A gravity model for world natural rubber trade

UNIVERSITY OF ECONOMICS HO CHI MINH CITY VIETNAM ERASMUS UNVERSITY ROTTERDAM INSTITUTE OF SOCIAL STUDIES THE NETHERLANDS VIETNAM – THE NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS A GRAVITY MODEL FOR WORLD NATURAL RUBBER TRADE BY VO CONG DANH MASTER OF ARTS IN DEVELOPMENT ECONOMICS HO CHI MINH CITY, OCTOBER 2016 ACKNOWLEDGEMENTS First and foremost, I would love to sincerely express my thanks to my thesis supervisor Dr Pham Thi Bich Ngoc who provided me with invaluable academic guidance Her research experiences, particularly with the issue of international trade helped me from building thesis proposal to the final thesis report with patient and timely guidance I would like to express my special thanks to Dr Tran Thi Thuy Hoa from Vietnam Rubber Association for her supports during writing this thesis I would also like to extend my thanks to all Vietnam – The Netherlands professors, staffs for their knowledge, helps and encouragement In particular, I am very grateful to MA Nguyen Son Kien and MA Trinh Hoang Viet who are also VNP students, their enthusiastic support gives me invaluable advices Finally, I would give the deepest thanks to my beloved parents who unconditionally support me in everything I do, including the completing of this dissertation I ABSTRACT The purpose of this empirical analysis is to investigate the determinants of world natural rubber trade flows based on gravity model among four largest exporting countries, namely, Thailand, Indonesia, Malaysia, Vietnam and their fourteen major importing partners: Brazil, Canada, China, France, Germany, India, Indonesia, Italia, Japan, Korea, Malaysia, Poland, Spain, Turkey and the US in the period of 2000 to 2013 The thesis uses panel data in a gravity model with different fixed effects and GMM model with exogenous instrument variables to control for endogeneity problems causing by the correlation among independent variables and the error term as well as time-varying multiple price term The empirical results suggest that gravity model are reliably applied to a single industry, at least for the case of world natural rubber bilateral trade flows Significant effects of economic size as well as and trade cost between trading countries with expected signs on world bilateral natural rubber trade reveals that both value and volume of bilateral natural rubber traded can be treated as alternative methods to explain the bilateral trade in gravity models even for disaggregated trade The research findings also reveal that world’s natural rubber trade flows are positively affected by car production of importing countries when controlling for bilateral fixed effects and year dummies It means that almost natural rubber in ASEAN countries are exported to countries with developed rubber downstream industry II TABLE OF CONTENTS ACKNOWLEDGEMENTS i ABSTRACT ii TABLE OF CONTENTS iii LIST OF TABLES v LIST OF FIGURES v CHAPTER INTRODUCTION 1.1 Problem statement 1.2 Research objectives 1.3 Research questions 1.4 Data and methodology 1.5 Thesis structure CHAPTER LITERATURE REVIEW 2.1 Trade theory 2.1.1 Absolute and Comparative advantage 2.1.2 Hecksher-Ohlin model 2.1.3 New trade theory 2.2 Gravity model 2.2.1 Traditional gravity equation 10 2.2.2 Anderson gravity equation 11 2.3 Empirical application of gravity model 12 2.3.1 Dependent Variable Specification 13 2.3.2 Traditional gravity variables for trade 14 2.4 Empirical research for natural rubber trade 19 2.5 Overview of world natural rubber industry 25 III CHAPTER DATA AND METHODOLOGY 29 3.1 Data Source 29 3.2 Model specification 31 3.3 Estimation strategy and correction method 32 3.3.1 Ordinary least square (OLS) 32 3.3.2 Fixed effect model (FEM) 33 3.3.3 Random effect model (REM) 34 3.3.4 Pooled OLS, Fixed effect or Random effect 34 3.3.5 Econometric problems 36 CHAPTER EMPIRICAL RESULTS 38 4.1 Descriptive statistics 38 4.2 Empirical results 41 4.3 GMM model to control the potential endogeinety 48 CHAPTER CONCLUSIONS 51 5.1 Main findings 51 5.2 Limitations and further