The Political Economy of Joining the AIIB

The Political Economy of Joining the AIIB Abstract This article analyzes the determinants of new prospective members of the Asian Infrastructure Investment Bank (AIIB). I argue that less democratic countries are more likely to apply, and that when deciding to join the institution countries learn from their neighbors and from any previous international organization (IO) interactions they have had with China. Building on detailed panel data covering the institution’s founding period, i.e. from October 24, 2014 to March 31, 2015, I fit both a probit model with time polynomials and a Cox duration model to identify country characteristics that correlate with joining the AIIB. I show that countries with lower polity scores or whose neighbors had already become members were more likely to join, and that the probability was higher for countries that belonged to China-led IOs. Lastly, I show that countries under-represented in the existing international financial system were also more likely to join. My findings highlight the importance of democracy in shaping the membership structure of the AIIB, and demonstrate how countries leverage information from their neighbors and from previous interactions with China to adjust the perceived risk of joining. They also provide the first modern-day empirical support for the contested multilateralism framework. Introduction That good infrastructure helps economic growth and poor infrastructure hinders it is well known.1 Equally well known is the fact that many of the developing countries, for example, in Asia and Africa have long suffered a basic infrastructure deficit.2 A report in 2017 by the Asian Development Bank (ADB) states that Asia will need to invest $22.6 trillion between 2016 and 2030 to maintain its growth momentum and eradicate poverty.3 When China proposed setting up the Asian Infrastructure Investment Bank (AIIB) to, ‘bring countries together to address the daunting infrastructure needs across Asia and beyond’,4 however, responses throughout the world were divided. Some countries, such as India and Mongolia, joined immediately. Others, including Australia and South Korea, were more reticent, and did not decide to join till a later stage. Meanwhile a few others, most notably the United States and Japan, declared that they would not even consider joining. Given the significance of the new institution and the urgent needs of the developing world, how did countries decide whether and when to join as prospective founding members? Figure 1 illustrates the expansion of AIIB membership between October 24, 2014 and March 31, 2015. During this period, 57 countries joined as prospective founding members. On the whole, it was countries close to China, such as Pakistan and Kazakhstan, that were the early joiners, and countries farther away that either joined later or decided not to. However, it is not easy through such high-level aggregate observations alone to pinpoint the relative importance of the various factors at play at the time a country was making the decision. Fig. 1. View largeDownload slide Geographic Expansion of the AIIB Membership. Note: The graph presents all 56 founding members of the AIIB other than China. Due to space constraints, I have not reported all the names. It can be clearly seen how AIIB membership expands geographically overtime (the x axis). Given its proximity to China, South Korea’s late joining makes it an outlier on the graph, not to say Japan’s no-show. Fig. 1. View largeDownload slide Geographic Expansion of the AIIB Membership. Note: The graph presents all 56 founding members of the AIIB other than China. Due to space constraints, I have not reported all the names. It can be clearly seen how AIIB membership expands geographically overtime (the x axis). Given its proximity to China, South Korea’s late joining makes it an outlier on the graph, not to say Japan’s no-show. In this article, I model the expansion of the AIIB prospective founding membership to analyze how countries weigh the various factors in this significant but uncertain process. I consider countries’ economic characteristics, such as GDP and GDP per capita; dyad traits, such as physical distance from China, and countries that had already become AIIB members; political characteristics, such as polity score and civil liberties; and existing international organization (IO) relationships, including the ADB, the Shanghai Cooperation Organization (SCO), and the World Bank. Building from an original dataset of membership expansion dates, I estimate both a probit model with time polynomials and a Cox duration model to analyze the probability of applying to become AIIB prospective founding members. To preview the results, I find that less democratic countries are more likely to join the AIIB as founding members. This result is robust across all the metrics: polity score, political constraints, civil liberties, and political rights.5 Consistent with our overall observation from Figure 1, I find that countries located closer to China are more likely to join than countries farther away. This physical barrier breaks down, however, when neighboring countries take the lead in joining the IO. For example, since the UK joined the AIIB, the 8017 kilometre distance between France and China has become less relevant for the French government. Similarly, Russia’s joining the IO has significantly reduced the distance between Bishkek and Beijing as far as Kyrgyzstan is concerned. This suggests that it is the intrinsic uncertainty involved in joining an IO, rather than physical distance from it, that affects countries’ calculations.6 Such uncertainty is lower for countries that have previously interacted with China in other China-led IOs, for example, the SCO. Finally, I demonstrate that countries under-represented in the current international financial system are more likely to join the AIIB as a fresh alternative. Facts show that China itself has long been under-represented in the current international financial system. As of 2015, China’s voting share for the International Monetary Fund (IMF) was a meagre 3.81%, substantially smaller than that for Germany, the UK, or France,7 even though China had become the world’s second largest economy. Similarly, the voting share for India, Bangladesh, Bhutan, and Sri Lanka combined amounted to less than half that for Germany. The creation of the AIIB and the positive responses of under-represented countries thus lend empirical support to the contested multilateralism framework.8 The rest of the article is organized as follows. ‘Background and theory’ section presents the background, theories, and related hypotheses. ‘Model formulation and data’ section presents the econometric model. ‘Data’ section presents the dataset and its summary statistics. ‘Empirical analysis’ section presents and discusses the results. Last section concludes. Background and Theory The AIIB is a multilateral financial institution which the Chinese government initiated in 2014. China is its largest contributing member. Fifty-seven countries have joined the AIIB as prospective founding members,9 and by June 16, 2017, 23 other countries, including Argentina, Belgium, and Canada, had joined as non-founding members. The AIIB’s membership thus quickly eclipsed that of the Japan-led ADB, which currently has 67 members. With an initial capitalization of $100 billion, and with members ranging from South Africa, Egypt, and Great Britain to Brazil and Chile, the AIIB stands out as an important IO, both economically and politically. At the same time, the sequential nature of the expansion of the AIIB’s membership presents researchers with an ideal opportunity to examine several important rational choice-based questions that can shed light on the AIIB expansion process, and provide guidance for the future expansion of similar IOs. IOs such as the AIIB are created to facilitate collective action and to lower transaction costs.10 Over the past few decades, IOs have proliferated, expanded their memberships, and extended their influence to include more and more new issue areas. The vast majority of these IOs, especially the most prominent ones such as the IMF and World Bank, were created by the world’s most developed countries. More recently, the developing world led by the BRICS countries also started to establish IOs to advance economic development and facilitate multilateral cooperation. The New Development Bank, the SCO, and the AIIB are but a few examples. Although joining an IO is a risky proposition filled with uncertainty, countries nevertheless come together under the common belief in mutual benefits for big and small players alike.11 Therefore, this article assumes that countries that stand to benefit from AIIB membership choose to join while others do not. This leads us to several testable hypotheses. Hypothesis 1: Compared with democracies, autocracies are more likely to join the AIIB. The first hypothesis is based on the observation that autocracies face fewer political constraints than democracies with respect to joining the China-led institution.12 Such political constraints have both domestic and international sources. Like any other international negotiation, joining the AIIB can be viewed as a two-level bargaining issue, whereby leaders of potential member countries negotiate with the Chinese government at the international level, and with their congress or parliament at the domestic level.13 The political constraints on democracies are likely to be greater at the domestic level because (i) leaders will need the approval of congress or parliament, and might fall into a ‘democratic gridlock’ and (ii) there is greater ideological distance between China and the prospective country.14 At the international level, in addition to the Chinese government, the United States also played an important role in shaping the membership structure, because the United States was highly sceptical of the China-initiated institution, and tried publicly to dissuade its Western allies from joining. Implicitly, the United States could also put substantial pressure on countries that rely on US foreign aid.15 Research has demonstrated that the United States exerts more influence over aid-receiving democracies than it does on aid-receiving autocracies, as aid linkages to a country’s decision on AIIB membership are more credible when directed toward democracies.16 Consequently, pressure from the United States, both public and in private, further shrinks the win-set for democratic governments, making democracies even less likely to join. The hypothesis connects this study to the burgeoning literature of autocracy promotion that investigates the effects of interactions between China and autocracies on the regime duration, and more generally on the literature on foreign aid and democratization.17 While most of the existing studies focus on autocracies that are recipients of Chinese loans and aid, my article examines a completely new angle: the different reaction of autocracies compared to that of democracies to a Chinese initiative. Viewed from the perspectives of both domestic political constraints and of international pressure from the United States, it is reasonable to expect that autocracies are more likely than democracies to join the AIIB.18 Next, I examine the role of uncertainty and learning in shaping the AIIB membership structure. As with other IOs, the process of joining the AIIB and the institution itself are filled with uncertainty. That China is a developing country and the AIIB is one of the first IOs China has initiated heightens the degree of uncertainty.19 How could countries rationally commit themselves to this new IO? Assuming that countries are risk averse, those that have greater confidence in the China-led institution, and in China in general, will have a higher utility in joining (Figure 2). Mathematically, I formulate this argument as follows: u(AIIB) = u(E(AIIBperformance)) = u((1 − p) · AIIBfail + p ·AIIBsucceed) where u is concave and twice differentiable, AIIBsucceed denotes AIIB’s success, AIIBfail represents the failure of the institution, and p is probability of AIIBsucceed. The assumption that countries are risk averse is reflected by the fact that u′ > 0 and u″ < 0.20 As countries are interdependent and information flows across borders, countries learn from their neighbors to mitigate the lack of information, similar to interorganizational learning in organizational learning theory.21 Consequently, a country’s confidence (p) in AIIB should increase if neighboring countries have already joined. Similarly, countries that have past experience with China could place a higher p on AIIBsucceed as well. As China is a developing country and its credentials in the international community are not fully established, at least as compared with the UK or the United States, previous interactions with China should play an important role in the decision-making process, and countries that have had such interactions are able to achieve a higher certainty equivalent for joining the AIIB. Fig. 2. View largeDownload slide Information and AIIB Performance. Note: As countries accumulate more information about the IO, E(AIIB performance) goes up and the purple dot (second from the left) moves toward the green dot (third from the left) along the utility curve. Fig. 2. View largeDownload slide Information and AIIB Performance. Note: As countries accumulate more information about the IO, E(AIIB performance) goes up and the purple dot (second from the left) moves toward the green dot (third from the left) along the utility curve. Hypothesis 2: Countries with neighbors that are joining or that have had previous interactions with China are more likely to join. Lastly, I examine the role of representation in the international financial system. Ideally, the structure of the system is determined by the distribution of capabilities across the countries.22 In reality, however, the structure of the system usually lags behind the evolution of capabilities, leading to the observation that some countries are over-represented and others under-represented.23 If fair representation is desirable, then I expect that countries which are currently under-represented would be able to increase their utility by joining a new IO (with the intuition that the marginal utility is for them substantially higher than that for fairly or over-represented countries). When a mismatch between power and IO resources occurs and is not addressed, and when existing IOs cannot meet global demands, for example, in fighting climate change, this will add to pressure for regime contestation, in the form either of regime shifting or regime creation.24 The China-initiated AIIB attests to this argument. As China’s power in the international system grows, it seeks greater representation in existing IOs like the IMF and the World Bank. Such aspirations, however, have been constantly thwarted, thus creating the chance for China to establish ‘a World Bank of its own’.25 Smaller countries that have no outside option and feel under-represented are not able to push for adjustment within existing IOs, and nor could they credibly establish new IOs.26 Therefore, the pent-up demand for representation should naturally lead them to become the earliest to participate in newly created IOs, such as the AIIB, both to signal dissatisfaction with the status quo and to secure a better representation in the new IOs. This leads to the third testable hypothesis. Hypothesis 3: Compared with well-represented countries, countries that are under-represented in the existing IOs are more likely to join, and to join early. Model Formulation and Data Model Formulation To test these hypotheses, I formulate a random utility model where countryi has the utility function specified as follows, and will join if its utility is greater than 0. The key variables of interest are Polity (Hypothesis 1), Neighborextensive(i, t), Neighborintensive(i, t) and IO (Hypothesis 2), and Representation(i) (Hypothesis 3). Neighborextensive(i, t) and Neighborintensive(i, t) are updated daily between October 24, 2014 and March 31, 2015. Where GDP is the country’s economic size, GDP per capita is the income level; Distance is the geographical distance between the country and China; IO variables are binary, indicating whether or not countryi is a member of an IO of interest; Asiai is a dummy variable indicating whether the countryi is an Asian country or not; Δt is a duration variable representing the number of days since the most recent country joined; and εi, t has a standard normal distribution and is i.i.d.27 I include country GDP and physical distance in the model because empirical observations show that countries close to China are the first ones to join, and large economies tend to join the AIIB earlier than smaller ones when controlling for geographic distances. I report this pattern in Figure 1. This suggests that GDP size and geographic distance weigh heavily in states’ calculations.28 Since the AIIB is an investment bank in nature, I expect that, all other things being equal, the probability of rich countries joining the AIIB should be higher than of poor countries, as they have the resources to contribute to the institution and are generally eager to ‘shape the trajectory’ of it.29 On the other hand, poor countries also have an incentive to join, as they will likely need the AIIB to help fund infrastructure projects.30 Therefore, I include both linear and quadratic terms of GDP per capita in the modelling, and expect the coefficient on the quadratic term to be positive. Neighborextensive(i, t) captures the extensiveness of the AIIB’s attraction for countryi at time t. It is defined as the number of neighboring countries already in the IO for countryi at time t. Neighborextensivei, t= Σj∈IO⁡Neighbor(i, j, t) Neighborintensive(i, t) captures the intensity of attraction of the AIIB for countryi at time t.31 Neighborintensivei, t=maxj∈IO⁡logGDP(j)Distance(i, j) Given that not all neighbors carry the same weight, this variable Neighborintensive is so constructed as to capture the effect of important neighbors: when a new member has joined the AIIB, Neighborextensive will be updated according to maxj∈IO⁡logGDP(j)Distance(i, j). By design, countries with small economies and countries located far away will not affect country i’s utility. In contrast with Neighborextensive, Neighborintensive is not restricted to physically contiguous countries. The UK’s application, for example, can thus affect the decisions of Germany and Italy through this channel. Polityi is the polity score of countryi. The higher the polity score, the more democratic the country is. Representationi measures how well represented the country is in the current international system, and is the key variable for testing the hypothesis that under-represented countries are the ones that join the AIIB, and that they join early. Δt and its polynomial terms are aimed at capturing time dependency.32 From the parameters λ1, λ2 and λ3, I will be able to test the existence of momentum effects.33 This is closely related to the hazard rate concept in the literature on duration analysis.34 The dynamics of the model play out as follows. In Period 1, (only) the founder joins the IO. Variables Neighborextensive and Neighborintensive are updated for each country. Countries with positive utility choose to join. The world enters Period 2, with Neighborextensive and Neighborintensive updated for countries not yet in the IO. Countries that have not yet joined calculate their utility for Period 2, and decide whether or not to join. So, in Period t, countryi that is not yet in the IO will decide again with updated Neighborextensive and Neighborintensive. There is a finite number of periods, as there is a deadline for applications for founding member status in the IO. The key assumption of the model is that countryj’s joining the IO will affect the subsequent calculations of all the non-Member States through two channels: Neighborextensive and Neighborintensive. In terms of marginal utility, this can be expressed as: Neighborextensive:ΔUi=β4, if country i and country j are continugous0, otherwise Neighborintensive:∂Ui∂Neighborintensive(i)=β5 Learning from the past, in the sense that countries that are already members of a China-led IO are more likely to join the AIIB, can be captured by Γ. I will estimate the random utility model in ‘Empirical analysis’ section using probit, but first let me introduce the data. Data The central item of data in my article consists in dates of application, which I present in Appendix B. Data on other variables are from standard sources. Importantly, I restrict my sample countries to United Nations Member States that recognize China (not Taiwan). Data with respect to countries that recognize Taiwan, listed in Appendix C8, are from Xinhua. The dataset comprises a total of 170 countries, and once one joins the AIIB it will drop out of the sample. Data on country GDP and GDP per capita come from the World Bank.35 Both GDP and GDP per capita are on log scale. Data on geographical distance and physical contiguity come from Centre d’Etudes Prospectives et d’Informations Internationales (CEPII).36 Distances are measured in kilometres, and in this article distance represents the natural logarithm of the physical distance. A key question I address in the article is whether or not less democratic countries are more likely to join the AIIB. For this