EXPLORING CHINA S STRATEGIES IN EXPANDING ITS FREE TRADE AGREEMENT(FTA) NETWORK

Transcription

1 EXPLORING CHINA S STRATEGIES IN EXPANDING ITS FREE TRADE AGREEMENT(FTA) NETWORK A Thesis submitted to the Faculty of the Graduate School of Arts and Sciences of Georgetown University in partial fulfillment of the requirements for the degree of Master of Public Policy in Public Policy By Xiran Liu, B.A. Washington, D.C. April 12, 2018

2 Copyright 2018 by Xiran Liu All Rights Reserved ii

3 EXPLORING CHINA S STRATEGIES IN EXPANDING ITS FREE TRADE AGREEMENT(FTA) NETWORK Xiran Liu, B.A. Thesis Advisor: Andreas Kern, Ph.D. ABSTRACT This paper discusses potential factors that lead China to engage in Free Trade Agreement (FTA) with certain partner countries. Previous literature has linked a country s interest in forming an FTA with a variety of internal, competitive, and emulative calculations. While multilateral trade arrangements, such as TPP and NAFTA, are suffering from skepticism and protectionist policies, China has taken positive steps to incorporate other countries in its FTA framework. My study explores whether China s likelihood of forming FTAs with certain countries is impacted by a group of factors, including bilateral trade relationship, natural resource abundance, openness to trade, political alignment in international affairs, and quality of governance. An analysis of available data sources shows that multiple factors can help explain China s tendency to form FTA partnership with certain countries. In addition, through a case study of the newly concluded China-Maldives FTA, this paper makes a preliminary attempt to explain how China integrates its FTA strategy into its broader geopolitical plan to assume global influence under new circumstances. iii

4 I would like to express my gratitude to my supervisor Dr. Andreas Kern for the useful comments, remarks and engagement through the learning process of this master thesis. I am also grateful to my loved ones, family and friends, who have supported me throughout entire process, both by keeping me encouraged and providing me with constructive feedbacks. AMOR EST VITAE ESSENTIA Xiran Liu iv

5 TABLE OF CONTENTS Introduction... 1 Background... 3 Literature Review... 8 Theoretical Model Data Empirical Model Results Conclusion and Policy Implications References v

6 LIST OF TABLES Table 1: China's FTA network (as of March Table 2: Outcome Variable and Explanatory Variables Table 3: Summary Statistics Table 4: Baseline Model Table 5: Disaggregate the Quality of Governance Indicator Table 6: Variations of the Dependent Variable LIST OF FIGURES Figure 1. China's Free Trade Agreement (FTA) Network... 7 vi

7 INTRODUCTION The past two years have witnessed Brexit, the Trump Administration s withdrawal from the Trans-Pacific Partnership (TPP) and the renegotiation of North American Free Trade Agreement (NAFTA), all in part attributed to the skepticism towards trade liberalization and preferential trade agreements (PTAs). Though it is widely recognized that PTAs--mainly, free trade agreements, and customs unions--have effectively advanced trade liberalization among a subset of countries and breaking deadlocks in the multilateral negotiations, there have been emerging concerns over whether they generate net welfare gains for the parties involved. Amid an uncertain global economy, China, though being a latecomer, has moved fast and continues to maintain its position in promoting free trade worldwide. A key component of this effort involves the free trade agreement (FTA) negotiations with other economies. From the initial version of ASEAN-China Free Trade Area (ACFTA) in 2002 to the most recent one with Maldives in 2017, China has signed 13 cross-country FTAs with a wide range of geographically and developmentally diverse countries and regional economic blocs. In the meantime, a great deal of joint feasibility research for future FTAs is under way. From an analytical perspective, the main question this paper hopes to answer is: how does China select its partners for FTAs? In particular, are there economic, political and geographical factors that China specifically consider when choosing/selecting FTA partners? My study will examine considerations that motivate China s overall FTA strategy and the selection of its partners. Employing a wide range of country-specific variables for the period that covers the whole range of China s FTA network building, I develop an empirical model for FTA formation between China and potential partner countries. Controlling for year dummies, the empirical results show strong 1

8 support for my hypothesis: that the probability of FTA formation increases as the partner country is geographically closer to China, has a relatively large economy, is more export dependent on China, and is sufficiently ranked in the regulatory quality and governance effectiveness. The findings remain robust to several model specifications. My paper proceeds as follows. In Section 2, I briefly review China s trade policy over time and the growth of its cross-country FTA network since In Section 3, I conduct a literature review on trade liberalization and the proliferation of free trade agreements, driving forces of a country s FTA interest, as well as China s underlying calculations in its FTA strategy. In Section 4, I present the main argument and derive a group of hypotheses of FTA formation on geographic distance, economic size, bilateral trade relation, natural resource abundance, political alignment with Beijing in international affairs, and quality of governance. In Section 5, I explain the data sources and measurements of my outcome variable of interest and covariates. Then, I discuss the model I use to fit the data, report the findings, and conduct robustness checks. Section 6 discusses my conclusions and their policy implications. 2

