Determinants of trade flows in APEC member economies

Nguyễn Khánh Doanh Tạp chí KHOA HỌC & CÔNG NGHỆ 81(05): 45 - 50 45 DETERMINANTS OF TRADE FLOWS IN APEC MEMBER ECONOMIES Nguyen Khanh Doanh* Thainguyen University of Economics and Business Administration - TNU ABSTRACT This paper attempts to analyze the determinants of trade flows among APEC member economies. Using the panel data analysis, the empirical results show a number of robust findings. First, GDP is one of the most important determinants of trade flows among APEC economies. Second, membership in an FTA would likely lead to an increase in trade among the member countries. Third, countries speaking common languages tend to trade more with each other. Fourth, distance remains a hindrance to trade flows even though technological innovations continue to spark reductions in transport costs. Fifth, countries having colonial relationship or sharing the same border have tendency to trade more with each other. Finally, for APEC economies, it might be that the economies-of-scale effect is greater than the absorption effect, which allows the advantages of economies of scale to be fully exploited. Efforts to increase the GDP of APEC member economies, enhance social infrastructure, improve language ability and reduce the cultural differences are suggested as remedies for the obstacles to freer flow of trade in the region. Keywords: Trade Flows, Fixed Effects Model, Random Effects Model, APEC INTRODUCTION* Economists have recognized that international trade has made significant contribution to and serve as an engine for economic growth. The major gains for countries liberalizing trade could be realized through improved efficiency as a result of greater competition, specialization and economies of scale, increased availability of imported inputs, and enhanced access to foreign technology. In terms of this vital contribution, member economies of the Asia-Pacific Economic Cooperation (APEC) are not exceptional since international trade is an important economic dynamic in this region. Founded in 1989, APEC aims at promoting free trade and economic cooperation throughout the Asia-Pacific region. With a population of about 2.7 billion people, APEC made up 50.1 percent of world GDP in 20094. In terms of trade performance, the member economies of APEC accounted for 39 percent of world exports and 40 percent of world imports in 19905. This figure increased to * 4 Based on the data from the United Nations Statistics Division 5 Based on data from the World Economic Outlook Database - IMF 45.6 percent and 45.4 percent respectively in 2009. Given the increasing importance of APEC member economies in the world trade, it is important to identify the main sources of international trade in this region. The main purpose of this paper is to investigate the determinants of trade among APEC member economies. Therefore, it is guided by the following research objectives:  To analyze the factors that affect the trade flows among APEC member economies.  To determine whether or not an FTA membership has the positive effect on trade flows.  To derive policy implications based on the empirical analysis. The paper is structured as follows: Section 2 presents the econometric method, including the fixed effects model and random effects model, which is used for the analysis in this paper. Section 3 describes the data used for the sample. Section 4 displays and interprets the regression results. Concluding remarks and policy implications are included in the final section. THEORETICAL FRAMEWORK 2.1. Analytical model To determine the trade performance of APEC member economies, I use the standard gravity Số hóa bởi Trung tâm Học liệu - Đại học Thái Nguyên Nguyễn Khánh Doanh Tạp chí KHOA HỌC & CÔNG NGHỆ 81(05): 45 - 50 46 model augmented by some new variables. The gravity model in international trade was developed by Tinbergen (1962), Poyhonen (1963) and Linnemann (1966) and applied in a number of empirical studies (Okubo, 2007; Disdier and Head, 2008; Magee, 2008; Wang et al., 2010). In its original form, it is assumed that the volume of bilateral trade between two countries is determined by their economic size and the