purpose, I use four alternative measures of democracy: polity scores from the Polity IV Project, political constraints, and political rights, and civil liberties from Freedom House.37 To study the effects of learning from IO interactions, I construct the binary variable IO. The variable IO will take value 1 if countryi is a member of an IO of interest. In this article, I will use the BRICS and the SCO. BRICS is an (informal) IO that consists of Brazil, Russia, India, China, and South Africa. All five countries are founding members, but joined at different times. The SCO was cofounded by China, Kazakhstan, Kyrgyzstan, Russia, Tajikistan, and Uzbekistan in 2001. I present the detailed membership information in Table 1.38 I code the variable SCO as 1 for the 18 countries other than China in the SCO, and 0 for all other countries. Here I do not distinguish between formal members, observer states, and dialogue partners. Table 1 SCO Membership SCO Status Country Member States China, India, Kazakhstan, Kyrgyzstan, Pakistan, Russia, Tajikistan, Uzbekistan Observer States Afghanistan, Belarus, Iran, Mongolia Dialogue Partners Armenia, Azerbaijan, Cambodia, Nepal, Sri Lanka, Turkey Guest Turkmenistan SCO Status Country Member States China, India, Kazakhstan, Kyrgyzstan, Pakistan, Russia, Tajikistan, Uzbekistan Observer States Afghanistan, Belarus, Iran, Mongolia Dialogue Partners Armenia, Azerbaijan, Cambodia, Nepal, Sri Lanka, Turkey Guest Turkmenistan Note: India and Pakistan became full members of the SCO in 2017, and Belarus became an observer member (up from dialogue partner) in 2015. Table 1 SCO Membership SCO Status Country Member States China, India, Kazakhstan, Kyrgyzstan, Pakistan, Russia, Tajikistan, Uzbekistan Observer States Afghanistan, Belarus, Iran, Mongolia Dialogue Partners Armenia, Azerbaijan, Cambodia, Nepal, Sri Lanka, Turkey Guest Turkmenistan SCO Status Country Member States China, India, Kazakhstan, Kyrgyzstan, Pakistan, Russia, Tajikistan, Uzbekistan Observer States Afghanistan, Belarus, Iran, Mongolia Dialogue Partners Armenia, Azerbaijan, Cambodia, Nepal, Sri Lanka, Turkey Guest Turkmenistan Note: India and Pakistan became full members of the SCO in 2017, and Belarus became an observer member (up from dialogue partner) in 2015. To test the contested multilateralism framework, I draw data from the ADB. Founded on December 19, 1966 the ADB is led by Japan and the United States. Information on ADB membership, including how it overlaps with and differs from AIIB membership, is presented in ‘Contested multilateralism: ADB members’ section. The variable representation is calculated, using ADB data, according to the country’s voting share to GDP ratio. A country is well represented in the ADB if it has a high share-to-GDP ratio, and will have a higher representation value. Raw data on countries’ physical neighbors have been obtained from the CEPII.39 To incorporate dynamics into the model, I update the Neighborextensive and Neighborintensive for each country according to the updated membership, which I base on Chinese Ministry of Finance (MOF) announcements.40 As will be elaborated on later, not all applications are public, but the Chinese MOF publicly announces all admissions. I assume that only publicly available information will enter into states’ calculation. Under this assumption, Iran’s application (dated March 30, 2015) will not affect Azerbaijan’s decision to apply on March 31, 2015. Similarly, Spain’s application on March 27, 2015 will not affect Portugal’s decision. Empirical Analysis In this section, I estimate the model using the AIIB dataset, and test the hypotheses, one-by-one, incrementally. I first estimate the model using the standard probit, and examine the empirical results on democracy. Second, I analyze the effects of learning on AIIB membership. Third, using the ADB as an example, I test the contested multilateralism framework. Fourth, to re-formulate the dataset into panel data, I estimate the same regressors using a Cox Partial Likelihood model as a robustness check. I will discuss the results in the same order. Democracy and AIIB Prospective Founding Membership The first column in Table 2 displays estimate from a static model, and only considers economic factors. This is similar to the standard gravity model in international trade, and aims at capturing static economic factors. The result shows that large countries and countries close to China are more likely to join. It is worth noting that, while the coefficients on GDP per capita point to the expected directions, they are not statistically significant. Table 2. The Expansion of AIIB Membership Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.0856*** 0.157*** 0.0918*** 0.111*** 0.111*** (0.006) (0.000) (0.005) (0.003) (0.003) Ln GDP per capita −0.193 −0.383 −0.273 −0.179 −0.153 (0.612) (0.387) (0.480) (0.677) (0.721) Ln GDP per capita2 0.0117 0.0211 0.0174 0.0143 0.0123 (0.603) (0.419) (0.447) (0.572) (0.628) Ln Distance −0.522*** −0.767*** −0.499*** −0.753*** −0.763*** (0.000) (0.000) (0.000) (0.000) (0.000) Polity −0.0274** (0.012) Political Constraints −0.497* (0.089) Civil Liberties 0.0826** (0.042) Political Rights 0.0692* (0.061) Δ −2.828*** −3.900*** −2.862*** −4.096*** −4.064*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 1.246*** 2.232*** 1.266*** 2.493*** 2.465*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.153*** −0.367*** −0.155*** −0.432*** −0.425*** (0.000) (0.003) (0.000) (0.000) (0.001) Constant 1.427 3.028 1.497 2.431 2.517 (0.461) (0.195) (0.443) (0.273) (0.258) Observations 21876 16618 21240 18899 18899 Pseudo R2 0.243 0.297 0.245 0.283 0.282 Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.0856*** 0.157*** 0.0918*** 0.111*** 0.111*** (0.006) (0.000) (0.005) (0.003) (0.003) Ln GDP per capita −0.193 −0.383 −0.273 −0.179 −0.153 (0.612) (0.387) (0.480) (0.677) (0.721) Ln GDP per capita2 0.0117 0.0211 0.0174 0.0143 0.0123 (0.603) (0.419) (0.447) (0.572) (0.628) Ln Distance −0.522*** −0.767*** −0.499*** −0.753*** −0.763*** (0.000) (0.000) (0.000) (0.000) (0.000) Polity −0.0274** (0.012) Political Constraints −0.497* (0.089) Civil Liberties 0.0826** (0.042) Political Rights 0.0692* (0.061) Δ −2.828*** −3.900*** −2.862*** −4.096*** −4.064*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 1.246*** 2.232*** 1.266*** 2.493*** 2.465*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.153*** −0.367*** −0.155*** −0.432*** −0.425*** (0.000) (0.003) (0.000) (0.000) (0.001) Constant 1.427 3.028 1.497 2.431 2.517 (0.461) (0.195) (0.443) (0.273) (0.258) Observations 21876 16618 21240 18899 18899 Pseudo R2 0.243 0.297 0.245 0.283 0.282 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Table 2. The Expansion of AIIB Membership Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.0856*** 0.157*** 0.0918*** 0.111*** 0.111*** (0.006) (0.000) (0.005) (0.003) (0.003) Ln GDP per capita −0.193 −0.383 −0.273 −0.179 −0.153 (0.612) (0.387) (0.480) (0.677) (0.721) Ln GDP per capita2 0.0117 0.0211 0.0174 0.0143 0.0123 (0.603) (0.419) (0.447) (0.572) (0.628) Ln Distance −0.522*** −0.767*** −0.499*** −0.753*** −0.763*** (0.000) (0.000) (0.000) (0.000) (0.000) Polity −0.0274** (0.012) Political Constraints −0.497* (0.089) Civil Liberties 0.0826** (0.042) Political Rights 0.0692* (0.061) Δ −2.828*** −3.900*** −2.862*** −4.096*** −4.064*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 1.246*** 2.232*** 1.266*** 2.493*** 2.465*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.153*** −0.367*** −0.155*** −0.432*** −0.425*** (0.000) (0.003) (0.000) (0.000) (0.001) Constant 1.427 3.028 1.497 2.431 2.517 (0.461) (0.195) (0.443) (0.273) (0.258) Observations 21876 16618 21240 18899 18899 Pseudo R2 0.243 0.297 0.245 0.283 0.282 Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.0856*** 0.157*** 0.0918*** 0.111*** 0.111*** (0.006) (0.000) (0.005) (0.003) (0.003) Ln GDP per capita −0.193 −0.383 −0.273 −0.179 −0.153 (0.612) (0.387) (0.480) (0.677) (0.721) Ln GDP per capita2 0.0117 0.0211 0.0174 0.0143 0.0123 (0.603) (0.419) (0.447) (0.572) (0.628) Ln Distance −0.522*** −0.767*** −0.499*** −0.753*** −0.763*** (0.000) (0.000) (0.000) (0.000) (0.000) Polity −0.0274** (0.012) Political Constraints −0.497* (0.089) Civil Liberties 0.0826** (0.042) Political Rights 0.0692* (0.061) Δ −2.828*** −3.900*** −2.862*** −4.096*** −4.064*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 1.246*** 2.232*** 1.266*** 2.493*** 2.465*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.153*** −0.367*** −0.155*** −0.432*** −0.425*** (0.000) (0.003) (0.000) (0.000) (0.001) Constant 1.427 3.028 1.497 2.431 2.517 (0.461) (0.195) (0.443) (0.273) (0.258) Observations 21876 16618 21240 18899 18899 Pseudo R2 0.243 0.297 0.245 0.283 0.282 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. In columns 2, 3, 4, and 5, I sequentially include four alternative democracy scores: polity, political constraints, political rights, and civil liberties. I find that countries with lower polity scores, lower political constraints, higher political rights scores, and higher civil liberties scores tend to join the AIIB.41 The coefficients are all significant at the 0.1 significance level. They are all consistent with my first hypothesis: less democratic countries are more likely to join the institution than democratic countries. Lastly, I examine time dependency. It will be interesting to know whether countries are more likely to follow other countries’ example by joining, or to wait. I answer this question through analyzing the shape of the polynomial function of time. In Figure 3, I calculate the expected probability of a country joining the AIIB as a function of its waiting time. The figure shows that the probability of joining decreases sharply as the waiting time grows. The function is not monotonic, though. There is a slight increase around Day 28. But overall, the result suggests that timing is very important, and that a country is most likely to join immediately upon following another country’s lead. Fig. 3 View largeDownload slide Joining Probability as a Function of Time (with a 95% Confidence Interval). Note: The 95% confidence interval is calculated using the method. To preserve its symmetric structure, I do not cutoff the regions below zero, but it should be understood that the probability of joining cannot be negative. Fig. 3 View largeDownload slide Joining Probability as a Function of Time (with a 95% Confidence Interval). Note: The 95% confidence interval is calculated using the method. To preserve its symmetric structure, I do not cutoff the regions below zero, but it should be understood that the probability of joining cannot be negative. Uncertainty, Learning, and the AIIB Founding Membership Next, I focus on the two groups of learning variables in the model: Neighborextensive and Neighborintensive, SCO, and BRICS (Table 3). As regards neighbors, I find that countries whose neighbor states have joined the AIIB are more likely to join, as are countries whose neighbors are important countries (i.e. Neighborintensive). This result accords with my argument that as information flows in from neighboring and nearby states, the risk of joining as perceived by countries decreases, and the likelihood that they will join increases. This also provides insight into why the UK’s joining the AIIB should have such impact: it not only affects neighboring countries, such as Ireland, but also reassures countries nearby, such as France, Germany, and to some extent Israel. Table 3 Dynamic Model of AIIB Membership Pr(Join = 1) 1 2 3 4 5 6 Ln GDP 0.157*** 0.192*** 0.213*** 0.235*** 0.196*** 0.197*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.001) Ln GDP per capita −0.383 −0.452 −0.352 −0.587 −0.790 −1.375** (0.387) (0.352) (0.466) (0.254) (0.128) (0.017) Ln GDP per capita2 0.0211 0.0243 0.0134 0.0277 0.0417 0.0740** (0.419) (0.391) (0.638) (0.359) (0.175) (0.029) Ln Distance −0.767*** −0.648*** −0.400*** −0.275* −0.374** 0.396 (0.000) (0.000) (0.007) (0.091) (0.024) (0.127) Polity −0.0274** −0.0230** −0.0331*** −0.0292** −0.0356*** −0.0263** (0.012) (0.034) (0.004) (0.013) (0.004) (0.037) Neighborextensive 0.231*** 0.162** 0.137** 0.122* 0.0256 (0.000) (0.010) (0.035) (0.065) (0.715) Neighborintensive 0.413*** 0.418*** 0.472*** 0.709*** (0.000) (0.000) (0.000) (0.000) SCO 0.442** 0.403* 0.307 (0.026) (0.050) (0.125) BRICS 1.097*** 1.200*** (0.002) (0.002) Asia 1.017*** (0.000) Δ −3.900*** −3.941*** −3.721*** −3.823*** −3.745*** −3.872*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 2.232*** 2.250*** 2.169*** 2.245*** 2.186*** 2.268*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.367*** −0.369*** −0.360*** −0.374*** −0.362*** −0.377*** (0.003) (0.003) (0.005) (0.004) (0.004) (0.005) Constant 3.028 1.298 −9.968*** −10.88*** −9.515** −19.07*** (0.195) (0.592) (0.009) (0.005) (0.015) (0.000) Observations 16618 16618 16618 16618 16618 16618 Pseudo R2 0.297 0.318 0.342 0.350 0.362 0.386 Pr(Join = 1) 1 2 3 4 5 6 Ln GDP 0.157*** 0.192*** 0.213*** 0.235*** 0.196*** 0.197*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.001) Ln GDP per capita −0.383 −0.452 −0.352 −0.587 −0.790 −1.375** (0.387) (0.352) (0.466) (0.254) (0.128) (0.017) Ln GDP per capita2 0.0211 0.0243 0.0134 0.0277 0.0417 0.0740** (0.419) (0.391) (0.638) (0.359) (0.175) (0.029) Ln Distance −0.767*** −0.648*** −0.400*** −0.275* −0.374** 0.396 (0.000) (0.000) (0.007) (0.091) (0.024) (0.127) Polity −0.0274** −0.0230** −0.0331*** −0.0292** −0.0356*** −0.0263** (0.012) (0.034) (0.004) (0.013) (0.004) (0.037) Neighborextensive 0.231*** 0.162** 0.137** 0.122* 0.0256 (0.000) (0.010) (0.035) (0.065) (0.715) Neighborintensive 0.413*** 0.418*** 0.472*** 0.709*** (0.000) (0.000) (0.000) (0.000) SCO 0.442** 0.403* 0.307 (0.026) (0.050) (0.125) BRICS 1.097*** 1.200*** (0.002) (0.002) Asia 1.017*** (0.000) Δ −3.900*** −3.941*** −3.721*** −3.823*** −3.745*** −3.872*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 2.232*** 2.250*** 2.169*** 2.245*** 2.186*** 2.268*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.367*** −0.369*** −0.360*** −0.374*** −0.362*** −0.377*** (0.003) (0.003) (0.005) (0.004) (0.004) (0.005) Constant 3.028 1.298 −9.968*** −10.88*** −9.515** −19.07*** (0.195) (0.592) (0.009) (0.005) (0.015) (0.000) Observations 16618 16618 16618 16618 16618 16618 Pseudo R2 0.297 0.318 0.342 0.350 0.362 0.386 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Table 3 Dynamic Model of AIIB Membership Pr(Join = 1) 1 2 3 4 5 6 Ln GDP 0.157*** 0.192*** 0.213*** 0.235*** 0.196*** 0.197*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.001) Ln GDP per capita −0.383 −0.452 −0.352 −0.587 −0.790 −1.375** (0.387) (0.352) (0.466) (0.254) (0.128) (0.017) Ln GDP per capita2 0.0211 0.0243 0.0134 0.0277 0.0417 0.0740** (0.419) (0.391) (0.638) (0.359) (0.175) (0.029) Ln Distance −0.767*** −0.648*** −0.400*** −0.275* −0.374** 0.396 (0.000) (0.000) (0.007) (0.091) (0.024) (0.127) Polity −0.0274** −0.0230** −0.0331*** −0.0292** −0.0356*** −0.0263** (0.012) (0.034) (0.004) (0.013) (0.004) (0.037) Neighborextensive 0.231*** 0.162** 0.137** 0.122* 0.0256 (0.000) (0.010) (0.035) (0.065) (0.715) Neighborintensive 0.413*** 0.418*** 0.472*** 0.709*** (0.000) (0.000) (0.000) (0.000) SCO 0.442** 0.403* 0.307 (0.026) (0.050) (0.125) BRICS 1.097*** 1.200*** (0.002) (0.002) Asia 1.017*** (0.000) Δ −3.900*** −3.941*** −3.721*** −3.823*** −3.745*** −3.872*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 2.232*** 2.250*** 2.169*** 2.245*** 2.186*** 2.268*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.367*** −0.369*** −0.360*** −0.374*** −0.362*** −0.377*** (0.003) (0.003) (0.005) (0.004) (0.004) (0.005) Constant 3.028 1.298 −9.968*** −10.88*** −9.515** −19.07*** (0.195) (0.592) (0.009) (0.005) (0.015) (0.000) Observations 16618 16618 16618 16618 16618 16618 Pseudo R2 0.297 0.318 0.342 0.350 0.362 0.386 Pr(Join = 1) 1 2 3 4 5 6 Ln GDP 0.157*** 0.192*** 0.213*** 0.235*** 0.196*** 0.197*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.001) Ln GDP per capita −0.383 −0.452 −0.352 −0.587 −0.790 −1.375** (0.387) (0.352) (0.466) (0.254) (0.128) (0.017) Ln GDP per capita2 0.0211 0.0243 0.0134 0.0277 0.0417 0.0740** (0.419) (0.391) (0.638) (0.359) (0.175) (0.029) Ln Distance −0.767*** −0.648*** −0.400*** −0.275* −0.374** 0.396 (0.000) (0.000) (0.007) (0.091) (0.024) (0.127) Polity −0.0274** −0.0230** −0.0331*** −0.0292** −0.0356*** −0.0263** (0.012) (0.034) (0.004) (0.013) (0.004) (0.037) Neighborextensive 0.231*** 0.162** 0.137** 0.122* 0.0256 (0.000) (0.010) (0.035) (0.065) (0.715) Neighborintensive 0.413*** 0.418*** 0.472*** 0.709*** (0.000) (0.000) (0.000) (0.000) SCO 0.442** 0.403* 0.307 (0.026) (0.050) (0.125) BRICS 1.097*** 1.200*** (0.002) (0.002) Asia 1.017*** (0.000) Δ −3.900*** −3.941*** −3.721*** −3.823*** −3.745*** −3.872*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 2.232*** 2.250*** 2.169*** 2.245*** 2.186*** 2.268*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.367*** −0.369*** −0.360*** −0.374*** −0.362*** −0.377*** (0.003) (0.003) (0.005) (0.004) (0.004) (0.005) Constant 3.028 1.298 −9.968*** −10.88*** −9.515** −19.07*** (0.195) (0.592) (0.009) (0.005) (0.015) (0.000) Observations 16618 16618 16618 16618 16618 16618 Pseudo R2 0.297 0.318 0.342 0.350 0.362 0.386 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. This has two implications. First, the effect of a regional power joining the AIIB is greater than that of an average country. Second, the Neighborintensive variable is also able to explain the observation that it is usually large countries (generally hegemons) that establish IOs. The most popular supply-side argument is that hegemons alone can reap enough benefits from providing public goods.42 Here Neighborintensive raises the demand-side argument: only large economies can garner enough initial momentum from neighboring countries and attract countries farther away through network effects (Figure 4). Fig. 4. View largeDownload slide The Estimated Probability of Joining as a Function of the Neighborintensive Variable. Fig. 4. View largeDownload slide The Estimated Probability of Joining as a Function of the Neighborintensive Variable. Looking now at the second group of learning variables, SCO and BRICS, I find that countries’ membership in these two organizations also increases the likelihood of their joining the AIIB. This suggests that information also flows from the past: countries that have had more interactions with China have higher confidence in the success of the new institution. From a policy point of view, this implies that as China now plays a more active role on the global stage, and gradually consolidates its credentials, countries will respond more positively to China-initiated IOs. Contested Multilateralism: ADB Members The AIIB almost directly competes with the ADB, and the memberships largely overlap. Figure 5 (upper) shows the membership of the two institutions. The AIIB has 57 members,43 the ADB has 67, and they have 42 members in common. Prominent countries that are members of the ADB but not of the AIIB include the United States, Japan, and Canada; the three large emerging economies of Russia, Brazil, and South Africa are in the AIIB but not the ADB. Since both institutions are development banks whose focus is on Asia, this presents an ideal case for testing the framework of contested multilateralism.44 Specifically, I test the hypothesis that dissatisfied members of the ADB are more likely to join the AIIB. To operationalize the idea, I test whether ADB members that have a low representation, calculated according to their voting share to GDP ratio, are more likely to join the institution. Figure 5 (lower) shows the distribution of the representation variable. Not surprisingly, China has the lowest representation; the representation for certain early AIIB joiners like the UK and Spain is also low. However, among those countries with a higher representation ratio, Ireland and Belgium applied relatively late, while Armenia and Turkmenistan have yet to make their application. I present the formal regression results in Table 4. Table 4. Contested Multilateralism: The Case of AIIB and ADB Pr(Join = 1) 1 2 3 4 Ln GDP 0.0735 0.0728 0.124 0.0415 (0.371) (0.401) (0.187) (0.689) Ln GDP per capita −3.997*** −3.928*** −3.097** −4.452*** (0.001) (0.002) (0.016) (0.003) Ln GDP per capita2 0.197*** 0.193*** 0.139** 0.219*** (0.003) (0.006) (0.049) (0.008) Ln Distance −0.379* −0.410* −0.0242 0.326 (0.099) (0.084) (0.926) (0.368) Representation −0.181*** −0.228*** −0.157*** −0.206*** (0.000) (0.000) (0.003) (0.001) Polity 0.00275 0.0116 −0.00416 0.0195 (0.894) (0.592) (0.846) (0.429) Neighborextensive 0.267*** −0.00279 −0.00456 (0.006) (0.980) (0.968) Neighborintensive 0.612*** 0.611*** (0.000) (0.000) SCO 0.194 (0.554) Asia 0.621 (0.149) Δ −3.548*** −3.340*** −3.344*** −3.270*** (0.000) (0.000) (0.000) (0.000) Δ2 1.729*** 1.598*** 1.645** 1.587** (0.005) (0.009) (0.012) (0.015) Δ3 −0.247** −0.226* −0.235* −0.224* (0.048) (0.070) (0.077) (0.090) Constant 20.17*** 20.08*** −0.365 3.861 (0.000) (0.001) (0.965) (0.656) Observations 4373 4373 4373 4373 Pseudo R2 0.346 0.363 0.405 0.415 Pr(Join = 1) 1 2 3 4 Ln GDP 0.0735 0.0728 0.124 0.0415 (0.371) (0.401) (0.187) (0.689) Ln GDP per capita −3.997*** −3.928*** −3.097** −4.452*** (0.001) (0.002) (0.016) (0.003) Ln GDP per capita2 0.197*** 0.193*** 0.139** 0.219*** (0.003) (0.006) (0.049) (0.008) Ln Distance −0.379* −0.410* −0.0242 0.326 (0.099) (0.084) (0.926) (0.368) Representation −0.181*** −0.228*** −0.157*** −0.206*** (0.000) (0.000) (0.003) (0.001) Polity 