9 BACKGROUND Overview of China s trade policy China is one of the latest in the Asia-Pacific region to integrate itself into the global economic system. For three decades after the establishment of the PRC in 1949, Beijing mainly followed the former Soviet Union s model and by 1978 import substitution had been a longstanding element of Chinese trade policy (Sheng, 2015). Due to the state s "self-sufficiency"-themed trade planning, exports were heavily controlled and imports were only seen as a response to domestic shortages. During that period, the most populous Asian country remained a negligible player in the global market. The situation changed with the "Reform and Opening-up" policy initiated by Deng Xiaoping in Beginning then, via a series of experimental practices and trade reforms, China has gradually shifted its development strategy towards more export-led growth. As the pace of trade liberalization accelerated, China s interest in joining the regional and multilateral economic cooperation increased. Its accession to the WTO in 2001 after 15 years back-and-forth negotiation process is one of the most noticeable achievements. China s transition towards a more open economy has been rewarding. From 1978 through 2016, the total trade volume grew more than 179 times, from US $20.6 billion to US $3.686 trillion. China is now the world s largest exporter and the second largest importer of goods. After the WTO entry, as Beijing continues its promotion of free trade within the multilateral scheme, it seems to be increasingly involved in regional and bilateral trade agreement negotiations. Policymakers acknowledge that, compared with multilateral mechanisms such as the WTO, smaller-scale trade frameworks enable countries to achieve consensus on entities, procedures and 3

10 implementation of competition laws (Su, 2015). As the stagnation of WTO talks leads more countries to conduct FTA negotiations, China has been learning the international economic game and actively catching up with the trend. China s FTA network Over the past 15 years, China s FTA network building has been remarkable. As of March 2018, China had 12 cross-country FTAs in effect [ASEAN, Chile, Pakistan, New Zealand, Singapore, Peru, Costa Rica, Iceland, Switzerland, South Korea, Australia, and Georgia]; it had recently concluded the preferential trade agreement with Maldives; and it was in active negotiations with another eight countries and regional blocs [RCEP, Gulf Nations, Japan-Korea, Sri Lanka, Israel, Norway, Moldova, and Mauritius]. Table 1, below, shows that, China s FTA partnerships are quite diversified in terms of developmental stage, size of economy, distance, and trade volume. While a large proportion of countries in the current network are in Asia and Oceania, China also has more distant trade partners such as Peru and Chile (See Figure1.). Many of China s non-asian/oceanian partners, current or potential, also have preferential trade agreements with EU or the United States. Traditionally resource-rich countries are well targeted, Peru and Chile for copper, India for iron ore and bauxite, Gulf Nations for oil and gas, and Australia for a wide range of natural resources, which reflects the country s security agenda and the spread of its resource imports need (Wilson, 2012). As for the import and export partner share, some of the FTA partners have outstanding trade relationships with China but others are just less significant. 4

11 (Continued) 5 Table 1: China's FTA network (as of March 2018 Partner Region Status Partner Share (%) 1 FTA with other economies Export Import EU U.S. ASEAN Southeast Asia In force (2005) X X Chile South America In force (2006) Y (2003) Y (2004) Upgrade (negotiation) Pakistan South Asia In force (2007) X X Second phase (negotiation) New Zealand Oceania In force (2008) X X Upgrade (negotiation) Singapore Southeast Asia In force (2009) X Y (2004) Upgrade (negotiation) Peru South America In force (2010) Y (2013) Y (2009) Upgrade (study) Costa Rica Central America In force (2011) Y (2013) Y (2006) Iceland Northern Europe In force (2014) Y (1973) X Upgrade (study) Switzerland Western Europe In force (2014) Y (1973) X Korea East Asia In force (2015) Y (2011) Y (2012) Australia Oceania In force (2015) X Y (2005) Georgia Eastern Europe Signed (2017) Y (2014) X Maldives Southeast Asia Signed (2017) X X RCEP Asia & Oceania Negotiation (since 2013) X X GCC Middle East Negotiation (since 2004) X X Japan-Korea East Asia Negotiation (since 2012) X X Sri Lanka Southeast Asia Negotiation (since 2014) X X Israel Middle East Negotiation (since 2016) Y (2000) Y (1985) Norway Northern Europe Negotiation (since 2007) Y (1973) X Moldova Eastern Europe Negotiation (since 2017) Y (2014) X Mauritius Africa Negotiation (since 2017) X X Panama Central America Study X Y (2012) Colombia South America Study X Y (2012) Fiji Oceania Study Y (2009) X 1 Source: World Integrated Trade Solution (2016 data)

12 Partner Region Status Partner Share (%) 1 FTA with other economies Export Import EU U.S. Nepal South Asia Study X X Papua New Guinea Oceania Study Y (2009) X Canada North America Study Y (2017) Y (1994) Bengal South Asia Study X X Mongolia East-Central Asia Study X X 6