studies 52 REFERENCE 54 APPENDIX 56 IV LIST OF TABLES Table 2.1: Top tyre manufacturers worldwide in 2012 based on the tyre sale 21 Table 2.2: Some literature reviews about determinants of international trade using gravity model and natural rubber trade 24 Table 3.1: Data source of variables and their expected sign 32 Table 4.1: Bilateral natural rubber trade in 2000 39 Table 4.2: Bilateral natural rubber trade in 2013 40 Table 4.3: Correlation table of variables 41 Table 4.4: VIF results 44 Table 4.5: Estimation Results for world natural rubber trade value 45 Table 4.6: Estimation Results for world natural rubber trade volume 46 Table 4.7: Panel gravity equations in levels using various specifications 48 Table 4.8: Estimation results using GMM for controlling endogeneity 50 LIST OF FIGURES Figure 3.1: Diagram of the comparison progress between FE & RE models 36 Figure 4.1: World’s natural rubber production 2011 – 2013 (Unit: Tonnes) 26 Figure 4.2: Top natural rubber exporting country 2015 in term of value 27 Figure 4.3: World’s natural rubber export 2011 – 2015 (Unit: Tonnes) 28 V CHAPTER INTRODUCTION 1.1 Problem statement The importance role of international trade in the world has been broadly studied in the many issues Many researchers have proved that increased participation in international market would motivate trade activities, which is an integral condition for further economic growth of any countries Joining global markets to connect with producers and buyers in many countries, trade provides an environmental for related parties to reduce the cost of production, attract foreign investment, enhance the value-added of their products and provide countries opportunity to move up global value chain Which is the recent strategic orientation followed by developing and emerging countries in Asia and ASEAN countries A fundamental factor behind their rapid economic growth has been their ability to strengthen the competitive productive and export abilities, first in traditional agricultural industry and then in labour-intensive manufactures The growing role of globalization today makes international trade become one of the most issues need to be discovered with multi-million negotiations take place every day However, a number of areas of international trade are still not thoroughly studied and current traditional theories of trade also need to be further employed in empirical application Although many trade theories primarily try to question why nations trade with one another? However, a quantitative question of how large and the destination of trade still remains as substantial field for further studies In recent years, gravity model has appeared as the most fashion and successful analytical instrument to quantify international trade flows, particularly thanks to the high explanatory ability and available data for easily collecting regarding to international trade statistics of goods According to Prehn and Brümmer (2011) a huge number of papers using gravity trade models to investigate the disaggregated data trade flows, however, the applications to a single commodity industry are still room for further study Particularly, investigating the determinants of bilateral trade flows for the specific industry in developing countries is a practical empirical task In fact, the analysis of bilateral trade flows is significantly essential for the natural rubber industry This is true for rubber importing nations, which are natural rubber main end-users, and rubber producing countries, which are more and more depending on natural rubber trade as main source of national income and employment generation Natural rubber is one of the most essential natural resource in ASEAN area which relates to foreign exchange earnings, unemployment reduction, and last but not least, material input to the industrial production in both home and foreign markets By 2013, rubber production of four largest producing countries, namely, Thailand, Indonesia, Malaysia, Vietnam, respectively, has a share of around 82.2 percent of world production, according to