0.00275 0.0116 −0.00416 0.0195 (0.894) (0.592) (0.846) (0.429) Neighborextensive 0.267*** −0.00279 −0.00456 (0.006) (0.980) (0.968) Neighborintensive 0.612*** 0.611*** (0.000) (0.000) SCO 0.194 (0.554) Asia 0.621 (0.149) Δ −3.548*** −3.340*** −3.344*** −3.270*** (0.000) (0.000) (0.000) (0.000) Δ2 1.729*** 1.598*** 1.645** 1.587** (0.005) (0.009) (0.012) (0.015) Δ3 −0.247** −0.226* −0.235* −0.224* (0.048) (0.070) (0.077) (0.090) Constant 20.17*** 20.08*** −0.365 3.861 (0.000) (0.001) (0.965) (0.656) Observations 4373 4373 4373 4373 Pseudo R2 0.346 0.363 0.405 0.415 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Table 4. Contested Multilateralism: The Case of AIIB and ADB Pr(Join = 1) 1 2 3 4 Ln GDP 0.0735 0.0728 0.124 0.0415 (0.371) (0.401) (0.187) (0.689) Ln GDP per capita −3.997*** −3.928*** −3.097** −4.452*** (0.001) (0.002) (0.016) (0.003) Ln GDP per capita2 0.197*** 0.193*** 0.139** 0.219*** (0.003) (0.006) (0.049) (0.008) Ln Distance −0.379* −0.410* −0.0242 0.326 (0.099) (0.084) (0.926) (0.368) Representation −0.181*** −0.228*** −0.157*** −0.206*** (0.000) (0.000) (0.003) (0.001) Polity 0.00275 0.0116 −0.00416 0.0195 (0.894) (0.592) (0.846) (0.429) Neighborextensive 0.267*** −0.00279 −0.00456 (0.006) (0.980) (0.968) Neighborintensive 0.612*** 0.611*** (0.000) (0.000) SCO 0.194 (0.554) Asia 0.621 (0.149) Δ −3.548*** −3.340*** −3.344*** −3.270*** (0.000) (0.000) (0.000) (0.000) Δ2 1.729*** 1.598*** 1.645** 1.587** (0.005) (0.009) (0.012) (0.015) Δ3 −0.247** −0.226* −0.235* −0.224* (0.048) (0.070) (0.077) (0.090) Constant 20.17*** 20.08*** −0.365 3.861 (0.000) (0.001) (0.965) (0.656) Observations 4373 4373 4373 4373 Pseudo R2 0.346 0.363 0.405 0.415 Pr(Join = 1) 1 2 3 4 Ln GDP 0.0735 0.0728 0.124 0.0415 (0.371) (0.401) (0.187) (0.689) Ln GDP per capita −3.997*** −3.928*** −3.097** −4.452*** (0.001) (0.002) (0.016) (0.003) Ln GDP per capita2 0.197*** 0.193*** 0.139** 0.219*** (0.003) (0.006) (0.049) (0.008) Ln Distance −0.379* −0.410* −0.0242 0.326 (0.099) (0.084) (0.926) (0.368) Representation −0.181*** −0.228*** −0.157*** −0.206*** (0.000) (0.000) (0.003) (0.001) Polity 0.00275 0.0116 −0.00416 0.0195 (0.894) (0.592) (0.846) (0.429) Neighborextensive 0.267*** −0.00279 −0.00456 (0.006) (0.980) (0.968) Neighborintensive 0.612*** 0.611*** (0.000) (0.000) SCO 0.194 (0.554) Asia 0.621 (0.149) Δ −3.548*** −3.340*** −3.344*** −3.270*** (0.000) (0.000) (0.000) (0.000) Δ2 1.729*** 1.598*** 1.645** 1.587** (0.005) (0.009) (0.012) (0.015) Δ3 −0.247** −0.226* −0.235* −0.224* (0.048) (0.070) (0.077) (0.090) Constant 20.17*** 20.08*** −0.365 3.861 (0.000) (0.001) (0.965) (0.656) Observations 4373 4373 4373 4373 Pseudo R2 0.346 0.363 0.405 0.415 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Fig. 5 View largeDownload slide AIIB and ADB: Contested Multilateralism. Note: The ratio for Myanmar is not shown, as the country’s GDP figures are not available. Fig. 5 View largeDownload slide AIIB and ADB: Contested Multilateralism. Note: The ratio for Myanmar is not shown, as the country’s GDP figures are not available. Across all specifications, the coefficient on representation remains negative, and is statistically significant at a 95% confidence level. This means that countries with low representation in the ADB are more likely to join the AIIB, thus confirming the argument that contested multilateralism results from the dissatisfaction of countries that have outside options. The AIIB is often described as ‘China’s World Bank’, which underlines the fact that to some extent it competes with the World Bank. For a robustness check, I also analyze how representation in the World Bank’s two main institutions—the International Bank for Reconstruction and Development (IBRD) and the International Development Association (IDA)— affects countries’ decision to join. Since there is a difference in distribution of voting power between the IBRD and the IDA, I calculate two representation metrics, one for each branch, and refer to them as IBRD Representation and IDA Representation, respectively. I find that member countries in these institutions that are under-represented are also more likely to join the AIIB, so lending further support to the contested multilateralism framework. I report the details of the results in Table 5.45 Table 5. Contested Multilateralism: Robustness Check with IBRD and IDA Pr(Join = 1) 1 2 3 Ln GDP 0.0326 0.136* 0.0393 (0.753) (0.053) (0.617) Ln GDP per capita −4.271*** −1.770*** −1.493** (0.004) (0.007) (0.031) Ln GDP per capita2 0.210** 0.0964** 0.0827** (0.011) (0.012) (0.041) Ln Distance 0.335 0.277 −0.0913 (0.353) (0.314) (0.764) Neighborextensive −0.00868 0.0524 0.123 (0.939) (0.481) (0.145) Neighborintensive 0.613*** 0.675*** 0.608*** (0.000) (0.000) (0.000) Polity 0.0176 −0.0213 −0.0274* (0.476) (0.111) (0.080) SCO 0.147 0.320 0.563** (0.658) (0.123) (0.018) BRICS 0 1.295*** 1.430*** (−) (0.001) (0.000) Asia 0.642 0.965*** 0.680** (0.134) (0.000) (0.017) (0.090) (0.006) (0.010) ADB Representation −0.198*** (0.002) IBRD Representation −0.0290 (0.163) IDA Representation −0.0242** (0.031) Δ −3.236*** −3.751*** −3.546*** (0.000) (0.000) (0.000) Δ2 1.573** 2.193*** 2.036*** (0.015) (0.000) (0.001) Δ3 −0.222* −0.365*** −0.333** (0.090) (0.006) (0.010) Constant 3.059 −13.97** −8.034 (0.723) (0.018) (0.229) Observations 4372 16618 15186 Pseudo R2 0.403 0.391 0.404 Pr(Join = 1) 1 2 3 Ln GDP 0.0326 0.136* 0.0393 (0.753) (0.053) (0.617) Ln GDP per capita −4.271*** −1.770*** −1.493** (0.004) (0.007) (0.031) Ln GDP per capita2 0.210** 0.0964** 0.0827** (0.011) (0.012) (0.041) Ln Distance 0.335 0.277 −0.0913 (0.353) (0.314) (0.764) Neighborextensive −0.00868 0.0524 0.123 (0.939) (0.481) (0.145) Neighborintensive 0.613*** 0.675*** 0.608*** (0.000) (0.000) (0.000) Polity 0.0176 −0.0213 −0.0274* (0.476) (0.111) (0.080) SCO 0.147 0.320 0.563** (0.658) (0.123) (0.018) BRICS 0 1.295*** 1.430*** (−) (0.001) (0.000) Asia 0.642 0.965*** 0.680** (0.134) (0.000) (0.017) (0.090) (0.006) (0.010) ADB Representation −0.198*** (0.002) IBRD Representation −0.0290 (0.163) IDA Representation −0.0242** (0.031) Δ −3.236*** −3.751*** −3.546*** (0.000) (0.000) (0.000) Δ2 1.573** 2.193*** 2.036*** (0.015) (0.000) (0.001) Δ3 −0.222* −0.365*** −0.333** (0.090) (0.006) (0.010) Constant 3.059 −13.97** −8.034 (0.723) (0.018) (0.229) Observations 4372 16618 15186 Pseudo R2 0.403 0.391 0.404 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Table 5. Contested Multilateralism: Robustness Check with IBRD and IDA Pr(Join = 1) 1 2 3 Ln GDP 0.0326 0.136* 0.0393 (0.753) (0.053) (0.617) Ln GDP per capita −4.271*** −1.770*** −1.493** (0.004) (0.007) (0.031) Ln GDP per capita2 0.210** 0.0964** 0.0827** (0.011) (0.012) (0.041) Ln Distance 0.335 0.277 −0.0913 (0.353) (0.314) (0.764) Neighborextensive −0.00868 0.0524 0.123 (0.939) (0.481) (0.145) Neighborintensive 0.613*** 0.675*** 0.608*** (0.000) (0.000) (0.000) Polity 0.0176 −0.0213 −0.0274* (0.476) (0.111) (0.080) SCO 0.147 0.320 0.563** (0.658) (0.123) (0.018) BRICS 0 1.295*** 1.430*** (−) (0.001) (0.000) Asia 0.642 0.965*** 0.680** (0.134) (0.000) (0.017) (0.090) (0.006) (0.010) ADB Representation −0.198*** (0.002) IBRD Representation −0.0290 (0.163) IDA Representation −0.0242** (0.031) Δ −3.236*** −3.751*** −3.546*** (0.000) (0.000) (0.000) Δ2 1.573** 2.193*** 2.036*** (0.015) (0.000) (0.001) Δ3 −0.222* −0.365*** −0.333** (0.090) (0.006) (0.010) Constant 3.059 −13.97** −8.034 (0.723) (0.018) (0.229) Observations 4372 16618 15186 Pseudo R2 0.403 0.391 0.404 Pr(Join = 1) 1 2 3 Ln GDP 0.0326 0.136* 0.0393 (0.753) (0.053) (0.617) Ln GDP per capita −4.271*** −1.770*** −1.493** (0.004) (0.007) (0.031) Ln GDP per capita2 0.210** 0.0964** 0.0827** (0.011) (0.012) (0.041) Ln Distance 0.335 0.277 −0.0913 (0.353) (0.314) (0.764) Neighborextensive −0.00868 0.0524 0.123 (0.939) (0.481) (0.145) Neighborintensive 0.613*** 0.675*** 0.608*** (0.000) (0.000) (0.000) Polity 0.0176 −0.0213 −0.0274* (0.476) (0.111) (0.080) SCO 0.147 0.320 0.563** (0.658) (0.123) (0.018) BRICS 0 1.295*** 1.430*** (−) (0.001) (0.000) Asia 0.642 0.965*** 0.680** (0.134) (0.000) (0.017) (0.090) (0.006) (0.010) ADB Representation −0.198*** (0.002) IBRD Representation −0.0290 (0.163) IDA Representation −0.0242** (0.031) Δ −3.236*** −3.751*** −3.546*** (0.000) (0.000) (0.000) Δ2 1.573** 2.193*** 2.036*** (0.015) (0.000) (0.001) Δ3 −0.222* −0.365*** −0.333** (0.090) (0.006) (0.010) Constant 3.059 −13.97** −8.034 (0.723) (0.018) (0.229) Observations 4372 16618 15186 Pseudo R2 0.403 0.391 0.404 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Cox Proportional Hazards Model So far, the major mechanisms underlying AIIB membership have been tested in a probit model. Here I test the robustness of the dynamic variables by placing them in a duration analysis framework, i.e. the wait duration before joining the AIIB. There is a natural link between binary data and duration data.46 Intuitively, countries that are more likely to join (with a higher xβ, and thus ϕ(xβ)) will have a shorter expected wait duration. This intuition can be captured by the Cox proportional hazard model: λit=λ0(t)exiβ where λt= p(t)P(T > t). If xiβ>xjβ, i.e. countryi is more likely to join than countryj, in the duration analysis framework, PTi<Tj=exiβexiβ+exjβ>0.5 with Ti and Tj denoting the waiting time for countryi and countryj, respectively. I provide details of the proof in Appendix A. With this natural transition, I can use the Cox Proportional Hazard model to test the robustness of the main results. The data are right censored. Once a country joins the AIIB, the remaining countries will enter a new period of observation with updated Neighborextensive and Neighborintensive. In the full specification (Column 4), the number of observations is 1615. The regression results are reported in Tables 6 and 7. I use the same set of variables and the same specifications as in Table 4, except that in this framework I do not include the time variables, as time dependency can be reflected in the hazard ratio.47 The number of observations is also much smaller than in the probit models, where data are constructed on a daily basis. Another difference here is that in the Cox Partial Likelihood framework, only ranking matters; the waiting time is not relevant. As several countries applied for prospective founding membership on the same day—Denmark and the Netherlands, for example, both applied on March 28, 2015—there are ties observed in the data. For robustness, I use both Efron’s method and Breslow’s method to break such ties. Table 6. Robustness Check (Efron): Wait Duration Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.302*** 0.397*** 0.428*** 0.0236 (0.011) (0.007) (0.001) (0.001) (0.911) Ln GDP per capita −0.655 −0.907 −1.018 −2.586* −2.457 (0.524) (0.420) (0.402) (0.063) (0.272) Ln GDP per capita2 0.0416 0.0548 0.0504 0.142* 0.107 (0.495) (0.412) (0.482) (0.080) (0.407) Ln Distance −1.172*** −1.444*** −0.727** 1.217** 0.731 (0.000) (0.000) (0.038) (0.045) (0.342) Polity −0.0604** −0.0731*** −0.0404 0.0303 (0.020) (0.007) (0.161) (0.519) Neighborextensive 0.391*** 0.134 −0.0213 (0.009) (0.406) (0.924) Neighborintensive 0.931*** 1.441*** 1.090*** (0.002) (0.000) (0.005) SCO 0.851* 0.115 (0.056) (0.834) Asia 2.217*** 1.342 (0.000) (0.153) ADB Representation −19.11* (0.050) Observations 2122 1615 1615 1615 391 Pseudo R2 0.117 0.142 0.203 0.248 0.172 Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.302*** 0.397*** 0.428*** 0.0236 (0.011) (0.007) (0.001) (0.001) (0.911) Ln GDP per capita −0.655 −0.907 −1.018 −2.586* −2.457 (0.524) (0.420) (0.402) (0.063) (0.272) Ln GDP per capita2 0.0416 0.0548 0.0504 0.142* 0.107 (0.495) (0.412) (0.482) (0.080) (0.407) Ln Distance −1.172*** −1.444*** −0.727** 1.217** 0.731 (0.000) (0.000) (0.038) (0.045) (0.342) Polity −0.0604** −0.0731*** −0.0404 0.0303 (0.020) (0.007) (0.161) (0.519) Neighborextensive 0.391*** 0.134 −0.0213 (0.009) (0.406) (0.924) Neighborintensive 0.931*** 1.441*** 1.090*** (0.002) (0.000) (0.005) SCO 0.851* 0.115 (0.056) (0.834) Asia 2.217*** 1.342 (0.000) (0.153) ADB Representation −19.11* (0.050) Observations 2122 1615 1615 1615 391 Pseudo R2 0.117 0.142 0.203 0.248 0.172 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Table 6. Robustness Check (Efron): Wait Duration Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.302*** 0.397*** 0.428*** 0.0236 (0.011) (0.007) (0.001) (0.001) (0.911) Ln GDP per capita −0.655 −0.907 −1.018 −2.586* −2.457 (0.524) (0.420) (0.402) (0.063) (0.272) Ln GDP per capita2 0.0416 0.0548 0.0504 0.142* 0.107 (0.495) (0.412) (0.482) (0.080) (0.407) Ln Distance −1.172*** −1.444*** −0.727** 1.217** 0.731 (0.000) (0.000) (0.038) (0.045) (0.342) Polity −0.0604** −0.0731*** −0.0404 0.0303 (0.020) (0.007) (0.161) (0.519) Neighborextensive 0.391*** 0.134 −0.0213 (0.009) (0.406) (0.924) Neighborintensive 0.931*** 1.441*** 1.090*** (0.002) (0.000) (0.005) SCO 0.851* 0.115 (0.056) (0.834) Asia 2.217*** 1.342 (0.000) (0.153) ADB Representation −19.11* (0.050) Observations 2122 1615 1615 1615 391 Pseudo R2 0.117 0.142 0.203 0.248 0.172 Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.302*** 0.397*** 0.428*** 0.0236 (0.011) (0.007) (0.001) (0.001) (0.911) Ln GDP per capita −0.655 −0.907 −1.018 −2.586* −2.457 (0.524) (0.420) (0.402) (0.063) (0.272) Ln GDP per capita2 0.0416 0.0548 0.0504 0.142* 0.107 (0.495) (0.412) (0.482) (0.080) (0.407) Ln Distance −1.172*** −1.444*** −0.727** 1.217** 0.731 (0.000) (0.000) (0.038) (0.045) (0.342) Polity −0.0604** −0.0731*** −0.0404 0.0303 (0.020) (0.007) (0.161) (0.519) Neighborextensive 0.391*** 0.134 −0.0213 (0.009) (0.406) (0.924) Neighborintensive 0.931*** 1.441*** 1.090*** (0.002) (0.000) (0.005) SCO 0.851* 0.115 (0.056) (0.834) Asia 2.217*** 1.342 (0.000) (0.153) ADB Representation −19.11* (0.050) Observations 2122 1615 1615 1615 391 Pseudo R2 0.117 0.142 0.203 0.248 0.172 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Table 7. Robustness Check (Breslow): Wait Duration Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.305*** 0.381*** 0.398*** 0.0494 (0.010) (0.006) (0.001) (0.001) (0.811) Ln GDP per capita −0.681 −0.952 −1.076 −2.359* −2.132 (0.511) (0.401) (0.381) (0.086) (0.331) Ln GDP per capita2 0.0442 0.0586 0.0572 0.133* 0.0965 (0.469) (0.379) (0.426) (0.097) (0.447) Ln Distance −1.125*** −1.325*** −0.648* 1.182* 1.019 (0.000) (0.000) (0.069) (0.055) (0.206) Polity −0.0547** −0.0658** −0.0345 0.0248 (0.032) (0.014) (0.230) (0.592) Neighborextensive 0.384** 0.137 −0.0125 (0.011) (0.400) (0.955) Neighborintensive 0.822*** 1.327*** 0.996*** (0.004) (0.000) (0.009) SCO 0.765* 0.000358 (0.087) (0.999) Asia 2.124*** 1.527 (0.001) (0.112) ADB Representation −13.77 (0.150) Observations 2122 1615 1615 1615 391 Pseudo R2 0.110 0.129 0.182 0.221 0.125 Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.305*** 0.381*** 0.398*** 0.0494 (0.010) (0.006) (0.001) (0.001) (0.811) Ln GDP per capita −0.681 −0.952 −1.076 −2.359* −2.132 (0.511) (0.401) (0.381) (0.086) (0.331) Ln GDP per capita2 0.0442 0.0586 0.0572 0.133* 0.0965 (0.469) (0.379) (0.426) (0.097) (0.447) Ln Distance −1.125*** −1.325*** −0.648* 1.182* 1.019 (0.000) (0.000) (0.069) (0.055) (0.206) Polity −0.0547** −0.0658** −0.0345 0.0248 (0.032) (0.014) (0.230) (0.592) Neighborextensive 0.384** 0.137 −0.0125 (0.011) (0.400) (0.955) Neighborintensive 0.822*** 1.327*** 0.996*** (0.004) (0.000) (0.009) SCO 0.765* 0.000358 (0.087) (0.999) Asia 2.124*** 1.527 (0.001) (0.112) ADB Representation −13.77 (0.150) Observations 2122 1615 1615 1615 391 Pseudo R2 0.110 0.129 0.182 0.221 0.125 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Table 7. Robustness Check (Breslow): Wait Duration Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.305*** 0.381*** 0.398*** 0.0494 (0.010) (0.006) (0.001) (0.001) (0.811) Ln GDP per capita −0.681 −0.952 −1.076 −2.359* −2.132 (0.511) (0.401) (0.381) (0.086) (0.331) Ln GDP per capita2 0.0442 0.0586 0.0572 0.133* 0.0965 (0.469) (0.379) (0.426) (0.097) (0.447) Ln Distance −1.125*** −1.325*** −0.648* 1.182* 1.019 (0.000) (0.000) (0.069) (0.055) (0.206) Polity −0.0547** −0.0658** −0.0345 0.0248 (0.032) (0.014) (0.230) (0.592) Neighborextensive 0.384** 0.137 −0.0125 (0.011) (0.400) (0.955) Neighborintensive 0.822*** 1.327*** 0.996*** (0.004) (0.000) (0.009) SCO 0.765* 0.000358 (0.087) (0.999) Asia 2.124*** 1.527 (0.001) (0.112) ADB Representation −13.77 (0.150) Observations 2122 1615 1615 1615 391 Pseudo R2 0.110 0.129 0.182 0.221 0.125 Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.305*** 0.381*** 0.398*** 0.0494 (0.010) (0.006) (0.001) (0.001) (0.811) Ln GDP per capita −0.681 −0.952 −1.076 −2.359* −2.132 (0.511) (0.401) (0.381) (0.086) (0.331) Ln GDP per capita2 0.0442 0.0586 0.0572 0.133* 0.0965 (0.469) (0.379) (0.426) (0.097) (0.447) Ln Distance −1.125*** −1.325*** −0.648* 1.182* 1.019 (0.000) (0.000) (0.069) (0.055) (0.206) Polity −0.0547** −0.0658** −0.0345 0.0248 (0.032) (0.014) (0.230) (0.592) Neighborextensive 0.384** 0.137 −0.0125 (0.011) (0.400) (0.955) Neighborintensive 0.822*** 1.327*** 0.996*** (0.004) (0.000) (0.009) SCO 0.765* 0.000358 (0.087) (0.999) Asia 2.124*** 1.527 (0.001) (0.112) ADB Representation −13.77 (0.150) Observations 2122 1615 1615 1615 391 Pseudo R2 0.110 0.129 0.182 0.221 0.125 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Overall, the results are very similar to the ones reported in the main results. I find that all the variables of interest, such as Neighborextensive, SCO, and Polity, retain their significance and point to the same direction as reported in Tables 3 and 4. But the coefficients now have a different interpretation: countries with a larger Neighborextensive in the AIIB tend to wait shorter periods, and so do countries with a large Neighborintensive value. But essentially, they are reading from the same mechanisms: less democratic countries are more likely to join, countries with neighbors in the AIIB are more likely to join, countries that are already members of the SCO are more likely to join, and ADB members who feel under-represented are more likely to join. Conclusion This article has provided a comprehensive modelling of the membership structure of the China-led AIIB. First, I have demonstrated that less democratic countries are more likely to join as founding members. From the perspective of risk, uncertainty, and learning, I have also shown that, when evaluating AIIB founding membership, countries learn from their neighbors and from their previous interactions with China. Third, through the case of the ADB, I have shown that countries under-represented in the ADB are more likely join the AIIB. Lastly, I have employed the Cox proportional hazard model as an alternative to demonstrate the robustness of the membership structure uncovered in this article. This article contributes to the IO literature in two aspects. First, most studies have focused on the decision-making of the leading state(s). The behavior of the participating countries is usually described as utility maximizing, but seldom receives serious attention. The phenomenon whereby China, a developing country, establishes an IO and invites developed countries to join is new. It is hence ideal for modelling the decision-making of participating countries. Second, my study has empirically tested and confirmed the contested multilateralism framework. Less represented countries (thus more dissatisfied) are shown to be more eager to join the new institution. This article also contributes to the literature on autocracy promotion and the debate on democracy and multilateralism. I find that less democratic countries are more eager to join the China-led institution than their more democratic counterparts. This could be explained by the argument (i) less democratic countries face fewer domestic political constraints on joining the China-led institution, and (ii) more democratic countries are more constrained by virtue of opposing pressure from the United States. Appendix A In this appendix, I prove that if countryi is more likely to join the IO (in the binary setting), then countryi is more likely to wait a shorter period of time before joining (in the proportional hazard duration setting). Mathematically, Φx1β>Φx2β ⇒Pr⁡T1<T2>0.5 where T denotes the wait duration. Proof: (i) The hazard function is defined as h(t) =  p(t)P(T≥t). (ii) The survival function is defined as S(t) =  PrT ≥ t= ∫t∞pτdτ, with S't=-pt. From (i) and (ii), I can derive the identify St=e-∫0th(τ)dτ. (iii) Introduce the Cox model as follows: ht=h0(t)exβ where x is the set of regressors, β is the vector of parameters, and h0(t) is the baseline. Then h2(t)h1(t)=ex2βex1β=γ2γ1⇒ S2=S1γ2γ1. Pr⁡(T1<T2)=∫0∞∫0∞p(t1, t2)dt2t1=∫0∞p(t1)∫t1∞p(t2)dt2t1=∫0∞p(t1)S2(t1)dt1=∫0∞p(t1)(S1(t1))γ2γ1dt1=−∫0∞S'(t1)(S1(t1))γ2γ1dt1=−γ1γ1+γ2(S1(t1))γ2γ1|0∞=γ1γ1+γ2=ex1βex1β+ex2β=11 + ex2β−x1β > 0.5 (iv) Appendix B Table B1. Dates of Application and Public Known Country Application Date Publicly Known Date2014 Month Day Year Month Day Year Indonesia 11 25 2014 11 15 2014 Maldives 12 17* 2014 12 31 2014 New Zealand 12 18 2014 1 1 2015 Saudi Arabia 12 31 2014 1 13 2015 Tajikistan 12 31 2014 1 13 2015 Jordan 1 24* 2015 2 7 2015 UK 3 12 2015 3 12 2015 Germany 3 17 2015 3 17 2015 France 3 17 2015 3 17 2015 Italy 3 17 2015 3 17 2015 Luxembourg 3 18 2015 3 18 2015 Switzerland 3 20 2015 3 20 2015 UAE 3 20* 2015 4 3 2015 Iran 3 20* 2015 4 3 2015 Turkey 3 26 2015 3 26 2015 Spain 3 27* 2015 4 11 2015 South Korea 3 27 2015 3 27 2015 Austria 3 27 2015 3 27 2015 Georgia 3 28 2015 3 28 2015 Denmark 3 28 2015 3 28 2015 Netherlands 3 28 2015 3 28 2015 Brazil 3 28 2015 3 28 