Estimation of the Export Demand Function using Bilateral Trade Data: The Case of Bangladesh

Estimation of the Export Demand Function using Bilateral Trade Data: The Case of Bangladesh MOHAMMED NUR ALBERT WIJEWEERA BRIAN DOLLERY Trade liberali...
Author: Rosalind Knight
0 downloads 0 Views 143KB Size
Estimation of the Export Demand Function using Bilateral Trade Data: The Case of Bangladesh MOHAMMED NUR ALBERT WIJEWEERA BRIAN DOLLERY Trade liberalization policy represents an important policy instrument for developing countries. However, existing empirical literature on the impact of trade liberalization is ambiguous and this uncertainty also extends to Bangladesh economy. Various studies have sought to estimate econometrically the export demand function for Bangladesh as a means of clarifying the effects of trade liberalization in that country. Unfortunately, no recent studies have examined the liberalization effect on disaggregated exports in Bangladesh. In order to remedy this neglect, this article examines the likely impacts of trade liberalization policies on the disaggregated export function in Bangladesh for the period 1973–2004. The main purpose is to examine whether bilateral export elasticities are significantly different between major trading partners. If this is indeed the case, then different policies should be implemented rather than a single trade policy to enhance exports. A secondary objective is to establish whether trade liberalization has had any impact on the export sector in Bangladesh. (JEL : C22, F13, F32, F43) Keywords: Bangladesh, Export Demand Functions, Trade Liberalization.

1. Introduction A substantial theoretical and empirical literature exists on the impact of trade liberalization on external trade in developing countries (Dornbusch 1992). This Acknowledgement: The authors would like to thank two anonymous referees for helpful comments on an earlier draft of the article. Mohammed Nur, Albert Wijeweera and Brian Dollery are with School of Economics, University of New England, Armidale, NSW, 2351 Australia. Albert Wijeweera is the corresponding author. Email: [email protected]

SOUTH ASIA ECONOMIC JOURNAL, 8:2 (2007): 249–264 SAGE Publications Los Angeles n London n New Delhi DOI: 10.1177/139156140700800204

n

Singapore

250 / MUHAMMED NUR, ALBERT WIJEWEERA

AND

BRIAN DOLLERY

literature has formed the basis for much of the policy advice offered to poor nations, with trade liberalization argued to expand both domestic and international markets by enhancing efficiency, innovation and technological adaptation. In essence, trade liberalization induces a substitution effect in consumption and production by changing relative prices, thereby reallocating resources towards economically efficient uses and raising national income in the process. Several studies have examined the impact of trade liberalization on exports and imports in different countries, but the findings are far from unanimous. Some papers show that various countries’ export performance improved after trade liberalization (Joshi and Little 1996; Thomas et al. 1991; Weiss 1992). However, other researchers have established little impact of trade liberalization on export growth (see, for instance, Agosin 1991; Clarke and Kirkpatrick 1992). In the specific case of Bangladesh, Dutta and Ahmed (2000), and Hossain and Alauddin (2005) examined the impact of trade liberalization on exports and imports. Hussain (2004), Choudhury (2001), Roy (1991), Shilpi (1990), Nguyen and Bhuiyan (1977) have estimated the export and import demand functions for Bangladesh. Despite this useful work, no recent studies have examined the liberalization effect of disaggregated exports in Bangladesh. In order to remedy this neglect, this article will examine the likely impacts of trade liberalization policies on the disaggregated export function in Bangladesh for the period 1973–2004. The article has two main aims. Our main purpose is to examine whether bilateral export elasticities are significantly different between trading partners. This question is important since trade policies that are based on aggregate export elasticities can be misleading if bilateral trade elasticities are different from those of an aggregate model. For instance, if the income elasticity derived from aggregate data is inelastic, it implies that whenever world income increases by 1 per cent, Bangladesh’s export volume will increase by less than 1 per cent. However, if we find that bilateral export income elasticity related to a specific national trading partner is elastic, then if income in this country grows by 1 per cent, export demand from Bangladesh to that particular country will increase by more than 1 per cent. Accordingly, trade policy can be tailored to meet the specific circumstances of particular trading partner nations rather than a single trade policy to enhance the aggregate export sector. A secondary objective is to establish whether trade liberalization has had any impact on the export sector in Bangladesh. To the best of our knowledge, and in stark contrast to our focus on the disaggregated export function, all the existing studies on Bangladesh rely exclusively on aggregated data to estimate the export demand function. The article itself is divided into seven main sections. Section 2 provides a synoptic description of trade policy in Bangladesh since independence in 1971

South Asia Economic Journal 8:2 (2007): 249–264

Estimation of Export Demand Function using Bilateral Trade Data / 251

and background information on export performance over this period. Section 3 sets out a brief review of the more important theoretical considerations underlying our study. Section 4 describes the salient international empirical literature as well as work done directly on export demand functions in Bangladesh. Section 5 deals with the model specification. Section 6 considers data requirements, variables and key descriptive statistics. The results are discussed in section 7. The paper ends with some brief concluding remarks in section 8.

