8 Is East Asia a natural trade bloc? The trade complementarity index, the intensity index, and the bias index

Transcription

1 8 Is East Asia a natural trade bloc? The trade complementarity index, the intensity index, and the bias index East Asian countries have become interdependent and economically integrated because of the growth of trade especially intra- industry trade and the development of the trade FDI nexus since the mid- 1980s. The increasing intra- regional trade and the proliferation of the trading blocs in East Asia after the late 1990s have also led to several interesting studies on the prospects of economic integration, which will be addressed in Chapter 9. But the drive for regional trading arrangements and economic integration is not motivated by economic rationales alone. As they coalesce, trade blocs must consider the relative competitiveness and complementarity among the prospective member countries. This chapter bridges the gap by examining whether East Asia is a natural trade bloc by empirically exploiting their respective trade complementarities. Though there are many, well- documented studies on the groupings of natural trading blocs in the Asia-Pacific region, 1 research on the underlying forces of the trade complementarities with their respective trading partners is sparse. In theory, under free trade, the rapid growth of exports from developing Asian countries to the world market must be based on their comparative advantages. 2 But whether their comparative advantages in the world market are complementary to or competitive with one another is yet to be determined empirically. If, as Razmi and Blecker (2008) argue, a larger number of developing countries have been competing with one another, rather than with OECD nations, in export markets, then any emerging trade bloc in a region would need to take into consideration the degree of complementarities and/or competitiveness among the members. In addition to the studies on the correlation coefficients of the revealed comparative advantage analyzed in Chapters 2 and 3, it is necessary to analyze the inter- country trade complementarities. Intuitively, under a free trade regime without policy distortions, a pair of countries which trade more intensively with each other than with the 300

2 Is East Asia a natural trade bloc? 301 rest of the world must be complementary with each other. Which pair of countries trades intensively in East and South Asia? What are the factors underlying the growth of trade intensity between countries in East Asia? To what extent are those factors determining the trade intensity increasing or decreasing over time? These interesting questions must be rigorously examined before one can argue for any economic rationales for the emerging trading blocs and/or economic integration in the region. Therefore, trade complementarity must be addressed before any natural trading partners can be identified in the region (Krugman, 1991). To study the fundamental market forces for the creation of any trade bloc, it is necessary to know what the factors underlying trade complementarities are and how they change over time. Theoretically, if trade complementarity between a pair of countries increases over time, and if their determinants are sustainable, then one could argue that, regardless of any probable policy manipulation of preferential trading arrangements and/or trade blocs, there is a natural tendency for some Asian countries to become natural trading partners. 3 On the other hand, if the creation of a trading bloc is driven by policies that do not consider the underlying factors of trading complementarities, then the sustainability of that trading bloc is subject to scrutiny unless it is completely manipulated by international politics. Once the hypothesis of trade complementarities is empirically proven, then it would make more sense to push for further economic integration among them in the near future. Though there is no consensus among economists on which nations are natural trading partners, 4 one can try to discern them through trade complementarity by investigating their export import specialties of commodities. In general, in the absence of any preferential trading arrangement, a pair of countries which trade much more intensively with each other by comparison with other trading partners in the world could be considered as natural trading partners. The underlying factors contributing to the natural trading partnership could be attributed to various trade theories. Conventional classical and or neoclassical theories tend to suggest that relative labor costs and/or factor endowments between nations must be attributes. Supporters of the product cycle theory would argue for the technological gap, whereas advocates of Linder s hypothesis would argue that similar tastes or levels of incomes among nations could be proxy determinants. In this chapter, the trade complementarities between nations are measured by the structure of export commodities between a pair of countries (Drysdale and Garnaut, 1982). The rationale is that, in the absence of preferential trading arrangements or other policy distortions, if country A s export concentration or specialty happens to be that of country B s

3 302 Trade and industrial development in East Asia imports, then A and B can be reasonably deemed natural trading partners. Since the exports of most Asian countries under study have been growing rapidly after their economies took off, one could test whether trade complementarities increased progressively in accordance with incomes per capita. In addition to exploring the international political economy of integration, results from this chapter should empirically provide some significant economic rationale for the formation of trading blocs, which bears important policy implications for economic integration in the region. 8.1 THE TRADE COMPLEMENTARITY INDEX, THE INTENSITY INDEX, AND THE BIAS INDEX Frankel et al. (1995) define what are natural, unnatural, and supernatural trading blocs. According to their definition, if in the absence of any preferential trading arrangement a pair of countries trade much more intensively with each other by comparison with all the other countries in the world, then they are generally considered as natural trading partners. Therefore, whether two countries are natural trading partners or not can be examined by the degree of their trade intensities. Intuitively, one would expect that strong complementarities exist between nations when marked differences in their comparative advantages are observed. The studies on the revealed comparative advantages in Chapters 2 and 3 could serve as a reflection of the differences and similarities of comparative advantage within the Asian countries. Recall from those chapters that none of the first tier of the NICs has really replaced Japan, even in the traditional labor- intensive products, in the world market. Hence, despite the changing dynamic comparative advantages in the developing Asian countries generalized in Akamatsu s flying geese paradigm, there are still some marked differences in their respective degrees of competitiveness, and these could be the source of trade complementarities between these developing Asian countries. 5 However, in examining trade relations between nations, one has to distinguish trade complementarities from trade intensity. According to Drysdale and Garnaut (1982), the trade intensity index measures the degree to which two countries trade more intensively with each other than with other trading partners. The trade complementarities index measures the extent to which the export pattern in one country matches the commodity mixture of import patterns of its trading partner. In fact, trade intensity between nations could be attributed to trade complementarities and trade bias. According to Drysdale and Garnaut (1982), the trade intensity index, Tij, is defined as the share of country