13 New Zealand China s FTA network building for less than 15 years has been remarkable. As of January 2018, it has signed 13 FTAs with countries and regional bodies geographically and developmentally diverse, covering Asia, Central and South America, Oceania, and Europe. Source: geodistance dataset, Centre d'etudes Prospectives et d'informations Internationales China s FTA network building for less than 15 years has been remarkable. As of January 2018, it has signed 13 FTAs with countries and regional bodies geographically and developmentally diverse, covering Asia, Central and South America, Oceania, and Europe. Iceland Switzerland Georgia Pakistan China Costa Rica ASEAN countries Maldives Singapore Peru 7 Chile Australia Figure 1 Source: China's geodistance Free Trade dataset Centre Agreement d'études (FTA) Prospectives Network et d'informations Internationales

14 LITERATURE REVIEW Trade liberalization and the proliferation of free trade agreements A free trade agreement (FTA) is an arrangement which establishes unimpeded exchange and flow of goods and services between trading partners regardless of national borders (Hinkelman, 2011). Dissatisfied with the protracted multilateral negotiations in the World Trade Organization (WTO), more countries have begun to consider FTAs as an alternative approach to deepening the liberalization of trade. The last two decades have witnessed rapid growth in the number of FTAs. With more than 400 in force so far, it is estimated that virtually every country in the world is now a partner in one or several such agreements (Stefanidou, 2012). Apart from the lack of progress in the multilateral talks, there are other explanations for the popularity of FTAs. Two commonly cited explanations are the spread of democracy globally and the quest for geopolitical stability (Mansfield, Milner, & Rosendorff, 2002; Martin et al., 2008; Vicard, 2012). Adding the third-nation effects, the "domino effect" proposed by Baldwin suggests the presence of contagion in the FTA formation: the signing or deepening of one FTA can induce excluded nations to sign their own FTAs, and this second-round of FTAs may in turn create their own trade diversion that may lead to more FTAs (Baldwin, 1993; Baldwin & Jaimovich, 2012). Forces driving a country s FTA interest Countries seek to negotiate trade agreements for a wide range of reasons. While there is no simple explanation behind this, current literature on regional FTA proliferation features three distinct hypotheses: independent decision-making, emulation, and competition: Independent decision-making suggests that a country launches an FTA policy due to internal considerations, either to seek a domestic political realignment or to gain a positive economic impact. On the one hand, preferential liberalization creates incentives for private sector 8

15 lobbies and governments to achieve competitive goals through the agreements (Urata, 2009). Governments need to balance the interests of different groups before an FTA negotiation starts. On the other hand, the economic impacts of an FTA are not unidirectional, if decomposed into "trade creation" and "trade diversion". The government will form an FTA if the perceived benefits exceed the costs (Baldwin & Venables, 1995; Urata, 2009). Emulation is the method in which governments embrace an FTA track following their socio-cultural peers or leading nations (Solís & Katada, 2009). Especially for risk-averse countries, emulation is a wise way to quickly expand the FTA network. For one thing, similar countries previous experiences will provide lessons about forming the country s FTA negotiation strategy. For another, big nations successful examples may serve as a reference and effectively reduce the cost of FTA policy making. Competition is the most popular frame used in analyzing a country s FTA interest. Countries will use FTAs as a means of competing with their economic rivals, to help them gain access to markets and capital, improve their advantage in the struggle for status and leadership, and help them challenge the existing norm of economic integration with alternative approaches (Solís & Katada, 2009). Compared with the emulative strategy, competing requires a government to be more proactive and selective. Economic, political, legal, and security calculations are normally included in the analysis, but the mapping of them is very context-specific. For example, to explain Chilean FTAs, Stallings (2008) compared Chile with three sets of competitors: Mexico, South American countries, and emerging markets, mainly in Asia, focusing on political and economic stability, good policies, openness, and the trade links with the rest of the region. In the case of the KORUS FTA, Koo (2008) pointed out that in addition to the goal of maximizing the gains from trade and investment, South Korea s competitive FTA strategy under President Roh 9

16 Tae Woo was to hedge against the growing uncertainties in Northeast Asia by cementing its security-embedded economic ties with the United States. Over time and among partners, it is hard to find a country that fits neatly to any of the above hypotheses. In some cases, a country might shift its strategy from one to another as its FTA network expands and its negotiation experiences accumulate. In other cases, these motivations might not be mutually exclusive. Countries must weigh costs and benefits of each partnership agreement before making a decision. China s story China's Ministry of Commerce (MOFCOM) states that "the Chinese government deems free trade agreements as a new platform to further opening up to the outside and speeding up domestic reforms, an effective approach to integrate into global economy and strengthen economic cooperation with other economies, as well as particularly an important supplement to the multilateral trading system." 2 Instead of referencing to economic growth or job creation, the government s statement places greater emphasis on FTAs role in securing domestic reforms, though clearly, this is not the single motivation that drives China s interest. Unlike many large trading economies such as Japan, Korea, and the U.S., China hasn't set any official FTA policies. This absence leaves room for surmises concerning the country's objectives and strategies. Scholars have made several attempts to identify the calculations behind Beijing s FTA network building. Reviewing China's first decade of free trade, Zhao and Webster (2011) conclude that China's FTA strategy focuses more on economic considerations than domestic interests or foreign policy. The chief objective of China s FTA strategy, they argue, is to 2 MOFCOM: accessed April 10,

