Trade, security, and authoritarianism: domestic politics of foreign policy making in China

ABSTRACT Could trade bring security and stability to international relations? China has become one of the most important players in the global economy without being democratized. In this article, I present a game-theoretic model in which an authoritarian state would behave more cooperatively in international relations if internationalists are empowered vis-à-vis hardliners in domestic politics. Reformists and internationalists are likely partners. The reformists have a strong incentive to use the free trade agreements’ (FTAs) stipulating domestic regulations as external pressure to advance structural reforms: they see intra-industry trade, enhanced by global value chains (GVCs), as a means to achieve sustainable economic growth. This article breaks down the broader claim that FTAs regulating GVCs-based intra-industry trade are connected to regional security in two steps: first, FTAs empower the reformist-internationalist coalition; and second, the empowered reformist internationalists make the state’s behavior more cooperative in the international sphere. The model’s equilibrium outcome implies that the FTAs that require the signatories to be committed to domestic economic reforms will promote international cooperation and strengthen regional security by empowering people who favor reform and internationalism in each state’s domestic politics. This means, if an authoritarian state is involved in the global economy and benefits from international economic transactions, like China does, then concluding and spreading FTAs requiring domestic economic reforms would be more effective policy to bring security and stability than promoting democratic transitions.


Introduction
Could trade bring security and stability to international relations? Realist theories argue that states' taking international relations as zero-sum interactions is unavoidable in the anarchic nature of the international system. Liberal theories argue, first, that economic interdependence in the market system produces common interests and therefore decreases conflict among states and decreases the role of military power and the insecurity it breeds 1 Institutions, then, both international and domestic, can mitigate the effects of anarchy and, as a result, there is opportunity for positive-sum, mutually beneficial cooperation 2 While realists conclude that international cooperation is difficult, and conflict and insecurity are enduring features of world politics, liberals argue that market interactions and institutions can strengthen each state's security by promoting international cooperation. The question of whether trade could bring global security and stability is particularly important to consider when evaluating the implications of China's rise on U.S. foreign policy, considering the United States has a long history of "supporting democratic transitions and economic liberalization around the world." 3 China has become one of the most important players in the global market economy without being democratized. Edward Steinfeld argues that, even without democratization, China's economic growth enhances the U.S.-led liberal international order because it means China is playing our game by our rules 4 In this article, I advance these liberal arguments by examining the security implications of the emerging intra-industry trade of manufacturing industries induced by the spread of global value chains (GVCs) -where manufacturers are procuring parts produced beyond national borders -of multinational corporations in the Asia-Pacific region 5 In manufacturing industries, it has become common practice for different stages of production to be located in different countries. The development of GVCs over the last two decades has brought a new international division of labor between developed and developing countries, which includes both authoritarian and democratic countries 6 I argue that GVCs based intra-industry trade brings increased international cooperation to the security relationships in the Asia-Pacific because each state is motivated to engage in institution building in both international and domestic politics.
This article proceeds as follows. The second section discusses the interrelationship between economy and security by introducing how the game-theoretic model accounts for the logic of the argument. The third section presents a basic gametheoretic model, laying out the model's main assumptions. While at the domestic level two factions are contending with each other for making coalitions, at the international level an authoritarian state and a democratic state are strategically interacting whereas there is uncertainty over the other player's payoffs. The fourth section sketches the main features of the equilibria that exist at the international level when the authoritarian country's leadership is divided over domestic economic reforms and foreign policies. This section highlights the incentives that actors have in this strategic setting, and particularly the choices that they face in initiating or responding to a cooperative gesture. The fifth section applies the model to domestic politics of Chinese foreign policy making. I argue that an FTA requiring domestic economic reforms will enhance regional security by empowering reformist internationalists vis-à-vis conservative hardliners in the balance of power of Chinese domestic politics even if China is not a member of such an FTA.