4 Is East Asia a natural trade bloc? 303 i s exports going to country j relative to the share of j s imports from the world. The complementarity index, Cij, measures the extent to which the export specialization in one country matches the import specialization in another country. Notationally, Tij can be represented by the following formula: T ij 5 X ij/x i M j /T where: X ij 5 country i s exports to country j X i 5 total exports in country i M j 5 total imports in country j T 5 total world trade. (8.1) In summary, the trade complementarity index, Cij, can be written as: C ij 5 ak a X ik X i b a T T k b a M jk M j b (8.2) where: k 5 individual commodities k X ik 5 country i s export of k commodity M jk 5 country j s imports of k commodities T k 5 the total trade of k product in the world T, X i, M j 5 the same meaning as above. Cij measures what the intensity index Tij would be if country i s exports of each commodity k were distributed throughout the world import market proportionally to each market s share of world import of commodity k. The trade bias index (Bij) is by defined as: B ij 5 X ij (X ik M jk ) /T k (8.3) A biased index of individual commodities could serve as a proxy of the relative factors which influence various trade flows on a bilateral basis (the Drysdale s resistance factors). 6 The weighted average of Bij measures the average effect of differential factors on intensity of bilateral trade. Drysdale and Garnaut (1982) illustrate that the trade intensity index, Tij, can be decomposed into a trade complementarity index, Cij, and a trade bias index, Bij. In fact, one can easily show that T ij 5 C ij B ij. 7

5 304 Trade and industrial development in East Asia This implies that the degree of trade intensity between any pair of countries can possibly be attributed to two major factors. The first is called the complementarity factor which describes how the commodity structure of exports in one country tends to converge to the product composition of another country s import mix. In that case, the commodity composition of its exports matches closely that of its trading partner s imports. In other words, one country s export concentration (specialization) matches well with the import concentration of another country. Hence, without policy distortions or preferential trading arrangement, a pair of countries with a relatively high complementarity index tends to be considered as natural trading partners. In analyzing the determinants of the trade complementarities, one could argue that as the income gaps between most developing Asian countries and the OECD become narrower over time, a strong association between relative income per capita and the degree of trade complementarity between them would push them closer to becoming natural trading partners. This argument, which is yet to be proved empirically, is similar to Linder s hypothesis of similar tastes and the overlapping of representative demand. The second factor that determines trade intensity can be attributed to the trade bias index, which measures the accessibility of country i s exports to country j s market relative to what might be expected from both countries shares of world trade in each commodity. The trade bias index could be attributed to discriminatory trade policies, preferential trading arrangements such as a free trade area (FTA), and political ties between exporting and importing countries such as and China. If Bij is greater than unity, then there is a positive or favorable bias for the bilateral trade. Drysdale and Garnaut (1982) point out that intensity of the trade index fails to make allowance for the varying commodity composition of countries foreign trade, because trade opportunity is limited by the degree of complementarities in the commodity composition of trade between trading partners. On the other hand, the trade complementarity index does take into account the degree of closeness of countries commodity structures relative to that of total world trade. Therefore, Cij is a better barometer than Tij for trade complementarities because the former can better reflect whether or not export specialization in one country matches the import concentration in another country. In general, two countries have a strong degree of complementarities if the calculated Cij is greater than unity. Otherwise, it will be less than 1. Hence, the threshold of Cij greater than or less than unity could serve as the benchmark of natural partners between a pair of countries. To measure the degree of trade complementarities between countries in

6 Is East Asia a natural trade bloc? 305 this chapter, the trade complementarity index will be analyzed econometrically, whereas the trade intensity index will be addressed as a reference on the degree of the intensity of trade flows between nations. 8.2 TRADE INTENSITY INDEX Comprehensive time series data on trade intensity index could reveal whether two countries have been trading with one another relatively more intensively than with the rest of the world. The trade FDI nexus described in Chapter 7 should have a strong impact on the trade intensity in East and South East Asia; the phenomenon of the triangular trade flow illustrates that the NICs, notably Korea and Taiwan, import technology and immediate goods from Japan, further processed either domestically or in China and ASEAN countries through outward FDIs, then re- export final products to the US. To some extent, this triangular trade flow could be reflected in the development of the relative trade intensity index. Since the end of WWII, Japan has been one of the major suppliers of production technology and intermediate products for many developing Asian countries, notably Korea and Taiwan. Japanese multinational enterprises (MNEs) coordinate their production and distribution networks through foreign direct investment in East and South Asian countries (Ando and Kimura, 2003). Hence, developing Asian countries must have been trading intensively with Japan as the major importing country, with the US as the largest exporting market and China as the largest assembly country. With its size of total GDP relative to the rest of Asian countries, China has become the largest trading partner for many Asian countries after the decade of the 1990s. Therefore, one could argue that there are three major players in the emerging trade blocs in Asian economic integration (Chapter 9). As a rule of thumb, the US serves as the market of final products (as a last resort), whereas Japan serves as an important provider of technology and intermediate goods, and China serves as the major assembly point to process and reprocess those intermediate goods before exporting final goods to the US and other OECD countries. The NICs and the ASEAN- 4 have been providing the global production network with their intermediate functions in the triangular trade flow at various stages of manufacturing production in accordance with their comparative advantages. Though the trade intensity index was calculated for pairs of trading partners among all ten countries under study, the results are too voluminous to be presented in full in this chapter. Therefore, analysis of the path and development of the trade intensity index will focus on China, Japan, and the US. 8

