The Evolution and Reshaping of Globalization: A Perspective Based on the Development of Regional Trade Agreements

China & World Economy / 51 71, Vol. 27, No. 1, 2019 51 The Evolution and Reshaping of Globalization: A Perspective Based on the Development of Regional Trade Agreements Xiaohua Bao, Xiaozhuo Wang* Abstract With the weakening role of the World Trade Organization multilateral trading system, the globalization pattern is moving toward regional economic integration. As a result, the number of regional trade agreements (RTAs) has rapidly increased. New trends in international economics and trade, such as the withdrawal of the US from the Trans- Pacific Partnership and the trade disputes between the US and China, have revealed the intention of the developed countries represented by the US to reshape the direction of globalization. This paper combines the relevant research conclusions and current stylized facts to examine the evolution and reshaping of globalization. We find that: (i) countries have different attitudes toward the recent round of globalization, which are related to changes in the patterns of income distribution within countries caused by the last round of globalization; and (ii) regional economic development is an effective way to reshape globalization. The self-strengthening effect of the hub country in the trade network has promoted global RTA expansion. Key words: economic globalization, hubness, income inequality, regional trade agreements JEL codes: F13, F15, P48 I. Introduction Globalization has become increasingly prominent since the 1980s, and has largely promoted trade and economic growth across the world. However, globalization results in unequal income distribution among different countries and also among different interest groups inside the same country. Thus, countries have different attitudes toward the recent round of globalization and have declared their intention to reshape the direction of globalization, which has promoted dynamic development. *Xiaohua Bao, Professor, College of Business, Shanghai University of Finance and Economics, China. Email: xhbao369@mail.shufe.edu.cn; Xiaozhuo Wang, PhD Candidate, College of Business, Shanghai University of Finance and Economics, China. Email: 475789627@qq.com. We thank the referees for their thoughtful comments and acknowledge the financial support from the Natural Science Foundation of China (No. 71673177), the Social Science Foundation of China (No. 18ZDA069) and the SHUFE Graduate Student Research Project (No. CXJJ-2016-339).

52 Xiaohua Bao, Xiaozhuo Wang / 51 71, Vol. 27, No. 1, 2019 Two stylized facts can be observed in the process of globalization. One is that the world trade pattern has shown a multipolar development trend. Since the establishment of the World Trade Organization (WTO), developing countries have entered a new stage of rapid trade liberalization. During this period, the average most favored nation tariff level of developing countries dropped from 14.76 percent in 1995 to 8.03 percent in 2015, a reduction of as much as 45.6 percent. 1 The process of globalization has greatly promoted trade in developing countries: the whole trade share of global exports was 22.5 percent in 1995 and rose to 45.1 percent in 2015 (Figure 1). As a result of the economic growth of developing countries, the world trade pattern has evolved from American unipolarity to American European bipolarity to American European Asian tripolarity. At present, the regionalization of international trade has formed three major trade bodies: the EU trade body with Germany as the core, the North American trade body with the US as the core, and the East Asian economic integration body 2 with China and Japan as the core. In 2015, the total trade volume of the three major trade bodies accounted for 14.1, 18.5 and 28.4 percent of global trade in goods, respectively, which together accounted for 60 percent of the total global trade in goods. The service trade volumes of the three major trade bodies accounted for 49.1, 16.3 and 14.4 percent of global trade in services, respectively, which together accounted for 79 percent of the total global trade in services. We can conclude that the development of global trade has formed an interdependent trade network, in which some countries have assumed a central role. Figure 1. Proportions of Developing and Developed Countries in Global Trade, 1962 2015 Source: UN Comtrade. 1 Authors calculation based on data from the World Bank website. 2 The East Asian economic integration body includes 10 countries of the Association of Southeast Asian Nations plus China, Japan and Korea.

Globalization and Development of Regional Trade Agreements 53 The other stylized fact is the proliferation of regional trade agreements (RTAs). The delay in the Doha Round negotiations caused the multilateral trading system to weaken and countries turned to RTA negotiations as a solution. Thus, the number of RTAs concluded has rapidly grown (Figure 2). Since the 1990s, a new wave of regional economic integration has emerged (Bergstrand et al., 2016), which has shown three distinct characteristics. First, regarding total growth, according to WTO statistics, more than 300 different RTAs were concluded globally between 1958 and 2015. By contrast, 12.7 times more RTAs have been concluded in the last 25 years. 3 Figure 2. Growth of Global RTAs, 1958 2015 Source: WTO RTA Database. Note: RTA, regional trade agreement. Second, regarding network development, the intricate global RTA distribution pattern looks like a spaghetti bowl with obvious network characteristics (Bhagwati, 2002). Network is reflected not only in the overlapping of various RTAs, but also in the network structure of hub and spoke with the hub country as the core on different continents. The hub countries in the different continents are the United States, China and Japan, and Germany, which correspond to the core countries of the three major plates in the trade network. Third, regarding the structural characteristics, RTA growth in recent years has shown new features that the traditional theory cannot explain. As shown in Table 1, RTAs can be divided into free trade agreement (FTA) and customs union (CU); bilateral 3 When calculating the number of RTAs, it is necessary to specify that the expansion of an agreement is counted as an agreement. For example, the EU experienced seven expansions from 1958 to 2015, in 1973, 1981, 1986, 1995, 2004, 2005 and 2013, consecutively.