2015 Australia 3 29 2015 3 29 2015 Finland 3 30 2015 3 30 2015 Russia 3 30 2015 3 30 2015 Norway 3 30* 2015 4 14 2015 Egypt 3 30 2015 3 30 2015 Kyrgyzstan 3 31 2015 3 31 2015 Malta 3 31* 2015 4 9 2015 Sweden 3 31 2015 3 31 2015 Israel 3 31* 2015 4 15 2015 Portugal 3 31 2015 3 31 2015 South Africa 3 31* 2015 4 15 2015 Azerbaijan 3 31* 2015 4 15 2015 Iceland 3 31 2015 3 31 2015 Poland 3 31* 2015 4 15 2015 Country Application Date Publicly Known Date2014 Month Day Year Month Day Year Indonesia 11 25 2014 11 15 2014 Maldives 12 17* 2014 12 31 2014 New Zealand 12 18 2014 1 1 2015 Saudi Arabia 12 31 2014 1 13 2015 Tajikistan 12 31 2014 1 13 2015 Jordan 1 24* 2015 2 7 2015 UK 3 12 2015 3 12 2015 Germany 3 17 2015 3 17 2015 France 3 17 2015 3 17 2015 Italy 3 17 2015 3 17 2015 Luxembourg 3 18 2015 3 18 2015 Switzerland 3 20 2015 3 20 2015 UAE 3 20* 2015 4 3 2015 Iran 3 20* 2015 4 3 2015 Turkey 3 26 2015 3 26 2015 Spain 3 27* 2015 4 11 2015 South Korea 3 27 2015 3 27 2015 Austria 3 27 2015 3 27 2015 Georgia 3 28 2015 3 28 2015 Denmark 3 28 2015 3 28 2015 Netherlands 3 28 2015 3 28 2015 Brazil 3 28 2015 3 28 2015 Australia 3 29 2015 3 29 2015 Finland 3 30 2015 3 30 2015 Russia 3 30 2015 3 30 2015 Norway 3 30* 2015 4 14 2015 Egypt 3 30 2015 3 30 2015 Kyrgyzstan 3 31 2015 3 31 2015 Malta 3 31* 2015 4 9 2015 Sweden 3 31 2015 3 31 2015 Israel 3 31* 2015 4 15 2015 Portugal 3 31 2015 3 31 2015 South Africa 3 31* 2015 4 15 2015 Azerbaijan 3 31* 2015 4 15 2015 Iceland 3 31 2015 3 31 2015 Poland 3 31* 2015 4 15 2015 Note: Estimated dates are marked with a ‘*’. The 21 countries, including China that signed the Memorandum of Understanding regarding the AIIB on October 24, 2014 in Beijing, and became prospective founding members are not listed here. The relevant documents and the code book are available on the author’s website. Table B1. Dates of Application and Public Known Country Application Date Publicly Known Date2014 Month Day Year Month Day Year Indonesia 11 25 2014 11 15 2014 Maldives 12 17* 2014 12 31 2014 New Zealand 12 18 2014 1 1 2015 Saudi Arabia 12 31 2014 1 13 2015 Tajikistan 12 31 2014 1 13 2015 Jordan 1 24* 2015 2 7 2015 UK 3 12 2015 3 12 2015 Germany 3 17 2015 3 17 2015 France 3 17 2015 3 17 2015 Italy 3 17 2015 3 17 2015 Luxembourg 3 18 2015 3 18 2015 Switzerland 3 20 2015 3 20 2015 UAE 3 20* 2015 4 3 2015 Iran 3 20* 2015 4 3 2015 Turkey 3 26 2015 3 26 2015 Spain 3 27* 2015 4 11 2015 South Korea 3 27 2015 3 27 2015 Austria 3 27 2015 3 27 2015 Georgia 3 28 2015 3 28 2015 Denmark 3 28 2015 3 28 2015 Netherlands 3 28 2015 3 28 2015 Brazil 3 28 2015 3 28 2015 Australia 3 29 2015 3 29 2015 Finland 3 30 2015 3 30 2015 Russia 3 30 2015 3 30 2015 Norway 3 30* 2015 4 14 2015 Egypt 3 30 2015 3 30 2015 Kyrgyzstan 3 31 2015 3 31 2015 Malta 3 31* 2015 4 9 2015 Sweden 3 31 2015 3 31 2015 Israel 3 31* 2015 4 15 2015 Portugal 3 31 2015 3 31 2015 South Africa 3 31* 2015 4 15 2015 Azerbaijan 3 31* 2015 4 15 2015 Iceland 3 31 2015 3 31 2015 Poland 3 31* 2015 4 15 2015 Country Application Date Publicly Known Date2014 Month Day Year Month Day Year Indonesia 11 25 2014 11 15 2014 Maldives 12 17* 2014 12 31 2014 New Zealand 12 18 2014 1 1 2015 Saudi Arabia 12 31 2014 1 13 2015 Tajikistan 12 31 2014 1 13 2015 Jordan 1 24* 2015 2 7 2015 UK 3 12 2015 3 12 2015 Germany 3 17 2015 3 17 2015 France 3 17 2015 3 17 2015 Italy 3 17 2015 3 17 2015 Luxembourg 3 18 2015 3 18 2015 Switzerland 3 20 2015 3 20 2015 UAE 3 20* 2015 4 3 2015 Iran 3 20* 2015 4 3 2015 Turkey 3 26 2015 3 26 2015 Spain 3 27* 2015 4 11 2015 South Korea 3 27 2015 3 27 2015 Austria 3 27 2015 3 27 2015 Georgia 3 28 2015 3 28 2015 Denmark 3 28 2015 3 28 2015 Netherlands 3 28 2015 3 28 2015 Brazil 3 28 2015 3 28 2015 Australia 3 29 2015 3 29 2015 Finland 3 30 2015 3 30 2015 Russia 3 30 2015 3 30 2015 Norway 3 30* 2015 4 14 2015 Egypt 3 30 2015 3 30 2015 Kyrgyzstan 3 31 2015 3 31 2015 Malta 3 31* 2015 4 9 2015 Sweden 3 31 2015 3 31 2015 Israel 3 31* 2015 4 15 2015 Portugal 3 31 2015 3 31 2015 South Africa 3 31* 2015 4 15 2015 Azerbaijan 3 31* 2015 4 15 2015 Iceland 3 31 2015 3 31 2015 Poland 3 31* 2015 4 15 2015 Note: Estimated dates are marked with a ‘*’. The 21 countries, including China that signed the Memorandum of Understanding regarding the AIIB on October 24, 2014 in Beijing, and became prospective founding members are not listed here. The relevant documents and the code book are available on the author’s website. Appendix C Table C1. United Nations Member States that Recognize Taiwan Burkina Faso, Belize, Dominican Republic El Salvador, Guatemala, Guinea Bissau Bissau Haiti, Honduras, Kiribati Nauru, Nicaragua, Paraguay Palau, Panama, Sao Tome, and Principe St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines Solomon Islands, Swaziland, Tuvalu Burkina Faso, Belize, Dominican Republic El Salvador, Guatemala, Guinea Bissau Bissau Haiti, Honduras, Kiribati Nauru, Nicaragua, Paraguay Palau, Panama, Sao Tome, and Principe St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines Solomon Islands, Swaziland, Tuvalu Table C1. United Nations Member States that Recognize Taiwan Burkina Faso, Belize, Dominican Republic El Salvador, Guatemala, Guinea Bissau Bissau Haiti, Honduras, Kiribati Nauru, Nicaragua, Paraguay Palau, Panama, Sao Tome, and Principe St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines Solomon Islands, Swaziland, Tuvalu Burkina Faso, Belize, Dominican Republic El Salvador, Guatemala, Guinea Bissau Bissau Haiti, Honduras, Kiribati Nauru, Nicaragua, Paraguay Palau, Panama, Sao Tome, and Principe St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines Solomon Islands, Swaziland, Tuvalu Appendix D Table D1. Main Variables and Sources of Raw Data Variable Explanation Source Ln GDP Log of country GDP World Bank Ln GDP per capita Log of country GDP per capita World Bank Ln Distance Log of distance between countryi and China CEPII Polity Democracy index Polity IV Project Political Constraints Measure of political constraints W. J. Henisz (2000) Political Rights Measure of political rights Freedom House Civil Liberties Measure of civil liberty Freedom House Taiwan A dummy on whether country i recognizes Taiwan Xinhua ADB Representation Ratio of ADB voting share to GDP ADB IBRD Representation Ratio of IBRD voting share to GDP World Bank IDA Representation Ratio of IDA voting share to GDP World Bank Variable Explanation Source Ln GDP Log of country GDP World Bank Ln GDP per capita Log of country GDP per capita World Bank Ln Distance Log of distance between countryi and China CEPII Polity Democracy index Polity IV Project Political Constraints Measure of political constraints W. J. Henisz (2000) Political Rights Measure of political rights Freedom House Civil Liberties Measure of civil liberty Freedom House Taiwan A dummy on whether country i recognizes Taiwan Xinhua ADB Representation Ratio of ADB voting share to GDP ADB IBRD Representation Ratio of IBRD voting share to GDP World Bank IDA Representation Ratio of IDA voting share to GDP World Bank Table D1. Main Variables and Sources of Raw Data Variable Explanation Source Ln GDP Log of country GDP World Bank Ln GDP per capita Log of country GDP per capita World Bank Ln Distance Log of distance between countryi and China CEPII Polity Democracy index Polity IV Project Political Constraints Measure of political constraints W. J. Henisz (2000) Political Rights Measure of political rights Freedom House Civil Liberties Measure of civil liberty Freedom House Taiwan A dummy on whether country i recognizes Taiwan Xinhua ADB Representation Ratio of ADB voting share to GDP ADB IBRD Representation Ratio of IBRD voting share to GDP World Bank IDA Representation Ratio of IDA voting share to GDP World Bank Variable Explanation Source Ln GDP Log of country GDP World Bank Ln GDP per capita Log of country GDP per capita World Bank Ln Distance Log of distance between countryi and China CEPII Polity Democracy index Polity IV Project Political Constraints Measure of political constraints W. J. Henisz (2000) Political Rights Measure of political rights Freedom House Civil Liberties Measure of civil liberty Freedom House Taiwan A dummy on whether country i recognizes Taiwan Xinhua ADB Representation Ratio of ADB voting share to GDP ADB IBRD Representation Ratio of IBRD voting share to GDP World Bank IDA Representation Ratio of IDA voting share to GDP World Bank Footnotes 1 James E. Rauch, ‘Bureaucracy, Infrastructure, and Economic Growth: Evidence from U.S. Cities During the Progressive Era’, American Economic Review, Vol. 85, No. 4 (1995), pp. 968–79; Sylvie Démurger, ‘Infrastructure Development and Economic Growth: An Explanation for Regional Disparities in China’, Journal of Comparative Economics, Vol. 29, No. 1 (2001), pp. 95–117. 2 Vivien Foster and Cecilia Briceño-Garmendia, eds., Africa’s Infrastructure: A Time for Transformation (Washington, D.C.: World Bank, 2015), pp. 47–49; Jeffrey Gutman, Amadou Sy, and Soumya Chattopadhyay, Financing African Infrastructure: Can the World Deliver? (Washington, D.C.: Brookings, 2015), p. 1. 3 Asian Development Bank, Meeting Asia’s Infrastructure Needs (Mandaluyong City: Asian Development Bank, 2017), https://www.adb.org/sites/default/files/publication/227496/special-report-infrastructure.pdf. 4 The original statement from the AIIB can be found at https://www.aiib.org/en/news-events/news/2017/20171219_001.html. 5 I use the present tense here because the findings are applicable to other IOs established by China and more generally by other developing countries. 6 Barbara Koremenos, ‘Contracting around International Uncertainty’, American Political Science Review, Vol. 99, No. 4 (2005), pp. 549–65. 7 Data on countries’ voting power can be found in the IMF 2015 Annual Report, https://www.imf.org/external/pubs/ft/ar/2015/eng/pdf/AR15-AppIV.pdf. 8 Phillip Y. Lipscy, Renegotiating the World Order: Institutional Change in International Relations (Cambridge: Cambridge University Press, 2017); Julia C. Morse and Robert O. Keohane, ‘Contested Multilateralism’, Review of International Organizations, Vol. 9, No. 4 (2014), pp. 385–412. 9 For AIIB membership information, see http://www.aiib.org/html/pagemembers. 10 For discussion on collective action, see Randall W. Stone, Branislav L. Slantchev, and Tamar R. London, ‘Choosing How to Cooperate: A Repeated Public-Goods Model of International Relations’, International Studies Quarterly, Vol. 52, No. 2 (2008), pp. 335–62; Randall W. Stone, ‘Institutions, Power, and Interdependence’, in Helen V. Milner and Andrew Moravcsik, eds., Power, Interdependence, and Nonstate Actors in World Politics (Princeton: Princeton University Press, 2009), pp. 31–49. For discussion on transaction costs, see Robert O. Keohane, ‘The Demand for International Regimes’, International Organization, Vol. 36, No. 2 (1982), pp. 325–55; Darren G. Hawkins, David A. Lake, Daniel L. Nielson, and Michael J. Tierney, ‘Delegation under Anarchy: States, International Organizations, and Principal-Agent Theory’, in Darren G. Hawkins, David A. Lake, Daniel L. Nielson, and Michael J. Tierney, eds., Delegation and Agency in International Organizations (Cambridge: Cambridge University Press, 2009), pp. 3–38. 11 Randall W. Stone, Controlling Institutions: International Organizations and the Global Economy (Cambridge: Cambridge University Press, 2011). 12 James D. Fearon, ‘Domestic Political Audiences and the Escalation of International Disputes’, American Political Science Review, Vol. 88, No. 3 (1994), pp. 577–92. 13 George W. Downs, David M. Rocke, and Peter N. Barsoom, ‘Managing the Evolution of Multilateralism’, International Organization, Vol. 52, No. 2 (1998), pp. 397–419; Robert D. Putnam, ‘Diplomacy and Domestic Politics: The Logic of Two-Level Games’, International Organization, Vol. 42, No. 3 (1988), pp. 427–60; Robert O. Keohane, ‘International Institutions: Two Approaches’, International Studies Quarterly, Vol. 32, No. 4 (1988), pp. 379–96. 14 For discussion on ‘domestic gridlock’, see Samuel Brazys and Diana Panke, ‘Why do States Change Positions in the United Nations General Assembly?’, International Political Science Review, Vol. 38, No. 1 (2017), pp. 70–84. For discussion on homophily, see Miller McPherson, Lynn Smith-Lovin, and James M Cook, ‘Birds of a Feather: Homophily in Social Networks’, Annual Review of Sociology, Vol. 27 (2001), pp. 415–44. 15 Alberto Alesina and David Dollar, ‘Who Gives Foreign Aid to Whom and Why?’, Journal of Economic Growth, Vol. 5, No. 1 (2000), pp. 33–63. 16 David B. Carter and Randall W. Stone, ‘Democracy and Multilateralism: The Case of Vote Buying in the UN General Assembly’, International Organization, Vol. 69, No. 1 (2015), pp. 1–33. 17 Julia Bader, ‘China, Autocratic Patron? An Empirical Investigation of China as a Factor in Autocratic Survival’, International Studies Quarterly, Vol. 59, No. 1 (2015), pp. 23–33; Julia Bader, ‘Propping up Dictators? Economic Cooperation from China and its Impact on Authoritarian Persistence in Party and Non-Party Regimes’, European Journal of Political Research, Vol. 54, No. 4 (2015), pp. 655–72; Julia Bader and Ursula Daxecker, ‘A Chinese Resource Curse? The Human Rights Effects of Oil Export Dependence on China versus the United States’, Journal of Peace Research, Vol. 52, No. 6 (2015), pp. 774–90. 18 For discussion on domestic political constraints, see Alexandre Debs and Jessica Chen Weiss, ‘Circumstances, Domestic Audiences, and Reputational Incentives in International Crisis Bargaining’, Journal of Conflict Resolution, Vol. 60, No. 3 (2014), pp. 403–33; Adam Przeworski, Susan C. Stokes, and Bernard Manin, eds., Democracy, Accountability, and Representation (Cambridge: Cambridge University Press, 1999). 19 An alternative mechanism is competition. While China is a developing country, it is also the world’s second largest economy, and poised to become the world’s largest economy in the next decade or so. Therefore, countries do have an incentive to develop and maintain a good relationship with China, and joining the AIIB early as a founding member is a good opportunity to do so. Unfortunately, there is no easy way to distinguish whether it is the flow of information or the flow of competition that gives rise to the observed neighbor effects. In this article, I will stick to the mechanism of information flow. 20 Andreu Mas-Colell, Michael D. Whinston, and Jerry R. Green, Microeconomic Theory (Oxford: Oxford University Press, 2012); John W. Pratt, ‘Risk Aversion in the Small and in the Large’, Econometrica, Vol. 32, No. 1/2 (1964), pp. 122–36. 21 For discussion on interdependence, see Robert O. Keohane and Joseph S. Nye, Jr., ‘Power and Interdependence Revisited’, International Organization, Vol. 41, No. 4 (1987), pp. 725–53. For discussion on transnational diffusion, see Fabrizio Gilardi, ‘Transnational Diffusion: Norms, Ideas, and Policies’, in Walter Carlsnaes, Thomas Risse, and Beth Simmons, eds., Handbook of International Relations (Thousand Oaks: SAGE Publications, 2012), pp. 453–77; Beth A. Simmons and Zachary Elkins, ‘The Globalization of Liberalization: Policy Diffusion in the International Political Economy’, American Political Science Review, Vol. 98, No. 1 (2004), pp. 171–89. For discussion on interorganizational learning, see Johan Bruneel and Bart Clarysse, ‘Learning from Experience and Learning from Others: How Congenital and Interorganizational Learning Substitute for Experiential Learning in Young Firm Internationalization’, Strategic Entrepreneurship Journal, No. 4 (2010), pp. 164–82. 22 Kenneth N. Waltz, Theory of International Politics (New York: McGraw-Hill Publishing Company, 1979); Robert Powell, In the Shadow of Power: States and Strategies in International Politics (Princeton: Princeton University Press, 1999); David A. Lake, Hierarchy in International Relations (Ithaca: Cornell University Press, 2009). 23 For discussion on how institutional changes are affected by an IO’s policy area, see Phillip Y. Lipscy, ‘Explaining Institutional Change: Policy Areas, Outside Options, and the Bretton Woods Institutions’, American Journal of Political Science, Vol. 59, No. 2 (2015), pp. 341–56. 24 Julia C. Morse and Robert O. Keohane, ‘Contested Multilateralism’, pp. 385–412; Joseph Jupille, Walter Mattli, and Duncan Snidal, Institutional Choice and Global Commerce (Cambridge: Cambridge University Press, 2013); Robert O. Keohane and David G. Victor, ‘The Regime Complex for Climate Change’, Perspectives on Politics, Vol. 9, No. 1 (2011), pp. 7–23. 25 G. John Ikenberry and Darren Lim, China’s Emerging Institutional Statecraft: The Asian Infrastructure Investment Bank and the Prospects for Counter-Hegemony (Washington, D.C.: Brookings, 2017), https://www.brookings.edu/wp-content/uploads/2017/04/chinas-emerging-institutional-statecraft.pdf. 26 The framework of contested multilateralism mostly focuses on the big powers. My work, in contrast, studies the decision making of the ‘small’ powers. 27 Δt is reset to 0 whenever a country joins and is incremented by 1 each day thereafter. 28 This set of variables is also standard in international trade literature. See, for example, Paul Krugman, ‘Scale Economies, Product Differentiation, and the Pattern of Trade’, American Economic Review, Vol. 70, No. 5 (1980), pp. 950–59; Thomas Chaney, ‘The Gravity Equation in International Trade: An Explanation’, Journal of Political Economy, Vol. 126, No. 1 (2018), pp. 150–77. 29 Phillip Y. Lipscy, ‘Who’s Afraid of the AIIB’, Foreign Affairs, 7 May, 2017, https://www.foreignaffairs.com/articles/china/2015-05-07/whos-afraid-aiib. 30 The latter argument is also supported by the latest AIIB data. As of February 2018, the AIIB has approved 25 projects, and all of these projects are located in member countries. For details of these projects, please see https://www.aiib.org/en/projects/approved/index.html. 31 Its design follows Thomas Chaney, ‘The Network Structure of International Trade’, American Economic Review, Vol. 104, No. 11 (2014), pp. 3600–34. The naming convention follows Thomas Chaney, ‘Distorted Gravity: The Intensive and Extensive Margins of International Trade’, American Economic Review, Vol. 98, No. 4 (2008), pp. 1707–21. 32 David B. Carter and Curtis S. Signorino, ‘Back to the Future: Modeling Time Dependence in Binary Data’, Political Analysis, Vol. 18, No. 3 (2010), pp. 271–92. 33 For alternative ways to capture time dependency, such as time dummies and/or cubic splines, see Nathaniel Beck, Jonathan N. Katz, and Richard Tucker, ‘Taking Time Seriously: Time-Series-Cross-Section Analysis with a Binary Dependent Variable’, American Journal of Political Science, Vol. 42, No. 4 (1998), pp. 1260–88. 34 Janet M. Box-Steffensmeier and Bradford S. Jones, Event History Modeling: A Guide for Social Scientists (Cambridge: Cambridge University Press, 2004). 35 I use GDP and GDP per capita data for the year 2013 rather than 2014 because data with respect to the latter year on many countries have not yet been reported. 36 http://www.cepii.fr/cepii/en/bdd_modele/bdd.asp. 37 For polity scores, see http://www.systemicpeace.org/polity/polity4.htm. For political constraints, see W. J. Henisz, ‘The Institutional Environment for Economic Growth’, Economics & Politics, Vol. 12, No. 1 (2000), pp. 1–31. For data on political rights and civil liberties, see https://freedomhouse.org/report/freedom-world-2016/table-scores. 38 http://eng.sectsco.org. 39 http://www.cepii.fr/cepii/en/bdd_modele/bdd.asp. 40 http://english.mofcom.gov.cn. 41 Please note that scores for political rights and civil liberties range between 1 and 7, with 1 representing the freest and 7 the least free. This explains why the coefficients of political rights and civil liberties bear different signs from those of polity scores, which range from −10 to 10, with −10 representing autocracy and 10 full democracy. 42 Stone, Slantchev, and London, ‘Choosing How to Cooperate’, pp. 31–49; Robert O. Keohane, After Hegemony: Cooperation and Discord in the World Political Economy (Princeton: Princeton University Press, 2005). 43 Here I only consider the 57 founding members, as they are the focus of this study. As of June 16, 2017, the number of approved memberships had risen to 80. 44 Julia C. Morse and Robert O. Keohane, ‘Contested Multilateralism’, pp. 385–412. 45 I find that the correlation between the variable Asia and representation in the World Bank to be negative. In particular, the negative correlation between Asia and IDA representation is statistically significant. This highlights the general under-representation of Asian countries in the world’s financial system. 46 Beck, Katz, and Tucker, ‘Taking Time Seriously’, pp. 1260–88; Kjell A. Doksum and Miriam Gasko, ‘On a Correspondence between Models in Binary Regression Analysis and in Survival Analysis’, International Statistical Review, Vol. 58, No. 3 (1990), pp. 243–52. 47 Beck, Katz, and Tucker, ‘Taking Time Seriously’, pp. 1260–88; Carter and Signorino, ‘Back to the Future’, pp. 271–92. © The Author(s) 2018. Published by Oxford University Press on behalf of The Institute of International Relations, Tsinghua University. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The Chinese Journal of International Politics Oxford University Press

The Political Economy of Joining the AIIB

Loading next page...
 
/lp/ou_press/the-political-economy-of-joining-the-aiib-exY8SfIclx
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of The Institute of International Relations, Tsinghua University. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
1750-8916
eISSN
1750-8924
D.O.I.