the Association of Natural Rubber Producing Countries (ANRPC) Base on the data collected from UN Comtrade, world natural rubber import value have grown dramatically in recent years, rising from about 3.8 USD billion in 2000 to almost 25.2 USD billion in 2013, in which the contribution of largest producing countries account for 78 percent to 84 percent, with an average annual growth rate of almost 15.6 percent (CAGR) However, the downsizing of the world’s trade of natural rubber in term of US value from percent for 2000 – 2010 reduce to -28 percent for 2011 – 2014, and expected would continue decline in the coming years, which is attributed to the consequence of uncontrolled rubber plantations in the period of 2005 – 2011 by Thailand, Indonesia, Malaysia, Vietnam and even China With the high proportion of natural rubber production joining international market, the study of natural rubber trade is importance for any policy related to rubber industry In this study a gravity model is developed to discover the possibility of using gravity model to explore the trade flows of world natural rubber and to test whether world natural rubber trade depend on the development level of downstream industries Although the natural rubber has an important role as a critical material input for the production of elastomers which are essential in automotive and general rubber goods industry since they can be almost distorted and still return to their original shape Due to this unique property, natural rubber is used by variety of industries to produce tires, gloves, condoms, mattress, equipment for general physical exercise, automobile spare parts such as mounts, springs, absorbers, and dampers, etc The empirical literature analyzing natural rubber international trade is still rather limited As few number of empirical analyses available in this area, existing researches just focus on investigating the determinants of natural rubber trade for a single country as Yusof (1988), Kannan (2013), Chawananon (2014) Therefore, we have mainly concentrated on applying the gravity model to explore the bilateral trade at disaggregated level between four largest natural rubber exporting countries including, Thailand, Indonesia, Malaysia and Vietnam which accounting for 82.2 percent of world production, with their fourteen major natural rubber importing countries which accounting for over 90 percent of world import, namely: Brazil, Canada, China, France, Germany, India, Italia, Japan, Korea, Malaysia, Poland, Spain, Turkey and the US in the period of 2000 to 2013 1.2 Research objectives From those important points stated above, the main objectives of the study are: Investigating whether world natural rubber bilateral trade depend on the development level of rubber downstream industries 1.3 Research questions To deal with these above objectives, the research questions need to be specified in this thesis are: (1) Do traditional gravity variables such as the economic size of country and bilateral geographical distance or trade cost have an expected significant impacts on bilateral trade between natural rubber trading partners? (2) Do factors such as Industrial Production Index (IPI) and car production of natural rubber importing countries have significantly positive impacts on bilateral natural rubber trade? 1.4 Data and methodology This study uses the annually bilateral natural rubber trade data collected panel data of four exporting countries and their fourteen major trading partners on United Nations Comtrade Database (UN Comtrade) from 2000 to 2013 The data of GDP and Population will be collected through World Bank Data of bilateral geographic distance and trade cost are taken from The United Nations Economic and Social Commission for Asia and the Pacific (ESCAP-World bank) database The data of country industrial production index (IPI) will be gathered from The International Monetary Fund (IMF), while the car productions are available on The International Organization of Motor Vehicle Manufacturers (OICA) website After acquiring the strongly balanced