7 306 Trade and industrial development in East Asia Tij Year Japan Singapore Taiwan Figure 8.1 Trade intensity index of Japan and the first tier of the NICs with China Figures 8.1 to 8.6 show the trade intensity index for Asian countries in the three major markets in China, Japan, and the US based on five- year intervals. The absolute numerical magnitude of the Tij is less significant than the path of its development, which illustrates the trend of trade intensity between a pair of trading partners. Tij could also illustrate the relative degree of trade intensity among a country s trading partners within the same time period. Hence, one should look at the following figures from both the time series and cross- sectional perspective. From Figure 8.1, one can see that, Korea, Taiwan, and to a lesser extent Japan, have been trading fairly intensively with China since China s economic reform and openness in Among the trading partners which have been trading with China, s trade intensity increased steadily from 1980 to 2000, and fluctuated highly after 2001, but it still ranked as the most intensive trading partner in Hence, is ranked the highest, trading most intensively with China, for obvious reasons. Next in the list in Figure 8.1 is Taiwan, whose trade intensity index also increased steadily after the 1990s. In spite of political entanglements between the governments in Beijing and Taipei, Taiwan experienced accelerating trade growth with China in the 1990s, and it reached the top two of intensive trading partners of China in Korea established its formal diplomatic ties with China in August Since then, Korea s

8 Is East Asia a natural trade bloc? Tij Year Malaysia Indonesia Thailand Figure 8.2 Trade intensity index of the ASEAN- 4 with China trade intensity with China has been well above 3, except during the worldwide recession in Korea maintained its position in the top three intensive trading partners with China through Among South East Asian countries, the trade intensity index with China was in the order of Singapore, Thailand, the, Indonesia, and Malaysia. Interestingly, Japan had a low trade intensity with China compared to Korea and Taiwan, and its trade intensity index fluctuated over time. This phenomenon was probably due to the size of its total exports relative to other Asian countries; while Japan s exports to China (Xij) could be substantially greater than other Asian countries, they may still be a relatively small percentage of its total exports to the world (Xi). Hence, the ratio Xij/Xi in equation (8.1) is still relatively smaller than those in Korea and Taiwan. In Japan s market, one can find that the ASEAN- 4 traded more intensively with Japan than with the NICs, mainly because of Japan s poor resource endowment relative to the ASEAN- 4. Another reason is the small trade volume in the ASEAN- 4 relative to that of the first tier of the NICs. Among them, Japan s most intensive trading partner was Indonesia, followed by the, Thailand, and Malaysia. Among the NICs, Korea traded more intensively with Japan than did Taiwan, followed by Singapore and then, which was ranked lowest. The reasons for Korea and Taiwan to trade relatively intensively with Japan than with and Singapore include the colonial legacies in these

9 308 Trade and industrial development in East Asia Tij Year Singapore Taiwan Figure 8.3 Trade intensity index of the first tier of the NICs with Japan 8 6 Tij Year Malaysia China Indonesia Thailand Figure 8.4 Trade intensity of China and the ASEAN- 4 with Japan two countries and the development trajectories in Korea and Taiwan, as argued in Chapters 2 and 3. It is also noted that China s trade intensity with Japan, though relatively higher than the trade intensities of the NICs, declined over time, as shown in Figures 8.3 and 8.4.

10 Is East Asia a natural trade bloc? Tij Year Japan Singapore Taiwan Figure 8.5 Trade intensity index of Japan and the first tier of the NICs with the US Asia s trade intensity with the US has undergone substantial changes since China adopted its export- led growth strategy; prior to the rise of China in the world market, all Asian countries had a trade intensity index greater than 1, except for Thailand, which had 0.97 in But the trade intensity index in most Asian countries declined after Even though Japan, Malaysia, and the still traded intensively with the US, their trade intensity indexes declined over time. The phenomenon of all Asian countries other than China experiencing declining trade intensity with the US reflects the shift of the trade structure after the accelerated development of the global production network and the global value chain, with China becoming the world assembly line and workshop (Chow, 2010). This phenomenal development diverted much export from Japan and the first tier of the NICs; notably, Korea and Taiwan detoured to China for reprocessing before destining their final products to the US. In 1990, all Asian countries, except for Indonesia, had a trade intensity index greater than 1 with the US. In 1990, all Asian countries, except for Indonesia, had a trade intensity index greater than 1 with the US. Except in China, trade intensity indexes in all Asian countries with the US in 2000 were lower than those in In 2006, only China, Japan, Malaysia, and the had a trade intensity index greater than 1 with the US.