The argument
The spread of GVCs has changed the rule-making of international trade. In the past, trade negotiations mainly discussed lowering tariffs and other trade barriers imposed at the national border to manage inter-industry trade in which nations trade raw materials and final goods among different industries. However, as more international economic interactions have become GVCs based trade, building rules on domestic institutions such as regulations over foreign direct investment, state-owned enterprise (SOE) reforms, and intellectual property rights is necessary to manage the intra-industry trade, shifting the focus of trade negotiations from national border measures to domestic regulations. In the meantime, the World Trade Organization (WTO) has failed to function as a forum to discuss the complicated and comprehensive rules needed to manage GVCs based trade. As a result, many nations -both developed and developing as well as both democratic and authoritarian -have concluded free trade agreements (FTAs) for this rule-making all over the world -especially in the Asia-Pacific region.
Meanwhile, in the domestic sphere, if a state is involved in the global economy and benefits from economic transactions, like China does, GVCs-based trade motivates reformists to make a coalition with internationalists, who appreciate a peaceful international environment 7 The reformists, who want to advance structural reforms -such as China's SOE reform -in the nation's domestic politics, understand the necessity to maintain good relations with their trading partners to benefit from their economic interactions, and therefore argue for cooperative foreign policy. In the meantime, structural reforms undermine the vested interests of the collusive and corrupt rentseeking scheme. As a result, reformists often face backlash from conservatives who enjoy the benefit from such vested interests 8 Thus, while reformists and internationalists are likely partners, conservatives are motivated to make a coalition with hardliners and to support aggressive and abrasive foreign policy to protect their vested interests. To confront the conservative-hardliner coalition, the reformists have a strong incentive to use FTAs requiring domestic economic reforms as a gaiatsu (literally meaning "foreign pressure") to advance the structural reforms. This article breaks down the broader claim that FTAs regulating GVCs based trade are connected to regional security in two steps: first, FTAs empower the reformist-internationalist coalition against the conservativehardliner coalition; and second, the empowered reformist internationalists make the nation's behavior more cooperative in the international sphere.
In this article, I develop a two-level game that examines political implications of FTAs, taking into account the strategic interdependence of the domestic and international levels. At the domestic level, the game captures the struggle for power between two groups in the authoritarian context where no institutional mechanisms constrain the fundamental part of decision-making of both domestic politics and international relations. In the model, there are two dimensions of coalition making: one is conservatives against reformists, and the other is hardliners against internationalists. The exact nature of the government is determined as a result of contests for coalition-building among contending factions over 7 Internationalism here is defined as the position to appreciate international cooperation and coordination, and is not related to the neo-conservative internationalism in the American foreign policy debate.. 8 Conservatism here is defined as the position to protect the rent-seeking scheme, and is not related to the conservative ideology in American politics. Thus, in this article I use "conservatism" interchangeably with "anti-reformism.". some domestic economic policy and foreign security policy 9 I assume that a new FTA requiring domestic economic reforms will undermine the vested interests of the rentseeking scheme. Given this assumption, the FTA brings a gaiatsu to advance the politically sensitive domestic economic reforms. In this way, it would empower people who favor reform and internationalism in each state's domestic politics, and make the state's behavior more cooperative. At the international level, states must decide whether to be aggressive or cooperative. In the model, there are two states: one is authoritarian and the other is democratic. Each state's leader is uncertain about the preference of the other. Possible types include hardliners, who have strong incentives to free ride the other state's cooperative stance by unilaterally taking aggressive policy, and internationalists, who would give a priority to cooperating to gain economic benefits from interdependent relations. The interaction at the international level highlights a number of problems identified in the literature on why international cooperation is so difficult under uncertainty. States are trapped in a security dilemma, which arises when efforts that states make to defend themselves lead other states to feel less secure and to fear that they will be attacked.
The central implications of this game-theoretic model are as follows. In the model's equilibrium, if an FTA empowers internationalists vis-à-vis hardliners in the power balance of domestic politics, it will make the authoritarian state's behavior more cooperative in international relations. Such an outcome is expected in some of the FTAs under certain conditions. If the FTA requires domestic economic reforms, it will undermine the rent-seeking scheme benefiting the conservative hardliners. The FTAs stipulating comprehensive rules on GVCs based trade will require the signatories to be committed to domestic economic reforms, imposing a gaiatsu on governments to reform the politically sensitive issues. For example, the SOE reform in China is one of those politically sensitive issues as particularistic vested interests are formed under the SOE system in the Chinese state capitalism. Therefore, the international rule-making on GVCs based trade will make each nation's behavior more cooperative in international relations by empowering the reformist-internationalist coalition vis-à-vis the conservative-hardliner coalition in domestic politics. Among the FTAs in the Asia-Pacific region, the Comprehensive and Progressive Agreement for Trans-Pacific Partnership (CPTPP), which requires domestic economic reforms, would serve for this purpose, while the Chinese-led trade agreement Regional Comprehensive Economic Partnership (RCEP), which does not include such a requirement, would not.
The broader implications of this game-theoretic model are concerned with the debate over whether and how capitalism -instead of democracy -may bring peace to the world. The literature of commercial liberalism -also called the capitalist peace theory -has argued that states will not fight war if they trade with each other because it is in neither state's best interest. Patrick McDonald takes this capitalist peace theory a step further arguing that capitalist institutions provide a foundation for peace by offering liberal domestic economic institutions such as property rights regimes 10 It is these competitive 9 Actors such as reformists versus conservatives and internationalists versus hardliners are set exclusively for each other.
However, this assumption does not mean that their policy positions are not changed. In fact, in the model discussed in this article, reformists may make a coalition with conservative hardliners when they have acquired certain vested interests.. 10 McDonald, The Invisible Hand of Peace.. market structures that constrain the political rent-seeking of leaders and make them more responsive to popular demands for peace. Concurring with McDonald's argument, in this article I suggest that, as the market-promoting domestic institutions bring peace to international relations, the FTA that requires signatories to commit to regulatory reforms will enhance regional security. I argue that international rule-making of the FTAs that stipulate domestic regulations will enhance regional security because such an FTA may motivate an authoritarian state involved in GVCs based intra-industry trade -such as China -to empower the internationalists who appreciate international cooperation with trading partners.

The game-theoretic model
The game-theoretic model captures strategic interactions on both the domestic and international levels. At the domestic level, I assume that the state's leadership is composed of two contending factions over domestic economic policy: conservatives and reformists. Initially, one of these groups is stronger than the other in the balance of power of the state's domestic politics. Each group chooses its coalition partner between the other contending factions over foreign security policy: hardliners and internationalists. Some type of international shock (e.g., the election of President Donald Trump, the signing of a new FTA, U.S. withdrawal from liberal international order, the global financial crisis, etc.) may change which group is in power domestically and what foreign policy the state takes. There are at least three potential outcomes by which this shock may shift the domestic balance of power and influence regional security. It might: (a) alter the domestic balance of power so much that it brings a new coalition to power, (b) trigger instead some concessions that leave conservatives in power but encourage them to engage in international cooperation, or (c) have no effect on the domestic balance of power or the foreign policy that the state pursues.
At the international level, there are two actors, an authoritarian state and a democratic state. The authoritarian state decides whether to be aggressive to expand its sphere of influence or to be cooperative to enjoy the economic benefit from interdependent relations with the democratic state. If the authoritarian state has taken an aggressive stance, then the democratic state will have to decide whether to retaliate or to yield to maintain the economic benefit from interdependent relations. If the democratic state retaliates, then the authoritarian state will decide whether to escalate the conflict or back down. The crucial dilemma for the democratic state is that it does not know if the authoritarian state is hardliner or internationalist, that is, whether the authoritarian state is inclined to escalate the conflict even if it means sacrificing the economic benefit from interdependent relations.