panel data, the Ordinary Least Squares (OLS), Fixed effects (FE) and Random effects (RE) model are employed to analyze the relationship between natural rubber trade flows and its explanatory variables The outcomes of different Fixed effects models are considered as main results Table 4.8: Estimation results using GMM for controlling endogeneity Trade value Trade volume (1) (2) LnGDPi 1.497*** (0.223) 1.234*** (0.240) LnGDPj 0.591** (0.270) 0.430* (0.245) LnPopi -0.338* (0.197) -0.159 (0.204) LnPopj -0.125 (0.169) -0.065 (0.201) Lndisij -0.521 (0.320) -0.289 (0.378) TCij -0.008*** (0.003) -0.009*** (0.003) Caryj 0.463* (0.241) 0.414* (0.226) IPI 0.003 (0.007) 0.003 (0.005) AR (1) 0.000 0.009 AR (2) 0.048 0.638 AR (3) 0.597 0.402 Hansen Test 0.163 0.292 Instruments L3.(LnGDPi, LnGDPj) L2.(LnGDPi, LnGDPj) Year dummy Yes Yes Note: *, **, *** correspond with 1%, 5% and 10% significance level 50 CHAPTER CONCLUSIONS The purpose of this paper is to investigate the Determinants of World Natural Rubber Trade Using Gravity Model A motivation for this paper is to examine the validity of using gravity model to analyze disaggregated sectoral data level, especially for world natural rubber trade By collecting the world natural rubber bilateral trade data of four largest exporting countries, namely, Vietnam, Thailand, Malaysia, and Indonesia with their fifteen major trading partners namely: Brazil, Canada, France, China, Germany, India, Italia, Japan, Korea, Malaysia, Poland, Spain, Turkey and The US in the period of 2000 – 2013, this study has success in drawing some important findings as follows Furthermore, suggestions for further researches could be implied in the next sections 5.1 Main findings The empirical results indicate that the gravity model is very effective in explaining world’s natural rubber bilateral trade flows when controlling for endogeneity problems by fixed effects and especially GMM model with the lag of LnGDP as instrument variable In more detail, the results from section 4.3 prove that the third lag of LnGDPi and LnGDPj are valid instrument variables for world bilateral natural rubber trade value, and, the second lag of LnGDPi and LnGDPj are valid instrument variables for world bilateral natural rubber trade volume The results from fixed effects and GMM model are quite consistent, in which the world’s natural rubber bilateral trade are positively determined by the economic size of both exporting and importing country, however, bilateral trade cost is proved to hinder trade It is demonstrated that the gravity model is also well applicable to the case of single industry with disaggregate data level, at least for natural rubber trade 51 The coefficient on the trade structure variables identifies that world’s natural rubber trade is affected not only by traditional gravity variables but also by factor other than gravity variables To be more detailed, natural rubber trade is hindered by the market size (population) of exporting countries, it confirms the findings of Rahman (2003), Hassan et al (2010) and Natale et al (2015) which is stated that poorer country tend to trade less with others The prominent findings of this thesis are natural rubber is tend to be exported to developed rubber downstream industry countries, to be more detailed, the coefficients of car production of importing countries are positive impacts to world natural rubber trade when eliminating the potential problems of endogeneity, however the impacts industrial production index on natural rubber trade is not significant To promote natural rubber export the governments of four largest exporting countries should focus on the countries with high level of rubber downstream industry development The evidence of small but significant negative effect of trade cost suggests that natural rubber exporting countries should focus on enhancing trade promotion with Asia countries which have increasingly relied on importing natural rubber in recent year and have low bilateral trade cost with ASEAN countries as well 5.2 Limitations and further studies This