11 310 Trade and industrial development in East Asia Tij Year Malaysia China Indonesia Thailand Figure 8.6 Trade intensity index of China and the ASEAN- 4 with the US 8.3 TRADE COMPLEMENTARITY INDEX FOR ALL EXPORT COMMODITIES In this section, the trade complementarity index for all export commodities for these ten Asian countries with the US, Japan, and China is analyzed. As noted before, if the trade complementarity index between a pair of trading countries is greater than 1, then one would argue that these two countries have complementarity. In the time series of the trade complementarity index for those trade- dependent economies over time, one can examine the development of trade complementarity for a pair of trading countries over time as well as the degree of trade complementarities between countries through the relative magnitudes of their complementarity indexes, Cij. Figures 8.7 and 8.8 show that trade complementarities between China and most Asian countries increased over time. Moreover, China tended to have higher trade complementarities with Japan and the first tier of the NICs than with the ASEAN- 4. In the 1990s when policy distortions of trade flows with China were gradually reduced, Korea and Taiwan were ranked as the top trading nations with China and had relatively higher Cij than all other Asian countries. In fact, after Korea formally built diplomatic ties with China in 1992, its trade complementarity index with China became greater than Japan s. Taiwan s trade complementarity with China increased steadily after its trade policy with China

12 Is East Asia a natural trade bloc? Cij Figure Year Japan Singapore Taiwan Cij for all export commodities for Japan and the first tier of the NICs with China Cij Year Malaysia Indonesia Thailand Figure 8.8 Cij for all export commodities for the ASEAN- 4 with China was gradually liberalized in the 1990s. Due to Taiwan s relatively faster growth of exports to China in the 1990s, China became its largest trading partner in As a result, Taiwan s trade complementarity index with China exceeded that of all other Asian countries. It is also notable that the

13 312 Trade and industrial development in East Asia Cij Figure Year Singapore Taiwan Cij for all export commodities for the first tier of the NICs with Japan unusually high complementarity index between and China has been gradually leveling off since the 1990s, when Korea and Taiwan became important trading partners for China. As shown in Figures , China and the ASEAN- 4 had higher complementarity with Japan than did most of the first tier of the NICs. The first tier of the NICs had Cij values less than 1 with Japan most of the time, whereas those of China and the ASEAN- 4 were generally greater than 1. This further confirms that, relatively speaking, China and the ASEAN- 4 are better- endowed than Korea and Taiwan with resources which match Japan s market demand. Among the ASEAN- 4, Indonesia had the highest trade complementarity with Japan, followed by Malaysia, the, and Thailand. Figures report the trade complementarity index with the US. All Asian countries, except for China, had a trade complementarity index greater than 1 in the early 1980s. Japan and the first tier of the NICs led the rest of the countries by having a relatively higher Cij than those of China and the ASEAN- 4. But the situation changed substantially in the second half of the 1980s after the rise of China, which has been penetrating the US market more aggressively than other Asian countries. China s trade complementarity index with the US passed the threshold of 1 in

14 Is East Asia a natural trade bloc? Cij Figure Year Malaysia Indonesia Thailand China Cij for all export commodities for the ASEAN- 4 and China with Japan 1987 and increased steadily in the following decade before it decelerated. In 1998, China became the most complementary trading partner with the US, having the highest Cij compared to other Asian countries, and it has maintained its status since then. Among the ASEAN- 4, the had the highest score of Cij with the US, followed by Malaysia, Indonesia, and Thailand. Nevertheless, the trade complementarity index of the ASEAN- 4, except for Thailand, declined after the 1990s, plausibly due to the rise of China. Japan s trade complementarity with the US was greater than the threshold of 1 from 1980 to For the first tier of the NICs, Korea, Taiwan, and Singapore had trade complementarity indexes with the US greater than 1 until the second half of the 1990s (Cij declined to less than 1 in 1995, 1998, and 1999 respectively in these three countries), whereas had Cij greater than 1 all the time except for Since the trade complementarity index is calculated from actual trade flows, which incorporate all existing trade policies, the relative shift in its magnitude reflects the changing status of the respective partnerships between trading countries under the existing trading framework. The period of observation includes several major developments, which were incorporated to generate the actual trade flows in those Asian countries. They include, but are not limited to, the trade liberalization undertaken

15 314 Trade and industrial development in East Asia Cij Figure Year Japan Singapore Taiwan Cij for all export commodities for Japan and the NICs with the US Cij Figure Year Malaysia Indonesia Thailand China Cij for all export commodities for ASEAN and China with the US