Outcomes and payoffs
The payoff assignment for the game at the international level is similar to that developed by Kenneth Schultz, as shown in Figure 1 11 There are four possible outcomes: mutual cooperation, mutual defection, unilateral defection by A, and unilateral defection by B. The game assumes that R > T > W > 0 12 The parameters α and β indicate how hawkish the actors are, with higher values meaning, ceteris paribus, a greater willingness to take an aggressive stance.
The international component of the game can be formalized in a simple incomplete information game, as shown in Figure 2. The nature node indicates whether the authoritarian state is hardliner or internationalist. The hardliner type's payoff for "war" from the conflict escalation, W A þ α, is higher than its payoff for benefiting from international commerce in a peaceful environment, R A . For the internationalist type, the economic benefit from interdependent relations is higher than the political benefit from fighting (i.e., R A > W A þ α) 13 Following the nature node, the authoritarian state has the first move, choosing either an aggressive stance or a cooperative stance. If the authoritarian state decides to take the cooperative stance, the game ends and both players get their "reward" payoffs, R A for the authoritarian state and R D for the democratic state. If the authoritarian state decides to take the aggressive stance, the democratic state has the option to retaliate the aggression or yield to it. If the democratic state decides to yield, the game ends by changing the balance of power in the authoritarian state's favor. In this acquiescence outcome, where the authoritarian state gets T A þ α and the democratic state gets 0, the democratic state loses payoffs (i.e., R D > 0). Meanwhile, whether the authoritarian state gains or loses payoffs depends on how much the authoritarian state is willing to sacrifice the status quo benefit from economic interdependence to gain the political benefit from the new balance of power more favorable vis-à-vis the democratic state (i.e., whether T A þ α is larger or smaller than R A depends on the value of α). If the democratic state decides to retaliate, the authoritarian state has the option to escalate the conflict or back down. If the authoritarian state backs down, then the game ends with changing the balance of power in the democratic state's favor. The payoff assignment at the back-down outcome is exactly opposite to that at the acquiescence outcome. In the back-down outcome, where the authoritarian state gets the payoff of 0 and the democratic state gets T D þ β, the authoritarian state loses payoffs from the decision to back down (i.e., R A > 0). Meanwhile, whether the democratic state gains or 12 I assume that R > T À W, following the logic in Schultz that "cooperation is relatively risky, in the sense that the spread between the payoffs that can be obtained from cooperating [R or 0] is large relative to the spread between the payoffs that can be obtained from defecting [T or W]." Ibid., 7.. 13 Thus, α > R A À W A for the hardliner type while α < R A À W A for the internationalist type.. loses payoffs depends on how much the democratic state is willing to sacrifice the status quo benefit from the new balance of power more favorable vis-à-vis the authoritarian state (i.e., whether T D þ β is larger or smaller than R D depends on the value of β). Finally, if the authoritarian state escalates, then both get the "war" payoffs, W A þ α for the authoritarian state and W D þ β for the democratic state.

The structure of the domestic level game
The nature node makes the authoritarian state internationalist with probability p on the right branch of the game tree, and hardliner with probability 1 À p on the left branch of the tree. As indicated by the information set linking the democratic state's decision nodes, the democratic state is not informed of what type of player it faces. However, the probability that the authoritarian state is internationalist, p, is known to the democratic state; and the greater p is, the more certain the democratic state is that the authoritarian Nature Authoritarian Authoritarian state is internationalist. This prior expectation regarding the authoritarian state's type can be a product of several things. If the authoritarian state has been involved in the global economy, this could result in a higher level of p. For instance, China has advanced market-oriented economic reforms since the 1980s, which has led China to benefit from economic interactions with democratic states such as the United States and Japan; and as a result, democratic states may expect China to maintain good relations with those democratic trading partners. The prior expectation could also be a result of a membership of international institutions. If the authoritarian state is enmeshed in U.S.-led international economic institutions such as the WTO, then it may have a stronger incentive to cooperate with the United States and other democratic states, which means a higher p. Finally, the prior expectation could be influenced by domestic politics of the authoritarian state. If those who appreciate the benefit from international cooperation are empowered in the authoritarian state's domestic politics, and if that information is credibly signaled to other states, then the democratic state will believe that the authoritarian state is more likely to be internationalist. For instance, if reformist internationalists are empowered vis-à-vis conservative hardliners in Chinese domestic politics, then the United States or Japan will expect China to be more likely to cooperate in international relations.
To capture the domestic component of the game, I assume that how likely the authoritarian state is the internationalist type (i.e., the value of p) is determined by factional politics in the authoritarian state's leadership. In particular, I assume that there are two contending factions: conservatives and reformists. The matrix in Figure 3 shows the incentives affecting each faction to form a coalition. Each faction has two strategies: coalition with hardliners (HL) and coalition with internationalists (INT).
Initially, conservatives are motivated to make a coalition with hardliners while reformists are motivated to make a coalition with internationalists. However, some international shock may trigger new coalition building for both factions. I assume that if conservatives switch their coalition partner from hardliners to internationalists

Hardliners (HL)
Internationalists ( reformist policy will be implemented regardless of what strategy reformists take. The utility each faction would gain from reformist policy is denoted M. Even the conservatives who oppose the reformist policy would gain M because the benefit of reformist policy is nonexcludable and nonrival 14 If neither faction changes its coalition partner, then reformist policy will be implemented with probability q and conservative policy will be implemented with probability 1 À q. This assumption implies that when conservative policy is implemented conservatives will gain the whole rents. The value of q reflects the initial balance of power between conservatives and reformists. A higher value of q indicates that reformists are more powerful vis-à-vis conservatives. Contrary to the benefit from reformist policy, the benefit from conservative policy is excludable and rival 15 Thus, the utility gained from the vested interest of the rent-seeking scheme, denoted V C for conservatives and V R for reformists, is given only to those who support the conservative policy. If both factions make coalitions with hardliners, conservative policy will be implemented and the available rents will be shared by the two factions. The conservatives' utility from the rent in this outcome, denoted V 0 C , is lower than the case where the conservatives can gain the whole rents (i.e., V 0 C < V C ). In this outcome, because the reformists support the conservative policy, the conservatives will have to share the rents with the reformists. I assume here that the conservatives give smaller rents to the reformists than the amount they keep (i.e., V R < V 0 C ). In particular, I assume that V R ¼ qV 0 C 16 This assumption implies that the split of the rent between the conservatives and the reformists is based on the initial balance of power between these two factions. In other words, the more powerful the reformists are, the higher share of the rents the reformists gain. Figure 4 describes the equilibrium strategies of the two factions as a set of thresholds along the continuum of possible circumstances, along with the association between the values of p and q 17 When the value gained from implementing reform policy is high, even conservatives would alter their coalition partner from the hardliners to the internationalists. As the value of reform policy vis-à-vis conservative policy is lower, conservatives would have less incentive to switch their coalition partner, and at a new equilibrium both conservatives and reformists would stay with their original coalition partners. When the value from reformist policy is even lower than a certain threshold, even reformists will join the coalition with hardliners to seek the rents generated by the vested interest. A formal description of the solution is in Appendix 1.