level of classification does not allow distinguishing among grades For example, natural rubber are grouped in one commodity which does not differentiate between Technically Specified Rubber (TSR), Ribber Smocked Sheet and Concentrated Latex The issue of lack of details in the HS classification for some rubber grades should be considered for further researches Moreover, we simply assume that the impact of trade cost on trade for natural rubber bilateral trade have the same direction with the impact of trade cost on trade 52 for agriculture sector If it is possible, further researches should use freight rate data for natural rubber, it is more representative of the actual transportation costs Due to the unavailable data, another limitation of this thesis is the using of car production as a proxy for the natural rubber consumption by tyre and automobile rubber spare part industries Further researches should use the variable of natural rubber consumption, it is more representative for the demand of natural rubber in importing countries 53 REFERENCE Amarasinghe, A A M D (2016) A Study on the Impact of Industrial Production Index (IPI) to Beverage, Food and Tobacco Sector Index with Special Reference to Colombo Stock Exchange Procedia Food Science, 6, 275-278 Anderson, J E., & Van Wincoop, E (2003) Gravity with gravitas: a solution to the border puzzle The American economic review, 93(1), 170-192 Baier, S L., &Bergstrand, J H (2007) Do free trade agreements actually increase members' international trade? Journal of international Economics, 71(1), 72-95 Barlow, C., Jayasuriya, S., & Tan, C S (2014) The world rubber industry Routledge Bergstrand, J H (1985) The gravity equation in international trade: some microeconomic foundations and empirical evidence The review of economics and statistics, 474-481 Beverelli, C., Lanz, R (2015) Workshop for empirical trade analysis http://www.unescap.org/ Blum, B S., & Goldfarb, A (2006) Does the internet defy the law of gravity? Journal of international economics, 70(2), 384-405 Buch, C M., Kleinert, J., &Toubal, F (2004) The distance puzzle: on the interpretation of the distance coefficient in gravity equations Economics Letters, 83(3), 293-298 Byers, J A (1997) Surface distance between two points of latitude and longitude USDA, Agricultural Research Service Retrieved November, 3, 2006 Stephen, B., Anke, H., & Jonathan, T (2001) GMM Estimation of Empirical Growth Models Economics Papers Deardorff, A (1998) Determinants of bilateral trade: does gravity work in a neoclassical world? In The regionalization of the world economy (pp 7-32) University of Chicago Press Dennis, A., & Shepherd, B (2011) Trade facilitation and export diversification The World Economy, 34(1), 101-122 Filippini, C., &Molini, V (2003) The determinants of East Asian trade flows: a gravity equation approach Journal of Asian Economics, 14(5), 695-711 Geraci, V J., &Prewo, W (1977) Bilateral trade flows and transport costs The Review of Economics and Statistics, 67-74 Hassan, M K., Sanchez, B A., & Hussain, M E (2010) Economic Performance of the OIC Countries and the prospect of an Islamic Common Market Journal of Economic Cooperation and Development, 31(2), 65-121 Huang, R R (2007) Distance and trade: Disentangling unfamiliarity effects and transport cost effects European Economic Review, 51(1), 161-181 Khan, S., Haq, I., & Khan, D (2013) An empirical analysis of Pakistan’s bilateral trade: A gravity model approach The Romanian Economic Journal, 16(48), 103-120 54 Khiyavi, P K., Moghaddasi, R., & Yazdani, S (2013) Investigation of Factors Affecting the International Trade of Agricultural Products in Developing Countries Life Science Journal, 10(3s), 409-414 Márquez-Ramos, L (2007) Understanding the determinants of international trade in African countries: An empirical analysis for Ghana and South Africa CESifo Venice Summer Institute Martínez-Zarzoso, I., & Nowak-Lehmann, F D (2004) Economic and geographical distance: Explaining Mercosur Sectoral Exports to the EU Open Economies Review, 15(3), 291-314 Natale, F., Borrello, A., &Motova, A (2015) Analysis of the determinants of international seafood trade using a gravity model Marine Policy, 60, 98-106 Novy, D (2012) International trade without CES: Estimating translog gravity Journal of International Economics, 89(2), 271-282 Park, H M (2011) Practical guides to panel data modeling: a step-by-step analysis using stata Public Management & Policy Analysis Program, International University of Japan Persson, M (2013) Trade facilitation and the extensive margin The Journal of International Trade & Economic Development, 22(5), 658-693 Petersen, M A (2009) Estimating standard errors in finance panel data sets: Comparing approaches Review of financial studies, 22(1), 435-480 Prehn, S., & Brümmer, B (2011) Estimation Issues in Single Commodity Gravity Trade Models In 2011 International Congress, August 30-September 2, 2011, Zurich, Switzerland (No 114776) European Association of Agricultural Economists Prentice, B E., Wang, Z., & Urbina, H J (1998) Derived demand for refrigerated truck transport: a Gravity Model analysis of Canadian pork exports to the United States Canadian Journal of Agricultural Economics/Revue canadienne d'agroeconomie, 46(3), 317-328 Rahman, M M (2003) A panel data analysis of Bangladesh’s trade: the gravity model approach In Proceedings of the 5th Annual Conference of the European Trade Study Group (ETSG2003) European Trade Study Group Roodman, D (2006) How to xtabond2: An introduction to difference and system GMM in Stata Center for Global Development working paper, (103) Sohn, Chan-Hyun, and Jinna Yoon (2001) "Does the gravity model fit Korea’s trade patterns?." Seoul: Korean Institute of Economic Policy Thai, T D (2006) A gravity model for trade between Vietnam and twenty-three European countries Vido, E., & Prentice, B E (2003) The use of proxy variables in economic gravity models: a cautionary note Transportation Quarterly, 57(1), 123-137 Wang, C., Wei, Y., & Liu, X (2010) Determinants of bilateral trade flows in OECD countries: evidence from gravity panel data models The World Economy, 33(7), 894-915 Yusof, M (1988) Malaysian natural rubber market model Pertanika, 11(3), 441-449 55 APPENDIX Regression results Table 4.4: Regression results for world natural rubber trade value OLS model Source SS df MS Model Residual 1383.91956 808.363877 716 172.989944 1.12899983 Total 2192.28343 724 3.02801579 lntij Coef lnyi lnyj lnpi lnpj tcij lndisij ipindex caryj _cons 1.27263 7269312 -.2301906 -.1957704 -.0079796 -.4251263 -.0034093 5315922 -15.21235 Std Err .0579368 0592249 0563439 0482822 0008725 0744648 0027417 0804869 1.828709 t 21.97 12.27 -4.09 -4.05 -9.15 -5.71 -1.24 6.60 -8.32 Number of obs F( 8, 716) Prob > F R-squared Adj R-squared Root MSE P>|t| 0.000 0.000 0.000 0.000 0.000 0.000 0.214 0.000 0.000 = = = = = = 725 153.22 0.0000 0.6313 0.6271 1.0625 [95% Conf Interval] 1.158883 610656 -.3408096 -.290562 -.0096925 -.5713218 -.008792 3735736 -18.80263 1.386376 8432065 -.1195716 -.1009789 -.0062666 -.2789309 0019734 6896108 -11.62208 56 Fixed effects model Fixed-effects (within) regression Group variable: id1 Number of obs Number of groups = = 725 55 R-sq: Obs per group: = avg = max = 13.2 14 within = 0.7294 between = 0.1185 overall = 0.0843 corr(u_i, Xb) F(7,663) Prob > F = -0.9633 lntij Coef lnyi lnyj lnpi lnpj tcij lndisij ipindex caryj _cons 1.452008 614162 -4.561391 4.090452 -.007031 0015502 7555151 -18.51085 sigma_u sigma_e rho 6.0823486 5250505 99260335 Std Err .1331687 1679363 9967504 9788775 0009529 (omitted) 0029841 1322968 21.62913 t P>|t| = = 255.33 0.0000 [95% Conf Interval] 10.90 3.66 -4.58 4.18 -7.38 0.000 0.000 0.000 0.000 0.000 1.190524 2844109 -6.518559 2.168379 -.0089021 1.713491 9439131 -2.604224 6.012525 -.00516 0.52 5.71 -0.86 0.604 0.000 0.392 -.0043092 495744 -60.98071 0074097 1.015286 23.959 (fraction of variance due to u_i) F-test result F test that all u_i=0: F(54, 663) = 44.50 Prob > F = 0.0000 57 Random effects model Random-effects GLS regression Group variable: id1 Number of obs Number of groups = = 725 55 R-sq: Obs per group: = avg = max = 13.2 14 within = 0.7147 between = 0.6010 overall = 0.6162 corr(u_i, X) Wald chi2(8) Prob > chi2 = (assumed) lntij Coef lnyi lnyj lnpi lnpj tcij lndisij ipindex caryj _cons 1.209143 7943479 -.3233388 -.0915576 -.0081867 -.4282928 -.0002664 7935838 -12.43794 0871868 1204891 1838385 1345769 0009106 2221237 002609 1184193 4.367507 sigma_u sigma_e rho 97788782 5250505 77622494 (fraction of variance due to u_i) Std Err z 13.87 6.59 -1.76 -0.68 -8.99 -1.93 -0.10 6.70 -2.85 P>|z| 0.000 0.000 0.079 0.496 0.000 0.054 0.919 0.000 0.004 = = 1744.35 0.0000 [95% Conf Interval] 1.03826 5581936 -.6836556 -.3553235 -.0099715 -.8636472 -.0053798 5614862 -20.9981 1.380026 1.030502 0369779 1722084 -.006402 0070615 0048471 1.025681 -3.877781 LM-test result Breusch and Pagan Lagrangian multiplier test for random effects lntij[id1,t] = Xb + u[id1] + e[id1,t] Estimated results: Var lntij e u Test: sd = sqrt(Var) 3.028016 275678 9562646 1.740119 5250505 9778878 Var(u) = chibar2(01) = Prob > chibar2 = 2258.09 0.0000 58 Hausman-test results Test: Ho: difference in coefficients not systematic chi2(7) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 37.11 Prob>chi2 = 0.0000 (V_b-V_B is not positive definite) Table 4.6: Regression results for world natural rubber trade volume OLS model Source SS df MS Model Residual 865.383669 966.486114 716 108.172959 1.34984094 Total 1831.86978 724 2.53020688 lnkg Coef lnyi lnyj lnpi lnpj tcij lndisij ipindex caryj _cons 8298621 6542796 -.0741529 -.2107101 -.0086454 -.230544 -.0110796 7625461 -2.549852 Std Err .0633504 0647588 0616086 0527936 000954 0814227 0029978 0880075 1.999581 t 13.10 10.10 -1.20 -3.99 -9.06 -2.83 -3.70 8.66 -1.28 Number of obs F( 8, 716) Prob > F R-squared Adj R-squared Root MSE P>|t| 0.000 0.000 0.229 0.000 0.000 0.005 0.000 0.000 0.203 = = = = = = 725 80.14 0.0000 0.4724 0.4665 1.1618 [95% Conf Interval] 7054874 5271397 -.195108 -.3143589 -.0105184 -.3903997 -.0169652 5897625 -6.475595 9542368 7814194 0468022 -.1070614 -.0067724 -.0706882 -.005194 9353297 1.375892 59 Fixed effects model Fixed-effects (within) regression Group variable: id1 Number of obs Number of groups = = 725 55 R-sq: Obs per group: = avg = max = 13.2 14 within = 0.3617 between = 0.0714 overall = 0.0367 corr(u_i, Xb) F(7,663) Prob > F = -0.9775 lnkg Coef lnyi lnyj lnpi lnpj tcij lndisij ipindex caryj _cons 5209237 1789096 -4.673779 5.222755 -.0071568 0029237 8208221 -.7449937 sigma_u sigma_e rho 7.2432259 51957831 99488072 Std Err .1317808 166186 986362 9686754 000943 (omitted) 002953 1309179 21.40371 t P>|t| = = 53.68 0.0000 [95% Conf Interval] 3.95 1.08 -4.74 5.39 -7.59 0.000 0.282 0.000 0.000 0.000 2621658 -.1474047 -6.610549 3.320714 -.0090084 7796817 505224 -2.737009 7.124797 -.0053053 0.99 6.27 -0.03 0.322 0.000 0.972 -.0028746 5637584 -42.77222 0087221 1.077886 41.28223 (fraction of variance due to u_i) F-test result F test that all u_i=0: F(54, 663) = 54.76 Prob > F = 0.0000 60 Random effects model Random-effects GLS regression Group variable: id1 Number of obs Number of groups = = 725 55 R-sq: Obs per group: = avg = max = 13.2 14 within = 0.3135 between = 0.4352 overall = 0.3845 corr(u_i, X) Wald chi2(8) Prob > chi2 = (assumed) lnkg Coef lnyi lnyj lnpi lnpj tcij lndisij ipindex caryj _cons 3185603 5074335 -.0531569 0558642 -.0084333 -.1004278 -.0017974 9777399 10.28383 0887908 1229449 1899165 138767 000923 2294222 0026517 1204016 4.50755 sigma_u sigma_e rho 98939159 51957831 7838331 (fraction of variance due to u_i) Std Err z 3.59 4.13 -0.28 0.40 -9.14 -0.44 -0.68 8.12 2.28 P>|z| 0.000 0.000 0.780 0.687 0.000 0.662 0.498 0.000 0.023 = = 346.89 0.0000 [95% Conf Interval] 1445335 266466 -.4253864 -.2161142 -.0102424 -.5500871 -.0069946 7417571 1.449199 4925871 7484011 3190727 3278426 -.0066242 3492315 0033999 1.213723 19.11847 61 LM-Test Breusch and Pagan Lagrangian multiplier test for random effects lnkg[id1,t] = Xb + u[id1] + e[id1,t] Estimated results: Var lnkg e u Test: sd = sqrt(Var) 2.530207 2699616 9788957 1.590662 5195783 9893916 Var(u) = chibar2(01) = Prob > chibar2 = 2096.33 0.0000 Hausman-test