2. Institutional Background Bangladesh has adopted three major trade policies since independence in 1971. In the first phase, import-substitution strategies through quantitative restrictions on imports, import licensing and strict exchange control measures were undertaken over the period of 1972–78. In the second phase that ran from 1979 to 1990, various programmes, such as fiscal, financial, trade and industrial policy reforms, public resource management and privatization, a structural adjustment policy and export-led industrialization, were adopted. The third phase (1990–2003) targeted domestic and foreign private investment, the development of a strong quality control mechanism for imported goods and easier raw material imports. Import licensing restrictions on intermediate and capital goods were mostly eliminated, tariffs were significantly reduced and full convertibility of foreign exchange earnings was established for current account transactions. To comply with the Uruguay Round Accord and agreements of the World Trade Organization (WTO), the programme of tariff reductions is still continuing in Bangladesh. Import duties of 150 per cent in 1992–93 were gradually reduced to 32.5 per cent by 2002–03 (Bangladesh Economic Review 2003). Although export incentives have been in place in Bangladesh since the late seventies, they have only been strengthened since 1986. At present, the major export policies consist of: removing restrictions on imports for exporters; reducing duty rates across the board; duty free importation of capital machinery and duty drawbacks on inputs; extended coverage of ‘back-to-back’ letters of credit and bonded warehouse facilities for exporters; income tax rebates on export earnings; import facilities for export commodities; additional infrastructure for export processing zones; and an export performance benefit scheme for exporters. These policies have induced changes in export performance over the sample period 1973–2004, and the value of exports in terms of the domestic currency has gradually increased. For instance, annual export volumes grew from US$ 489 million in 1973 to around US$ 6,800 million in 2003 (Bangladesh Economic Review 2004). Figure 1 shows a comparison between World and Bangladesh

South Asia Economic Journal 8:2 (2007): 249–264

252 / MUHAMMED NUR, ALBERT WIJEWEERA

AND

BRIAN DOLLERY

Export Growth Rates over the period 1974–2004. It indicates that the export growth has been positive since 1988, except for 2002 and 2003. It also shows that the average growth in exports for Bangladesh is higher than the average global export growth for the sample period, except during the period 1972–82. FIGURE 1 Comparison between World and Banagladesh Export Growth Rates (1974–2004) 60

Percentage

40

20

0 1973

–20

1978

1983

1988

1993

1998

2003

Year World Export Growth Rate Bangladesh Export Growth Rate

Source: Based on International Financial Statistical Yearbook (various issues).

The main purpose of this article is to estimate the export demand function for the top five exporting countries and thereby to compare price and income elasticities between these countries. Accordingly, Bangladesh’s export destinations are important. Table 1 shows the destination of exports from Bangladesh from 1972–73 to 2003–2004. The major portion of export earnings for Bangladesh came from the United States (US). In the fiscal year 2003–04, 26 per cent of total exports were to the US, a fall from 35 per cent in the fiscal year 1992–93. The US, Germany, the United Kingdom (UK) and France collectively accounted for over half of Bangladesh’s total exports. Table 1 also reveals that the relative importance of some countries, notably the US and Japan, have declined in recent years.

3. Theoretical Considerations on Trade Liberalization Trade liberalization has been extensively studied, both in developed and developing countries. Little et al. (1970) and Balassa (1971) were the pioneers of modern South Asia Economic Journal 8:2 (2007): 249–264

Estimation of Export Demand Function using Bilateral Trade Data / 253 TABLE 1 Bangladesh’s Exports to Major Destinations 1972–73 to 2003–04 (US$ million) Country United States United Kingdom Germany France Belgium Italy Netherlands Canada Japan Rest of world Total

1972–73 71.39 (20.49) 26.55 (7.62) 10.54 (3.03) 8.74 (2.51) 22.99 (6.60) 14.36 (4.12) 8.14 (2.34) 7.03 (2.01) 6.70 (1.92) 171.98 (49.36) 348.42 (100)

1982–83

1992–93

2003–04

78.86 (11.49) 30.96 (4.51) 13.75 (2.00) 7.26 (1.06) 30.29 (4.41) 32.14 (4.68) 12.79 (1.86) 6.68 (0.97) 45.01 (6.56) 428.86 (62.46) 686.60 (100)

822.51 (34.52) 183.42 (7.70) 216.21 (9.07) 127.36 (5.35) 83.14 (3.49) 83.14 (3.49) 85.80 (3.60) 44.38 (1.86) 53.31 (2.24) 629.36 (26.41) 2,382.89 (100)