Equilibria in the domestic level game
When the expected value of the benefit from reformist policy is very high, both conservatives and reformists would have a dominant strategy in making a coalition with internationalists. As a result, conservatives would join the coalition of reformists and internationalists. Then, once the expected value of the benefit from reformist policy is lower than a certain threshold, conservatives would have an incentive to deviate from 14 This assumption implies that reformist policy has the universal nature of public goods.. 15 This assumption implies that conservative policy has the particularistic nature of private goods.. 16 I implicitly assume that V C ¼ 1 þ q ð ÞV 0 C .. 17 I write outcomes using the notation: (Conservatives' strategy, Reformists' strategy).. the coalition with internationalists and go back to make the original coalition with hardliners. In this case, reformists would still stay with the coalition with internationalists. Finally, if the expected value of the benefit from reformist policy is even lower, both conservatives and reformists would have a dominant strategy in making a coalition with hardliners. In this case, reformists would stop counting on the benefit from reformist policy but instead join the conservative-hardliner coalition to seek the benefit from the vested interests of the collusive and corrupt rent-seeking scheme.

Initiating cooperation: equilibrium behavior in the international level game
The concept I employ to solve the international level game is the perfect Bayesian equilibrium, requiring that beliefs be updated on the equilibrium path according to Bayes' rule. There are two types of Bayesian equilibria in the game: one separating equilibrium and one semi-separating equilibrium 18 A formal description of this game's equilibria is discussed in Appendix 2.
At the very last node of the game, on the left branch where the authoritarian state is hardliner (i.e., α > R A À W A ), the authoritarian state will escalate the conflict. On the right branch where the authoritarian state is internationalist (i.e., α < R A À W A ), the decision depends on the value of α, which means how pacifist the internationalist type of authoritarian state is. If the authoritarian state is pacifist internationalist (i.e., α < À W A ), then it will back down. If the authoritarian state is moderate internationalist (i.e., R A À W A > α > À W A ), then, it will escalate the conflict. Now consider the democratic state's strategy. Suppose that p 0 is the democratic state's perception of the updated probability that the authoritarian state is internationalist once the democratic state has seen that the authoritarian state has taken an aggressive stance. First, I consider the case where the internationalist type of authoritarian state is pacifist internationalist. In this case, the democratic state's expected payoff from retaliating is p 0 T D þ 1 À p 0 ð ÞW D þ β whereas its expected payoff from yielding is 0. Thus, the democratic state will retaliate if p 0 exceeds a critical threshold p � defined in the following identity: p � ; À W D À β T D À W D . This condition implies that the strategy to be taken by the democratic state would be different depending on three types of the democratic state: hardliner, moderate, and pacifist. If the democratic state is hardliner (i.e., β > À W D ), it will always retaliates regardless of the expectation of how the authoritarian state would respond in the next node. If the democratic state is moderate (i.e., À T D < β < À W D ), it will retaliate when p 0 > p � expecting that the authoritarian state is likely to back down if the democratic state retaliates, while it will yield when p 0 < p � expecting that the authoritarian state is likely to escalate if the democratic state retaliates. If the democratic state is pacifist (i.e., β � À T D ), it will always yield regardless of the value of p 0 . Next, I consider the case where the internationalist type of authoritarian state is moderate internationalist. In this case, the strategy taken by the democratic state does not depend on the value of p 0 . Because the authoritarian state would escalate the conflict whether it is hardliner or internationalist, the democratic state's expected payoff from retaliating is W D þ β. Thus, the democratic state's strategy depends on which type the democratic state is. Namely, the democratic state will retaliate if it is hardliner, while it will yield if it is moderate or pacifist.
Continuing up the tree, I consider the authoritarian state's decision to take an aggressive stance or a cooperative stance toward the democratic state. First, in the separating equilibrium, only the hardliner type of authoritarian state will take an aggressive stance while the internationalist type will take a cooperative stance. Formally, this means that p 0 ¼ 0, which is smaller than p � . Thus, in the separating equilibrium, the moderate type of democratic state will yield regardless of the type of the authoritarian state. This equilibrium is a perfect Bayesian equilibrium, because the hardliner type of authoritarian state does not have an incentive to deviate from taking the aggressive stance while the internationalist type does not have an incentive to deviate from taking the cooperative stance, given that the hardliner type of democratic state would retaliate and the moderate or pacifist type would yield when the authoritarian state has taken the aggressive stance 19 The separating equilibrium implies the following proposition: Proposition 1: Regardless of the type of the democratic state (i.e., the value of β), the internationalist type of authoritarian state always has an incentive to initiate cooperation with the democratic state while the hardliner type never has an incentive to do so. In other words, the authoritarian state's type determines whether the authoritarian state initiates cooperation.
Second, in the semi-separating equilibrium, both the authoritarian state and the democratic state randomize their strategy, taking a mixed strategy. Formally, in the semiseparating equilibrium, the hardliner type will always take an aggressive stance but the internationalist type will take an aggressive stance with probability b, and the democratic state will retaliate with the probability s if the authoritarian state has taken the aggressive 19 The only exception is the case where the authoritarian state is less pacifist moderate internationalist (i.e., R A À T A < α < R A À W A ). In this case, the internationalist type has an incentive to deviate from taking the cooperative stance given that the democratic state will choose to yield..