10.1093/cjip/poy006
Publisher site
See Article on Publisher Site

Abstract

Abstract This article analyzes the determinants of new prospective members of the Asian Infrastructure Investment Bank (AIIB). I argue that less democratic countries are more likely to apply, and that when deciding to join the institution countries learn from their neighbors and from any previous international organization (IO) interactions they have had with China. Building on detailed panel data covering the institution’s founding period, i.e. from October 24, 2014 to March 31, 2015, I fit both a probit model with time polynomials and a Cox duration model to identify country characteristics that correlate with joining the AIIB. I show that countries with lower polity scores or whose neighbors had already become members were more likely to join, and that the probability was higher for countries that belonged to China-led IOs. Lastly, I show that countries under-represented in the existing international financial system were also more likely to join. My findings highlight the importance of democracy in shaping the membership structure of the AIIB, and demonstrate how countries leverage information from their neighbors and from previous interactions with China to adjust the perceived risk of joining. They also provide the first modern-day empirical support for the contested multilateralism framework. Introduction That good infrastructure helps economic growth and poor infrastructure hinders it is well known.1 Equally well known is the fact that many of the developing countries, for example, in Asia and Africa have long suffered a basic infrastructure deficit.2 A report in 2017 by the Asian Development Bank (ADB) states that Asia will need to invest $22.6 trillion between 2016 and 2030 to maintain its growth momentum and eradicate poverty.3 When China proposed setting up the Asian Infrastructure Investment Bank (AIIB) to, ‘bring countries together to address the daunting infrastructure needs across Asia and beyond’,4 however, responses throughout the world were divided. Some countries, such as India and Mongolia, joined immediately. Others, including Australia and South Korea, were more reticent, and did not decide to join till a later stage. Meanwhile a few others, most notably the United States and Japan, declared that they would not even consider joining. Given the significance of the new institution and the urgent needs of the developing world, how did countries decide whether and when to join as prospective founding members? Figure 1 illustrates the expansion of AIIB membership between October 24, 2014 and March 31, 2015. During this period, 57 countries joined as prospective founding members. On the whole, it was countries close to China, such as Pakistan and Kazakhstan, that were the early joiners, and countries farther away that either joined later or decided not to. However, it is not easy through such high-level aggregate observations alone to pinpoint the relative importance of the various factors at play at the time a country was making the decision. Fig. 1. View largeDownload slide Geographic Expansion of the AIIB Membership. Note: The graph presents all 56 founding members of the AIIB other than China. Due to space constraints, I have not reported all the names. It can be clearly seen how AIIB membership expands geographically overtime (the x axis). Given its proximity to China, South Korea’s late joining makes it an outlier on the graph, not to say Japan’s no-show. Fig. 1. View largeDownload slide Geographic Expansion of the AIIB Membership. Note: The graph presents all 56 founding members of the AIIB other than China. Due to space constraints, I have not reported all the names. It can be clearly seen how AIIB membership expands geographically overtime (the x axis). Given its proximity to China, South Korea’s late joining makes it an outlier on the graph, not to say Japan’s no-show. In this article, I model the expansion of the AIIB prospective founding membership to analyze how countries weigh the various factors in this significant but uncertain process. I consider countries’ economic characteristics, such as GDP and GDP per capita; dyad traits, such as physical distance from China, and countries that had already become AIIB members; political characteristics, such as polity score and civil liberties; and existing international organization (IO) relationships, including the ADB, the Shanghai Cooperation Organization (SCO), and the World Bank. Building from an original dataset of membership expansion dates, I estimate both a probit model with time polynomials and a Cox duration model to analyze the probability of applying to become AIIB prospective founding members. To preview the results, I find that less democratic countries are more likely to join the AIIB as founding members. This result is robust across all the metrics: polity score, political constraints, civil liberties, and political rights.5 Consistent with our overall observation from Figure 1, I find that countries located closer to China are more likely to join than countries farther away. This physical barrier breaks down, however, when neighboring countries take the lead in joining the IO. For example, since the UK joined the AIIB, the 8017 kilometre distance between France and China has become less relevant for the French government. Similarly, Russia’s joining the IO has significantly reduced the distance between Bishkek and Beijing as far as Kyrgyzstan is concerned. This suggests that it is the intrinsic uncertainty involved in joining an IO, rather than physical distance from it, that affects countries’ calculations.6 Such uncertainty is lower for countries that have previously interacted with China in other China-led IOs, for example, the SCO. Finally, I demonstrate that countries under-represented in the current international financial system are more likely to join the AIIB as a fresh alternative. Facts show that China itself has long been under-represented in the current international financial system. As of 2015, China’s voting share for the International Monetary Fund (IMF) was a meagre 3.81%, substantially smaller than that for Germany, the UK, or France,7 even though China had become the world’s second largest economy. Similarly, the voting share for India, Bangladesh, Bhutan, and Sri Lanka combined amounted to less than half that for Germany. The creation of the AIIB and the positive responses of under-represented countries thus lend empirical support to the contested multilateralism framework.8 The rest of the article is organized as follows. ‘Background and theory’ section presents the background, theories, and related hypotheses. ‘Model formulation and data’ section presents the econometric model. ‘Data’ section presents the dataset and its summary statistics. ‘Empirical analysis’ section presents and discusses the results. Last section concludes. Background and Theory The AIIB is a multilateral financial institution which the Chinese government initiated in 2014. China is its largest contributing member. Fifty-seven countries have joined the AIIB as prospective founding members,9 and by June 16, 2017, 23 other countries, including Argentina, Belgium, and Canada, had joined as non-founding members. The AIIB’s membership thus quickly eclipsed that of the Japan-led ADB, which currently has 67 members. With an initial capitalization of $100 billion, and with members ranging from South Africa, Egypt, and Great Britain to Brazil and Chile, the AIIB stands out as an important IO, both economically and politically. At the same time, the sequential nature of the expansion of the AIIB’s membership presents researchers with an ideal opportunity to examine several important rational choice-based questions that can shed light on the AIIB expansion process, and provide guidance for the future expansion of similar IOs. IOs such as the AIIB are created to facilitate collective action and to lower transaction costs.10 Over the past few decades, IOs have proliferated, expanded their memberships, and extended their influence to include more and more new issue areas. The vast majority of these IOs, especially the most prominent ones such as the IMF and World Bank, were created by the world’s most developed countries. More recently, the developing world led by the BRICS countries also started to establish IOs to advance economic development and facilitate multilateral cooperation. The New Development Bank, the SCO, and the AIIB are but a few examples. Although joining an IO is a risky proposition filled with uncertainty, countries nevertheless come together under the common belief in mutual benefits for big and small players alike.11 Therefore, this article assumes that countries that stand to benefit from AIIB membership choose to join while others do not. This leads us to several testable hypotheses. Hypothesis 1: Compared with democracies, autocracies are more likely to join the AIIB. The first hypothesis is based on the observation that autocracies face fewer political constraints than democracies with respect to joining the China-led institution.12 Such political constraints have both domestic and international sources. Like any other international negotiation, joining the AIIB can be viewed as a two-level bargaining issue, whereby leaders of potential member countries negotiate with the Chinese government at the international level, and with their congress or parliament at the domestic level.13 The political constraints on democracies are likely to be greater at the domestic level because (i) leaders will need the approval of congress or parliament, and might fall into a ‘democratic gridlock’ and (ii) there is greater ideological distance between China and the prospective country.14 At the international level, in addition to the Chinese government, the United States also played an important role in shaping the membership structure, because the United States was highly sceptical of the China-initiated institution, and tried publicly to dissuade its Western allies from joining. Implicitly, the United States could also put substantial pressure on countries that rely on US foreign aid.15 Research has demonstrated that the United States exerts more influence over aid-receiving democracies than it does on aid-receiving autocracies, as aid linkages to a country’s decision on AIIB membership are more credible when directed toward democracies.16 Consequently, pressure from the United States, both public and in private, further shrinks the win-set for democratic governments, making democracies even less likely to join. The hypothesis connects this study to the burgeoning literature of autocracy promotion that investigates the effects of interactions between China and autocracies on the regime duration, and more generally on the literature on foreign aid and democratization.17 While most of the existing studies focus on autocracies that are recipients of Chinese loans and aid, my article examines a completely new angle: the different reaction of autocracies compared to that of democracies to a Chinese initiative. Viewed from the perspectives of both domestic political constraints and of international pressure from the United States, it is reasonable to expect that autocracies are more likely than democracies to join the AIIB.18 Next, I examine the role of uncertainty and learning in shaping the AIIB membership structure. As with other IOs, the process of joining the AIIB and the institution itself are filled with uncertainty. That China is a developing country and the AIIB is one of the first IOs China has initiated heightens the degree of uncertainty.19 How could countries rationally commit themselves to this new IO? Assuming that countries are risk averse, those that have greater confidence in the China-led institution, and in China in general, will have a higher utility in joining (Figure 2). Mathematically, I formulate this argument as follows: u(AIIB) = u(E(AIIBperformance)) = u((1 − p) · AIIBfail + p ·AIIBsucceed) where u is concave and twice differentiable, AIIBsucceed denotes AIIB’s success, AIIBfail represents the failure of the institution, and p is probability of AIIBsucceed. The assumption that countries are risk averse is reflected by the fact that u′ > 0 and u″ < 0.20 As countries are interdependent and information flows across borders, countries learn from their neighbors to mitigate the lack of information, similar to interorganizational learning in organizational learning theory.21 Consequently, a country’s confidence (p) in AIIB should increase if neighboring countries have already joined. Similarly, countries that have past experience with China could place a higher p on AIIBsucceed as well. As China is a developing country and its credentials in the international community are not fully established, at least as compared with the UK or the United States, previous interactions with China should play an important role in the decision-making process, and countries that have had such interactions are able to achieve a higher certainty equivalent for joining the AIIB. Fig. 2. View largeDownload slide Information and AIIB Performance. Note: As countries accumulate more information about the IO, E(AIIB performance) goes up and the purple dot (second from the left) moves toward the green dot (third from the left) along the utility curve. Fig. 2. View largeDownload slide Information and AIIB Performance. Note: As countries accumulate more information about the IO, E(AIIB performance) goes up and the purple dot (second from the left) moves toward the green dot (third from the left) along the utility curve. Hypothesis 2: Countries with neighbors that are joining or that have had previous interactions with China are more likely to join. Lastly, I examine the role of representation in the international financial system. Ideally, the structure of the system is determined by the distribution of capabilities across the countries.22 In reality, however, the structure of the system usually lags behind the evolution of capabilities, leading to the observation that some countries are over-represented and others under-represented.23 If fair representation is desirable, then I expect that countries which are currently under-represented would be able to increase their utility by joining a new IO (with the intuition that the marginal utility is for them substantially higher than that for fairly or over-represented countries). When a mismatch between power and IO resources occurs and is not addressed, and when existing IOs cannot meet global demands, for example, in fighting climate change, this will add to pressure for regime contestation, in the form either of regime shifting or regime creation.24 The China-initiated AIIB attests to this argument. As China’s power in the international system grows, it seeks greater representation in existing IOs like the IMF and the World Bank. Such aspirations, however, have been constantly thwarted, thus creating the chance for China to establish ‘a World Bank of its own’.25 Smaller countries that have no outside option and feel under-represented are not able to push for adjustment within existing IOs, and nor could they credibly establish new IOs.26 Therefore, the pent-up demand for representation should naturally lead them to become the earliest to participate in newly created IOs, such as the AIIB, both to signal dissatisfaction with the status quo and to secure a better representation in the new IOs. This leads to the third testable hypothesis. Hypothesis 3: Compared with well-represented countries, countries that are under-represented in the existing IOs are more likely to join, and to join early. Model Formulation and Data Model Formulation To test these hypotheses, I formulate a random utility model where countryi has the utility function specified as follows, and will join if its utility is greater than 0. The key variables of interest are Polity (Hypothesis 1), Neighborextensive(i, t), Neighborintensive(i, t) and IO (Hypothesis 2), and Representation(i) (Hypothesis 3). Neighborextensive(i, t) and Neighborintensive(i, t) are updated daily between October 24, 2014 and March 31, 2015. Where GDP is the country’s economic size, GDP per capita is the income level; Distance is the geographical distance between the country and China; IO variables are binary, indicating whether or not countryi is a member of an IO of interest; Asiai is a dummy variable indicating whether the countryi is an Asian country or not; Δt is a duration variable representing the number of days since the most recent country joined; and εi, t has a standard normal distribution and is i.i.d.27 I include country GDP and physical distance in the model because empirical observations show that countries close to China are the first ones to join, and large economies tend to join the AIIB earlier than smaller ones when controlling for geographic distances. I report this pattern in Figure 1. This suggests that GDP size and geographic distance weigh heavily in states’ calculations.28 Since the AIIB is an investment bank in nature, I expect that, all other things being equal, the probability of rich countries joining the AIIB should be higher than of poor countries, as they have the resources to contribute to the institution and are generally eager to ‘shape the trajectory’ of it.29 On the other hand, poor countries also have an incentive to join, as they will likely need the AIIB to help fund infrastructure projects.30 Therefore, I include both linear and quadratic terms of GDP per capita in the modelling, and expect the coefficient on the quadratic term to be positive. Neighborextensive(i, t) captures the extensiveness of the AIIB’s attraction for countryi at time t. It is defined as the number of neighboring countries already in the IO for countryi at time t. Neighborextensivei, t= Σj∈IO⁡Neighbor(i, j, t) Neighborintensive(i, t) captures the intensity of attraction of the AIIB for countryi at time t.31 Neighborintensivei, t=maxj∈IO⁡logGDP(j)Distance(i, j) Given that not all neighbors carry the same weight, this variable Neighborintensive is so constructed as to capture the effect of important neighbors: when a new member has joined the AIIB, Neighborextensive will be updated according to maxj∈IO⁡logGDP(j)Distance(i, j). By design, countries with small economies and countries located far away will not affect country i’s utility. In contrast with Neighborextensive, Neighborintensive is not restricted to physically contiguous countries. The UK’s application, for example, can thus affect the decisions of Germany and Italy through this channel. Polityi is the polity score of countryi. The higher the polity score, the more democratic the country is. Representationi measures how well represented the country is in the current international system, and is the key variable for testing the hypothesis that under-represented countries are the ones that join the AIIB, and that they join early. Δt and its polynomial terms are aimed at capturing time dependency.32 From the parameters λ1, λ2 and λ3, I will be able to test the existence of momentum effects.33 This is closely related to the hazard rate concept in the literature on duration analysis.34 The dynamics of the model play out as follows. In Period 1, (only) the founder joins the IO. Variables Neighborextensive and Neighborintensive are updated for each country. Countries with positive utility choose to join. The world enters Period 2, with Neighborextensive and Neighborintensive updated for countries not yet in the IO. Countries that have not yet joined calculate their utility for Period 2, and decide whether or not to join. So, in Period t, countryi that is not yet in the IO will decide again with updated Neighborextensive and Neighborintensive. There is a finite number of periods, as there is a deadline for applications for founding member status in the IO. The key assumption of the model is that countryj’s joining the IO will affect the subsequent calculations of all the non-Member States through two channels: Neighborextensive and Neighborintensive. In terms of marginal utility, this can be expressed as: Neighborextensive:ΔUi=β4, if country i and country j are continugous0, otherwise Neighborintensive:∂Ui∂Neighborintensive(i)=β5 Learning from the past, in the sense that countries that are already members of a China-led IO are more likely to join the AIIB, can be captured by Γ. I will estimate the random utility model in ‘Empirical analysis’ section using probit, but first let me introduce the data. Data The central item of data in my article consists in dates of application, which I present in Appendix B. Data on other variables are from standard sources. Importantly, I restrict my sample countries to United Nations Member States that recognize China (not Taiwan). Data with respect to countries that recognize Taiwan, listed in Appendix C8, are from Xinhua. The dataset comprises a total of 170 countries, and once one joins the AIIB it will drop out of the sample. Data on country GDP and GDP per capita come from the World Bank.35 Both GDP and GDP per capita are on log scale. Data on geographical distance and physical contiguity come from Centre d’Etudes Prospectives et d’Informations Internationales (CEPII).36 Distances are measured in kilometres, and in this article distance represents the natural logarithm of the physical distance. A key question I address in the article is whether or not less democratic