output growth while depreciation of real exchange rate has, in fact, a

very constructive effect. Morley (1995) on 28 developing countries

notes that depreciations in the real exchange rate value at level reduce

output over a period of two years. Gala and Lucinda (2006) argue that

the productivity differential may have an important role on the impact of

real exchange levels on per capita real income growth rates. These

results show that 10 percent real exchange rate devaluation given

everything else being constant average growth rates could be higher by

0.122 per cent. Rodrik (2008) suggests that undervaluation (a high real

exchange rate) estimates result in economic growth. This may be the

case particularly in developing countries where tradable goods suffer

disproportionately from the distortions that keep poor countries from

converging. However, Kamin and Klau (1998) estimate the impact of

devaluation on 27 countries and find no evidence of contractionary

impact in the long run, contradicting the conventional view that

devaluations are expansionary.

6 Aizenman and Chrichton (2006) evaluate the impact of international reserves, terms

of trade shocks and the capital flows on the real exchange rate (REER). The major

effect is on the Asian and oil exporting countries.


Apart from the above studies, a large number of other set of studies

conducted, support the proposition that exchange rate shocks lead to

negative effect on output, Particularly for Latin American countries (for

example; Rogers and Wang, 1995; Santaella and Vela, 1996; Copelman

and Warner, 1995; and Kamin and Rogers, 2000; Rodriquez and Diaz,

1995). similar results are found for Peru, and in Hoffmaist and Vesh

(1996) for Uruguay. There is hardly any study which suggests high

depreciation combined with high level of output and high appreciation

of exchange rate with high level of depressed output (Kamin and

Rogers, 2000).

Most of the studies have used the Vector Auto Regression (VAR)

mechanism to find the inter-relationship between exchange rates with

output and prices in different countries contexts. Case studies such as

Ndung’u (1993, 1997), by estimating a six variable VAR on the data set

of Kenya, discover the link between the rate of inflation and exchange

rate explaining each other. Montiel (1989) by using VAR model for

Argentina, Brazil and Israel observes that exchange rate movements

explain inflation. Dornbush et. al. (1990) find that real exchange rate is

an important source of inflation in Argentina, Brazil, Peru, and Mexico

but not in Bolivia. Inflation seems to be inertial with regard to exchange

rate and is being determined through demand shocks. Exchange rate and

inflation are also studied in several other countries context (see e.g.,

Kamin, 1996; Odedokum, 1997; London, 1989; Cannetti and Greene,

1991; Calvo et. al., 1995; Elbadawi, 1990).

The available evidence is quite rich in its content and methodological

rigorousness. It addresses and formulates the impact and channels

through which the oil price induces the changes in domestic dynamics.

To the best of authors’ knowledge, hardly any significant attempt has

ever been made towards the understanding of these channels in the

context of small open economy of Oman. This study fills the gap.

3. Methodology

Structural Vector Auto Regression (SVAR) model is considered a very

useful approach to find the link between the oil price shocks and the

domestic economy. The SVAR has a number of advantages. It identifies

the structural shocks through innovations with identifiable restrictions



and thereafter, generalizes the impact through impulse response

functions and variance decompositions, capable to trace out individual

shocks and variances on each variable. The study considers the SVAR

model initially with 6-7 variables based on theoretical relationship

among the set of variables in a VAR system. The choices of variables

are based on the procedure adopted to incorporate the vector of

endogenous variables used by (McCarthy, 2000; and Hahn, 2003).

Supply shocks are identified and derived from changes in global oil

prices and through changes in external balances. The demand side

effects traced in changes in real output growth, changes in government

consumption expenditures. Interest rate, international prices, real

effective exchange rate, money supply and interest rate variables are

used to allow the effects of monetary policy and fiscal policy responses

on inflation and output. Real effective exchange rate is used primarily to

avoid bilateral exchange rate vis-à-vis the US dollar. This will give us a

leverage to measure the extent to which the country’s trade dependence

with other countries. The use of this variable shall provide the real

currency appreciation or depreciation or a gain or loss in the price

competitiveness. It also provides the extent to which the country is

facing the inflationary pressure through imports. CPI is used for

domestic inflation.

The structural model is identified by imposing zero restrictions on the

number of endogenous variables based on the theoretical relationship of

the endogenous variables. The changes in oil prices are ordered first

because the oil price variable is likely to affect all the other variables in

the system. Real output gap is placed second and the interest rate and

money supply (M2) is ordered fourth. The use of money supply in place

of interest rate turns out to be a better choice. It is more reasonable to

measure monetary policy shocks to contemporaneous effects after

exchange rate variants or pass through on monetary variables reflected

after changes in oil and output variables. The policy reaction function is

assumed to flow from the changes in oil prices to GDP gap,

subsequently affects the monetary and fiscal variables. Real effective

exchange rate variable is placed ahead of government consumption

expenditures and domestic prices. This implies that the real effective

exchange rate responds contemporaneously to supply and demand

management policies of the government towards containing the effects

of the shocks.


3.1 The Model

Based on the above interrelationships among the set of variables the

vector of economic variables is expressed in a dynamic framework. The

framework is expressed on the similar lines expressed by (Vinh and

Fujita, 2007; Blanchard and Watson, 1986; Bernanke, 1986; Odusola

and Akinlo, 2002; Hen, 2003; and McCarthy, 2006 among others). The

methodology indicates the specified and identified restrictions on each

equation to estimate the joint behaviour or dynamic nature of

interrelationship through time of vectors of economic variables. The set

of variables is written;

n

Yt = Σ Аi Xt + βut (1)

i=0

Here, Yt is defined as a set of vector of observation of dependent in an

(n x 1) vector at time t, Аi is the matrix of coefficients ( n x n) X

vector of lagged values. The ut is a disturbance term of (n x 1) vector in

the system and β is an (n x n) vector of matrix of coefficients of

disturbances to the dependent Yt vector. For reduced term it is

expressed as;

N

Yt = Σ Ci Xt-1 + εt (2)

i=0

εt = Gut,

Where, vector of Ci = (1- А0)-1

, and vector of Аi and G= (1-А0)-1

β

matrices of coefficients and disturbances are estimated. Whereas, the

restriction on the structure of coefficient matrices А0 and β disturbance

matrices are imposed in order to derive the policy analysis. The

structural disturbances (ut ) are derived from the reduced form equation

2 (εt), β matrices are assumed to be a diagonal matrix and whereas А0

is always used to be lower triangular. This is a direct causal ordering of

the variables. One can directly relate the structural and reduced format

in the form of:

ut = β -1

(1- А0) εt (3)



When β is an identity matrix, it would be easy to calculate the structural

disturbance terms when the system has enough information to estimate

the non-zero elements of the А0 matrix along with the unknown

variances of the vector ut . The available information consists of n

n(n+1/2)/2. Distinct sample covariance is derived from the covariance

matrix of the residuals of reduced form. Naturally А0 is a lower

triangular matrix and the β is an identity matrix. One can easily

considers as an order condition of identification of non-zero elements of

А0 which must not exceed n (n-1)/2, the condition of degrees of freedom

when the number of structural disturbances is calculated.

The process involves the estimation of innovations in unrestricted VAR

first, and thereafter, the identified postulating structure for A0 is to be

estimated by using the estimation method of moments as suggested by

(Bernanke, 1986; Odusola and Akinlo, 2001). This can be expressed as;

Ŝ = (1-Ã0) M (1-Ã0)-1

(4)

Where Ŝ = μμ’ and M = (Σέεtέεt)/T which is an estimated covariance

matrix of shocks and sample covariance matrix of the residuals,

respectively. The matrix Ã0 element is diagonal matrix of fundamental

shocks of Ŝ. This matrix as earlier defined as an n(n+1)/2 distinct

elements of the symmetric matrix M. Therefore, with number of

equations variances to estimate in an identifiable system of n (n-1)/2

nonzero elements of the Ã0 matrix is going to be estimated. To best

capture the stylized structure of fundamentals of Oman economy, the

necessary identification scheme is identified.

3.2 The Reduced Form Model

The identification scheme in equation 5 table 2 indicates specified

restrictions which are adopted in order to capture the joint behaviour of

interrelationship among the set of variables and their innovations.


Table 2: Reduced Form Model

u OCOP OCOP 1 0 0 0 0 0 ε OCOP

(5)

u RGDP RGDP α21 1 α23 0 0 0 ε RGDP

u REER = REER + α31 0 1 0 α35 0 = ε REER

u M2 Int rate/M2 0 α42 α43 1 α35 0 ε M2

u GCE GCE 0 α52 α53 0 1 0 ε GCE

u CPI CPI 0 α 63 α64 α65 1 ε CPI

The specifications are indicated as constants of the six vector variables

and αij represent the coefficients. As reflected in the equation, the

innovations in the crude oil prices are entirely due to its own shock

innovations and do not necessarily depend upon the innovations from

any other variable in the model. Innovations in the RGDP depend on its

own innovations and the innovations stem out of oil prices and real

effective exchange rate.

Since Oman is a small open economy, its output is largely determined

by the strength of its commodity exports. Therefore, any change in

output is stemed out of the changes in crude oil price shall reflect

through the strength and weakness of its real exchange rate.

3.3 Data Sources and Construction of Variables

International Financial Statistics (IFS) annual data series from 1980 to

2010 is used for the estimation. The construction of variables,

measurements and data sources are listed in table 3. All variables are

taken into logarithm form and measured at lag difference of the actual

values and the inferences to be drawn from the VAR. The variables

could be sensitive. The particular sensitivities are apparent in the level

of first differences or in inclusion and exclusion of time trend. Time

series properties are carefully evaluated.



Table 3: Summary of Variables, Measurements and Data Sources

Name of

variable

Variable

Measurements

Explanation Data Sources

Oman Crude

Oil Prices

(OCOP)

Proxy for Real

average global oil

prices (base year

2005 prices) in $

Crude oil prices are connected

with the external balances

International

Financial

Statistics (IFS)

annual data

series from

1980 to 2010

Output

R(GDP),

Real Gross

Domestic Product

Adjusted with normal GDP

with deflator base year 2005,

Nominal GDP was obtained

from IFS 2005 base prices

IFS 1980-

2010

Money

Supply (M2)

Money Supply,

Nominal Money

supply in millions

of Rials

Monetary variable IFS 1980-

2010

Real

Effective

Exchange

Rate (REER)

The real effective

exchange rate is

the measure of

price adjusted

trade weighted

exchange rate. The

trade weights are

calculated from

the relative trade

share (imports +

exports) taken

from the direction

of trade statistics.

The total weight in particular

year is equal to one. REER is

constructed by taking the trade

weighted share of trade with

selected trading partners equal

to one adjusted with respective

whole sale prices of trading

partners with CPI of Oman

adjusted by using 2005 base

prices (increase in the value of

real effective exchange rate

shows the real depreciation,

decrease shows the real

appreciation. REER is defined

based on IMF defined

methodology.

Direction of

trade statistics.

1980-2010

Consumer

Price Index

(CPI),

Commodity price

index is a proxy

measure of

domestic inflation

Adjusted with 2005 base year.

Consumer price index adjusted

in order to get the simple

index-missing data is adjusted

with GDP deflator trend in CPI

IFS 1980-

2010

Government

Consumption

Expenditures

(GCE)

Percentage of

GDP

Government consumption

expenditure is the government

consumption expenditure-

proxy for fiscal policy

IFS 1980-

2010

Short term

interest rate

Short-term

discount rate end

of the period

Short term discount

rate/deposit rate as proxy for

monetary policy variable.

IFS 1980-

2010


4. Estimation and Result Discussion

Augmented Dickey Fuller (ADF) and Philips- Perron (PP) tests at level

and first difference are conducted. The first differencing variables are

integrated at different orders. Variables such as Real Gross Domestic

Product (RGDP), Government Consumption Expenditures (GCE) and

Consumer Price Index (CPI) variables are adjusted as rates of change

rather levels. Other variables such as Average Oman Crude Oil Price

(OCOP), Money Supply (M2) and Real Effective Exchange Rate

(REER) are used as level. Since the data series are in annual form, no

seasonal trend is observed and this has been verified and checked. Serial

correlation test is performed by using Lagrange Multiplier (LM)

statistics to check the robustness of ADF tests. Lag lengths are

determined by using Akaike Information Criterion (AIC).

The VAR model is estimated with log differences of six variables by

using yearly data with two lags in each equation. The two year lag

period is estimated over the period 1980-20107. Most of the variables

are found to be cointegarated. Therefore, the Vector Error Correction

model is used. The model allows the long term behaviour of the set of

endogenous variables to converge and cointegrating the long term

equilibrium relationship along with their short term dynamics. The

cointegration relationship is tested by using Johansen Cointegration Test

(1995). Four co-integrating vectors linking each other are found.

Cointegrating vectors are tested based on the trace and max-eigen

statistics. The cointegration relationship among the variables is

presented in the panel of table 4.

7 Charemza and Deadman (1992) suggest use of cointegrating vector for REER

variable, when it is used in long series, in short series, the variable may turns out to be

inconsistent in theoretical terms.



Table 4: Summary of the Reduced Form Estimation

The coefficient signs indicate the adjustments to the long run deviations.

Unlike the oil prices, the domestic prices are adjusted from a long run

path in a relatively longer time period than the adjustment in

goveronment expenditure which is relatively quick to adjust from its

deviations. The adjustment coefficient corresponds -0.04 per cent for

exchange rate, while for government consumption expenditure

coefficient is -0.288 and for real output, it is -0.171. Coefficients

suggest that government expenditure is relatively quick to adjust to the

equilibrium path, as compared to real output and real oil prices.

However, the real output responds to the external shocks relatively

slower than the response to government consumption expenditure. The

adjustment process indicates that output is positively responding to the

external shocks. Relatively slow adjustment in real oil prices and

domestic prices to equilibrium path indicate the long run relationship

among the two variables. Oman economy is a small economy, its

exportable share of commodities volume is very small compared to the

world demand. It seems less likely that oil prices of Oman crude affects

the global oil prices.

4.1 The Structural Model and Impulse Response Functions

The SVAR model suggested with structural restrictions in equation 5

and table 2 is estimated by using the VEC model. The results are

presented in the table 5. Structural response function of SVAR in a form

of coefficients in table 5 and Figure 3 indicates the generalized impulse

response functions.

Test statistics Average

Crude

Price of

Oil

RGDP Money

Supply

M2

REER Govt.

Cons Exp.

CPI

Co integration

equation

-0.04

(-0.34)

-0.171

(-2.45)

0.613

(2.66)

5.796

(3.30)

-0.288

(-2.34)

-0.005

(-0.06)

Goodness of fit statistics

Adjusted R2 0.46 0.21 0.52 0.51 0.64 0.18

SEE 0.02 0.00 0.08 4.77 0.02 0.01


Table 5: Structural VAR Regression Impulse Response Function of

SVAR ε OCOP = 0.05μ COP

(7.4)***

6.1

ε RGDP = 0.38 OCOP

(4.14)***

6.2

ε REER = 0.09 OCOP

(0.02)

4.50GCE 6.3

(0.47)

ε M2 = -2.67RGDP

(7.88)***

-0.02REER

(2.09)**

-0.25GCE 6.4

(2.1)**

ε GCE = +1.89RGDP

(3.85)***

-0.04REER

(1.51)

6.5

ε CPI = -1.34RGDP

(-3.13)***

-0.002REER

(0.88)

-0.32M2 6.6

(-2.4)***

The coefficients in table 5 are structural impulse response functions of

SVAR. These coefficients are driven from the vector error correction

and cointegrated series with structural restrictions. These coefficients

may not appear to be very precise. This may be possible because of the

estimation techniques and the nature of standard errors as cautioned by

(Bernanke, 1986; Calomiris and Hubbard, 1989, Turner, 199; Kiguel,

Lizondo and O’Connell, 1997).

4.2 Discussion

The results suggest that innovations in crude oil prices positively affect

the real output innovations (refer equation 6.1 in table 5). One

percentage point variation in real crude oil prices significantly impact

the real output by 0.38 per cent in the same direction. This means every

10 per cent increase in oil prices positively influence the real output by

3.8 per cent in long run. The innovations in real output also significantly

impact the changes in the money supply. However, these changes are

inconsistent with changes in real output growth. In the long run, shocks

in both real output and money supply become smooth. Changes in

output are primarily driven by changes in oil prices, whereas the

changes in output induce increase in money supply.



Figure 3: Response to Generalized One S.D Innovations to 2S.E


One percentage point increase in real Oman crude oil prices affects the

real exchange rate by 0.09 per cent, but this effect is statistically very

insignificant. This suggests that changes in global oil prices, in fact, do

affect the real exchange rate, but that effect is a very minor. Innovations

in real effective exchange rate affect the government consumption

expenditures positively but insignificantly. This result is consistent with

the trends in government consumption expenditures which are usually

increasing at the time of positive external balances. Innovations in real

exchange rates are having a negative and negligible impact on money

supply expansion. Changes in real output negatively affect the money

supply and it is quite clear that real output growth induces by the

external trade revenues but the output is not influenced by the changes

in domestic money supply. Innovations in real output infuence the

government consumption expenditures significantly and positively. One

percentage point increase in real output increases the government

consumption expenditure by 1.89 per cent. Innovations in output growth

affect the fiscal side of government consumption expenditure positively

and significantly. Real appreciation in exchange rate has a negligible

effect on prices. It means that real appreciation of exchange rate effect

on domestic prices is nominal. Because the domestic prices are not

much affected by appreciation or depreciation in real exchange rate. In

figure 3 response to generalized one standard deviation innovations to

two standard errors suggest consistency with the structural coefficients

in table 5. The results suggest negative response of external balances to

a real effective exchange rate. While real effective exchange rate has a

positive effect on real interest rate.

Pegged exchange rate policy seems to be performing well in anchoring

the inflationary expectation in Oman.. However, the changes in real

output significantly affect the prices negatively. Monetary expansion has

a negative effect on domestic prices. This also suggests that

expansionary monetary policy traces the pace of output and balances the

inflationary expectations. Oman has relatively stable prices in last few

decades. Every one per cent increases in money supply adjust the pace

of inflationary expectations by 0.32 per cent. Increase in output is

accompanied by decline in prices. This suggests that stabilization

policies works well with a mixture of expansion in output growth

coupled with monetary expansion accommodating the price changes.



The results are in line with the established theoretical propositions and

the evidence discussed earlier. The innovations in real effective

exchange rate not necessarily affect the consumer prices. However,

innovations to real output affect prices negatively. The relationship of

output innovations with government consumption expenditure positively

affect output which leads to an expansionary fiscal policy.

5. Conclusion

The study has examined the relationship of real oil prices shocks and its

channels to domestic macroeconomic dynamics of small open economy

of Oman. It has established the extent of interlinks of external and

domestic structural variants and their inter relationships. The dynamic

structural vector autoregressive econometric model along with stylized

structural variants of Oman economy has shed lights on the

interrelationship of macro variables and trace out their dynamic

moments. The results are consistent with the theory and provide the

important insights into the policy directions.

The results imply that crude oil price emerges to be a very significant

variable inducing output extends the monetary and government

consumption expenditures. In other words, changes in oil prices induce

output and consequent response to fiscal and monetary policy responses.

Monetary and fiscal policies anchor the global price shock and the long

run inflationary pressures. Monetary variables such as money supply

responses to changes in output and affects prices. The changes in

exchange rate are presumably positive shock to terms of trade or

external favourable effect. Results from impulse response functions

imply that innovation in output is influenced by oil prices. Exchange

rate, in fact, do appreciates due to increase in oil prices; the response to

fiscal expansion is derived out of changes in oil prices which is

straightforward theoretical result. Structural impulse response

coefficients indicate consistency with the earlier results.

Output and demand management policies in Oman economy are largely

dependent on the external factors, particularly oil prices. The shocks to

oil prices may likely affect the demand management policies. It seems in

the long run changes in oil prices determine the output and subsequent

fiscal and monetary policies which serve well to contain the inflationary


expectations and maintain positive external balances. In short run

fluctuations in oil prices or global imbalances though are contained well

in mixture of stabilization policies, continuation of a mixture of fiscal

and monetary stabilization policies may serve well the purpose of the

domestic dynamics of Oman economy. However, in the long run, over

reliance on stabilization policies in days of global imbalances may

provide fewer options to contain the external shocks. Therefore, the

exchange rate policy of pegging with dollar may be questionable in the

longer term.



References:

Abiad, A. G., Balakrishnan, R., Koeva Brooks, P., Leigh, D., & Tytell, I.

(2009), "What’s the damage? Medium-term output dynamics after

banking crises". IMF working papers, 1-37.

Agénor, P. R. (1991), "Output, devaluation and the real exchange rate in

developing countries". Weltwirtschaftliches Archiv, 127(1), 18-41.

Ahmed, S., & Loungani, P. (1999), Business cycles in emerging market

economies. Manuscript, IMF and Board of Governors of the Fed.

Aizenman, J., & Riera-Crichton, D. (2008), "Real exchange rate and

international reserves in an era of growing financial and trade

integration," The Review of Economics and Statistics, 90(4), 812-815.

Allsopp, C. (2006), "Why is the Macroeconomic Impact of Oil Prices

Different this Time?," In Oxford Energy Forum (Vol. 66, p. 20). The

Oxford Institute for Energy Studies.

Amano, R. A., & Van Norden, S. (1995), "Terms of trade and real

exchange rates: the Canadian evidence" Journal of International Money

and Finance, 14(1), 83-104.

Bayoumi, T., & Symansky, S (1994), The Robustness of Equilibrium

Exchange Rate Calculations to Alternative Assumptions and

Methodologies, (eds) John Williamson, equilibrium exchange rates.

Peterson Institute, 1994,pp. 19–59.

Bahmani-Oskooee, M., & Kutan, A. M. (2008), "Are devaluations

contractionary in emerging economies of Eastern Europe?". Economic

Change and Restructuring, 41(1), 61-74.

Carvalho Filho, I. E., & Bems, R. (2009), "Exchange rate assessments:

methodologies for oil exporting countries", IMF Working Papers, 1-35.

Bhattacharya, R. (2003), "Sources of variation in regional economies,"

The Annals of Regional science, 37(2), 291-302.


Blanchard, O. J., & Watson, M. W. (1986), "Are business cycles all

alike?," In The American business cycle: Continuity and change (pp.

123-180). University of Chicago Press.

Bernanke, B. S. (1986), "Alternative explanations of the money-income

correlation" In Carnegie-Rochester conference series on public policy

(Vol. 25, pp. 49-99). North-Holland.

Calvo, G. A., Reinhart, C. M., & Vegh, C. A. (1995), "Targeting the real

exchange rate: theory and evidence," Journal of Development

Economics, 47(1), 97-133.

Canetti, E. (1991), Monetary growth and exchange rate depreciation as

causes of inflation in African countries: An empirical analysis. (eds)

Centre for Economic Research on Africa Research Monograph Series,

School of Business-Montclair State University, New Jersey.

Cashin, P., Céspedes, L. F., & Sahay, R. (2004), "Commodity currencies

and the real exchange rate," Journal of Development Economics, 75(1),

239-268.

Clarida, Richard., Gali, Jordi., Gertler, Mark (1999), "The Science of

Monetary Policy: A New Keynesian Perspective," Journal of Economic

Literature, 37/ 2 : 1661-1707.

Calomiris, C. W., & Hubbard, R. G. (1989), "Price flexibility, credit

availability, and economic fluctuations: evidence from the United States,

1894-1909," The Quarterly Journal of Economics, 429-452.

Central Bank of Oman (2014), Annual Report 2014, Sultanate of Oman

Charemza, W.W. and Deadman, D.F. (1992), New Directions in

Econometric Practice, Edward Elgar: Aldershot

Chen, Yu-chin. Kenneth, Rogoff. (2003), “Commodity currencies,”

Journal of International Economics, 60/1: 133-160.

http://www.nber.org/people/mark_gertler

http://ideas.repec.org/a/eee/inecon/v60y2003i1p133-160.html

http://ideas.repec.org/s/eee/inecon.html



Copelman, M., & Werner, A. M. (1995), "The monetary transmission

mechanism in Mexico" (No. 521). Board of Governors of the Federal

Reserve System.

Coudert, V., Couharde, C., & Mignon, V. (2008), "Do terms of trade

drive real exchange rates? Comparing oil and commodity currencies,"

Centre d'Etudes Prospectives et d'Informations Internationales (CEPII),

Paris, 32.

Dornbusch, R., Sturzenegger, F., Wolf, H., Fischer, S., & Barro, R. J.

(1990), "Extreme inflation: dynamics and stabilization," Brookings

Papers on Economic Activity, 1990(2), 1-84.

Edwards, S. (1989), Real Exchange Rates, Devaluation, and

Adjustment: Exchange Rate Policy in Developing Countries,

Cambridge, Mass.: MIT Press

Elbadawi, I. A. (1990), "Inflationary process, stabilization and the role

of public expenditure in Uganda," Washington, DC: World Bank.

Gala, P., & Lucinda, C. R. (2006), "Exchange rate misalignment and

growth: old and new econometric evidence," Revista Economia, 7(4),

165-187.

Ghosh, A. R., Gulde, A. M., Ostry, J. D., & Wolf, H. C. (1997), "Does

the nominal exchange rate regime matter?" (No. w5874). National

Bureau of Economic Research.

Habib, M. M., & Kalamova, M. M. (2007), "Are there oil currencies?

The real exchange rate of oil exporting countries," Working Paper, No.

839, European Central Bank.

Hakro, A. N., & Omezzine, A. M. (2010), "Macroeconomic effects of

oil and food price shocks to the oman economy," Middle Eastern

Finance and Economics, 6(2010), 72-89.

Hirata, H., Kim, S. H., & Kose, M. A. (2004), "Integration and

fluctuations: the case of MENA," Emerging Markets Finance and

Trade, 40(6), 48-67.


Hoffmaister, A. W., & Roldós, J. E. (1997), "Are business cycles

different in Asia and Latin America?"Working Paper No. 97/9,

International Monetary Fund.

Hoffmaister, A. W., & Végh, C. A. (1996), "Disinflation and the

recession-now-versus-recession-later hypothesis: evidence from

Uruguay," Staff Papers-International Monetary Fund, 355-394.

Hahn, E. (2003), "Pass-through of external shocks to euro area

inflation," Working paper, 243, European Central Bank.

Isard, P., & Faruqee, H. (1998), "Exchange rate assessment: extension of

the macroeconomic balance approach" (Vol. 167). International

monetary fund.

Ito, T., & Sato, K. (2006), "Exchange rate changes and inflation in post-

crisis Asian economies: VAR analysis of the exchange rate pass-

through" (No. w12395). National Bureau of Economic Research.

Jbili, A., & Kramarenko, V. (2003), "Choosing exchange regimes in the

Middle East and North Africa," International Monetary Fund

Kamin, S. B., & Klau, M. (1998), "Some multi-country evidence on the

effects of real exchange rates on output," FRB International Finance

Discussion Paper, (611).

Kamin, S. B. (1996), "Real Exchange rates and Inflation in exchange-

Rate based Stabilizations: an empirical examination," International

Finance Discussion Papers, (554).

Kamin, S. B., & Rogers, J. H. (2000), "Output and the real exchange

rate in developing countries: an application to Mexico," Journal of

development economics, 61(1), 85-109.

Kiguel, M. A., Lizondo, J. S., & O'Connell, S. A. (Eds.). (1997),

Parallel Exchange Rates in Developing Countries. St. Martin's Press.



Kim, S. H., & Ahn, H. D. (2005), "Dynamics of open economy business

cycle models: the case of Korea," Korea Development Review, 1(1),

157-84.

Korhonen, I. and T. Juurikkala (2007), "Equilibrium exchange rates in

oil-dependent countries," BOFIT Discussion Papers No. 8.

Kose, M. A. (2002), "Explaining business cycles in small open

economies:‘How much do world prices matter?," Journal of

International Economics, 56(2), 299-327.

London, A. (1989), "Money, inflation and adjustment policy in Africa:

some further evidence," African Development Review, 1(1), 87-111.

MacDonald, R., & Ricci, L. A. (2001). "PPP and the Balassa Samuelson

effect: The role of the distribution sector," (Vol. 442). International

Monetary Fund.

Makdisi, S., Fattah, Z., & Limam, I. (2006), "Determinants of Growth in

the MENA Countries," Contributions to economic analysis, 278, 31-60.

McCarthy, J. (2000), "Pass-through of exchange rates and import prices

to domestic inflation in some industrialized economies," FRB of New

York Staff Report, (111).

Mendoza, E. G. (1995), "The terms of trade, the real exchange rate, and

economic fluctuations," International Economic Review, 101-137.

Montiel, P. J. (1989), "Empirical analysis of high-inflation episodes in

Argentina, Brazil, and Israel," Staff Papers-International Monetary

Fund, 527-549.

Morley, S. A. (1995),. "Structural adjustment and the determinants of

poverty in Latin America," Coping with austerity: Poverty and

inequality in Latin America, 42.

Ndung’u, Njuguna. (1993), Dynamics of the Inflationary Process in

Kenya, Göteborg: University of Göteborg.


Ndung'u, N. S. (1997), "Price and Exchange Rate Dynamics in Kenya:

An Empirical Investigation," (No. 58).African Economic Resarch

Consortium (AERC)African Economic Research Consortium, AERC

Research Paper /58.

Odusola, A. F., & Akinlo, A. E. (2001),"Output, inflation, and exchange

rate in developing countries: An application to Nigeria," The Developing

Economies, 39(2), 199-222.

Odedokun, M. O. (1997)," Dynamics of inflation in Sub-Saharan Africa:

the role of foreign inflation, official and parallel market exchange rates,

and monetary growth," Applied Financial Economics, 7(4), 395-402.

Ricci, M. L. A., Lee, M. J., & Milesi-Ferretti, M. G. M. (2008), "Real

exchange rates and fundamentals: A cross-country perspective," (No. 8-

13). International Monetary Fund.

Rodriguez, G. H., & Gazani, G. D. (1995), "Fluctuaciones

macroeconómicas en la economía peruana" Banco Central de la

República Dominicana.

Rodrik, D. (2008), "Normalizing industrial policy," International Bank

for Reconstruction and Development/The World Bank.

Rogers, J. H., & Wang, P. (1995), "Output, inflation, and stabilization in

a small open economy: evidence from Mexico," journal of Development

Economics, 46(2), 271-293.

Roubini, Nouriel, and Brad Setser. (2004), "The US as a net debtor: The

sustainability of the U.S. external imbalances," New York University.

Mimeograph.

Santaella, J. A., & Vela, A. (1996),"The 1987 Mexican disinflation

program: an exchange rate-based stabilization?" International Monetary

Fund.



Schneider, M., & Tornell, A. (2004), "Balance sheet effects, bailout

guarantees and financial crises," The Review of Economic Studies, 71(3),

883-913.

Setser, B. (2007), "The case for exchange rate flexibility in oil-exporting

economies," (No. PB07-8). Peterson Institute for International

Economics.

Shahin, Wassim and Elias El-Achkar,( 2010), "Regional Monetary

Coordination between GCC Monetary Union and Other MENA

Countries," paper presented at the workshop, "The Evolving

International Role of the GCC," Economies at the Mediterranean

Research Meeting of the European University Institute.

Times of Oman, (2008), Daily Times of Oman, Muscat, 16 March,

Muscat, Sultanate of Oman.

Turner, P. M. (1993),. "A structural vector autoregression model of the

UK business cycle," Scottish Journal of Political Economy, 40(2), 143-

164.

Vinh, Nguyen Thi Thuy, and Seiichi Fujita (2001), "The impact of real

exchange rate on output and inflation in Vietnam: A VAR approach,"

Pesaran, M., Y.Shin and R. Smith (2001). "Bounds testing approaches to

the analysis of level relationships", Journal of Applied Econometrics 16

(2007): 289-326.

Williamson, J. (1994), Estimating equilibrium exchange rates. Peterson

Institute.


The Impacts of Foreign Labour Entry on the Labour Productivity

in the Malaysian Manufacturing Sector

Nur Sabrina Mohd Palel

1, Rahmah Ismail

2 and Abdul Hair Awang

3

Improvement and strengthening of labour productivity is an important

approach to accelerate the growth of the manufacturing sector in Malaysia.

This study attempts to analyses the impacts of the entry of foreign workers on

the labour productivity of the manufacturing sector in Malaysia. The analysis

of this study employs the dynamic panel data method which combines time

series and cross-section data. The data used was from the year 1990 to 2008,

covering 15 selected sub-sectors in the Malaysia manufacturing sector. Core to

the analysis in the study is the Pooled Mean Group (PMG) estimation model.

The study found that foreign labour, local labour, capital intensity and foreign

direct investment (FDI) have positive and significant effects on the labour

productivity growth. The study differentiates between local and foreign labour

into categories of skilled and unskilled labour. The findings indicated that

unskilled foreign and local labour are negatively and significantly affect the

growth of labour productivity in the long run. Inversely, skilled local and

foreign labour had a significant and positive impact on the labour productivity

growth. However, the contribution of foreign labour on labour productivity is

smaller compared to the local labour.