54 Xiaohua Bao, Xiaozhuo Wang / 51 71, Vol. 27, No. 1, 2019 and multilateral; South South, North North and South North; and cross-continental and intracontinental according to the degree of integration, type of agreement, development level of the member countries and whether the contracting countries are crosscontinental. The early theory of CUs suggests that it is easier for a country to form an RTA when there is a large number of member countries of similar economic size located in close proximity (Viner, 1950), such as the EU. However, during 1990 2015, bilateral agreements accounted for 82 percent of total RTAs. The number of cross-regional RTAs also grew: RTAs between developing and developed countries accounted for 53 percent. Countries are undergoing the process of regional economic integration in different forms. Table 1. Structural Characteristics of Regional Trade Agreements (RTAs), 1958 2015 Total RTA integration Number of members Development level Continent RTA FTA CU Multilateral Bilateral S S N N S N Cross- Intra- 1958 1989 22 15 7 14 8 4 6 12 6 16 1990 2015 281 257 24 42 239 114 19 148 92 189 Total 303 272 31 56 247 118 25 160 98 205 Ratio of each type of RTAs to total RTAs (%) 100 89.8 10.2 18.5 81.5 38.9 8.3 52.8 32.3 67.7 Source: WTO website. Notes: FTA, free trade agreement; CU, customs union; S S, South South; N N, North North; S N, South North. In fact, for both developed and developing countries, regional economic integration is a general trend. Japan has just reached an FTA with the EU; the US and the EU are also working on negotiations of zero-tariff trade. By the end of 2015, China had signed 14 RTAs, all of which were service, bilateral FTAs. 4 Moreover, China has established 4 According to the Ministry of Commerce website, in 2018, China will have 10 RTA promotion negotiations, including: China Chile (upgrade), China GCC (Gulf Cooperation Council), China Israel, China Japan Korea, China Moldova, China Norway, second stage negotiation of China Pakistan, China Singapore (upgrade), China Sri Lanka, Regional Comprehensive Economic Partnership. Ten FTA advancement feasibility studies are also being conducted, including: China Colombia, China Fiji, China India, China Mauritius, China Mongolia, China Nepal, China Palestine, China Panama, China Peru (upgrade), China Switzerland (upgrade).

Globalization and Development of Regional Trade Agreements 55 different levels of partnerships with 72 countries and organizations. Building an RTA network was an important national strategy for China to build an open economy. However, with the deepening of globalization, certain interest groups in some countries have been damaged and new trends have emerged in the international economy and trade, such as US withdrawal from the Trans-Pacific Partnership and the trade disputes between China and the US, which have revealed the intention of developed countries represented by the US to reshape the direction of globalization. What has driven the dynamic evolution of globalization from a multilateral to a regional pattern? Why are the majority of developing countries actively participating in economic integration while some developed countries show a protectionist tendency of national interest first? How will the direction of economic globalization be reshaped? These realistic problems have set our theoretical thinking. Combining the relevant research conclusions and the characteristics of the empirical facts, this paper examines the evolution and reshaping of the globalization pattern. We find that the reason for the dynamic evolution of globalization is the uneven distribution of interests. The establishment of bilateral and regional trade agreements will be a useful complement to the multilateral trading system in the process of reshaping globalization. It is also an important initiative for countries to actively participate in the formulation of international economic and trade rules and strive for bargaining power in global economic governance. This paper has important implications for understanding the dynamic evolution of the globalization pattern, dealing with the protectionist tendencies and building a network of global RTAs. The paper s contributions are twofold. First, it links the existing literature to clarify the mechanism of how trade liberalization influences the formation of trade policy through income inequality. We discuss the reasons for the dynamic evolution of the globalization pattern, pointing out that developing countries are the beneficiaries of the integration process, thus supporting more trade liberalization, while developed countries are the losers, who in turn show a protectionist tendency. Second, using hubness to characterize the relative position of countries in the hub spoke trade network, the paper points out that, in addition to political and economic factors, the self-reinforcing effect of the hub country will promote the global expansion of RTAs. Regional economic integration will become an effective way to reinvigorate globalization in a multilateral trading system. The rest of this paper is organized as follows: Section II analyzes the causes for the evolution of globalization pattern; Section III proposes an effective way to reshape globalization; and Section IV concludes and discusses the relevant policy implications.