1,966.58 (25.87) 898.21 (11.81) 1298.54 (17.08) 552.96 (7.27) 326.95 (4.30) 315.93 (4.16) 290.44 (3.82) 284.33 (3.74) 118.16 (1.55) 1550.89 (20.40) 7,602.99 (100)

Source: Bangladesh Economic Review (2004).

multi-country investigations on trade orientation and economic performance in developing countries. By using the effective rate of protection, Cordon (1966) tried to find comparative evidence on how the structure of protection on intermediate and final goods affected relative profitability of sectoral value-added products. A number of studies have shown that the effective rate of protection is negative in many developing countries for the agricultural sector. Little et al. (1970) and Balassa (1971) demonstrated that the degree of protection is significantly higher for the manufacturing sector. They argued that the most important consequences of a protectionist policy are the worsening of income distribution, a reduction in savings, an increase in the rate of unemployment and a very low rate of capacity utilization. Bhagwati (1988) examined trade orientation, which is measured by the degree by which the protective (and incentive) structure in a country is biased against exports. The degree of bias was defined as the ratio of the effective exchange rate of imports index to the effective exchange rate of exports index. If the degree of bias is smaller than one, this means that the country could have followed an export promotion strategy. If the degree of bias takes the value of unity, this indicates a neutral trade regime. Where the degree of bias is greater than one, this implies South Asia Economic Journal 8:2 (2007): 249–264

254 / MUHAMMED NUR, ALBERT WIJEWEERA

AND

BRIAN DOLLERY

that the country in question has followed an import substitution policy. According to Bhagwati (1988), trade liberalization is defined as any policy that reduces the degree of anti-export bias.

4. Empirical Considerations As we indicated earlier in the article, a substantial empirical literature exists on the impact of trade liberalization, although this work has yet to generate consensus. Several studies are significant in the context of the empirical analysis of export demand functions for Bangladesh. In the first place, Marquez and McNeilly (1988) examined income and price elasticities for exports for non-OPEC (Organization of Petroleum Exporting Countries) developing countries using quarterly data for 1973–84. This study was based on the two-stage least square estimation technique. Relative import prices, real income and lagged endogenous variables were the main explanatory variables. They found that income elasticity for exports of non-OPEC developing countries to industrial countries varied from 1.4 to 1.9. Furthermore, Senhadji and Montenegro (1999) studied export demand functions using cross-country analysis. They found that the average long-run price and income elasticities are 1 and 1.5 respectively. They also found that the African income elasticity of exports was lower than its Asian counterpart. Paulino and Thirlwall (2004) examined the impact of trade liberalization on exports, imports and the balance of payments for 22 developing countries. They tested the impact and significance of trade liberalization using different estimation techniques, such as fixed effects and the generalized method of moments (GMM) for panel data analysis. They found that the impact on import growth was greater than on export growth after trade liberalization. They also found income elasticities of import demand for imports and exports increased by equal amounts, but the price elasticity of demand for imports was greater than for exports. They concluded that the balance of trade deteriorated after trade liberalization. Lopez (2005) examined the effect of trade liberalization on exports, imports, the balance of payments and growth for the Mexican economy, covering the period 1973–2000. An autoregressive distributed lag (ARDL) estimation technique was used to identify the long-run relationship between the dependent and independent variables. He found that the price and income elasticities of demand and error correction coefficient were significant for both the export and import demand functions. He also found that trade liberalization positively affected export and import growth. However, growth in imports was higher than growth in exports. In addition to this body of work, several studies have been undertaken on the Bangladesh economy. For example, Roy (1991) estimated an export demand function for Bangladesh over the sample period 1976–77 to 1986–87. He found that there was a strong effect of world demand, the exchange rate, an efficiency South Asia Economic Journal 8:2 (2007): 249–264

Estimation of Export Demand Function using Bilateral Trade Data / 255

index, the effective rate of assistance and other non-price factors on export performance. He also found the exchange rate or trade-weighted nominal export exchange rate or effective rate of assistance coefficients positive and statistically significant. In addition, his article indicated that devaluation had some positive effects on the export function. Ahmed (2000) examined the response of Bangladesh’s export demand function over the period 1974–95. Two types of modelling approaches—cointegration and error correction—were applied in this study. Only one cointegration relationship among three variables (that is, real export, relative price and real exchange rate) was established while the coefficient of error correction showed a high speed of convergence to equilibrium. The results indicated that the liberalization policies were not fully effective. Using updated data, Choudhury (2001) estimated an export demand function of Bangladesh for the period 1972–73 to 1997–98. His results suggested that the demand for exports was highly income elastic and significant. Hossain and Alauddin (2005) examined the impact of trade liberalization on exports in the Bangladesh economy based on the ARDL model cointegration technique and found that there existed a short-run mechanism leading to the long-run equilibrium.