1. Introduction

Labour productivity is an important, frequently emphasised element in

any sector as part of an effort to keep a sector competitive in the global

market. The growth of the industrial sector has been successful in

increasing export and this effectively led to the innovation of new

1 MEc Student, Faculty of Economics and Management, Universiti Kebangsaan

Malaysia Email:nur02862yahoo.com 2 Senior Lecturer, Faculty of Economics and Management, Universiti Kebangsaan

Malaysia Email:[email protected] 3 Senior Lecturer, Faculty of Social Sciences and Humanities,Universiti Kebangsaan

Malaysia Email: [email protected]

30 The Impacts of Foreign Labour Entry on the Labour Productivity


technologies and improvements in the internal management of firms.

Since independence, Malaysia has experienced encouraging phases of

economic development through various economic development policies

and strategies. In fact, before the Asian financial crisis in the year 1997,

the country is known as one of the rapidly growing economies on par

with the developed economies in East Asia such as Japan, South Korea,

Taiwan and Hong Kong (Ishak Yussof & Nor Aini, 2009). A rapidly

growing manufacturing sector can speed up the growth of an economy

due to its share in the Gross Domestic Product (GDP), job creation and

foreign exchange. One of the most effective ways to further improve this

sector is by improving labour productivity. The logic is simple; a higher

productivity means that the same number of inputs can produce higher

quantities of output. Sahar (2002) explains that a high labour

productivity in the manufacturing sector can help create sustainable

growth and development of a country.

Economic growth is an important determinant of national development

and these are two interrelated concepts. A country can grow without

developing, but to develop, a country needs to grow. Economic growth

requires a combination of a number of factors, including labour, land,

capital and technology being employed efficiently without waste.

Labour - local and foreign - is an important part of this equation. In

general, the rationale behind employing foreign labour is to fill the gap

create by the shortage of labour supply in the local labour market

especially in sectors like farming, construction, service, industrial and

manufacturing. This is a short term solution according to Preibisch

(2007). Nevertheless, the output generated through the foreign labour

contributes positively to the country's export.

It must be noted that foreign labour has contributed to Malaysia's

economic growth, especially with regard to the GDP and aggregate

expenditure. The government has effectively prevented excessive

increases in wages and inflation. The wages paid to the foreign workers

are considerably lower than the what the locals receive. Despite the

relatively lower wages, the foreign labour positively contribute to

increase in demand (Borjas, 2006). However, being a short-term

solution, the economy cannot continuously depend on the contributions

of foreign workers. With the country looking to become a high earning

economy by 2020, a switch of strategy is important and thus, rather than


relying on the number of inputs, more emphasis should be given to

increasing productivity of input.

An influx of foreign workers can affect an economy both positively and

negatively. Its influence on the labour market depends on its role,

whether as a substitute or a complement to the local workers. Borjas

(1993) found a negative impact of foreign workers. However, he

stressed that such negative influence is only valid for unskilled foreign

labour. The direction of the effects of foreign worker entry on a market

depends on their productivity. Highly productive, skilled foreign worker

can immensely contribute to an (Hercowitz et al. 1999). On the other

hand, unskilled foreign workers may have problems adapting and may

end up needing more help than contributing. It is worth noting that

several studies have been conducted on this topic but the results have

been inconsistent. As such, the issue is very much inconclusive. It is

therefore imperative for researchers to keep on studying the subject and

more importantly on the factors that make the results so inconsistent.

Compared to more developed economies like China, Singapore and

South Korea, our labour productivity is still relatively modest (Malaysia

Productivity Corporation, 2013). A shortage of labour during the period

of the 7th Malaysia Plan (7MP) has tightened the labour market and

pushed wages up. As part of the effort to reduce the shortage, a

programme was initiated to allow for foreign workers entry. It was,

however, at that time, unclear how the influx of skilled and unskilled

foreign workers can generate growth of the manufacturing sector in the

longer term.

Overall, the study aims to contribute to the literature on this topic

through its measurement method which was used to estimate the impact

of foreign workers entry on labour productivity. The method is known

as Autoregressive Distributed Lag (ARDL) dynamic panel test method

using Pooled Mean Group (PMG) estimation model. This method

allows the researcher to observe the relationship between the

independent and dependent variables in the short and long terms.

Furthermore, this study provides a more detailed insight by dividing the

foreign labour into skilled and unskilled categories.

The objective of this study is to analyse the short and long term effects

of foreign worker entry in general on the productivity of labour. The



study also sought to examine the influences of the use of foreign and

local, skilled and unskilled labour on the productivity of labour. This

paper is organised into five sections, namely introduction, literature

review, methodologies, research findings and the last part from this

study is conclusions and suggestions.

1.1 Trend of Foreign Labour in Malaysia

Rapid economic development has led to rapid changes in the labour

market. With demand for labour increasing more than the supply, there

was a need to address this shortage by allowing entry of foreign workers

into the domestic market. As a result, there has been a steady influx of

foreign workers from various countries such as Indonesia, India, Nepal,

Bangladesh and Filipina. Table 1 summarises the development of the

entry of foreign workers into Malaysia from the year 2007 to 2011. It is

clear, however, that while entry is still occurring, the numbers were

declining steadily. Manufacturing sector recorded the highest inflow of

foreign workers from year 2007 to year 2011 compared to other sectors.

Local Electrical and Electronics (E &E) industry was the main sector

contributing 55.9 percent of the country's exports and employs 28.8

percent of the national labour force (Prime Minister's Department,

2012). This industry has also successfully developed the ability and

skills for the manufacturing sector, consumer electronics, and electronic

and electrical components. Productivity generated by foreign labour has

helped increase total exports and consequently contribute to the surplus

in the balance of payments. Strong export revenue growth is directly

driven by high export value. Therefore, the enhancement and

strengthening of productivity is one approach that can be taken to

accelerate the growth of the manufacturing sector in Malaysia.


Table 1:Number of foreign labour in Malaysia by sector , 2007-2011 (’000).

Sector 2007 2008 2009 2010 2011

Total 2,044,805 2,062,596 1,918,146 1,817,871 1,5730,61

Agriculture 165,698

(8.1)

186,967

(9)

181,660

(9.4)

231,515

(12.7)

152,325

(9.6)

Farming 337,503

(16.5)

333,900

(16.1)

318,250

(16.5)

266,196

(14.6)

299,217

(19)

Manufacturing 733,372

(35.8)

728,867

(35.3)

663,667

(34.5)

672,823

(37)

580,820

(36.9)

Construction 293,509

(14.3)

306,873

(14.8)

299,575

(15.6)

235,010

(12.9)

223,688

(14.2)

Service 200,428

(9.8)

212,630

(10.3)

203,639

(10.6)

165,258

(9)

132,919

(8.4)

House maid 314,295

(15.3)

293,359

(14.2)

251 355

(13.1)

247,069

(13.5)

184,092

(11.7)

Source: Ministry of Home Affairs, various years.

Note: Values in bracket are percentage.

In the duration of the 9MP, the government was committed to reform the

labour market, with particular emphasis on the increased mobility of

labour and increasing the skills of the workforce. Labour market reforms

were essential to provide a platform for the country to continue to grow

towards becoming a high income economy.

Table 2: Number of foreign labour by skill (’000) for the year 1990-2008

Source: Department of Statistics Malaysia, various years.

Table 2 reports that the number of unskilled foreign workers has more

than doubled in the 2000s compared to the 1990s (from 86,847 in 1990

to 198,876 in 2000). This influx of unskilled foreign workers can be

attributed to the unwillingness of the locals to be involved in

occupations which has the features of 3D (Dirty, Dangerous, Difficult).

Year Skilled Unskilled

1990 9,154 86,847

1995 7,016 104,665

2000 6,986 198,876

2005 9,022 336,055

2008 11,245 421,985



Meanwhile, the number of skilled foreign labour has shown a steady

decrease from the year 1990 until 2000 but steadily increased again until

2008. The skilled foreign workers are complements to the local skilled

workers and a combination of both is needed to facilitate the growth of

the economy towards becoming more competitive.


Productivity refers to the comparison between the output and the input

used to produce it. The inputs for a business are resources within an

organisation which include human capital, finances, etc. while output is

the product produced using the available input (Braid 1983; Prokopenko

1987). Most of the definitions of productivity are related to the

measurement of efficiency of input in the production of output and the

measurement of effectiveness by observing the ratio between the actual

output and the projected output (Pritchard, 1995).

Peri (2012) used the Cobb Douglas function to examine the long-term

effects of the entry of foreign workers on the American labour market

for the year 1960, 2000 and 2006. The findings indicate that foreign

workers have a positive effect on the specialisation of local workers.

Moreover, no evidence that foreign labour crowded-out employment,

rather it encourages efficient specialisation and encourages better use of

technology among the less skilled workers.

Huber et al. (2010), on the contrary, explains the entry of foreign

workers in a more negative light. They explained that locals normally

view foreigners as a threat to their employability. With increasing

population and labour supply, such view is natural. They also studied on

the roles of skilled foreign workers and how qualified they are for their

jobs. They noted a positive contribution from the foreign workers on

labour productivity in skill-intensive industries.

According to Zaleha et al (2011), the increasing entry of foreign workers

into Malaysia is related to its strong economic growth. The Cobb

Douglas production function was used as the basic to form their labour

productivity model with the data ranging from the year 1972 to the year

2005. Their findings indicate that foreign workers contribute positively

to the productivity of labour in this country with improvements of 0.172

percent in labour productivity with a 1 percent increase in foreign


labour. However, their findings also showed a negative effect between

labour productivity and capital labour ratio. This shows that the

Malaysian manufacturing sector is still labour intensive in nature, since

any increase in capital usage,will leads to increase in capital labour ratio

and this negatively affects labour productivity.

Llull (2008), meanwhile, is of the opinion that an immigration of

workers into a country adversely affect the country's productivity. The

argument is that the entry of foreign worker into a firm lowers its

average wage. The estimation took into account the effects of inverse

causality. The researcher also analysed the impact of foreign labour on a

country's per capita GDP. The findings showed a negative and

insignificant impact on productivity. On the contrary, Chia (2011)

argues that foreign workers can increase a country's GDP and

simultaneously fill the needs for workers in Singapore. However, the

study also mentioned that dependence on foreign workers can slow

down economic restructuring and ultimately affect national productivity

adversely. At the end of the study, the researcher advised the

Singaporean government to reduce its dependency on foreign workers to

improve its productivity.

FDI is an important role in the development process in many countries.

FDI generally provides capital and technology to developing countries.

Hale and Long (2006) found that there were positive effects on labour

productivity as a result of the spill-over effects of FDI. They used a

survey of 1,500 firms in China to determine whether there are

technology spill-overs from foreign firms to domestic firms in the same

city and the same industry. The same outcome was found from the study

by Salim and Bloch (2009), found that FDI is an important contributor

to the growth of the chemical industry in Indonesia.Tanna (2009)

studied the relations between FDI and the changes in productivity of

banks. Using data between the years 2000 and 2004 which was obtained

through observation method covering 566 commercial banks, the

analysis was conducted in two stages - firstly, a non-parametric

Malmquist method was used to explain the changes in TFP for bank's

efficiencies of scale, and secondly to explain the changes in technology.

An analysis was also conducted to examine the influences of FDI on

productivity. The study found that FDI has a negative effect in the short

term, but affect productivity positively in the long term. This finding is

consistent with the Malmquist analysis.



Meanwhile, Mohammad Sharif Karimi and Zulkormain (2009) examine

the relationship between FDI and productivity growth using the Todo-

Yamamoto test to understand the causality relation and bounds testing

(ARDL). They used data between 1970 and 2005 in Malaysia. They

found that there is no strong evidence for bi-directional causality and

long-term relationship between FDI and productivity growth.

Thangavelu and Owyong (2003) studied the relationship between export

performance and productivity growth in the Singapore manufacturing

sector. A panel data of 10 major industries in the manufacturing sector

were analysed for the duration between 1974 and 1995. The findings

suggest that growth in labour productivity helps improve export growth

for sub-sectors of selected industries. They also found that FDI-intensive

industries contribute more to labour productivity in the manufacturing

sector than industries which are not FDI-intensive.

Rahmah et al (2012) states that globalisation and technological

advancement increase the demand for high-quality labour. The

researcher examined the impact of globalisation on labour productivity

in the manufacturing sector in Malaysia. This study examined 5 sub-

industries, namely production, processing and preservation of meat,

fisheries, fruits, vegetables, oil and fat (151), manufacturing of refined

petroleum products (232), basic chemical manufacturing (241), iron and

steel manufacturing, office machinery manufacturing, accounting and

computing machinery (300) as well as the manufacturing of electric

valves and tubes and other electronic components, component (321).

Data used include duration of time from the year 1985 to 2007. This

study uses the data panel method choosing between fixed effects to

examine the relationship between labour productivity and capital labour

ratio, export import ratio, FDI, technology transfer and foreign labour.

The study found that globalisation indicators such as FDI and openness

of an economy has a negative and significant effect on labour

productivity. Meanwhile, capital labour ratio significantly influences

labour productivity growth in the manufacturing sector in general as

well as its sub-sectors.

3. Methodology

The analysis in this study uses dynamic panel data methods that

combine time series data (t) and cross-sectional data (n) with t larger

than n. n is the 15 sub-sectors in the manufacturing sector based on 5-


digit Standard Industrial Classification Malaysia (MSIC), while t is 19

years from the year 1990 until 2008. The data used in this study are a

secondary data obtained from the Manufacturing Industry Survey

conducted by the Department of Statistics (DOS). Other data used were

obtained from the Ministry of International Trade and Industry (MITI)

and Malaysian Industrial Development Authority (MIDA). The data was

analysed using the software Stata 10.1. The approach used in the

analysis was Pooled Mean Group (PMG) regression. The study also

employed the Hausman Test to help value the statistical model between

the Mean Group (MG) and PMG to choose the more suitable model for

the available data. To test the stationary of data, the unit root test was

conducted based on the standards of Augmented Dickey Fuller (ADF)

and Philips Perron (PP). Labour productivity is derived from the total

output divided by total labour. Next, the productivity obtained was

assigned as the dependent variable while independent variables are

composed of capital intensity, FDI and labour types of foreign and

domestic labour. For each category of labour, they will be split into two

groups - skilled and unskilled workers.

The dependent variable in this study is labour productivity. In this study,

the output value (production) of 15 sub-sectors of manufacturing is used,

with real production value deduced with Producer Price Index (PPI) year

2000=100 as the base year. The independent variable used is capital

intensity (KL) which is the total fixed asset owned by firms in January

divided by the number of labour involved in the manufacturing sub-

sector, FDI, total number of foreign labour and local labour. The skill

category variable is divided into skilled and unskilled labour. Skills refer

to their abilities and education levels.

3.1 Model Specification

Productivity can be defined as the value or the quantity of output that

can be generated by all units of input. Outputs are products or service

produced by the organisation. In other words, productivity is a concept

that describes the relationship between the output produced by an

organization with the inputs used. In a nutshell, it measures the

efficiency and effectiveness of each unit of input.



To produce a labour productivity model, this analysis employs the Cobb

Douglas production function as basic which can be written as follow:

Y =A Kβ1

Lβ2

(1)

With, Y is total output, A, β1 and β2 are the parameters, K is value of

capital stock and L is total number of labour. The assumption made is

that β1+β2≠ 1, i.e. returns to scale. Two scenarios can take place ;

β1+β2>1 i.e increased returns to scale (IRS) or β1+β2<1 i.e. decreased

returns to scale (DRS). Marginal labour production can be explained

using equation (1) on labour;

𝜕𝑌

𝜕𝐿 = β2 AK

β1 L

β2-1 =

1

𝐿 β2 AK

β1 L

β2 = β2

𝐴𝐾𝛽1𝐿𝛽2

𝐿 = β2

𝑌

𝐿 (2)

Or, 𝜕𝑌

𝜕𝐿= β2

𝑌

𝐿 (3)

Quantity Y/L is the average productivity of labour. It therefore becomes

clear that average production of labour Y/L is the labour productivity.

Hence, the labour productivity equation is as follows:

β2𝑌

𝐿=

𝜕𝑌

𝜕𝐿

𝑌

𝐿=

𝜕𝑌

𝜕𝐿

1

𝛽2 (4)

Replace 𝜕𝑌

𝜕𝐿 from equation (2), thus equation (4) becomes

𝑌

𝐿= β2

𝐴𝐾𝛽1𝐿𝛽2

𝐿

1

𝛽2 = AK

β1 L

β2-1 𝑌

𝐿 = A (

𝐾

𝐿)

𝛽1

L β1+β2-1

(5)

In the form of a logarithm, equation (5) can be written as :

In (𝑌

𝐿)= In A + β1 In (

𝐾

𝐿) + (β1 + β2-1)In L (6)

3.2 Estimation Model

a. ARDL Model

In this study, the dynamic panel data analysis involves a large number of

cross-sectional data (n) and time series data (t) observations. PMG based


on Autoregressive Distributed Lag (ARDL) panel model can determine

the short-term and long-term relationship of a model. By using this

estimation, the intersection, the slope coefficient and standard deviation

are allowed to differentiate the entire group. Assume ARDL (p,

q1,…..qk), the dynamic panel equation can be written as below: -

(7)

With yit as the dependent variable i.e. productivity (Y/L), Xit as vektor k

x 1 as the qualifier variable, µi representing effects of specific groups

(fixed effects), 𝜙𝑖 as the multiplier for dependent variable lag, βi as the

multiplier vector k x 1 qualifier variable, λ*i j multiplier for dependent

variable at lagged first-differences and yi,j is the vector multiplier k x 1

for qualifier variable at lagged first-differences and the value lagged and

i is the manufacturing sub-sector and t is the year.

The main assumption of the ARDL model is that 𝑢𝑖𝑗 is independently

distributed with mean value equals to 0 and standadrd deviation 𝛿2> 0.

It further assumes that error correction term (ec) 𝜙𝑖< 0 for all i ,

meaning that there exist a long-term relationship between 𝑦𝑖𝑡and 𝑥𝑖𝑡.

The long-term relationship can also be written as follows:

𝑦𝑖𝑡 = 𝜃′𝑖𝑗𝑥𝑖𝑗 + Ƞ𝑖𝑗𝑖 = 1,2, … . . 𝑁; 𝑡 = 1,2, … … 𝑇 (8)

With 𝜃′𝑖𝑗 = −𝛽𝑡′

𝜙𝑖 being the multiplier vector k x 1 which is the long-

term coeeficient and Ƞ𝑖𝑗 is stationary with the probability of non-zero

mean which involves fixed effects. Equation (7) can be re-written in the

VECM (Vector Error Correction Model ) system as follows:

∆𝑦𝑖𝑡 = 𝜙𝑖Ƞ𝑖,𝑡−1 + ∑ 𝜆𝑖𝑗

𝑝−1

𝑗=1

∆𝑦𝑖,𝑡−𝑗 + ∑ 𝛿′𝑖,𝑗

𝑞−1

𝑗=0

∆𝑥𝑖,𝑡−𝑗 + µ𝑖

+ µ𝑖𝑡 (9)

With Ƞ𝑖,𝑡−1 as the error variable generated from the long term equation

in (8), 𝜙𝑖 as the error correction term to adjust the balance in the long

term. If 𝜙𝑖 =0, then there is no long term relationship between the

dependent and independent variables. The parameter must be expected

1

1

1

0

,

'

,

*

1,

'

1,

p

j

q

j

itijtiijjtiijtiitiiit uxyxyy



to be significantly negative under the main assumption which shows that

the variables returns to balance in the long term.

Intervals order in the ARDL model is determined using either the

information criteria of Akaike (AIC) or Schwartz Bayesian (SBC)

before the chosen model is estimated using the ordinary least squares

method. Although the estimated value obtained is the same, the standard

deviation estimation of the model chosen by AIC is smaller. However,

the interval order chosen by AIC is higher that the one chosen by SBC.

b. Estimation Model

To identify the short-term and long-term relationships between foreign

labour entry on labour productivity, the model below was estimated:-

Model 1

(10)

Model 2

(11)

1,

31,21,1

1,1

0 '''ln

ti

titi

tii

it L

KInInFwInLw

L

Y

L

YIn

jtij

p

j

q

j

q

j

jtij

jti

jti InFwInLwL

YInFDI

,,21

1

1

1

0

1

0

,,11

,

11,4 **ln'

tjti

q

j

j

jti

q

j

j InFDIL

KIn 1,

1

0

,41

,

1

0

,31 **

1,31,21,

'

1

1,2

0 ''ln

tititi

tii

it

InskillFwwInunskillLInskillLwL

Y

L

YIn

jti

p

j

j

ti

titiL

YIn

L

KInInFDIwInunskilLF

,

1

1

2

1,

61,51,4 '''

1

0

,32*

,

1

0

,22*

1

0

,,12*

q

j

jtijjti

q

j

j

q

j

jtij InskillFwwInunskillLInskillLw

tjti

q

j

j

jti

q

j

jjti

q

j

j InFDIL

KInwInunskillL 2,

1

0

,62*

,

1

0

,52*

,

1

0

,42*


c. Unit Root Test

The unit root test is conducted to observe the stationary level of every

variable tested. A variable is said to be stationary if the mean and

variance is constant over time. It can be stationary either at the levels or

difference. Every variable in a regression equation should be stationary

at the same level, i.e. either being stationary at level or difference, for

instance at the first difference. This condition must be fulfilled for the

estimation to be valid. Otherwise, a false regression estimation is

produced which may produce good estimation results, but in reality

there exists no relationship. In this study, the Augmented Dickey Fuller

(ADF) and Philip Perrons (PP) methods of unit root test are used.

d. Hausman Test

Hausman Test is an econometric statistics named in conjunction with

Jerry A. Hausman. Hausman Test is applied to choose the estimation

Mean Group (MG) or Pooled Mean Group (PMG). If under the null

hypothesis, the difference in the estimated coefficients between the MG

and PMG are not much different, or in other words the value of chi-

square (χ2) not significant, then the PMG is more efficient. This test

helps evaluate if a statistical model corresponds to the data used.

4. Research Findings

4.1 Labour Productivity Growth in the Malaysian Manufacturing

Sector

A finding on the growth of labour productivity in the manufacturing

sector in Malaysia is shown in Table 3. On the whole, labour

productivity growth in the manufacturing sector experienced uneven

growth.



Table 3: Labour productivity growth for the manufacturing sector in Malaysia,

1990-2008

Source: Department of statistics Malaysia, various years.

For the duration 1990-1995, the economy recorded a growth in labour

productivity of 20.7 percent. As for duration 1996-2000 labour

productivity growth declined by 0.4. percent. The sharp fall in domestic

demand following the 1997 financial crisis contributed to the decline in

labour productivity growth (National Productivity Corporation, 2002).

Government initiatives through the Eighth Malaysia Plan (8MP) to

enhance labour productivity include encouraging higher private

investment in research and development (R & D), increasing higher

education enrolment, increasing the number of skilled workers and

knowledgeable workforce, improve skills and capacity related to

technology and promote the use of information and communication

technology (ICT). As a result, the contribution of labour productivity

rebounded in 2001-2005 period by 34.5 percent. The increase was about

17.8 percent higher compared to the duration 1996-2000. However, for

the duration 2006-2008 once again there is a decline in labour

productivity due to the global economic slowdown in year 2008.

a. Unit Root Test Analysis

Based on the analysis as reported in Table 4, the unit root test for both

procedures ADF and PP show that all the variables are stationary in the

first difference I(1) at both without trend and trend i.e. at 1 percent, 5

percent and 10 percent significance levels. This shows that a false

regression can be avoided since all the variables are stationary at first

level of differentiation with trend. Since the panel data was not

stationary at level I(0) but was stationary at first difference, there is a

possible long-term relationship between panel data. These findings

corroborate those of Husin Abdullah and Ferayuliani Yuliyusman

(2011) and Widiyant et al (2012).

Duration Productivity value (RM'000) Growth rate (%)

1990-1995 40659.94 20.7

1996-2000 32949.09 16.7

2001-2005 67724.83 34.5

2006-2008 54827.64 27.9


b. Hausman Test Analysis

Based on the analysis, the results of the Hausman Test for the first

model are shown in Table 5. It was found that for the first model the chi

squared (chi2) is 0.16, Prob> chi2 = 0.9971. This means that Hausman

Test is not significant. The PMG estimation model is more efficient than

the MG. Meanwhile, the results of the Hausman Test for the second

model are shown in Table 6b. The chi squared (chi2) is 0.36 and Prob>

chi2 = 0.9990. Thus, the PMG estimation model is more efficient than

the MG estimation model.

c. Analysis of Dynamic panel data test estimation Pooled Mean

Group (PMG) for First Model

PMG estimation has been conducted and the results from the first PMG

estimation model are shown in Table 5. Results from the first estimation

model using ARDL (0,2,3,0,1) found that in the short term, the error

correction term (ec) is negative 0.19 and is significant at significance

level of 5 percent. The negative ec value reflects the existence of long-

term relationships in the model. In the short term, the results showed

that only the variable KL give significant and positive results in relation

to labour productivity. A 1 percent increase in KL will increase labour

productivity by 0.46 percent. The variable FW, LW, and FDI are

negatively related and do not significantly affect labour productivity.

According Hercowitz et al. (1999) the negative contribution of foreign

and domestic labour on labour productivity growth is only in the short

run. This is because they need to take time to adapt to the labour market

or a new job. FDI inflows into the manufacturing sector in Malaysia

bring with it technology from the country of origin. In the short run,

labour is less able to absorb and apply the technology brought in, which

means that the technology cannot be used efficiently. This tends to

change over time.

In the long run, all variables studied showed significant effects with

labour productivity at the significance level of 1 percent. This is

observable in Table 4 for FW where the coefficient value is positive

0.17. In other words, a 1 percent increase in FW increases labour

productivity by 0.17 percent. The positive effects of foreign labour

supports the findings by Peri (2012) and Kangasniemi et al. (2009). FDI

also positively contributes to labour productivity in the long term with a



0.14 percent increase in productivity with any 1 percent increase in FDI.

Ram and Zhang (2002), who used data from the year 1990 cross-

sectional data, also found the same relationship. The variables LW and

KL also showed the same relationship with coefficients of 1.15 and

0.289. The findings on the variable KL disagrees with the findings in

Zaleha et al (2011). They argued that KL and labour productivity are

inversely related.

d. Analysis of Dynamic Panel Data Test Pooled Mean Group

(PMG) Etimation for the Second Model

In the second estimation model, the findings have been shown in Tables

6a and 6b. Foreign labour and local labour were broken down into two

types of labour by skill type for each type of labour. Foreign and local

labour who work in administrative and professional occupations,

technical and supervisory are classified into skilled labour. While the

labour involved in clerical, general employment and production are

classified unskilled labour. The results from both estimation models

using ARDL (0,0,1,0,0,2,0) found that in the short term, the ec is

negative 0.24 and is significant at 5 percent significance level. The

negative ec value reflects the existence of long-term relationships in the

model studied. In other words, the existence of a long term relationship

between the independent variable and labour productivity in the second

estimation model.

In the short run, the analysis showed that skilled foreign workers

(SkillFW), skilled local labour (SkillLW), unskilled foreign labour

(UnskillFW) and unskilled local labour (UnskillLW) all have negative

but not significant influence on labour productivity. These results

suggest that there is no difference between the productivity of both

labour for skilled and unskilled category in influencing labour

productivity growth. For the category of skilled labour, these findings

are not consistent with the human capital theory where skilled workers

are able to contribute to higher levels of productivity. Level of skills

possessed by the skilled labour are mostly at the most basic level and is

more operations and production oriented, and this may reduce their

influence on the productivity of the organization represented (Rahmah

Ismail et al, 2003) .Besides, variables KL and FDI show positive

influence but are also not significant.


However, in the long run, the study found that all types of labour

according to skill levels indicate relationship with labour productivity

and are all significant at 1 percent significance level. The results

obtained are that SkillFW and SkillLW are positively related to

productivity growth and the values of the coefficient are 0.035 and 0.45

respectively. According Rahmah Ismail et al. (2003), professional

foreign labours are necessary because they can motivate increase in

output. In fact, the recruitment of skilled and experienced,

knowledgeable foreign worker in the manufacturing industry is essential

for smoothing and accelerating the process of transfer of modern

technology. On the other hand, UnskillFW and UnskillLW showed a

negative relationship with productivity growth. This can be seen from

the obtained coefficients of -0.13 and -0.086. The negative coefficient

for the variable UnskillFW and UnskillLW means that a 1 percent

increase of UnskillFW and UnskillLW will reduce labour productivity

by 0.13 percent and 0.086 percent respectively. George Borjas (2006),

using case studies in the United States, concluded that migrant workers

contribute a negative impact on labour productivity of American,

especially those with low skills. Similarly, KL and FDI respectively

relate positively with labour productivity and the coefficients are 0.36

and 0.0085.

5. Conclusions and Suggestions

The study found that overall, in the short run, local and foreign labour

force labour contribute negatively but are both not significant to the

growth of labour productivity. In contrast, in the long run, both have a

positive relationship with labour productivity growth. The labour

productivity increase is a good omen toward achieving a high-income

country status by the year 2020, due to the increase in total output

produced. When broken down by type of labour skill categories, only

migrant labourers and skilled local labour have positive and significant

relationship with labour productivity in the long term. In other words,

foreign and local recruitment of skilled labour is necessary because they

can lead to increase in productivity. Meanwhile, with regard to variable

KL, it was found that in the long run, it has a positive impact on labour

productivity growth of the manufacturing sector. Similarly, the FDI

variable indicated a positive relationship with labour productivity in the

long term. The findings support previous studies which argue that FDI



intensive industries contribute more to productivity than those which are

not FDI intensive ( Thangavelu and Owyong, 2003).

The study concludes that high-skilled foreign workers are needed to help

develop the manufacturing sector in Malaysia. Recruitment of unskilled

foreign labour should be reduced and replaced with the use of local

labour to fill labour shortages. In order to decrease the dependence on

foreign labour, especially unskilled foreign labour, the government has

undertaken several efforts. Among them are efforts to attract expertise

from other countries and Malaysians who are working in other countries

to return home and work in Malaysia to improve the various sectors

especially the R&D sector. These measures should be strengthened

from time to time so that results can be obtained continuously and the

implementation can be more effective. The demand for knowledge

workers, a category that includes senior officials and managers,

professionals, technicians and associate professionals is expected to

increase at an average rate of 2.5 percent per annum (Malaysia, 2006).

The introduction of a minimum wage by the government does not only

benefit local labour, it is also aimed at reducing Malaysia's dependence

on foreign labour especially unskilled foreign labour. Higher minimum

wages lead to increased cost to the employer. Such circumstances would

indirectly force employers to reduce the use of foreign labour and thus

pave the way to greater employment of local labour. Moreover, in

encouraging more local labour into the workforce in labour, especially

in the farming sector, the government is always trying to upgrade

facilities and provide the basic infrastructure including residential estate

(National Productivity Corporation, 2011).

This study also presents a number of steps that can be taken into account

to assist policy makers in designing a strategy to reduce dependence on

foreign labour. The researcher suggests that the policy maker step up

efforts to streamline the recruitment of foreign labour and the levy

system, revise the wage system, and provide better benefits to increase

the chances of retaining local labour in selected sectors. The government

should review the policies, strategies, laws and procedures relating to

the employment and wages of highly skilled foreign expertise of in

select occupations. Firms are also encouraged to implement

productivity-linked salary system. Through innovation and exploitation

of new ideas, added value can be obtained using the same human capital


and the same amount of other resources. Investment in science, research

and education can also serve as an engine of innovation for the

economy.

FDI will be continue to be a catalyst to improve the R&D ability and as

a source of technology transfer. Private sector actors are urged to forge

strategic alliances with foreign partners to ensure that their R&D

activities are not out of touch with the outside world. The government

should continue to identify and provide assistance to multinational

companies with R&D capabilities in strategic areas to invest in

Malaysia. In addition, the incentive mechanism for FDI should be

revised to give priority to those with new, updated R&D capabilities and

with value-added to be located in Malaysia.



References

Aliya Rosa., Rahmah Ismail., & Noorasiah Sulaiman,.(2012).

Globalisation and Labour Productivity in the Malaysian Manufacturing

Sector. Review of Economics & Finance, 2, 76-86.

Borjas, G. J. (1993). Immigration Policy, National Origin, and

Immigrant Skills: A Comparison of Canada and the United States.

National Bureau of Economic Research. No. 3691

Borjas, G. J. (2006). Native internal migration and the labour market

impact of immigration. Journal of Human Resources, 41(2), 221-258

Chia, S.Y. (2011).Foreign labour in Singapore: trends, policies, impact

and challenge. Philippine Institute for Development Studies, No.2011-

24.

Chuang, Y. C., & Lin, C. M. (1999). Foreign direct investment, R&D

and spillover efficiency: Evidence from Taiwan's manufacturing firms.