countries are more likely to join the AIIB. For this purpose, I use four alternative measures of democracy: polity scores from the Polity IV Project, political constraints, and political rights, and civil liberties from Freedom House.37 To study the effects of learning from IO interactions, I construct the binary variable IO. The variable IO will take value 1 if countryi is a member of an IO of interest. In this article, I will use the BRICS and the SCO. BRICS is an (informal) IO that consists of Brazil, Russia, India, China, and South Africa. All five countries are founding members, but joined at different times. The SCO was cofounded by China, Kazakhstan, Kyrgyzstan, Russia, Tajikistan, and Uzbekistan in 2001. I present the detailed membership information in Table 1.38 I code the variable SCO as 1 for the 18 countries other than China in the SCO, and 0 for all other countries. Here I do not distinguish between formal members, observer states, and dialogue partners. Table 1 SCO Membership SCO Status Country Member States China, India, Kazakhstan, Kyrgyzstan, Pakistan, Russia, Tajikistan, Uzbekistan Observer States Afghanistan, Belarus, Iran, Mongolia Dialogue Partners Armenia, Azerbaijan, Cambodia, Nepal, Sri Lanka, Turkey Guest Turkmenistan SCO Status Country Member States China, India, Kazakhstan, Kyrgyzstan, Pakistan, Russia, Tajikistan, Uzbekistan Observer States Afghanistan, Belarus, Iran, Mongolia Dialogue Partners Armenia, Azerbaijan, Cambodia, Nepal, Sri Lanka, Turkey Guest Turkmenistan Note: India and Pakistan became full members of the SCO in 2017, and Belarus became an observer member (up from dialogue partner) in 2015. Table 1 SCO Membership SCO Status Country Member States China, India, Kazakhstan, Kyrgyzstan, Pakistan, Russia, Tajikistan, Uzbekistan Observer States Afghanistan, Belarus, Iran, Mongolia Dialogue Partners Armenia, Azerbaijan, Cambodia, Nepal, Sri Lanka, Turkey Guest Turkmenistan SCO Status Country Member States China, India, Kazakhstan, Kyrgyzstan, Pakistan, Russia, Tajikistan, Uzbekistan Observer States Afghanistan, Belarus, Iran, Mongolia Dialogue Partners Armenia, Azerbaijan, Cambodia, Nepal, Sri Lanka, Turkey Guest Turkmenistan Note: India and Pakistan became full members of the SCO in 2017, and Belarus became an observer member (up from dialogue partner) in 2015. To test the contested multilateralism framework, I draw data from the ADB. Founded on December 19, 1966 the ADB is led by Japan and the United States. Information on ADB membership, including how it overlaps with and differs from AIIB membership, is presented in ‘Contested multilateralism: ADB members’ section. The variable representation is calculated, using ADB data, according to the country’s voting share to GDP ratio. A country is well represented in the ADB if it has a high share-to-GDP ratio, and will have a higher representation value. Raw data on countries’ physical neighbors have been obtained from the CEPII.39 To incorporate dynamics into the model, I update the Neighborextensive and Neighborintensive for each country according to the updated membership, which I base on Chinese Ministry of Finance (MOF) announcements.40 As will be elaborated on later, not all applications are public, but the Chinese MOF publicly announces all admissions. I assume that only publicly available information will enter into states’ calculation. Under this assumption, Iran’s application (dated March 30, 2015) will not affect Azerbaijan’s decision to apply on March 31, 2015. Similarly, Spain’s application on March 27, 2015 will not affect Portugal’s decision. Empirical Analysis In this section, I estimate the model using the AIIB dataset, and test the hypotheses, one-by-one, incrementally. I first estimate the model using the standard probit, and examine the empirical results on democracy. Second, I analyze the effects of learning on AIIB membership. Third, using the ADB as an example, I test the contested multilateralism framework. Fourth, to re-formulate the dataset into panel data, I estimate the same regressors using a Cox Partial Likelihood model as a robustness check. I will discuss the results in the same order. Democracy and AIIB Prospective Founding Membership The first column in Table 2 displays estimate from a static model, and only considers economic factors. This is similar to the standard gravity model in international trade, and aims at capturing static economic factors. The result shows that large countries and countries close to China are more likely to join. It is worth noting that, while the coefficients on GDP per capita point to the expected directions, they are not statistically significant. Table 2. The Expansion of AIIB Membership Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.0856*** 0.157*** 0.0918*** 0.111*** 0.111*** (0.006) (0.000) (0.005) (0.003) (0.003) Ln GDP per capita −0.193 −0.383 −0.273 −0.179 −0.153 (0.612) (0.387) (0.480) (0.677) (0.721) Ln GDP per capita2 0.0117 0.0211 0.0174 0.0143 0.0123 (0.603) (0.419) (0.447) (0.572) (0.628) Ln Distance −0.522*** −0.767*** −0.499*** −0.753*** −0.763*** (0.000) (0.000) (0.000) (0.000) (0.000) Polity −0.0274** (0.012) Political Constraints −0.497* (0.089) Civil Liberties 0.0826** (0.042) Political Rights 0.0692* (0.061) Δ −2.828*** −3.900*** −2.862*** −4.096*** −4.064*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 1.246*** 2.232*** 1.266*** 2.493*** 2.465*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.153*** −0.367*** −0.155*** −0.432*** −0.425*** (0.000) (0.003) (0.000) (0.000) (0.001) Constant 1.427 3.028 1.497 2.431 2.517 (0.461) (0.195) (0.443) (0.273) (0.258) Observations 21876 16618 21240 18899 18899 Pseudo R2 0.243 0.297 0.245 0.283 0.282 Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.0856*** 0.157*** 0.0918*** 0.111*** 0.111*** (0.006) (0.000) (0.005) (0.003) (0.003) Ln GDP per capita −0.193 −0.383 −0.273 −0.179 −0.153 (0.612) (0.387) (0.480) (0.677) (0.721) Ln GDP per capita2 0.0117 0.0211 0.0174 0.0143 0.0123 (0.603) (0.419) (0.447) (0.572) (0.628) Ln Distance −0.522*** −0.767*** −0.499*** −0.753*** −0.763*** (0.000) (0.000) (0.000) (0.000) (0.000) Polity −0.0274** (0.012) Political Constraints −0.497* (0.089) Civil Liberties 0.0826** (0.042) Political Rights 0.0692* (0.061) Δ −2.828*** −3.900*** −2.862*** −4.096*** −4.064*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 1.246*** 2.232*** 1.266*** 2.493*** 2.465*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.153*** −0.367*** −0.155*** −0.432*** −0.425*** (0.000) (0.003) (0.000) (0.000) (0.001) Constant 1.427 3.028 1.497 2.431 2.517 (0.461) (0.195) (0.443) (0.273) (0.258) Observations 21876 16618 21240 18899 18899 Pseudo R2 0.243 0.297 0.245 0.283 0.282 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Table 2. The Expansion of AIIB Membership Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.0856*** 0.157*** 0.0918*** 0.111*** 0.111*** (0.006) (0.000) (0.005) (0.003) (0.003) Ln GDP per capita −0.193 −0.383 −0.273 −0.179 −0.153 (0.612) (0.387) (0.480) (0.677) (0.721) Ln GDP per capita2 0.0117 0.0211 0.0174 0.0143 0.0123 (0.603) (0.419) (0.447) (0.572) (0.628) Ln Distance −0.522*** −0.767*** −0.499*** −0.753*** −0.763*** (0.000) (0.000) (0.000) (0.000) (0.000) Polity −0.0274** (0.012) Political Constraints −0.497* (0.089) Civil Liberties 0.0826** (0.042) Political Rights 0.0692* (0.061) Δ −2.828*** −3.900*** −2.862*** −4.096*** −4.064*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 1.246*** 2.232*** 1.266*** 2.493*** 2.465*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.153*** −0.367*** −0.155*** −0.432*** −0.425*** (0.000) (0.003) (0.000) (0.000) (0.001) Constant 1.427 3.028 1.497 2.431 2.517 (0.461) (0.195) (0.443) (0.273) (0.258) Observations 21876 16618 21240 18899 18899 Pseudo R2 0.243 0.297 0.245 0.283 0.282 Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.0856*** 0.157*** 0.0918*** 0.111*** 0.111*** (0.006) (0.000) (0.005) (0.003) (0.003) Ln GDP per capita −0.193 −0.383 −0.273 −0.179 −0.153 (0.612) (0.387) (0.480) (0.677) (0.721) Ln GDP per capita2 0.0117 0.0211 0.0174 0.0143 0.0123 (0.603) (0.419) (0.447) (0.572) (0.628) Ln Distance −0.522*** −0.767*** −0.499*** −0.753*** −0.763*** (0.000) (0.000) (0.000) (0.000) (0.000) Polity −0.0274** (0.012) Political Constraints −0.497* (0.089) Civil Liberties 0.0826** (0.042) Political Rights 0.0692* (0.061) Δ −2.828*** −3.900*** −2.862*** −4.096*** −4.064*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 1.246*** 2.232*** 1.266*** 2.493*** 2.465*** (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.153*** −0.367*** −0.155*** −0.432*** −0.425*** (0.000) (0.003) (0.000) (0.000) (0.001) Constant 1.427 3.028 1.497 2.431 2.517 (0.461) (0.195) (0.443) (0.273) (0.258) Observations 21876 16618 21240 18899 18899 Pseudo R2 0.243 0.297 0.245 0.283 0.282 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. In columns 2, 3, 4, and 5, I sequentially include four alternative democracy scores: polity, political constraints, political rights, and civil liberties. I find that countries with lower polity scores, lower political constraints, higher political rights scores, and higher civil liberties scores tend to join the AIIB.41 The coefficients are all significant at the 0.1 significance level. They are all consistent with my first hypothesis: less democratic countries are more likely to join the institution than democratic countries. Lastly, I examine time dependency. It will be interesting to know whether countries are more likely to follow other countries’ example by joining, or to wait. I answer this question through analyzing the shape of the polynomial function of time. In Figure 3, I calculate the expected probability of a country joining the AIIB as a function of its waiting time. The figure shows that the probability of joining decreases sharply as the waiting time grows. The function is not monotonic, though. There is a slight increase around Day 28. But overall, the result suggests that timing is very important, and that a country is most likely to join immediately upon following another country’s lead. Fig. 3 View largeDownload slide Joining Probability as a Function of Time (with a 95% Confidence Interval). Note: The 95% confidence interval is calculated using the method. To preserve its symmetric structure, I do not cutoff the regions below zero, but it should be understood that the probability of joining cannot be negative. Fig. 3 View largeDownload slide Joining Probability as a Function of Time (with a 95% Confidence Interval). Note: The 95% confidence interval is calculated using the method. To preserve its symmetric structure, I do not cutoff the regions below zero, but it should be understood that the probability of joining cannot be negative. Uncertainty, Learning, and the AIIB Founding Membership Next, I focus on the two groups of learning variables in the model: Neighborextensive and Neighborintensive, SCO, and BRICS (Table 3). As regards neighbors, I find that countries whose neighbor states have joined the AIIB are more likely to join, as are countries whose neighbors are important countries (i.e. Neighborintensive). This result accords with my argument that as information flows in from neighboring and nearby states, the risk of joining as perceived by countries decreases, and the likelihood that they will join increases. This also provides insight into why the UK’s joining the AIIB should have such impact: it not only affects neighboring countries, such as Ireland, but also reassures countries nearby, such as France, Germany, and to some extent Israel. Table 3 Dynamic Model of AIIB Membership Pr(Join = 1) 1 2 3 4 5 6 Ln GDP 0.157*** 0.192*** 0.213*** 0.235*** 0.196*** 0.197*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.001) Ln GDP per capita −0.383 −0.452 −0.352 −0.587 −0.790 −1.375** (0.387) (0.352) (0.466) (0.254) (0.128) (0.017) Ln GDP per capita2 0.0211 0.0243 0.0134 0.0277 0.0417 0.0740** (0.419) (0.391) (0.638) (0.359) (0.175) (0.029) Ln Distance −0.767*** −0.648*** −0.400*** −0.275* −0.374** 0.396 (0.000) (0.000) (0.007) (0.091) (0.024) (0.127) Polity −0.0274** −0.0230** −0.0331*** −0.0292** −0.0356*** −0.0263** (0.012) (0.034) (0.004) (0.013) (0.004) (0.037) Neighborextensive 0.231*** 0.162** 0.137** 0.122* 0.0256 (0.000) (0.010) (0.035) (0.065) (0.715) Neighborintensive 0.413*** 0.418*** 0.472*** 0.709*** (0.000) (0.000) (0.000) (0.000) SCO 0.442** 0.403* 0.307 (0.026) (0.050) (0.125) BRICS 1.097*** 1.200*** (0.002) (0.002) Asia 1.017*** (0.000) Δ −3.900*** −3.941*** −3.721*** −3.823*** −3.745*** −3.872*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 2.232*** 2.250*** 2.169*** 2.245*** 2.186*** 2.268*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.367*** −0.369*** −0.360*** −0.374*** −0.362*** −0.377*** (0.003) (0.003) (0.005) (0.004) (0.004) (0.005) Constant 3.028 1.298 −9.968*** −10.88*** −9.515** −19.07*** (0.195) (0.592) (0.009) (0.005) (0.015) (0.000) Observations 16618 16618 16618 16618 16618 16618 Pseudo R2 0.297 0.318 0.342 0.350 0.362 0.386 Pr(Join = 1) 1 2 3 4 5 6 Ln GDP 0.157*** 0.192*** 0.213*** 0.235*** 0.196*** 0.197*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.001) Ln GDP per capita −0.383 −0.452 −0.352 −0.587 −0.790 −1.375** (0.387) (0.352) (0.466) (0.254) (0.128) (0.017) Ln GDP per capita2 0.0211 0.0243 0.0134 0.0277 0.0417 0.0740** (0.419) (0.391) (0.638) (0.359) (0.175) (0.029) Ln Distance −0.767*** −0.648*** −0.400*** −0.275* −0.374** 0.396 (0.000) (0.000) (0.007) (0.091) (0.024) (0.127) Polity −0.0274** −0.0230** −0.0331*** −0.0292** −0.0356*** −0.0263** (0.012) (0.034) (0.004) (0.013) (0.004) (0.037) Neighborextensive 0.231*** 0.162** 0.137** 0.122* 0.0256 (0.000) (0.010) (0.035) (0.065) (0.715) Neighborintensive 0.413*** 0.418*** 0.472*** 0.709*** (0.000) (0.000) (0.000) (0.000) SCO 0.442** 0.403* 0.307 (0.026) (0.050) (0.125) BRICS 1.097*** 1.200*** (0.002) (0.002) Asia 1.017*** (0.000) Δ −3.900*** −3.941*** −3.721*** −3.823*** −3.745*** −3.872*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 2.232*** 2.250*** 2.169*** 2.245*** 2.186*** 2.268*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.367*** −0.369*** −0.360*** −0.374*** −0.362*** −0.377*** (0.003) (0.003) (0.005) (0.004) (0.004) (0.005) Constant 3.028 1.298 −9.968*** −10.88*** −9.515** −19.07*** (0.195) (0.592) (0.009) (0.005) (0.015) (0.000) Observations 16618 16618 16618 16618 16618 16618 Pseudo R2 0.297 0.318 0.342 0.350 0.362 0.386 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Table 3 Dynamic Model of AIIB Membership Pr(Join = 1) 1 2 3 4 5 6 Ln GDP 0.157*** 0.192*** 0.213*** 0.235*** 0.196*** 0.197*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.001) Ln GDP per capita −0.383 −0.452 −0.352 −0.587 −0.790 −1.375** (0.387) (0.352) (0.466) (0.254) (0.128) (0.017) Ln GDP per capita2 0.0211 0.0243 0.0134 0.0277 0.0417 0.0740** (0.419) (0.391) (0.638) (0.359) (0.175) (0.029) Ln Distance −0.767*** −0.648*** −0.400*** −0.275* −0.374** 0.396 (0.000) (0.000) (0.007) (0.091) (0.024) (0.127) Polity −0.0274** −0.0230** −0.0331*** −0.0292** −0.0356*** −0.0263** (0.012) (0.034) (0.004) (0.013) (0.004) (0.037) Neighborextensive 0.231*** 0.162** 0.137** 0.122* 0.0256 (0.000) (0.010) (0.035) (0.065) (0.715) Neighborintensive 0.413*** 0.418*** 0.472*** 0.709*** (0.000) (0.000) (0.000) (0.000) SCO 0.442** 0.403* 0.307 (0.026) (0.050) (0.125) BRICS 1.097*** 1.200*** (0.002) (0.002) Asia 1.017*** (0.000) Δ −3.900*** −3.941*** −3.721*** −3.823*** −3.745*** −3.872*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 2.232*** 2.250*** 2.169*** 2.245*** 2.186*** 2.268*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.367*** −0.369*** −0.360*** −0.374*** −0.362*** −0.377*** (0.003) (0.003) (0.005) (0.004) (0.004) (0.005) Constant 3.028 1.298 −9.968*** −10.88*** −9.515** −19.07*** (0.195) (0.592) (0.009) (0.005) (0.015) (0.000) Observations 16618 16618 16618 16618 16618 16618 Pseudo R2 0.297 0.318 0.342 0.350 0.362 0.386 Pr(Join = 1) 1 2 3 4 5 6 Ln GDP 0.157*** 0.192*** 0.213*** 0.235*** 0.196*** 0.197*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.001) Ln GDP per capita −0.383 −0.452 −0.352 −0.587 −0.790 −1.375** (0.387) (0.352) (0.466) (0.254) (0.128) (0.017) Ln GDP per capita2 0.0211 0.0243 0.0134 0.0277 0.0417 0.0740** (0.419) (0.391) (0.638) (0.359) (0.175) (0.029) Ln Distance −0.767*** −0.648*** −0.400*** −0.275* −0.374** 0.396 (0.000) (0.000) (0.007) (0.091) (0.024) (0.127) Polity −0.0274** −0.0230** −0.0331*** −0.0292** −0.0356*** −0.0263** (0.012) (0.034) (0.004) (0.013) (0.004) (0.037) Neighborextensive 0.231*** 0.162** 0.137** 0.122* 0.0256 (0.000) (0.010) (0.035) (0.065) (0.715) Neighborintensive 0.413*** 0.418*** 0.472*** 0.709*** (0.000) (0.000) (0.000) (0.000) SCO 0.442** 0.403* 0.307 (0.026) (0.050) (0.125) BRICS 1.097*** 1.200*** (0.002) (0.002) Asia 1.017*** (0.000) Δ −3.900*** −3.941*** −3.721*** −3.823*** −3.745*** −3.872*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ2 2.232*** 2.250*** 2.169*** 2.245*** 2.186*** 2.268*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Δ3 −0.367*** −0.369*** −0.360*** −0.374*** −0.362*** −0.377*** (0.003) (0.003) (0.005) (0.004) (0.004) (0.005) Constant 3.028 1.298 −9.968*** −10.88*** −9.515** −19.07*** (0.195) (0.592) (0.009) (0.005) (0.015) (0.000) Observations 16618 16618 16618 16618 16618 16618 Pseudo R2 0.297 0.318 0.342 0.350 0.362 0.386 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. This has two implications. First, the effect of a regional power joining the AIIB is greater than that of an average country. Second, the Neighborintensive variable is also able to explain the observation that it is usually large countries (generally hegemons) that establish IOs. The most popular supply-side argument is that hegemons alone can reap enough benefits from providing public goods.42 Here Neighborintensive raises the demand-side argument: only large economies can garner enough initial momentum from neighboring countries and attract countries farther away through network effects (Figure 4). Fig. 4. View largeDownload slide The Estimated Probability of Joining as a Function of the Neighborintensive Variable. Fig. 4. View largeDownload slide The Estimated Probability of Joining as a Function of the Neighborintensive Variable. Looking now at the second group of learning variables, SCO and BRICS, I find that countries’ membership in these two organizations also increases the likelihood of their joining the AIIB. This suggests that information also flows from the past: countries that have had more interactions with China have higher confidence in the success of the new institution. From a policy point of view, this implies that as China now plays a more active role on the global stage, and gradually consolidates its credentials, countries will respond more positively to China-initiated IOs. Contested Multilateralism: ADB Members The AIIB almost directly competes with the ADB, and the memberships largely overlap. Figure 5 (upper) shows the membership of the two institutions. The AIIB has 57 members,43 the ADB has 67, and they have 42 members in common. Prominent countries that are members of the ADB but not of the AIIB include the United States, Japan, and Canada; the three large emerging economies of Russia, Brazil, and South Africa are in the AIIB but not the ADB. Since both institutions are development banks whose focus is on Asia, this presents an ideal case for testing the framework of contested multilateralism.44 Specifically, I test the hypothesis that dissatisfied members of the ADB are more likely to join the AIIB. To operationalize the idea, I test whether ADB members that have a low representation, calculated according to their voting share to GDP ratio, are more likely to join the institution. Figure 5 (lower) shows the distribution of the representation variable. Not surprisingly, China has the lowest representation; the representation for certain early AIIB joiners like the UK and Spain is also low. However, among those countries with a higher representation ratio, Ireland and Belgium applied relatively late, while Armenia and Turkmenistan have yet to make their application. I present the formal regression results in Table 4. Table 4. Contested Multilateralism: The Case of AIIB and ADB Pr(Join = 1) 1 2 3 4 Ln GDP 0.0735 0.0728 0.124 0.0415 (0.371) (0.401) (0.187) (0.689) Ln GDP per capita −3.997*** −3.928*** −3.097** −4.452*** (0.001) (0.002) (0.016) (0.003) Ln GDP per capita2 0.197*** 0.193*** 0.139** 0.219*** (0.003) (0.006) (0.049) (0.008) Ln Distance −0.379* −0.410* −0.0242 0.326 (0.099) (0.084) (0.926) (0.368) Representation −0.181*** −0.228*** −0.157*** −0.206*** (0.000) (0.000) (0.003) (0.001) Polity 0.00275 0.0116 −0.00416 0.0195 (0.894) (0.592) (0.846) (0.429) Neighborextensive 0.267*** −0.00279 −0.00456 (0.006) (0.980) (0.968) Neighborintensive 0.612*** 0.611*** (0.000) (0.000) SCO 0.194 (0.554) Asia 0.621 (0.149) Δ −3.548*** −3.340*** −3.344*** −3.270*** (0.000) (0.000) (0.000) (0.000) Δ2 1.729*** 1.598*** 1.645** 1.587** (0.005) (0.009) (0.012) (0.015) Δ3 −0.247** −0.226* −0.235* −0.224* (0.048) (0.070) (0.077) (0.090) Constant 20.17*** 20.08*** −0.365 3.861 (0.000) (0.001) (0.965) (0.656) Observations 4373 4373 4373 4373 Pseudo R2 0.346 0.363 0.405 0.415 Pr(Join = 1) 1 2 3 4 Ln GDP 0.0735 0.0728 0.124 0.0415 (0.371) (0.401) (0.187) (0.689) Ln GDP per capita −3.997*** −3.928*** −3.097** −4.452*** (0.001) (0.002) (0.016) (0.003) Ln GDP per capita2 0.197*** 0.193*** 0.139** 0.219*** (0.003) (0.006) (0.049) (0.008) Ln Distance −0.379* −0.410* −0.0242 