Impact of asean-plus-one ftas: A gravity approach

Nội dung text: Impact of asean-plus-one ftas: A gravity approach

1. IMPACT OF ASEAN-PLUS-ONE FTAS: A GRAVITY APPROACH * Phuong Bui Thi Hang 1 ABSTRACT: This study examines the trade effect of ASEAN-plus-one free trade agreements with 6 partners, including Japan, China, Korea, India, Australia and New Zealand. The gravity model has been used with the inclusion of country-pair effect and country-and-time effect to prove significant impacts of ASEAN-plus-one FTAs on trade of members and trade of the outsiders. Statistical evidence has been found that AIFTA has the highest trade effect among FTAs. AJFTA and ACFTA had pure trade creation in terms of exports for all members, whereas AKFTA and AANZFTA had pure trade creation in terms of imports. Keywords: FTA, ASEAN partner, trade impact 1. INTRODUCTION When international trade increasingly emerges, there is a trend to reduce trade barriers and open the economy to the world. Most countries seek partnership within their neighbourhood or with crucial trading partners all over the world. As a result, there has been a wide variety of free trade agreements (FTAs) signed so far. As recorded by WTO Secretariat, FTAs have become prevalent for the last decade, especially since 2009. It is a general belief that FTAs will promote trade among contracting parties thanks to trade creation effect. However, that belief is not true all the time. Some research have approved the opposite. The increasing number of FTAs does not ensure the increase in intra-trade. Therefore it is essential to assess the trade effect, including trade creation and trade diversion, of FTAs after they came into force. (Source: ASEAN Secretariat) Figure 1. ASEAN’s top trading partners * Banking Academy, 12 Chua Boc, Hanoi, 100000, Vietnam, Corresponding author. Tel.: +84000000000.E-mail address: aaa@edu.vn
2. 514 HỘI THẢO KHOA HỌC QUỐC TẾ KHỞI NGHIỆP ĐỔI MỚI SÁNG TẠO QUỐC GIA Within Asian area, Association of Southeast Asian Nations (ASEAN) has played an active role to form FTAs. Established in 1967, ASEAN originally included 5 countries (Indonesia, Malaysia, Philippines, Singapore and Thailand) and had Brunei to join later. Together 6 members formed ASEAN FTA (AFTA) in 1992 and extended the membership to 4 other countries in the area. Since then, ASEAN continuously formed FTAs with its major partners. Up to now, it has successfully cooperated with 6 dialogue partners by 5 FTAs : ASEAN-China FTA (ACFTA) took effect in 2005, ASEAN-Korea FTA (AKFTA) in 2007, ASEAN- Japan FTA (AJFTA) in 2009, ASEAN-Australia-New Zealand FTA (AANZFTA) in 2010 and ASEAN-India (AIFTA) in 2010. These 6 partners have been among list of top trading partners of ASEAN for years. As shown in Table 1, the tariff rates under AANZFTA were reduced the most, with the average of 95.7% has committed to be removed. ACFTA ranks second with 94.7%, followed by AKFTA with 94.5% and AJFTA with 92.8%. AIFTA is far behind with only 79.6%. Given such different tariff elimination ratios, ASEAN-plus-one FTAs will surely have different trade impacts. Table 1. Tariff Elimination Ratios in ASEAN+1FTA ASEAN + ASEAN member China Korea Japan Australia-New Zealand India Average Brunei 98.3 99.2 97.7 99.2 85.3 95.9 Cambodia 89.9 97.1 85.7 89.1 88.4 90 Indonesia 92.3 91.2 91.2 93.7 48.7 83.4 Lao 97.6 90 86.9 91.9 80.1 89.3 Malaysia 93.4 95.5 94.1 97.4 79.8 92 Myanmar 94.5 92.2 85.2 88.1 76.6 87.3 Philippines 93 99 97.4 95.1 80.9 93.1 Singapore 100 100 100 100 100 100 Thailand 93.5 95.6 96.8 98.9 78.1 92.6 Vietnam N/A 89.4 94.4 94.8 79.5 89.5 94.7 94.9 92.9 94.8 79.7 China 94.1 Korea 90.5 Japan 91.9 Australia 100 New Zealand 100 India 78.8 Averages 94.7 94.5 92.8 95.7 79.6 (Source: Kuno (2012)) Although such partners’ crucial roles in trading relationship with ASEAN are indisputable, it is necessary to assess the trade impacts, including trade creation and trade diversion, of these FTAs to member countries after many years into force. This study is to examine whether each individual ASEAN FTAs led to trade creation or trade diversion to members and which one had the biggest effect. 