16 Is East Asia a natural trade bloc? 315 in various countries such as China s openness in , Taiwan s gradual trade liberalization with China since the mid- 1980s, the currency appreciations after the Plaza Accord in 1985, Japan s trade tensions with the US in the 1980s, the phasing out of the generalized system of preference for the NICs in the US market in 1989, Korean trade relations after its formal diplomatic ties with China in 1992, and the overall drive for globalization. Therefore, the measurement of trade complementarities in this section has incorporated many of these crucial factors and better reflects the actual status of trading partnerships in the region. 8.4 COMPLEMENTARITY INDEX FOR MANUFACTURED PRODUCTS In this section, the trade complementarity index is calculated for manufactured commodities only. In general, manufactured commodities include the commodities under SITC 5 to 8 but not SITC 68 (non- ferrous metals). Therefore, the product groups under SITC 0 4, 9, and 68 are excluded from analysis in this section. Nevertheless, because manufactured exports dominated the trade flows in most Asian countries, the complementarity index Cij for manufactured products is highly correlated with the Cij for all export commodities. 9 In China s market, one can find that, year by year, the Cij of manufactured products for the pairs of China is relatively higher than the Cij of all export commodities. Obviously, because the products under SITC 0 4 and 68 have been excluded, tended to have higher trade complementarities with China mainly because it does not have many natural resource- based and primary commodities to trade with China. The same phenomenon also existed in Taiwan, and to a lesser extent in Japan and Korea. The differences between the two sets of Cij in the ASEAN- 4 are less significant than those in the NICs. Indonesia seems to have lower Cij in its manufactured products than in its overall commodities in the early 1990s. But the differences between these two sets of Cij gradually leveled off after the 1990s, as in other ASEAN countries, as shown in Figures 8.13 and The situation in Japan is quite different, as reported in Figures 8.15 and 8.16; except for Indonesia and Malaysia, all these Asian countries tended to have relatively lower Cij for manufactured products than for all export commodities with Japan. This implies that for most Asian countries, the trade complementarity with Japan was enhanced by including the resource- based and primary commodities due to the factor endowments in Japan. For Indonesia and Malaysia, the Cij for manufactured products with Japan was higher than that for all export commodities after For

17 316 Trade and industrial development in East Asia Cij Japan Taiwan Singapore Figure 8.13 Cij for manufactured products: Japan and the NICs with China 4 3 Cij Malaysia Indonesia Thailand Figure 8.14 Cij for manufactured products: ASEAN- 4 with China

18 Is East Asia a natural trade bloc? Cij Singapore Taiwan Figure 8.15 Cij for manufactured products: the NICs with Japan 4 3 Cij Malaysia China Indonesia Thailand Figure 8.16 Cij for manufactured products: China and the ASEAN- 4 with Japan

19 318 Trade and industrial development in East Asia Cij Japan Taiwan Singapore Figure 8.17 Cij for manufactured products: Japan and the NICs with the US Thailand, the Cij for manufactured products was also either higher or at a similar level to that for all export commodities (higher in 2001, 2002, and 2003). This is an interesting finding which further illustrates the shifting trade structure of Japan after the Plaza Accord. As noted before, Japan deliberately expanded its outward FDI in South East Asia after the yen was appreciated from Japanese MNEs used the location advantages by investing their labor- intensive sectors in South East Asian countries, and they re- imported some of those finished manufactured products back to Japan. Therefore, in terms of trade complementarities, Indonesia and Malaysia had higher Cij with Japan in manufactured products than in overall export commodities. The exceptional case among the ASEAN- 4 is the, probably due to its less favorable investment environment than those in other ASEAN countries. In the US market, Figures 8.17 and 8.18 show that Japan, Taiwan, and to a lesser extent Korea tended to have higher Cij for manufactured products than for overall export commodities. After 1985, most of their trade complementarity indexes for manufactured products were either higher, or more or less the same, compared to those for overall export commodities with the US. Given their relatively poor resource endowments compared to the US, Japan, Korea, and Taiwan tended to have higher trade complementarities with the US in manufactured products than in overall export commodities. But the situation is quite different in China and the

20 Is East Asia a natural trade bloc? Cij Malaysia China Indonesia Thailand Figure 8.18 Cij for manufactured products: China and ASEAN- 4 with the US ASEAN- 4; their trade complementarities with the US are higher in all export commodities by including the resource- based and primary commodities rather than just the manufactured products. 8.5 REGRESSION MODEL ON Cij Regression analysis of trade complementarity between a pair of partners will be conducted by using the Cij as the dependent variable. One advantage of using the trade complementarity index as the dependent variable, rather than the trade intensity index, is that it nets out the bias index and thereby sterilizes any possible trade distortions caused by the general trade intensity index. This section will examine whether trade complementarity between any pair of countries could be explained by the ratio of their relative incomes per capita, their real exchange rates, and the market size of the importing country. Choosing the ratio of relative income per capita between trading partners is justified by the argument that, if the degree of trade complementarity is sustainable over time, then it is generally expected that the relative income gap would have significant impacts on trade complementarities. To some extent, this model moves beyond the argument of the Linder hypothesis (Linder, 1961) in that trades between nations are determined by their similarity taste and/or income levels.