0.326 (0.099) (0.084) (0.926) (0.368) Representation −0.181*** −0.228*** −0.157*** −0.206*** (0.000) (0.000) (0.003) (0.001) Polity 0.00275 0.0116 −0.00416 0.0195 (0.894) (0.592) (0.846) (0.429) Neighborextensive 0.267*** −0.00279 −0.00456 (0.006) (0.980) (0.968) Neighborintensive 0.612*** 0.611*** (0.000) (0.000) SCO 0.194 (0.554) Asia 0.621 (0.149) Δ −3.548*** −3.340*** −3.344*** −3.270*** (0.000) (0.000) (0.000) (0.000) Δ2 1.729*** 1.598*** 1.645** 1.587** (0.005) (0.009) (0.012) (0.015) Δ3 −0.247** −0.226* −0.235* −0.224* (0.048) (0.070) (0.077) (0.090) Constant 20.17*** 20.08*** −0.365 3.861 (0.000) (0.001) (0.965) (0.656) Observations 4373 4373 4373 4373 Pseudo R2 0.346 0.363 0.405 0.415 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Table 4. Contested Multilateralism: The Case of AIIB and ADB Pr(Join = 1) 1 2 3 4 Ln GDP 0.0735 0.0728 0.124 0.0415 (0.371) (0.401) (0.187) (0.689) Ln GDP per capita −3.997*** −3.928*** −3.097** −4.452*** (0.001) (0.002) (0.016) (0.003) Ln GDP per capita2 0.197*** 0.193*** 0.139** 0.219*** (0.003) (0.006) (0.049) (0.008) Ln Distance −0.379* −0.410* −0.0242 0.326 (0.099) (0.084) (0.926) (0.368) Representation −0.181*** −0.228*** −0.157*** −0.206*** (0.000) (0.000) (0.003) (0.001) Polity 0.00275 0.0116 −0.00416 0.0195 (0.894) (0.592) (0.846) (0.429) Neighborextensive 0.267*** −0.00279 −0.00456 (0.006) (0.980) (0.968) Neighborintensive 0.612*** 0.611*** (0.000) (0.000) SCO 0.194 (0.554) Asia 0.621 (0.149) Δ −3.548*** −3.340*** −3.344*** −3.270*** (0.000) (0.000) (0.000) (0.000) Δ2 1.729*** 1.598*** 1.645** 1.587** (0.005) (0.009) (0.012) (0.015) Δ3 −0.247** −0.226* −0.235* −0.224* (0.048) (0.070) (0.077) (0.090) Constant 20.17*** 20.08*** −0.365 3.861 (0.000) (0.001) (0.965) (0.656) Observations 4373 4373 4373 4373 Pseudo R2 0.346 0.363 0.405 0.415 Pr(Join = 1) 1 2 3 4 Ln GDP 0.0735 0.0728 0.124 0.0415 (0.371) (0.401) (0.187) (0.689) Ln GDP per capita −3.997*** −3.928*** −3.097** −4.452*** (0.001) (0.002) (0.016) (0.003) Ln GDP per capita2 0.197*** 0.193*** 0.139** 0.219*** (0.003) (0.006) (0.049) (0.008) Ln Distance −0.379* −0.410* −0.0242 0.326 (0.099) (0.084) (0.926) (0.368) Representation −0.181*** −0.228*** −0.157*** −0.206*** (0.000) (0.000) (0.003) (0.001) Polity 0.00275 0.0116 −0.00416 0.0195 (0.894) (0.592) (0.846) (0.429) Neighborextensive 0.267*** −0.00279 −0.00456 (0.006) (0.980) (0.968) Neighborintensive 0.612*** 0.611*** (0.000) (0.000) SCO 0.194 (0.554) Asia 0.621 (0.149) Δ −3.548*** −3.340*** −3.344*** −3.270*** (0.000) (0.000) (0.000) (0.000) Δ2 1.729*** 1.598*** 1.645** 1.587** (0.005) (0.009) (0.012) (0.015) Δ3 −0.247** −0.226* −0.235* −0.224* (0.048) (0.070) (0.077) (0.090) Constant 20.17*** 20.08*** −0.365 3.861 (0.000) (0.001) (0.965) (0.656) Observations 4373 4373 4373 4373 Pseudo R2 0.346 0.363 0.405 0.415 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Fig. 5 View largeDownload slide AIIB and ADB: Contested Multilateralism. Note: The ratio for Myanmar is not shown, as the country’s GDP figures are not available. Fig. 5 View largeDownload slide AIIB and ADB: Contested Multilateralism. Note: The ratio for Myanmar is not shown, as the country’s GDP figures are not available. Across all specifications, the coefficient on representation remains negative, and is statistically significant at a 95% confidence level. This means that countries with low representation in the ADB are more likely to join the AIIB, thus confirming the argument that contested multilateralism results from the dissatisfaction of countries that have outside options. The AIIB is often described as ‘China’s World Bank’, which underlines the fact that to some extent it competes with the World Bank. For a robustness check, I also analyze how representation in the World Bank’s two main institutions—the International Bank for Reconstruction and Development (IBRD) and the International Development Association (IDA)— affects countries’ decision to join. Since there is a difference in distribution of voting power between the IBRD and the IDA, I calculate two representation metrics, one for each branch, and refer to them as IBRD Representation and IDA Representation, respectively. I find that member countries in these institutions that are under-represented are also more likely to join the AIIB, so lending further support to the contested multilateralism framework. I report the details of the results in Table 5.45 Table 5. Contested Multilateralism: Robustness Check with IBRD and IDA Pr(Join = 1) 1 2 3 Ln GDP 0.0326 0.136* 0.0393 (0.753) (0.053) (0.617) Ln GDP per capita −4.271*** −1.770*** −1.493** (0.004) (0.007) (0.031) Ln GDP per capita2 0.210** 0.0964** 0.0827** (0.011) (0.012) (0.041) Ln Distance 0.335 0.277 −0.0913 (0.353) (0.314) (0.764) Neighborextensive −0.00868 0.0524 0.123 (0.939) (0.481) (0.145) Neighborintensive 0.613*** 0.675*** 0.608*** (0.000) (0.000) (0.000) Polity 0.0176 −0.0213 −0.0274* (0.476) (0.111) (0.080) SCO 0.147 0.320 0.563** (0.658) (0.123) (0.018) BRICS 0 1.295*** 1.430*** (−) (0.001) (0.000) Asia 0.642 0.965*** 0.680** (0.134) (0.000) (0.017) (0.090) (0.006) (0.010) ADB Representation −0.198*** (0.002) IBRD Representation −0.0290 (0.163) IDA Representation −0.0242** (0.031) Δ −3.236*** −3.751*** −3.546*** (0.000) (0.000) (0.000) Δ2 1.573** 2.193*** 2.036*** (0.015) (0.000) (0.001) Δ3 −0.222* −0.365*** −0.333** (0.090) (0.006) (0.010) Constant 3.059 −13.97** −8.034 (0.723) (0.018) (0.229) Observations 4372 16618 15186 Pseudo R2 0.403 0.391 0.404 Pr(Join = 1) 1 2 3 Ln GDP 0.0326 0.136* 0.0393 (0.753) (0.053) (0.617) Ln GDP per capita −4.271*** −1.770*** −1.493** (0.004) (0.007) (0.031) Ln GDP per capita2 0.210** 0.0964** 0.0827** (0.011) (0.012) (0.041) Ln Distance 0.335 0.277 −0.0913 (0.353) (0.314) (0.764) Neighborextensive −0.00868 0.0524 0.123 (0.939) (0.481) (0.145) Neighborintensive 0.613*** 0.675*** 0.608*** (0.000) (0.000) (0.000) Polity 0.0176 −0.0213 −0.0274* (0.476) (0.111) (0.080) SCO 0.147 0.320 0.563** (0.658) (0.123) (0.018) BRICS 0 1.295*** 1.430*** (−) (0.001) (0.000) Asia 0.642 0.965*** 0.680** (0.134) (0.000) (0.017) (0.090) (0.006) (0.010) ADB Representation −0.198*** (0.002) IBRD Representation −0.0290 (0.163) IDA Representation −0.0242** (0.031) Δ −3.236*** −3.751*** −3.546*** (0.000) (0.000) (0.000) Δ2 1.573** 2.193*** 2.036*** (0.015) (0.000) (0.001) Δ3 −0.222* −0.365*** −0.333** (0.090) (0.006) (0.010) Constant 3.059 −13.97** −8.034 (0.723) (0.018) (0.229) Observations 4372 16618 15186 Pseudo R2 0.403 0.391 0.404 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Table 5. Contested Multilateralism: Robustness Check with IBRD and IDA Pr(Join = 1) 1 2 3 Ln GDP 0.0326 0.136* 0.0393 (0.753) (0.053) (0.617) Ln GDP per capita −4.271*** −1.770*** −1.493** (0.004) (0.007) (0.031) Ln GDP per capita2 0.210** 0.0964** 0.0827** (0.011) (0.012) (0.041) Ln Distance 0.335 0.277 −0.0913 (0.353) (0.314) (0.764) Neighborextensive −0.00868 0.0524 0.123 (0.939) (0.481) (0.145) Neighborintensive 0.613*** 0.675*** 0.608*** (0.000) (0.000) (0.000) Polity 0.0176 −0.0213 −0.0274* (0.476) (0.111) (0.080) SCO 0.147 0.320 0.563** (0.658) (0.123) (0.018) BRICS 0 1.295*** 1.430*** (−) (0.001) (0.000) Asia 0.642 0.965*** 0.680** (0.134) (0.000) (0.017) (0.090) (0.006) (0.010) ADB Representation −0.198*** (0.002) IBRD Representation −0.0290 (0.163) IDA Representation −0.0242** (0.031) Δ −3.236*** −3.751*** −3.546*** (0.000) (0.000) (0.000) Δ2 1.573** 2.193*** 2.036*** (0.015) (0.000) (0.001) Δ3 −0.222* −0.365*** −0.333** (0.090) (0.006) (0.010) Constant 3.059 −13.97** −8.034 (0.723) (0.018) (0.229) Observations 4372 16618 15186 Pseudo R2 0.403 0.391 0.404 Pr(Join = 1) 1 2 3 Ln GDP 0.0326 0.136* 0.0393 (0.753) (0.053) (0.617) Ln GDP per capita −4.271*** −1.770*** −1.493** (0.004) (0.007) (0.031) Ln GDP per capita2 0.210** 0.0964** 0.0827** (0.011) (0.012) (0.041) Ln Distance 0.335 0.277 −0.0913 (0.353) (0.314) (0.764) Neighborextensive −0.00868 0.0524 0.123 (0.939) (0.481) (0.145) Neighborintensive 0.613*** 0.675*** 0.608*** (0.000) (0.000) (0.000) Polity 0.0176 −0.0213 −0.0274* (0.476) (0.111) (0.080) SCO 0.147 0.320 0.563** (0.658) (0.123) (0.018) BRICS 0 1.295*** 1.430*** (−) (0.001) (0.000) Asia 0.642 0.965*** 0.680** (0.134) (0.000) (0.017) (0.090) (0.006) (0.010) ADB Representation −0.198*** (0.002) IBRD Representation −0.0290 (0.163) IDA Representation −0.0242** (0.031) Δ −3.236*** −3.751*** −3.546*** (0.000) (0.000) (0.000) Δ2 1.573** 2.193*** 2.036*** (0.015) (0.000) (0.001) Δ3 −0.222* −0.365*** −0.333** (0.090) (0.006) (0.010) Constant 3.059 −13.97** −8.034 (0.723) (0.018) (0.229) Observations 4372 16618 15186 Pseudo R2 0.403 0.391 0.404 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Cox Proportional Hazards Model So far, the major mechanisms underlying AIIB membership have been tested in a probit model. Here I test the robustness of the dynamic variables by placing them in a duration analysis framework, i.e. the wait duration before joining the AIIB. There is a natural link between binary data and duration data.46 Intuitively, countries that are more likely to join (with a higher xβ, and thus ϕ(xβ)) will have a shorter expected wait duration. This intuition can be captured by the Cox proportional hazard model: λit=λ0(t)exiβ where λt= p(t)P(T > t). If xiβ>xjβ, i.e. countryi is more likely to join than countryj, in the duration analysis framework, PTi<Tj=exiβexiβ+exjβ>0.5 with Ti and Tj denoting the waiting time for countryi and countryj, respectively. I provide details of the proof in Appendix A. With this natural transition, I can use the Cox Proportional Hazard model to test the robustness of the main results. The data are right censored. Once a country joins the AIIB, the remaining countries will enter a new period of observation with updated Neighborextensive and Neighborintensive. In the full specification (Column 4), the number of observations is 1615. The regression results are reported in Tables 6 and 7. I use the same set of variables and the same specifications as in Table 4, except that in this framework I do not include the time variables, as time dependency can be reflected in the hazard ratio.47 The number of observations is also much smaller than in the probit models, where data are constructed on a daily basis. Another difference here is that in the Cox Partial Likelihood framework, only ranking matters; the waiting time is not relevant. As several countries applied for prospective founding membership on the same day—Denmark and the Netherlands, for example, both applied on March 28, 2015—there are ties observed in the data. For robustness, I use both Efron’s method and Breslow’s method to break such ties. Table 6. Robustness Check (Efron): Wait Duration Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.302*** 0.397*** 0.428*** 0.0236 (0.011) (0.007) (0.001) (0.001) (0.911) Ln GDP per capita −0.655 −0.907 −1.018 −2.586* −2.457 (0.524) (0.420) (0.402) (0.063) (0.272) Ln GDP per capita2 0.0416 0.0548 0.0504 0.142* 0.107 (0.495) (0.412) (0.482) (0.080) (0.407) Ln Distance −1.172*** −1.444*** −0.727** 1.217** 0.731 (0.000) (0.000) (0.038) (0.045) (0.342) Polity −0.0604** −0.0731*** −0.0404 0.0303 (0.020) (0.007) (0.161) (0.519) Neighborextensive 0.391*** 0.134 −0.0213 (0.009) (0.406) (0.924) Neighborintensive 0.931*** 1.441*** 1.090*** (0.002) (0.000) (0.005) SCO 0.851* 0.115 (0.056) (0.834) Asia 2.217*** 1.342 (0.000) (0.153) ADB Representation −19.11* (0.050) Observations 2122 1615 1615 1615 391 Pseudo R2 0.117 0.142 0.203 0.248 0.172 Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.302*** 0.397*** 0.428*** 0.0236 (0.011) (0.007) (0.001) (0.001) (0.911) Ln GDP per capita −0.655 −0.907 −1.018 −2.586* −2.457 (0.524) (0.420) (0.402) (0.063) (0.272) Ln GDP per capita2 0.0416 0.0548 0.0504 0.142* 0.107 (0.495) (0.412) (0.482) (0.080) (0.407) Ln Distance −1.172*** −1.444*** −0.727** 1.217** 0.731 (0.000) (0.000) (0.038) (0.045) (0.342) Polity −0.0604** −0.0731*** −0.0404 0.0303 (0.020) (0.007) (0.161) (0.519) Neighborextensive 0.391*** 0.134 −0.0213 (0.009) (0.406) (0.924) Neighborintensive 0.931*** 1.441*** 1.090*** (0.002) (0.000) (0.005) SCO 0.851* 0.115 (0.056) (0.834) Asia 2.217*** 1.342 (0.000) (0.153) ADB Representation −19.11* (0.050) Observations 2122 1615 1615 1615 391 Pseudo R2 0.117 0.142 0.203 0.248 0.172 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Table 6. Robustness Check (Efron): Wait Duration Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.302*** 0.397*** 0.428*** 0.0236 (0.011) (0.007) (0.001) (0.001) (0.911) Ln GDP per capita −0.655 −0.907 −1.018 −2.586* −2.457 (0.524) (0.420) (0.402) (0.063) (0.272) Ln GDP per capita2 0.0416 0.0548 0.0504 0.142* 0.107 (0.495) (0.412) (0.482) (0.080) (0.407) Ln Distance −1.172*** −1.444*** −0.727** 1.217** 0.731 (0.000) (0.000) (0.038) (0.045) (0.342) Polity −0.0604** −0.0731*** −0.0404 0.0303 (0.020) (0.007) (0.161) (0.519) Neighborextensive 0.391*** 0.134 −0.0213 (0.009) (0.406) (0.924) Neighborintensive 0.931*** 1.441*** 1.090*** (0.002) (0.000) (0.005) SCO 0.851* 0.115 (0.056) (0.834) Asia 2.217*** 1.342 (0.000) (0.153) ADB Representation −19.11* (0.050) Observations 2122 1615 1615 1615 391 Pseudo R2 0.117 0.142 0.203 0.248 0.172 Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.302*** 0.397*** 0.428*** 0.0236 (0.011) (0.007) (0.001) (0.001) (0.911) Ln GDP per capita −0.655 −0.907 −1.018 −2.586* −2.457 (0.524) (0.420) (0.402) (0.063) (0.272) Ln GDP per capita2 0.0416 0.0548 0.0504 0.142* 0.107 (0.495) (0.412) (0.482) (0.080) (0.407) Ln Distance −1.172*** −1.444*** −0.727** 1.217** 0.731 (0.000) (0.000) (0.038) (0.045) (0.342) Polity −0.0604** −0.0731*** −0.0404 0.0303 (0.020) (0.007) (0.161) (0.519) Neighborextensive 0.391*** 0.134 −0.0213 (0.009) (0.406) (0.924) Neighborintensive 0.931*** 1.441*** 1.090*** (0.002) (0.000) (0.005) SCO 0.851* 0.115 (0.056) (0.834) Asia 2.217*** 1.342 (0.000) (0.153) ADB Representation −19.11* (0.050) Observations 2122 1615 1615 1615 391 Pseudo R2 0.117 0.142 0.203 0.248 0.172 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Table 7. Robustness Check (Breslow): Wait Duration Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.305*** 0.381*** 0.398*** 0.0494 (0.010) (0.006) (0.001) (0.001) (0.811) Ln GDP per capita −0.681 −0.952 −1.076 −2.359* −2.132 (0.511) (0.401) (0.381) (0.086) (0.331) Ln GDP per capita2 0.0442 0.0586 0.0572 0.133* 0.0965 (0.469) (0.379) (0.426) (0.097) (0.447) Ln Distance −1.125*** −1.325*** −0.648* 1.182* 1.019 (0.000) (0.000) (0.069) (0.055) (0.206) Polity −0.0547** −0.0658** −0.0345 0.0248 (0.032) (0.014) (0.230) (0.592) Neighborextensive 0.384** 0.137 −0.0125 (0.011) (0.400) (0.955) Neighborintensive 0.822*** 1.327*** 0.996*** (0.004) (0.000) (0.009) SCO 0.765* 0.000358 (0.087) (0.999) Asia 2.124*** 1.527 (0.001) (0.112) ADB Representation −13.77 (0.150) Observations 2122 1615 1615 1615 391 Pseudo R2 0.110 0.129 0.182 0.221 0.125 Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.305*** 0.381*** 0.398*** 0.0494 (0.010) (0.006) (0.001) (0.001) (0.811) Ln GDP per capita −0.681 −0.952 −1.076 −2.359* −2.132 (0.511) (0.401) (0.381) (0.086) (0.331) Ln GDP per capita2 0.0442 0.0586 0.0572 0.133* 0.0965 (0.469) (0.379) (0.426) (0.097) (0.447) Ln Distance −1.125*** −1.325*** −0.648* 1.182* 1.019 (0.000) (0.000) (0.069) (0.055) (0.206) Polity −0.0547** −0.0658** −0.0345 0.0248 (0.032) (0.014) (0.230) (0.592) Neighborextensive 0.384** 0.137 −0.0125 (0.011) (0.400) (0.955) Neighborintensive 0.822*** 1.327*** 0.996*** (0.004) (0.000) (0.009) SCO 0.765* 0.000358 (0.087) (0.999) Asia 2.124*** 1.527 (0.001) (0.112) ADB Representation −13.77 (0.150) Observations 2122 1615 1615 1615 391 Pseudo R2 0.110 0.129 0.182 0.221 0.125 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Table 7. Robustness Check (Breslow): Wait Duration Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.305*** 0.381*** 0.398*** 0.0494 (0.010) (0.006) (0.001) (0.001) (0.811) Ln GDP per capita −0.681 −0.952 −1.076 −2.359* −2.132 (0.511) (0.401) (0.381) (0.086) (0.331) Ln GDP per capita2 0.0442 0.0586 0.0572 0.133* 0.0965 (0.469) (0.379) (0.426) (0.097) (0.447) Ln Distance −1.125*** −1.325*** −0.648* 1.182* 1.019 (0.000) (0.000) (0.069) (0.055) (0.206) Polity −0.0547** −0.0658** −0.0345 0.0248 (0.032) (0.014) (0.230) (0.592) Neighborextensive 0.384** 0.137 −0.0125 (0.011) (0.400) (0.955) Neighborintensive 0.822*** 1.327*** 0.996*** (0.004) (0.000) (0.009) SCO 0.765* 0.000358 (0.087) (0.999) Asia 2.124*** 1.527 (0.001) (0.112) ADB Representation −13.77 (0.150) Observations 2122 1615 1615 1615 391 Pseudo R2 0.110 0.129 0.182 0.221 0.125 Pr(Join = 1) 1 2 3 4 5 Ln GDP 0.211** 0.305*** 0.381*** 0.398*** 0.0494 (0.010) (0.006) (0.001) (0.001) (0.811) Ln GDP per capita −0.681 −0.952 −1.076 −2.359* −2.132 (0.511) (0.401) (0.381) (0.086) (0.331) Ln GDP per capita2 0.0442 0.0586 0.0572 0.133* 0.0965 (0.469) (0.379) (0.426) (0.097) (0.447) Ln Distance −1.125*** −1.325*** −0.648* 1.182* 1.019 (0.000) (0.000) (0.069) (0.055) (0.206) Polity −0.0547** −0.0658** −0.0345 0.0248 (0.032) (0.014) (0.230) (0.592) Neighborextensive 0.384** 0.137 −0.0125 (0.011) (0.400) (0.955) Neighborintensive 0.822*** 1.327*** 0.996*** (0.004) (0.000) (0.009) SCO 0.765* 0.000358 (0.087) (0.999) Asia 2.124*** 1.527 (0.001) (0.112) ADB Representation −13.77 (0.150) Observations 2122 1615 1615 1615 391 Pseudo R2 0.110 0.129 0.182 0.221 0.125 p-values in parentheses * p < 0.10. ** p < 0.05. *** p < 0.010. Overall, the results are very similar to the ones reported in the main results. I find that all the variables of interest, such as Neighborextensive, SCO, and Polity, retain their significance and point to the same direction as reported in Tables 3 and 4. But the coefficients now have a different interpretation: countries with a larger Neighborextensive in the AIIB tend to wait shorter periods, and so do countries with a large Neighborintensive value. But essentially, they are reading from the same mechanisms: less democratic countries are more likely to join, countries with neighbors in the AIIB are more likely to join, countries that are already members of the SCO are more likely to join, and ADB members who feel under-represented are more likely to join. Conclusion This article has provided a comprehensive modelling of the membership structure of the China-led AIIB. First, I have demonstrated that less democratic countries are more likely to join as founding members. From the perspective of risk, uncertainty, and learning, I have also shown that, when evaluating AIIB founding membership, countries learn from their neighbors and from their previous interactions with China. Third, through the case of the ADB, I have shown that countries under-represented in the ADB are more likely join the AIIB. Lastly, I have employed the Cox proportional hazard model as an alternative to demonstrate the robustness of the membership structure uncovered in this article. This article contributes to the IO literature in two aspects. First, most studies have focused on the decision-making of the leading state(s). The behavior of the participating countries is usually described as utility maximizing, but seldom receives serious attention. The phenomenon whereby China, a developing country, establishes an IO and invites developed countries to join is new. It is hence ideal for modelling the decision-making of participating countries. Second, my study has empirically tested and confirmed the contested multilateralism framework. Less represented countries (thus more dissatisfied) are shown to be more eager to join the new institution. This article also contributes to the literature on autocracy promotion and the debate on democracy and multilateralism. I find that less democratic countries are more eager to join the China-led institution than their more democratic counterparts. This could be explained by the argument (i) less democratic countries face fewer domestic political constraints on joining the China-led institution, and (ii) more democratic countries are more constrained by virtue of opposing pressure from the United States. Appendix A In this appendix, I prove that if countryi is more likely to join the IO (in the binary setting), then countryi is more likely to wait a shorter period of time before joining (in the proportional hazard duration setting). Mathematically, Φx1β>Φx2β ⇒Pr⁡T1<T2>0.5 where T denotes the wait duration. Proof: (i) The hazard function is defined as h(t) =  p(t)P(T≥t). (ii) The survival function is defined as S(t) =  PrT ≥ t= ∫t∞pτdτ, with S't=-pt. From (i) and (ii), I can derive the identify St=e-∫0th(τ)dτ. (iii) Introduce the Cox model as follows: ht=h0(t)exβ where x is the set of regressors, β is the vector of parameters, and h0(t) is the baseline. Then h2(t)h1(t)=ex2βex1β=γ2γ1⇒ S2=S1γ2γ1. Pr⁡(T1<T2)=∫0∞∫0∞p(t1, t2)dt2t1=∫0∞p(t1)∫t1∞p(t2)dt2t1=∫0∞p(t1)S2(t1)dt1=∫0∞p(t1)(S1(t1))γ2γ1dt1=−∫0∞S'(t1)(S1(t1))γ2γ1dt1=−γ1γ1+γ2(S1(t1))γ2γ1|0∞=γ1γ1+γ2=ex1βex1β+ex2β=11 + ex2β−x1β > 0.5 (iv) Appendix B Table B1. Dates of Application and Public Known Country Application Date Publicly Known Date2014 Month Day Year Month Day Year Indonesia 11 25 2014 11 15 2014 Maldives 12 17* 2014 12 31 2014 New Zealand 12 18 2014 1 1 2015 Saudi Arabia 12 31 2014 1 13 2015 Tajikistan 12 31 2014 1 13 2015 Jordan 1 24* 2015 2 7 2015 UK 3 12 2015 3 12 2015 Germany 3 17 2015 3 17 2015 France 3 17 2015 3 17 2015 Italy 3 17 2015 3 17 2015 Luxembourg 3 18 2015 3 18 2015 Switzerland 3 20 2015 3 20 2015 UAE 3 20* 2015 4 3 2015 Iran 3 20* 2015 4 3 2015 Turkey 3 26 2015 3 26 2015 