56 Xiaohua Bao, Xiaozhuo Wang / 51 71, Vol. 27, No. 1, 2019 II. Evolution of the Globalization Pattern Economic globalization is a dynamic development process and a major background for global economic development. In this process, some developed countries have shown signs of rising trade protectionism in recent years, which has led people to question the driving force behind the dynamic evolution of the globalization pattern. In 2016, the US election year, three topics were the most popular among candidates and the public: China, globalization and the 1 percent (the wealthiest 1 percent of the country). Former World Bank economist Branko Milanovic believes that these three topics reflect that the current globalization trend has enabled the rapid rise of two developing countries, China and India, which has greatly reduced world inequality; however, within many developed countries, the polarization between rich and poor is expanding. As a result, people often misunderstand that globalization has exacerbated domestic inequalities. In his book, Global Inequality: A New Approach for the Age of Globalization, Milanovic (2016) examines the distribution of global income from 1988 to 2008 and reveals that the victims of the recent round of globalization, that is, the middle and lower middle classes in high-income countries, have suffered from the stagnation of real income for more than 20 years. How does this affect the formulation of a country s domestic trade policy? There are two relatively independent threads of literature related to this topic: one is the impact of trade liberalization on income distribution, and the other is the impact of income inequality on trade policy formulation. We first review the main conclusions and try to connect them using income inequality as a bridge to explain the impact of trade liberalization on trade policy. 5 1. The Impact of Trade Liberalization on Income Inequality From a macro perspective, trade affects income inequality through factor endowment and employment structure. First, the Stolper Samuelson (SS) theorem suggests that trade will increase the return to the factor that is used most intensively in the production of the export good, and decrease the return to the other factor. According to this theory, trade would increase the income of low-skilled workers in developing countries and the income of high-skilled workers would decline. Therefore, trade would lower the income gap in developing countries. However, empirical findings for developing countries are inconsistent with this theory. For example, based on data from 1960 to 1990, Barro 5 The aggregate data does not support country-level analysis, so we focus on the differences between developed and developing countries in the following sections.

Globalization and Development of Regional Trade Agreements 57 (2000) showed that trade openness would increase income inequality within the country, especially for countries in Africa and Latin America. Kanbur and Zhang (2005) also found that trade openness has deepened income inequality in developing countries, especially China. The theory and the empirical evidence are inconsistent because, from a theoretical point of view, SS theory can only be applied to examine the problem of income distribution in a two-sector and two-factor setting. On the other hand, from an empirical point of view, trade openness, used in most empirical literature, is measured by trade dependence, that is, the ratio of total import and export to GDP. However, a country s trade dependence is affected by factors such as the level of economic development, market size and domestic consumption demand. Thus, trade dependence cannot fully capture changes in trade openness. Another way to measure trade openness is by measuring government interventions, such as tariffs. With this alternative measure, Dutt et al. (2009) showed that trade openness and unemployment are negatively correlated. 6 Second, trade affects income distribution through employment restructuring. Jaumotte et al. (2008) used a sample of 51 developed and developing countries from 1981 to 2003 to examine the impact of trade dependence or tariffs on income inequality for different groups of countries. They concluded that the rise in agricultural openness in developing countries reduced domestic income inequality because opening up the agricultural sector raised the income of farmers who depend on agriculture for their livelihood. At the same time, unemployed farmers moved to manufacturing or service industries and their marginal output increased, thus increasing the income of the lower classes and reducing the domestic income gap. Further, Demir et al. (2012) used a sample of 55 developing countries from 1981 to 2005 to show that the impact of the trade structure 7 on income inequality within the country depends on the employment structure. When the manufacturing employment ratio is lower than a specific value, the increase in the proportion of manufacturing exports will increase the degree of income inequality within the country. When the manufacturing employment ratio is higher than a certain value, more workers will earn higher wages because of trade, thus narrowing the gap of income inequality. The literature emphasizes that if the expansion of trade openness is beneficial to the majority of the population, inequality will disappear; otherwise it will rise. 6 Under the Heckscher Ohlin (HO) model, an increase in trade restrictions will increase unemployment in labor-rich countries but reduce unemployment in capital-rich countries. 7 For global trade, the international trade structure refers to what products a country (or economy) exports to another country, commonly known as trade flows. In order to facilitate understanding, this paper defines the trade structure by the proportion of manufacturing exports to the exports of all products.