2. LITERATURE REVIEW There has been a wide range of empirical research in assessment on trade effects of FTA. As a pioneering work, Viner (1950) laid the foundation for concept of “trade creation” and “trade diversion”. “Trade creation” was defined as the reduction in production of a less efficient country which was replaced by imports from a more efficient country in the partnership. “Trade diversion” was posed when imports
3. INTERNATIONAL CONFERENCE STARTUP AND INNOVATION NATION 515 previously coming from an outsider was replaced by imports from a partner. In order to measure these trade effects, gravity model was introduced and developed by Tinbergen (1962) and has become the formal method with many extensions afterwards. The gravity model measures bilateral trade between two countries based on the economy size and distance of two countries. Economy size can be measured by the gross domestic product (GDP) and the population of the country. The formula (1) is as follows: (1) Where is trade flow from country I to country j, and are GDP of country I and country j, respectively, is the distance between two countries. Many studies have applied gravity model to analyse the trade impacts by taking natural logarithm of the equation (1) and add some variables capturing the common effects of two countries such as common border, common official language and FTAs. So the equation (2) is as follows: (2) Where is natural log of exported value from exporter i to importer j at time t (in thousand USD); and are the gross domestic product of exporter i and importer j at time t, respectively; and are the populations of exporter i and importer j at time t, respectively; is the distance from the capital city of exporter i to the capital city of importer j and is fixed over time; , , , and are all dummy variables: equals to 1 if exporter i and importer j have the same official language and equals to 0 otherwise, equals to 1 if exporter i and importer j share the common border and equals to 0 otherwise, is the error term. For each ASEAN-plus-one FTA, we put 3 dummy variables into the regression, including equals to 1 if both exporter i and importer j are in that particular FTAs after that FTA came into force and equals to 0 otherwise, equals to 1 if exporter i is in that FTA and importer j is not and equals to 0 otherwise, equals to 1 if importer j is in that FTA and exporter i is not and equals to 0 otherwise. So for example, in order to examine the trade effects of AJFTA which came into force at the beginning of 2009, we include equals to 1 if both exporter i and importer j are in AJFTAs since 2009, equals to 1 if exporter i is in AJFTA and importer j is not, equals to 1 if importer j is in that FTA and exporter i is not. Similarly, we take 2005 as a threshold for ACFTA, 2007 for AKFTA, 2010 for both AIFTA and AANZFTA. It is expected that GDP and Population will have a positive relationship with exported value as they indicate the economy size. The larger economy size is, the more demand for goods is. So are the exports. While distance is expected to decrease exports (as it increases trade costs such as transportation cost and delivery time), common language and border should have positive sign because they promote trade. Anderson et al (2003) found that estimations based on standard gravity model suffer from omitted variable bias. They developed a method to estimate a theoretical gravity equation efficiently by introducing multilateral trade resistance factors. Multilateral trade resistance includes trade barriers that each individual