Spain 3 27* 2015 4 11 2015 South Korea 3 27 2015 3 27 2015 Austria 3 27 2015 3 27 2015 Georgia 3 28 2015 3 28 2015 Denmark 3 28 2015 3 28 2015 Netherlands 3 28 2015 3 28 2015 Brazil 3 28 2015 3 28 2015 Australia 3 29 2015 3 29 2015 Finland 3 30 2015 3 30 2015 Russia 3 30 2015 3 30 2015 Norway 3 30* 2015 4 14 2015 Egypt 3 30 2015 3 30 2015 Kyrgyzstan 3 31 2015 3 31 2015 Malta 3 31* 2015 4 9 2015 Sweden 3 31 2015 3 31 2015 Israel 3 31* 2015 4 15 2015 Portugal 3 31 2015 3 31 2015 South Africa 3 31* 2015 4 15 2015 Azerbaijan 3 31* 2015 4 15 2015 Iceland 3 31 2015 3 31 2015 Poland 3 31* 2015 4 15 2015 Country Application Date Publicly Known Date2014 Month Day Year Month Day Year Indonesia 11 25 2014 11 15 2014 Maldives 12 17* 2014 12 31 2014 New Zealand 12 18 2014 1 1 2015 Saudi Arabia 12 31 2014 1 13 2015 Tajikistan 12 31 2014 1 13 2015 Jordan 1 24* 2015 2 7 2015 UK 3 12 2015 3 12 2015 Germany 3 17 2015 3 17 2015 France 3 17 2015 3 17 2015 Italy 3 17 2015 3 17 2015 Luxembourg 3 18 2015 3 18 2015 Switzerland 3 20 2015 3 20 2015 UAE 3 20* 2015 4 3 2015 Iran 3 20* 2015 4 3 2015 Turkey 3 26 2015 3 26 2015 Spain 3 27* 2015 4 11 2015 South Korea 3 27 2015 3 27 2015 Austria 3 27 2015 3 27 2015 Georgia 3 28 2015 3 28 2015 Denmark 3 28 2015 3 28 2015 Netherlands 3 28 2015 3 28 2015 Brazil 3 28 2015 3 28 2015 Australia 3 29 2015 3 29 2015 Finland 3 30 2015 3 30 2015 Russia 3 30 2015 3 30 2015 Norway 3 30* 2015 4 14 2015 Egypt 3 30 2015 3 30 2015 Kyrgyzstan 3 31 2015 3 31 2015 Malta 3 31* 2015 4 9 2015 Sweden 3 31 2015 3 31 2015 Israel 3 31* 2015 4 15 2015 Portugal 3 31 2015 3 31 2015 South Africa 3 31* 2015 4 15 2015 Azerbaijan 3 31* 2015 4 15 2015 Iceland 3 31 2015 3 31 2015 Poland 3 31* 2015 4 15 2015 Note: Estimated dates are marked with a ‘*’. The 21 countries, including China that signed the Memorandum of Understanding regarding the AIIB on October 24, 2014 in Beijing, and became prospective founding members are not listed here. The relevant documents and the code book are available on the author’s website. Table B1. Dates of Application and Public Known Country Application Date Publicly Known Date2014 Month Day Year Month Day Year Indonesia 11 25 2014 11 15 2014 Maldives 12 17* 2014 12 31 2014 New Zealand 12 18 2014 1 1 2015 Saudi Arabia 12 31 2014 1 13 2015 Tajikistan 12 31 2014 1 13 2015 Jordan 1 24* 2015 2 7 2015 UK 3 12 2015 3 12 2015 Germany 3 17 2015 3 17 2015 France 3 17 2015 3 17 2015 Italy 3 17 2015 3 17 2015 Luxembourg 3 18 2015 3 18 2015 Switzerland 3 20 2015 3 20 2015 UAE 3 20* 2015 4 3 2015 Iran 3 20* 2015 4 3 2015 Turkey 3 26 2015 3 26 2015 Spain 3 27* 2015 4 11 2015 South Korea 3 27 2015 3 27 2015 Austria 3 27 2015 3 27 2015 Georgia 3 28 2015 3 28 2015 Denmark 3 28 2015 3 28 2015 Netherlands 3 28 2015 3 28 2015 Brazil 3 28 2015 3 28 2015 Australia 3 29 2015 3 29 2015 Finland 3 30 2015 3 30 2015 Russia 3 30 2015 3 30 2015 Norway 3 30* 2015 4 14 2015 Egypt 3 30 2015 3 30 2015 Kyrgyzstan 3 31 2015 3 31 2015 Malta 3 31* 2015 4 9 2015 Sweden 3 31 2015 3 31 2015 Israel 3 31* 2015 4 15 2015 Portugal 3 31 2015 3 31 2015 South Africa 3 31* 2015 4 15 2015 Azerbaijan 3 31* 2015 4 15 2015 Iceland 3 31 2015 3 31 2015 Poland 3 31* 2015 4 15 2015 Country Application Date Publicly Known Date2014 Month Day Year Month Day Year Indonesia 11 25 2014 11 15 2014 Maldives 12 17* 2014 12 31 2014 New Zealand 12 18 2014 1 1 2015 Saudi Arabia 12 31 2014 1 13 2015 Tajikistan 12 31 2014 1 13 2015 Jordan 1 24* 2015 2 7 2015 UK 3 12 2015 3 12 2015 Germany 3 17 2015 3 17 2015 France 3 17 2015 3 17 2015 Italy 3 17 2015 3 17 2015 Luxembourg 3 18 2015 3 18 2015 Switzerland 3 20 2015 3 20 2015 UAE 3 20* 2015 4 3 2015 Iran 3 20* 2015 4 3 2015 Turkey 3 26 2015 3 26 2015 Spain 3 27* 2015 4 11 2015 South Korea 3 27 2015 3 27 2015 Austria 3 27 2015 3 27 2015 Georgia 3 28 2015 3 28 2015 Denmark 3 28 2015 3 28 2015 Netherlands 3 28 2015 3 28 2015 Brazil 3 28 2015 3 28 2015 Australia 3 29 2015 3 29 2015 Finland 3 30 2015 3 30 2015 Russia 3 30 2015 3 30 2015 Norway 3 30* 2015 4 14 2015 Egypt 3 30 2015 3 30 2015 Kyrgyzstan 3 31 2015 3 31 2015 Malta 3 31* 2015 4 9 2015 Sweden 3 31 2015 3 31 2015 Israel 3 31* 2015 4 15 2015 Portugal 3 31 2015 3 31 2015 South Africa 3 31* 2015 4 15 2015 Azerbaijan 3 31* 2015 4 15 2015 Iceland 3 31 2015 3 31 2015 Poland 3 31* 2015 4 15 2015 Note: Estimated dates are marked with a ‘*’. The 21 countries, including China that signed the Memorandum of Understanding regarding the AIIB on October 24, 2014 in Beijing, and became prospective founding members are not listed here. The relevant documents and the code book are available on the author’s website. Appendix C Table C1. United Nations Member States that Recognize Taiwan Burkina Faso, Belize, Dominican Republic El Salvador, Guatemala, Guinea Bissau Bissau Haiti, Honduras, Kiribati Nauru, Nicaragua, Paraguay Palau, Panama, Sao Tome, and Principe St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines Solomon Islands, Swaziland, Tuvalu Burkina Faso, Belize, Dominican Republic El Salvador, Guatemala, Guinea Bissau Bissau Haiti, Honduras, Kiribati Nauru, Nicaragua, Paraguay Palau, Panama, Sao Tome, and Principe St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines Solomon Islands, Swaziland, Tuvalu Table C1. United Nations Member States that Recognize Taiwan Burkina Faso, Belize, Dominican Republic El Salvador, Guatemala, Guinea Bissau Bissau Haiti, Honduras, Kiribati Nauru, Nicaragua, Paraguay Palau, Panama, Sao Tome, and Principe St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines Solomon Islands, Swaziland, Tuvalu Burkina Faso, Belize, Dominican Republic El Salvador, Guatemala, Guinea Bissau Bissau Haiti, Honduras, Kiribati Nauru, Nicaragua, Paraguay Palau, Panama, Sao Tome, and Principe St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines Solomon Islands, Swaziland, Tuvalu Appendix D Table D1. Main Variables and Sources of Raw Data Variable Explanation Source Ln GDP Log of country GDP World Bank Ln GDP per capita Log of country GDP per capita World Bank Ln Distance Log of distance between countryi and China CEPII Polity Democracy index Polity IV Project Political Constraints Measure of political constraints W. J. Henisz (2000) Political Rights Measure of political rights Freedom House Civil Liberties Measure of civil liberty Freedom House Taiwan A dummy on whether country i recognizes Taiwan Xinhua ADB Representation Ratio of ADB voting share to GDP ADB IBRD Representation Ratio of IBRD voting share to GDP World Bank IDA Representation Ratio of IDA voting share to GDP World Bank Variable Explanation Source Ln GDP Log of country GDP World Bank Ln GDP per capita Log of country GDP per capita World Bank Ln Distance Log of distance between countryi and China CEPII Polity Democracy index Polity IV Project Political Constraints Measure of political constraints W. J. Henisz (2000) Political Rights Measure of political rights Freedom House Civil Liberties Measure of civil liberty Freedom House Taiwan A dummy on whether country i recognizes Taiwan Xinhua ADB Representation Ratio of ADB voting share to GDP ADB IBRD Representation Ratio of IBRD voting share to GDP World Bank IDA Representation Ratio of IDA voting share to GDP World Bank Table D1. Main Variables and Sources of Raw Data Variable Explanation Source Ln GDP Log of country GDP World Bank Ln GDP per capita Log of country GDP per capita World Bank Ln Distance Log of distance between countryi and China CEPII Polity Democracy index Polity IV Project Political Constraints Measure of political constraints W. J. Henisz (2000) Political Rights Measure of political rights Freedom House Civil Liberties Measure of civil liberty Freedom House Taiwan A dummy on whether country i recognizes Taiwan Xinhua ADB Representation Ratio of ADB voting share to GDP ADB IBRD Representation Ratio of IBRD voting share to GDP World Bank IDA Representation Ratio of IDA voting share to GDP World Bank Variable Explanation Source Ln GDP Log of country GDP World Bank Ln GDP per capita Log of country GDP per capita World Bank Ln Distance Log of distance between countryi and China CEPII Polity Democracy index Polity IV Project Political Constraints Measure of political constraints W. J. Henisz (2000) Political Rights Measure of political rights Freedom House Civil Liberties Measure of civil liberty Freedom House Taiwan A dummy on whether country i recognizes Taiwan Xinhua ADB Representation Ratio of ADB voting share to GDP ADB IBRD Representation Ratio of IBRD voting share to GDP World Bank IDA Representation Ratio of IDA voting share to GDP World Bank Footnotes 1 James E. Rauch, ‘Bureaucracy, Infrastructure, and Economic Growth: Evidence from U.S. Cities During the Progressive Era’, American Economic Review, Vol. 85, No. 4 (1995), pp. 968–79; Sylvie Démurger, ‘Infrastructure Development and Economic Growth: An Explanation for Regional Disparities in China’, Journal of Comparative Economics, Vol. 29, No. 1 (2001), pp. 95–117. 2 Vivien Foster and Cecilia Briceño-Garmendia, eds., Africa’s Infrastructure: A Time for Transformation (Washington, D.C.: World Bank, 2015), pp. 47–49; Jeffrey Gutman, Amadou Sy, and Soumya Chattopadhyay, Financing African Infrastructure: Can the World Deliver? (Washington, D.C.: Brookings, 2015), p. 1. 3 Asian Development Bank, Meeting Asia’s Infrastructure Needs (Mandaluyong City: Asian Development Bank, 2017), https://www.adb.org/sites/default/files/publication/227496/special-report-infrastructure.pdf. 4 The original statement from the AIIB can be found at https://www.aiib.org/en/news-events/news/2017/20171219_001.html. 5 I use the present tense here because the findings are applicable to other IOs established by China and more generally by other developing countries. 6 Barbara Koremenos, ‘Contracting around International Uncertainty’, American Political Science Review, Vol. 99, No. 4 (2005), pp. 549–65. 7 Data on countries’ voting power can be found in the IMF 2015 Annual Report, https://www.imf.org/external/pubs/ft/ar/2015/eng/pdf/AR15-AppIV.pdf. 8 Phillip Y. Lipscy, Renegotiating the World Order: Institutional Change in International Relations (Cambridge: Cambridge University Press, 2017); Julia C. Morse and Robert O. Keohane, ‘Contested Multilateralism’, Review of International Organizations, Vol. 9, No. 4 (2014), pp. 385–412. 9 For AIIB membership information, see http://www.aiib.org/html/pagemembers. 10 For discussion on collective action, see Randall W. Stone, Branislav L. Slantchev, and Tamar R. London, ‘Choosing How to Cooperate: A Repeated Public-Goods Model of International Relations’, International Studies Quarterly, Vol. 52, No. 2 (2008), pp. 335–62; Randall W. Stone, ‘Institutions, Power, and Interdependence’, in Helen V. Milner and Andrew Moravcsik, eds., Power, Interdependence, and Nonstate Actors in World Politics (Princeton: Princeton University Press, 2009), pp. 31–49. For discussion on transaction costs, see Robert O. Keohane, ‘The Demand for International Regimes’, International Organization, Vol. 36, No. 2 (1982), pp. 325–55; Darren G. Hawkins, David A. Lake, Daniel L. Nielson, and Michael J. Tierney, ‘Delegation under Anarchy: States, International Organizations, and Principal-Agent Theory’, in Darren G. Hawkins, David A. Lake, Daniel L. Nielson, and Michael J. Tierney, eds., Delegation and Agency in International Organizations (Cambridge: Cambridge University Press, 2009), pp. 3–38. 11 Randall W. Stone, Controlling Institutions: International Organizations and the Global Economy (Cambridge: Cambridge University Press, 2011). 12 James D. Fearon, ‘Domestic Political Audiences and the Escalation of International Disputes’, American Political Science Review, Vol. 88, No. 3 (1994), pp. 577–92. 13 George W. Downs, David M. Rocke, and Peter N. Barsoom, ‘Managing the Evolution of Multilateralism’, International Organization, Vol. 52, No. 2 (1998), pp. 397–419; Robert D. Putnam, ‘Diplomacy and Domestic Politics: The Logic of Two-Level Games’, International Organization, Vol. 42, No. 3 (1988), pp. 427–60; Robert O. Keohane, ‘International Institutions: Two Approaches’, International Studies Quarterly, Vol. 32, No. 4 (1988), pp. 379–96. 14 For discussion on ‘domestic gridlock’, see Samuel Brazys and Diana Panke, ‘Why do States Change Positions in the United Nations General Assembly?’, International Political Science Review, Vol. 38, No. 1 (2017), pp. 70–84. For discussion on homophily, see Miller McPherson, Lynn Smith-Lovin, and James M Cook, ‘Birds of a Feather: Homophily in Social Networks’, Annual Review of Sociology, Vol. 27 (2001), pp. 415–44. 15 Alberto Alesina and David Dollar, ‘Who Gives Foreign Aid to Whom and Why?’, Journal of Economic Growth, Vol. 5, No. 1 (2000), pp. 33–63. 16 David B. Carter and Randall W. Stone, ‘Democracy and Multilateralism: The Case of Vote Buying in the UN General Assembly’, International Organization, Vol. 69, No. 1 (2015), pp. 1–33. 17 Julia Bader, ‘China, Autocratic Patron? An Empirical Investigation of China as a Factor in Autocratic Survival’, International Studies Quarterly, Vol. 59, No. 1 (2015), pp. 23–33; Julia Bader, ‘Propping up Dictators? Economic Cooperation from China and its Impact on Authoritarian Persistence in Party and Non-Party Regimes’, European Journal of Political Research, Vol. 54, No. 4 (2015), pp. 655–72; Julia Bader and Ursula Daxecker, ‘A Chinese Resource Curse? The Human Rights Effects of Oil Export Dependence on China versus the United States’, Journal of Peace Research, Vol. 52, No. 6 (2015), pp. 774–90. 18 For discussion on domestic political constraints, see Alexandre Debs and Jessica Chen Weiss, ‘Circumstances, Domestic Audiences, and Reputational Incentives in International Crisis Bargaining’, Journal of Conflict Resolution, Vol. 60, No. 3 (2014), pp. 403–33; Adam Przeworski, Susan C. Stokes, and Bernard Manin, eds., Democracy, Accountability, and Representation (Cambridge: Cambridge University Press, 1999). 19 An alternative mechanism is competition. While China is a developing country, it is also the world’s second largest economy, and poised to become the world’s largest economy in the next decade or so. Therefore, countries do have an incentive to develop and maintain a good relationship with China, and joining the AIIB early as a founding member is a good opportunity to do so. Unfortunately, there is no easy way to distinguish whether it is the flow of information or the flow of competition that gives rise to the observed neighbor effects. In this article, I will stick to the mechanism of information flow. 20 Andreu Mas-Colell, Michael D. Whinston, and Jerry R. Green, Microeconomic Theory (Oxford: Oxford University Press, 2012); John W. Pratt, ‘Risk Aversion in the Small and in the Large’, Econometrica, Vol. 32, No. 1/2 (1964), pp. 122–36. 21 For discussion on interdependence, see Robert O. Keohane and Joseph S. Nye, Jr., ‘Power and Interdependence Revisited’, International Organization, Vol. 41, No. 4 (1987), pp. 725–53. For discussion on transnational diffusion, see Fabrizio Gilardi, ‘Transnational Diffusion: Norms, Ideas, and Policies’, in Walter Carlsnaes, Thomas Risse, and Beth Simmons, eds., Handbook of International Relations (Thousand Oaks: SAGE Publications, 2012), pp. 453–77; Beth A. Simmons and Zachary Elkins, ‘The Globalization of Liberalization: Policy Diffusion in the International Political Economy’, American Political Science Review, Vol. 98, No. 1 (2004), pp. 171–89. For discussion on interorganizational learning, see Johan Bruneel and Bart Clarysse, ‘Learning from Experience and Learning from Others: How Congenital and Interorganizational Learning Substitute for Experiential Learning in Young Firm Internationalization’, Strategic Entrepreneurship Journal, No. 4 (2010), pp. 164–82. 22 Kenneth N. Waltz, Theory of International Politics (New York: McGraw-Hill Publishing Company, 1979); Robert Powell, In the Shadow of Power: States and Strategies in International Politics (Princeton: Princeton University Press, 1999); David A. Lake, Hierarchy in International Relations (Ithaca: Cornell University Press, 2009). 23 For discussion on how institutional changes are affected by an IO’s policy area, see Phillip Y. Lipscy, ‘Explaining Institutional Change: Policy Areas, Outside Options, and the Bretton Woods Institutions’, American Journal of Political Science, Vol. 59, No. 2 (2015), pp. 341–56. 24 Julia C. Morse and Robert O. Keohane, ‘Contested Multilateralism’, pp. 385–412; Joseph Jupille, Walter Mattli, and Duncan Snidal, Institutional Choice and Global Commerce (Cambridge: Cambridge University Press, 2013); Robert O. Keohane and David G. Victor, ‘The Regime Complex for Climate Change’, Perspectives on Politics, Vol. 9, No. 1 (2011), pp. 7–23. 25 G. John Ikenberry and Darren Lim, China’s Emerging Institutional Statecraft: The Asian Infrastructure Investment Bank and the Prospects for Counter-Hegemony (Washington, D.C.: Brookings, 2017), https://www.brookings.edu/wp-content/uploads/2017/04/chinas-emerging-institutional-statecraft.pdf. 26 The framework of contested multilateralism mostly focuses on the big powers. My work, in contrast, studies the decision making of the ‘small’ powers. 27 Δt is reset to 0 whenever a country joins and is incremented by 1 each day thereafter. 28 This set of variables is also standard in international trade literature. See, for example, Paul Krugman, ‘Scale Economies, Product Differentiation, and the Pattern of Trade’, American Economic Review, Vol. 70, No. 5 (1980), pp. 950–59; Thomas Chaney, ‘The Gravity Equation in International Trade: An Explanation’, Journal of Political Economy, Vol. 126, No. 1 (2018), pp. 150–77. 29 Phillip Y. Lipscy, ‘Who’s Afraid of the AIIB’, Foreign Affairs, 7 May, 2017, https://www.foreignaffairs.com/articles/china/2015-05-07/whos-afraid-aiib. 30 The latter argument is also supported by the latest AIIB data. As of February 2018, the AIIB has approved 25 projects, and all of these projects are located in member countries. For details of these projects, please see https://www.aiib.org/en/projects/approved/index.html. 31 Its design follows Thomas Chaney, ‘The Network Structure of International Trade’, American Economic Review, Vol. 104, No. 11 (2014), pp. 3600–34. The naming convention follows Thomas Chaney, ‘Distorted Gravity: The Intensive and Extensive Margins of International Trade’, American Economic Review, Vol. 98, No. 4 (2008), pp. 1707–21. 32 David B. Carter and Curtis S. Signorino, ‘Back to the Future: Modeling Time Dependence in Binary Data’, Political Analysis, Vol. 18, No. 3 (2010), pp. 271–92. 33 For alternative ways to capture time dependency, such as time dummies and/or cubic splines, see Nathaniel Beck, Jonathan N. Katz, and Richard Tucker, ‘Taking Time Seriously: Time-Series-Cross-Section Analysis with a Binary Dependent Variable’, American Journal of Political Science, Vol. 42, No. 4 (1998), pp. 1260–88. 34 Janet M. Box-Steffensmeier and Bradford S. Jones, Event History Modeling: A Guide for Social Scientists (Cambridge: Cambridge University Press, 2004). 35 I use GDP and GDP per capita data for the year 2013 rather than 2014 because data with respect to the latter year on many countries have not yet been reported. 36 http://www.cepii.fr/cepii/en/bdd_modele/bdd.asp. 37 For polity scores, see http://www.systemicpeace.org/polity/polity4.htm. For political constraints, see W. J. Henisz, ‘The Institutional Environment for Economic Growth’, Economics & Politics, Vol. 12, No. 1 (2000), pp. 1–31. For data on political rights and civil liberties, see https://freedomhouse.org/report/freedom-world-2016/table-scores. 38 http://eng.sectsco.org. 39 http://www.cepii.fr/cepii/en/bdd_modele/bdd.asp. 40 http://english.mofcom.gov.cn. 41 Please note that scores for political rights and civil liberties range between 1 and 7, with 1 representing the freest and 7 the least free. This explains why the coefficients of political rights and civil liberties bear different signs from those of polity scores, which range from −10 to 10, with −10 representing autocracy and 10 full democracy. 42 Stone, Slantchev, and London, ‘Choosing How to Cooperate’, pp. 31–49; Robert O. Keohane, After Hegemony: Cooperation and Discord in the World Political Economy (Princeton: Princeton University Press, 2005). 43 Here I only consider the 57 founding members, as they are the focus of this study. As of June 16, 2017, the number of approved memberships had risen to 80. 44 Julia C. Morse and Robert O. Keohane, ‘Contested Multilateralism’, pp. 385–412. 45 I find that the correlation between the variable Asia and representation in the World Bank to be negative. In particular, the negative correlation between Asia and IDA representation is statistically significant. This highlights the general under-representation of Asian countries in the world’s financial system. 46 Beck, Katz, and Tucker, ‘Taking Time Seriously’, pp. 1260–88; Kjell A. Doksum and Miriam Gasko, ‘On a Correspondence between Models in Binary Regression Analysis and in Survival Analysis’, International Statistical Review, Vol. 58, No. 3 (1990), pp. 243–52. 47 Beck, Katz, and Tucker, ‘Taking Time Seriously’, pp. 1260–88; Carter and Signorino, ‘Back to the Future’, pp. 271–92. © The Author(s) 2018. Published by Oxford University Press on behalf of The Institute of International Relations, Tsinghua University. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

The Chinese Journal of International PoliticsOxford University Press

Published: May 22, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off