17 maintain the nation s rapid economic growth (Zhao & Webster, 2011). Yang (2008) considers that China's FTA strategy, either the selection of partners or the speed of negotiations, is determined by the country s effort to compete for realistic relative gains, including economic, political, and legal competition within regional and multilateral frameworks. China's geo-political aspirations may also influence China's FTA calculations. Sigurðsson (2014) summarizes three geo-political factors that might drive China s pursuit of free trade: to gain the favor of its neighbors and prove the peaceful nature of its rise; to compete with East Asian rivals, notably Japan and the U.S.; and to be in a better position to counter U.S. influence and unilateralism. Other perspectives such as value chain dynamics and political signaling are also mentioned to explain China s FTA strategy. The existing theoretical literature has already offered a range of explanations for China s preferential trading agenda, but the discussions are still limited to the qualitative measurement, and there is no systematic research on the micro foundations of FTA partner choice. This study aims to add to the literature on China s calculations. It uses regression methods to empirically test the hypotheses regarding considerations behind China s FTA network building, and particularly its choice of FTA trading partners. Besides, through a case study of the newly concluded China- Maldives FTA, this paper also makes a preliminary attempt to explain how China integrates its FTA strategy into its broader geopolitical plan to assume global influence under new circumstances. 11

18 THEORETICAL MODEL The selection of preferential trading partners normally reflects the multidimensional consideration beneath a country s FTA strategy. Based on the above survey of literature, I apply the competition and emulation hypotheses to investigate China s FTA strategy and try to identify characteristics that could determine the Asian country s partner preference. From the competitive perspective, we can consider a country s trade relations with China, political alignment with Beijing in international affairs, abundance of natural resources, and quality of governance as influential factors in the FTA formation. As for emulative calculations, in its initial stage of FTA partnership building, China would be inclined to negotiate preferential trading proposals with countries that are more open to trade. As can be observed in Table 1, when China is pursuing free trade relationships beyond the Asia-Pacific region, countries that already have trade agreements with big economies, such as EU and the U.S., are easily targeted. I set the country s FTA status with China as the dependent variable, measured by whether a country has an FTA with China or not. Thus, a mathematical explanation of the initial theoretical model goes as follows: FTA status with China= F (bilateral trade relations, natural resource abundance, governance quality, political alignment with China, trade openness) The hypotheses to test regarding China s choice of FTA partners are that China would prefer FTAs with countries that have closer trade ties with it; are more open to trade; have more natural resources; have better governance; and share more similar stances in the international affairs. Below I further develop each of these hypotheses. 12

19 Bilateral trade relations We expect that higher trade interdependence between China and potential partner countries is associated with a higher probability of FTA formation. The bilateral trade relations relate to the FTA via two mechanisms. First, FTA s cost reduction implications. With large amounts of trade happening between two countries, exporters of both sides often face a lot of trade barriers and lengthy compliance procedures. The formation of a trade agreement would eliminate tariff rates and lower non-tariff barriers, effectively reducing related administrative and transaction costs. Therefore, two countries who already have a higher level of trade relations would be more motivated to advance trade liberalization and further secure access to the key export markets. Second, good will to address the trade disputes. Closer trade ties tend to have positive effects on trade disputes, as countries use its constrained resources more to launch disputes against its main export partners (Bown, 2005; Li & Qiu, 2014). This also creates incentives for countries to resort to FTAs, especially the dispute settlement mechanism contained, to address the problem. Thus we can assume a positive correlation between bilateral trade relations and the probability of FTA ratification. I employ two pairs of variables to measure bilateral trade relations, one based on the gravity model, and the other of trade share. Economic size and distance: The gravity model 3 is the most commonly applied econometric technique to predict bilateral trade flows(tinbergen, 1962). Based on an analogy of the law of gravity in physics, it relates trade between two countries to their economic size (i.e. average real GDPs) and the distance between them. According to the model, pairs of countries 3! "# = % & () ' & (+ *, where F is the trade flow, M is the economic mass of each country, D is the distance and G, (- '* is a constant. 13

20 trade more if they are closer to each other in terms of geographic distance and if their economies are larger. Import and export partner share: The import partner share is the percentage of imports from the country of interest to the country under study (in this case, China) in the total imports of the destination. The export partner share is the share of exports to the country accounting for China s total export. The two variables represent the relative importance of the bilateral trade relationship between two countries. Openness to trade We expect that a country s openness to trade would also influence China s choice of trading partners, mainly due to the emulative calculations and hub effect. Some countries that already have preferential trade relations with big economies can be easily targeted especially when China is pursuing free trade relationships beyond the Asia-Pacific region. For one thing, China can borrow experiences from the big economies previous trade negotiations and more efficiently reach the deal. For another, these countries are often members of other preferential trading networks, making the spoke effect more important than their domestic market size in the future trade relations. By forming FTAs with these partners, China could potentially join the bigger markets created by the preferential trading arrangements already in place, which is a highly cost-effective way of exploring new markets (Kong, 2012). We use the trade openness variable to capture this calculation. Trade openness: A common way to measure how much one country relies on international trade is the trade openness index, which adds imports and exports in goods and services and divides this sum by GDP. The larger the ratio, the more the country is exposed to international trade. 14