The Journal of Development Studies, 35(4), 117-137.

Hale, G., & Long, C. X. (2006). What determines technological

spillovers of foreign direct investment: Evidence from China. Economic

Growth Center, Yale University.

Hausman, J. A. (1978). Specification tests in econometrics.

Econometrica: Journal of the Econometric Society, 1251-1271.

Hercowitz, Z., Lavi, Y., & Melnick. (1999). The Impact of

Macroeconomic Factors on Productivity in Israel, 1960-96. Bank of

Israel Economic Review 72: 103-124

Huber, P., Landesmann, M., Robinson, C., & Stehrer, R. (2010).

Migrants skills and productivity: A European Perspective. National

Institute Economic Review, 213(1), R20-R34.

Husin Abdullah & Ferayuliani Yuliyusman. (2011). The effects of

economic freedom on economic growth: Empirical Study in Indonesia,

Hong Kong, Malaysia, Singapore and the United States. Proceeding

PERKEM VI, jilid 1 (2011) 410 – 423.


Ishak Yussof & Nor Aini Haji Idris (2009). Policy and strategy of

economic development of Malaysia: An assessment of the economy

towards balanced development, edited by Nor Aini Haji Idris dan Ishak

Yussof. Bangi: Printed by University Kebangsaan Malaysia.

Prime Minister's Department of Malaysia.Economics Transformations

Program Report (ETP), (2012). Kuala Lumpur. Printed by National

Malaysia.

Department of Statistics Malaysia. Malaysia Manufacturing Industries

Survey Report, various years. Kuala Lumpur. Printed by National

Malaysia.

Kangasniemi, M., Ivars, M. M., Robinson, C., & Martínez, L. S. (2009).

The economic impact of migration: Productivity analysis for Spain and

the United Kingdom. Documentos de trabajo (Fundación BBVA), (10), 1

Llull, J (2008). The impact of immigration on productivity. Center for

Monetary and Financial Studies (CEMFI), No. 0802 Malaysia (1996).

Seven Malaysia Plan , 1996-2000. Kuala Lumpur: Printed by National

Malaysia.

Malaysia (2006). Ninth Malaysia Plan, 2006-2010. Kuala Lumpur:

Printed by National Malaysia.

Malaysia (2011). Ten Malaysia Plan , 2011-2015. Kuala Lumpur:

Printed by National Malaysia.

Widiyanti, M. (2012). Monetary with Interest Rate Policy and Impact on

Prices: Empirical Evidence from Malaysia. Proceedings PERKEM VII,

jilid 1 (2012) 91 – 100.

Mohammad Sharif Karimi & Zulkormain Yusop (2009).FDI and

economic Growth in Malaysia.Faculty of Economics and Management,

University Putra Malaysia.

Malaysia Productivity Corporation .(2011). Productivity Report. Kuala

Lumpur: Printed by National Malaysia.



Malaysia Productivity Corporation. (2013). Productivity Report. Kuala

Lumpur: Printed by National Malaysia.

Peri, G. (2012). The effect of immigration on productivity: Evidence

from US states. Review of Economics and Statistics, 94(1), 348-358.

Pesaran, M. H., Shin, Y., & Smith, R. J. (1999).Bounds Testing

Approaches to the Analysis Long Run Relationship. Unpublished

Manuscript. University of Cambridge.(http// www.econ.cam.ac.uk/

faculty pesaran.

Preibisch, K. L. (2007). Local produce, foreign labour: Labour mobility

programs and global trade competitiveness in Canada. Rural Sociology,

72(3), 418-449.

Pritchard, R. D. (1995). Productivity measurement and improvement:

Organizational case studies. Praegar.Publisher Prokopenko, J. (1987).

Productivity management: A Practical Handbook. Geneva:ILO

Rosario-Braid, F. (1983). Communication strategies for productivity

improvement Tokyo: Asian Productivity Organization.

Rahmah Ismail, Nasri Bachtiar, Zulkifly Osman & Zulridah Mohd Noor

(2003). The role of foreign labour to output growth, employment

opportunities and wages in the manufacturing sector in Malaysia.

Journal of Economics Malaysia 37:103-128.

Ram, R., & Zhang, K. H. (2002).Foreign direct investment and

economic growth: Evidence from Cross‐Country data for the 1990s.

Economic Development and Cultural Change, 51(1), 205-215.

Sahar, M .(2002). Labour productivity: An important business strategy

in manufacturing. Integrated Manufacturing Systems, 13(6),435-438

Salim, R. A., & Bloch, H. (2009).Does foreign direct investment lead to

productivity spillovers? Firm level evidence from Indonesia.World

Development, 37(12), 1861-1876.


Thangavelu, S. M., & Owyong, D. T. (2003).The impact of export

growth and scale economies on productivity in Singapore's

manufacturing industries. Journal of Economic Studies, 30(6), 623-635

Zaleha M.N., Noraini I., Rusmawati S., & Suhaila A.J. (2011).The

Impact of foreign workers on labour productivity in Malaysia

manufacturing sector. Int. Journal of Economics and Management

5(1): 169-178.



Table 4: Results for Unit root test ADF and PP

Variable ADF Philips Perron (PP)

Without Trend Trend Without Trend Trend

Level

InLW -1.6890 -2.0667 -1.7100 1.8953

(0.2300) (0.2369) (0.2299) (0.2370)

InFW -0.5659 -1.7679 -0.2466 -1.6063

(0.1933) (0.2164) (0.1932) (0.2064)

lnSkillLW -2.6450 -3.7204 -2.6445 -3.7200

(0.2131) (0.2600) (0.2100) (0.2555)

lnSkillFW -3.0741 -3.6774 -3.8573 -4.5720

(0.2436) (0.2524) (0.2440) (0.2500)

lnUnSkillLW -1.3650 1.7900 -1.4161 1.8000

(0.1700) (0.1744) (0.1672) (0.1744)

lnUnskillFW 0.3158 -2.0631 0.8032 -1.9858

(0.0881) (0.1774) (0.0900) (0.1774)

lnKL -2.4111 -2.0066 -2.3556 -1.9332

(0.1850) (0.2313) (0.1845) (0.2310)

lnFDI -4.4193 -4.2835 -4.4202 -4.2838

(0.2490) (0.2578) (0.2491) (0.2577)

Difference

Without Trend Trend Without Trend Trend

InLW -7.5267 -10.8000 -6.6781 12.1981

(0.1886)*** (0.1420)*** (0.1890)*** (0.1410)***

InFW -5.5500 -6.3010 -5.5400 -7.8247

(0.2413)** (0.2414)** (0.2403)** (0.2424)***

lnSkillLW -6.7111 -6.5311 -12.3400 -16.0802

(0.2235)*** (0.2310)** (0.2235)*** (0.2206)**

lnSkillFW -7.0468 -8.1492 -7.1000 -8.6363

(0.2709)*** (0.2715)*** (0.2710)*** (0.2710)***

lnUnSkillLW -4.0310 -4.0861 -4.0312 -3.2977

(0.2578)* (0.2650)* (0.2577)* (0.2648)*

lnUnskillFW -4.1910 -4.5616 -2.6666 -3.7104

(0.2673)* (0.2940)** (0.2700)* (0.2939)**

lnKL -4.8425 -4.4287 -5.0787 -12.1316

(0.2536)** (0.5767)* (0.2540)* (0.2528)***

lnFDI -6.9469 -6.7336 -16.9683 -18.3393

(0.2196)*** (0.2269)** (0.2197)*** (0.2270)**

Note: ***, **, and *is significant at 1% , 5%, and 10% significance level. Upper value

is the coefficient value, the value in bracket is the standard deviation.


Table 5: Results of Labour productivity estimation using first model

Pooled Mean Group (PMG).

DEPENDENT VARIABLE:

Productivity Growth (In Y/L)

Model 01:

ARDL (0,2,3,0,1)

Short term effect

∆lnFW -0.0037

(0.0573)

∆lnLW -0.3251

(0.418)

∆lnKL 0.4552

(0.1046)***

∆lnFDI -0.001

(0.0167)

Constant 0.5745

(0.2266)**

Error Correction Term (ec) -0.1861

(0.043)**

Long term effect

lnFW 0.1711

(0.0516)***

lnLW 1.1519

(0.3374)***

lnKL 0.2894

(0.0547)***

lnFDI 0.1384

(0.0192)***

Hausman Test chi2(4)= (b-B)'[(V_b-V_B)^(-1)](b-B)

= 0.16

Prob>chi2 =0.9971

Note: Lag order is chosen based on AIC (Akaine Information Criteria).*** Significant

at 1% significance level, ** Significant at 5% significance level and *Significant at

10% significance level. Upper value is the coefficient value, the value in bracket is

the standard deviation.



Table 6a: Results of Labour productivity estimation using second model

Pooled Mean Group (PMG).


at 1% level, ** Significant at 5% significance level and *Significant at 10%

significance level. Upper value is the coefficient value, the value in bracket is the

standard deviation.

DEPENDENT VARIABLE

Productivity growth (In Y/L)

Model 02:

ARDL (0, 0, 1, 0, 0, 2, 0)

Short-term effect

∆InSkillFW

-0.0251

(0.0454)

∆InUnskillFW -0.0283

(0.0452)

∆InSkillLW -0.4145

(0.5161)

∆UnskillLW -0.0214

(0.0515)

∆InKL 0.0374

(0.1004)

∆InFDI 0.0136

(0.0291)

Constant 0.4105

(0.1907)**

Error Correction Term (ec) -0.2417

(0.1527)**


Table 6b: Results of Labour productivity estimation using second

model Pooled Mean Group (PMG)


at 1% level, ** Significant at 5% significance level and *Significant at 10%

significance level. Upper value is the coefficient value, the value in bracket is the

standard deviation.

DEPENDENT VARIABLE

Productivity Gowth (In Y/L)

Model 02:

ARDL (0,0,1,0,0,2,0)

Long term effect

InSkillFW 0.0345

(0.0068)***

InUnskillFW -0.125

(0.0360)***

InSkillLW 0.447

(0.0103)***

InUnskillLW -0.0863

(0.0151)***

InKL 0.3614

(0.0105)***

InFDI 0.0085

(0.0028)**

Hausman Test chi2(6) =(b-B)'[(V_b-V_B)^(-1)](b-B)

=0.38

Prob>chi2=0.9990


The Real Effect of Government Debt:

Evidence from the Malaysian Economy


The results demonstrate that there is a long-run relationship between federal

government debt and economic growth in Malaysia. In addition, our findings

are of great interest since there is evidence of a non-linear relationship

between the federal government debt and economic growth, which suggests

the optimal level of debt that the government should hold. Hence, the

accumulation of federal government debt is positively associated with

Malaysia’s economic growth up to an optimal level. While an additional

increase in federal government debt beyond the optimal level has inversely

contributed to the Malaysian economy.

1. Introduction

There are several lessons to be learnt from the recent sovereign debt

crisis that has affected most of the European economies. A rise in public

debt and country-specific problems are among the factors behind the

European Financial Crisis. Ireland is facing a banking crisis while Spain

is experiencing a housing bubble. In addition, Greece, Italy and Portugal

are involved in fiscal mismanagement. Thus, they have drawn their

economies into sovereign debt crises and are still in the process of

recovery four years after the crisis erupted in 2008. On the other hand,

current development shows that Japan, Greece, Italy, Portugal and

Ireland are among the top-listed countries with very high levels of public

debt (IMF 2012). As at the end of 2011, Japan held a public debt of

about 229.77 per cent of gross domestic product (GDP) and this was

expected to hit almost 240 per cent of its growth. In addition, Greece,

Italy, Portugal and Ireland held public debts (as percentages of GDP) of

about 160.81 per cent, 120.11 per cent, 106.79 per cent and 104.95 per

cent respectively. Furthermore, in response to this issue, the

Organisation for Economic Cooperation and Development (OECD) has

Faculty of Economics & Muamalat, Universiti Sains Islam Malaysia

Email: [email protected]

58 The Real Effect of Government Debt:


called on countries to cut their public debts to prudent levels of around

50 per cent of GDP in order to cope with future challenges including

health, long-term care and pensions.2

Looking at the roots of the problem, it can be seen that persistent large

deficits, used primarily to finance public sector operating expenses, have

resulted in high ratios of debt to GDP, thus driving Greece and other

European countries into debt overhang problems. In the midst of this

phenomenon, some developing countries are showing signs or

symptoms of the debt overhang problem; this highlights the question of

whether government debt has benefited economies (Cecchetti, Mohanty

and Zampolli 2011; Baum, Checherita-Westphal and Rother 2012;

Presbitero 2010; Caner, Grennes and Koehler-Geib 2010). Thus has

underlined the urgency of investigating the issue of the heavy stock of

public debt in developing economies, since neither developed nor

developing countries are immune to the public debt sustainability issue.

Malaysia, a small open economy, has recorded a fiscal deficit position

since its independence in 1957 except for the period 1993-1997.

Furthermore, the country has recorded 13 consecutive years of fiscal

deficits since the year 1997. This condition was leading Malaysia to

accumulate a stock of indebtedness regardless of domestic or

international capital markets since, by continuing to run budget deficits,

the country would have a high stock of debt (as depicted in Appendix

1). In addition, the federal government debt was financed from domestic

and foreign funding, which constituted approximately 96.2 per cent and

3.8 per cent of the gross borrowing respectively.3 As at the end of 2011,

total federal government debt was recorded at RM 455,745 million,

which is equivalent to 53.8 per cent of Gross Domestic Product (GDP)

(Malaysia Economic Planning Unit 2011). This position has already

been reached and is slightly higher than the prudent cut-off point of

public debt-holding that has been set for developed economies.

However, no one size fits all. This highlighted the importance of

2 Furthermore, the European Union has set a ceiling for public debt at 60 per cent of

GDP. By the same token, OECD highlights the importance of macro prudential

supervision on an individual-country basis. 3 Meanwhile, the federal government financing came mainly from domestic sources

through the issuance of Malaysian Government Securities (MGS) and Government

Investment Issues (GIIs) where the major shareholders were Employee Provident Fund

(EPF), foreign investors, banking institutions and insurance companies.


analysis on a country-by-country basis before it was too late and the

country had already passed its optimal level, becoming trapped in a debt

overhang situation, being in default and, to a lesser extent, witnessing

the eruption of a sovereign debt crisis.

In principle, if borrowing has been allocated efficiently, a country will

benefit from it since debt financing of public spending can make a

positive contribution to productive investment and ultimately to

economic growth (Miller and Foster 2012). In contrast, a country with

high levels of debt will face the probability of a debt overhang problem,

to a lesser extent, or default or bankruptcy. As such, the increasing level

of government indebtedness raises the issue of the effectiveness of the

fiscal policy formulated by the government. Furthermore, it leads to the

issue of whether the borrowing is efficiently and productively allocated

to the economy through development projects which in return will

generate sustainable economic growth.4

A country that accumulates large stock of debt and could affect the

ability to repay its debts is more likely involved in a debt overhang

situation. There is limited study conducted to investigate the effect of

debt overhang on various economies. Recent study conducted by

Reinhart et al. 2012, identify 26 public debt overhang episodes in 22

advanced economies since the early 1800s. The study found that growth

effects are significant even in the many episodes where debtor countries

were able to secure continual access to capital markets at relatively low real

interest rates. On the other spectrum, Brown and Lane (2011) conduct a

study to assess the effect of debt overhang to economic activity in

Emerging Europe countries. Debt overhang could be a threat to activity

in the tradable sector in the more advanced economies of the region. In

addition, the debt overhang emerge at different levels of indebtedness

depending on the country characteristics namely institution, policies and

4 The highest deficit was recorded in 2009 and was due to the global economic

slowdown as the external sector collapsed and the business community remained

cautious and risk-averse. Weak private investment, sluggish exports performance and

higher expenditure incurred are due to the implementation of the stimulus packages,

resulting in a weak financial position in the fiscal position (Malaysia Economic

Planning Unit 2009).



access to private capital (Cordella et al. 2005). Sound institutions can

probably affect the ability to service debt in times of crisis. As such,

with limited but growing study, the effect of debt overhang situation

might differ depending on the economies.

As a consequence, the increasing level of the stock of government debt

has raised concerns and leads to the question of whether a country with

high levels of government debt is still sustainable. Furthermore, this

problem could threaten the developing economies, especially their

banking sectors, in the event of sovereign debt crises. Thus, this

situation has intensified interest and drawn attention to the long-lasting

implications of policy action for the country’s government debt position.

A real picture of Malaysia’s public debt position is important for policy

formulation as well as for investors’ ability to strategize their investment

decisions. Considering the growth in the literature on developed

countries, an attempt to investigate Malaysia’s public debt position is

feasible since, as a small open economy, Malaysia also faces a high risk

of vulnerability and uncertainty in its economy. Focusing on a long-

horizon return, the objective of this paper is to analyze the real effect of

government debt on Malaysia’s economy. This study is a contribution to

the literature on the Malaysian government’s debt position after a long

episode of fiscal deficits. Thus, this study attempts to fill this gap in the

literature. The paper is laid out as follows. Section II offers a brief

overview of the literature on external debt. Section III outlines the data

and methodology, while the empirical results are presented in section

IV. Section V concludes the paper.

2. Theoretical and empirical evidence

Over the past decades, academics and policy-makers have shown a

consistent interest in investigating and developing the theory on the link

between debt and economic growth. However, a limited but growing

number of studies have been examining the role of public debt in a

country’s economic growth.5 The discussion on the impact of public

debt on a country’s growth has produced a single conclusion on the

adverse impact of public debt on growth (Modigliani 1961; Adam and

Bevan 2005; Aizenman et al. 2007). Based on the aggregate model

5 On the other hand, there is a vast amount of literature focusing on the impact of

external debt in generating countries’ economic growth.


developed by Modigliani (1961), the accumulation of government debt

will have a positive impact on growth if the increase in debt is

accompanied by government expenditure on productive public capital

formation. In other words, the debt will benefit the economy if it is

capable of generating a stream of real income for future generations and

vice versa (Modigliani 1961). In addition, Modigliani stated that holding

too much public debt will affect the country through the crowding-out

effect, which will lead to the debt overhang problem, as explained by

Krugman (1988). On the other hand, by setting up a simple overlapping

generation (OLG) model of savings, Adam and Bevan (2005) found

evidence of interactive effects between deficits and debt stock, with high

debt stock exacerbating the adverse consequences of high deficits.

Furthermore, the fiscal deficits would be growth-enhancing if financed by

limited seigniorage, growth-inhibiting if financed by domestic debt, and

have an opposite flow and stock effect if financed by external loans. In

addition, Aizenman et al. (2007) examined the optimal public investment

and fiscal policy for countries subject to binding on tax and debt

capacities. They found that the public debt-to-GDP ratio should be held

constant in the economy, adding that public sector borrowing to finance

the accumulation of public capital goods may allow the economy to reach

a long-run optimal growth path faster (Aizenman et al. 2007).

Several empirical works have focused on the impact of public debt on a

country’s economic growth, while proposing a non-linear analysis

(Reinhart and Rogoff 2010; Pattillo 2004; Baum et al. 2012; Cecchetti et

al. 2012; Presbitero 2010; Schclarek 2004). Most of the literature found

that the tipping point of government debt should be held at around 64

per cent to 100 per cent of GDP depending on the size and the

development stage of the economy. In particular, Reinhart and Rogoff

(2010) conducted a study on 20 developed countries over the period

1790-2009 to investigate the relationship between public debt and long-

term real GDP growth. The findings suggest that the relationship

between public debt and economic growth is relatively weak below the

threshold of 90 per cent of GDP; however, above the 90 per cent level

the median growth rate falls by one per cent. By the same token,

Checherita and Rother (2010) examined the impact of government debt

on economic growth in twelve Euro-area countries for a period spanning

about 40 years from 1970 and found a non-linear impact of debt with 90

to 100 per cent of GDP estimated as the optimal level. On the other



hand, recent studies conducted for almost the same sample, for the

period 1990-2010, suggest that the short-run impact of debt on growth is

positively and statistically significant up to 67 per cent debt to GDP

(Baum et al. 2012). Meanwhile, according to Baum et al. (2012), beyond

the limits the positive impact of government debt decreases to around

zero and loses its significant impact on economic growth. However, no

robust evidence on the relationship between government debt and

economic growth has been found for 24 industrial countries (Schclarek

2004). In addition, results from a dataset of 18 OECD countries over the

period 1980-2010 support the view that, beyond 85 per cent debt to

GDP, debt is a drag on growth (Cecchetti et al. 2012). Meanwhile,

Presbitero (2010) complemented the existing studies by focusing on the

developing countries. Over the period 1990-2007, the results show that

public debt has a negative impact on output growth up to a threshold of

90 per cent of GDP; beyond this level, its effect becomes irrelevant. An

interesting study by Caner et al. (2010) estimates the threshold debt

levels based on annual datasets of 101 developing and developing

countries from 1980 to 2008. The results established a threshold of 77

per cent public debt-to-GDP ratio. Beyond the threshold level, each

additional percentage point of debt costs 0.017 percentage points of

annual real growth. In addition, results for a sample of emerging

economies are even more pronounced with the estimated thresholds

found to be at 64 per cent debt-to-GDP ratio; above the threshold level

each additional percentage point of public debt amounts to 0.02

percentage points. Thus, inspired by the notion that no one size fits all,

this paper will contribute to the literature by focusing on an individual

developing country, namely Malaysia.

3. Model, Method and Data

A detailed analysis of the effect of federal government debt on

Malaysia’s economic growth will provide evidence on the real scenario

of Malaysia’s federal government debt position. In addition, with the

application of several econometric procedures, the optimal level of

public debt that the country should hold will be estimated. Thus, to

investigate whether the federal government has contributed to economic

growth, the basic growth model to be estimated is

ttt εβXY 0 (1)


where Y is the dependent variable, X is k-vector of regressors, and the

subscripts t =1,….,T identify the time dimensions. Y represents the real

GDP per capita and X includes investment rate, labour force rate and

federal government debt, while t represent the error term. The real GDP

per capita (dependent variable) is a proxy of economic growth. In

addition, the independent variables include investment rate, labour force

rate and federal government debt to represent the rates of growth of

factor inputs in the production function, and openness captures for

government policy. This paper also estimates the direct link between

federal government debt and investment rate to provide an additional

insight into the effect of federal government debt on economic growth

via capital accumulation. The estimated investment model is

ttt XI 0 (2)

I represents the investment rate while X is labour force, openness and

federal government debt. The investment rate represents the growth of

the economy, labour force rate and federal government debt represent

the rates of growth of factor inputs in the production function, and

openness represents the government policy. Since the observations are

on a quarterly basis, for the maximum order of the lags in the ARDL

model, a lag order of 4 is chosen.

The procedure starts with the Ordinary Least Squares estimation as a

benchmark for the analysis. Next, the analysis proceeds with the

cointegration tests. The Autoregressive Distributed Lag (ARDL)

cointegration bound test developed by Pesaran et al. (2001) will be

employed. The bound test developed by Pesaran et al. (2001) is the

Wald test (F-statistic version of the bound testing approaches) for the

lagged level variables in the right-hand side of an Unrestricted Error

Correction Model (UECM). This procedure involved two stages before

the long-run relationship could be established. In addition, the null

hypothesis of a non-cointegrating relation (Ho: δ1= δ2= δ3=…= δn = 0) is

tested by performing a joint significance test on the lagged level

variables. The first stage of the ARDL approach involved the F-test in

which the asymptotic distribution of the F-statistic is non-standard under

the null hypothesis of no cointegrating relationship between the

examined variables, irrespective of whether the explanatory variables



are purely I(0) or I(1). Under the conventional levels of significance

such as 10 per cent, 5 per cent and 1 per cent, if the statistic from a Wald

test falls outside the critical bounds value (lower and upper values), a

conclusive inference can be made without considering the order of

integration of the explanatory variables. If the F-statistic exceeds the

upper critical bound, the null hypothesis of no cointegrating relationship

can be rejected. However, if the test statistic (F-statistic) falls below the

lower critical bound, then the null of non-cointegration cannot be

rejected. If the F-statistic falls between the upper and lower bounds, a

conclusive inference cannot be made. Next, the second stage of the

ARDL approach involves an estimation of the coefficients on the long-

run cointegrating relationship and the corresponding error correction

model. The lagged error correction term (et-1) derived from the error

correction model is an important element in the dynamics of the

cointegrated system as it allows for adjustment back to the long-term

equilibrium relationship given a deviation from the last year.

In addition, in order to gather as much as information as possible on the

effect of the federal government debt on Malaysia’s economic growth,

this paper also employs a causality test which is an extension of the

Granger causality test that applies the bootstrapping method with

endogenous lag length proposed by Hacker and Hatemi-J (2010). The

causal relationship between federal government debt and other

economic variables will provide evidence on the impact of federal

government debt on the economy. As such, this analysis should at least

provide some indication of the impact of the current implemented

policy. A standard Granger causality test on the first difference is

performed to find the direction of causality. The Granger test is

performed on

itit

n

i

i GDPFGDFGD 1

1

(3)

itit

n

i

i FGDGDPGDP 1

1

(4)

itit

n

i

i GDPDEFDEF 1

1

(5)


itit

n

i

i DEFGDPGDP 1

1

(6)

itit

n

i

i FGDINVINV 1

1

(7)

itit

n

i

i INVFGDFGD 1

1

(8)

itit

n

i

i DEFINVINV 1

1

(9)

itit

n

i

i INVDEFDEF 1

1

(10)

where GDP represents the real GDP per capita and FGD signifies the

federal government debt. In addition, the INV and DEF represent the

investment rate and federal government deficits respectively. In standard

procedure when applying the Granger causality test, the lag length is

assumed to be known beforehand where the preselection of the lag order

may affect the distribution of the test statistics. Thus, Hacker and

Hatemi-J (2010) suggests endogenously determining the lag length

choice including the use of the bootstrapping method. Furthermore, the

bootstrapping method appears to have better size properties, seems to be

robust to the existence of autoregressive conditional heteroscedasticity

(ARCH), and appears to have more power compared to the asymptotic

test for the same actual size (Hacker and Hatemi-J 2010).6

Next, this paper employs the test proposed by Hansen (2000) which

tests the null hypothesis of a linear regression against a threshold

regression analysis. In the form of the thresholds model,

ttt xy '

1 tq (11)

titt xy '

2 tq (12)

6 Thanks to Professor Abdulnasser Hatemi-J and Scott Hacker for the GAUSS routine.



where tq is the threshold variable, which is federal government debt. In

addition, the threshold variable could be part of the regressors and it is

used to split the sample into two regimes. Meanwhile ty is economic

growth measured by real GDP per capita. tx is 1p vector of

independent variables which include investment rate, openness and

federal government debt, and t is a regression error.

Hansen (2000) has developed a threshold model estimator that considers

the least squares estimations. Furthermore, by providing an asymptotic

simulation test of the null of linearity against the alternative of a

threshold, this method also computed a confidence interval by inverting

the likelihood ratio statistics. Hansen (2000) also proposes an F-test

bootstrap (heteroscedasticity-consistent) procedure to test the null of

linearity. Since the threshold value is not identified under the null, the

p-values are computed by a fixed bootstrap method. The sample consists

of economics data from the period 1996Q1-2011Q4. The data are

collected from the IMF/IFS statistics and the Monthly Bulletin of the

Central Bank of Malaysia. Details of the variables are attached in

Appendix 5.

4. Results and discussion

4.1 Descriptive analysis

Table 1 provides descriptive statistics on the main variables employed in

this study. The variables include real GDP per capita, labour force,

investment, openness, federal government debt and federal government

expenditure. The descriptive statistics consist of mean, standard

deviation, maximum values and minimum values. Table 1 shows that

there are substantial variations for all variables. The real GDP per capita

ranges from RM2,785 to RM7,090 with a mean value of RM4,454. In

addition, the labour force and investment show small variations with

lower values of standard deviation, which indicates the dispersion from

the mean. Meanwhile the openness variable ranges from RM95,889 to

RM330,817, with a mean of RM206,075. By the same token, with a

significant variation indicated by its standard deviation of RM108,776,

the mean value of federal government debt is RM210,478.


Table 1. Descriptive Statistics

Mean Standard Deviation

Min Max

Real GDP per capita 4,454 1,110 2,785 7,090

Labour force 10,253 1,095 8,469 12,730

Investment 27,075 5,988 15,189 39,019

Openness 206,238 71,614 95,889 330,817

Federal government debt 210,478 108,776 83,533 456,127

Note: All figures are in RM Million.

The investigation starts by analyzing the composition of Malaysia’s

government debt. Figure 1 shows the composition of federal government

debt over the period 1970-2012. The figure shows a stable pattern with

the domestic debt accounting for about 85 per cent of total federal

government debt in 1970, increasing to 96 per cent in 2011. In addition,

the borrowing is funded from domestic financial institutions in

Malaysia, including banks, financial institutions and social security

institutions.

Figure 1. The composition of Malaysia Federal Government Debt

0

50,000

100,000

150,000

200,000

250,000

300,000

350,000

400,000

450,000

19

70

19

73

19

76

19

79

19

82

19

85

19

88

19

91

19

94

19

97

20

00

20

03

20

06

20

09

RM

Mill

ion

Year

Domestic debt External debt



In addition, about 63 per cent and 25 per cent are in the form of

government securities and government investment issues (GII’s)

respectively (as at the end of 2011). Furthermore, the maturity of the

government debt is mainly seen over the maturity periods of 6 to 10

years and 4 to 5 years, covering 38 per cent and 29 per cent respectively

of the total government securities debt.

4.2 Empirical analysis

The empirical analysis of this paper starts by estimating a standard

ordinary least squares test on the standard baseline debt-growth model to

establish a reference point. The results are shown in Table 2. By

following the general-to-specific methodology, the results show an

estimation of a model where the real GDP per capita and investment

signify the economic growth. The estimated model of real GDP per

capita as the dependent variable is shown in columns (1) to (5), while

columns (6) to (10) show the results of the estimates where investment

proxy is the dependent variable. The results in columns (1), (2) and (3)

show an insignificant effect of federal government debt on the country’s

economic growth (which is proxied by the real GDP per capita as the

dependent variable). However, estimates that only consider federal

government debt as the independent variable reveal a positive and

significant effect (at 5 per cent significance level) on the country’s

economic growth.

On the other hand, with the investment rate variable representing the

country’s growth, there are positive and significant (at 5 per cent

significance level) effects of federal government debt on the country’s

growth, as depicted in columns (6), (7) and (10). Meanwhile, the results

also reveal a positive and significant (at 5 per cent significance level)

effect of trade openness in explaining Malaysia’s economic growth, as

shown in columns (1) to (3). In addition, we try to include the federal

government debt squared in columns (3) and (8) to investigate the

potential of the non-linear effect or debt-Laffer curve relationship.

Columns (3) and (8) demonstrate that the federal government debt has a

significant effect on the country’s economic growth. In addition, the

federal government debt ^2 variable is significant at 5 per cent

significance level, with a negative effect on the growth rate of the

country’s income. This may implies evidence of an inverted-U-shaped

relationship in the federal government debt-growth model. The inverted-


U relationship explains that an increase in debt stock has a positive

effect on economic growth until it achieves its optimal level (up to a

certain level). Beyond the threshold level, an increase in the stock of

indebtedness is associated with a negative effect on economic growth.

The negative effect could be related in cases where it has not been

efficiently allocated to investment and if there is too much debt-holding

that might squeeze the investment through debt repayment. In contrast,

the results show that the federal government deficit plays an

insignificant role in explaining the variability in economic growth.

However, these results should be interpreted with caution since the

diagnostic test shows a sign of bias where the results may be suffering

from major econometrics problems including serial correlation,

functional form and heteroscedasticity problems.

We proceed by estimating equations (1) and (2) with cointegration

technique to examine the long-run cointegration relationship between

economic growth and federal government debt with the inclusion of

other explanatory variables.

70 T

he

Rea

l E

ffec

t of

Gover

nm

ent

Deb

t:

Evid

ence

fro

m t

he

Mal

aysi

an E

con

om

y

Ta

ble

2.