stance. In this equilibrium, s ¼ T A þαÀ R A T A þα
if the authoritarian state is pacifist internationalist, and s ¼ T A þαÀ R A T A À W A if it is moderate internationalist, which means that the probability that the democratic state retaliates depends on neither the democratic state's expectation of the authoritarian state's type (i.e., the value of p) nor the democratic state's preference ordering (i.e., the value of β or the democratic state's payoffs), but depends on the authoritarian state's preference ordering (i.e., the value of α and the authoritarian state's payoffs). In the meantime, , which means that the probability that the internationalist type of authoritarian state takes the aggressive stance is negatively associated with the probability that the democratic state expects the authoritarian state to be internationalist (i.e., the value of p), and negatively associated with the democratic state's appreciation of conflict escalation (i.e., the value of β). That is, if the democratic state benefits more from conflicting relations, ceteris paribus the internationalist type of authoritarian state is more likely to take the cooperative stance. However, the magnitude of the democratic state's bellicose preference inducing the authoritarian state's cooperation is smaller than that of the authoritarian state becoming more internationalist. As a result, the overall probability that the authoritarian state initiates cooperation, defined as ρ C ;p 1 À b ð Þ, will increase as the democratic state expects the authoritarian state more likely to be internationalist (i.e., ρ C will increase when p increases). This equilibrium outcome has the following implications for my argument. An FTA requiring domestic economic reforms undermines vested interests based on the collusive and corrupt rentseeking scheme and hence empower the reformist internationalists vis-à-vis the conservative hardliners, which means the increase in the value of p through the increase in the value of M relative to V. As a result, the authoritarian state will be more likely to initiate cooperation in international relations, strengthening regional security.
In the meantime, the probability that the democratic state yields to the authoritarian state's aggression, defined as ρ Y ;pb 1 À s ð Þ þ 1 À p ð Þ 1 À s ð Þ, will decrease as the authoritarian state is more likely to be internationalist. In other words, the more likely the authoritarian state is hardliner, the more likely the democratic state yields. This equilibrium outcome implies the following proposition: Proposition 2: One of the democratic state's equilibrium strategies is first to empower hardliners vis-à-vis internationalists in the authoritarian state's domestic politics (i.e., the decrease in the value of p), then to raise the probability that the authoritarian state takes an aggressive stance, and in the end to yield to the authoritarian state's aggression. Although the democratic state is apparently aggressive in this equilibrium, it will compromise with the authoritarian state and the balance of power is changed in the authoritarian state's favor, weakening regional security.

The "beyond-economics" importance of free trade agreements
To provide some empirical background for this game-theoretic model, I discuss how FTAs would influence regional security in the Asia-Pacific, with special attention to the impact on Chinese domestic politics and foreign policy making. FTAs requiring domestic economic reforms would enhance regional security and increase international cooperation, because such an FTA would change the balance of power in each state's domestic politics in favor of reformist internationalists. Reformists who see GVCs based trade as a means to achieve sustainable economic growth would have a strong incentive to use the FTAs as a gaiatsu to advance domestic economic reforms when facing the backlash from conservatives who give a priority to protecting their vested interests of the collusive and corrupt rent-seeking scheme. The reformists would appreciate a peaceful and cooperative international environment while the conservatives would use aggressive and abrasive foreign policy as a means to resist the reformist agenda that would undermine their vested interests. Therefore, the empowered reformists would make the nation's behavior more cooperative in the international sphere by making a coalition with the internationalists in the domestic sphere.
My goal here is to try to reinterpret Chinese domestic politics over foreign policy making by showing how reformists and internationalists make coalitions to confront the coalitions of conservatives and hardliners. The discussion here is built on Julian Gewirtz's argument that a coalition of Chinese reformers and Western economists was what made it possible for the post-Mao economic reform to emerge in the 1980s 20 Gewirtz suggests that the origin of China's current prosperity is not nationalism but internationalism.
Donald Trump's election to be U.S. President in 2016 added two complications to the debate about Chinese elite politics and its implications on China's foreign economic policy. First, it is now salient to ask the question of whether the U.S.-led international order is dead as anti-internationalism is rising in the United States and in other parts of the world. In Liberal Leviathan, published in 2011 as a response to the George W. Bush administration's Iraq War, John Ikenberry discussed the possibility of a breakdown of liberal international order but argued that the breakdown would be unlikely because the United States had benefitted from the U.S.-led liberal international order 21 However, since the election of President Trump the breakdown has become a much more likely scenario. Second, the implications drawn from the model suggest the changing political economy of the Asia-Pacific region and the world due to China's economic rise, the relative decline of U.S. economic power, and deepened regional integration 22 Backed by newly acquired economic power, China has certainly been interested in lowering U.S. influence in East Asia, and has taken initiative in forming international order and breaking ties between the United States and its allies. Although the Trump administration was using provocative rhetoric, the administration's unpredictable nature raised a suspicion that the United States could suddenly make a backdoor deal with China by promising to lower the commitment of the United States to regional security in East Asia. This possibility diminished the credibility of the commitment of the United States, but the Trump administration did not prioritize the reputation of the United States among its regional allies. This scenario is consistent with the equilibria of the model examined in this article.