distance between them (Fidrmuc, 2009). Since then, the gravity model has been widely used and increasingly improved in empirical studies of international trade. In addition, more variables have been incorporated to the model to account for trade flows. The final regression equation is as follows: (1) Where: Tijt is the bilateral trade between country i and country j at the time t.6 GDPit is Gross Domestic Product of country i at the time t. GDPjt is Gross Domestic Product of country j at the time t. POPit is the population of country i at the time t. POPjt is the population of country j at the time t. DISTij: Distance between the capital city of country i and the capital city of country j. Boderij is a dummy variable that equals 1 if both country i and country j share a common border, and zero otherwise. Languageij is a dummy variable that equals 1 if both country i and country j speak the same language, and zero otherwise. Colonyij is a dummy variable that equals 1 if country i had ever been colonized by country j or vice versa, and zero otherwise. FTAijt is a dummy variable that equals 1 if both country i and country j belong to a FTA at the time t, and zero otherwise. :ijtu Residual term 6 Drawing on Musila (2005), the dependent variable log(Tijt + 1) is used instead of log Tijt in order to include the observations of zero measurable trade. For the estimation purpose, the equation 1 is expressed in log-linear form as follows: (2) According to the gravity assumptions, the coefficients on GDP (GDPit and GDPjt) are positive (See e.g. Magee, 2008; Martinez- Zarzoso et al., 2009). The parameters on population (POPit and POPjt) could be positive or negative depending on whether the absorption effect or the economies of scale effect is dominant (See Linnemann, 1966; Endoh, 2000; Martinez-Zarzoso and Nowak- Lehmann, 2003; Koo et al., 2006; Magee, 2008). Distance between trading partners (DISTij) reflects the cost of international transactions of goods and services and are expected to affect trade negatively (See, e.g., Lee and Shin, 2006; Martinez-Zarzoso et al., 2008). Therefore, the sign of the coefficient for DISTij variable is expected to be negative. Since linguistic affinity, ex-colony and commonly shared borders tend to reduce cultural distance and therefore encourage bilateral trade, it is expected that the coefficients for these three dummy variables are positive (Peridy, 2005; Lee and Shin, 2006). Finally, a dummy variable is included to capture the integration effect of the FTA. The coefficient on FTA could be negative or positive depending on a case-by-case basis (Koo et al., 2006; Lee and Shin, 2006; Baier and Bergstrand, 2007; Jayasinghe and Sarker, 2007; Gil-Pareja et al., 2008). A positive and significant coefficient on the FTA dummy could imply that its members have traded with each other more than the hypothetical level predicted by basic explanatory variables. 2.2. Method of estimation In this paper, two techniques are employed, including the fixed effects model and random effects model. The fixed effects model allows for country-pair heterogeneity and gives each country-pair its own intercept. The equation for fixed effects model is expressed in the following form: Số hóa bởi Trung tâm Học liệu - Đại học Thái Nguyên Nguyễn Khánh Doanh Tạp chí KHOA HỌC & CÔNG NGHỆ 81(05): 45 - 50 47 (3) Where: β0ij indicates that each country-pair has its own intercept. The fixed effects estimates can help us reduce potential specification errors from omitting important variables. One shortcoming of this model, however, is that it does not allow for time-invariant variables to be included7. Therefore, we include the random effects model in order to incorporate differences between cross-sectional entities by allowing the intercept to change, as in the fixed effects model, but the amount of change is random. The random effects model is expressed as follows: (4) Where: β0 is the mean intercept, and wijt is composite error term (wijt = µij + uijt). µij is a random unobserved bilateral effect (which is cross-section or country-pair error component), and uijt is the remaining error (which is the combined time series and cross- section