Res

ult

s O

f O

rdin

ary L

east

Squar

es E

stim

atio

n o

n t

he

Impac

t of

Fed

eral

Go

ver

nm

ent

Deb

t an

d D

efic

its

to t

he

Eco

nom

y

R

eal

GD

P P

er C

ap

ita a

s th

e D

ep

en

den

t V

ari

ab

le

Inv

estm

en

t as

the

Dep

end

en

t V

ari

ab

le

(1

) (2

) (3

) (4

) (5

) (6

) (7

) (8

) (9

) (1

0)

ln (

Inves

tmen

t)

0.2

34

(0.0

59

)*

0.2

79

(0.0

58

)*

0.2

15

(0.0

60

)*

0.3

100

(0.0

57

)*

ln (

Lab

ou

r)

-1.3

21

(0.4

75

)8

-1.3

18

(0.4

72

)*

-0.8

50

(0.4

83

)**

-0.6

74

(0.2

57

)*

-1

.18

9

(1.0

46

) -1

.23

3

(1.0

49

) 0

.15

5

(1.0

51

) 0

.77

1

(0.5

73

)

ln (

Tra

de

Op

enn

ess)

0

.61

1

(0.0

80

)*

0.6

10

(0.0

79

)*

0.5

46

(0.0

79

)*

0.6

68

(0.0

73

)

0.0

10

(0.1

78

) 0

.01

7

(0.1

78

) -0

.14

8

(0.1

72

) 0

.19

1

(0.1

63

)

ln (

Fed

eral

go

ver

nm

ent

deb

t)

0.1

79

(0.1

11

) 0

.17

3

(0.1

11

) 0

.17

5

(0.1

05

)

0.4

08

(0.0

35

)*

0.5

27

(0.2

38

)*

0.5

19

(0.2

39

)*

0.4

13

(0.2

22

)

0.2

86

(0.0

49

)*

ln (

Fed

eral

go

ver

nm

ent

deb

t ^2

)

-0

.00

4

(0.0

01

)*

-0.0

10

(0.0

03

)*

Fed

eral

go

ver

nm

ent

def

icit

s -0

.00

0

(0.0

00

)

-0

.00

0

(0.0

00

)

0.0

00

(0.0

00

)

0

.00

0

(0.0

00

)

Inte

rcep

t 8

.06

2

(3.3

33

)*

8.0

46

(3.3

13

)*

5.3

12

(3.3

19

) 3

.31

0

(1.5

71

)*

3.4

217

(0.3

65

) 1

4.6

18

(7.1

62

)*

15

.053

(7.1

79

)*

5.6

71

(7.1

83

) 0

.72

1

(3.5

55

) 6

.71

1

(0.6

05

)

No

of

ob

serv

atio

ns

64

64

64

64

64

64

64

64

64

64

Ad

just

ed R

-Squ

ared

0

.89

5

0.8

96

0.9

05

0.8

92

0.7

44

0.4

08

0.4

04

0.4

92

0.3

70

0.4

09

D

iagn

ost

ic t

est

Ser

ial

Co

rrel

atio

n

4

1.6

47

*

41

.297

*

40

.066

*

42

.854

*

46

.908

*

47

.261

*

46

.140

*

40

.730

*

48

.647

*

46

.353

*

Fu

nct

ion

al F

orm

0

.30

6

0.3

68

0.7

45

0.1

02

17

.204

*

3.6

73

**

6.4

37

*

0.3

59

10

.298

*

4.0

33

*

Het

ero

sced

asti

city

11

.016

*

11

.127

*

8.7

78

*

9.1

15

*

23

.983

*

12

.474

*

13

.536

*

6.2

99

*

21

.999

*

16

.029

*

No

tes:

* a

nd

** d

eno

te s

ignif

ican

t at

5 a

nd

10

per

cen

t si

gnif

ican

ce l

evel

s. N

um

ber

s in

bra

cket

s re

pre

sent

the

rob

ust

sta

nd

ard

err

or.

The

seri

al c

orr

elat

ion t

est

is b

ased

on L

agra

nge

mu

ltip

lier

test

of

resi

dual

ser

ial

corr

elat

ion,

the

funct

ional

fo

rm t

est

is

bas

ed o

n R

am

sey’s

tes

t,

and

the

hete

rosc

edas

tici

ty t

est

is b

ased

on t

he

regre

ssio

n o

f sq

uar

ed r

esi

dual

s o

n s

quar

ed f

itte

d v

alue.

)4(

2

)1(2

)1(

2


A maximum lag of 4 (since our data involved a quarterly series) is

imposed for both specifications with real GDP per capita and investment

as the dependent variables. Results of the F-test are presented in Table 3,

where the real GDP per capita is the dependent variable, and the

computed F-statistics exceed the critical bound (at the 5 per cent

significance level) described by Pesaran et al. (2001) at a lag length of 3

for both estimated models (Model 1 only includes federal government

debt as the independent variable and model 2 includes other explanatory

variables). However, with the critical values provided by Narayan

(2004), which are robust for a small sample size, the computed F-

statistics exceed the critical bound (at the 5 per cent significance level)

at lag lengths of 3 and 4, implying a rejection of the null hypothesis of

no cointegration on the two estimated models.

Table 3. Long- Run Cointegration Test

Real GDP per capita as

dependent variable Investment as dependent variable

Model 1 Model 2 Model 3 Model 4

Independent variable (s)

Federal government debt

investment, labour force, trade penness and federal government debt

federal government

labour force, trade penness and federal government debt

F-statistic of bound test Lag 1 Lag 2 Lag 3 Lag 4

3.581 5.640 7.609* 4.342*

1.926 3.288 4.184* 2.721*

1.172 1.482 2.783 2.429

4.192* 6.064* 5.236* 6.421*

Pesaran et al. (2001) critical values 5 per cent 10 per cent

(4.934,5.764) (4.042,4.788)

(2.850,4.049) (2.425,3.574)

(4.934,5.764) (4.042,4.788)

(2.850,4.049) (2.425,3.574)

Narayan (2004) critical values 5 percent 10 percent

(3.803,4.363) (3.127,3.650)

(2.743,3.792) (2.323, 3.273)

(3.803,4.363) (3.127,3.650)

(2.962,3.910) (2.496,3.346)

Notes: * and ** denote significant at 5 and 10 per cent significance levels.

On the other hand, the computed F-statistics exceed the critical bounds

of Pesaran et al. (2001) and Narayan (2004) at the 5 per cent

significance level for model 4 where investment is the dependent



variable with the inclusion of labour force, trade openness and federal

government debt as the independent variables. This suggests a rejection

of the null hypothesis of no cointegration on the estimated model. Once

a long-run cointegration relationship has been established, we proceed to

the coefficients of the estimated model, which are revealed in Table 4.

In model 1, the federal government debt is found to have a positive and

significant (at 5 per cent significance level) effect in explaining the

variation in the country’s economic growth. However, the robustness

test indicates that the estimated model would be able to reject the null

of no serial correlation and heteroscedasticity, thus suggesting that the

estimation is biased and inefficient.

Table 4. Results of Cointegration Tests Between Gowth and Federal

Government Debt

ln(Real GDP per capita) as dependent variable

Model 1 ARDL (1,0)

Model 2 ARDL

(1,1,0,4,0)

ln (Investment) 0.312 (0.136)*

ln (Labour) -0.847 (1.179)

ln (Trade Openness) 1.279 (0.330)*

ln (Federal government debt) 0.316 (0.147)*

-0.424 (0.377)

Intercept 4.597 (1.805)*

2.650 (8.676)

Error correction term -0.122 (0.070)**

-0.232 (0.078)*

Diagnostic test Serial Correlation )4(2 15.247* 5.577 Functional Form )1(2 1.500 0.030 Heteroscedasticity )1(2 3.589** 0.066 No of observations 64 64 Adjusted R-Squared 0.922 0.658

Notes: * and ** denote significant at 5 and 10 per cent significance levels. Numbers in

brackets represent the robust standard error. The critical values are provided by

Pesaran et al. (2001), unrestricted intercept and no trend. All models include intercept

in the estimation. The null hypothesis of F-test is no long-run relationship. Numbers in

brackets represent the standard error. The ARDL model is selected based on Schwarz

Bayesian Criterion (SBC). The serial correlation test is based on Lagrange multiplier

test of residual serial correlation, the functional form test is based on Ramsey’s test,

and the heteroscedasticity test is based on the regression of squared residuals on

squared fitted value.


Meanwhile, despite the significant role of investment and trade openness

in explaining the country’s economic growth in model 2, there is no

evidence to support the role of federal government debt in the country’s

economic growth. In addition, this model best represents the real

scenario of Malaysia since none of the test statistics could reject the null

of no serial correlation, functional form and heteroscedasticity in the

model, hence suggesting that the estimation analysis is unbiased and

efficient. The error correction term coefficient is estimated at -0.122 and

-0.232 for model 1 and model 2 respectively, is statistically significant,

and has the correct sign, ensuring that the long-run equilibrium is

attainable. This suggests that economic growth is adjusting in a slow

phase, ranging from about 12.2 per cent to 23.2 per cent changes in the

explanatory variables before reaching its equilibrium.

To provide an in-depth investigation of the role of federal government

debt in the country’s economic growth, this paper tries to estimate from

different perspectives where the growth is measured by the investment

variable. Following the results of F-statistics that lie above the critical

upper bound and lower bound of Pesaran et al. (2001) and Narayan

(2004), the estimation of model 4 is presented in Table 5. In line with

the previous estimate, the federal government debt variable is found to

be insignificant in explaining Malaysia’s economic growth. Even though

none of the test statistics could reject the null of no serial correlation,

functional form and heteroscedasticity in the model, which implies that

the estimation analysis is unbiased and efficient, there is no evidence of

the role played by federal government debt in the country’s economic

growth.



Table 5. Results of cointegration tests between growth and federal

government debt

ln(Investment) as dependent variable

Model 4

ARDL

(1,0,2,2)

ln (Labour)

1.094

(2.448)

ln (Trade Openness) 0.660

(0.099)*

ln (Federal government debt) -0.155

(0.619)

Intercept -5.934

(17.646)

Error correction term -0.196

(0.057)*

Diagnostic test

Serial Correlation )4(2 2.034

Functional Form )1(2 0.007

Heteroscedasticity )1(2 1.106

No of observations 64

Adjusted R-Squared 0.474

Notes: * and ** denote significant at 5 and 10 per cent significance levels. Numbers in

brackets represent the robust standard error. The critical values are provided by

Pesaran et al. (2001), unrestricted intercept and no trend. All models include intercept

in the estimation. The null hypothesis of F-test is no long-run relationship. Numbers in

brackets represent the standard error. The ARDL model is selected based on Schwarz

Bayesian Criterion (SBC). The serial correlation test is based on Lagrange multiplier

test of residual serial correlation, the functional form test is based on Ramsey’s test,

and the heteroscedasticity test is based on the regression of squared residuals on

squared fitted values.

Under the linear assumption, this paper also examines the possible

existence of a short-run causality relationship between the interest

variables. The results are presented in Table 6. There is no evidence of

causality between the GDP and federal government debt or vice versa.

In addition there is no evidence of causal direction between investment

rate and federal government debt. On the other hand, the reported Wald

statistics are significant at 5 per cent and 10 per cent significance levels,

implying a rejection of the null hypothesis that real GDP per capita does


not cause the primary deficits. In other words, it can be argued that past

values of the real GDP per capita contribute to the prediction of the

fiscal position. In addition, the results also reveal a causal effect of

investment rate on the primary deficit at 5 per cent significance level.

These results would imply that any changes in the investment rate cause

the changes in the primary deficit position. Thus, growth is a

requirement for fiscal restraint in Malaysia.

Table 6. Results of bootstrap test for causality with endogenous lag

length choice

W-statistics Critical values Lag length

5 percent 10 percent

GDP => Federal government

debt

0.713 4.016 2.835 1

Federal government debt =>

GDP

0.265 4.044 2.769 1

GDP => Fiscal position 6.238* 4.047 2.890 1

Fiscal position => GDP 1.213 4.140 2.853 1

Federal government debt =>

Investment

0.289 3.998 2.769 1

Investment => Federal

government debt

0.107 4.030 2.847 1

Fiscal position => Investment 2.203 6.594 4.818 1

Investment => Fiscal position 7.946* 6.292 4.703 1

Notes: * and ** denote significant at 5 per cent and 10 per cent significance levels

respectively. The lag length selection is based on Schwarz's Bayesian criterion (SBC).

On the other spectrum, this paper uses an econometrics tool to

investigate the descriptive findings on the possible existence of a non-

linear effect of federal government debt. By employing the threshold

method of Hansen (2000), with 10,000 bootstrap replications, the results

for F-statistics and the p-value for the threshold model are reported in

Table 7. The F-statistics and the bootstrap p-value suggest a rejection of

the null of no thresholds effect, at 5 per cent significance level, of

federal government debt on economic growth. This shows the existence

of an inverted-U-shaped relationship between the federal government

debt stock and growth. The inverted-U relationship explains that an

increase in debt stock has a positive effect on economic growth until it



achieves the optimal level (up to a certain level). Beyond the threshold

level, an increase in the stock of indebtedness is associated with

negative federal government debt of RM362, 386 million for the overall

period of estimation.6

In other words, the result reveals that an increase of federal government

debt below RM362, 386 million is associated with an increase in

Malaysia’s economic growth. As the stock of federal government debt

increases, it is associated with a negative effect on the economy. The

empirical results obtained in this study suggest that Malaysia should

hold the federal government debt within the limit of RM362, 386

million. Intuitively, it can be seen that, with the current stock of federal

government indebtedness of RM456,128 (as at the end of 2011),

Malaysia has accumulated debt at about 25.92 per cent higher than the

optimal debt level. Furthermore, Malaysia has held the debt higher than

the optimal amount for the last eight quarters and is positioned in the

‘bad’ section of the “Laffer Curve”, which implies that accumulating

more borrowings would raise the risk of being trapped in the debt-

overhang situation.

6 The estimation has also been conducted with the variable transformed into natural

logarithmic terms. The results do not vary sensibly and are attached in Appendix 3.


Table 7. Results of threshold regression: Federal government debt as a

threshold variable between growth and federal government debt

Real GDP per capita as dependent variable

F-test statistics 368.50 443.91

Bootstrap p-value 0.000 0.000

113,260iq 17.386,362iq

Investment 0.033

0.005)*

Labour force -0.005

0.0740)

Openness 0.010

0.001)*

Federal government debt 0.014

(0.000)*

0.005

(0.001)*

Intercept 1818.3

(82.70)*

745.221

(641.321)

No of observations 46 56

R-Squared 0.949 0.982

113,260iq 17.386,362iq

Investment 0.001

(0.009)

Labour force -0.117

(0.111)

Openness 0.013

(0.003)*

Federal government debt -0.010

(0.001)*

-0.002

(0.003)

Intercept 9340.98

(501.06)*

3190.87

(671.84)*


R-Squared 0.685 0.963

Notes: * and ** denote significant at 5 and 10 per cent significance levels. The null

hypothesis is no threshold relationship. Numbers in brackets represent the standard

error.

To check the stability of the estimated parameters, this paper also

performs a Cumulative Sum of Recursive Residual test and a Recursive

Coefficient test, as depicted in Figure 2 and Figure 3 respectively. The

graphs show that none of the lines exceed the critical lines of 5 per cent



significance level, implying a non-rejection of the null hypothesis of

stability. In other words, the estimated equation is stable over the period

of study with a 5 per cent significance level.

Figure 2. Plot of cumulative sum of recursive residuals


Figure 3. Plot of recursive coefficient and the standard errors

Notes: the LINV, LLABOR, LOPEN, LFGDP represent investment, labour, trade openness and

federal government debt respectively.



5. Conclusion and policy implication

In this paper, the continuous increase in Malaysia’s federal government

debt as a result of a long episode of fiscal deficits (since 1957, with the

exception of the period 1993-1997) has underlined the urgency of

analyzing this issue. Countries with too much public debt may

potentially be trapped in a debt overhang situation, which could lead to a

default condition. Even worse, this would be associated with sovereign

debt for a series of countries in default. Thus, the objective of this paper

is to investigate the role of federal government debt in Malaysia’s

economic growth. Preliminary analysis shows that no role is played by

the federal government debt in Malaysia’s economic growth. However,

further analysis shows that there are non-linear relationships between

public and economic growth, thus suggesting an inverted-U-shaped

relationship. In other words, the accumulation of federal government

debt is positively associated with Malaysia’s economic growth up to an

optimal level. A novel aspect of this article is its recommendation of an

optimal level of federal government debt that the Malaysian government

should hold with respect to its economic growth rate. Furthermore, an

additional increase in federal government debt beyond the optimal level

has inversely contributed to the Malaysian economy. Moreover, the

findings also demonstrate that growth is a requirement for fiscal restraint

in Malaysia even though the analysis involves short-run causal effect.

The policy should consider a growth-driven approach to narrow the

deficit (or to reach a surplus fiscal position), thus reducing the stock of

federal government debt. This is important since, at this point of

analysis, the Malaysian federal government debt is in the ‘bad’ section

of the Laffer Curve where additional debt has an adverse effect on

growth.


References

Adam, C.S., and Bevan, D.L. 2005. Fiscal deficits and growth in developing

countries. Journal of Public Economies 4: 571-597.

Aizenman, J., Kletzer, K., and Pinto, B. 2007. Economic Growth with

Constraint on Tax

Revenues and Public Debt: Implications for Fiscal Policy and Cross-Country

Differences. NBER Working Paper No. 12750. Cambridge, MA.

Baum, A., Chevherita-Westphal, C., and Rother, P. 2012. Debt and Growth:

New Evidence for the Euro Area. European Central Bank Working Paper

Series No. 1450.

Brown, M. and Lane, P.R. 2011. Debt Overhang in Emerging Europe. Policy

Research Working Paper No. 5784.

Caner, M., Grennes, T., and Koehler-Geib, F., 2011. Finding the tipping point-

when sovereign debt turns bad. In: Braga and G. Vincelette, ed. Sovereign debt

and financial crisis, 63-76. Washington: World Bank.

Cecchetti, S.G., Mohanty, M.S., and Zampolli, F., 2011. The Real Effects of

Debt. BIS Working Papers No. 352.

Central bank of Malaysia, Monthly Bulletin 2012, Malaysia.

Central bank of Malaysia, 2010 Annual Report, Malaysia.

Checherita, C. and Rother, P., 2010. The Impact of High and Government Debt

On Economic Growth: An Empirical Investigation for the Euro Area.

European Central Bank Working Paper Series No. 1237.

Cordella, T., Ricci, L. A., & Ruiz-Arranz, M. 2005. Debt overhang or debt

irrelevance? Revisiting the debt-growth link. IMF Working Paper WP/05/223.

Hacker, S. and Hatemi-J, A. 2012. A bootstrap test for causality with

endogenous lag length choice: Theory and application in finance. Journal of

Economic Studies 39: 144-160.

Hansen, B. E. 2000. Sample splitting and threshold estimation. Econometrica

68: 575-603.



International Monetary Fund. 2012. Global Financial Stability Report. A

Report by the Monetary and Capital Markets Department on Market

Developments and Issues. International Monetary Fund, Washington DC.

Krugman, P. 1988. Financing vs. forgiving a debt overhang. Journal of

Development Economics 29: 253-268.

Malaysia Economic Planning Unit, Malaysia 2011/2012 Economic Report,

Malaysia.

Malaysia Economic Planning Unit, Malaysia 2009/2010 Economic Report,

Malaysia.

Miller, T. and Foster, J.D. 2012. Public debt, Economic freedom and growth.

In: Miller,

T, Holmes, K. R., and Feulner, E.J., ed. 2012 Index of Economic Freedom, 45-

55. United States: The Heritage Foundation.

Modigliani, F., 1961. Long-run implication of alternative fiscal policies and the

burden of the national debt. Economic Journal 71:730-755.

Pattillo, C., Poirson, H., and Ricci, L. 2004. What are the Channels Through

which External Debt Affects Growth?. IMF Working Paper No. WP/04/15,

International Monetary Fund, Washington, DC.

Pesaran, M. H., Shin, Y., and Smith, R. J. 2001. Bounds testing approaches to

the analysis of long-run relationships. Journal of Applied Econometrics 16:

289-326.

Presbitero, A.F. 2010. Total Public Debt and Growth in Developing Countries.

MOFIR Working Paper No. 44.

Reinhart, C.M. and Rogoff, K.S. 2010. Growth in a Time of Debt, NBER

Working Paper No. 15639. Cambridge, MA.

Reinhart, C.M, Reihnart, V.R. and Rogoff, K.S. 2012. Debt overhangs: Past

and Present, NBER Working Paper No. 18015. Cambridge, MA.

Schclarek, A. 2004. Debt and Economic Growth in Developing and Industrial

Countries. Lund University Working Paper No. 34.


APPENDIX

Appendix 1. The pattern of economic growth and federal government

debt

Source: Monthly Bulletin, Central Bank of Malaysia.



Appendix 2: Currency composition of federal government debt

RM 83.62%

USD 10.84%

Yen 4.77%

Others 0.78%

1999Q4

RM 96.03%

USD 2.45%

Yen 1.51% Others

0.02%

2011Q4


Appendix 3. Results of threshold regression: Federal government debt

as a threshold variable (in natural logarithmic term)

ln(Real GDP per capita) as dependent variable

F-test statistics 184.43 304.66

Bootstrap p-value 0.000 0.000

469.12iq 800.12iq

ln(Investment) 0.235

(0.024)*

ln(Labour force) -0.175

(0.239)

ln(Openness) 0.392

(0.037)*

ln(Federal government debt) 0.516

(0.023)*

0.226

(0.054)*

Intercept 2.148

(0.023)*

0.111

(1.756)


R-Squared 0.92 0.983

469.12iq 800.12iq

ln(Investment) 0.000

(0.053)*

ln(Labour force) -0.333

(0.268)

ln(Openness) 0.834

(0.210)*

ln(Federal government debt) -0.647

(0.078)*

-0.174

(0.282)

Intercept 16.898

(1.000)*

3.345

(0.853)*


R-Squared 0.70 0.966

Notes: * and ** denote significant at 5 and 10 per cent significance levels. The null

hypothesis is no threshold relationship. Numbers in brackets represent the standard

error.



Appendix 4. Breitung (2001) non-linear cointegration test

Real GDP per capita as

dependent variable

Investment as dependent

variable

Rank test 0.020* 0.019*

Score test 7.495* 1.240

Notes: * and ** denote significant at 5 and 10 per cent significance levels. Critical

values for the rank test statistics are from Breitung (2001) and the null hypothesis of no

non-linear cointegration is rejected for a test statistic value smaller than the critical

value.

Appendix 5: List of variables

Variables Descriptions Sources

Real GDP per capita Real Gross Domestic product

per capita (2000 constant

prices)

Monthly Bulletin, Central

Bank of Malaysia

Investment Gross fixed capital formation

(in RM Million)


Bank of Malaysia

Labour force Total labour force International Monetary

Fund/International Financial

Statistics

Openness Trade openness (exports plus

imports)


Bank of Malaysia

XX Exports of goods and

services


Bank of Malaysia

MM Imports of goods and

services


Bank of Malaysia

Federal government

debt

Total Federal Government

external debt (in RM

Million)


Bank of Malaysia

Fiscal surplus/ deficit (in RM Million) Monthly Bulletin, Central

Bank of Malaysia


Environmental Kuznets Curve for Deforestation in Indonesia:

An ARDL Bounds Testing Approach

Efendi Agus Waluyo1 and Taku Terawaki

2

The main objective of this study is to empirically demonstrate the inverse

U-shaped relationship, which is generally called the environmental Kuznets

curve (EKC), between economic development and deforestation rate in

Indonesia. For this purpose, we analyzed time-series data for Indonesia over 46

years from 1962 to 2007 with the autoregressive distributed lag (ARDL)

bounds testing approach to cointegration. Results support the long-run

inverted-U relationship, which implies that, while the deforestation rate

increases at the initial stage of economic growth, it declines after a threshold

point. The income turning point of the EKC was calculated to be US$ 990.4.

These findings derived solely from the time-series data for Indonesia provide

helpful information for the Indonesian government and policy-makers in the

sense that it explicitly indicates the specific tendency for that country.

1. Introduction

A “grow first, clean up later” approach, which means that only the

economic growth is targeted with little regard for its environmental

impact, is the basic strategy that have been taken by many developing

countries. Unfortunately, the rapid economic growth in this strategy has

often caused unprecedented environmental degradation in an early stage

of the growth especially. Tropical deforestation is one of the examples.

Since the forestry sector is a major contributor to the economy in the

developing countries most part of whose land is covered in forest, the

initial economic growth in those countries have naturally a direct and

negative impact on the forest ecosystem. Flood damage occurring in all

parts of the world will be one piece of clear evidence showing the fact.

There is no doubt that it is one of the greatest concerns of many

1 Forestry Research Institute, Forestry Research and Development Agency (FORDA),

The Ministry of Forestry Republic of Indonesia E-mail address: [email protected] 2 College of Economics, Ritsumeikan University, Japan

mailto:[email protected]

88 Environmental Kuznets Curve for Deforestation in Indonesia:


developing countries to know whether “grow first, clean up later” is a

costly strategy in a long-run view, and whether it can be a threat against

the sustainability of growth itself.

Indonesia, which has the most extensive forest area in the ASEAN

nations, has also suffered from massive and rapid destruction of the

forests for the last few decades, while the country has experienced the

economic growth acceleration by extracting natural resources in its

“grow first, clean up later” strategy. The Food and Agriculture

Organization of the United Nations (FAO) shows that about 30 % of the

forest cover area in Indonesia had been lost during the period 1962-2007,

although the GDP per capita had increased by five times and more

during the same time. However, such a negative correlation between

these two economic and environmental indicators is not always applied.

The inverse U-shaped relationship between them, which is generally

called “environmental Kuznets curve (EKC)”, is a key concept in

understanding the impact of economic growth on deforestation. The

evidence of the existence of EKC for deforestation would encourage

developing countries facing the problem of serious forest loss to

advance their economic development.

The main objective of this study is to empirically demonstrate the

inverse U-shaped relationship between economic development and

deforestation rate in Indonesia. We also figure out the turning point at

which the increase in income level does not lead to the increase in

deforestation rate. For these purposes, time-series data for Indonesia

over 46 years from 1962 to 2007 was analyzed with the autoregressive

distributed lag (ARDL) bounds testing approach to cointegration,

developed by Pesaran and Shin (1998) recently. Although many relevant

previous studies have so far tested the EKC hypothesis for deforestation

with cross-country or panel data (Shafik and Bandyopadhyay, 1992;

Koop and Tole, 1999; Bhattarai and Hammig, 2001; Culas, 2007), to our

knowledge, no study exists as yet that has shown the existence of the

EKC on deforestation by using the data for a single country. The

findings derived solely from the time-series data for Indonesia would

provide helpful information for the Indonesian government and

policy-makers in the sense that it explicitly indicates the specific

tendency for that country.


The paper consists of six sections. Following this introduction, the

second section provides the literature review on previous EKC studies.

The third section explains the empirical model specified in this study

and the data employed for the analysis. The ARDL bounds testing

procedure will be described in the forth section. The fifth section

discusses the results and discussion. The key findings are summarized in

the sixth section, and then we conclude this paper with further

discussion.

2. Literature review on the EKC

The environmental Kuznets curve is a theoretical concept that describes

the relationship between income growth and environmental degradation.

The term is named for Simon Kuznets (1955) who proposed that a

connection between economic growth and income equality is shaped as

an inverted U. This inverted U-shape hypothesis for environmental

indicators were first examined by Grossman and Krueger (1991). They

found in their study that the concentrations of two air pollutants out of

three (sulfur dioxide and “smoke”) increases at a low level of national

income and decreases at a higher level of income by analyzing the data

for 42 countries. Studies on the EKC following Grossman and Krueger

(1991) have exhibited an inconsistent tendency. Panayotou (1995) and

Song et al. (2008) showed that the EKC hypothesis was supported for all

the pollutants employed in those studies in common, while Grossman

and Krueger (1995), Akbostanchi et al. (2009), Shaw et al. (2010)

demonstrated that the inverted U is not necessarily described for all the

environmental indicators. More recently, it is reported that the inverse

U-shaped relationship regarding carbon dioxide was accepted for China

(Jalil and Mahmud, 2009; Jalil and Feridun, 2011) and for France (Iwata

et al., 2010), but it was rejected for Turkey (Akbostanchi et al., 2009;

Ozturk and Acaravci, 2010) and for Russia (Pao et al., 2011). We also

have to note that these inconsistent results could be caused by the other

factors, such as model specification and employed variables (Stern,

2004).

The EKC studies regarding deforestation have also produced various

findings. In the two pioneering papers, mixed results are reported.

Shafik and Bandyopadhyay (1992) failed to explain the EKC



relationship between income and two types of deforestation indicators

(annual deforestation and total deforestation). In Panayotou (1995)1, on

the contrary, an inverse U-shaped relationship appeared to hold between

forest area and GDP per capita from the cross-sectional data covering 68

countries. The income turning point was estimated to be about US$ 800

in his research. Several studies that investigated the existence of the

EKC for each continent also do not show regular patterns. While

Bhattarai and Hammig (2001) suggested that there was a strong

evidence of the EKC relationship between income and deforestation for

all the three continents of Latin America, Africa, and Asia, Cropper &

Griffiths (1994) indicated that the EKC hypothesis was supported for

Latin America and Africa, but not for Asia. In the more recent research

by Culas (2007), the inverted-U shape was statistically accepted only for

Latin America. The existence of the EKC for Africa or Asia did not

result in being significant in that study. In addition, Koop and Tole

(1999) were unable to reject the hypothesis that the country-specific

coefficients of GDP and GDP squared were vary across the countries for

each of Latin America, Africa, and Asia. This implies the necessity of

estimating an individual EKC with the data for each single country, as

well as the difficulty of obtaining a single EKC relationship among all

the countries in a region.

3. Model specification and data

This study estimates the deforestation equation that describes the factors

affecting deforestation in Indonesia, by using time-series data for that

country over 46 years from 1962 to 2007. The empirical model is

specified as

(1) 𝐷𝐸𝐹 = 𝛼 + 𝛽1𝐺𝐷𝑃𝑡 + 𝛽2(𝐺𝐷𝑃𝑡)2 + 𝛽3𝑃𝑂𝑃𝐺𝑅𝑊𝑡

+ 𝛽4𝑅𝑃𝑂𝑃𝑡 + 𝛽5𝐴𝐺𝐼𝑡 + 𝛽6𝐴𝐺𝐿𝑡 + 𝛽7𝑅𝑊𝑂𝑂𝐷𝑡

+ 𝛽8𝐹𝑂𝑅𝐸𝑋𝑃𝑡 + 𝑢𝑡 ,

where DEF is the annual rate of deforestation, GDP is gross domestic

product per capita, POPGRW is population growth, RPOP is rural

population, AGI is agricultural index, AGL is agricultural land area,

RWOOD is roundwood production, FOREXP is forest products export,

1 The original paper was published in 1993 (Panayotou, 1993).


and ut is a stochastic error term. The subscript t refers to year t. The DEF,

which is the dependent variable in equation (1), is calculated as

(2) DEF = (Ft-1 – Ft) / Ft-1

where F is forest cover area. The data on forest cover area comes from

FAOSTAT released by FAO. For the GDP, which is the most important

data in the explanatory variables, we employ the real GDP per capita

converted into US dollars that is obtained from World Bank. The EKC

hypothesis for deforestation in Indonesia would be accepted, when the

coefficient of GDP is positive and the coefficient of GDP2 is negative in

equation (1).

As carried out by many previous EKC studies on deforestation, we also

include variables other than income as explanatory variables, because

the causes of deforestation are considered to be complex and interlinked.

In our analysis, the significances of population, agricultural, and forestry

factors, as well as income, are inspected. First, the variables of

population growth (POPGRW) and rural population (RPOP) are

included in the model to examine the impact of population pressure on

deforestation. Those variables have been widely used in previous

empirical studies (Cropper and Griffiths, 1994; Bhattari and Hammig,

2001; Barbier and Burgess, 2001; Culas, 2007). We obtained the data on

population growth from World Bank, and the data on rural population

from FAOSTAT. Population pressure can increase the demand for forest

products or alternative land uses that causes deforestation, but it might

also work so as to reduce the deforestation, inducing technological

progress or institutional changes in agricultural or forestry sector (Culas,

2007). Second, we add the variables of agricultural land area (AGL) and

agricultural production index2 (AGI) to the list of explanatory variables.

The purpose of adding them is to illustrate how the deforestation in

Indonesia is connected with the increase in agricultural production. The

two major strategies to promote agricultural production are the

expansion of agricultural land into forests and technological

improvement in agriculture. The AGL and AGI are employed as proxy

variables for them, respectively. Third, the model comprises roundwood

2 This index shows the relative level of the aggregate volume of agricultural

production for each year in comparison with that of the base period of 1999-2001.



production (RWOOD) and forest products export (FOREXP) as the

variables expressing the forestry factors of deforestation. These can be

the direct determinants that raise the deforestation rate. The sources of

data on agricultural and forestry variables are all FAOSTAT.

4. The ARDL bounds testing procedure

The ARDL bounds testing approach to cointegration developed by

Pesaran and Shin (1998) is often applied by EKC studies in recent years

(Jalil and Mahmud, 2009; Shahbaz et al., 2010; Iwata et al., 2010; Jalil

and Feridun, 2011). While many macroeconomic variables are

integrated of order zero (I(0)) or one (I(1)), this approach is applicable,

even in the case that explanatory variables have different orders of

integration, as long as it is less than two. In addition, it is argued that the

ARDL approach to cointegration gives better results for small sample

data, as compared to other techniques, such as Engle and Granger (1987)

and Johansen and Juselius (1990) (Haug, 2002).