5. Model Specification The export performance of a given country depends inter alia on competitiveness, which is measured as the price of the country’s exports relative to the foreign price of related goods and the level of international income. According to standard economic theory, the export demand function relates the quantity of exports demanded to foreign income and the real exchange rate. By assuming price and income elasticities of demand are constant, the export demand function for Bangladesh at time t may be written as: α

1  Pf E  α EXPOit = A   GDPit 2 P  d  it or

EXPOit = A ( REXC )it 1 GDPitα 2 α

(1)

(2)

where i = 1,2…….5, denoting the five major exporting countries of Bangladesh, which are Belgium, France, Germany, the UK and the US. EXPOit is the level of real export of Bangladesh with respect to ith countries at time t; A is a  Pf  constant;  P  is foreign and domestic (Bangladesh) prices for the ith countries  d it South Asia Economic Journal 8:2 (2007): 249–264

256 / MUHAMMED NUR, ALBERT WIJEWEERA

AND

BRIAN DOLLERY

at time t; REXC is the bilateral real exchange rate between Bangladesh and selected countries at time t. GDPit is the real gross domestic product of ith countries at time t. Orthodox economic theory predicts that the price elasticity of demand for export is positive (a1 > 0) and the income elasticity of demand for export is positive (a2 > 0). This approach raises the important question as to whether the assumption of a standard downward-sloping demand curve is applicable in the case of the Bangladesh market, where it could be argued, in general, it is the buyer that sets the price. Nevertheless, the downward-sloped demand curve is as yet the most practical way developed to examine the issue econometrically. Almost all published works in this area have used a downward-sloping demand curve or a variant of that approach, regardless of whether it is investigating a large country or a small country (see, for instance, the seminal paper by Houthakker and Magee 1969; Bleaney, 1999; Paulino and Thirlwall 2004, among many other studies). We thus follow this approach despite its shortcomings. Taking the logs and differentiating with respect to time, the export function can be written as: EXPOit = α 0 + α1 REXCit + α 2GDPit

(3)

Following the literature, we have used double logarithmic equations throughout the article. This gives superior fit and is easier to interpret. However, the specification assumes a constant elasticity throughout the time period, which can be questionable. Further, we have used Engle and Granger cointegration method for estimating the model. This method assumes only one long-term linear relationship or cointegration relationship between the variables. Following Paulino and Thirlwall (2004), we introduce a dummy variable to capture the effects of the trade liberalization policy. The dummy variable takes the value of zero prior to the liberalization period and one after trade liberalization. Although it was in 1986 that the Bangladesh government first adopted a liberalization policy, the effectiveness of liberalization seems to have only begun in 1992. A dummy variable takes the values 0 for 1973–91 and 1 for 1992–2004. Accordingly, the extended export demand function can be expressed as equation (4). EXPOit = α 0 + α1 REXCit + α 2 GDPit + α 3 D + ε it

(4)

This procedure means that only one dummy variable has been used which implies that it is not possible to identify which particular policy measure was more effective. During the time period under consideration, a number of policy changes have been introduced by authorities. Even trade liberalization was introduced in several stages. However, we have only introduced one dummy variable for testing whether liberalization has any impact on the export performance in South Asia Economic Journal 8:2 (2007): 249–264

Estimation of Export Demand Function using Bilateral Trade Data / 257

Bangladesh. The consequence of this is that the impact of other omitted policy variables is shown via the single dummy variable. Omission of the dummies is due to three main reasons. First, with a small data set, a more parsimonious model is preferred. Second, our main purpose is to see whether export price and income elasticities are different between the top five purchasing countries. The impact of trade liberalization is thus, only a by-product of this procedure. Third, but most important, our initial estimations show that incorporating more dummies into the model produces counterintuitive results indicating that particular specifications are inappropriate.

6. Data Requirements, Variables and Key Descriptive Statistics This article aims to examine whether bilateral export elasticities are significantly different between top trading partners and to see whether trade liberalization has had any impact on the export sector in Bangladesh. For these reasons, annual data for selected variables has been gathered from the annual reports and relevant publications of various government agencies of Bangladesh and other salient countries. The time period of the analysis is limited to 32 years; from 1973 to 2004. In general, Bangladesh’s major export countries are the UK, the US, Germany, China and Thailand. Although China and Thailand are among the top five exporting countries to Bangladesh, these two countries could not be included in this study due to lack of data for the whole period. Fortunately, the omission of Thailand and China does not create any statistical problems because we estimate the export demand function for each country separately and not as a panel or using methods such as seemingly unrelated regression. An alternative approach would be to include China and Thailand by running the same model for a shorter period of time and explaining any differences in the outcome. Unfortunately, we are not able to do this with the current data set that we possess. However, as explained before, the results will not affect the coefficients of already estimated equations because they are estimated as separate equations. The alternative approach would thus only provide income and price elasticities for China and Thailand. Accordingly, we have included the next two major exporting countries— Belgium and France—in the analysis. These countries accounted for about 66.33 per cent of total Bangladesh’s exports in 2004. Variables used in this study include real Gross Domestic Product (GDP), the export price index, nominal exchange rate, consumer price index, import duties ratio, and the values of exports of Bangladesh from the selected countries. The GDP and GDP deflator data are taken from the International Financial Statistics series published by the International Monetary Fund (IMF). Original GDP data were in the own currencies (billion) of various countries. These data South Asia Economic Journal 8:2 (2007): 249–264