21 Natural resource abundance For resource-importing states, obtaining preferential resource access is an important consideration in their FTA agenda, as a continuous and stable supply of needed resources at a reasonable price is essential to the long-run of their economic health (Wilson, 2012). We expect a positive relationship between a partner country s natural resource abundance and the possibility that China forms an FTA with it. From an internal consideration, a resource-related FTA can help address the gap between domestic supply and demand of natural resources in China (Yang, 2008). With a large population, China s per capita possession of natural resources is far lower than the world average, but its demand for resources rapidly increases at the fascinating speed of modernization and industrialization. A competitive dynamic to this motive is also at play among regional peers. China, Japan, and South Korea all rely their economic growth on the resource security. To avoid being disadvantaged by the concessions obtained by others, each government attempts to best, or at least match, the others in negotiations of resource-related FTAs. I use the variable "Total Natural Resource Rents to GDP" to measure a country s natural resource abundance. Natural resource rents (% of GDP): an indicator that measures total natural resources rents as a share of the gross domestic product (GDP) of a given country. Total natural resources rents are the sum of oil rents, natural gas rents, coal rents (hard and soft), mineral rents, and forest rents. Quality of governance The quality of governance plays an important role in the formation of FTAs through two main channels: the internal checks and balances, and the credibility of the government s policy. In any regime type, political leaders depend on the support of constituents to stay in power. Therefore, the negotiation process of a trade agreement reflects the balancing of diverse domestic interests, 15

22 especially of those major players. The more the checks and resistance from different stakeholders and sectors, the weaker the incentives of a government to ratify an FTA (Mansfield & Milner, 2010). We expect that the level of checks and balances in a country has negative effects on its willingness to negotiate a deal with other countries. In the meantime, regarding trade, a country s institutional quality normally reflects its credibility and effectiveness both in the goods and service provision and in the legal enforcement. Thus we assume that one country s governance quality is positively correlated with the other country s willingness to build a preferential trading relationship with it. I adopt the Bank s Worldwide Governance Indicators for the measurement purpose. Quality of governance: I use the six indicators of good governance developed by the World Bank to measure the quality of governance: Voice and accountability (VA) political, civil and human rights, Political stability and absence of violence (PV) likelihood of violent threats to, or changes in, government, Government Effectiveness (GE) competence of bureaucracy and quality of public service delivery, Regulatory quality(rq) incidence of market-unfriendly policies, Rule of Law (RL) quality of contract enforcement, police, and courts, and likelihood of crime and violence, and Control of Corruption (CC) exercise of public power for private gains. Political alignment in international affairs I argue that countries with more political affinity with Beijing in international affairs are more likely to form FTAs with China. Economic and trade interdependence is strongly correlated with international politics; the higher benefits two countries can gain from trade, the higher the costs they will bear if engaging in a potential conflict (Einstein, 2017). Therefore, it is reasonable to assume that two countries are more willing to deepen their trade relationships (i.e. the formation of preferential trading arrangements) if they are politically aligned in international affairs, because 16

23 the latter helps ensure the continuity of the arrangement enforcement. I adopt the UN General Assembly voting records in line with China to capture this consideration. UNGA voting records in line with China: The variable "Percentage share of votes aligned with China" is an index of agreement between the state and the China in the UN general assembly voting, which equals 1 if a country always agrees with China, and 0 if it always votes against. 17

24 DATA For the empirical study, I construct a comprehensive dataset of country-specific variables for the period from 2002 to 2015, which covers the whole range of China s FTA network building. Dependent Variable I use the information of China s FTA network from MOFCOM to build the dichotomous dependent variable of free trade agreements (FTA). The dummy FTA is equal to 1 if two countries signed an FTA before January 2018 and 0 otherwise. As the study looks at ex ante factors, there is no variation of the variable s value before and after the signing of agreements. Independent Variable(s) The independent variables of interest consist of each country s profile in terms of bilateral trade relations with China, economic performance, natural resource rents contribution to the GDP, openness to trade, and quality of governance. They are from various data sources. CEPII provides a gravity dataset for world pairs of countries, including GDP and capital-to-capital distance information. The World Integrated Trade Solution (WITS) database of the World Bank provides information about the trade partner share (i.e. import partner share & export partner share) between each two countries. To estimate a country s governance quality, I adopt the Bank s Worldwide Governance Indicators data, which reports the countries performance in six different dimensions of governance. Using these data, I also generate an aggregated quality of governance variable "qgov". The ideal point estimates from the UNGA Voting Data provides me with the index of agreement between the country and China in the UN General Assembly voting, the "PctAgreeChina". In addition, the Bank s indicators "Trade (% of GDP)" and "Total natural 18