The first step of this ARDL approach is to establish the long-run

relationship among variables by estimating an unrestricted error

correction model. In this study, the model is specified as

(3) ∆𝐷𝐸𝐹𝑡

= 𝛼 + 𝛽0𝐷𝐸𝐹𝑡−1 + 𝛽1𝐺𝐷𝑃𝑡−1 + 𝛽2(𝐺𝐷𝑃𝑡−1)2

+ 𝛽3𝑃𝑂𝑃𝐺𝑅𝑊𝑡−1 + 𝛽4𝑅𝑃𝑂𝑃𝑡−1 + 𝛽5𝐴𝐺𝐼𝑡−1

+ 𝛽6𝐴𝐺𝐿𝑡−1 + 𝛽7𝑅𝑊𝑂𝑂𝐷𝑡−1 + 𝛽8𝐹𝑂𝑅𝐸𝑋𝑃𝑡−1

+ ∑ 𝛿𝑖𝛥𝐷𝐸𝐹𝑡−𝑖

𝑝

𝑖=1+ ∑ 𝜃𝑖𝛥𝐺𝐷𝑃𝑡−𝑖

𝑝

𝑖=1

+ ∑ 𝜇𝑖𝛥(𝐺𝐷𝑃𝑡−𝑖)2

𝑝

𝑖=1+ ∑ 𝜋𝑖𝛥𝑃𝑂𝑃𝐺𝑅𝑊𝑡−𝑖

𝑝

𝑖=1

+ ∑ 𝜌𝑖𝛥𝑅𝑃𝑂𝑃𝑡−𝑖

𝑝

𝑖=1+ ∑ 𝜎𝑖𝛥𝐴𝐺𝐼𝑡−𝑖

𝑝

𝑖=1

+ ∑ 𝜏𝑖𝛥𝐴𝐺𝐿𝑡−𝑖

𝑝

𝑖=1+ ∑ 𝜑𝑖𝛥𝑅𝑊𝑂𝑂𝐷𝑡−𝑖

𝑝

𝑖=1

+ ∑ 𝜔𝑖𝛥𝐹𝑂𝑅𝐸𝑋𝑃𝑡−𝑖

𝑝

𝑖=1+ 𝑢𝑡,

where α is the drift component, and ut is the white noise error

component. The null hypothesis that there is no cointegration among the


variables is expressed as β0=β1=β2=β3=β4=β5=β6=β7=β8.

We can conclude that there is a cointegration relationship among them,

if the calculated F-statistics is more than the upper critical bound given

by Pesaran et al., (2001). If the F-statistics is lower than the lower

critical bound, then it is judged that there is no cointegration. The

decision regarding cointegration will be inconclusive, when the

F-statistic lies within the upper and lower critical bounds.

Once a cointegration relationship among the variables is established, the

next step is to obtain the long-run equilibrium equation for deforestation

and its determinants. We can derive the reduced-form solution from

equation (3) as

(4) 𝐷𝐸𝐹𝑡 = 𝜆0 + 𝜆1𝐺𝐷𝑃𝑡 + 𝜆2(𝐺𝐷𝑃𝑡)2 + 𝜆3𝑃𝑂𝑃𝐺𝑅𝑊𝑡

+ 𝜆4𝑅𝑃𝑂𝑃𝑡 + 𝜆5𝐴𝐺𝐼𝑡 + 𝜆6𝐴𝐺𝐿𝑡 + 𝜆7𝑅𝑊𝑂𝑂𝐷𝑡

+ 𝜆8𝐹𝑂𝑅𝐸𝑋𝑃𝑡 + 𝑢𝑡 ,

where 𝜆0 = − 𝛼 𝛽0⁄ , 𝜆1 = − 𝛽1 𝛽0⁄ , 𝜆2 = − 𝛽2 𝛽0⁄ , 𝜆3 = − 𝛽3 𝛽0⁄ ,

𝜆4 = − 𝛽4 𝛽0⁄ , 𝜆5 = − 𝛽5 𝛽0⁄ , 𝜆6 = − 𝛽6 𝛽0⁄ , 𝜆7 = − 𝛽7 𝛽0⁄ , and

𝜆8 = − 𝛽8 𝛽0⁄ . On the other hand, the short-run dynamics is described in

the form of an error correction model (ECM) as

(5) ∆𝐷𝐸𝐹𝑡

= ∑ 𝛿𝑖𝛥𝐷𝐸𝐹𝑡−𝑖

𝑝

𝑖=1+ ∑ 𝜃𝑖𝛥𝐺𝐷𝑃𝑡−𝑖

𝑝

𝑖=1

+ ∑ 𝜇𝑖𝛥(𝐺𝐷𝑃𝑡−𝑖)2

𝑝

𝑖=1+ ∑ 𝜋𝑖𝛥𝑃𝑂𝑃𝐺𝑅𝑊𝑡−𝑖

𝑝

𝑖=1

+ ∑ 𝜌𝑖𝛥𝑅𝑃𝑂𝑃𝑡−𝑖

𝑝

𝑖=1+ ∑ 𝜎𝑖𝛥𝐴𝐺𝐼𝑡−𝑖

𝑝

𝑖=1

+ ∑ 𝜏𝑖𝛥𝐴𝐺𝐿𝑡−𝑖

𝑝

𝑖=1+ ∑ 𝜑𝑖𝛥𝑅𝑊𝑂𝑂𝐷𝑡−𝑖

𝑝

𝑖=1

+ ∑ 𝜔𝑖𝛥𝐹𝑂𝑅𝐸𝑋𝑃𝑡−𝑖

𝑝

𝑖=1+ ψ 𝐸𝐶𝑇𝑡−1 + 𝑢𝑡 ,

where the ECT is an error correction term.

To evaluate the goodness of fit of the model, we use several criteria.

These include classical assumption test, R-squared and adjusted



R-squared, lowest standard error of regression, lowest AIC, lowest SIC,

and model stability test. The technique employed to test model stability

is cumulative sum (CUSUM) and cumulative sum of squares

(CUSUMSQ). If the plots of CUSUM and CUSUMSQ statistics stay

within the critical bounds of 5 % level of significance, the null

hypothesis that all of the coefficients in the given regression are stable

cannot be rejected.

5. Results and discussion

5.1. Preliminary Examination

This study performs the conventional Augmented Dicky Fuller (ADF)

test and the Philip-Perron (PP) test to ensure that none of the variables

are I(2) or beyond. Table 1 shows the results of these unit root tests for

each variable. In the ADF test, the Swarchz Baysian Criterion (SBC)

was used to determine the optimal lag length. As shown in Table 1, the

results of ADF tests indicate that most of the variables are non-stationary

and have a unit root, while only the population growth (POPGRW) and

rural population (RPOP) are stationary at level. The results of PP tests

are also consistent with those of ADF tests. We can conclude from these

results that there is only a mixture of I(0) and I(1) among underlying

regressors. Hence, the ARDL bounds testing approach to cointegration

can be applied in this analysis (Duasa, 2007).


Table 1. Results of Unit Root Tests

Variable ADF test at level ADF test at first difference

None Intercept Trend

and

intercept

None Intercept Trend and

intercept

DEF 0.8275 0.8311 0.0825 0.0001*** 0.0008*** 0.0046 ***

GDP 1.0000 0.9984 0.3002 0.0008 *** 0.0003 *** 0.0012 ***

GDP2 0.9999 0.9998 0.8206 0.0004 *** 0.0008 *** 0.0011 ***

POPGWR 0.0003*** 0.2921 0.3893 0.2443 0.0672 * 0.6523

RPOP 0.0001*** 0.0014*** 0.5224 0.1673 0.9972 0.2186

AGI 1.0000 1.0000 0.8896 0.3503 0.0000 *** 0.0000 ***

AGL 0.9606 0.9575 0.6402 0.0000 *** 0.0000 *** 0.0002 ***

RWOOD 0.0000*** 0.9868 0.7543 0.0155 ** 0.0000 *** 0.0000 ***

FOREXP 0.9625 0.9396 0.2940 0.0000 *** 0.0000 *** 0.0000 ***

PP test at level PP test at first difference

DEF 0.9184 0.9025 0.4585 0.0001 *** 0.0016 *** 0.0096 ***

GDP 1.0000 0.9984 0.4418 0.0010 *** 0.0004 *** 0.0015 ***

GDP2 0.9997 0.9995 0.8903 0.0004 *** 0.0007 *** 0.0014 ***

POPGWR 0.0372 ** 0.9916 0.1150 0.3569 0.5262 0.8978

RPOP 0.9429 0.0342** 1.0000 0.3358 0.9925 0.1833

AGI 1.0000 1.0000 0.8701 0.0012 *** 0.0000 *** 0.0000 ***

AGL 0.9505 0.9410 0.6402 0.0000 *** 0.0000 *** 0.0003 ***

RWOOD 0.0000 *** 0.9906 0.7626 0.0001 *** 0.0000 *** 0.0000 ***

FOREXP 0.9769 0.9590 0.3034 0.0000 *** 0.0000 *** 0.0000 ***

Note1: Reported values are p-values for testing the null hypothesis that the variable has unit root.

Note 2: The symbols ***, **, and * indicate 1, 5, and 10 percent of significance, respectively.

The step of discovering the long-run relationship among explanatory

variables requires an adequate lag length of them in order to remove any

serial correlation. The optimum lag length of vector autoregressive

(VAR) is usually selected based on AIC, SBC, and likelihood ratio (LR)

test statistic. From the values of each criterion presented in Table 2, we

can choose order 2 in this study. Pesaran and Shin (1998) and Narayan

(2005) also have suggested that we should choose 2 as the maximum

order of lags for annual data in the ARDL.



Table 2. Selection Criteria for Lag Length

Lag LogL LR FPE AIC SC HQ

0 -2847.4 NA 4.04E+46 132.8558 133.2244 132.9918

1 -2445.04 617.5763 1.41E+40 117.9088 21.5951* 119.2682

2 -2307.94 3.0454* 1.73e+39* 5.2994* 122.3032 17.8822*

Note 1: The symbol * indicates the lag order selected by the criterion.

Note 2: LR: Sequential modified LR test statistic (each test at 5% level); FPE: Final

prediction error; AIC: Akaike information criterion; SC : Schwarz information

criterion; HQ: Hannan-Quinn information criterion

Table 3. Test Results of Granger Causality

Null Hypothesis F-Statistic Prob.

GDP does not Granger Cause DEF 6.45599 0.0038

DEF does not Granger Cause GDP 0.31788 0.7296

GDP2 does not Granger Cause DEF

3.80436 0.0309

DEF does not Granger Cause GDP2

0.79056 0.4607

A major concern in the analysis of EKC hypothesis is whether the GDP

has an impact on the deforestation or the contrary. We therefore

conducted the Granger causality tests to ascertain the direction of

causality, before testing the cointegration. As presented in Table 3, the

results indicate that GDP per capita Granger causes deforestation rate in

the long run at the 1 % level of significance. The same results were also

observed for GDP per capita squared at the 5 % level.

5.2. Results of bounds testing

The F-statistic calculated under equation (3) for testing whether the

variables are cointegrated or not was 1.515 as shown in Table 4. This

value is lower than the lower bound critical value at the 10 % level of

significance given by both Pesaran et al. (2001) and Narayan (2005),

which implies that null hypothesis of no cointegration cannot be rejected.

Hence, it is concluded that DEF is not cointegrated with GDP, GDP2,

POPGRW, POPDEN, AGI, AGL, RWOOD, FOREXP.


Table 4. Result of Cointegration Test (including all the variables)

Equation DEF = f (GDP, GDP2, POPGRW, RPOP, AGI, AGL, RWOOD, FOREXP)

Lag structure ARDL (2,2,2,2,2,2,2,2,2)

F-statistic 1.515425

Significance

level

Bound critical values of Case III (Unrestricted intercept and no trend)

Pesaran et al. (2001) Narayan (2005)

I(0) I(1) I(0) I(1)

1% 3.41 4.68 4.030 5.598

5% 2.62 3.79 2.922 4.268

10% 2.26 3.35 2.458 3.647

Since no evidence of cointegration was detected among all the variables,

we next attempted to narrow the variables down.We finally selected the

variables of RPOP, AGI, and RWOOD, as well as GDP and GDP2,

based on the statistical significance of those variables. The F-statistic

calculated again (=3.564) was higher than the upper bound critical value

provided by Pesaran et al. (2001) at the 10 % level of significance. This

implies that the null hypothesis of no cointegration can be rejected at the

10 % level, but it has not been supported yet according to the Narayan’s

critical values.

Then, we eliminated insignificant variables, except for the level

variables and the intercept, with the general-to-specific approach

(Krolzig and Hendry, 2001). The results are displayed in Table 5. The

F-statistic (=4.183) is higher than the upper critical value at the 5 %

level of significance given by Pesaran et al. (2001), and at the 10 %

level by Narayan (2005). It suggests that the variables are cointegrated,

and confirms the existence of long-run relationship among them.



Table 5. Result of Cointegration Test (excluding POPGRW, AGL, FOREXP)

Equation DEF = f (GDP, GDP2, RPOP, AGI, RWOOD)

Lag structure ARDL (2,1,1,1,1,0)

F-statistic 4.182707

Significance

level

Bound critical values of Case III (Unrestricted intercept and no trend)

Pesaran et al. (2001) Narayan (2005)

I(0) I(1) I(0) I(1)

1% 3.41 4.68 4.030 5.598

5% 2.62 3.79 2.922 4.268

10% 2.26 3.35 2.458 3.647

5.3. Long-run and short-run coefficient estimates

The estimated long-run coefficients are presented in Table 6. All the

explanatory variables included in this equation significantly affect the

deforestation rate. The positive coefficient of GDP and the negative

coefficient of GDP2 support the existence of the inverse U-shaped

relationship between economic growth and deforestation rate. This

finding is consistent with the empirical evidence of Panayotou (1995)

and Bhattarai and Hammig (2001). The income turning point (ITP) is

calculated to be US$ 990.4 from these coefficient estimates. This value

lies within the range of the GDP data set employed in this analysis,

which suggests that the ITP has already been reached in Indonesia. This

result is in line with several previous studies showing that ITP for

deforestation is placed within the sample range (Panayotou, 1995;

Kallbekken, 2000; Bhattarai and Hammig, 2001). Panayotou (1995) also

found that the ITP for deforestation in developing countries was

US$ 823. The value of ITP obtained in this research is extremely close

to the Panayotou’s finding.

The coefficient relating rural population to deforestation rate was

negative and significant at the 1% level. This suggests that an increase in

rural population in Indonesia tends to decrease the deforestation rate.

The same tendency was also found in several previous studies. In

Cropper and Griffiths (1994), and Bhattarai and Hammig (2001), rural

population density had a negative effect on the deforestation in Asian


region. In addition, Reis and Guzman (1994) obtained the negative sign

of the coefficient of rural population in the case of Amazon deforestation.

Culas (2007) also detected a negative coefficient of population density.

Templeton and Scherr (1999) noted that population pressure on forest

resources will increase at first, but it will change along with efficiency in

production processes into the direction of the conservation of the

remaining forest resources. This result might be related to the

technological or institutional innovation induced from population

pressure.

As to the agricultural indicators, the AGI significantly affects the

deforestation rate in a negative way. This implies that an increase in

agricultural production does not promote the conversion of forest lands

to agricultural lands, and that the increase has been led by improving

technology in agriculture. Technological progress in agriculture must

reduce the pressure on land demand and slow down the speed of

deforestation.

Roundwood production was also significantly connected to the

deforestation rate in Indonesia. This negative coefficient of RWOOD

indicates that the deforestation rate decreases with increasing log

production. Allen and Barnes (1985) that examined the effect of wood

use on forest area change over 1968-78 in developing countries, also

found a negative coefficient of wood use variable. This result may be

closely associated with that the data on roundwood production used in

this study is legally reported one. The roundwood products reported

legally are probably the ones that come from the forest managed

sustainably, and thus they cannot be a cause of deforestation.

Unfortunately, many log products have not been officially reported and

some of them are illegally produced. There is also evidence that large

amounts of timber traded in the world market are harvested illegally

(Hembery et al., 2007). The increase in illegal logs may decrease the

production of legal logs, and may cause higher deforestation at the same

time.



Table 6. Estimation Results of Long-run Model

Dependent Variable = DEF

Variable Coefficient T-statistic Prob.

Intercept 8.7418016 4.010585*** 0.0003

GDP 0.0266419 4.066877*** 0.0003

GDP2 -1.345E-05 -3.404278*** 0.0018

RPOP -0.0001057 -3.582405*** 0.0011

AGI -0.0332558 -2.893453*** 0.0068

RWOOD -2.092E-08 -3.141136*** 0.0036

Diagnostic Checks

Jarque-Bera 1.2114 (0.5457)

Serial Correlation LM 0.5612 (0.5764)

Heterocedasticity Test 0.9185 (0.5349)

Note1: ARDL (2,1,1,1,1,0) was selected on the basis of AIC.

Note2: The symbol *** indicates 1 percent of significance.

The three diagnostic tests of LM test, normality test of residual term,

and White heteroscedasticity test was also conducted in this step. The

results show that the long-run model has passed all the diagnostic tests

successfully. This indicates that there is no serial correlation, the residual

term is normally distributed, and there is no evidence of White

heteroscedasticity.

The results of short-run dynamics are presented in Table 7. The signs of

coefficients of GDP and GDP2 support the EKC hypothesis at the 1 %

level of significance. Only roundwood production variable was

insignificant, which implies that the change in log production does not

affect the change in deforestation in the short run. The coefficient of

lagged ECT is statistically highly significant, and its sign and size are

also reasonable, since it is generally required to be greater than -1 and

less than 0. The coefficient estimate of that variable suggests that

deviation from long-run equilibrium is corrected by nearly 57 % within

a year.


Table 7. Estimation Results of Short-run Model

Dependent Variable = ΔDEF

Variable Coefficie

nt

T-statistic Prob.

Intercept 0.463402 4.046573*** 0.0002

ΔGDP 0.014194 3.554423*** 0.0010

ΔGDP2 -7.52E-06 -3.168890*** 0.0030

ΔRPOP 0.000505 5.236923*** 0.0000

ΔAGI -0.029510 -3.919570*** 0.0004

ECT(-1) -0.568800 -5.676490*** 0.0000

Diagnostic Checks

Jarque-Bera 0.8867 (0.6418)

Serial Correlation LM 0.7740 (0.4689)

Heterocedasticity test 0.9760 (0.4634)

Note1: ARDL (2,1,1,1,1,0) was selected on the basis of AIC.

Note2: The symbol *** indicates 1 percent of significance.

Note3: ECT = DEF - 0.0266419*GDP + 1.345E-05*GDP2 + 0.0001057*RPOP +

0.0332558*AGI + 2.092E-08*RWOOD - 8.7418016

The last stage of ARDL bounds testing approach is to check the stability

of parameter estimates included in the model. In order to test the

stability, cumulative sum (CUSUM) and cumulative sum of squares

(CUSUMSQ) tests are generally performed. Figure 1 exhibits the plots

of CUSUM and CUSUMSQ, respectively. We can see from this figure

that the statistics are well within the critical bounds, which means that

all the parameter estimates in the model are stable.



Figure 1. Plots of Cumulative Sum (CUSUM) and Cumulative Sum of Squares

(CUSUMSQ) of Recursive Residuals

6. Conclusion

Out results support the long-run inverted-U relationship between

economic growth and deforestation rate in Indonesia. It implies that,

while the deforestation rate increases at the initial stage of economic

growth, it declines after a threshold point. Regarding previous study

conducted by several researchers (Shafik & Bandyopadhyay, 1992;

Cropper & Griffiths, 1994; Koop & Tole, 1999; Bhattarai & Hammig,

2001; Barbier & Burgess, 2001; Culas, 2007), the results of EKC for

deforestation in Asia are still debatable due to variety of data and

methodology. All of them used cross-country analyses and panel data

analyses. In addition, some studies used relatively small sample size

that might made the EKC hypothesis was not supported in some

studies in Asia. The income turning point of the EKC was calculated

to be US$ 990.4. This estimated ITP lies within the range of the data

on GDP employed in this analysis, which means that the ITP has

already been reached in Indonesia. In addition, we found that rural

population, agricultural index, and roundwood production have a

negative and significant impact on the deforestation. These results

suggest in order that (1) the deforestation might be restrained by

technological or institutional innovation in agricultural or forestry

sectors induced from population pressure in rural area, (2)

technological progress in agriculture reduce the pressure on land

demand, and then would slow down the speed of deforestation, and

(3) there is a possibility that the increase in “illegal” logs, which are

not reported officially, cause higher deforestation. The analysis of

-20

-15

-10

-5

0

5

10

15

20

76 78 80 82 84 86 88 90 92 94 96 98 00 02 04 06

CUSUM 5% Significance

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

76 78 80 82 84 86 88 90 92 94 96 98 00 02 04 06

CUSUM of Squares 5% Significance


short-run dynamics also reveals that the deviation from the long-run

equilibrium is quickly adjusted.

While the EKC results obtained from cross-country information

clearly shows an inverse U-shaped relationship as a whole between

economic development and environmental degradation, they would be

insufficient for each developing country facing the issue of whether

the development of the country is sustainable or not to be

optimistically confident that “grow first, clean up later” strategy will

work well in that country. As described above, several previous

studies analyzing these data, in fact, have revealed that there exist

different EKCs among continents. In terms of practical policy-making

on sustainable development, it would be necessary to test the

existence of the specific EKC for each country. This study definitely

demonstrates the usefulness of adopting the ARDL approach in the

evaluation of EKC hypothesis for a single country.

Due to data restriction, it was not possible to use provincial data in

this study. Using provincial data would provide a more valuable

insight into policy making, because they can introduce the effect of

region to the model. Another weakness of this study is that there is no

policy variable in the empirical model. Policy variables that could be

included are, for example, international environmental agreement,

enforcement of environmental legislation, and green project policy

(e.g. forest and land rehabilitation movement). The addition of these

variables would facilitate explaining what kinds of policies can

reduce deforestation.



References

Akbostanci, E., Türüt-Asik, S., Tunc, G. I., 2009. The relationship

between income and environment in Turkey: Is there an environmental

Kuznets curve? Energy Policy, 37 (2), 861-867.

Allen, J.C., Barnes, D.F., 1985. The causes of deforestation in

developing countries. Annals of the Association of American

Geographers, 75 (2), 163-184.

Barbier, E.B., Burgess, J.C., 2001. The economics of tropical

deforestation. Journal of Economic Survey, 15 (3), 413-433.

Bhattarai, M., Hammig, M,. 2001. Institutions and the environmental

Kuznets Curve for deforestation: A cross-country analysis for Latin

America, Africa, and Asia. World Development, 29 (6), 995-1010.

Cropper, M., Griffits, C., 1994. The interaction of population growth and

environmental quality. The American Economic Review, 84 (2), 250-254.

Culas, R.J., 2007. Deforestation and the environmental Kuznets curve:

An institutional perspective. Ecological Economics, 61 (2-3), 429-437.

Duasa, J., 2007. Determinants of Malaysian trade balance: An ARDL

bound testing approach. Journal of Economic Cooperation, 28 (3),

21-50.

Engle, R. F., Granger, C. W. J., 1987. Cointegration and error correction

representation: estimation and testing. Econometrica, 55 (2), 251-276.

Grossman, G. M, Krueger, A. B., 1991. Environmental impacts of a

North American free trade agreement. NBER Working Paper 3914. MA:

National Bureau of Economic Research, Cambridge.

Grossman, G.M., Krueger, A.B., 1995. Economic growth and the

environment. Quarterly Journal of Economics, 110 (2), 353–377.


Haug, A.A., 2002. Temporal aggregation and the power of cointegration

test: A Monte Carlo study. Oxford Bulletin of Economics and Statistic,

64 (4), 399-412.

Hembery, R., Jenkins, A., White, G., Richards B., 2007. Illegal logging:

Cut it out! The UK’s role in the trade in illegal timber and wood product.

WWF UK Illegal logging report.

Iwata H., Okada, K., Samreth, S., 2010. Empirical study on the

environmental Kuznets curve for CO2 in France: The role of nuclear

energy. Energy Policy, 38 (8), 4057–4063.

Jalil, A., Mahmud, S.F., 2009. Environmental Kuznets curve for CO2

emissions: A cointegration analysis for China. Energy Policy, 37 (12),

5167-5172.

Jalil, A., Feridun, M., 2011. The impact of growth, energy and financial

development on the environment in China: A cointegration analysis.

Energy Economics, 33 (2), 284-291.

Johansen, S., Juselius, K., 1990. Maximum likelihood estimation and

inference on cointegration with applications to the demand for money.

Oxford Bulletin of Economics and Statistics, 52 (2), 169-210.

Kallbekken, S., 2000. An alternative environmental Kuznets curve

approach to deforestation. Disertation in Environmental Economics and

Environmental Management. University of York, UK.

Koop, G., Tole, L., 1999. Is the an environmental Kuznets Curve for

deforestation? Journal of Development Economics, 58 (1), 231-244.

Krolzig, H-M., Hendry, D.F., 2001. Computer automation of

general-to-specific model selection procedures. Journal of Economic

Dynamics and Control, 25 (6-7), 831-866.

Kuznets, S., 1955. Economic growth and income inequality. The

American Economic Review 45 (1), 1-28.



Narayan, P.K., 2005. The saving and investment nexus for China:

Evidence from cointegration tests. Applied Economics, 37 (17),

1979-1990.

Ozturk, I., Acaravci, A., 2010. CO2 emissions, energy consumption and

economic growth in Turkey. Renewable and Sustainable Energy Reviews,

14 (9), 3008-3014.

Panayotou, T., 1993. Empirical Tests and Policy Analysis of

Environmental Degradation at Different Stages of Economic

Development. Working Paper WP238: Technology and Employment

Programme, ILO.

Panayotou, T., 1995. Environmental degradation at different stages of

economic development. In: Ahmed, I. and J.A. Doeleman (Eds.),

Beyond Rio: The environmental crisis and sustainable livelihoods in the

third world. ILO, MacMillan Press Ltd., London, pp. 13-36.

Pao, H.-T., Yu, H.-C., Yang, Y.-H., 2011. Modelling the CO2 emissions,

energy use, and economic growth in Russia. Energy, 36 (8), 5094-5100.

Pesaran, M.H., Shin, Y., 1998. An autoregressive distributed lag

modeling approach to cointegration analysis. A revised version of a

paper presented at the Symposium at the Centennial of Ragnar Frisch,

The Norwegian Academy of Science and Letters, Oslo, March 3-5,

1995. In: Strom, S. (Eds.), Econometrics and Economic Theory in the

20th

Century. Cambridge University Press, UK, pp. 371-413.

Pesaran, M.H., Shin, Y., Smith, R.J., 2001. Bounds testing approaches to

analysis of level relationship. Journal of Applied Econometrics, 16 (3),

289-326.

Reis, E., Guzman, R., 1994. An econometric model of Amazon

deforestation. In: Brown, K., Pearce, D.W. (Eds.), The Causes of

Tropical Deforestation: The economic and statistical analysis of factors

giving rise to the loss of the tropical. UBC Press, Vancouver, Canada, pp.

172-191.


Shafik, N., Bandyopadhyay, S., 1992. Economic growth and

environmental quality: Time-series and cross-country evidence. Policy

Research Working Paper. World Development Report.

Shahbaz, M., Jamil, A., Dube, S., 2010. Environmental Kuznets curve

(EKC): Time series evidence from Portugal. MPRA Paper No.27443, 15

October.

Shaw, D., Pang, A., Lin, C-C., Hung., M-F., 2010. Economic growth and

air quality in China. Environmental Economics and Policy Studies, 12

(3), 179-196.

Song, T., Zheng, T., Tong, L., 2008. An empirical test of the

environmental Kuznets curve in China: A panel cointegration approach.

China Economic Review, 19 (3), 381-392.

Stern, D.I., 2004. The rise and fall of the environmental Kuznets curve.

World Development, 32 (8), 1419–1439.

Templeton, S.R., Scherr, S.J., 1999. Effect of demographic and related

microeconomic change on land quality in hills and mountains of

developing countries. World Development, 27 (6), 903-918.


Determination of the Degree of Development and the Impact of the

Information Environment on the Formation of A System of Social

Control in Procurement Under the Russian Contract System (Method

of Content Analysis of Information Resources on the Internet)

N.A. Mamedova

1 and A.N. Baykova

2

The article is a result of research on the effect of environment on the

development of information system of social control in procurement.

Informatization process of social relations largely determines how certain

trends of social activity will be popular and durable. To determine the degree

of maturity of information content on the theme of social control in

procurement in Russia, the study used the method of content analysis of

information resources. To improve the quality of research results were

analyzed according to the semiotic and conceptual and thematic units of

content.

Introduction

Content analysis is a method of research aimed at quantitative analysis

of texts and text arrays for subsequent meaningful interpretation of

numerical patterns identified. The primary mechanism of the method is

the identification and measurement of the frequency of the use of formal

or substantive components were analyzed in total individual text or data

array. The degree of frequency of use in this sense becomes the desired

indicator. The method is used to visualize the dynamics of the intuitions

of the text content (information file) on a repetitive manner, estimates,

opinions and other forms of expression in order to be able to organize

these intuitive feelings, give them meaningful and reasoned explanation

of the subject. Using the method involves the development of targeted

approaches to the collection of data, representing the contextual textual

evidence on which the feeling of repetition and repetition frequency

based. However, the potential of the method of content analysis may be

1 Moscow State University of Economics, Statistics and Informatics (MESI), Moscow

Email: [email protected] (first author) 2 Moscow State University of Economics, Statistics and Informatics (MESI), Moscow


mailto:[email protected]

110 Determination of the Degree of Development and the Impact of the

Information Environment on the Formation of a System of Social Control

in Procurement Under the Russian Contract System

(Method of Content Analysis of Information Resources on the Internet)

represented by a much wider. Using the method, the researcher can not

only streamline their understanding of the text (information files), but

also to justify their conclusions, to interpret the author's position on the

nature of used them formal elements or structures, revealing even more

than the author would like to put into words. In this regard, the content

analysis method called the "scientific method of reading between the

lines."

Thus, a systematic approach to the study of the context, the desire for

objectification of the data analysis of the text or information set is non-

exclusive characteristic of content analysis and observed with the use of

other methods of word processing. However, the establishment of

quality parameters of the text (information files) through a quantitative

measurement of the formal elements and establish the relation between

the detected quantitative indicators is a characteristic feature of content

analysis as a method of scientific knowledge in the methods of analysis

and processing of texts. Therefore, the need to clarify the definition of

the method. Content analysis is a systematic quantitative analysis,

evaluation and interpretation of the form and content of information

sources. This type of analysis is primarily quantitative analysis reveals

quantitative regularities of repetition components of text, while in the

analysis may reveal structural patterns, qualitative laws by classification,

ranking and establishing a causal relationship between the obtained

results of content analysis.

Over other methods of analysis and processing of text content analysis

method has several advantages. For example, advantage is its

adaptability, which manifests itself in the possibility of using the method

without limitation as to the minimum and the maximum volume of the

analyzed information. Processability method also evident in its effective

integration with many other methods, in relation to which the method of

content analysis, and can act as a primary, and as collateral. This method

is most suitable for the primary processing of large volumes of

information and adjusting the flow of information on the stage of

collecting and processing raw data for both theoretical and applied

research. Also indisputable advantage is the possibility of formalization

and computerization of the process and the results of the content

analysis, due to its characteristics as a quantitative method of analysis

and text processing.


Results and Discussion

As mentioned above, the mechanism of the method of content analysis

is to calculate the frequency of occurrence of certain components in the

analyzed text or data arrays, which is complemented by the

identification of qualitative relationships (statistical methods) and

structural relationships (through the analysis of structural relationships

between the components). The result of the method is considered to

justify the existence of some analytes in quantitative and qualitative

characteristics. Obviously, the effectiveness of the method is due to the

choice of components, i.e., the choice of units of analysis. Requirements

for the unity of content analysis are obvious enough. Firstly, it should be

easily identified in the text. Secondly, a unit of content analysis should

enable the semantic interpretation, that is to be interesting and useful

content in the scale of the study. The tasks of the content analysis

conducted for the purpose of the study were a content analysis on the

structural-semiotic units (keywords) and the conceptual and thematic

units. To conduct content analysis were used open sources of

information and communication on the Internet. To clarify the

parameters of the sampling information it was decided to use the data set

generated by the search engines. When choosing a search engine we

used the following criteria: a high degree of relevance of search results;

optimal set of advanced search functions; low percentage of references

to duplicate content; absence (very low percentage) relevance of links

aimed at commercial sites. With all of these criteria best search engine

was determined search system Google.

To conduct content analysis on the structural-semiotic units and specific

to the research conducted as keywords accepted words the phrase "social

control in procurement."