result Test: Ho: difference in coefficients not systematic chi2(7) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 47.08 Prob>chi2 = 0.0000 (V_b-V_B is not positive definite) 62 GMM model for trade value Dynamic panel-data estimation, two-step system GMM Group variable: id1 Time variable : year Number of instruments = 52 F(9, 54) = 43.89 Prob > F = 0.000 lntij Coef lnyi lnyj lnpi lnpj lndisij tcij ipindex caryj year _cons 1.496911 590728 -.3377503 -.1251367 -.520909 -.0078305 0027899 4627677 -.0373875 57.79291 Number of obs Number of groups Obs per group: avg max Corrected Std Err .2230807 2703586 1970562 1687007 3196323 0025163 0067108 2412122 0392001 75.57982 t 6.71 2.18 -1.71 -0.74 -1.63 -3.11 0.42 1.92 -0.95 0.76 P>|t| 0.000 0.033 0.092 0.461 0.109 0.003 0.679 0.060 0.344 0.448 = = = = = 725 55 13.18 14 [95% Conf Interval] 1.049661 0486916 -.7328242 -.4633612 -1.161733 -.0128753 -.0106643 -.0208336 -.115979 -93.73552 1.944161 1.132764 0573236 2130879 1199151 -.0027857 0162442 9463689 0412039 209.3213 Instruments for first differences equation Standard D.(lnpi lnpj lndisij tcij ipindex caryj year) GMM-type (missing=0, separate instruments for each period unless collapsed) L2.(L.lnyi L.lnyj) Instruments for levels equation Standard lnpi lnpj lndisij tcij ipindex caryj year _cons GMM-type (missing=0, separate instruments for each period unless collapsed) DL.(L.lnyi L.lnyj) Arellano-Bond test for AR(1) in first differences: z = Arellano-Bond test for AR(2) in first differences: z = Arellano-Bond test for AR(3) in first differences: z = Sargan test of (Not robust, Hansen test of (Robust, but overid restrictions: chi2(42) = 624.30 but not weakened by many instruments.) overid restrictions: chi2(42) = 50.93 weakened by many instruments.) -3.62 -1.97 0.53 Pr > z = Pr > z = Pr > z = 0.000 0.048 0.597 Prob > chi2 = 0.000 Prob > chi2 = 0.163 63 GMM model for trade volume Group variable: id1 Time variable : year Number of instruments = 56 F(9, 54) = 14.25 Prob > F = 0.000 lnkg Coef lnyi lnyj lnpi lnpj lndisij tcij ipindex caryj year _cons 1.233726 4302558 -.1590619 -.0652148 -.2887382 -.008988 0030809 4144135 -.1475906 282.8428 Number of obs Number of groups Obs per group: avg max Corrected Std Err .2404988 2452826 2043102 2013203 3775115 0026418 0054007 2257387 0408952 79.54464 t 5.13 1.75 -0.78 -0.32 -0.76 -3.40 0.57 1.84 -3.61 3.56 P>|t| 0.000 0.085 0.440 0.747 0.448 0.001 0.571 0.072 0.001 0.001 = = = = = 725 55 13.18 14 [95% Conf Interval] 7515552 -.0615062 -.5686792 -.4688376 -1.045603 -.0142845 -.0077468 -.0381653 -.2295805 123.3654 1.715897 9220178 2505553 3384081 4681267 -.0036916 0139086 8669923 -.0656006 442.3202 Instruments for first differences equation Standard D.(lnpi lnpj lndisij tcij ipindex caryj year) GMM-type (missing=0, separate instruments for each period unless collapsed) L2.(lnyi lnyj) Instruments for levels equation Standard lnpi lnpj lndisij tcij ipindex caryj year _cons GMM-type (missing=0, separate instruments for each period unless collapsed) DL.(lnyi lnyj) Arellano-Bond test for AR(1) in first differences: z = Arellano-Bond test for AR(2) in first differences: z = Arellano-Bond test for AR(3) in first differences: z = Sargan test of (Not robust, Hansen test of (Robust, but overid restrictions: chi2(46) = 885.20 but not weakened by many instruments.) overid restrictions: chi2(46) = 50.74 weakened by many instruments.) -2.61 0.47 0.84 Pr > z = Pr > z = Pr > z = 0.009 0.638 0.402 Prob > chi2 = 0.000 Prob > chi2 = 0.292 64 ... bilateral natural rubber traded can be treated as alternative methods to explain the bilateral trade in gravity models even for disaggregated trade The research findings also reveal that world s... international trade using gravity model and natural rubber trade 24 Table 3.1: Data source of variables and their expected sign 32 Table 4.1: Bilateral natural rubber trade in... fundamental idea behind the comparative advantage It is no doubt that international trade is fundamentally explained via comparative advantage However the comparative advantage based model was still