Internationalists meet reformists
The model distinguishes between the internationalist type of actor and the hardliner type. According to Miyamoto Yūji -former Japanese ambassador to China -the Chinese leadership is divided into reformist internationalists (kokusai kyōchō kaikaku-ha) and conservative hardliners (taigai kyōkō hoshu-ha) over how to respond to internal issues such as maintaining social stability, as well as external issues such as responding to globalized world politics 23 The 2008 global economic crisis empowered the hardliners, and since then the hardliners and the internationalists have been in conflict over China's foreign policy. At the same time, Miyamoto suggests that the reformists on domestic policies and the internationalists on foreign policies are aligned, while the conservatives resisting domestic economic reforms are aligned with the hardliners arguing for abrasive foreign policies. Reformist internationalists argue that to maintain one-party rule China should achieve sustainable economic growth by committing to SOE reform, and that China should join FTAs requiring domestic economic reforms, which would impose a gaiatsu on China regarding politically difficult reforms. By contrast, conservative hardliners insist that agreeing to reform in the name of trade would jeopardize oneparty rule, and hence that China should adopt policies that project its power even if they cause friction while promoting nationalism.
In the game-theoretic model, empowered internationalists bring more likelihood of international cooperation. As the value of p increases, the authoritarian state is more likely to take the cooperative stance. Translating this to Chinese domestic politics over FTAs, the reformist internationalists who argue that China should commit to domestic economic reforms and cooperative foreign policy are empowered. Perhaps the most important point to note is that the effect of empowering the reformist internationalists would occur even if China is not an immediate signatory of the FTA. China is not a signatory of the CPTPP. However, the model's equilibrium outcome suggests that as long as China has the option to join, the CPTPP would still empower the reformist internationalists in debates regarding economic reforms. Hence, the potential to sign the FTA is enough to strengthen regional security. By contrast, any policy lowering the value of p would empower the conservative hardliners and undermine regional security. The failure to found the CPTPP or any other FTA requiring domestic economic reforms would continue to empower the conservative hardliners and weaken regional security.
In the 1980s, Deng Xiaoping introduced the concept of "keeping a low profile" (taoguang yanghui) as a pillar of Chinese foreign policy, proclaiming that China should take a cooperative foreign policy to introduce the market economy. He supported reformist ideas and policies even as he purged those who were sympathetic to the demand for democratization. For example, taking the opportunity to announce reformist Hu Yaobang's purge in 1987, Deng expressed his intention to break with the economic ideology championed by conservative Chen Yun 24 Deng punished Hu for being sympathetic to democratization, while he himself supported market-oriented economic reforms. In the game-theoretic model, Deng's gestures can be conceived as an increase in the value of p, empowering the internationalists vis-à-vis the hardliners. As the model demonstrates, if internationalists are empowered vis-à-vis hardliners the authoritarian state will be more likely to take the cooperative stance. Reformists like Zhao Ziyang were empowered by Deng's support for economic reforms in the 1980s but, as predicted, faced the backlash from the conservatives. Responding to the backlash, Chinese reformers and Western economists formed a coalition, making economic reform possible 25 The Tiananmen democratization movement in 1989 reminded Deng that the market economy would lead to increasing popular demands for democratization. After this democratization movement, conservative leaders launched a nationalist movement called the Patriotic Education Campaign, forging a coalition with hardliners on foreign policy issues 26 The new President Jiang Zemin was different from the reformist internationalists such as Deng, Hu, or Zhao. His commitment to reformism and internationalism was much weaker 27 He employed the cooptation strategy, encouraging former officials and former SOE managers to start businesses by using their political connections, and this strategy prevented the market economy from threatening China's one-party rule 28 In this way, the Chinese Communist Party (CCP) created the state capitalist system where the CCP would distribute the rent to various societal groups 29 Those who benefit from the state capitalist system do not want to see the SOE reform that would diminish the benefit of the revolving door built between the CCP and SOEs.
Although Hu Jintao, who succeeded Jiang, tried to curb the corruption rooted in the state capitalist system by raising the slogan of "harmonious society" (hexie shehui), he was never able to consolidate his power base to achieve this goal 30 Hu's reform floundered due to backlash from the conservative hardliners, but he did have one success. In 2006, he accused Chen Liangyu, a conservative close to Jiang, of a serious violation of disciplines (i.e., corruption), dismissed him from his position as Party Secretary of Shanghai, and suspended him from taking any other position in the CCP. In 2007, Jiang promoted Xi Jinping into Party Secretary of Shanghai as Chen's successor; Xi felt indebted to Jiang for the promotion 31 Xi had been empowered by Jiang, and as a result, Hu was unable to select his favored candidate (i.e., Li Keqiang) to be his top potential successor in the National Party Congress of 2007, but had to select Xi 32 Xi originally advocated greater market reform to diminish the role of SOEs when his administration started in 2013. Having observed that China had maintained rapid economic growth since 1978 primarily because of private firms, he had a good reason to expand the private sector and shrink the state's role in the economy, although SOEs still played major roles in the Chinese state capitalist economic system 33 However, his SOE reform has made no progress and has instead intensified the role of the CCP in the decision-making of SOEs 34 Following Xi's deviation from the SOE reform, Nicholas Lardy, who contended in 2014 that markets had driven China's economic growth, now argues that resurgent state dominance has begun to diminish the vital role of the market 25 Ibid., 28.. 26 Ibid., 234.. 27 Ibid., 257-61.. 28 Chen and Dickson, Allies of the State; Tsai, Capitalism Without Democracy.. 29 Huang, Capitalism with Chinese Characteristics; Wright, Accepting Authoritarianism.. 30 Kokubun,Miyamoto,Shū Kinpei no Chūgoku,[69][70][71]. 31 Kokubun,86.. 32 Ibid., 219.. 33 Lardy, Markets over Mao.. 34 Economy,The Third Revolution,chap. 4.. and private firms in the Chinese economy 35 Wendy Leutert shows that there are at least three obstacles facing the SOE reform: difficulty to determine the timing and method of reform, mismatched executive incentives, and the complexity of intra-firm obstacles 36 Indeed, SOEs play a central role in China's industrial policy both at home and abroad such as in the Belt and Road Initiative (BRI), and as a result companies have developed a symbiotic relationship with the state 37 In short, since the 1990s the CCP has used a cooptation strategy to maintain popular support for one-party rule. The power struggle over distributing the economic rents became especially severe after Deng's death in 1997, when China lost its last charismatic leader who had participated in the Communist Revolution. Such a severe intra-party power struggle over distributing rents among the collective leadership was a necessary consequence of maintaining one-party rule while advancing a market-oriented economy. However, the CCP has to face the dilemma that real economic reform, which would undermine the vested interests based on the collusive and corrupt rent-seeking scheme, will be necessary to achieve sustainable economic growth. The reformist internationalists and the conservative hardliners have opposing stakes in domestic economic reforms, making the idea of an FTA requiring domestic economic reforms a sensitive issue in the intra-party power struggle.
Moreover, these vested interests are rooted in patron-client relationships between businesses and bureaucrats in various departments of municipal governments that manage SOEs 38 By competing in offering businesses government funding and tax breaks, local bureaucrats have been able to attain political achievements under the cadre evaluation system, gain authority over policies, and consolidate patron-client relationships with businesses. As a result, many of these firms, which found their own bureaucratic patrons, are not motivated to upgrade and raise their productivity, and are instead competing in a "race to the bottom" for predatory pricing. Moreover, confronting the reality that China has less secured property rights protection, as do other authoritarian countries, private entrepreneurs have a strong incentive to form collusive relationships with local bureaucrats to protect their wealth against state predation 39 Consequently, since the 1990s the major source of economic growth in China has shifted from rural entrepreneurship to state capitalism, even though there has been continuous growth of private capitalism in urban areas 40 Critics of the above argument focusing on the CCP's cooptation strategy point out that the number of private firms increased rapidly in the 1990s 41 However, as Yasheng Huang argues, although private entrepreneurs in the 1980s contributed significantly to rapid development, those since the 1990s have not 42 While most studies only include the firms that are formally registered as "privately managed enterprises" (siying qiye) in the private entrepreneurs, Huang also includes small-scale unregistered individual household enterprises (getihu qiye) in that category. These small-scale household enterprises usually try 35 Lardy, The State Strikes Back.. 36 Leutert, "Challenges Ahead in China's Reform of State-Owned Enterprises.". 37 Ye, The Belt Road and Beyond.. 38 Chen, Manipulating Globalization.. 39  to avoid interactions with the state 43 In the meantime, many of the registered private entrepreneurs were successful in business by using their political ties that had been established from their previous professional experience as an SOE manager or as a government official 44 Not surprisingly, those nominally private entrepreneurs who have strong ties to the CCP do not demand democratization but support the current oneparty rule 45 Because the CCP successfully created the collusive and corrupt vested interests of the nominally private entrepreneurs, nominal privatization of the SOEs in the 1990s formed the basis of the CCP's cooptation strategy under state capitalism. Financial backing from the Chinese government allows Chinese SOEs to take extra risks and make them less interested in raising workers' productivity 46 China's outbound investment is dominated by SOEs, which take advantage of better access to credit and monopolistic power in the domestic market, and then find it difficult to compete globally 47 In short, the Chinese economy needs structural reforms such as the SOE reform to increase productivity in the long run, but Xi has prioritized political control over economic efficiency. Applying this argument to the game-theoretic model, the increasing number of nominally private firms and Xi's giving a priority to political control over economic efficiency would indicate the decrease in the value of p.