Public control of procurement is defined by the law mechanism for the

rights of citizens, public associations, associations of legal entities to

control the legitimacy of the state and municipal customers. At its core,

social control in procurement is a form of interaction between civil

society and the state to ensure the functioning of the legal order of the

contract system in the procurement of goods, works and services for

state and municipal needs. The objectives of social control in

procurement include: a full submission of information, the development

of civil society, reducing corruption, development of markets and trade





relations. The implementation of these tasks is possible through the

creation of an open information space in which interactions (functional

level set), all participants of the contract system and society.

The legal basis for the functioning of social control in procurement in

Russia is so basic pieces of legislation: Federal Law № 112-FL "On the

basis of public control" and the Federal Law № 44-FL "On the contract

system in the procurement of goods, works and services for state and

municipal needs". Federal Law № 112-FL establishes the legal basis for

the organization and implementation of public control over the activities

of state authorities, local government, state and municipal organizations,

other agencies and organizations engaged in accordance with federal

laws separate public authority. In turn, the Federal Law № 44-FL

regulates relations directed at providing state and municipal needs in

order to improve efficiency, effectiveness of the procurement of goods,

works and services, ensure openness and transparency of such

purchases, the prevention of corruption and other abuses in the field of

procurement. And as one of the types of control procurement establishes

public oversight.

Thus, the legal framework of public control in procurement functions in

Russia, however, the activity of the subjects of public control is

insufficient. This study aims to assess the importance of social control in

procurement by analyzing the information field and placed it open

sources of information.

To evaluate the frequency of use of keywords in the information

resources, the results of extended context search were distributed in

accordance with the classification of information sources. In accordance

with the above keyword structural-semiotic research source was defined

thematic filter information resources (sites, portals). As classification

criterion, it was decided to use the theme filter in combination with the

degree of importance of the topic in the structure and content of web

pages, the totality of which was formed by the results of the extended

context search. Thus, information sources are classified as follows:

site (portal), specializing in the topic;

site (portal) wide profile information (main theme - procurement);


site (portal) independent profile - specialization relating missing

(news, panoramic, education);

site (portal) government or municipal authority;

site (portal) social organization, specializing on the subject of

procurement;

site (portal) social organization wide information profile;

site (portal) reference and the legal system;

links to educational materials, bills.

As the advanced search options in the search engine Google has taken

the following mandatory parameters (Table 1).

Advanced search options on the structural-semiotic units

Name of the parameter Basic / additional search

terms

The value of the selected

parameter

Select Keyword Keywords Public control procurement

Choice phrases Phrases Procurement for the needs

Search pages in the selected

language

Language web page In Russian

Search pages created in a

particular country

Country of creating web

pages

Russia

Search pages created or

updated within the

specified time

Date of creation / update Any

Search by text, title or

address of a page, as well

as links to them

Location words Anywhere on the page

Safe Search Blocking inappropriate

content

The function is activated

Search pages and files of a

certain format

File Format Either

Search for pages that are

free to use, distribute and

modify

Rights of use With any license





Materials and methods

Content analysis on the structural-semiotic units.

To evaluate the results obtained by sampling information into account

results in a certain time interval. Thus, the search results were classified

according to the three stages of the legal regulation of the procurement

activities in the Russian Federation: the information resources available

and updated up to 2012; Information Resources posted and updated in

the period 2012-2013; Information Resources posted and updated in

2014. Thus, the search covers a period of №94 federal law, the

transitional period for the formation of the contract system (2012-2013)

and the period of commencement of the contract system, regulated by

federal law №44 (2014). The total number of information sources that

match the search parameters, the source is 133. This number is the result

of contextual search and does not include the hidden results (results are

homogeneous (similar) presented the results of the search context).

Besides the classification of search results for the sources of information

and limitation of accommodation (update) web page for evaluating

advanced context sensitive searches used the following methods of

ranking results on the basis of scoring. The first direction ranking -

ranking Compliance common sense content of Internet pages (text)

keywords, i.e. structural-semiotic unit conducted a content analysis. This

approach allows taking into account the ranking parameter randomness

use keywords. The second direction of ranking is ranking in order of

importance of keywords in the content of the web page (text). Ranging

in this direction allows considering setting the frequencies of keywords

on a page (text). The order of distribution of GMAT results extended

context of keyword search in accordance with a first direction of the

ranking (in the parameter randomness use keywords) is defined the

following:

If the text is devoted to the topic of procurement, and the

keywords used in the description of social control as the main or

additional topics, and are not used in passing (in the list, listing,

for specifying, when specifying links to legal documents, etc.) -

assigned a score of three points;


If the text is devoted to the topic of procurement, but the key

words are used without description, detail that is casual (in the list,

listing, for specification, when specifying links to legal

documents, etc.) - assigned a score of two points;

If no text is devoted to the topic of procurement, and key words

are used without the description, detail that is casual (in the list,


documents, etc.) - assigned a score of one point.

The order of distribution of GMAT results extended context of keyword

search in accordance with the second direction ranking parameter

frequency of the use of keywords on a page is defined as follows:

If the text is devoted to the topic of social control in procurement

(permanently connected and used keywords - frequency of use of

keywords high) - assigned a score of three points;

If the text is partly devoted to the theme of social control, this

topic is the subject of this section in the overall structure of the

text (the frequency of the use of keywords is high or low

throughout the text, but always use a high frequency section (part

of the text) - assigned a score of two points;

If the text is devoted to the topic of the conjugate with the subject

of public scrutiny in procurement (the theme of fighting

corruption, the development of the contract system, the activities

of public associations, public policy and other topics); If keywords

are used without specification, detailing, that is casual (in the list,


documents, etc.) - assigned a score of one point.

As a result of the application of the methods of classification and

ranking of the sample to the resulting information by keyword "social

control in procurement" produced the following results of content

analysis on the structural-semiotic units:

1. Score from the structure of sources (indicator - the frequency of

the use of keywords by source type information). The most

frequently keywords "social control", "social control in





procurement" are used in the texts of analytical research papers on

the subject of public scrutiny and news reports in periodicals. The

analysis showed that in these materials the theme of social control

is the main. Social control in the framework of a contractual

change legislation considered in the comparative analysis of №44

and №94 of the federal law on the Ministry of the Russian

Government, in reference and legal systems, blogs. Sites of private

companies inform legislative change contract system, in particular,

attention is paid to public control. Keywords are also used in

textbooks on marketing and logistics, but the subject of public

scrutiny in procurement is not disclosed. Thus, periodicals,

reference and legal systems, the sites of governmental and

nongovernmental organizations are the main sources of

information on the topic of social control in procurement.

2. Evaluation of data on prescription accommodation (update)

information (parameter - the frequency of use of keywords). The

theme of social control has gained considerable resonance after the

reform of the legislation on the contract system in procurement.

Basically, all the texts that reveal the theme of social control in the

procurement and use of words containing contextual search, dated

2013, 2014 year. Most of the studied sources that use keywords at

least three times, contain outdated information, the texts published

before 2012.

3. Score from the dynamics of important topics (index - frequency of

use of keywords in the data stream). In consideration of

cumulative sources of social control is the theme of an

independent object, when keywords are used more than 10 times.

Public control is an additional theme when considering purchasing

system where keywords are used from 1 to 10 times (the amount

of text were taken into account). The use of keywords can also be

random when the overall context of the text does not match the

subject in question. Context analysis sources revealed that the

subject of public scrutiny in procurement basically not an

independent object of consumption (56.39% of the total number of

search results). In the context of the theme of the contract system

in the procurement of the theme of social control in the area of

procurement used in 34.58 of the total number of search results.

As an independent object use theme of social control is 9.02% of

the total number of search results (according to the table 2).


Results of content analysis on the structural-semiotic units

Number of occurrences

of keywords Number of sources

The proportion of the

total number of

sources,%

0 75 56,39

1-10 46 34,58

over 10 12 9,02

Total: 133 100

4. Evaluation of changes in the flow of information (index - the

quality of information). The theme of social control in

procurement is the main source of the 28 examined 133 sources.

Partially information devoted to public control in 45 sources.

While noting the coincidence of the main content of the text in

four cases (for example, the presentation of the text of the federal

law №44. Remains a source of no significance for the research

topics of social control and, in most cases, provided the

information they are not devoted to the topic of procurement.

Content analysis on the conceptual and thematic units.

As for the previous direction of content analysis of the direction

characterized by the use of keywords relating to ongoing investigations.

At this stage, phase content analysis Keywords are the words of the

phrase "social control." A narrower definition of the form of keywords

defined in order to expand the search capabilities and take into account

when analyzing the large number of results.

Conceptual-thematic units used in the content analysis of the words of

phrases that are directly or indirectly associated with the keyword. They

may have a direct association as disclosed subject of public scrutiny or

accompany the implementation of mechanisms of social control in

procurement practice. Indirect association occurs when the phenomena

associated with the theme of social control in the area of procurement,

are more extensive independent phenomenon. As a result, the sample





conceptual and thematic units for conducting content analysis were

formed by the following list:

anti-corruption expertise;

public examination;

public discussion;

public Policy;

the fight against corruption;

legal literacy;

citizenship;

civic engagement;

public initiative;

transparency of procedures;

transparency.

This list of conceptual-thematic units was formed by interviewing

through interviewing participants, simultaneously satisfying the

following requirements: non-professionally with procurement; not

members of associations and unions of legal entities; are not state or

municipal employees. In the process of interviewing was formed by a

set of associative phrases, while in the sample for the content analysis

was selected phrases that have the highest number of repetitions of the

participants.

Analyzing the list of conceptual and thematic units, it should be noted

that the sample included units that characterize the mechanisms of social

control - public debate, public examination, anti-corruption expertise, as

well as units that characterize the necessary conditions for the exercise

of social control - information openness, transparency procedures, and

public initiative. Also in the sample units were related to social control,

as a general private, ie units, which depends on the content and

implementation of the function of social control - civil, activity,

citizenship, legal literacy, public policy, the fight against corruption.

Thus, the final sample phrases for search are presented by keywords and

the conceptual and thematic units. Aware of the fact that the number

of results advanced context sensitive searches can vary over time,

despite the installed options, you must record the results on a specific


date. This date for the study is conducted August 1, 2014. The causes of

variation of search results are the following: adding information

resources in the Internet; update pages of existing information resources,

change the contents of the page; deleting pages of existing information

resources.

Implementation of a content analysis on the conceptual and thematic

units aims at the following objectives:

definition of search results for keywords "social control" and each

of the selected conceptual and thematic units (simple search -

search for one phrase);

definition of search results when combining keywords with each

of the selected conceptual and thematic units (advanced search -

search by two phrases);

defining relations of the results of keyword search and the search

results on the conceptual and thematic units (appraisal phrases

popularity to organize a search query);

the definition of (deviation) of the results of simple and complex

search (evaluation of search results for the pair relevance).

In assessing the results of the extended context based search described

above, the search parameters were used the following approach. The

first approach is to compare the result of a keyword search, and search

results on the conceptual and thematic units. Thus, each the result of

associative combinations compared with the results of keyword search

that allowed determining the ratio of the frequency of use of key words

and concepts and thematic units. The second approach is the comparison

of the results found for the conceptual and thematic units (the results of

a simple search) with the results of a complex search (search for the pair

relevance). Thus, each search result on each conceptual-thematic units

compared with the corresponding result of a complex search (when the

search query is generated using combinations of keywords phrases

conceptual-thematic units). It is possible to determine the link between

individual conceptual and thematic units and keywords.

Calculation of the ratio of the search result by keyword and conceptual-

thematic units held by calculating the difference between the number of

search results for keywords and the number of results for each





conceptual-thematic unity. For the absolute value was taken the number

of results based on keywords. At the date of August 10, 2014 the

number of results extended context of keyword search "social control"

was 1359 results. Calculation of the ratio of the results of simple and

complex search is carried out by calculating the percentage of the result

of a simple search to a result of a complex search. 100% was made the

number of results found for each simple conceptual and thematic unity.

Using the extended context search in the Google search engine, taking

into account the above parameters search yielded the following results

(Table 3).

The results of calculating the ratio of the search result by keyword and

conceptual-thematic units can be interpreted as follows. The greater the

deviation of the search result by the conceptual and thematic unity of the

result of a keyword search, the more popular, more semantic isolation

has the phrase conceptual-thematic units as compared to the keyword.

That is, the frequency of use in the context of a search query phrases

such high popularity due to the conceptual and thematic units, a wide

scope of use. Analysis of results of the extended context of search

(Table 3) showed that high semantic isolation have the following

conceptual and thematic units "transparent procedures", "public policy",

"public discussion", "information transparency", "public initiative".

Average values of semantic isolation showed the following conceptual

and thematic units, "citizenship", "civic engagement", "public

examination". Low values of semantic isolation showed the following

conceptual and thematic units, "anti-corruption expertise", "anti-

corruption", "legal literacy". These results also show the degree of

popularity of the conceptual and thematic units (within the parameters of

the search) in information and communication on the Internet, compared

with the popular keywords, and also among themselves.


The results of the extended context of keyword search and the conceptual

and thematic units

№

Option search

(keyword and

conceptual-

thematic units)

Search result

(considering the

received settings)

(simple search)

Result paired

relevance to the

keyword

(advanced

search)

Evaluation of

deviation (the

ratio of the

search result by

keyword and

conceptual-

thematic units)

Evaluation of

deviation (the

ratio of the

simple search

result with the

result of a

complex search

(search for the

pair relevance))

1 2 3 4 5 6

1 public control

1350 — 0 —

2

anti-corruption

expertise 652 153 - 698 23,5 %

3

public

examination 2300 117 + 950 5,1 %

4 public discussion

17100 402 + 15750 2,4 %

5 public policy

22100 452 + 20750 2, 0 %

6

the fight against

corruption 679 75 - 671 11,0 %

7 legal literacy

547 24 - 803 4,4 %

8 civil position

4100 74 + 2750 1,8 %

9 civic activity

3660 85 + 2310 2,3 %

10 public initiative

13400 178 + 12050 1,3 %

11

transparency of

procedures 30200 221 + 28850 0,7 %

12

information

transparency 13700 114 + 12350 0,8 %





The results of calculating the ratio of the results of simple and complex

searches can be interpreted as follows. Application of a complex search

query, comprising several phrases are always significantly narrows the

search area and the number of search results, respectively, since the

number of search results inversely proportional to the search terms. For

example, the number of results simple search on the word "fairy tale"

gives 23,900,000 results. Using complex search for "tales of Pushkin" is

1 300 000 results. The deviation in this case is 5.4%, despite the fact that

a simple search for "Pushkin" shows the result in 13.9 million hits. The

ratio of the result of complex and simple search can be very different,

vary from 100% to close to zero, but the main value of establishing a

relationship is that it gives definition to diagnose the connection

between the search queries.

To conduct content analysis, we define the following relation between

the result of complex and simple search. The smaller the amount of

deviation of the results of complex search (pair relevance - two phrases)

on the number of results on the appropriate conceptual and thematic

unity (ie, the larger percentage), the more stable relationship is observed

between the keywords "social control" and a separate conceptual and

thematic unit. In this case the deviation shows that the number of results

the search query in two phrases (keywords and conceptual-thematic

unity) slightly corrects (narrows) the number of results a search query on

a specific conceptual and thematic unity. Thus, a large number of

coincidences between the results of simple and complex research

suggests that in the information space, there is a significant number of

pages in the content of which there are complex search phrases desired

interconnected common sense.

Conversely, the greater the amount of deviation of the results of a

complex search (steam relevance - two phrases) on the number of results

on the appropriate conceptual and thematic unity (i.e., the smaller

percentage), the less stable relationship is observed between the

keywords "social control" and separate conceptual and thematic unity.

This dependence shows that the result of a complex search in large

communication distorts the results of a simple search on the conceptual

and thematic unity. For the purpose of this content analysis of this result

indicates that the number of pages of information and communication on

the Internet, which are the same and have the common sense of the


phrase keywords "social control" and the phrase on a separate

conceptual and thematic unity, slightly. This result demonstrates a low

degree of correlation between complex search phrases.

Because the number of conceptual-thematic units is defined as the

absolute number of the final sample, the analysis of the ratio of the

complex and the simple search is carried out exclusively within the

sample without relations with other possible search queries. Analyzing

the results of the extended context search in the table 3, it should be

noted that the total number of results regarding the high degree of

dependence is observed for the conceptual and thematic units' anti-

corruption expertise "(23.5%)," Fighting Corruption "(11.0% deviation)

and "public examination" (5.1%). All other results show a deviation of

5% or less. Indicators with an average degree of dependence (from 5%

to 2%) should be attributed to the following indicators conceptual-

thematic units "legal literacy" (4.4%), "public discussion" (2.4%), "civic

engagement" (2.3%), "public policy" (2.0%). Low degree of dependence

(below 2%) was observed for the conceptual and thematic units

"citizenship" (1.8%), "public initiative" (1.3%), "information

transparency" (0.8%), "the transparency of procedures "(0.7%).

If we compare the results obtained by the definition of semantic

isolation conceptual-thematic units of keywords and the results to

determine the quality of the connection between the conceptual and

thematic units and keywords, you can estimate the direction of the

relation between the results (Table 4)





Correlation of the results found for the extended context of semantic

degree of isolation and connection quality

The degree of semantic

isolation results simple

search

The results of

calculating the ratio of

the results of simple

search

The results of

calculating the ratio of

the results of simple

and complex search

The quality of

communication

between the results of

simple and complex

search

The high degree of sense

of separateness

"Transparent

procedures", "public

policy", "public

discussion",

"information

transparency", "public

initiative"

"Anti-corruption

expertise", "anti-

corruption", "public

examination"

High-quality

communication

The average degree of

sense of separateness

"Citizenship", "civic

engagement", "public

examination"

"Legal literacy", "public

discussion", "civic

engagement", "public

policy"

Average quality

communication

Low degree of sense of

separateness

"Anti-corruption

expertise", "anti-

corruption", "legal

literacy"

"Citizenship", "public

initiative", "information

transparency",

"transparent procedures"

Low quality

communication

Analysis of the results tables 4 to determine the presence inversely

proportional relationship between the degree of isolation of conceptual-

semantic unit’s thematic content analysis and quality of communication

between the key words and concepts and thematic units. One cannot

speak of the absolute values of the inverse proportional relationship, but

a definite trend correlation results advanced context sensitive searches

traced. Errors occur due to errors of the order of distribution of the

search results between levels (error application techniques comparison

purposes), and also due to the error operation of search when

determining the relevance of search results steam.

Establish the presence or absence of communication will allow

conducting correlation and regression analysis. To do this, we need to

calculate the correlation, linear regression to build and test the

hypothesis according to two related quantities. In our case, the

associated values are the results of a simple search and complex search

results. Recall that the condition is a simple search advanced contextual

search on the set parameters in one phrase (keyword "social control" and

the conceptual and thematic units). In turn, the complex search condition


is advanced contextual search on the set parameters on the phrase pair,

which consists of a keyword and a separate conceptual and thematic

units. For correlation and regression analysis is not required to use the

results of simple keyword search, analysis is carried out using only the

results of a simple search on the conceptual and thematic units. Thus,

there is a related sample of 11 pairs of values - the results of a simple

search (X) and the results of complex search (Y). Required:

calculate the correlation coefficient;

test the hypothesis according to the random variables X and Y,

with a confidence level of 98-99% (significance level - 0,02-0,01

respectively);

to calculate the coefficients of the linear regression;

build a scatter plot (correlation field) and the graph of the

regression line.

Calculation of the correlation coefficient.

The correlation coefficient is a measure of the probability of mutual

influence of two random variables. The correlation coefficient can take

specific values (-1 to +1). If the absolute value is closer to 1, it indicates

a strong connection between the values, and if closer to 0 - what it says

about the weak coupling or its absence. If the absolute value of the

correlation coefficient takes a value of 1, then we can talk about the

functional connection between random variables, that is, when one value

can be expressed by other means of a mathematical function. Compute

the correlation coefficient can be as follows:

,

xy x y

xy

x y

M MM

SSR

(1.1) To calculate the correlation coefficient is necessary to make a

table of values xk2, yk2 and xkyk (Table 5).





The values for the calculation of the correlation coefficient

Number

of search

options

(k)

Context search

options

conceptual and

thematic units

The result

of a simple

search (X)

The result of

a complex

search by

relevance (Y)

𝐗𝟐 𝐘𝟐 XY

1 2 3 4 5 6 7

1 anti-corruption

expertise

652 153 425104 23409 99756

2 public

examination

2300 117 5290000 13689 269100

3 public

discussion

17100 402 292410000 161604 6874200

4 public policy 22100 452 488410000 204304 9989200

5 the fight against

corruption

679 75 461041 5625 50925

6 legal literacy 547 24 299209 576 13128

7 civil position 4100 74 16810000 5476 303400

8 civic activity 3660 85 13395600 7225 311100

9 public initiative 13400 178 179560000 31684 2385200

10 transparency of

procedures

30200 221 912040000 48841 6674200

11 information

transparency

13700 114 187690000 12996 1561800

THE SUM OF

THE RESULTS 108438 1895 2096790954 515429 28532009

It is necessary to carry out calculations using formulas 1.2. and 1.3. The

calculations are shown in Table 6.

1

1 n

kxk

xn

M

, 1

1n

kyk

yn

M

, 1

1n

k kxyk

xyn

M

(1.2)

2 2 2

1

1n

k xxk

x Mn

S

, 2 2 2

1

1n

k yyk

y Mn

S

(1.3)

Journ

al o

f E

conom

ic C

ooper

atio

n a

nd D

evel

opm

ent

1

27

Cal

cula

tio

n o

f q

uan

titi

es b

y t

he

form

ula

s 1

.2.

and

1.3

.

№

The

proc

edur

e fo

r

calc

ulat

ing

Mx

My

S x2

S y

2

Mx

y

S x2

∗ S y

2

√S x

2∗

Sy2

R

xy

1 Ca

lcul

atio

n M

x;

My

; Mx

y; (

divi

ding

the

sum

of

the

resu

lts o

n th

e

num

ber

of s

earc

h

optio

ns)

9036

,5

157,

9166

6666

67

2377

667,

4166

6667

2 Ca

lcul

atio

n M

x;

My

; Mx

y;

(cal

cula

tion

of a

squa

re v

alue

Mx;

My

; Mx

y)

8165

8332

,25

2493

7,67

3611

111

2

5653

3023

4427

8,36

3 Ca

lcul

atio

n S x

2; S

y2;

(div

idin

g th

e su

m o

f

the

squa

re o

f th

e

resu

lts o

n th

e

num

ber

of s

earc

h

optio

ns a

nd

subt

ract

ing

the

squa

re s

um o

f th

e

resu

lts)

9307

4247

,25

1801

4,74

3055

5556

128 D

eter

min

atio

n o

f th

e D

egre

e of

Dev

elopm

ent

and t

he

Impac

t of

the

Info

rmat

ion E

nvir

on

men

t o

n t

he

Fo

rmat

ion

of

a

Syst

em o

f S

oci

al C

ontr

ol

in P

rocu

rem

ent

Under

the

Russ

ian C

ontr

act

Syst

em

(M

ethod o

f C

onte

nt

Anal

ysi

s of

Info

rmat

ion R

esourc

es o

n t

he

Inte

rnet

)

Cal

cula

tio

n o

f q

uan

titi

es b

y t

he

form

ula

s 1

.2.

and

1.3

. (C

ont.

)

4

Calc

ulat

ion

S x2

∗ S y

2

(пр

ои

звед

ени

е

S x2

на

S y2

)

1676

7086

4929

8

5 Sq

uare

root

extr

actio

n (t

he

squa

re ro

ot o

f th

e

prod

uct

of S

x2∗

S y2

)

1294

877,

8511

1106

6 Ca

lcul

atio

n of

the

valu

e of

the

corr

elat

ion

coef

ficie

nt

(Rx

y)

(Mx

y−

Mx*

My)/√

S x2

∗ S

y2

0,73

4164

5836


Thus, by calculating the correlation coefficient determined the random

variables, which amounted to 0.734. The resulting value represents the

presence of a stable connection between random variables. In our case

diagnosed stable relationship between the results of a simple search and

complex search results. However, since the estimate is made of the

correlation coefficient for the final sample of random values and may

therefore deviate from its general significance beyond the final sample

of random variables, it is necessary to check the significance of the

correlation coefficient. Thus, it is necessary to test the hypothesis of

random variables depending on the final sample.

Testing the hypothesis of random variables depending on the final

sample using the Student's t test of significance (t-test).

The random variable t follows t-distribution, so the table t-distribution

should find the critical value of the criterion (t kr.α) for a given level of

significance α (in our case α = 0,01 or 0,02). Calculate the random

variable t-criterion can be defined as:

,

2

,

2

1

xy

xy

R nt

R

(1.4)

If calculated according to formula 1.4. value t (in absolute value) will be

less than the critical value of the criterion (found on the table of the t-

distribution), the dependence between random variables X and Y is not.

If on the contrary, we must conclude that the experimental data of the

calculation of the correlation coefficient does not contradict the

hypothesis about the dependence of random variables.

Testing the significance of the correlation coefficient (to test the

hypothesis according to random variables) is shown in Table 7.

Verification made under the formula:

𝑡 = (𝑅𝑥𝑦 ∗ √𝑛 − 2)/√1 − 𝑅𝑥𝑦2 (1.5)

𝑅𝑥𝑦 = 0,713474472 𝑛 = 11





Test the significance of the correlation coefficient

Indicator Correlation

factor

(𝑅𝑥𝑦)

𝑅𝑥𝑦2 Number

of earch

options

(n)

Number of

degrees of

freedom of

Student's

(n-2)

√𝑛 − 2 Calculation

√1 − 𝑅𝑥𝑦2

Calculation t

criterion

value of

the index

0,7341645836 0,5389976358 11 9 3 0,6789715489 3,2438675147

Analyzing the results of testing the hypothesis according to random

variables should be noted that with high probability the experimental

data for the calculation of the correlation coefficient does not contradict

this hypothesis. This means that the result of the calculation of the

correlation coefficient of a stable correlation between random variables

finite sample can be extended to a general correlation coefficient outside

the final sample of random variables. To further study the nature of the

relationship between random variables necessary to calculate the

coefficient of the linear regression equation.

Calculation of the coefficient of the linear regression equation.

Linear regression equation is an equation of the line, about describing

the dependence between random variables X and Y. If we consider that

the value of X is free, and Y - a function of x, the regression equation

should be written as follows:

𝑌 = 𝑎 + 𝑏 ∗ 𝑀𝑥 (1.6)

In this case, the value of X (the simple search results) is indeed free

value, in turn, the magnitude of Y (complex search results) is completely

dependent on the value of X as a basis of the search results is the

complex combination of keywords and a notional focal units.

To calculate the coefficients of the linear regression - permanent

regression coefficient (a) and variable regression coefficient (b), it is

necessary to make calculations using formulas 1.7 and 1.8.

𝑏 = 𝑅𝑥𝑦𝛿𝑦

𝛿𝑥= 𝑅𝑥𝑦

𝑆𝑦

𝑆𝑥 (1.7)

𝑎 = 𝑀𝑦 − 𝑏 ∗ 𝑀𝑥 (1.8)


The procedure for calculating the coefficients of the linear regression is

shown in Table 8.

Calculation of the linear regression equation: 𝑌 = 𝑎 + 𝑏 ∗ 𝑋

Procedure for calculating the linear regression equation.

𝑀𝑥 𝑀𝑦 𝑅𝑥𝑦 𝑆𝑥2 𝑆𝑦

2 𝑆𝑥2/𝑆𝑦

2 √𝑆𝑥

2/𝑆𝑦2

coefficien

t a

coefficient

b

903

6,5

157,91666

66667

0,73416

45836

930742

47,25

18014,7430

555556

0,00019

35524

0,01391

23108

0,01021

39258

65,61852

58786

Linear regression equation has the form: 𝑌 = 65,6185258768 + 0,0102139258 ∗ 𝑋

To evaluate the prediction error is also necessary to make calculations,

the procedure for calculating prediction error Y for a given value of x in

accordance formulated 1.9 and 1.10 shown in the table.

𝛿𝑦/𝑥 = 𝛿𝑦√1 − 𝑅𝑥𝑦2 = 𝑆𝑦√1 − 𝑅𝑥𝑦

2 (1.9)

𝛿𝑦/𝑥 = 𝛿𝑦/𝑥

𝑀𝑦100% (1.10)

An error estimate for the linear regression equation.

Absolute error: 𝜎𝑥/𝑦 = √𝑆𝑥2 ∗ √1 − 𝑅𝑥𝑦

2

Relative error: 𝛿𝑦/𝑥 = (𝜎𝑥𝑦

𝑀𝑦) ∗ 100%

The calculation of prediction error for a given value Y X

𝑆𝑦 𝑅𝑥𝑦2 𝜎𝑥/𝑦 𝛿𝑦/𝑥

134,219011528 0,5389976358 91,1308901462 57,71%

Error equation:

𝜎𝑥/𝑦 = 91,1308901462

Further investigation of the dependence of random variables involves the

construction of scatter plots (correlation of the field) and the graph of the

regression line. The scattering diagram - a graphical representation of the

corresponding pairs of random variables X and Y in the form of points in





a plane with coordinates axes X and Y. In the same coordinate system and

construct a graph of the regression line. A special feature is the choice of

construction diagrams and scale of initial points on the axes, provides a

visual diagram. The procedure and results of the calculation of the

regression line points: point A with coordinates (Хmin; Ymin) and point B

with coordinates (Хmax; Ymax) is shown in Table 9.

Verifying plotting the regression line is performed by applying a

coordinate system point average values of the random variables X and Y

coordinates (Мх;Му).

Calculation of the regression line and the points on the linear

regression equation.

𝑌 = 65,6185258768 + 0,0102139258 ∗ 𝑋

Compute point on the regression line equation of the linear regression

Хmin Ymin Хmax Ymax

547 71,2055433074 30200 374,0790860334

Coordinates of the regression line:

A (547;71) ; B (30200;374)

Scatter plot and graph the regression line shown in Figure 1.

Figure 1. Scatter diagram (correlation field) and the graph of the regression line

B

A

Y

X

5

4

3

1

0

2

3 6

0

9 12 15 18 21 24 27 30


The distribution of points in pairs of random variables shows that they

correspond to the direction of the graph of the regression line, however,

the degree of approximation to the scattering points of the regression

line is not sufficient, which reduces the quality of the linear regression

equation. However, validation plotting a linear regression on the

coordinate system point average values of the random variables X and Y

coordinates (Мх;Му) showed perfect agreement with the coordinates of

the point (Мх;Му) generated from the linear regression line. Thus, the

lack of proximity of the scattering points (coordinates of points on the

results of a pair of simple and complex search) is due to an insufficient

number of random variables finite sample.

Findings

The results of correlation and regression analysis confirmed the presence

of a stable connection between the results of advanced context sensitive

searches as part of the content analysis showed an association between

the results of keyword search and conceptual and thematic units. It

speaks to the uniformity of information flow, which brings together

thematic pages on the Internet for all investigated units of content

analysis. Thus, the purpose of the content analysis by determining the

degree of development and the impact of the information environment

on the formation and implementation of methods of social control in

procurement under the Russian contract system achieved. The results of

the content analysis can be used to further study the dynamics of

changes in the interim results of content analysis to determine the areas

of development, structure and content of the flow of information on

topics related to information support of public control in procurement

and contract system development.





References

The Federal Law of 21.07.2014 № 212-FL "On the basis of public

control in the Russian Federation"

The Federal Law of 05.04.2013 № 44-FL (ed. By 04.06.2015) "On the

contract system in the procurement of goods, works and services for

state and municipal needs"


Bilateral Trade through Official Channel between India and

Bangladesh: An Analysis with the Use of Time Series Forecasting

Models

Muhammad Mahboob Ali1 and Anita Medhekar

2

Bilateral trade between India and Bangladesh will be mutually beneficial to

both the countries and improve welfare as per trade theory. This study has tried

to forecast impact of trade between two countries considering the time period

1991-2014.The researchers compared Autoregressive Moving Average

(ARMA) and Autoregressive Integrated Moving Average Model (ARIMA)

after testing the significance of Augmented Dicky fuller model’s intercept of

explanatory variables.While using ARIMA model the study found that best-fit

estimates were exports to India. However, forecasting indicates that import

from India to Bangladesh also had a positive impact over the time period 1991-

2014.By engaging in bilateral trade with India, Bangladeshi producers and

suppliers ought to be concerned about attaining long term sustainability in their

business, by improving quality of the products so that export can be raised in a

competitive manner. Market access to India will be beneficial only with having

competitive advantage, in bilateral trade. To export products, India should

allow duty free access to certain products from Bangladesh in which it has

competitive advantage. To maintain Pareto optimality in the bilateral trade,

Bangladesh should import products from India at competitive prices. This will

help to promote and nurture bilateral trade relations, ensure sustainability of

business and mutually benefit both the countries through free trade agreement.