4. 516 HỘI THẢO KHOA HỌC QUỐC TẾ KHỞI NGHIỆP ĐỔI MỚI SÁNG TẠO QUỐC GIA country faces when trading with all of its partners. To solve the problem of multilateral trade resistance, they suggested to use country-specific dummies. Baldwin and Taglioni (2006) used country-specific effects to capture all the time-invariant individual effects of exporters and importers that are omitted from the model rest of the model specifications, such as preferences and institutional differences. Moreover, Kepaptsoglou et al (2010) pointed out that the fixed effect model (FEM) is dominant among empirical research using gravity model in comparison with the random effect model (REM). In order to choose which econometric model should be applied, Hausman test will be conducted. It is noted that FEM cannot measure the impact of such time-invariant variables as distance, language and adjacency. So such variables are taken out of our estimation FEM model. Then the equation (3) is as follows: (3) Where is the country-pair effect and is the time effects that control for omitted variables which are in-variant for all trade flows but vary over time. Beir et al (2007) argued that omitted variables are the main cause for endogeneity bias in estimating FTA effects with cross-sectional data. So the paper suggested to use panel data and fixed effects to deal with the bias. However, the time-invarient fixed effects are insufficient to capture the unobservable factors such as time-varying multilateral resistance terms. Therefore they used country-and-time effects in addition to country-pair fixed effects to obtain unbiased estimates. Martinez et al (2009) also introduced individual country dummies in cross-sectional data estimation and bilateral fixed effects as well as country-and-time effects in panel data estimation to eliminate the endogeneity bias. The equation (4) is as follows: (4) Where is the importer-and-time effect and is the exporter-and-time effect. Many researches have indicated that trade creation can be offset against trade diversion because trade within FTA members (intra-bloc trade) grows and trade outside the FTA (extra-bloc trade) falls at the same time. In other words, an increase in intra-bloc exports is paired with a decrease in imports from extra-bloc. As a result, to capture net trade effects, it is necessary to include variables to examine trade among members and non-members in addition to variables capturing intra-bloc trade. Following Sloaga-Winters (2001), Carrere (2006), DeRosa (2007), Martinez (2009) and many other study, three dummy variables , and are added into all our estimation models. According to Martinez (2009), the scenarios of trade effects of FTAs are expressed in Table 2.
5. INTERNATIONAL CONFERENCE STARTUP AND INNOVATION NATION 517 Table 2. Scenarios of Trade Creation and Trade Diversion effects of FTAs Intra-bloc Extra-bloc trade trade Magnitude Trade effects ( ) ( ) > 0 Pure trade creation in terms of export > 0 > Trade creation + Export diversion Export 0 Export expansion of extra-bloc trade 0 Pure trade creation in terms of import > 0 > Trade creation + Import diversion Import 0 Import expansion of extra-bloc trade < 0 Import diversion + import contraction of < 0 intra-bloc (Source: Martinez et al, 2009) Given the three dummy variables, both intra-bloc and extra bloc trade are captured. According to Sloaga-Winters (2001), the trade effect is expressed by coefficients of those variables: (5) Where is the change in trade between exporter i and importer j in percentage, e is the natural number e. 3. DATA AND METHODOLOGY This study uses a panel dataset covering 26 countries during the period of 20 years from 1997 to 2016 at both aggregated and disaggregated levels of products. The 26 trading partners include 10 ASEAN countries, 6 partners which have signed free trade agreements with ASEAN and 10 other top trading partners of ASEAN in 2016 (Table 3). GDP (at current USD) and population data are derived from World Bank Development Indicator Database. Since Myanmar’s GDP is only available from 2000, the missing data is taken from World Economic Outlook Data (WEO, IMF). Distance between capital cities, language and adjacency dummy are derived from CEPII database. Exported values are collected from UN Comtrade at total trade.