Is China a responsible stakeholder?
When Xi Jinping succeeded Hu Jintao in 2012, he inherited the collusive and corrupt vested interests under state capitalism, the problem that Hu was unable to solve. Now that China is the second largest economy in the world, whether China is a responsible stakeholder is a 64,000 USD question, and the answer depends on whether Xi is reformist internationalist or conservative hardliner. The most optimistic answer suggests that China benefits from the current U.S.-led liberal international order and hence that Xi has plenty of reasons to be reformist internationalist 48 In the 1990s, immediately after the end of the Cold War, the United States was confident in its ability to solve many, if not all, of the problems in the world. However, the 9/11 terrorist attacks in 2001 combined with the subsequent failure of the Iraq War shook the confidence of the United States. The United States would have to prioritize global issues. Later, the Barack Obama administration prioritized Asia in its famous pivot to Asia. Then US Secretary of State Hilary Clinton wrote in Foreign Policy: The background of this renewed interest in Asia is the dynamic economic development of the Asia-Pacific region and the rise of China in both the security and economic aspects 49 Thus, Thomas Finger, political scientist and former Chairman of the National Intelligence Council, argued that by incorporating China into the U.S.-led international order the pivot to Asia would help the United States maintain supremacy in post-Cold War international politics 50 In this scenario, any policy incorporating China into the U.S.-led international order, like encouraging China to join an 43 Tsai, Capitalism Without Democracy.. 44 Chen and Dickson, Allies of the State, 36.. 45 Wright, Accepting Authoritarianism.. 46 Shi, "The Political Economy of China's Outward Direct Investment.". 47 Li and Zeng,"To Join or Not to Join?". 48 For example, Lardy, Markets over Mao; Steinfeld, Playing Our Game.. 49 Clinton, "America's Pacific Century.". 50 Author's interview with Thomas Fingar, Dallas, TX, November 5 2014.. FTA -such as the CPTPP -can be interpreted as an increase in p in accordance with the game-theoretic model. Of course, no FTA will alter China's worldview or ambitions. However, proposing an FTA requiring domestic economic reforms will positively influence China's intention by empowering the reformist internationalists vis-à-vis the conservative hardliners in domestic politics. Therefore, such an FTA brings security and stability in the Asia-Pacific by making China's behavior more cooperative.
Conversely, the less optimistic answer suggests that China's goal is building a Chineseled international order, and hence China has chosen a policy of aggression by challenging the U.S.-led international order. For example, in the South China Sea, China has ignored exclusive economic zones (EEZs) originally agreed upon by the surrounding nations. Patrick Walsh, former commander of the U.S. Navy Pacific Fleet, said that he asked a Chinese military man of the legitimacy of this declaration, and he responded: "Because we claim it." 51 This answer suggests that China's goal is building a Chinese-led international order and hence China will reject being incorporated into the U.S.-led international order. The current EEZs were determined when China's resources were limited, and therefore, from China's standpoint the EEZs should be drawn taking China's own new strength into consideration.
I propose a third answer, which focuses on China's attempt to build a Chinese-led international order. This answer would be the same as the second one if conservative hardliners are empowered in Chinese domestic politics; however, it would be similar to the first one if reformist internationalists are empowered. China has established international economic institutions, such as the New Development Bank (formerly referred to as the BRICS Development Bank) and the Asian Infrastructure Investment Bank (AIIB), and led the BRI. China has also taken the initiative in the negotiation of the RCEP to compete with the Trans-Pacific Partnership (TPP). China has not participated in the CPTPP because it is not ready to commit to the required domestic economic reforms such as the SOE reform. However, an official of the Office of the United States Trade Representative (USTR) said that China was interested in joining the TPP and the official argued that the United States should keep the possibility open for China to join the TPP in the future 52 This comment suggests that the Chinese leadership is divided and that those who argue for advancing the "real economic reform" and implementing cooperative foreign policy would intend to use the CPTPP as a gaiatsu to empower the reformist internationalists vis-à-vis the conservative hardliners in Chinese domestic politics.
Today, Xi's nationalist slogans, such as the "Chinese dream" (Zhongguo meng) and the "great rejuvenation of the Chinese nation" (Zhonghua minzu weida fuxing), make other nations in the Asia-Pacific region doubt whether China will behave as a responsible stakeholder for regional stability and security. Elizabeth Economy argues that Xi gives the highest priority to tightening his grip over anything 53 In the domestic sphere, Xi tightens his grip on the control of political power and social stability by cracking down on dissidents and curbing corruption. Meanwhile, in the international sphere, he tightens his grip on the control of the international order by projecting China's power in whatever way he can. According to Economy, China projected its power by establishing the AIIB, expanding its sphere of influence in Central Asia and more broadly in Eurasia by leading the Shanghai Cooperation Organization, and trying to build a Chinese-led security system by undermining U.S. influence in the Asia-Pacific.
What implications does building a Chinese-led international order have on regional security and stability in the Asia-Pacific? The division between the reformist internationalists and the conservative hardliners suggests that they both attempt to use the formation of Chinese-led international institutions to their own advantage. On the one hand, the reformist internationalists think that administrating an international development bank like the AIIB will provide China with an opportunity to learn the know-how of loaning for infrastructure projects in developing countries. Developed countries participating in the AIIB, such as Germany and the United Kingdom, will work as a gaiatsu to implement reformist policies to refine China's market economic institutions. In the model, this would translate as an increase of p. On the other hand, the conservative hardliners think that infrastructure projects invested by the AIIB will create demands for the Chinese manufacturers and construction companies that have excessive productive capacities. China's state capitalist system has allowed many zombie companies to survive thanks to political connections under the cooptation strategy. Thus, if taken as a cornucopia of reward for zombie companies, the AIIB will empower the conservative hardliners vis-à-vis the reformist internationalists in domestic politics and make China's behavior more aggressive in international relations, which can be modeled as a decrease in the value of p.