258 / MUHAMMED NUR, ALBERT WIJEWEERA

AND

BRIAN DOLLERY

were then converted from domestic currency to their US dollar equivalent. The value of exports for the top five exporting countries has been taken from the Bangladesh Economic Review (2004), published by the Ministry of Finance, Bangladesh. The data for exchange rates between Taka and foreign currencies were taken from the Economic Trends published by Bangladesh Bank. However, these data were for only the period 1993–2004. To get a complete series for 1972–2004, the exchange rate between Taka and foreign currencies used in this study employed a triangular arbitrage procedure. Triangular arbitrage involves three currencies, such as the Taka, the UK pound and the US dollar. All three currencies are traded in the US, Britain and Bangladesh. Suppose that the following prices exist in all three financial centres: US$ 1 is equal to 59.513 Taka and US$ 1 is equal to 0.5459 pounds. In this case, according to triangular arbitrage, £1 will be equal to 109.01814 Taka (59.513/0.5459). By following this triangular arbitrage method, a bilateral exchange rate between Bangladesh and other selected countries has been calculated for this article. Since Belgium, France and Germany began using the Euro during the latter part of the sample period, we had to adopt the methodology introduced by IMF to convert the currencies. IMF Statistics convert the Euro multiplying by 40.3399, 6.55955 and 1.95583 for Belgium, France and Germany respectively. This assists our estimation because exchange rates for all the trading partners are expressed in terms of US$. The consumer price index (CPI) data (base year 2000 = 100) are taken from International Financial Statistics. In case of Bangladesh and Germany, there were two base periods. A spliced index is used to convert the values into the same base year. We calculate the real exchange rate (RER) using the following standard formula.

 CPI i RERi, Bangladesh = eTaka / ithcountries ∗   CPI Bangladesh 

   

(5)

7. Results and Discussion We primarily use popular Engle and Granger residual based test (1987) to examine the cointegration relationships among the variables in the export demand function. Prior to the application of cointegration techniques, it is essential to test for the unit roots. We use Augment Dickey-Fuller (ADF) and PhillipsPerron (PP) tests for that purpose. The results are shown in Table 2. The results clearly suggest that all three variables contain a unit root in levels, but stationary in first differences. In other words, the variables used in this study are integrated of order one or I(1). South Asia Economic Journal 8:2 (2007): 249–264

Estimation of Export Demand Function using Bilateral Trade Data / 259 TABLE 2 Unit Root Test for Stationary and Non-Stationary for Major Exporting Countries Level Countries Belgium

Variables

LBEXO LBGDP LBREXC France LFEXO LFGDP LFREXC Germany LGEXO LGGDP LGREXC UK LUKEXO LUKGDP LUKREXC US LUSEXO LUSGDP LUSREXC

First Difference

ADF

PP

Conclusion

–1.33 –1.630 –1.824 –0.328 –1.66 –1.77 –0.613 –1.15 1.02∗ –0.467 0.96∗ –1.42 –0.428 0.3936 –2.25

–1.18 –0.92 -1.86 –0.45 –0.84 –1.64 –0.88 –1.13 1.32∗ –0.52 2.56∗ –1.41 –0.40 1.02 –2.63

NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS

ADF –6.71 –3.36 –5.08 –5.51 –3.72 –6.42 –7.73 –3.86 –4.42 ∗ –4.96 –3.29 ∗ –6.54 –5.80 –4.39 –6.44

PPC

Conclusion

–11.9 –3.15 –5.11 –5.49 –3.66 –6.26 –8.26 –3.76 –4.43 ∗ –4.96 –2.99 ∗ –5.53 –5.82 –6.23 –5.91

S S S S S S S S S S S S S S S

Note: All variables are in logarithmic form, EXO = Export, RGDP = Real Gross Domestic Product, REXC = Bilateral Real Exchange Rate, NS = Non-Stationary, S = Stationary, Critical Values at 5% level = –2.96, Majority of the tests used constant, no trend specification. ∗Indicates no trend, no constant specification. DF critical value at 5% level of significance is –1.95.