6. 518 HỘI THẢO KHOA HỌC QUỐC TẾ KHỞI NGHIỆP ĐỔI MỚI SÁNG TẠO QUỐC GIA Table 3. List of Trading partners under estimation 10- ASEAN countries 6- FTA partners 10- Top trading partners Brunei Australia Belgium Indonesia China Canada Cambodia India France Lao Japan Germany Myanmar Korea Hong Kong Malaysia New Zealand Mexico Philippines Netherland Singapore UAE Thailand United Kingdom Vietnam United States of America Table 4 shows statistical summary of all variables included in our models. As in Table 4, the minimum value of exports is 0 which raises the issue of zero trade. However, zero trade is not a critical issue in our study. Only 13 out of 12063 observations, accounting for 0.11% su ered from zero values so it cannot statistically affect our results. ff Table 4. Summary statistics Number of Variable Mean Std. Dev. Min Max Observations Exported value 12063 8495114 2.61E+07 0 4.10E+08 GDPi 12063 1.66E+12 3.00E+12 3.65E+09 1.86E+13 GDPj 12063 1.57E+12 2.95E+12 1.28E+09 1.86E+13 POPi 12063 1.69E+08 3.38E+08 312038 1.38E+09 POPj 12063 1.54E+08 3.24E+08 312038 1.38E+09 Distance 12063 7611.521 4790.6 315.5433 19263.88 Language 12063 0.077676 0.267671 0 1 Adjacency 12063 0.062091 0.24133 0 1 AJFTA_1 12063 0.070795 0.256493 0 1 AJFTA_2 12063 0.105198 0.306821 0 1 AJFTA_3 12063 0.107270 0.309469 0 1 ACFTA_1 12063 0.099561 0.299426 0 1 ACFTA_2 12063 0.148388 0.355499 0 1 ACFTA_3 12063 0.161734 0.368222 0 1 AKFTA_1 12063 0.085468 0.279588 0 1 AKFTA_2 12063 0.127414 0.33345 0 1 AKFTA_3 12063 0.134461 0.341161 0 1 AIFTA_1 12063 0.063334 0.243573 0 1 AIFTA_2 12063 0.094172 0.29208 0 1 AIFTA_3 12063 0.093758 0.291504 0 1 AANZFTA_1 12063 0.076018 0.265037 0 1 AANZFTA_2 12063 0.095996 0.294598 0 1 AANZFTA_3 12063 0.095333 0.293686 0 1 (Source: Author’s calculation)
7. INTERNATIONAL CONFERENCE STARTUP AND INNOVATION NATION 519 4. MAIN RESULTS After applying the specified gravity model and running the regressions, we have the main results for aggregated trade expressed in Table A1,A2,A3,A4,A5 (Annex) for AJFTA, ACFTA, AKFTA, AIFTA and AANZFTA respectively. Each table has 4 columns. Column (1) presents the results of pooled OLS method using equation (1). Hausman tests have been conducted to choose between FEM and REM and in all cases. As expected, chi2 are all large so we reject the null hypothesis which is difference in coefficients is not systematic. In other words, the FEM is more suitable to estimate than the REM. So column (2) presents the results with FEM. Then we conduct the FEM with time effect and country-pair effect as in Column (3). Finally the FEM estimation with country-pair and country-and-time effect is in Column (4). The results indicate that GDP of both exporter i and importer j significantly affect exports: the larger GDPs, the more exports. Population of importer j has negative relationship with exports and is statistically significant for all cases. It is because larger population means larger domestic market, richer resource endowment and more diversified products, which leads to little dependence on international trade. On the contrary, coefficient of population of importerj is ambiguous. As expected, distance between two countries discourages trade while common language and shared border help promote trade significantly. It is quite surprising that most of 3 coefficients of FTA_1, FTA_2 