25 resource rents (% of GDP)" are also used to measure a country s openness to trade and richness of natural resources. A summary of variables is shown in Table 2. Table 2: Outcome Variable and Explanatory Variables Notation Description Measurement Sources A dummy variable which equals 1 if China FTA A country s FTA status and the counterpart share with China by Jan a Free Trade Agreement and equals 0 otherwise by Jan logdist Distance between China and the host country (capital-to-capital) In kilometers China Ministry of Commerce CEPII distance dataset loggdp Log of GDP ln (US dollars) CEPII gravity dataset Importpartnershare Exportpartnershare The share of imports from the country accounting for China s total imports The share of exports to the country accounting for China s total export In percentage In percentage World Integrated Trade Solution World Integrated Trade Solution natural The sum of oil, natural gas, coal (hard and soft), mineral, and forest rents contribution to GDP Percentage of GDP World Bank vae, pve, gee, rqe, rle, cce qgov tradeopenness PctAgreeChina Individual governance indicators for six dimensions of governance An aggregated measurement of the quality of governance Index for a country s trade openness Lijphart s index of agreement between the state and the China in the UN general assembly voting Ranging from approximately -2.5 to 2.5 [2+(cce+pve+vae +rle +rqe +gee )/6]/4 Exports plus imports as a share of GDP (in percentage) This equals 1 if a state always agrees with China, 0 if it always votes the other way. If one state votes yes and the other abstains, the vote is coded as.5. World Bank World Bank Ideal point data 19

26 EMPIRICAL MODEL In this section I describe the regression methodology. The model this paper employs is the qualitative choice model of McFadden, which explains or predicts a country-pair s choice of whether enter or not a preferential trading accord using a latent dependent variable. y is a latent variable, representing the probability that China signs the FTA with a country, which is unobservable. x is the vector of explanatory variables; β is the vector of parameters for each independent variable; and ε is the error term. y = / 0 + xβ + ε (1) The outcome of interest, a country s FTA status with China, is observable. We define an indicator variable FTA, which takes the value 1 if China and the counterpart share a free trade agreement, and 0 otherwise. Then we have the function below: FTA= > (2) Using probit model as the link function, we get the following model, where ; is the standard normal cumulative distribution function, which ensures that P(FTA=1) lies between 0 and 1:!<= = ;(/ 0 +?/ (3) 20

27 RESULTS In this section, I present the empirical findings for the sets of predictive regressions. Tables 3 to 6 report the outputs of the statistical analysis. 4 Table 3 shows the summary statistics of the dependent and explanatory variables. Table 3: Summary Statistics Variable Observations Mean Std.Dev Min Max FTA Log of Distance Log of GDP Natural Resource Rents Export Partner Share Import Partner Share Trade Openness Voice & Accountability Political Stability & Absence of Violence Governance Effectiveness Regulatory Quality Rule of Law Control of Corruption Governance (Aggregated Indicator) PctAgreeChina FTA & Negotiation FTA & Negotiation & Study Note: The table provides summary statistics for independent variables averaged over the years 2002 to 2015; and for dependent variables as of January For convenience, the year dummies involved in all the specification models are not listed in the table. 21

28 Table 4 introduces my baseline results. Following the theoretical framework, I specified eight probit models in the first set of regressions. Column (1) uses only GDP as the independent variable. Column (2) includes GDP and geographical distance between two countries as regressors to test for the gravity model. In columns (3) and (4), I use the difference of GDP between two countries and natural resource rents accounted for a country s GDP as the independent variable respectively. Column (5) tests the correlation between bilateral trade relationships and the formation of FTA. Column (6) adds the regressor of a country s openness to trade to (5). In column (7), I tested the correlation between the aggregated quality of governance indicator and my dependent variable. As for column (8), I combined all the previous independent variables in the regression and added the percentage share of UN votes in line with China, which serves as the baseline model of my study. In each case, nearly all key independent variables show statistically significant correlations with the dependent variable. The first specification shows that a country s GDP is positively correlated with the probability that China forms an FTA with it. The estimation is statistically significant at 1% level. The effect remains significant (at 1% level) after adding the country s geographic distance with China, which is a good proof of our theoretical line linking the bilateral trade relations and the occurrence of an FTA. Countries are more likely to form an FTA with China if their economic sizes are bigger and are geographically closer to China. In the third column, the coefficient of GDP difference on the likelihood of forming an FTA between China and other countries is also statistically significant with a positive value (at 5% level), indicating that China tends to form FTAs with countries whose economic sizes are smaller. As for the fourth specification, a country s rents from natural resource industry are negatively correlated with the 22

29 possibility of forming an FTA with China. Natural resource is not a major priority in China s current FTA network strategy. In the sixth column, when regressing trade partner share on the possibility of an FTA formation, the coefficients turn out to be statistically significant (at 1% level) and with different signs. The probability that China forms an FTA with a country is positively correlated with the country s import partner share with China and negatively correlated with its export partner share with China. In other words, China s current preferential trading partners are more likely to be countries that export a lot to China and less likely to be countries that import a lot from China. Adding the variable "openness to trade" to the model, the estimates sign remains unchanged. The coefficient of the new variable is also significantly positive (at 1% level), suggesting that a country is more likely to form an FTA with China when it is more open to trade. The correlation between the quality of governance and the probability of FTA formation with China also turns out to be salient and positive (at 1% level), suggesting that countries with better level of governance are more likely to ratify an FTA with China. When including all the above variables and the percentage share of UN votes aligned with China in one regression, I get the baseline model. According to the estimation, the signs of the coefficients remain consistent with the previous specifications, and are all statistically significant as well. The estimate of a country s UN voting record in line with China is also positive and statistically significant. Countries that take similar stance with Beijing in global affairs are more likely to form FTAs with China. Table 5 presents the variation of the baseline model, where I disaggregated the quality of governance indicator, and individually test the six governance dimensions on the probability of the FTA ratification, controlling for other independent variables. As is shown in columns (1) (3) (4) (5) and (6), the coefficients of five dimensions turn out to be statistically significant and positive 23