58 Xiaohua Bao, Xiaozhuo Wang / 51 71, Vol. 27, No. 1, 2019 From a micro perspective, according to the theory of firm heterogeneity, the size of export firms under trade liberalization would expand and employment would increase. Compared to non-export firms, export firms would pay higher wages, thus the wage gap in the industry would widen. Further, because workers are also heterogeneous, trade liberalization would exert an impact on workers wage inequality. Han et al. (2012) examined the impact of globalization on wage inequality using Chinese Urban Household Survey data from 1988 to 2008. The results showed that the Southern Tour speech 8 and WTO accession have increased inequality among China s regions, with the regions affected more by trade liberalization experiencing a greater wage rise. Autor et al. (2013) analyzed the increase in US import competition from China between 1990 and 2007, which reduced the wages of workers in the manufacturing industry. Chetverikov et al. (2016) found that trade shocks at commuting zone-level 9 particularly negatively affect the wages of workers in the bottom four wage deciles; thus income inequality among domestic workers has widened. In this literature, employment ratios are used as weights when examining the impact of China s import shocks on the wages of local US workers, indicating that the consideration of how many people benefit from trade plays an important role in examining the impact of trade on income inequality. 2. The Impact of Income Inequality on Trade Policy The decision-making process of trade policy can be summarized as follows: the effects of trade-induced income distribution encourage participants in political markets to generate demand and supply for new trade policies that ultimately reach equilibrium with tariffs as the price clearing the political market. The protection for sale model states that trade policy is a product for sale in the political market. Interest groups in each industry provide political contributions to the government to lobby for policies that are beneficial to them (Grossman and Helpman, 1995). Medium voter theory clarifies the way in which political parties gain political support, that is, the party that wins the medium voter will eventually win. The process of globalization has produced two 8 In 1992, Deng Xiaoping visited Guangzhou, Shenzhen, Zhuhai and Shanghai, where he delivered a series of speeches, emphasizing the importance of openness. After Deng s tour, the central and local governments began to reduce barriers to foreign trade and direct investment. Specific policies include the establishment of Pudong as a new open economic development zone, offering tax incentives to foreign investors, reducing the administrative procedures for foreign investment projects, and providing favorable import and export policies to foreign firms (Han et al., 2012). 9 Commuting zones, which encompass all metropolitan and nonmetropolitan areas in the US, are logical geographic units for defining local labor markets.

Globalization and Development of Regional Trade Agreements 59 groups: those who have benefitted and those who have been harmed. The struggle to distribute the benefits between the two groups affects the choice of trade policy. For example, Mayer (1984) applied the medium voter theory to the Heckscher Ohlin (HO) theory and concluded that the medium voters will get the optimal tariff level according to their interests. Further, Alesina and Rodrik (1994) quantified the study of income inequality from a theoretical point of view, suggesting that inequality can be regarded as the income gap between voters with a median capital labor ratio and those with an average capital labor ratio. Moreover, from an empirical point of view, they show that income inequality can be represented by the Gini coefficient. Dutt and Mitra (2002) adopted the variables of income inequality used by Alesina and Rodrik (1994) and found that rising income inequality increases tariffs in capital-rich economies, but decreases tariffs in capital-scarce economies. 3. The Influence Mechanism of Trade Liberalization on Trade Policy Based on HO and the intermediate voter theories, trade liberalization will affect the formulation of a country s trade policy by adjusting its internal income distribution. Using global tariffs, Gini coefficients and trade data from 1980 to 2013, we empirically examine the impact mechanism of trade liberalization on trade policy and further confirm our theoretical predictions. Figure 3 describes the relationship between trade liberalization and income inequality and Figure 4 the relationship between income inequality and tariffs in developing and developed countries. Among them, the index of trade liberalization is adjusted by the employment structure, defined as the proportion of a country s manufacturing exports multiplied by the proportion of manufacturing employment. Income inequality is measured by the Gini coefficient, and the country with a larger Gini coefficient has greater income inequality. Taking developing countries as an example, Figures 3(a) and Figure 4(a), respectively, show that the trade liberalization index of developing countries is negatively correlated with the Gini coefficient, and the Gini coefficient is positively correlated with tariffs. This means that as the proportions of manufacturing exports and manufacturing employment increase, the income inequality in developing countries declines (Demir et al., 2012), and thus these countries are more inclined to formulate lower tariffs. Similarly, as shown in Figures 3(b) and 4(b), respectively, the trade liberalization index of developed countries is positively correlated with the Gini coefficient, and the Gini coefficient is positively correlated with tariffs (Dutt and Mitra, 2002). As the proportions of manufacturing exports and manufacturing employment increase, the level of domestic income inequality in developed countries rises, and these countries tend to set higher tariffs.

60 Xiaohua Bao, Xiaozhuo Wang / 51 71, Vol. 27, No. 1, 2019 Figure 3. Relationship between Trade Liberalization and Income Inequality, 1980 2013: (a) Developing and (b) Developed Countries Sources: Trade structure and employment structure data were obtained from World Bank World Development Indicators. Gini coefficients were obtained from the World Income Inequality Database, version 3.3, September 2015. Notes: Trade liberalization is defined as the proportion of a country s manufacturing exports (in total exports) multiplied by the proportion of manufacturing employment (in total employment). Income inequality is measured by the Gini coefficient. Figure 4. Relationship between Income Inequality and Tariffs, 1980 2013: (a) Developing and (b) Developed Countries Source: Trade structure and employment structure data were obtained from World Bank World Development Indicators. Gini coefficients were obtained from the World Income Inequality Database, version 3.3, September 2015. Note: Income inequality is measured by the Gini coefficient. Trade liberalization, which has been affected by the employment structure, has reduced income inequality in developing countries, leading most people to benefit from globalization. This is because most developing countries have a large number of agricultural workers. An increase in manufacturing exports leads to labor transfer from