and FTA_3 in column (1), (2), (3) of Table A1,A2,A3,A4,A5 are all negative, which means export (import) diversion and contraction of intra-FTA and there is no trade creation after AJFTA. As discussed earlier, estimations in column (1), (2), (3) are more likely to be endogeneity-biased due to omitted variables. To fix the problem of endogeneity, the country-pair and country-and-time effect is introduced in column (4). The inclusion of such effects in estimation column (4) control for all determinants which vary with each individual exporter and individual importer in each year and also control for constant-over-time effect of two countries, thus overcoming the problem of endogeneity. The trade impact of 5 ASEAN-plus-one FTAs is summarized in Table 5. Table 5. The trade impact of 5 ASEAN-plus-one FTAs AJFTA ACFTA AKFTA AIFTA AANZFTA (1) (2) (3) (4) (5) FTA_1 1.069 0.453 0.506 1.279 0.533 (0.224) (0.205) (0.205) (0.186) (0.202) FTA_2 1.148 0.645 0.160 0.821 0.0573 (0.216) (0.160) (0.131) (0.115) (0.180) FTA_3 0.0773 -0.0424 0.537 0.552 0.646 (0.189) (0.135) (0.152) (0.150) (0.152) Robust standard errors in parentheses (Source: Author’s calculation) As in Table 5, we can see that trade impact of AJFTA and ACFTA are quite similar: coefficients of FTA_1 and FTA_2 are both positive and statistically significant while that of FTA_3 is not significant. It implies that FTAs between Japan - ASEAN and China – ASEAN have affected intra-bloc and extra-bloc trade positively. The FTAs has created pure trade creation in terms of exports. The net effect of two significant variables of AJFTA and ACFTA are 2.22 and 1.1, respectively. According to equation (5), AJFTA members exported 8.2% more than normal level of trade predicted by the gravity model whereas ACFTA members exported 2% more than normal. This estimated result also proves that AJFTA has stronger trade impact than ACFTA as it brings more trade creation. However, these trade effects are not very high. This is probably due to the active role of Japan and China in forming bilateral partnership with each ASEAN partners. In fact, Japan signed Economic
8. 520 HỘI THẢO KHOA HỌC QUỐC TẾ KHỞI NGHIỆP ĐỔI MỚI SÁNG TẠO QUỐC GIA Partnership Agreements (EPA) with 7 individual ASEAN countries before AJFTA took effect, including Singapore (JSEPA) in Jan 2002, Malaysia (MJEPA) in Dec 2005, Philippines (JPEPA) in Sep 2006, Thailand (JTEPA) in Apr 2007, Brunei (BJEPA) in June 2007, Indonesia (JIEPA) in July 2008, Vietnam (JVEPA) in Dec 2008. Such agreements weakened the trade effects of ASEAN-plus-one FTAs. It is interesting that AKFTA and AANZFTA both create the same effect in terms of import trade creation. The coefficients of FTA_1 and FTA_3 are both positive and statistically significant, implying that AKFTA and AANZFTA promote trade within the bloc and induce the imports from outside-bloc partners. The results indicate that AKFTA and AANZFTA members exported 1.84% and 2.25% more than normal, respectively. Out of 5 FTAs, only AIFTA has 3 coefficients of FTA_1, FTA_2 and FTA_3 significantly positive. Although the tariff elimination ratios of AIFTA are the lowest, it creates the most trade. It is interesting that AIFTA has pure trade creation effect in terms of both exports and imports. AIFTA created 13.18% more than normal trade, letting other ASEAN-plus-one FTAs far behind. 