7.1 Testing for Residuals Since all the variables are I(1), the Engle and Granger (1987) method can be used to examine the cointegration relationships among the variables. This method consists of two major steps. First, the best linear model is estimated and residuals of the models are saved. These residuals are then tested for stationarity. If the residuals represent white noise, it can be said that all the variables in the model are cointegrated or there exits a long-run relationship in the model. The best linear long-run regression results and related residual tests are shown in Table 3 and Table 4, respectively. Table 3 indicates that residuals from the linear regressions are white noise. Hence, there is a linear combination of EXPO, REXC and GDP that is stationary. In other words, all the variables are cointegrated in these five disaggregated export functions for Belgium, France, Germany, United Kingdom and United States. Accordingly, there exists a long-run equilibrium relationship for all five equations. This means that we can meaningfully interpret the results that are shown in Table 4 and conduct hypothesis testing on the coefficients.

South Asia Economic Journal 8:2 (2007): 249–264

260 / MUHAMMED NUR, ALBERT WIJEWEERA

AND

BRIAN DOLLERY

TABLE 3 Engle-Granger and Phillips-Perron’s Residual-based Test (Major Exporting Countries of Bangladesh) Country Belgium France Germany United Kingdom United States

ADF

Critical Values

PP

Critical Values

Conclusion

–4.00 –2.87 –3.82 –2.81 –4.35

–2.76 –2.76 –2.76 –2.76 –2.76

–3.91 –2.91 –3.82 –2.86 –4.40

–2.76 –2.76 –2.76 –2.76 –2.76

Stationary Stationary Stationary Stationary Stationary

TABLE 4 Cointegration Results for Top Five Exporting Countries Country Belgium France Germany United Kingdom United States

Estimated Equation −1.33 − 0.24 REXC + 0.06GDP + 0.78 D ∗ (4.83) (0.37) (0.40) (0.15) ∗ −23.81 +0.08 REXC + 1.60 GDP∗∗ + 1.51 D ∗ (8.22) (0.51) (0.62) (0.25) −33.55 ∗ −1.13 REXC + 2.57GDP ∗ +0.95 D ∗ (9.56) (1.07) (0.88) (0.35) −1.13 − 0.32 REXC + 0.18 GDP + 1.21 D ∗ (3.57) (0.29) (0.32) (0.17) −35.69 ∗ −1.62 REXC ∗ +2.79GDP ∗ +0.24 D (4.70) (0.32) (0.35) (0.16)

R2

DW

F

0.67

1.31

18.97

0.82

0.80

45.08

0.82

1.29

44.45

0.78

0.78

34.59

0.92

1.43

110.93

Note: ∗Significant at 1% levels. ∗∗Significant at 5% levels. Figures in parentheses are standard error.

7.2 Disaggregated Export Functions Long-run equilibrium results for the top five exporting countries are shown in the Table 4. As expected, the estimation produces consistently positive export income elasticity for all five exporting countries. The magnitude of the export income elasticity varies among the countries. For instance, in the case of France, Germany and the US, we observe elastic export income elasticity. By contrast, it is low in magnitude for the UK and Belgium. This implies that if the real GDP of, say, France, increases by 1 per cent, then exports from Bangladesh will increase by more than 1 per cent. The elasticities of the real exchange rate of selected exporting countries with respect to Bangladesh’s exports are negative and have unexpected signs, except for France. The coefficients of the intercept dummy variables, which have been used to represent the liberalization shift, produce expected positive signs and South Asia Economic Journal 8:2 (2007): 249–264

Estimation of Export Demand Function using Bilateral Trade Data / 261

are significant. However, the effect of trade liberalization on exports is smaller in magnitude for the US. It is important to consider the question of why the elasticities and the impacts of trade liberalization varied among the trading partners. As pointed out, trade liberalization has significant impacts on all of the trading partners, except the US. The US has been a major purchaser from Bangladesh for a long period of time. Hence, the introduction of liberalized policies did not make significant difference in the bilateral trade with the US. As can be seen in Table 5, the share of Bangladeshi imports to total US imports has not changed significantly over the sample period. For instance, while the share increased from 0.05 per cent in 1992 to 0.13 per cent in 2003 in the case of France, it only increased from 0.148 to 0.150 per cent in the case of the US. This could be a reason for the insignificant trade liberalization dummy for the USA. TABLE 5 The Ratio of Bangladesh Imports to Total Import (Percentage) Country Belgium France Germany UK USA

1973 0.14858 0.03286 0.02621 0.10211 0.12129

1992 0.068565237 0.055292055 0.05555054 0.096379133 0.148488951

2003 0.124701 0.133661 0.193279 0.216124 0.150648

Source: Authors’ calculations based on both IMF and Bangladesh Economic Review data. The ratios in Table 5 are expressed in percentage terms.