Globalization and Development of Regional Trade Agreements 61 agriculture to manufacturing where more workers earn higher wages because of trade, which increases the income of the lowest-income workers, and thus narrows the gap in income inequality. Further, although most developing countries are not democratic, when formulating trade policies the government still has to consider how many people will benefit from globalization, which in turn will lead to more open trade policies. However, developed countries are capital-rich, and their exports of manufacturing products are mostly capital-intensive products. According to the SS theorem, exports of manufacturing industries in developed countries will increase the income of workers using capital-intensive factors, while the income of workers using capital-scarce factors will decrease, thus widening the income gap. Global trade liberalization has affected the labor market and increased the unemployment rate of low-skilled labor in developed countries, rendering the situation of the lower class more difficult. Polarization between the political, business and intellectual elites and the low-educated working class is widening. The median voter approach to trade policy determination (within the HO framework), as in Mayer (1984), predicts that when the economy s overall relative endowment is maintained, rising inequality will increase trade barriers in developed countries. In general, countries around the world benefit from export trade, and are more inclined to support trade liberalization; therefore, the trend in globalization is irreversible. However, developed and developing countries have benefited differently from the last round of WTO-led economic globalization, leading to different attitudes toward a new round of globalization. After joining the WTO, China quickly integrated into the global market and the continuous expansion of exports has enabled China to gain dividends of economic globalization. Developing countries such as China have become the biggest beneficiaries of the WTO multilateral trading system. At the same time, the expansion of China s exports has intensified manufacturing unemployment in the EU and the US, which has widened the gap between rich and poor (Autor et al., 2013) and subsequently harmed the lower classes in the EU, the US and other countries. Thus, developed countries such as the US have shown a negative attitude toward the globalization dominated by the WTO multilateral trading system. They have called for reform of the international economic governance system and the adjustment of economic and trade rules to reshape the direction of economic globalization. III. Reshaping the Globalization Pattern The RTA is a complement to globalization, and an effective way to reshape globalization is to conclude RTAs. Since Viner s (1950) pioneering research, the study of global RTA development has formed three main streams. The first stream is centered on CU

62 Xiaohua Bao, Xiaozhuo Wang / 51 71, Vol. 27, No. 1, 2019 theory, with the motivation of RTA formation mainly discussed from economic factors. The second stream is represented by the theory of protection for sale, which discusses RTA conclusions from political factors. The third stream focuses on structural factors to explore the recent development of the RTA network. On the basis of reviewing the three research streams, we will explore the driving forces of the RTA network development from the perspective of hub spoke structure in the following. 1. The Impact of Economic Factors on the Formation of Regional Trade Agreements The pioneering theory that first proposed the explanation of RTA formation is the theory of CU (Viner, 1950). In CU theory, Viner proposed the concepts of trade creation and trade transfer. He expected that with a large number of member countries of similar economic size located in close proximity, the welfare benefits of trade creation will be greater than the losses caused by trade transfer, thus these countries find it easier to form RTAs. Kemp and Wan (1976) argued that in CUs, tariffs are adjusted externally to keep non-members welfare unaffected. Therefore, when members are allowed to adjust their external tariffs, neither the members nor the non-members will be negatively affected. These conclusions reveal win win conditions for the allies and outsiders. With the rapid development of international trade theory, the theoretical study of the welfare effect of CUs began to surpass the basic assumptions of perfect competition in the commodity and factor markets, and gradually introduced the assumption of increasing returns to scale and imperfect competition. This development overcomes the limitations of comparative advantage theory and focuses on product differences and trade benefits from economies of scale. According to Wonnacott and Lutz (1989), in order to reduce the effect of trade transfer, members of RTAs strive to reduce production costs and generate economies of scale. Although the CU theory provides a diversified perspective to analyze the effects of RTAs on welfare, its analytical method rests on partial equilibrium analysis, which focuses only on the welfare of the RTA members. From the perspective of general equilibrium theory, the relationship between RTA formation and the relative transportation cost, the national output and the relative factor endowment ratio can be explained. Baier and Bergstrand (2004) first systematically examined the impact of economic factors on the formation of RTAs, and laid the foundation for empirical research on RTA formation mechanisms. They found that it is easier for two countries to form an RTA when they are larger and more similar in economic size, have greater differences in factor endowments but smaller differences from the rest of the world, and are located closer to each other. The intrinsic mechanism is that governments