U.S.-China relations under the Trump administration
Since Donald Trump was elected president of the United States in 2016, the United States has become the epicenter of uncertainty against liberal international order -the president uses 140-character (now 280-character) messages on Twitter to announce policy. Although the Trump administration's decision-making lacks consistency, unfortunately it is consistently anti-internationalist and anti-institutionalist. Moreover, Trump is unpredictable. I suspect that this is because his agenda is not based on strategy, but instead is based on his psychology. New York Times columnist David Brooks wrote: "It's not clear if Trump is combative because he sees the world as dangerous or if he sees the world as dangerous because it justifies his combativeness 54 Either way, Trumpism is a posture that leads to the now familiar cycle of threat perception, insult, enemy-making, aggrievement, self-pity, assault, and counterassault." While many analysts have struggled to identify a strategy behind his erratic announcements, it seems much more likely that he chooses his stance based on what will satisfy his ego. Trump is also a hard-core anti-institutionalist. Even if some of the voices in his administration who do not share his view try to explain how rulemaking through multilateral institutions benefits U.S. strategic interests, he will not listen to their advice because his ears do not respond to strategy but only to what satisfies his ego. Furthermore, he demonizes the economic regulations that make the market system function. He also attacks the political institutions such as the separation of powers and freedom of speech that make democracy work, because those institutions hurt his ego. Unfortunately, the less civil he is, the more attractive his rhetoric is to his supporters, as 54 Brooks, "A Gift for Donald Trump.". his anti-institutional attitude and lack of civility are criticized by those who, he tells his supporters, look down upon them 55 Given the political climate discussed above, it has been increasingly possible that the United States will withdraw from its commitment to regional security in East Asia in exchange for a superficial compromise from China on economic issues, such as importing more products from the United States, voluntary export restraints to the United States, or the appreciation of the renminbi to U.S. dollars. This kind of "deal" may satisfy the interests of both China and the United States at the cost of regional stability. While China will be satisfied with the decreased presence of the United States in East Asia, the United States will be satisfied with maintaining economic interactions with China. Although U.S. credibility will plummet, as mentioned before, this might not concern the Trump administration. I do not claim that this scenario is inevitable but suggest that it is consistent with the model's equilibrium. As the Trump administration's decisionmaking was increasingly influenced by anti-internationalists such as Robert Lighthizer, Peter Navarro, and Wilbur Ross, internationalist security experts such as James Mattis and H. R. McMaster were removed from office, and this scenario became more likely.

Conclusion
The overall implication of the game-theoretic model and case study presented here is that an FTA requiring domestic economic reforms can enhance regional security in international relations by empowering reformist internationalists vis-à-vis conservative hardliners in domestic politics. The model suggests that this type of FTA makes the authoritarian state behave cooperatively in international relations. This process can be seen at work in the political economy of China's foreign policy making since it started the market-oriented reform in the 1980s. The model suggests a way to make an authoritarian state to behave cooperatively and peacefully without democratization.
The model provides a theoretical underpinning for the security implications of the emerging intra-industry trade of manufacturing industries induced by the spread of GVCs. The development of GVCs in the Asia-Pacific region over the last two decades has brought a new international division of labor between developed and developing countries, which also include both authoritarian and democratic states. Thus, in the international sphere, the states in the Asia-Pacific have a strong incentive to establish internationally adopted rules by concluding FTAs to manage GVCs based intra-industry trade. Meanwhile, in the domestic sphere, governments in the Asia-Pacific have a strong incentive to use the FTAs as a gaiatsu to advance domestic reforms to achieve sustainable economic growth. Skeptics of the TPP have often overlooked the new trend of international trade in the Asia-Pacific. They assume that the mainstream of international trade is still inter-industry, and argue that the tariffs are already low, making the TPP unnecessary. However, in reality, the CPTPP is necessary economically to promote GVCs based intra-industry trade in the Asia-Pacific, and it is necessary politically to strengthen regional security by empowering reformist internationalists vis-à-vis conservative hardliners. 55 Brooks, "When Politics Becomes Your Idol.".

The Separating Equilibrium
In the separating equilibrium, p 0 ¼ 0, which means p 0 < p � . In this equilibrium, the hardliner type of authoritarian state will take an aggressive stance while the internationalist type of authoritarian state will take a cooperative stance. This is a Bayesian perfect equilibrium because it satisfies the following conditions: (1) For the hardliner type, the authoritarian state's expected payoffs are defined as follows: (1) For the internationalist type, the authoritarian state's expected payoffs are defined as follows: Therefore, the internationalist type of the authoritarian state does not have an incentive to deviate from taking the cooperative stance, except for the case of R A À T A < α < R A À W A .

The Pooling Equilibrium
In the pooling equilibrium, p 0 ¼ p. In this equilibrium, both types of the authoritarian state will take an aggressive stance. If p > p � , the democratic state will retaliate, and the authoritarian state will respond by escalating if it is hardliner or moderate internationalist and by backing down if it is pacifist internationalist. If p < p � , the democratic state will yield. This equilibrium is a Bayesian perfect equilibrium if and only if p < p � and R A À T A < α < R A À W A . Otherwise, the internationalist type has an incentive to deviate from the aggressive stance regardless of the value of β.
(1) If p > p � , the internationalist type's expected payoffs are defined as follows: Therefore, the internationalist type has an incentive to deviate from the aggressive stance if p < p � , except for the case of R A À T A < α < R A À W A .