Bilateral income elasticities are also different in magnitude. The export income elasticity is more than one for France, Germany and for the US. The size of the economies could be a factor behind this. For instance, when the US economy grows by 1 per cent, Bangladeshi exports to the US will increase by 2.79 per cent. It is not surprising that the largest economy of the world has the biggest impact on the export performance in Bangladesh. Another possibility for the different income elasticities is the type of products that are exported to each of these countries. Another reason for the varying elasticities can be the type of goods that these countries import from Bangladesh. We can observe that income elasticity is significant and large in magnitude for France and Germany. Compared to the rest of the countries in the sample, more than 90 per cent of imports from Bangladesh are related to textiles (Knitwear and Woven Garments). In general, they are highly income elastic goods. Major export commodities for each of the country are provided in Table 6. Table 7 compares the income and price elasticities of this study with those of previous studies of Bangladesh. The estimated elasticities of export demand function are consistent with the existing literature, with the exception of price elasticities. South Asia Economic Journal 8:2 (2007): 249–264

262 / MUHAMMED NUR, ALBERT WIJEWEERA

AND

BRIAN DOLLERY

TABLE 6 Composition of Export Commodities to Major Purchasing Countries Country Belgium France Germany UK USA

Export Commodities (2003–04) Knitwear (32.58%), Shrimps (28.85%), Woven Garments (25.35%), Jute and Jute Manufactured (6.92%) Knitwear (52.94%), Woven Garments (39.59%), Home Textile (1.72%), Footwear (leather) (1.82%) Knitwear (41.95%), Woven Garments (52.80%), Shrimps (1.73%), Home Textile (0.78%) Woven Garments (42.53%), Knitwear (32.22%), Shrimps (8.92%), Home textile (6.12%), Bi-cycle (2.94%) Woven Garments (70.58%), Knitwear (12.00%), Shrimps (6.46%), Cap (3.63%), Tent (1.24%)

Source: Export Promotion Bureau (2004–05). TABLE 7 Price Elasticities and Income Elasticities for Export and Import Demand Functions of Bangladesh Export Function References

Price Elasticity

Nguyen and Bhuyan (1977) Shilpi (1990) Roy (1993) Ahmed (2000) Choudhury (2001) Present Study

0.337 0.375 0.18–0.48 0.65 0.18 –0.24 to (–1.62)

Income Elasticity 0.848 0.069 to 3.51 0.24–1.15 0.41 0.65 0.06 to 2.79

Source: Authors’ compilation from Nguyen and Bhuyan (1977), Shilpi (1990), Roy (1993), Ahmed (2000) and Choudhury (2001).

8. Concluding Remarks The empirical estimations conducted in this article allow us to draw various useful conclusions. In the first place, the Engle and Granger cointegration test results suggest a long-run equilibrium relationship between Bangladesh and each of its top five exporting countries, each covered by separate bilateral trade operations, all involving Bangladesh as the supplier. Second, export income elasticity is positive and elastic for France, Germany and the United States. This indicates that as economic growth occurs in these countries, exports from Bangladesh will grow at a greater rate. From a policy perspective, this suggests that continuous monitoring of business cycle movements in these countries could identify opportunities for improving the balance of payments position in Bangladesh, by both securing additional export business and by adjusting macroeconomic policy to meet changed circumstances. Third, export income South Asia Economic Journal 8:2 (2007): 249–264

Estimation of Export Demand Function using Bilateral Trade Data / 263

elasticity is positive, but inelastic, for the United Kingdom and Belgium. Accordingly, although favourable economic conditions in these countries will enhance export earnings in Bangladesh, the rate of growth of these export earnings will be relatively small. It follows that research and development should focus on identifying further export opportunities in these nations. The fourth important finding is the negative export price elasticity for all top five exporting countries, except France. Unanticipated signs with insignificant coefficients might be attributed to the limitations of the model, spelled out earlier in the model specification section. Finally, we find that the impact of trade liberalization on exports is highly significant and positive. However, the impact of trade liberalization is rather small in the American case. In general, the different export income and price elasticities established in this article seem to suggest that a single external trade policy to improve the external sector will not prove efficacious for Bangladeshi policy makers. Instead of a uniform approach, a good understanding of the socio-economic environment and business climate in the export recipient countries and tailor-made policies for each country, would better assist in enhancing export earnings for Bangladesh. References AGOSIN, M. (1991), ‘Trade Policy Reform and Economic Performance: A Review of the Issues and Some Preliminary Evidence’, Discussion Paper No: 41, UNCTAD. AHMED, N. (2000), ‘Export Responses to Trade Liberalization in Bangladesh: A Cointegration Analysis’, Applied Economics, August, 32(8): 1077–84. BALASSA , B. (1971), ‘Trade Policies in Developing Countries’, American Economic Review, 61(2): 178–87. BANGLADESH ECONOMIC REVIEW (2003), Ministry of Finance, Dhaka, Bangladesh. BANGLADESH ECONOMIC REVIEW (2004), Ministry of Finance, Dhaka, Bangladesh BHAGWATI, J.N. (1988), ‘Export Promoting Trade Strategies, Issues and Evidence’, The World Bank Research Observer’, 3(1): 27–58. BLEANEY, M. (1999), ‘Trade Reform, Macroeconomic Performance and Export Growth in Ten Latin American Countries, 1979–95’, Journal of International Trade and Economic Development, 8(1): 89–105. CHOUDHURY, R.F. (2001), ‘Export Demand Function of Bangladesh: An Econometrics Exercise’, The Bangladesh Development Studies, September, 27(3): 68–90. CLARKE , R. and C. K IRKPATRICK (1992), ‘Trade Policy Reform and Economics Performance in Developing Countries: Assessing the Empirical Evidence’, in R. Adhikari, C. Kirkpatrick and J. Weiss (eds), Industrial and Trade Policy Reform in Developing Countries. Manchester: Manchester University Press. CORDON, W.M. (1966), ‘The Structure of a Tariff System and Effective Protection Rate’, Journal of Political Economy, 74(3): 221–37. DORNBUSCH, R. (1992), ‘The Case for Trade Liberalization in Developing Countries’, Journal of Economic Perspectives, 6(12): 69–85. DUTTA, D. and N. AHMED (2000), ‘An Aggregate Import Demand Function for India: A Cointegration Approach’, Applied Economics, 31(4): 465–72.