Globalization and Development of Regional Trade Agreements 63 are assumed to maximize their citizens economic welfare, and the net welfare gain or loss of two countries from forming an RTA depends on the trade creation versus trade diversion effects; when the former is greater than the latter, an RTA will be formed. Many scholars have verified these conclusions with panel data from a dynamic perspective (Egger and Larch, 2008; Baldwin and Jaimovich, 2012; Baier et al., 2014). 2. The Impact of Political Factors on RTA Formation The classic literature on RTA formation from the perspective of political economics is protection for sale by Grossman and Helpman (1995). The decision on tariffs depends not only on the welfare of all people, but also on the political contributions to the government (Grossman and Helpman, 1995; Krishna, 1998; Maggi and Rodríguez-Clare, 2007). In particular, Grossman and Helpman (1995) and Krishna (1998) suggest that an RTA is more likely to be formed when it enhances protection for member countries. However, the objective function in the protection for sale model only considers political contributions and national welfare, without considering the impact of electoral competition (Yu, 2009). In reality, the ruling party will inevitably seek more support in the election and make decisions that are beneficial to them. Therefore, the trade policy decision should include the politician s consideration of campaign support, and the election factor should be incorporated into the model of endogenous tariff decisions to better reflect the reality (Wang et al., 2011). Hillman (1989) argued that the government s objective function (political support function) is the sum of the welfare of workers and capitalists, and equilibrium tariffs can be obtained by maximizing the objective function. The results show that in countries with abundant capital, a left-wing government 10 will adopt import protection policies. In countries with abundant labor, the left-wing government often adopts a policy to reduce import protection. Dutt and Mitra (2005) confirmed Hillman s (1989) conclusions with a sample of 138 countries from 1980 to 1989. 3. Influence of Structural Factors on RTA Formation The signing of various cross-regional trade agreements has led more and more scholars to realize that global RTAs have entered a period of rapid networking development. The network means that the relationship between any two economies, i and j, is not only affected by them alone, but also by other economies. Therefore, the characteristics of interdependence are the key to network research. Egger and Larch (2008) confirmed that a past RTA will have a positive effect on future RTAs signed, and this promotion effect 10 This paper assumes that the government is left-wing, which usually supports workers as well as egalitarianism, while a right-wing government more actively supports capitalists.

64 Xiaohua Bao, Xiaozhuo Wang / 51 71, Vol. 27, No. 1, 2019 will decrease as the distance between the two countries increases. They divide RTAs interdependencies into domino and wave effects. The former focuses on new countries joining existing RTAs; the latter on forming new RTAs among non-member countries. Baldwin and Jaimovich (2012) obtained similar conclusions using spatial econometric methods. Which of these economic factors and structural factors have a greater impact on RTA formation? Chen and Joshi (2010) found that the interdependence of RTAs can explain most RTA formation, but much of this cannot be predicted by the economic characteristics of the countries. Regarding the relative importance of two sources of RTAs interdependences, Baier et al. (2014) used empirical data from 146 countries in 1960 2005 to find that the domino effect is greater than the wave effect. Baldwin s (2009) theoretical study examined the logic of the formation of hub spoke structure in RTA networks. He set up a hub spoke structure system containing one hub country and three spoke countries for cost benefit analysis. He showed that the attraction of the hub country to the spoke country comes from the potential gains of the spoke country after the formation of the RTA between the two countries, and the size of attraction comes from trade dependence. With advancement in economic globalization, the trade relations of various countries have also been woven into an inseparable global trade network. In this context, countries will not be limited to geographical location, economic size and other factors when signing RTAs more consideration will be given to the potential benefits of signing RTAs, especially trade benefits. There is a similar hub spoke structure in the global trade network, which may promote global RTA formation. The hub spoke structure can be regarded as a relatively stable structure in the trade network; whether a country can become a potential hub country can be measured by the index of hubness. Based on Dixit Stiglitz s monopolistic competition model, Baldwin (2009) measures the hubness of a country based on trade in goods: the attractiveness of any country to a partner country is the size of the potential gains in trade to the partner if an RTA is formed. The specific definition of the hubness of country j to country i is as follows: hub ji exportij import ji = (1 - ). (1) export import i j In Equation (1), j represents the importing country and i the exporting country. The first part represents the proportion of exports from country i to country j in total exports of country i, while the second part represents the proportion of country j s imports from countries other than i in total imports of country j. Economically, this means that the more important the export to country j is to country i, and the lower the relative

Globalization and Development of Regional Trade Agreements 65 importance of country i in the import of country j, the stronger the ability of country j to form a hub relative to country i. We measure and calculate the hubness of each country within three continents relative to the other countries in each continent for 1995 and 2015. Figures 5 7 depict the hubness of representative countries within Europe, America and Asia in 1995 and 2015. In 1995, the hub countries in Europe, America and Asia are Germany, the US and Japan, respectively; while in 2015, the hub countries are Germany, the US and China, respectively. Figure 5. Hubness of Selected Countries in America: (a) 1995 and (b) 2015 Source: Authors calculation based on data from UN Comtrade. Note: The value of hubness is calculated according to Equation (1). Figure 6. Hubness of Selected Countries in Europe: (a) 1995 and (b) 2015 Source: Authors calculation based on data from UN Comtrade. Note: The value of hubness is calculated according to Equation (1).