error component). The random effects model requires that µ ij ~ (0,σ2µ), uijt ~ (0,σ2µ), the µij is independent of the uijt, and the explanatory variables have to be independent of the µij and the uijt for all cross-sections (ij) and time periods (t). The advantage of random effects model is that both time-series and cross-sectional variations are used. The method adopted is the GLS random-effects. DATA This paper uses the panel data for 19 APEC member economies over the period of 16 years, from 1994 through 20098. They include Australia, Canada, Chile, China, Hong Kong- 7 Examples of time-invariant variables include distance, border, etc. 8 Brunei Darussaram and Chinese Taipei are excluded from the sample due to lack of data on these two economies. The period 1994-2009 is chosen because data before this period are not available for all observations. China, Indonesia, Japan, Korea, Malaysia, Mexico, New Zealand, Philippines, Papua New Guinea, Peru, Russia, Singapore, Thailand, United States and Vietnam. Yearly total trade between two countries is obtained from the IMF-Direction of Trade Statistics (CD-ROM). Data on GDP and population are extracted from World Economic Outlook Database - IMF and the United Nations Statistics Division. The distance between the two capital cities is available at Indo.com. Finally, information regarding language and colonial relationship is obtained from the Economist Intelligence Unit. EMPIRICAL RESULTS The summary of statistics is displayed in Table 1. The samples include trade flows from 19 APEC member economies for the period 1994- 2009, leading to 2,736 observations. Estimates of the bilateral trade flows using equations 3 and 4 are presented in Table 2. As explained above, in the fixed effects model, the variables log DISTij and Borderij are dropped because these variables are time- invariant. Because of this reason, the interpretation of the results will be on the basis of random effects model. As a result of fixed effects and random effects models, the gravity model fits the data well, providing explanation for the major variation in bilateral trade. Most of the coefficients are estimated as expected and three of them are statistically significant at the 0.01 significance level. According to the results of the fixed effects model, GDPit and GDPjt turn out to be the most important explanatory variables, not unexpectedly. As indicated in Table 2, the coefficients of log GDPit and log GDPjt are positive and statistically significant. This suggests that GDP growth in APEC member economies would trigger and accelerate the expansion of trade. This result is consistent with trade theory and empirical studies undertaken by Magee (2008) and Hatab et al. (2010). The estimated value of 0.627 means that, holding Số hóa bởi Trung tâm Học liệu - Đại học Thái Nguyên Nguyễn Khánh Doanh Tạp chí KHOA HỌC & CÔNG NGHỆ 81(05): 45 - 50 48 constant for other variables, a 100 percent increase in country i’s GDP would lead to an increase in its trade by 62.7 percent. Likewise, an increase in country j’s GDP by 100 percent would result in 71 percent increase in country j’s trade. One of the important issues in this paper is the impact of FTA on bilateral trade flows. The estimated coefficient on the FTA dummy variable is positive and statistically significant. Therefore, membership in an FTA would lead to an increase in bilateral trade with member countries of that FTA. The estimated value of 0.154 means that a pair of countries that joins an FTA will likely experience an increase in bilateral trade between them by a roughly 16.6 percent, with other variables held constant. Although the estimated coefficient of log POPit and log POPjt are statistically insignificant, its positive value could be indicative that the economies-of-scale effect is greater than the absorption effect, which allows the advantages of economies of scale to be fully exploited. So, in the case of APEC, large population might promote a division of labor and allow more industries to reach efficient economies of scale. Thus, opportunities for trade with foreign partners in a wide variety of goods will increase. Table 1. Summary of Statistics Variable