5. CONCLUSION The study proved that the bias problem of endogeneity can be fixed with the inclusion of country-pair effect and country-and-time effect. It found statistical evidence that ASEAN-plus-one FTAs had significant impacts on trade of members, even though the effects are quite small. Among 5 FTAs, AIFTA has the highest trade effect due to India and ASEAN members do not have other bilateral or regional agreements. AJFTA and ACFTA had pure trade creation in terms of exports for all members, whereas AKFTA and AANZFTA had pure trade creation in terms of imports. The results from this gravity approach suggest that we should focus more on and materialize the trade opportunities with AIFTA. At the same time, AKFTA and AANZFTA should be examined to understand whether they only promote imports from other outsiders and how to induce exports of FTA members. Appendix Appendix A1. Gravity estimation for aggregated trade for AJFTA (1) (2) (3) (4) VARIABLES lnX lnX lnX lnX lnGDPi 0.917 0.828 0.971 (0.0400) (0.0234) (0.0602) lnGDPj 1.025 0.715 0.839 (0.0396) (0.0248) (0.0665) lnPOPi 0.0730 -0.0111 -0.0162 (0.0463) (0.0489) (0.0764) lnPOPj -0.106 -0.996 -0.920 (0.0445) (0.174) (0.456) lnDist -1.084 (0.0985) lang 0.218 (0.156) adj 0.647 (0.327) ajfta1 -0.320 -0.195 -0.236 1.069 (0.214) (0.0370) (0.0875) (0.224) ajfta2 -0.278 -0.0644 -0.0492 1.148 (0.124) (0.0303) (0.0783) (0.216)
9. INTERNATIONAL CONFERENCE STARTUP AND INNOVATION NATION 521 ajfta3 -0.577 -0.0438 -0.0416 0.0773 (0.0920) (0.0297) (0.0623) (0.189) Constant -28.51 -10.27 -18.45 12.66 (1.626) (2.688) (7.688) (0.144) Observations 12,050 12,050 12,050 12,050 R-squared 0.720 0.487 0.493 0.643 ols YES fe YES fe,ij YES fe,ij,it,jt YES Robust standard errors in parentheses, p<0.01, p<0.05, * p<0.1 (Source: Author’s calculation) Appendix A2. Gravity estimation for aggregated trade for ACFTA (1) (2) (3) (4) VARIABLES lnX lnX lnX lnX lnGDPi 0.948 0.793 0.959 (0.0397) (0.0265) (0.0665) lnGDPj 0.993 0.753 0.894 (0.0382) (0.0264) (0.0692) lnPOPi 0.0568 -0.0286 -0.0198 (0.0471) (0.0488) (0.0861) lnPOPj -0.0805* -0.959 -0.701 (0.0433) (0.173) (0.462) lnDist -1.072 (0.0954) lang 0.281* (0.155) adj 0.631* (0.323) acfta1 -0.0272 -0.175 -0.271 0.453 (0.207) (0.0406) (0.0831) (0.205) acfta2 0.124 0.0155 0.0155 0.645 (0.125) (0.0343) (0.0837) (0.160) acfta3 -0.465 -0.125 -0.141* -0.0424 (0.1000) (0.0324) (0.0738) (0.135) Constant -28.81 -10.66 -23.28 12.84 (1.578) (2.637) (7.907) (0.0972) Observations 12,050 12,050 12,050 12,050 R-squared 0.719 0.487 0.493 0.642 ols YES fe YES fe,ij YES fe,ij,it,jt YES Robust standard errors in parentheses, p<0.01, p<0.05, * p<0.1 (Source: Author’s calculation)
10. 522 HỘI THẢO KHOA HỌC QUỐC TẾ KHỞI NGHIỆP ĐỔI MỚI SÁNG TẠO QUỐC GIA Appendix A3. Gravity estimation for aggregated trade for AKFTA (1) (2) (3) (4) VARIABLES lnX lnX lnX lnX lnGDPi 0.937 0.830 0.973 (0.0403) (0.0246) (0.0620) lnGDPj 1.009 0.731 0.852 (0.0391) (0.0254) (0.0671) lnPOPi 0.0676 -0.0153 -0.0141 (0.0463) (0.0489) (0.0771) lnPOPj -0.0981 -0.979 -0.812* (0.0441) (0.175) (0.467) lnDist -1.065 (0.0962) lang 0.259* (0.157) adj 0.638 (0.321) akfta1 -0.0258 -0.225 -0.254 0.506 (0.212) (0.0386) (0.0859) (0.205) akfta2 -0.0262 -0.0556* -0.0168 0.160 (0.128) (0.0318) (0.0831) (0.131) akfta3 -0.540 -0.0795 -0.0561 0.537 (0.0953) (0.0311) (0.0686) (0.152) Constant -28.87 -10.94 -20.76 12.85 (1.615) (2.687) (8.042) (0.0897) Observations 12,050 12,050 12,050 12,050 R-squared 0.719 0.488 0.493 0.643 ols YES fe YES fe,ij YES fe,ij,it,jt YES Robust standard errors in parentheses, p<0.01, p<0.05, * p<0.1 (Source: Author’s calculation) Appendix A4. Gravity estimation for aggregated trade for AIFTA (1) (2) (3) (4) VARIABLES lnX lnX lnX lnX lnGDPi 0.899 0.843 0.986 (0.0402) (0.0255) (0.0663) lnGDPj 1.027 0.673 0.801 (0.0393) (0.0255) (0.0681) lnPOPi 0.0833* -0.0174 -0.0157 (0.0466) (0.0491) (0.0732) lnPOPj -0.102 -1.274 -1.145 (0.0440) (0.179) (0.464) lnDist -1.101 (0.0988) lang 0.246 (0.151) adj 0.660