21 320 Trade and industrial development in East Asia Under a free trade regime, trade flows between countries could grow sustainably only if both countries are complementary to each other. Since Asian countries have been growing faster than other developing countries in Latin America and Africa in the past decades, it is justifiable to use the relative income per capita as one of the independent variables in the regression. On the other hand, trade complementarities between nations must be influenced by the relative prices between them. Hence, the real exchange-rate between a pair of trading partners must be one of the important determinants in the regression model. Therefore, both the income ratios and real exchange rates are included as the explanatory variables in this model. The market size could be one of the explanatory variables to reflect that trade complementarity between a pair of trading partners could depend, among other factors, on the market potentials of the trading partners. Other things being equal, the larger the market size, the greater the trade complementarity between a pair of trading partners. The model is a modified gravity model (Krugman, 1991; Frankel, 1992) with three significant differences from the existing literature. The first is to use the innovative complementarity index rather than the trade volume or trade intensity as the dependent variable. The second is to include the variable of real exchange rates, which was generally excluded from the gravity model in some existing literature such as Krugman (1991). The third is that the income ratio, rather than the product of exporter s and importer s per capita incomes, is one of the explanatory variables. The regression model hypothesizes that the trade complementarity index between two countries i and j, Cij, as a function of the relative income per capita (PCIi /PCIj) and their real exchange rates [(XRi/WPIi)/(XRj/ WPIj)]. By taking natural logarithms for all variables, the following regression model is obtained: ln 5 C ij 5 f aln PCI i, ln XR i/wpi i b PCI j XR j /WPI j (8.4) where: PCIi 5 per capita income in country i PCIj 5 per capita income in country j XRi 5 unit of currency i per unit of currency i XRj 5 unit of currency j per unit of currency j WPIi 5 wholesale price index in country i WPIj 5 wholesale price index in country j. Data of per capita income are the per capita income in constant international dollars (Chain Index), whereas XRi/XRj is the cross exchange rate derived from the exchange rates between currencies i and j and the

22 Is East Asia a natural trade bloc? 321 US dollar. Both sets of data are downloaded from Penn World Tables 6.8. Data for WPI were derived from the World Development Indicators and Global Development Finance of the World Bank, except for Taiwan, whose data are from the Taiwan Statistical Data Book. 8.6 EMPIRICAL RESULTS There are 90 sets of regression models conducted in this study. Results from the Ordinary Least Squares (OLS) are reported in Appendix A8.2 in Tables A8.2a to A8.2j. There are three important findings from the ordinary least squares (OLS) on the Cij for each pair of trading partners. The first is to confirm that relative income per capita between a pair of countries is important in determining the trade complementarity. Out of the 90 regression models, nearly two- thirds (58) have shown that the variable of relative income per capita, ln(pci/pcj), is significant at various p-values. Since the income ratio in the regression model is defined as the exporter s income per capita (PCi) relative to that of the importer s income per capita (PCj), the positive and negative signs of the estimated coefficient ln(pci/ PCj) have to be interpreted by incorporating the numerical magnitudes of relative income per capita between exporting and importing countries. The second finding is that the real exchange rate between the exporting and importing countries is also highly significant: 55 out of the 90 regressions show that the estimated coefficients of the real exchange rates are statistically significant at various levels. This implies that for a pair of trading partners to be complementary, it is imperative to have more realistic exchange rates. The third finding seems more relevant to the determination of trading partners than the previous ones. Theoretically, the trade complementarity index for two countries Cij should be the same or close to Cji. That is to say that if country i has strong trade complementarity with country j, then country j must also have strong trade complementarity with country i. But the empirical evidence from the regression models does not fully support that conventional wisdom at all. It implies that there are some countries i whose trade is complementary with country j, but j s trade may not be complementary with i. This finding will have strong policy implications in determining the natural trading partner, to be addressed in Chapter 9. The less- than- fully robust findings in this section could be explained by the fact that the trend of Cij is far more constant than the trade intensity index, as reported in section 8.2, whereas income gaps, with few exceptions, have been narrowed down much more rapidly between the trading partners since the late 1980s.

23 322 Trade and industrial development in East Asia 8.7 POOLING THE Cij IN THE THREE MAJOR MARKETS To investigate further the determinants of the trade complementarity, a pooling of all the exporters in the three major importing countries i.e. China, Japan, and the US is conducted in this section. To measure whether the market size of the importing country is an important factor in determining the trade complementarity, the total gross domestic product (GDP) of the importing country (GDPj) is included as another independent variable in the regression. An alternative method of measuring the market size is to take the product of the GDP in both exporting and importing countries as an explanatory variable. The rationale of adopting this variant of market size is to try to deal with the imperfect competition in modern trade theory, which is empirically well established in bilateral trade regressions, as Hejazi and Safarian argued (2001, p. 435). 10 Therefore, two alternative regression models are utilized in the following way: ln C ij 5 f aln PCI i PCI j, ln XR i/wpi i XR j /WPI j, GDP j b (8.5) ln C ij 5 f aln PCI i PCI j, ln XR i/wpi i XR j /WPI j, GDP i GDP j b (8.6) where GDPi and GDPj are the total GDP in countries i and j. All other variables have the same meanings as in (8.4). Before proceeding with the pooling, the Phillips Perron unit root test is conducted. The results are reported in Appendix A8.3. The pooling results on regressing Cij are reported in Tables 8.1 to 8.3. Both OLS and fixed effect models are conducted using (8.4), (8.5), and (8.6). It is interesting that in the pooling regressions of their trade complementarity indexes Cij, the three major markets respond quite differently to those parameters hypothesized. In China s market, none of these three explanatory variables the income gap, the real exchange rates, and market size are significant. But in Japan s market, the exchange rate variable is significant at the 95% level (p, 0.05) under OLS in regression models (8.5) and (8.6). In the US market, the gap of per capita income is significant in both the OLS and fixed effect model under (8.4) and (8.6), as shown in Table 8.3. Hence, in spite of the fact that the gap of per capita income between exporting and importing countries has been narrowed by the diversity of income and economic growth, there is still robust evidence that as