30 Constant *** *** *** *** *** *** *** *** (0.368) (0.831) (1.119) (0.131) (0.132) (0.148) (0.154) (12.76) Observations 2,263 2,263 2,263 2,258 2,279 2,194 2,293 2,186 Pseudo R-squared Log Likelihood Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Table 4: Baseline Model (1) (2) (3) (4) (5) (6) (7) (8) GDP & GDP Natural Trade Share & Quality of All Variables Dependent Variable: FTA GDP only Trade Share Distance Difference Resource Rents Openness Governance Included 24 loggdp *** *** *** (0.0143) (0.0151) (0.0282) logdist *** *** (0.0728) (0.0955) loggdpdiff ** *** (0.0398) (0.441) natural *** *** ( ) ( ) ImportPartnerShare *** *** *** (0.0606) (0.0587) (0.0807) ExportPartnerShare *** *** *** (0.0958) (0.0905) (0.123) tradeopenness *** *** ( ) ( ) qgov *** *** (0.145) (0.292) PctAgreeChina *** (0.402)

31 25 Table 5: Disaggregate the Quality of Governance Indicator Control of Corruption (0.0546) (1) (2) (3) (4) (5) (6) Dependent Variable: FTA Voice & Political Stability & Government Regulatory Accountability Absence of Violence Effectiveness Quality Rule of Law loggdp *** *** *** *** *** *** (0.0279) (0.0286) (0.0295) (0.0290) (0.0279) (0.0279) logdist *** *** *** *** *** *** (0.0906) (0.0939) (0.0939) (0.0955) (0.0931) (0.0960) loggdpdiff *** *** *** *** *** *** (0.459) (0.445) (0.436) (0.415) (0.441) (0.440) natural *** *** *** *** *** *** ( ) ( ) ( ) ( ) ( ) ( ) ImportPartnerShare *** *** *** *** *** *** (0.0840) (0.0833) (0.0810) (0.0795) (0.0811) (0.0796) ExportPartnerShare *** *** *** *** *** *** (0.128) (0.126) (0.123) (0.122) (0.123) (0.122) tradeopenness *** *** ** *** *** ( ) ( ) ( ) ( ) ( ) ( ) PctAgreeChina *** *** *** *** *** *** (0.411) (0.343) (0.355) (0.413) (0.369) (0.346) vae *** (0.0626) pve (0.0549) gee *** (0.0653) rqe *** (0.0742) rle *** (0.0633) cce *** Constant *** *** *** *** *** *** (13.23) (12.90) (12.63) (12.01) (12.75) (12.70) Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

32 (all at 1% level). Voice & Accountability captures the citizens ability to impact government s behavior. The result shows that, controlling for other factors, the probability of ratifying an FTA with China increases when a country is more trusted by its citizens and represents better their preference. Government Effectiveness captures perceptions of the quality of public services and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government's commitment to such policies. Consistent with the hypothesis, more independence of bureaucracy from political pressures, greater credibility of government policy, and better policy implementation of a country all correspond to higher probability of concluding an FTA with China. Regulatory Quality captures perceptions of the ability of the government to formulate and implement sound policies and regulations that permit and promote private sector development. The higher a country s regulatory quality, the sounder policies it can launch to enable trade development, and the more likelihood that China shares a preferential trade agreement with the country. Rule of Law captures perceptions of the extent to which agents have confidence in and abide by the rules of society. China is more likely to sign an FTA with a country when it has a better quality of contract enforcement, property rights, the police, and the courts. Control of Corruption captures perceptions of the extent to which public power is exercised for private gain. Lack of corruption indicates lower costs of market entry and fewer compliance risks, which may explain its positive correlation with the FTA formation with China. 5 Table 6 expands the scope of our model to include pending or proposed FTAs. The first column is again our baseline model. In Column (2), I replace the dependent variable with a new one, FTA & Negotiation, that includes countries either have or are under the negotiation stage of 5 The definition of each indicator is included in the Worldwide Governance Indicators: Methodology and Analytical Issues by Daniel Kaufmann, Aart Kraay, and Massimo Mastruzzi. 26