66 Xiaohua Bao, Xiaozhuo Wang / 51 71, Vol. 27, No. 1, 2019 Figure 7. Hubness of Selected Countries in Asia: (a) 1995 and (b) 2015 Source: Authors calculation based on data from UN Comtrade. Note: The value of hubness is calculated according to Equation (1). Since the growth rate of service trade has surpassed the trade of goods in recent years, the focus of negotiations on global trade rules has gradually shifted from the field of goods trade to service trade. For example, the purpose of the Trade in Service Agreement is to establish rules for service trade liberalization that reflect 21st century demands in trade areas. Whether it is the Regional Comprehensive Economic Partnership, the China Japan Korea Free Trade Area and other multilateral FTA negotiations, or the China US, China EU and other bilateral investment agreement negotiations, all involve issues of service trade openness and related issues of service industry market access. Since 1990, more than half of the global RTAs signed have been service RTAs. 11 After accession to the WTO and by the end of 2015, China had signed 14 RTAs, all of which are service RTAs. If the early RTA network was established based on the network relationship of goods trade, it can be expected that the service trade network will play an increasingly important role in the construction of current and future RTA networks. Similar to the hubness of goods trade, we measure the hubness of service trade in representative countries of the three continents. As shown in Figure 8, we find that the US is the hub of the American trade both in services and goods. In Europe, Germany also enjoys hub status both in service and goods trade. However, in Asia, the largest hub country in goods trade is China, while the largest hub country in service trade is Japan, and China s hub status in service trade ranks third. 11 The service RTA defined in this paper refers to the RTA on service trade, including only service trade or trade in both goods and services.

Globalization and Development of Regional Trade Agreements 67 Figure 8. Top Five Countries in America, Europe and Asia by Hubness in Goods and Service Trade in 2015: (a) Goods in America, (b) Service in America, (c) Goods in Europe, (d) Service in Europe, (e) Goods in Asia and (f) Service in Asia Source: Authors calculation based on data from UN Comtrade. Note: Hubness is calculated according to Equation (1). Hubness can reflect the relative position of a country in the trade network. However, the question we are interested in is: What is the relationship between hubness and RTA formation? In order to answer this question, in Figure 9, we use our results to draw a scatter plot between hubness and RTA formation in goods and service trade. The horizontal axis represents the average hubness of a country relative to all other countries and is standardized. The vertical axis represents the cumulative number of RTAs signed by a country. As can be seen in Figure 9, there is a positive correlation between hubness and RTA formation. This reflects that the greater the hubness, the more likely RTAs will be formed. Because of the self-reinforcing effect of the hub country, its trading partners

68 Xiaohua Bao, Xiaozhuo Wang / 51 71, Vol. 27, No. 1, 2019 tend to be concerned about being marginalized in the network, and thus are inclined to join an existing RTA or organize a new RTA to compete with existing ones, both of which have promoted the global expansion of RTAs. Figure 9. Relationship between Hubness and RTA Formation: (a) Goods and (b) Service Trade Sources: Regional Trade Agreement (RTA) data obtained from WTO; trade data obtained from UN Comtrade, World Development Indicators and the WTO. Notes: The number of RTAs in goods trade is the cumulative sum during 1973 2015. For service trade, it is from 1985 to 2015. The data of hubness in goods are from 1973 to 2015, while hubness in service data are from 1985 to 2015. IV. Conclusions and Policy Implications This paper combines the relevant research conclusions and current stylized facts to examine the evolution and reshaping of globalization. We find that: (i) because of the uneven distribution of benefits in the process of globalization, countries have different attitudes toward the recent round of globalization. In the process of global trade liberalization, the interests of certain groups in some developed countries have been harmed and they are showing signs of rising protectionism. Developing countries have significantly benefited from trade liberalization and thus support globalization. (ii) Regional economic development is an effective way to reshape globalization. The self-strengthening effect of the hub country in the trade network has promoted global RTA expansion. In view of this, countries face challenges to improve their hubness in the context of globalization evolution and deal with the rise of trade protectionism. We propose the following policy recommendations to address these concerns. First, countries should expand their total imports and at the same time increase the diversity of import sources, thus reducing the constraint by a particular import

Globalization and Development of Regional Trade Agreements 69 source country. When countries expand exports, avoiding excessive concentration of export destinations will also help to increase the countries hubness. In the dynamic development of the global trade network, countries should build their own hub spoke network system, and seek to strengthen their hub status. Therefore, taking the initiative in trade and investment liberalization has great practical significance for developing countries. Second, when deepening the reform of external trade liberalization, countries should also promote the reform of internal marketization. In particular, developing countries should facilitate development in their trade and employment structures, and ensure that the employment creation effect of trade liberalization plays a more effective role. Third, countries can deal with the rise of trade protectionism by concluding RTAs and work to build a new trade relationship with the aim of cooperation. The trade disputes between the US and China can be seen as a big country game. However, non-cooperative games under asymmetric information often lead to the prisoner s dilemma. In fact, the gains from trade disputes are less than gains from cooperation between the two countries. Particularly in the context of global value chains, global market dependence is deepening, and changes in the interests of either country will affect the interests of the other. Thus, countries should negotiate to avoid the termsof-trade driven prisoner s dilemma, and the US and China need to establish a new type of major-country relationship to achieve the best outcome in competition and cooperation. References Alesina, A. and D. Rodrik, 1994, Distributive politics and economic growth, The Quarterly Journal of Economics, Vol. 109, No. 2, pp. 465 90. Autor, D. H., D. Dorn and G. H. Hanson, 2013, The China syndrome: Local labor market effects of import competition in the United States, American Economic Review, Vol. 103, No. 6, pp. 2121 68. Baier, S. L. and J. H. Bergstrand, 2004, Economic determinants of free trade agreements, Journal of International Economics, Vol. 64, No. 1, pp. 29 63. Baier, S. L., J. H. Bergstrand and R. Mariutto, 2014, Economic determinants of free trade agreements revisited: Distinguishing sources of interdependence, Review of International Economics, Vol. 22, No. 1, pp. 31 58. Baldwin, R. E., 2009, The Spoke Trap: Hub and spoke bilateralism in East Asia, NCCR Trade Working Paper No 2009/28, Swiss National Center for Competence in Research (NCCR),