24 Is East Asia a natural trade bloc? 323 Table 8.1 Pooling of all exports in the Chinese market Importer: China d.ln(pci_i/ PCI_j) d.ln ((XR_i/WPI_i)/ (XR_j/WPI_j)) (8.4) OLS d.ln(c_ij) b/se/t (8.4) FE d.ln(c_ij) b/se/t (8.5) OLS d.ln(c_ij) b/se/t (8.5) FE d.ln(c_ij) b/se/t (8.6) OLS d.ln(c_ij) b/se/t (8.6) FE d.ln(c_ij) b/se/t (0.092) (0.001) ( 0.024) ( 0.153) (0.172) (0.145) ( 0.240) ( 0.449) ( 0.273) ( 0.448) ( 0.298) ( 0.477) d.ln(gdp_j) ( 0.161) ( 0.429) d.ln(gdp_ij) ( 0.273) ( 0.604) constant (1.083) (1.374) (0.566) (1.959) (0.669) (2.124) N AdjR R2 within R2 overall Notes: t - values are in parentheses. economies in East and South East Asia continue to grow, the US could be considered a trading partner for many of them. The elasticity of trade complementarity with respect to the income gap ranged from 0.35 to This implies that as the level of per capita income in those East and South East Asian countries catches up and their income gaps with the US narrow by 1%, the degree of their trade complementarity will increase from 0.35 to SUMMARY AND CONCLUSION The trade intensity index can measure the degree to which two countries trade more intensively with each other than with other partners, but it cannot necessarily measure their trade complementarities. After netting out

25 324 Trade and industrial development in East Asia Table 8.2 Pooling of all exports in the Japanese market Importer: Japan d.ln(pci_i/ PCI_j) d.ln ((XR_i/WPI_i)/ (XR_j/WPI_j)) (8.4) OLS d.ln(c_ij) b/se/t (8.4) FE d.ln(c_ij) b/se/t (8.5) OLS d.ln(c_ij) b/se/t (8.5) FE d.ln(c_ij) b/se/t (8.6) OLS d.ln(c_ij) b/se/t (8.6) FE d.ln(c_ij) b/se/t (0.950) (0.419) (1.009) (0.430) (0.698) (0.384) 0.090* * * (2.460) (1.863) (2.480) (1.855) (2.452) (1.845) d.ln(gdp_j) (0.412) (0.393) d.ln(gdp_ij) (0.088) (0.278) constant ( 0.327) ( 0.281) ( 0.522) ( 0.365) ( 0.273) ( 0.292) N AdjR R2 within R2 overall Notes: t - values are in parentheses. * p,0.05, ** p,0.01, *** p, the trade bias index, the trade complementarity index measures the degree that export specialty in one country matches import specialty in another country. Hence, trade complementarity between a pair of trading partners is one of the key determinants of a natural trading bloc. The empirical studies in this chapter detail the development of trade complementarities in the three major markets for the trade- dependent countries in East Asia China, Japan, and the US for all export commodities and manufacturing products from the mid- 1980s to the beginning of the 21st century. Regression analyses on the trade complementarities index for 90 pairs of trading partners show that the gap of income per capita between exporting and importing countries, and the real exchange rates, are fairly significant in determining the degree of trade complementarity. This implies that, other things being equal, as the income gap becomes narrower a

26 Is East Asia a natural trade bloc? 325 Table 8.3 Pooling of all exports in the US market Importer: USA d.ln(pci_i/ PCI_j) d.ln ((XR_i/WPI_i)/ (XR_j/WPI_j)) (8.4) OLS d.ln(c_ij) b/se/t (8.4) FE d.ln(c_ij) b/se/t (8.5) OLS d.ln(c_ij) b/se/t (8.5) FE d.ln(c_ij) b/se/t (8.6) OLS d.ln(c_ij) b/se/t (8.6) FE d.ln(c_ij) b/se/t 0.351** 0.368* *** 0.712** (3.214) (3.019) (1.269) (0.847) (5.347) (4.151) (1.056) (1.540) (0.901) (0.727) (0.970) (0.867) d.ln(gdp_j) 1.203*** 1.233** ( 4.698) ( 4.211) d.ln(gdp_ij) 0.558*** 0.584** ( 4.419) ( 4.156) constant 0.022*** 0.023*** ( 4.135) ( 8.128) (1.932) (2.110) (1.788) (2.146) N AdjR R2 within R2 overall Notes: t - values are in parentheses. * p,0.05, ** p,0.01, *** p, pair of trading partners will become more complementary. It also implies that to maintain a more realistic exchange rate regime between a pair of trading partners, it is conducive for them to become more complementary. But the 90 regression models also show that trade complementarity is not reciprocal. That is, country i may have strong trade complementarity with country j, but j s trade may not necessarily be complementary with i s. The pooling of all exporting countries in the three major markets shows that three major markets i.e. China, Japan, and the US respond differently to the parameters hypothesized. In China s market, none of the three variables of income gap, exchange rate, and market size are significant in pooling the trade complementarities index for all exporting countries. For Japan s market the real exchange rate is significant, whereas in the US market the income gap is significant. This implies that maintaining