South Asia Economic Journal 8:2 (2007): 249–264

264 / MUHAMMED NUR, ALBERT WIJEWEERA

AND

BRIAN DOLLERY

ENGLE, R.F. and C.W.J. GRANGER (1987), ‘Cointegration and Error Correction: Representation, Estimation and Testing’, Econometrica, 55(March): 251–76. EXPORT PROMOTION BUREAU (2005), Bangladesh Export Statistics. Bangladesh. HOSSAIN, M.A. and M. ALAUDDIN (2005), ‘Trade Liberalization in Bangladesh: The Process and Its Impact on Macro Variables Particularly Export Expansion’, The Journal of Developing Areas, 39(1): 127–150. HOUTHAKKER, H.S. and MAGEE, S.P. (1969), ‘ Income and Price Elasticities in World Trade’, Review of Economics and Statistics, 51(2): 11–125. HUSSAIN, M.N. (2004), ‘Estimation of Import Demand Functions for Bangladesh Economy’, Economic Affairs, 50(4): 225–33. IMF (INTERNATIONAL MONETARY FUND) (2004), International Financial Statistics Yearbook. Joshi, V. and L.M.D. Little (1996), India’s Economic Reforms 1991–2001. Oxford: Oxford University Press. LITTLE, I.M.D., T. SCITOVSKY and M. SCOTT (1970), Industry and Trade in Some Developing Countries: A Comparative Study. London: Oxford University Press. LOPEZ, P.P. (2005), ‘The Impact of Trade Liberalisation on Exports, Imports, the Balance of Payments and Growth: The Case of Mexico’, Journal of Post Keynesian Economics, 27(4): 595–619. MARQUEZ, J. and C. MCNEILLY (1988), ‘Income and Price Elasticities for Exports of Developing Countries’, The Review of Economics and Statistics, 70(2): 306–14. NGUYEN, D.T. and A.R. BHUYAN (1977), ‘Elasticities of Export and Import Demand in Some South Asian Countries’, The Bangladesh Development Studies, 2: 133–52. PAULINO, A.S. and A.P. THIRLWALL (2004), ‘The Impact of Trade Liberalization on Exports, Imports, and the Balance of Payments of Developing Countries’, The Economic Journal, 114(1): 50–72. ROY, D.K. (1991), ‘Export Performance of Bangladesh: A Constant Market Share Analysis’, The Bangladesh Development Studies, December, 19(4): 63–81. ——— (1993), ‘Impact of Incentives on Export Performance of Bangladesh: A Preliminary Assessment’, The Bangladesh Development Studies, June, 21(2): 25–44. SENHADJI, A.S. and C. MONTENEGRO (1999), ‘Time Series Analysis of Export Demand Equations: A Cross-Country Analysis’, IMF Staff Papers, June, 46(3): 236–68. SHILPI, F.J. (1990), Estimating Income and Price Elasticities of Imports and Exports of Bangladesh. Research Report No.2, Bangladesh Institute of Development Studies, Dhaka. THOMAS, V., J. NASH and S. EDWARDS (1991), Best Practices in Trade Policy Reform. Oxford: Oxford University Press.. WEISS, J. (1992), ‘Export Response to Trade Reforms: Recent Mexican Experience’, Development Policy Review, 10(1): 45–60.

South Asia Economic Journal 8:2 (2007): 249–264

Suggest Documents