11. INTERNATIONAL CONFERENCE STARTUP AND INNOVATION NATION 523 (0.327) aifta1 -0.550 0.000212 -0.102 1.279 (0.216) (0.0387) (0.108) (0.186) aifta2 -0.364 -0.0476 -0.0669 0.821 (0.125) (0.0324) (0.0782) (0.115) aifta3 -0.450 0.0837 0.0493 0.552 (0.0863) (0.0323) (0.0668) (0.150) Constant -28.20 -4.578* -13.98* 13.00 (1.609) (2.750) (7.887) (0.116) Observations 12,050 12,050 12,050 12,050 R-squared 0.720 0.486 0.492 0.642 ols YES fe YES fe,ij YES fe,ij,it,jt YES Robust standard errors in parentheses, p<0.01, p<0.05, * p<0.1 (Source: Author’s calculation) Appendix A5. Gravity estimation for aggregated trade for AANZFTA (1) (2) (3) (4) VARIABLES lnX lnX lnX lnX lnGDPi 0.914 0.881 1.074 (0.0402) (0.0250) (0.0630) lnGDPj 1.032 0.682 0.851 (0.0403) (0.0251) (0.0680) lnPOPi 0.0746 0.00777 0.0160 (0.0461) (0.0489) (0.0466) lnPOPj -0.113 -1.174 -0.810* (0.0451) (0.180) (0.475) lnDist -1.066 (0.0978) lang 0.255* (0.154) adj 0.663 (0.324) anzfta1 -0.207 -0.218 -0.421 0.533 (0.194) (0.0358) (0.0920) (0.202) anzfta2 -0.324 -0.127 -0.212 0.0573 (0.125) (0.0314) (0.0771) (0.180) anzfta3 -0.535 0.0691 -0.0493 0.646 (0.0927) (0.0317) (0.0696) (0.152) Constant -28.73 -8.004 -23.96 12.41 (1.649) (2.775) (8.204) (0.210) Observations 12,050 12,050 12,050 12,050 R-squared 0.719 0.489 0.496 0.643 ols YES fe YES fe,ij YES fe,ij,it,jt YES Robust standard errors in parentheses, p<0.01, p<0.05, * p<0.1 (Source: Author’s calculation)
12. 524 HỘI THẢO KHOA HỌC QUỐC TẾ KHỞI NGHIỆP ĐỔI MỚI SÁNG TẠO QUỐC GIA REFERENCES Book: Krugman, Paul R, Maurice Obstfeld and Marc J. Melitz. (2012). International Economics: Theory and Policy (6th ed.). Boston: Pearson Addison-Wesley. Journals: Anderson, J. E., & Van Wincoop, E. (2003), Gravity with gravitas: A solution to the border puzzle, American Economic Review, 93(1), 170–192 ASEAN Secretariat (2016), ASEAN Economic Community Chartbook 2016 Baier, S. L., & Bergstrand, J. H. (2007), Do free trade agreements actually increase members’ international trade?, Journal of International Economics, 71, 72–95 Balassa, B. (1967), Trade creation and trade diversion in the European common market, Economic Journal, 77, 1–21 Baldwin, R., & Taglioni, D. (2006), Gravity for dummies and dummies for gravity equations, Working paper 12516, Cambridge, National Bureau of Economic Research Carrère, C. (2006), Revisiting the effects of regional trade agreements on trade flows with proper specification of the gravity model, European Economic Review, 50(2), 223–247 DeRosa D.A. (2007), The Trade Effects of Preferential Arrangements: New Evidence from the Australia Productivity Commission, IIE Working Paper No. 07-01 Endoh, M. (1999), Trade creation and trade diversion in the EEC, the LAFTA and the CMEA: 1960–1994, Applied Economics, 31, 207–216 Hiroyuki Taguchi (2015) Trade creation and diversion effects of ASEAN-plus-one free trade agreements, Economics Bulletin, Volume 35, Issue 3, 1856-1866 Kepaptsoglou, K., Karlaftis, M. K., & Tsamboulas, D. (2010), The gravity model specification for modelling international trade flows and free trade agreement effects: A 10-year review of empirical studies, The Open Economics Journal, 3, 1–13. Kuno, Arata (2012), Revealing Tariff Structures under the ASEAN+1 FTAs and ATIGA: Implications for a Possible ASEAN++ FTA, in Chang Jae Lee, Shujiro Urata and Ikumo Isono (eds.) Comprehensive Mapping of FTAs in ASEAN and East Asia Phase II. ERIA Martínez-Zarzoso, I., Nowak-Lehmann, D. F., & Horsewood, N. (2009), Are regional trading agreements beneficial? Static and dynamic panel gravity models, North American Journal of Economics and Finance, 20(1), 46–65 Soloaga, I., & Winters, A. (2001), Regionalism in the nineties: What effects on trade?, North American Journal of Economics and Finance, 12(1), 1–29 Tinbergen, J. (1962), Shaping the World Economy: Suggestions for an International Economic Policy, The Twentieth Century Fund, New York. Viner, J. (1950), Customs union theory, New York: Carnegie Endowment for International Peace Yang, S. & Maztinez, I. (2014), A panel data analysis of trade creation and trade diversion effects: The case of ASEAN–China Free Trade Area, China Economic Review, 29, 138-151