Observations Mean Std. Dev. Minimum Maximum Log GDPit 2736 3.10 1.13 0.00 5.77 Log GDPjt 2736 2.46 0.68 0.49 4.16 Log POPit 2736 2.34 0.83 0.49 4.16 Log POPjt 2736 1.58 0.69 0.53 3.13 Log DISTij 2736 4.63 0.62 3.53 6.69 Borderij 2736 3.83 0.37 2.48 4.29 Languageij 2736 0.04 0.20 0.00 1.00 Colonyij 2736 0.25 0.43 0.00 1.00 FTAij 2736 0.06 0.25 0.00 1.00 rta 2736 0.12 0.33 0.00 1.00 Source: Statistical result Note: Std. Dev. stands for standard deviation. Table 2 Gravity Equations Explaing Total Trade Explanatory Variables Fixed Effects Model Random Effects Model Coefficient t-statistic Coefficient z-statistic Constant -1.630* (-2.49) 4.447** (9.12) Log GDPit 0.627** (15.51) 0.660** (21.24) Log GDPjt 0.710** (17.35) 0.766** (24.70) Log POPit 0.037 (0.15) 0.009 (0.16) Log POPjt 0.307 (1.90) 0.009 (0.14) Log DISTij - - -1.287** (-12.43) Borderij - - 0.033 (0.18) Languageij 0.085 (0.38) 0.344** (4.18) Colonyij 0.038 (0.17) 0.083 (0.67) FTAij 0.154** (3.62) 0.138** (3.39) Number of observations 2736 2736 R-square (overall) 0.542 0.804 Source: Regression results Note: * Significant at the 0.05 level; ** Significant at the 0.01 level. Số hóa bởi Trung tâm Học liệu - Đại học Thái Nguyên Nguyễn Khánh Doanh Tạp chí KHOA HỌC & CÔNG NGHỆ 81(05): 45 - 50 49 In the random effects model, the results are relatively similar to those of the fixed effects model, meaning that a very high degree of explanation is achieved. The basic variables of gravity equation behave as the model predicts. As the data reveal, most of the coefficients are statistically significant at 0.01 significant level, except POPit, POPjt, Borderij and Colonyij. Again, GDP (GDPit and GDPjt) proves to be the most important explanatory variables. According to the model estimation, an increase in GDP of country i by 100 would likely result in an increase in country i’s trade by 66 percent, with other variables controlled. Similarly, a 100 percent increase in country j’s GDP would lead to a 76.6 percent increase in country j’s trade, with other variables being kept constant. The estimated coefficient on the log of bilateral distance is negative and statistically significant. This means that distant countries tend to trade less with each other. According to the estimation, an increase in bilateral distance by 100 percent leads to a 72.4 percent decline in bilateral trade. The coefficient on Languageij is positive and highly significant. This suggests that countries speak the same language have tendency to trade more with each other. Specifically, it is estimated that two countries speaking the same language are likely to trade more with each other by 41 percent. Apart from POPit and POPjt, the model also finds the traditional positive signs on Borderij and Colonyij. Although statistically insignificant, the positive value could be indicative that an ex-common colonizer could raise trade by 8.7 percent, while a commonly share border could increase trade by 3.4 percent. CONCLUSION This paper attempts to analyze the determinants of trade flows among APEC member economies. Using the panel data analysis with the fixed and random effects models, the empirical results show a number of robust findings. First, GDP is one of the most important determinants of trade flows among APEC economies. Countries with higher level of GDP tend to trade more because higher level of exporting country’s GDP indicates higher level of production for exports, while higher level of importing country’s GDP suggests higher level of demand for imports. Second, membership in an FTA would likely lead to an increase in trade among the member countries. Third, countries speaking common languages tend to trade more with each other since they can facilitate easier transactions and reduce the cost of doing business (e.g. translations and disputes). Fourth, distance remains a hindrance to trade flows even though technological innovations continue to spark reductions in transport costs. Integration and globalization have enhanced communication, broken down cultural barriers, and facilitated transactions. However, they have not reduced the importance of physical distance. Fifth, although being statistically insignificant, the