70 Xiaohua Bao, Xiaozhuo Wang / 51 71, Vol. 27, No. 1, 2019 Geneva. Baldwin, R. E. and D. Jaimovich, 2012, Are free trade agreements contagious? Journal of International Economics, Vol. 88, No. 1, pp. 1 16. Barro, R., 2000, Inequality and growth in a panel of countries, Journal of Economic Growth, Vol. 5, No. 1, pp. 5 32. Bergstrand, J. H., P. Egger and M. Larch, 2016, Economic determinants of the timing of preferential trade agreement formations and enlargements, Economic Inquiry, Vol. 54, No. 1, pp. 315 41. Bhagwati, J., 2002, The wind of the hundred days: How Washington mismanaged globalization, Asian-Pacific Economic Literature, Vol. 16, No. 1, pp. 169 71. Chen, M. X. and S. Joshi, 2010, Third-country effects on the formation of free trade agreements, Journal of International Economics, Vol. 82, No. 2, pp. 238 48. Chetverikov, D., B. Larsen and C. Palmer, 2016, IV quantile regression for group-level treatments, with an application to the distributional effects of trade, Econometrica, Vol. 84, No. 2, pp. 809 33. Demir, F., J. D. Ju and Y. Zhou, 2012, Income inequality and structures of international trade, Asia-Pacific Journal of Accounting & Economics, Vol. 19, No. 2, pp. 1 14. Dutt, P. and D. Mitra, 2002, Endogenous trade policy through majority voting: An empirical investigation, Journal of International Economics, Vol. 58, No. 1, pp. 107 33. Dutt, P. and D. Mitra, 2005, Political ideology and endogenous trade policy: An empirical investigation, The Review of Economics and Statistics, Vol. 87, No. 1, pp. 59 72. Dutt, P., D. Mitra and P. Ranjan, 2009, International trade and unemployment: Theory and crossnational evidence, Journal of International Economics, Vol. 78, No. 1, pp. 32 44. Egger, P. and M. Larch, 2008, Interdependent preferential trade agreement memberships: An empirical analysis, Journal of International Economics, Vol. 76, No. 2, pp. 384 99. Grossman, G. M. and E. Helpman, 1995, The politics of free trade agreements, American Economic Review, Vol. 85, No. 4, pp. 667 90. Han, J., R. J. Liu and J. S. Zhang, 2012, Globalization and wage inequality: Evidence from urban China, Journal of International Economics, Vol. 87, No. 2, pp. 288 97. Hillman, A. L., 1989, The political economy of protection, Forum for Social Economics, Vol. 10, pp. 86 100. Jaumotte, F., S. Lall and C. Papageorgiou, 2008, Rising income inequality: Technology, or trade and financial globalization? IMF Working Paper WP/08/185, International Monetary Fund, Washington DC. Kanbur, R. and X. B. Zhang, 2005, Fifty years of regional inequality in China: A journey through central planning, reform and openness, Review of Development Economics, Vol. 9, No. 1, pp. 87 106. Kemp, M. C. and H. Y. Wan Jr., 1976, Elementary proposition concerning the formation of

Globalization and Development of Regional Trade Agreements 71 customs unions, Journal of International Economics, Vol. 6, No. 1, pp. 95 7. Krishna, P., 1998, Regionalism and multilateralism: A political economy approach, The Quarterly Journal of Economics, Vol. 113, No. 1, pp. 227 51. Maggi, G. and A. Rodríguez-Clare, 2007, A political-economy theory of trade agreements, American Economic Review, Vol. 97, No. 4, pp. 1374 406. Mayer, W., 1984, Endogenous tariff formation, American Economic Review, Vol. 74, No. 5, pp. 970 85. Milanovic, B., 2016, Global Inequality: A New Approach for the Age of Globalization, Cambridge, MA: Harvard University Press. Viner, J., 1950, The Customs Union Issue, New York: Carnegie Endowment for International Peace. Wang X. S., K. W. Li and S. X. Xie, 2011, How is trade policy formulated: An endogenous protection model that includes political contributions, campaign support, and power delegation, Shijie Jingji (The Journal of World Economy), No. 10, pp. 107 26. Wonnacott, P. and M. Lutz, 1989, Is there a case for free trade areas? in J. Schott, ed, Free Trade Areas and U.S. Trade Policy, Washington DC: Institute of International Economics, pp. 59 84. Yu, M. J., 2009, Political Economic Analysis of International Trade: Theoretical Models and Econometric Empirical Studies, Beijing: Peking University Press (in Chinese). (Edited by Yuanfang Li)