33 an FTA with China. In Column (3), I further expand the range of countries covered by the dependent variable, adding countries that are in the exploratory phase of China s FTA network. According to the estimates, most of the coefficients signs stay unchanged, which shows the robustness of the results under different specifications. In the second specification, the coefficient of natural resource rents changes from negative at 1% level to statistically insignificant, largely due to the potential addition of GCC countries and Norway to the partnership. The coefficient of openness to trade has also changed. It is insignificant in the second estimation, and becomes significant and negative in the third estimation (at 10% level), reflecting that China is seeking preferential trading partners beyond those countries traditionally open to trade. Dependent Variable(s) Table 6: Variations of the Dependent Variable (1) (2) (3) FTA FTA & Under Negotiation FTA & Under Negotiation & Study loggdp *** *** *** (0.0282) (0.0265) (0.0224) logdist *** *** *** (0.0955) (0.0862) (0.0836) loggdpdiff *** *** *** (0.441) (0.388) (0.240) natural *** ( ) ( ) ( ) ImportPartnerShare *** *** *** (0.0807) (0.0770) (0.0653) ExportPartnerShare *** *** *** (0.123) (0.0519) (0.0415) tradeopenness *** * ( ) ( ) ( ) qgov *** *** *** (0.292) (0.281) (0.252) PctAgreeChina *** *** *** (0.402) (0.400) (0.333) Constant *** *** *** (12.76) (11.19) (6.952) Observations 2,186 2,186 2,186 Pseudo R-squared Log Likelihood Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 27

34 Despite the support from the data, this research struggles to overcome several limitations. First, in alignment with current literature, the diplomatic dimension is particularly relevant in the discussion of China s FTA partnership building. While we recognize that there is a special order in which China classifies its bilateral diplomatic relationship with other countries, there is a lack of available data to quantify such an aspect. The data availability problem also hinders the exploration of other political consideration s effect on the FTA partner choice. Secondly, the empirical method used in this study suffers a small sample size. The significance of the relationships is contestable when the number of the units of analysis is small. For example, according to the estimation of our baseline model, the coefficient of natural resource rents (% of GDP) is negative and statistically significant, but it makes no sense to conclude that China is more likely to form an FTA with countries of less natural resource production. Instead, the negative result is attributed to the fact that FTA negotiations with some most resource-rich countries are not yet concluded. In addition, when constructing the dependent variable, the treatment of country blocs as several individual and separate countries also affects the effectiveness of the empirical testing. 28

35 CONCLUSION AND POLICY IMPLICATIONS The present investigation made use of a panel of 166 countries over 14 years to test different indices effects on the formation/negotiation of FTAs with China. My empirical findings complement the existing literature on the discussion of China s FTA strategies. These findings suggest that, at the current stage, China s FTA partners are more likely to be countries that are close to China geographically, with an economic size that is relatively large but not similar to China s, occupying a bigger share of China s total import from the world, sharing more similar stances in international affairs with China, and more open to trade. Voice & accountability, government effectiveness, regulatory quality, rule of law, and control of corruption also help explain China s preference in its FTA partnership building. As China expands its FTA network, it begins to seek preferential trading relationships with countries that are not traditionally open to trade. The natural resource richness, at this stage of China s development though, does not appear to be a robust indicator of its selection of FTA partner. My study contributes to the existing literature by empirically testing a group of hypotheses regarding considerations behind China s FTA network building, and particularly its partner choice. The coefficients of some economic factors in my model turn out to be statistically significant; however, the overall explanatory power of the model is still not perfect. The analytical approach of the paper can be enhanced by adding a series of case studies to further explain China s calculations in expanding its FTA network. Countries often implant a wide range of non-economic calculations beneath these ostensibly economic agreements, driving the direction of national policy decisions. China is not an exception. The lack of analytically well-grounded assessments of my issue is actually due to the fact that geopolitical motivations are inherently difficult to quantify. 29

36 However, efforts continue to integrate socio-political factors into existing data frameworks. I am optimistic that such efforts will shed lights on future studies of this kind. At this point, it is also worth noting that the study of China s FTA strategies and priorities contains both limitations and prospects. As China has been deepening its global engagement with the Belt and Road Initiative as well as changing its diplomatic priorities, new factors will need to be taken into consideration when thinking about how China designs its preferential trading agenda. Bearing these transitions in mind, readers need to be aware that a conclusion of the past cannot be simply applied to predict the future. The recently concluded China-Maldives FTA, for example, is such a new type of arrangement in this context. Maldives was not a priority in China s foreign policy until 2011, when China opened its first embassy in its capital Male after almost 40 years establishment of diplomatic relations. 6 There is also an extreme difference in scale between two countries, Maldives being one of the smallest economies of Asia and China the largest. As for the natural resource abundance and quality of governance, Maldives is not ranked well. Besides, the FTA with China is Maldives first with any foreign country. There are three possible reasons for this unlikely partnership based on my model. First, the signing of this agreement is the latest achievement of "promoting the construction of Free Trade Areas and an open world economy" put forward at the 19th CPC National Congress as well as a significant move to consolidate and strengthen the two countries all-round and friendly cooperative partnership facing the future. 7 Second, Maldives participation in the Maritime Belt Road Initiative illustrates Beijing s intention to engage the 6 Ministry of Foreign Affairs of the People s Republic of China: /t shtml, accessed April 10, MOFCOM: accessed April 10,