positive values of the coefficient on Colonyij indicate that countries also tend to trade more with their ex-colonizers since they are more familiar with the cultural backgrounds and modes of doing business. Similarly, countries which share the same border have tendency to trade more with each other. Finally, for APEC economies, it might be the case that the economies-of-scale effect is greater than the absorption effect, which allows the advantages of economies of scale to be fully exploited. Efforts to increase the GDP of APEC member economies, enhance social infrastructure, improve language ability and reduce the cultural differences are suggested as remedies for the obstacles to freer flow of trade in the region. REFERENCES [1].Baier, S. and Bergstrand, J. H. (2007), ‘Do free trade agreements actually increase members' international trade?,’ Journal of International Economics 71 (1): 72–95 Số hóa bởi Trung tâm Học liệu - Đại học Thái Nguyên Nguyễn Khánh Doanh Tạp chí KHOA HỌC & CÔNG NGHỆ 81(05): 45 - 50 50 [2].Disdier, A-C. and Head, K. (2008), ‘The Puzzling Persistence of the Distance Effect on Bilateral Trade,’ Review of Economics and Statistics 90 (1): 37-48. [3].Endoh, M. (2000), ‘The Transition of Postwar Asia-Pacific Trade Relations,’ Journal of Asian Economics 10 (4): 571-589. [4].Fidrmuc, J. (2009), ‘Gravity Models in Integrated Panel,’ Empirical Economics 37 (2): 435-446. [5].Gil-Pareja, S., Llorca-Vivero, R. And Martinez-Serrano, J. A. (2008), ‘Trade Effects of Monetary Agreements: Evidence from OECD Countries,’ European Economic Review 52 (4): 733-755. [6].Hatab, A. A., Romstad, E. and Huo, X. (2010), ‘Determinants of Egyptian Agricultural Exports: A Gravity Model Approach,’ Modern Economy 1 (3): 134-143. [7].Jayasinghe, S. and Sarker, R. (2007), ‘Effects of Regional Trade Agreements on Trade in Agrifood Products: Evidence from Gravity Modeling Using Disaggregated Data,’ Review of Agricultural Economics 30 (1): 61-81. [8].Koo, W.W., Kennedy, P. L. and Skripnitchenko, A. (2006), ‘Regional Preferential Trade Agreements: Trade Creation and Diversion Effects,’ Review of Agricultural Economics 28 (3): 408-415. [9].Lee, J-W. and Shin, K. (2006), ‘Does Regionalism Lead to More Global Trade Integration in East Asia,’ North American Journal of Economics and Finance 17 (3): 283-301. [10]. Linnemann, H. (1966), An Econometric Study of International Trade Flows, Amsterdam: North Holland Publishing Co. [11]. Magee, C. S. P. (2008), ‘New Measures of Trade Creation and Trade Diversion,’ Journal of International Economics 75 (2): 349-362. [12]. Martinez-Zarzoso, I. and Nowak-Lehmann, F. (2003), ‘Augmented Gravity Model: An Empirical Application to Mercosur-European Union Trade Flows,’ Journal of Applied Economics 6 (2): 291-316. [13]. Martinez-Zarzoso, I., Perez-Garcia, E. M. and Suarez-Burguet, C. (2008), ‘Do Transport Costs have a Differential Effect on Trade at the Sectoral Level?,’ Applied Economics 40 (24): 3145-3157. [14]. Martinez-Zarzoso, I., Felicitas, N-L. D. and Horsewood, N. (2009), ‘Are Regional Trading Agreement Beneficial? Static and Dynamic Panel Gravity Models,’ North American Journal of Economics and Finance 20 (1): 46-65. [15]. Musila, J. W. (2005), ‘The Intensity of Trade Creation and Trade Diversion in COMESA, ECCAS and ECOWAS: A Comparative Analysis,’ Journal of African Economies 14 (1): 117-141. [16]. Okubo, T. (2007), ‘Trade Bloc Formation in Inter-war Japan: A Gravity Model Analysis,’ Journal of Japanese International Economies 21 (2): 214-236. [17]. Peridy, N. (2005), ‘The Trade Effects of the Euro-Mediterranean Partnership: What Are the Lessons for ASEAN Countries,’ Journal of Asian Economics 16 (1): 125-139. [18]. Poyhonen, P. (1963), ‘A Tentative Model for the Volume of Trade between Countries,’ Weltwirtschaftliches Archiv 90: 93-100. [19]. Tinbergen, J. (1962), Shaping the World Economy: Suggestions for An International Economic Polity, New York: Twentieth Century Fund. [20]. Wang, C., Wei, Y. and Liu, X. (2010), ‘Determinants of Bilateral Trade Flows in OECD Countries: Evidence from Gravity Panel Data Models,’ The World Economy 33 (7): 894-915. Số hóa bởi Trung tâm Học liệu - Đại học Thái Nguyên