1. Introduction

Trade between India and Bangladesh has a long history. Besides

bilateral trade with India, informal trade also plays an important role in

Bangladesh. Trade between these two countries is not only creating

value but acting as a value chain, which is a corner stone for improving

1 Professor, Faculty of Business and Economics, and Director, Institutional Quality

Assurance cell, Daffodil International University, Bangladesh. The author is former

Vice Chancellor of Presidency University, Bangladesh. Email: [email protected] 2 Senior Lecturer of Economics, Central Queensland University, Australia.


http://r.search.yahoo.com/_ylt=A86.J3Wh_YJUISEAXc0PxQt.;_ylu=X3oDMTByNzhwY2hkBHNlYwNzcgRwb3MDMgRjb2xvA2dxMQR2dGlkAw--/RV=2/RE=1417899554/RO=10/RU=http%3a%2f%2fwww.forecastingsolutions.com%2farima.html/RK=0/RS=uWeGj2yjMYcYeeabnpo8gm.sYaA-

136 Bilateral Trade through Official Channel between India and

Bangladesh: An Analysis with the Use of Time Series Forecasting Models

bilateral relationships. However, given the geographical proximity,

relationship with India is historical, cultural, social and cordial to have

any significant impact on the two economies. Therefore increasing bi-

lateral trade between the two neighbouring countries is very essential for

employment generation, economic development and growth.

Furthermore, according to Ankit(2015) India’s foreign policy would

appear to the Commonwealth Relations Office to be ‘a picture not only

of an ever enlarging sphere of regional co-operation but also of

expanding Indian ambitions’ is somewhat not valid in this century .Both

the countries should maintain diplomatic relationship not only for geo-

political reasons to mutually benefit from each other to ensure

sustainable economic development and growth through South Asian

Association of Preferential Trade Agreements (SAPTA)and the Bay of

Bengal initiative for Multi-Sectorial Technical and Economic

Cooperation (BIMSTEC),but also to fight against global and regional

terrorism, irradiate poverty, and improve welfare for millions. There are

three factors that are identified in case of Bangladesh, which have a

negative influence for encouraging trade: ease of doing business,

unrecorded informal trade and high transaction cost of engaging in trade.

(1) Ease of Doing Business: ‘Ease of Doing Business Index (EDBI) for

Bangladesh in 2014 was last measured at 173 and for India 142 out of

189 countries [see World Bank, (2014)]. As such Bangladesh still has a

long way to go for further development of easing business procedure,

which is to provide business friendly environment in this competitive

global state of the 21stcentury. Likewise India should also reduce its

protectionist trade policies between the South Asian nations. (2)

Informal Trade: World Bank (2015) stated that since the independence

in December 1971, there has been a substantial increase in informal

unrecorded trade across the India-Bangladesh land borders, and a

number of studies both in Bangladesh and in India have dealt with

different aspects of it. (3) Transaction Cost: The trade diverted through

the formal channels provide customs revenue, and this would be higher

if administrative and other reforms reduce the scope for corrupt

practices. Better infrastructure, faster clearance times and reduced

transaction costs would also improve the prospects of Bangladesh

exporters finding niche market in India, especially if they rely on

importing inputs from India, where there is two way border crossing for

trade [see World Bank (2015)].


Key factors that unite Bangladesh and India as identified by

Government of India (2013) are the following:

Both the countries share a common heritage- language,

civilisation, colonial history, social and, economic history.

India and Bangladesh have common interest and share a

common heritage for music, classical dance, literature, poetry

and the creative arts.

With Bangladesh, India shares not only a common history of

struggle for freedom and liberation but also enduring feelings of

both fraternal as well as family ties.

India played a major role in emergence of Independent

Bangladesh during the 1971 war, and it was also the first country

to recognize Bangladesh as a separate independent nation.

There have been major issues such as illegal migration, border,

water sharing disputes, Moore Island, which have had a negative

impact on bilateral trade in the recent years.

Ahmed(2015) argued that the Indian Prime Minister Narendra Modi

visited Bangladesh in June, 2015 for a mere 36 hours, but left an impact,

big enough to wipe away mistrust that had crept in the Indo-Bangladesh

relationship over the decades. The Indian Express (2015) quoting Joshi,

President of the Federation of Industry and Commerce, North-Eastern

Region of India commented that the agreement during Modi’s visit to

Bangladesh was on infrastructure development, which focuses on

connectivity by road, rail, air, river, sea, transmission lines, petroleum

pipelines and digital links. This development will have a multiplier

effect on the economy and provide a real boost to cross-border trade

between Bangladesh and North Easternpart of India. Sharma (2015)

reported that an investment of US$ 2 billion line of credit is extended to

Bangladesh, further to current US$ 1 billion. Most importantly, if

boundary disputes aresettled then the foundation to build trust for

infrastructure development is laid. This will be mutually beneficial for

trade between the two countries. During the visit most of the 22 bilateral

agreements were signed aimed at boosting trade and transport links.

Further, Indian corporate entities like Reliance Power and Ambani

Group signed agreements to invest around $5 billion which will help

Bangladesh to generate extra 4600 MW of power, to meet its electricity

demand.



Main objective as neighbouring countries should be to try to resolve all

cross-border disputes peacefully, without forgetting the 1971 war of

independence and soldiers who sacrificed their life from both sides for

independence, in order to mutually benefit from trade relationships and

maintain peace and harmony across border for economic development

and growth. According to sources of ministry of external affairs, there

has been progress in Indo-Bangladesh relations as per the following

initiatives taken by the government on both sides:

High level of recent contacts at government level, exchange and

visits.

Wide ranging people-to-people interaction at cultural and social

networking level.

Indian High commission in Bangladesh issues about half a

million visas every year and thousands of Bangladeshi students

study in India on self-financing basis.

Recipients of over one hundred annual Government of India

scholarships.

Bangladesh Prime Minister Sheikh Hasina in 2011, along with

Indian Prime Minister Dr. M. Singh announced the

commencement of 24-hour access across the Tin Bigha corridor

to Dahagram and Angorpota enclaves, as well as duty-free

import of 46 textile items (subsequently expanded to all items,

except 25 items) from Bangladesh to India.

Common vision for rural development, health, education, clean

drinking water and sanitation, people’s empowerment and

economic development issues discussed at the government level.

India has always helped Bangladesh in its hour of need with aid

worth over Taka 250 crore (over US $ 37 million) to help it cope

with natural disasters and floods in 2007-08.

Supply of 1,000 MT of skimmed milk powder, and 40,000

Million Tons of rice.

India completed and handed over 2,649 core shelters in the

affected villages in Bagerhat district in southern Bangladesh.

Line of Credit Agreement was signed in Dhaka on August 7,

2010 between EXIM Bank of India and Government of

Bangladesh. India has extended a line of credit of US$1 billion to

Bangladesh for a range of infrastructure and development

projects, including railway infrastructure, supply of BG


locomotives and passenger coaches, procurement of buses, and

dredging projects.

January 29 -2012, NTPC and BPDB set up a Joint Venture for

the establishment of a 1320-MW coal-based power plant in

Bagerhat district, Khulna at an estimated cost of $1.5 billion and

is to be commissioned by 2016.

India offers 100 places under ITEC and 35 under Technical

Cooperation Scheme of Colombo Plan every year to Bangladesh.

In the last three years (2006-07 to 2009-10), 414 participants

from Bangladesh underwent training in India under ITEC

Programme and Technical Cooperation Scheme of Colombo

Plan. Government of India gave Muktijoddha Scholarship to 200

Higher Secondary-level students and 478 Graduate-level

students. Further, in 2011 three Bangladesh Diplomats were also

imparted training at Foreign Service Institute in India.

Such bilateral Cultural Exchange Programme (CEP) 2009-2012 between

Bangladesh and India provides the platform for fruitful exchanges; given

the shared history and commonality of language; cultural exchanges

form an important bond of friendship between the people of two

countries. Special emphasis has been laid on promotion of cultural

exchanges in the fields of music, theatre, art, sports, painting and books,

such as: Joint celebrations of 150th anniversary of Rabindranath Tagore;

to honour the Indian friends of Bangladesh for their contribution to the

1971 Liberation War; through student-teacher exchange programmes

and reciprocal programmes of cooperation; promote people to people

exchanges, 100 scholarships are being granted by ICCR every year to

students from Bangladesh, and; In the year 2013-14, totalexport from

Bangladesh to India was worth US$ 457 Million, whiletotal import from

India to Bangladesh was worth US$ 5514 Million and total trade with

India was US$ 5971 Million.



Picture 1: Geographical location of Bangladesh in South Asia

Source: Google Map

According to Bammi (2010), India being geographically close to

Bangladesh and a larger country is an important partner from trade and

economic point of view, given the benefit of ease of travel, similar

culture, language and transportation between the two countries as seen

from Picture-1.Sutherland (2012) argued that indeed the 21st century

rush to promote bilateral trade agreements has been accompanied by a

rise of protectionism. Srinivasan and Vani (2009) argued that keeping in

mind that one cannot infer welfare effects directly from the trade

creation and trade diversion effects of preferential trade; they interpret

their results from the coefficient estimates from their gravity model of

export, import and total trade flows as broadly indicating that the pursuit

of preferential trade agreements is counterproductive. They concluded

that India’s superior policy option continues to be unilateral and

multilateral trade liberalization.

The research question for this study is to investigate whether it is

possible to have an optimal formal trade arrangement between the two

neighbouring countries India and Bangladesh? This is elaborated below

based on preliminary analysis of the data. Asteriou and Hall (2007)


argued that ARMA models can only be made on time series Yt that are

stationary. Time series is not constant over time, which means that the

series are non-stationary. If, after first differencing, a series is stationary,

then the series is also called integrated to order one, and denoted 1(1) –

which completes the abbreviation ARIMA.This paper is structured as

follows. Following the introduction, section two provides literature

review. Section three provides methodology followed by section four

which discusses the results of empirical analysis. Section five provides

conclusions, implications and future research directions.


Krugman (1980) rightly asserts that if two countries have the same

composition of demand, the larger country will be a net exporter of the

products whose production involves economies of scale. Bangladesh

capacity to attain competitiveness in trade with India is a big question,

due to low capacity building, as well as lack of competitiveness in the

formal trade and huge balance of trade deficit. Wangwe (1993) elements

of the new trade theories which are relevant to trade and development

issues pertaining to developing countries can be applied to India and

Bangladesh: the conception of process of narrowing the technology gap

between the developed and developing countries; implications on the

conception of North-South technology-related negotiations; therole of

multinational activities in the developing countries; intra-south trade and

investments; industrial dynamics and attainment of competitiveness; and

the role of governmentpolicy in enhancing competitiveness in the

economy. Etheir (2001) described that countries tend to trade a lot with

their neighbours, so it is sometimes said that current regional initiatives

between the two countries, often involving neighbours, are therefore

likely to be benign. Hassan (2002), emphasised that the geographical

proximity of India along with the increasing familiarity of Bangladesh’s

importers to India’s production capacities, which in recent years have

become globally competitive both in terms of price as well as quality

has made Indian products increasingly competitive in Bangladesh’s

market. He suggested that the only way to increase the volume of intra-

regional trade and reduce the trade deficit with India, Bangladesh should

take the following steps: devalue its currency, seek reduction in tariff

and non-tariff barriers on exports to India, stop cross-border smuggling

activities, eliminate structural and political rigidities and conflicts,

encourage more Indian investment into Bangladesh and make the



SAPTA more meaningful, effective and operational tool to reap the

benefits from integration.

Moreover there are positive benefits to the countries and the world, by

engaging in regional, bi-lateral and multi-lateral free trade [see Khan

(1999), Dutta (1999),and Jain (1999)]. However, intra-SAARC trade is

very small (Hassan, Mehanna and Bashar 2001) compared to other

regional blocks like ASEAN and NAFTA. Further, Hassan (2001)

empirical study, using the gravity model of international trade for years

1996-1997, concluded that the proportion of intra-regional trade

between the South Asian block of countries is very small due to “normal

outcomes or unexplored opportunity” (p.264), and if increased, can have

significant welfare improving benefits along with implementing

supporting policies by the governments to encourage preferential trade

agreement under South Asian Association of Regional Cooperation

(SAARC).Hassan (2001) results supports the argument that small

countries depend more on trade then larger and diversified countries;

poor countries trade less with each other, than with the rich countries,

and countries sharing common border, trade more with each other.

Therefore regional economic cooperation between these two countries

should be encouraged, given the similar social-economic and cultural

conditions. The major export items from a small country Bangladesh, to

a comparatively large country India within the SAARC region as

recorded by Dhaka Chamber of Commerce are the following: Woven

Garments; Knitwear; Home Textile; Agri-Products; Frozen Food;

Leather & Leather Products; Footwear; Raw Jute; Jute Goods; Bicycle

Major Import Items to Bangladesh from India are: Cotton (all types),

cotton yarn / thread and cotton fabrics; Cereals; Vehicles other than

railway or tramway rolling- stock and parts and accessories thereof;

Residues and waste from the food industries, prepared animal fodder;

Nuclear reactor, boilers, machinery and mechanical appliances parts;

Iron and steel; Edible vegetables and certain roots and tubers; Organic

chemicals ;Mineral fuels, mineral oils and products of their distillation,

bituminous substances, mineral waxes; Plastics and articles thereof;

Tanning or dyeing extracts, tannins and their derivatives, dyes, pigments

and other colouring matter, paints and varnishes, putty and other

mastics, inks; Salt, sulphur, earths and stone, plastering materials, lime

and cement [see Dhaka Chamber of Commerce (2015)].


Further items exported by Bangladesh are: Electrical machinery and

equipment and parts thereof, sound recorders and reproducers, television

image and sound recorders and producers and parts and accessories of

such articles; Man-made staple fibers; Dairy produce, birds' eggs natural

honey, edible products of animal origin, not elsewhere specified or

included; Coffee, tea, mate and spices; Rubber and articles thereof ;

Edible fruit and nuts, peel of citrus fruit or melons; Man-made

filaments; strip and the link of man-made textile materials ;Aluminium

and articles thereof; Knitted or crocheted fabrics; Inorganic chemicals,

organic or inorganic compounds of precious metal, of rare earth metals,

of radioactive elements for isotopes; Paper and paper board, articles of

paper pulp; Essential oils and resinoids; perfumery, cosmetic or toilet

preparation; Oil seeds and oleaginous fruits; miscellaneous grains, seeds

and fruits; industrial or medicinal plants; straw and fodder and

Pharmaceutical products [see Dhaka Chamber of Commerce (2015)].

However, the largest trading partners of Bangladesh are European Union

and North America in terms of legal exports and India for legal imports

since late 1990’s [see Hassan (2001)].

World Bank also reported in the year 2002 surveys that smuggled goods

were imported from India to Bangladesh during 2002/03 were worth

approximately $500 million, or about 40% of recorded imports from

India, and approximately 30% of total imports (recorded plus smuggled)

from India[see World Bank, (2006)].Pohit and Taneja (2003) argued that

informal trade continues to thrive because the transacting environment

of formal and informal trading arrangements gives rise to lower

transaction costs in the informal channel. Srinivasan (2002) depicted

that unless the transacting environment improves significantly for

formal traders, informal trade will continue to co-exist along with formal

trade.

Hossain and Rashid (1999) argued from their empirical study, that

Bangladesh’s trade with India is neither fair nor competitive due to trade

barriers .Bhagwati (1995) described that the restrictiveness of trade

barriers is therefore likely to have increased as required. Such elasticity

and also selectivity are in fact characteristics of the “administered”

protections embodied in antidumping actions and Voluntary Exports

Restraints (VERs) which make them both a preferred instrument of

protection by industry and also a serious barrier to free trade. Further,

the transacting environment of formal trade agreement between India



and Bangladesh indicates that the inefficiencies of the trade regimes

give rise to rent seeking activities by the authorities, bureaucrats and

politicians. That is formal traders prefer to use mechanisms of informal

trading to settle disputes[see Pohit and Taneja (2003)].

Ahmed (2006) very rightly pointed out that firstly, Bangladesh-India

relationship is faced with certain puzzles which need to be addressed

professionally and without any animosity. Secondly,the regional and

global scenarios have transformed the Indo-Bangladesh relationship in

several key areas, both for the good and the bad, therefore, not fully

realising the benefits of bilateral trade between neighbours who have

significant historical advantage. Khan and Khan (2003) suggest to have

open regionalism, that is outward oriented development policies by

merging regional trading blocs and harmonising domestic economic

policies with global economic policies, and extend the SAARC

integration to West Asia in Iran, to Burma in the East, to benefit from

trade, investment opportunities and develop economic links with the

world to improve on its socio-economic indicators.

Bhuyan (2006) observed that the root cause of Bangladesh’s trade

imbalance with India is the country’s narrow production base in both

exports and import substitutes. The country’s industrial sector being in a

rudimentary stage of development, cannot meet the growing demand of

the domestic market. The result is the country’s acute dependence on

imported supplies to meet domestic demand. Export production in

Bangladesh is also narrowly based and not diversified. Most of the

products that Bangladesh may exports to India are already produced by

India for domestic consumption and exports. Further, protectionist trade

policies of India prevent imports of few products from Bangladesh,

which are believed to have comparative advantage. World Bank (2006)

study noted that Bangladesh perennial large bilateral trade deficit with

India might be a cause for concern, but it has not led to any balance of

payments problem for Bangladesh mainly because of regular trade

surpluses with trading partners as US and EU which compensates for

these deficits with India. The large volume of informal/illegal trade

remains a problem though across the borders with Bangladesh. Sikdar,

Mohnen and Chakraborty (2006),have suggested a Free Trade Area

(FTA) option, and the governments of both these countries need to think

about this seriously. Such a free trade arrangement is likely to go a long

way towards deeper integration of the two South Asian countries, such


as freeing of trade in services, free flow of investment, trade facilitation,

harmonization and mutual recognition of standards and coordination of

macro-economic policies and solving other disputes related to border

and resources. In particular, it will produce substantial benefits for the

Bangladeshe conomy by improving its overall competitiveness through

access to the marketing network, skill and technology of Indian

manufacturers and trading partners. Pursell and Sattar (2006) also

argued that informal trade between India and Bangladesh have

consistently found a similar pattern, to the pattern of formal trade, which

is large volumes of goods being smuggled from India to Bangladesh, but

much smaller volumes being smuggled in the opposite direction. This

generally concludes that there is also a substantial Indian trade surplus

on informal account, which is confirmed by this study and is consistent

with the findings in the literature.

Sikdar, Mohnen and Chakraborty (2006) described that Bangladesh’s

trade deficit with India has increased substantially since the start of this

21stcentury. This has given rise to concerns at the government policy

level, as well as public perception of deteriorating relationships.

Moreover, Basu and Datta (2007) noted that Bangladesh has export

similarity with India and hence faces high export competitiveness. The

lack of match between Bangladesh export and Indian import also

generates a constraint of complementarity. Both the countries use

different trade-related indices like RCA and Cosine measures to

examine the extent of trade similarity and complementarity in inter-

industry bilateral trade. The possibility of intra-industry trade between

the two countries is also studied with the help of G-L indices. Export has

been found to be of random nature and trade deficit has a perverse

relation with exchange rate, driven by flow of foreign exchange

remittances from abroad. They suggested that Bangladesh should pursue

an appropriate exchange rate policy and aim at increased diversification

in her export structure, in order to avoid Dutch disease and to reduce the

bilateral trade deficit. The effect of falling exchange rate can be positive

on one hand as it increases exports, but also it can increase trade deficit

[Islam, Kham and Ishak (2013)].

De and Bhattacharyya (2007) advocated that India and Bangladesh need

to minimize transaction costs arising due to trade by removing visible

and invisible barriers to trade. Countries can tackle transaction costs

only through improved and integrated trading infrastructure, which is



responsible for faster movement of goods and services across the

countries. Dutta (2007) observed that Bangladesh has a large trade

deficit with India which has been increasing on average at the rate of 9.5

per cent annually, along with large volumes of informal imports from

India across the land border, to avoid Bangladesh import duties.

Suranovic (2010) pointed out that the “competitive market” creates an

incentive to satisfy consumer desires and demands. This is the ultimate

goal of any economic system. The greater the competition between trade

incentives, the greater will be the potential surplus generated out of the

process. Thus, a competitive market promotes the incentives that result

in greater economic efficiency. Alam, Uddin, Alam and Malakar (2009) argued that the statistical result

of Purchasing Power Parity (PPP) for Bangladesh with India and China

shows that the price of foreign country (India or China) has no

significant impacts on bilateral exchange rate and the price of home

country (Bangladesh) has opposite behaviour that PPP warranted.

Further, Rahman, Khan, Nabi and Paul (2010) opined that a number of

initiatives could be taken to stimulate bilateral trade between the two

countries. As the analysis has shown, abolition of sensitive list is likely

to have only an insignificant adverse impact on the Indian economy; but

also, mere duty‐free-quota‐free (DF‐QF) market access to India is not

likely to enhance Bangladesh’s export to India in any significant way.

Under these circumstances India should be persuaded to provide duty‐free market access for all exports originating from Bangladesh, and

likewise Bangladesh should put renewed emphasis on diversification of

her export basket in the Indian market, which is possible under SAPTA

and BIMSTEC (alliance of South and South East Asian countries).

Islam (2011) commented that North-Eastern region of India bordering

with Bangladesh, should be explored in a ‘creative fashion’ as

Bangladesh enjoys certain locational and comparative advantages with

regard to the North-Eastern part of India. On the other hand, closer

economic integration and physical connectivity with Bangladesh would

not just reduce the economic isolation of the region, but more

importantly these would also reduce the isolation of North-East with the

Indian mainland. This is what perhaps underlies in the India’s ‘Looking

at East’ policy. The development of Indian North-East is inextricably

linked with India’s political, economic, social and security issues with

the bordering nations to the East. The development of the North Eastern


region would, serve India’s strategic purpose. Both the nations should

therefore, recognise the fact that Bangladesh’s economic development is

in India’s interest and similarly India’s development and prosperity of

the North-East region, is in Bangladesh’s interest.

Acharya and Marwaha (2012) recommended that to develop and build

technological capacity, huge investments in research and development

and innovation is required. Hence the signing of Bilateral Investment

Promotion and Protection Agreement (BIPPA) between India and

Bangladesh was the right step in this direction to encourage Indian

investment into Bangladesh. Nevertheless, there are certain issues which

need to be addressed for creating a favourable investment climate in

Bangladesh: Developing single window clearance for new business

proposals; setting an Industrial Park for India in Bangladesh outside

EPZ with all the needed infrastructure facilities; upgrading the tax

holiday system; and harmonizing HS Code System. Bhagwati and

Srinivasan (2002) commented that it is difficult to agree with the many

critics of free trade that see the heavy hand of such globalization

casting its evil spell on the poorest of the poor countries. The empirical

truth seems to be exactly the opposite, that is international trade is

mutually beneficial.

De, Raihan and Kathuria (2012) described that countries like

Bangladesh and India can benefit greatly from opportunities created for

trade through economic cooperation. The scope for trade expansion

between the two countries depends partly on their trade

complementarities, which is relatively limited, but growing; partly on

account of their economic imbalance. The other driver of bilateral trade

is intra-industry trade between India and Bangladesh. This has the

potential to grow significantly, since trade in similar product lines has

been growing, and that could deepen production and supply chain

networks between the two countries. Due to the infrastructural

bottlenecks at the border is affecting India-Bangladesh bilateral trade.

This is due to the following reasons as discussed by Acharya and

Marwaha (2012):

Inadequate traffic planning which causes congestion at the sea

ports. Port congestion results in demurrage, which hikes the cost

of production.



Irregular and inadequate supply of electricity has forced many

firms to rely on power from captive generators which further

aggravates production costs.

Lack of adequate infrastructure facilities at the Petropole Border

where majority of the trade is routed through the Petropole

(Indian side) - Benapole Border.

The road - rail connectivity is poor and there is lack of

alternative transport options.

Limited air cargo and container service (especially from / to

Benapole and Darshana)

Inadequate and limited facilities at the existing Land Border

Stations due to undeveloped infrastructure, causing delays in

valuation and clearance at Land Customs Stations.

Inadequate warehousing, cargo handling equipment, customs and

immigration facilities, and means of communication at some of

the road and rail based land ports on both sides. Bangladesh

Land Port Authority (BLPA) has recently taken initiatives to

develop the necessary infrastructures through the public-private

partnership.

Rahman, Ahamad, Islam and Amin, (2012) suggested that Bangladesh’s

policymakers should give higher priority to increase domestic

agricultural production and supply side capacities in items which have

already demonstrated their export potential in the Indian market. In this

regard, closer collaboration between research institutions of Bangladesh

and India will enable Bangladesh to access and benefit from transfer of

modern agricultural technology from India.Basher (2013) empirically

found that Bangladesh’s exports to India are highly responsive to

changes in the competitiveness of the country as reflected in real

exchange rate movements. Assuming Ceteris Paribus, a one percent

increase in competitiveness is likely to increase Bangladesh’s export to

India by about 8 percent. A one percent increase of Indian GDP is found

to be associated with0.8 percent rise in exports from Bangladesh to

India. Their findings indicate that improved competitiveness and

economic growth are significant predictors for exports. While policy

induced measures such as exchange rate management can be a difficult

option, as enhanced external competitiveness can be achieved through

tackling supply side hindrances.


Bhardwaj (2014) argued that FTA between Bangladesh and India has to

be signed on a priority basis because the huge trade gap has always been

a matter of concern between the two countries. Indian formal exports to

Bangladesh amounted to about US$4 billion (with an additional US$4

billion of ‘informal’ exports), while Bangladesh’s total exports to India

was around US$350 million. However, unless the para-tariff and non-

tariff barriers are completely removed, the trade cooperation will be

below its full potential level. Bown (2014) was concerned whether a

multilateral system with fully enforceable, time-invariant, free trade

would be possible or even desirable in the long run.Further, Razzaque

and Basnett, (2014), argued that implementation of a comprehension

regional integration program as well as improved trade facilitation

measurers, effective transport infrastructure networks, ICT connectivity,

and co-operation in such areas as services trade, investment and energy

can unleash major avenues of trade for regional as well as global

markets.

Rather and Gupta (2014) observed that the annual value of informal

exports to Bangladesh from India in the year 2000 was estimated at

between $1 billion. Mostly popular consumer goods of international

quality are also informally traded at the border areas between the two

countries. It is quite obvious that informal trade between the two

countries does not take place because of trade policy distortions. The

informal traders usually engage in illegal trade to avoid the problems

they face while transacting legal channels due to rent seeking behaviour

and corruption. Therefore it is possible, that even in a zero duty regime

some informal trade would persist between countries across borders.

Rashid (2014) observed that although India has granted Bangladesh

duty-free access to all items except tobacco and liquor, there exist

reportedly several types of local duties which are around 15 per cent and

this discourages exports from Bangladesh to India. Thus non-tariff

measures are turned into non-tariff barriers while complying with

sanitary and phyto-sanitary measures and technical barriers to trade.

Gaurav, Bharti and Sinha (2015) suggest that in the current context of

ongoing trade negotiations between Bangladesh-Sri Lanka and Indo-

Bangladesh bilateral FTA, it is advisable to reduce sensitive list of

commodities. This is because it gives freshimpetus in terms of providing

new technology, expansion of the international markets, and new

opportunities for investment in both the countries.The local business



entrepreneurs in Bangladesh raise the fear of losing local industry and

agro-activities, but Bangladesh may also realize the intra-SAARC trade,

differently. Instead of trade competition, Bangladesh may look for intra-

industry/intra-business compliments as is evident in the case of India-Sri

Lanka Free Trade Agreement (ISFTA).

Apart from the purely economic factors, non-tariff barriers like delays at

major trade routes, customs ‘harassment’, visa-related issues, rigid

bureaucracy and infrastructural hurdles have ailed the trading relations

between India and Bangladesh[see Mukherjee (2015)].Therefore from

the literature review it can be concluded that there is a significant

amount of unofficial/informal trade between the two countries, high

transaction cost with absence of business friendly environment to

engage in official trade, thus, resulting in trade imbalance between

Bangladesh and India. As such, a research gap was identified and this

research is undertaken to see the prospects of bilateral trade through

official channel between two countries considering forecasting of trade

between India and Bangladesh. The study has been undertaken with the

following objectives:

To assess the current bilateral trade situation between India and

Bangladesh

To examine an optimal trade arrangement between India and

Bangladesh

To provide forecasting about future trade between two countries.

3. Methodology

Box-Jenkins methodology is applied in this paper. It requires a long time

series generally covering at least 50 observations; the researchers could

only use data for the last 25 years. There are two reasons for using this

short span of data - first, information were not systematically available

prior to 1990, and secondly, the researchers wanted to avoid structural

break. The latest major structural break is noticeable in 1990 financial

year (FY).The study ignores introduction of floating exchange rate in the

year 2003 (May) due to two reasons: i) in our model we do not consider

exchange rate as an independent variable; ii)exchange rate is yet to be

fully determined by the market based mechanism in Bangladesh. Time

period of the study is from 1990 to 2014.

http://link.springer.com/search?dc.title=Free+Trade&facet-content-type=ReferenceWorkEntry&sortOrder=relevance


3.1 Data Sources and Data Examination

Export and import data by commodities and by destinations are

published quarterly by the Bangladesh Bank in its Quarterly Export

Receipts and Quarterly Import Payments and by the Ministry of Finance

through its Bangladesh Economic Review and Export Promotion

Bureau. The researchers have collected all statistical data from the above

mentioned sources. Correlation with Exchange Rate and Ratio of Trade

between India and Total world (in %), Total (World) Trade with

Bangladesh and Total (India) Trade will be determined.

However, depending on the dependent variable, time period willchange.

Therefore Augmented Dickey-Fuller test of unit root will be used to do

further tests. Augmented Dickey-Fuller test will examines whether a

unit root is present in an autoregressive model and whether there is

autocorrelation in the residuals. Unit root tests will be used whether

trending data can be first differenced or regressed on deterministic

functions of time to cause to be the data stationary.

The researchers compared autoregressive moving average (ARMA) or

Autoregressive Integrated Moving Average Model (ARIMA) after

testing the significance of Augmented Dicky Fuller model’s intercept of

explanatory variables. ARMA model may be examined in the model as

the study is working with time series data with fewer terms overall than

either through an MA and / or an AR model by themselves.

Autoregressive integrated moving average (ARIMA) model indicates

stationary and non-stationary time series. The study will also see trend

stationary and difference stationary. Stationary refers to mean, variance,

and autocorrelation over the time period will be constant, to find best

model-fit.

The best model-fit will be represented based on Box-Jenkins cycle of

identification-estimation-diagnostic checking and forecasting. The study

has taken natural logarithm of the variables. Depending on the data

series (exports, imports and ration) the researchers proceeded to decide

about ARMA or ARIMA equations of the model to follow an

Autoregressive (AR) process and Moving Average (MA).AR model

consists of lagged terms of the time series itself. Moving Average (MA)

is a lagging indicator that cannot predict new trends, but can

authenticate trends which is recognized. As such, even after using AR

https://en.wikipedia.org/wiki/Unit_root

https://en.wikipedia.org/wiki/Autoregressive

http://r.search.yahoo.com/_ylt=A86.J3Wh_YJUISEAXc0PxQt.;_ylu=X3oDMTByNzhwY2hkBHNlYwNzcgRwb3MDMgRjb2xvA2dxMQR2dGlkAw--/RV=2/RE=1417899554/RO=10/RU=http%3a%2f%2fwww.forecastingsolutions.com%2farima.html/RK=0/RS=uWeGj2yjMYcYeeabnpo8gm.sYaA-

http://www.investopedia.com/terms/l/laggingindicator.asp



terms, there can be serious correlation at various lags. Thereby, in the

next step we will estimate the model with different AR and MA terms,

keeping in view the properties of residuals like independence,

homoscedasticity and normality. A general ARIMA model could be

written as:

Yt = α0 + α1Yt-1 + … … + αnYt-n + β1εt-1. .. Bmεt-m … (1)

If white noise error term problem arises then in this model

autoregressive filter, which is used in the long term and moving average

filter which will be used in the short term. Integration filter refers to

stochastic trend. Autoregressive models are important to assess

stationary time series. Moving average models are appropriate for

stationary time series. Depending on sign of autocorrelation, partial

autocorrelation function (PACF) and autocorrelation function (ACF)

will be determined. This study will also forecast errors and these errors

will depict the quality of the forecasting model. For forecasting

evaluation it is essential to first determine Root Mean Squared Error,

Mean Absolute Error, and Mean Absolute Percent Error, Theil

inequality coefficient which will be re-scaled by bias, variance and

covariance.

3.2 Preliminary Analysis

Based on preliminary analysis of the data and tables below the research

question for this study was identified. The research question for this

study is to investigate whether it is possible to have an optimal formal

trade arrangement between the two neighbouring countries India and

Bangladesh?Figure-1 below illustrates the export from India to

Bangladesh which is near zero.