27 326 Trade and industrial development in East Asia a realistic exchange rate regime in East and South East Asia would be conducive to trade complementarities with Japan. Moreover, as economic growth rates in East and South East Asian countries continue to exceed those in the US, and as the income gap narrows over time, those tradedependent countries become more complementary with the US. NOTES 1. See Kreinin and Plummer (1994), Frankel et al. (1995). 2. For the revealed comparative advantage of the newly industrialized countries (NICs) and the ASEAN- 4 in the Organisation for Economic Co- operation and Development (OECD) countries, see Chow and Kellman (1993), Chapter In this chapter, the natural trading partners refer to a pair of countries whose trade intensity is generated through a free trade regime rather than policy- driven preferential trading arrangements. 4. The present study only deals with natural trading partners based on trade complementarity rather than the formulation of natural trading blocs. For the arguments on natural, unnatural, and super- natural trading blocs, see Frankel et al. (1995). 5. Since the revealed comparative advantage (RCA) index was calculated from the actual trade data in the world markets, discriminatory treatment on importing countries could mislead the actual differences on their trade competitiveness. This serves as a strong justification to examine the complementarities which excludes bias factors of trade by netting out the trade bias index as illustrated below. 6. Drysdale s resistances include both objective and subjective factors on determining bilateral trade. Objective resistances can be decomposed into transport and other costs of overcoming geographic distance, and the costs of overcoming official barriers to trade. Subjective resistances include imperfect information available to businessmen, internal constraints on profit maximization, and particular processes that firms undertake in international business. 7. This argument was highlighted by Anderson (1983). 8. Trade intensity index results for all other Asian countries are available upon request. 9. The correlation coefficients between these two sets of Cij (manufactured products versus all export commodities) for the same countries in China s market are all positive and significant at the 99% level of significance, except for Indonesia. In Japan s market, the correlation coefficients are positive and significant for China,, Singapore, and Taiwan. In the US market, the correlation coefficients are all positive and significant for all countries but Thailand. 10. A Markusen and Maskus model (2001) used the sum of total GDP in home and host countries as a measurement of market size in their regression model. Yet, this author found that coefficients of the sum of GDP are much less than the product of the GDP in Chow (2011). REFERENCES Anderson, Kym Prospects for Trade Growth Among Pacific Basin Countries. Developing Economies, 21(4), pp Ando, Mitsuyo and Kimura, Fukunari The Formation of International Production and Distribution Network in East Asao. National Bureau of Economic Research Working Papers series

28 Is East Asia a natural trade bloc? 327 Chow, Peter C.Y China as the World Market and/or the World Factory in the Global Economy. In China and the World Economy New Perspectives on China s Economic Rise after Three Decades of Reform. Berliner China Hefte / Berlin History and Society, No. 37. pp Chow, Peter C.Y The Effect of Outward Foreign Direct Investment on Home Country s Export: A Case Study on Taiwan, Journal of International Trade and Economic Development, forthcoming. Chow, Peter C.Y. and Kellman, M Trade the Engine of Growth in East Asia. New York: Oxford University Press. Drysdale, P. and Garnaut, R Trade Intensities and the Analysis of Bilateral Trade Flows in a Many- Country World: A Survey. Hitotusbashi Journal of Economics, 22(2). Frankel, Jeffery A Is Japan Creating A Yen Bloc In East Asia and the Pacific? Cambridge, MA: National Bureau of Economic Research Working Paper Series No Frankel, Jeffrey, Stein, Ernesto, and Wei, Shang- jin Trading Blocs and the Americas: the Natural, the Unnatural, and the Super- natural. Journal of Development Economics, 47, pp Hejazi, W. and Safarian, A.E The Complementarity between US Foreign Direct Investment Stock and Trade. Atlantic Economic Journal, 29(4), Kreinin, M.E. and Plummer, M.G Natural Economic Blocs: An Alternative Formulation. International Trade Journal, 8(2), pp Krugman, Paul The Move Toward Free Trade Zones. In Policy Implications of Trade and Currency Zones, A Symposium sponsored by The Federal Reserve Bank of Kansas City. pp Linder, B.S An Essay on Trade and Transformation. New York: John Wiley & Sons. Markusen, James R. and Maskus, Keith E A Unified Approach to Intra- Industry Trade and Direct Foreign Investment. NBER Working Paper Razmi, A. and Blecker, R Developing Country Exports of Manufactures: Moving up the Ladder to Escape Fallacy of Composition? Journal of Development Studies, 44(1), pp