Figure 1: Export from India to Bangladesh and World Exports

(Source: Based on Authors’ Data Analysis)

Figure-2, illustrates the import to India from Bangladesh which has been

rising since the beginning of this century, however still far less than that

of Rest of the World.

Figure 2: Import to India from Bangladesh and World


Balance of trade position of world countries and India’s with

Bangladesh shows that, it is low compared with the Rest of the World,

as shown in Figure-3.

0

10000

20000

30000

40000

50000

0 10 20 30

Mill

ion

USD

Year

World Export Million$

IndianExport Million$

0

10000

20000

30000

40000

50000

0 10 20 30

Mill

ion

USD

Year

India Import Million USD

World Import MillionUSD



Figure 3: Balance of Trade with India and World


Figure-4 illustrates the total trade of Bangladesh with the Rest of the

World countries is significantly on the rise, mainly in the ready-made

branded garment sector, where as trade with India is growing very

slowly, almost insignificant.

Figure 4: Total Trade of Bangladesh with India and World


0

10000

20000

30000

40000

50000

60000

70000

80000

90000

Mill

ion

USD

Total Trade with India

0

50000

100000

FY 1

99

0

FY 1

99

3

FY 1

99

6

FY 1

99

9

FY 2

00

2

FY 2

00

5

FY 2

00

8

FY 2

01

1

FY2

01

4

Total Trade with World

Mill

ion

USD


Ratio of trade between India and Total World, and its relationship with

the exchange rate is illustrated in Figure-5, which shows that

Bangladesh trade with India is growing slowly but far below the Total

Trade with Rest of the World.

Figure 5: Ratio of Trade between India and Total World and Exchange rate


Preliminary analysis suggest that if proper business friendly

environment is provided for formal trade development, reduced

transaction costs, along with transport infrastructure development, and

no tariff and non-tariff barriers to trade then both the countries can

engage in formal trade and mutually benefit from bilateral trade

agreement within the SARRC region.

4. Empirical Analysis

This section covers empirical results as discussed below:

4.1 Pearson correlation results

Pearson correlation result between Ratio of Trade between India

and Total World (in %) and exchange rate is found as 0.580

which is significant at 1% level. This indicates that there are

moderate positive correlation between ratio of trade between

India and total world (in %) and exchange rate which implies

0

20

40

60

80

100

FY 1

99

0FY

19

91

FY 1

99

2FY

19

93

FY 1

99

4FY

19

95

FY 1

99

6FY

19

97

FY 1

99

8FY

19

99

FY 2

00

0FY

20

01

FY 2

00

2FY

20

03

FY 2

00

4FY

20

05

FY 2

00

6FY

20

07

FY 2

00

8FY

20

09

FY 2

01

0FY

20

11

FY2

01

2FY

20

13

FY2

01

4

Ratio of Total trade of Bangladesh in between India and World(%)

I1 USD=?BDT



that one variables increase or decrease will have impact on other

variables.

Pearson correlation result between Total World Trade with

Bangladesh and Exchange rate is found as 0.904 which is

significant at 1% level of significance. This indicates that there is

strong positive correlation between Total World Trade with

Bangladesh and exchange rate, which implies that if one variable

increases or decreases will have impact on another variable.

Pearson correlation result between Total (India) Trade with

Bangladesh and Exchange rate is found as 0.910 which is

significant at 1% level of significance. This indicates that there is

strong positive correlation between Total (India) Trade with

Bangladesh and exchange rate which implies that one variables

increase or decrease will have impact on another variable.

4.2 Unitroot to determine stationarity

Augmented Dickey Fuller (ADF) test was applied to the data to check

for stationarity and the results are place in Table 1 below. From Table-1

it can be seen that at level, both ‘export to India’ (LEXPORTIND) and

‘imports from India’ (LIMPORTIND), both in log terms did not pass the

ADF test indicating that the variables are non-stationary. Nonetheless,

‘TRADE RATIO’ in logarithm passed the test at 5 percent level

showing that it is stationary at level.

Results of Stationarity test are reported in Table-1 below:

Table1:Result of the regression equation of Augmented Dicky

Variable Level(ADF with

intercept)

First Difference Type of Test

LEXPORTIND 4.348887 -3.172237** ARIMA

LIMPORTIND -1.486076 -5.409664 ***

ARIMA

LTRADE_RATIO -3.714755**

-6.599574*** ARMA

** 5% level***1% level Source: Based on Authors’ Data Analysis


At first difference, both the variables could reject the null hypothesis of

unit root at 5 percent and 1 percent level respectively. This means that

the variables are stationary at first difference. Thereby, we may

conclude that both exports to India and imports from India variables are

integrated of order 1, i.e. I (1). From Table-1, we can observe that

Augmented Dicky Fuller Model’s intercept of LEXPRTIND and

LIMPORTTIND is not significant. As such the study will test ARIMA

model .But for the LTRADE -RATIO‘s Augmented Dicky Fuller

Model’s intercept is significant at 5% level for which we shall test

ARMA model. In Table-1 the stationarity test indicates that first

difference of export is negative but significant at 5% level. In case of

import it is negative, but significant at 1% level. In case of trade ratio

intercept is significant at 5% level, while first difference is significant at

1% level although the sign is negative. Of the several alternatives, the

best estimate for exports and imports with India is reported in Table-2,

showing the results of ARIMA Model- Exports to India. Further, details

related to Table-2 are given in appendix as Table: 1.

Table2: Results of ARIMA Model: Exports to India Sample 1990-2014

LEXPORTIND

=

2312 +0.5148

AR(3)

+0.4849

AR(5)

+1.2820

MA(1)

+0.9032

MA(2)

t-statistics 0.001 2.63 2.15 11.24 10.74

Adjusted R2= 0.9353 F= 58.78

Source: Based on Authors’ Data Analysis

Table-3, reports the results of ARIMA Model- Imports from India. It can

be seen that all the t-test results were acceptable with high adjusted R2

and F-statistics. The forecasting power of the model is good. The

predictive power of the model indicates that actual and predicted values

have high level of close match. Detail of Table-3 is given in appendix as

Table: 2.



Table 3: Results of ARIMA Model: Imports from India - Sample 1990-

2014

LIMPORTI

ND

13.6 +0.9814AR

(1)

-0.9044MA

(1)

+0.5165MA

(2)

-0.5645MA

(4)

t-statistics 0.72 18.11 -5.03 3.36 -6.72

Adjusted R2= 0.9508 F =97.57


Table-4 below reports results of ARMA Model: Imports from India.

From the above Table-4, it can be observed that the equation adjusted

R2 is quite good and F statistics is acceptable, because as per ARMA

model it indicates that export to India is good as the equation-2 indicates

significant. But import and trade ratio indicates insignificant. Detail of

Table-4 is given in appendix as Table: 3.

Table 4: Results of ARMA Model: Trade Ratio - Sample 1990-2014

LTRADE_RATIO 2.13 +0.4509AR

(2)

-1.072MA

(1)

-1.1294MA

(2)

t-statistics 42.3 10.88 -2.45 -2.49

Adjusted R2= 0.7905 F =24.9


4.4 Forecast Evaluation

Table-5 evaluates the forecast results on exports and imports with India.

It reports the various measures of forecasting errors, viz., root mean

squared, mean absolute error, mean absolute percentage error, and Theil

coefficient.


Table 5: Forecast Evaluation

LEXPORTINDF LIMPORTINDF

Root Mean Squared Error 0.546811 0.204959

Mean Absolute Error 0.461016 0.171026

Mean Abs Percent Error 11.46903 2.505528

Theil Inequality Coefficient 0.057629 0.014459

Bias Proportion 0.710818 0.008751

Variance Proportion 0.045676 0.028413

Covariance proportion 0.243506 0.962836


It may be noted that our forecast is limited and it could not be extended

for out-of-sample, as number of the observation are made in only 14

years. From Table-5 it is discernible that the calculated value of RMSE

is almost of the same magnitude as that of MAE. They can be equal if

all errors are exactly the same. The smaller the reported errors, the better

will be the forecasting ability of the model. The Theil coefficient is also

less <1 one. It does not necessarily lead to acceptance of the model, but

does indicate that it performs better than other models.

From Table-5, it can be observed that imports from Bangladesh will be

relatively better if there is an increase in exports from Bangladesh to

India. Root Mean Squared Error is a measure of standard deviation

(SD). In case of exports SD is 0.546811 while for imports SD is

0.204959.Root Mean Squared Error for import is relatively better. Mean

absolute percentage error (MAPE), is determining prediction accuracy

of a forecasting method. As such MAPE for exports is 11.46903 while

MAPE for imports is 2.505528. Therefore, imports have a better

predictive accuracy. Theil inequality coefficient gives information about

accuracy of forecasting method for exports to India, which will be

0.057629, while in case of imports it will be 0.014459.It indicates that

imports from India are good for Bangladesh. Bias proportion which

indicates systematic error of 0.710818 is observed for exports to India,

while for imports is 0.008751.



Bias indicated systematic error and as for import value is close to zero

so it is relatively better than export. In case of variance proportion,

which pointed out the capability of the forecasts to replicate degree of

variability in the variable to be forecast, we observed that for exports it

is 0.045676 while for imports it is 0.028413.As the variance proportion

of export is large then the import, so it means that actual series has

fluctuated considerably, whereas the forecast has not. Covariance

proportion which measures unsystematic error indicates that for exports

it is 0.243506 while for imports it is 0.972836. Covariance of imports is

higher than exports, so highest proportion of inequality is relatively

good. An overall result from aforesaid forecasting evaluation differs

from our previous findings in Table: 2, as Import is relatively better than

export in Table:5.

5. Conclusions, Implications and Future Research Directions

The study concludes that in the forecasting model, import is good for

Bangladesh. Bangladesh being a small economy is less competitive than

India, for which formal trade in terms of imports can create value chain

between the two countries. This is due to the fact that while importing

from the India being a neighbouring country, transportation costs are

low. On the other hand as per the ARIMA model, best fit equation is

exports to India. Bangladesh should put emphasis on exporting products

to India as per their demand at a completive price, as well as maintain

high quality of the product. Suranovic (2010) comments on creating

competitive market will lead to efficiency in bilateral trade for

Bangladesh. However, until today, quality of goods exported from

Bangladesh to India are not up to the standard and therefore, is not well

equipped to compete in the Indian market. Further, Hassan (2001)

asserts that the reason for Bangladesh’s low intra-regional trade within

the SAARC region is due to not producing goods that are demanded by

the SAARC countries, low level of industrialisation and diversification

of the industry. Hassan, Mehanna and Bashar (2001), also concluded

from their study that within the framework of SAARC regional block ,

South Asian Preferential trade opportunities should be explored for

economic cooperation, potential for trade liberalisation policies and

concessions to reap the gains from mutually beneficial trade. However,

due to the absence of complementarity in production, resource base,

financial limitations, political tensions, there is low volume of intra-


regional trade which can be mutually beneficial to the SAARC or

BIMSTEC countries.

If Bangladesh can export readymade high quality branded clothing to

the western world, then the questions is why the same quality is not

maintained for exports to India. Excellence in maintaining first world

quality and benchmarking for the standard of the products at

international level by Bangladesh, with low cost, may create competitive

advantage to sustain favourable bilateral trade position with India, which

will ultimately narrow down balance of trade position. Further,

Wangwe’s (1993) view regarding the role of government policy in

promoting competitiveness in the economy of Bangladesh, for creating

competiveness in the production process will assist to achieve

favourable trade situation in the future. However, social and political

stability, zero-tolerance to terrorism, and building trust between two

nations especially at the political leadership level is extremely

significant to manage mutual benefits from bi-lateral free trade

agreement (FTA).Srinivasan (2002) suggestions should be considered

by the policy makers so that efficiency and effectiveness as well as

competitiveness in the bilateral trade through official channel should be

improved, so that informal trade can be reduced. Trading through

official channel should to be free from bureaucratic delays and rent

seeking behaviour.

Formal trade between Bangladesh and India is forecasted to be

economically and mutually beneficial creating a win-win situation, if the

least cost combination is used to produce products, efficient

infrastructure and transportation system across border is provided,

business friendly environment, avoiding bureaucratic delay, red tape and

corruption is reduced, which can further ensure free trade and fair

pricing along with “partnership and cooperation” [see Singh (2014)]

for economic development. Krugman’s (1980) view should be

considered by the policy makers of Bangladesh as economies of scale

can be attained in the production process. Withdrawal of protection in

terms of trade barriers (tariff and non-tariff barriers) are not sufficient

for Bangladesh unless and until Bangladesh can earn competitiveness in

bilateral trade. Non-tariff measures are turned into non-tariff barriers by

India, while complying with sanitary and phyto-sanitary measures and

technical barriers to trade, which is not effective for enhancing trade.



To reach sustainability in the long run through bilateral trade with India,

suppliers and manufacturer from Bangladesh ought to be concerned

about promoting development of world quality products, engage in

product differentiation, continuous innovation, attract new customers,

reliability of supply, investing and nurturing a sustainable business

enterprise through FTA.Indian foreign policy must also be business

friendly, towards gradually improving the bilateral trade relationship

which is supposed to be a combination of regional cooperation, as well

as meet the needs of the Indian market as also indicated by Ankit

(2015).Production intensity of Bangladesh in the industrial sector, both

export oriented industries and import substitution industries should be

raised. Thus, bilateral cooperation between the two countries will have

positive impact on creating economic efficiency and effectiveness. If

formal trade can be increased through efficient and effective services,

including infrastructural development, reducing transaction cost as

advocated by World Bank (2015), along with opportunities for cross-

border electricity trading [see Chattopadhyay and Fernando (2011)] will

not only result in rise in earnings from bilateral trade, but also meet the

increase in demand for power. However, India should withdraw tariff

and non-tariff barriers completely for SAPTA to work.

Given that Indian companies will have comparative advantage, mutually

agreed protective mechanisms can be put in place the transitional period,

for disadvantaged companies of Bangladesh. In the globalized regional

economy, India should give Bangladeshi exporters duty free access to

their market to a certain level, based on goods in demand. Innovative

and dynamic product lines, with assuring quality control of products

should be attained in Bangladesh in context of superior goods and

diversified exportable commodities should be produced by the country

to attain competitive advantage. Benchmarking of Bangladeshi product

with global standard should be obtained, as well as cost reduction and

establishing strategic alliances between both the country’s business

partners is required.

Bhagwati and Srinivasan (2002) advocated in their empirical findings

that globalization process is helpful for poor country. FTA between the

South Asian countries will be beneficial for which the platform of South

Asian Association of Regional Cooperation (SAARC) can be used, but

unfortunately it is not as successful as other regional trading agreements

such as ASEAN and EU. SAARC’s objective of economic cooperation


for promoting accelerated economic growth to improve the welfare and

quality of life of the people in South Asia has not been met. It is just a

cultural platform and not at all very effective in building trust and strong

business relationships in South Asia or having fruitful trade negotiations

under SAPTA. In order to build political, social, and economic ties to

mutually benefit the SAARC as well as BIMSTEC region, it is essential

to start first with building trust, disarm and focus on economic

development, outward looking economic policies to attract foreign

investment, improve institutions and promote growth through bilateral

preferential trade agreements. It seems that for India and Bangladesh

intra-regional trade is not as important, and they are more biased

towards trading with Rest of the World.

Current study is conducted based on secondary sources of trade data for

Bangladesh. For in-depth study, primary source can be used in future

research as to look at the informal trade between Bangladesh and its

bordering countries. Bilateral trade through official channel is beneficial

for both the countries and there is a wide scope to increase volume of

trade between the two countries through a collaborative approach,

private partnerships and capitalising on the huge potential benefit from

trade by providing business friendly environment, reduce informal trade

(smuggling) and transaction costs of doing business. It is thus essential

to build a sound foundation of trust, along with political wisdom to

eliminate all political conflicts and issues at all levels, in order to build

capacity, reap the benefits from economies of scale in trade flows,

attract direct investment from India in human and physical capital,

technology transfer, innovation and creation of economic efficiency ,for

which active cooperation, partnerships, and stability is required to

revitalize and rejuvenate the relationship between the two countries in

the 21st century.



References

Acharya, L. and Marwaha, A. (2012), Status Paper on India-

Bangladesh Economic Relations, FICCI Report, December, pp.2-85.

Ahmed, I. (2006), “Bangladesh-India Relations: The context of SAARC

and the emerging global scenario,” Proceedings of a Conference, Asian

Affairs, 28 (2), 46-62.

Ahmed, K. A. (2015), “View from Bangladesh: Ten take-aways from

the Modi visit,” South Asian Free Media Association, June 10, 2015.

Alam, K.A., Uddin, M.G.S. and Alam, M.M., Malakar, B. (2009),

“Trade Patterns of Bangladesh with India and China: An Empirical

Evidence of the PPP Theory,” Journal of Regional Economic Studies, 2,

26-34.

Ankit, R. (2015), “In the Twilight of Empire: Two Impressions of

Britain and India at the United Nations, 1945-1947,”Journal of South

Asian Studies, Routledge, 30, 1-15.

Asteriou, D. and Hill, S. G. (2007),Applied Econometrics-A modern

Approach using Eviews and Microfit, Revised edition, Palgrave,

Macmillan, UK, 230-247.

Bammi, Y. M. (2010),India Bangladesh Relations: The Way Ahead, Vij

Books India Pvt Ltd., 51-54.

Basu, S. and Datta, D. (2007),“India-Bangladesh Trade Relations:

Problem of Bilateral Deficit,” Indian Economic Review, New Series,

42(1), 111-129.

Basher, Md. Abdul (2013),Indo-Bangla Trade: Composition, Trends and

Way Forward, Commonwealth Secretariat, April, 4-21.

Bhardwaj, S. (2014), An Agenda for the New Government: Policy

Options for India in Bangladesh, Institute of Peace and Conflict Studies,

IPCS issue brief no. 251, June.

http://www.southasianmedia.net/blog/view-from-bangladesh-ten-take-aways-from-the-modi-visit/

http://www.southasianmedia.net/blog/view-from-bangladesh-ten-take-aways-from-the-modi-visit/

http://www.google.com.bd/url?sa=t&rct=j&q=&esrc=s&source=web&cd=35&cad=rja&uact=8&ved=0CDYQFjAEOB4&url=http%3A%2F%2Fwww.southasianmedia.net%2Fjournals%2F&ei=EHqdVbH-Ds-xuATomYGgDA&usg=AFQjCNENMQ7Xh1LzpJjaGESKeetmwM5Nqg&sig2=8jRX-HO6xYozQ8cewyzaeQ


Bhagwati, J. (1995), “Trade and Wages: Choosing among Alternative

Explanation,” Economic Policy Review, January, 42-47.

Bhagwati, J. and Srinivasan, T. N. (2002), “Trade and Poverty in the

Poor Countries,” The American Economic Review, 92(2), 180-183.

Bhuyan, A.R. (2006), “Bangladesh-India Trade Relations Prospects of a

Bilateral FTA,”Thoughts on Economics, 18(2),8-34.

Bown, C. P. (2014),Trade Policy Instruments over Time, Policy

Research Working Paper 6757, The World Bank, January, 2-23.

Chattopadhyay, D. and Fernando, P. N. (2011),“Cross-Border Power

Trading in South Asia: It’s Time to Raise The Game,”The Electricity

Journal, 10,1040-6190.

De, P. and Bhattacharyay, B. N. (2007),Prospects of India–Bangladesh

Economic Cooperation: Implications for South Asian Regional

Cooperation, ADB Institute Discussion Paper, No. 78, September, 1-36.

De, P., Raihan, S. and Kathuria, S. (2012),Unlocking Bangladesh-India

Trade Emerging Potential and the Way Forward, Policy Research

Working Paper, No 6155, The World Bank -South Asia Region

Economic Policy and Poverty Sector, August,1-33.

Dhaka Chamber of Commerce (2015),Viewed on 5th

July, 2015,

Available at:http://www.dhakachamber.com/home/saarc_trade: http:

//www.dhakachamber.com/home/saarc_trade.

Dutta, P. (2010),India-Bangladesh Relations issues, problems and recent

developments, Institute of Peace and Conflict Studies, IPCS Special

Report-97, September, 1-10.

Dutta, M. (1999), Economic regionalization in the Asia-Pacific:

challenges to economic cooperation, Edward Elgar Publishing, London.

Ethier, W. J. (2001), Regional Regionalism”, Regionalism and

Globalization-Theory and Practice, Lahiri, Sajal (Editor), Routledge,

London, 3-14.



Gaurav, K., Bharti N. andSinha, P. (2015),ISFTA: Lessons for

Bangladesh, Chatterjee, S. Singh, N.P., Goyal , D.P., andGupta, N.

(Editors). Managing in Recovering Markets, Springer India, 351-367,

viewed on 5th July, 2015, Available at: http://www.dhakachamber.com

/home/saarc_trade.

Government of India. (2013), Foreign Relationship with Bangladesh, ,

Viewed on 2nd

November, 2015,Available at: http://www.mea.gov.

in/Portal/ForeignRelation/Bangladesh_Brief.pdf

Hassan, M. Kabir. (2001), “Is SAARC a Viable Economic Block?

Evidence from Gravity Model,” Journal of Asian Economics, 12, 263-

290.

Hassan, M. Kabir, Mehanna, Rock-Antoine, & Bashar, S. Abul. (2001),

Regional Cooperation in Trade, Finance and Investment among SAARC

Countries: The Bangladesh perspective, Draft November 25, 2001,

Viewed on 2nd

November, 2015, Available at: http://www.syedbasher.

org/published/2002_ToE.pdf.

Hassan, M. Kabir. (2002),“Trade With India and Trade Policies of

Bangladesh,” Chapter 10 in, Towards Greater Sub-Regional Economic

Cooperation: Limitations, Obstacles and Benefits, Edited by Forrest E.

Cookson and A.K.M. Shamsul Alam, University Press Limited (UPL),

Dhaka, Bangladesh, 2002: 349-401.

Hossain, A. and Rashid, S. (1999), “The Political Economy of

Bangladesh's Large and Growing Trade Deficits with India,” The

Pakistan Development Review, 38(1), 25-68.

Islam, M.M. (2011), “Trade cooperation between Bangladesh and India

with Special Reference to the North-East India,” Dialogue, April-June,

12(4).

Islam, R. M. G., Khan, M. T. andIshak, A. (2013), “Bilateral and

International Trade of Bangladesh and India: Effect of Falling Exchange

Rate of Indian Rupee,” European Journal of Business and Management,

5(27), 33-39.

http://link.springer.com/search?facet-author=%22Kumar+Gaurav%22

http://link.springer.com/search?facet-author=%22Nalin+Bharti%22

http://link.springer.com/search?facet-author=%22Priyanka+Sinha%22

http://link.springer.com/search?facet-author=%22S.+Chatterjee%22

http://link.springer.com/search?facet-author=%22N.P.+Singh%22

http://link.springer.com/search?facet-author=%22D.P.+Goyal%22

http://link.springer.com/search?facet-author=%22Narain+Gupta%22

http://link.springer.com/book/10.1007/978-81-322-1979-8

http://www.dhakachamber.com/home/saarc_trade

http://www.dhakachamber.com/home/saarc_trade

http://www.syedbasher/

http://econpapers.repec.org/article/pidjournl/

http://econpapers.repec.org/article/pidjournl/


Jain, S.C. (1999), “Prospects for a South Asian free trade agreement:

problems and challenges,” International Business Review, 8, 399-419.

Khan, S. M. (1999), “South Asian Association for Regional

Cooperation,” Journal of Asian Economics, 10, 489-495.

Khan, S. M. and Khan, Z. S. (2003),“Asian economic integration: a

perspective on South Asia,” Journal of Asian Economics, 13, 767-785.

Krugman, P. (1980), “Scale Economies, Product Differentiation, and the

Pattern of Trade,” The American Economic Review, 70(5), 950-959.

Mukherjee, D. (2015), India-Bangladesh: new trade links, Viewed on 6

July, 2015, Available at: http://www.fii-news.com/india-bangladesh-

trade-links/.

Pohit, S.and Taneja, N. (2003), “India's Informal Trade with

Bangladesh: A Qualitative Assessment,” World Economy, Blackwell

Publishing Ltd., 26(8), 1187-1214.

Pursell, G. andSattar, Z. (2006), India-Bangladesh Bilateral Trade and

Potential Free Trade Agreement, Bangladesh Development Series Paper

No: 13, The World Bank, World Bank Office, Dhaka Bangladesh, 4-5.

Rahman, M., Ahamad, M. G. Islam, A K M. and Amin, N. M.A.

(2012),Agricultural Trade between Bangladesh and India-An Analysis

of Trends, Trading Patterns and Determinants, CPD-CMI Working

Paper 3, Centre for Policy Dialogue, September, 1-42.

Rahman,M., Khan,A.R., Nabi, A., andPaul, T.K. (2010),Bangladesh’s

Export Opportunities in the Indian Market: Addressing Barriers and

Strategies for Future, CPD Working Paper Series,Paper-90,July,pp.1-26.

Rashid, HarunUr (2014), Bangladesh-India trade talks, The Daily Star,

March 26, Working Paper No. 232.

Rather, Z.A. and Gupta, D. (2014), “India-Bangladesh Bilateral Trade:

Problems and Prospects,” International Affairs and Global Strategy, 22,

42-48.

http://www.researchgate.net/profile/Sanjib_Pohit

http://www.researchgate.net/profile/Sanjib_Pohit

http://www.researchgate.net/journal/0378-5920_World_Economy

http://cpd.org.bd/index.php/paper-90-bangladeshs-export-opportunities-in-the-indian-market-addressing-barriers-and-strategies-for-future/



http://cpd.org.bd/index.php/category/cpd-publications/working-papers/cpd-working-paper-series/



Razzaque, M. A. and Basnett, Y. (2014), “Regional Integration in South

Asia: Trends, Challenges and Prospects,” Commonwealth Secretariat, 1-

17.

Sharma, R. (2015), Narendra Modi's $2 billion loan to Bangladesh, The

Independent, 10 June.

Sikdar, C. R. andChakraborty, M. (2006), “Bilateral Trade between

India and Bangladesh: A General Equilibrium Approach,”Economic

Systems Research, 18(3), 257-279.

Singh, N. (2014), “India and Development Partnership: Special

reference to Bangladesh in 21st century,”Elsevier -Procedia, Social and

Behavioral Science, 157,137-142.

Srinivasan, T.N. andVani, A. (2009),India in the Global and Regional

Trade: Determinants of Aggregate and Bilateral Trade Flows and Firms

Decision to Export, Indian Council for Research on International

Economic Relations, February Report, 1-22.

Srinivasan, T. N. (2002), Trade, Finance, and Investment in South

Asia,Berghahn Books, 164-165.

Suranovic, S. (2010), A Moderate Compromise: Economic Policy

Choice in an Era of Globalization, Palgrave McMillan, UK.

Sutherland, P. (2012), The Bilateral Free Trade, Project Syndicate.

Available at: http://www.project-syndicate.org/commentary/the-doha-

round-and-the-decline-of-the-world-trade-organization-by-peter-

sutherland.

The Indian Express (2015), NE trade bodies happy over PM Modi’s

Bangladesh trade pacts, June, 9, 2015.

Wangwe, S. (1993),New Trade Theories and Developing Countries:

Policy and Technological Implications, UNU/INTECH Working Paper

No. 7, June, 1-24.

https://books.google.com.bd/url?id=JK49YUikMhgC&pg=PA164&q=http://www.berghahnbooks.com&clientid=ca-print-berghahn&linkid=1&usg=AFQjCNFimEoEEexI_FsQEUMt_qtqV97z3Q&source=gbs_pub_info_r

http://www.project-syndicate.org/commentary/the-doha-round-and-the-decline-of-the-world-trade-organization-by-peter-sutherland.




World Bank (2006), India-Bangladesh Bilateral Trade and Potential Free

Trade Agreement, Bangladesh Development Series, Paper No: 13, The

World Bank Office, Dhaka, December, 1-87.

World Bank (2014), Economic Indicators, Viewed on 15th March, 2015,

Available at: http://data.worldbank.org/indicator/IC.BUS.EASE.XQ)

World Bank (2015), Informal and illegal trade: dimensions, trends,

composition, and the role of domestic indirect taxes, Chapter 8, 57-63,

viewed on 15th March, 2015, Available at: http://siteresources. worldbank

.org/SOUTHASIAEXT/Resources/223546-1168296540386/ch8.pdf.



APPENDIX

Detail Regression Results

Below Table 1 reports the result of regression equation considering

dependent variable log of export to India where we did ARIMA model:

Table:1 Result of 1st Regression Equation on Export to India

Dependent Variable: LEXPORTIND

Method: Least Squares

Sample (adjusted): 1995 2014

Included observations: 20 after adjustments

Convergence achieved after 219 iterations

MA Backcast: 1993 1994

Variable Coefficient Std. Error t-Statistic Prob.

C 2312.358 1647090. 0.001404 0.9989

AR(3) 0.514779 0.196091 2.625210 0.0222

AR(5) 0.484941 0.225852 2.147164 0.0529

MA(1) 1.281965 0.114039 11.24143 0.0000

MA(2) 0.903175 0.084096 10.73978 0.0000

R-squared 0.951447 Mean dependent var 4.434991

Adjusted R-squared 0.935263 S.D. dependent var 0.941622

S.E. of regression 0.239582 Akaike info criterion 0.220086

Sum squared resid 0.688794 Schwarz criterion 0.465149

Log likelihood 3.129265 Hannan-Quinn criter. 0.244446

F-statistic 58.78808 Durbin-Watson stat 1.239795

Prob(F-statistic) 0.000000

Inverted AR Roots 1.00 .16+.74i .16-.74i -.66+.64i

-.66-.64i

Inverted MA Roots -.64-.70i -.64+.70i


From the above regression equation it can be noted that log of export

from India is dependent variable. Sample period of the equation is for

the period of 1995 to 2014. R square and adjusted R square indicates


that the equation fits well. Autoregressive -3 is significant at 5% level of

significance. Autoregressive -5 is significant at 10% level of

significance. Moving average (1) and moving average (2) indicates

significance at 1% level of significance. Durbin Watson statistics

indicates autocorrelation; F statistics is significant at 1% level of

significance.

Below given Table - 2 reports the results of another regression equation

considering dependent variable as log of import to India where we did

ARIMA model:

Table:2- Result of Regression Equation on Import to India

Dependent Variable: LIMPORTIND





MA Backcast: 1987 1990


C 13.63823 18.75394 0.727219 0.4776

AR (1) 0.981427 0.054200 18.10737 0.0000

MA (1) 0.904408 0.179821 -5.029495 0.0001

MA (2) 0.516486 0.153668 3.361054 0.0040

MA (4) 0.564490 0.084048 -6.716257 0.0000



S.E. of regression 0.192833 Akaike info criter -0.249727

Sum squared resid 0.594953 Schwarz criterion -0.001032

Log likelihood 7.622137 Hannan-Quinn criter -0.195754


Prob (F-statistic) 0.000000

Inverted AR Roots .98 .16+.74i .16-.74i -.66+.64i

Inverted MA Roots .98 .27+.9 .27-.92i -.62




From the above regression equation it can be noted that log of imports

from India is dependent variable. Sample period of the equation is for

the period of 1991 to 2014. R square and adjusted R square indicates

that the equation fits well. Autoregressive -1 is significant at 1% level of

significance. Moving average (1), moving average (2), moving average

(4) indicates significance at 1% level of significance. Durbin Watson

statistics indicates no autocorrelation; F statistics is significant at 1%

level of significance. Table 3 below reports the result of regression

equation considering dependent variable as log of Trade ratio where we

did ARMA model:

Table:3 Result of the regression equation on Trade Ratio

Dependent Variable: LTRADE_RATIO





MA Backcast: OFF (Roots of MA process too large)


C 2.131033 0.050321 42.34847 0.0000

AR (2) 0.450850 0.041450 10.87699 0.0000

MA (1) -1.072187 0.437441 -2.451047 0.0261

MA (2) -1.129435 0.453246 -2.491880 0.0241



S.E. of regression 0.086372 Akaike info criter -1.883441

Sum squared resid 0.119363 Schwarz criterion -1.684295

Log likelihood 22.83441 Hannan-Quinn criter -1.844566


Prob (F-statistic) 0.000003

Inverted AR Roots .67 -.67

Inverted MA Roots 1.73 -.65

Estimated MA process is noninvertible



From the Table: 3 regression equations it can be noted that log of trade

rate is dependent variable. Sample period of the equation is for the

period of 1992 to 2014. R square and adjusted R square indicates that

the equation fits well. Autoregressive -2 is significant at 1% level of

significance. Moving average (1), moving average (2), moving average

(4) indicates significance at 5% level of significance. Durbin Watson

statistics indicates that autocorrelation is significant at 1% level of

significance.

Documents

Jour nal of Economic C ooperation C · PDF fileJour nal of Economic C ooperation ... The Real Effect of Government Debt: ... Bangladeshi producers and suppliers ought to