BIPOLARITY,MULTIPOLARITY, AND FREE TRADE JOANNE GOWA University of Pennsylvania Iecent literature typically attributes the relative scarcity of open international markets to the prisoner's dilemma structure of state preferences with respect to trade.I argue that the prisoner's dilemma representation does not reflect the most critical aspect of free trade agreements in an anarchic international system:security externalities.Explicit consideration of these effects suggests that a bipolar international political system has an advantage relative to its multipolar counterpart with respect to the opening of markets among states.Less credible exit threats and stronger incentives to engage in altruism within its alliances explain the advantage of a two-power system. Lhe real income and illustrative rather than systematic em- gains that motivate free trade are also the pirical referents are used.These restric- source of the security externalities that tions are appropriate given my purpose: impede it.Their ability to internalize to demonstrate analytically the advantage these effects makes military alliances the of a bipolar international political system natural basis of agreements to open inter- with respect to free trade. national markets.The evolutionary pros- pects of alliances vary,however:ones that are the products of bipolar systems are more likely to evolve into free trade Open Markets,Prisoner's coalitions than are their multipolar Dilemmas,and State Power counterparts.I argue that less credible exit threats and stronger incentives to engage The earliest variant of hegemonic sta- in alliance altruism explain the advantage bility theory bases its representation of in- of a two-power system. ternational trade as a Prisoner's Dilemma Several important limits of the argu- (PD)on the public good literature ment should be made clear immediately.I (Kindleberger 1973).It claims that a stable argue at the systemic level:I attempt to system of international free trade involves isolate the political and economic incen- the supply of a public good.Such goods tives for free trade that occur at the level are joint in supply and nonexcludable of the international system.I do not, that is,(1)any individual's consumption therefore,consider the impact of unit- of these goods does not preclude their level factors on the pursuit of these incen- consumption by others;and (2)no indi- tives.For example,neither the role of vidual can be excluded or prevented from special interest groups nor the organiza- consuming such goods,whether he or she tion of domestic exchange via hierarchies has contributed to their production or instead of markets is considered (Doyle not.As a result,the preferences of each 1986);Frieden 1988;Ruggie 1982).The member of a large,or of any short-lived, conditions I assume are those of standard group facing a public good problem con- international trade theory (Ethier 1983), form to a PD.The corresponding payoff AMERICAN POLITICAL SCIENCE REVIEW VOLUME 83 NO.4 DECEMBER 1989BIPOLARITY, MULTIPOLARITY, AND FREE TRADE TOANNE GOWA Recent literature typically attributes the reLaioe scarcity of open international markets to the prisoner's dilemma structure of state preferences with respect to trade. I argue that the prisoner2 dilemma representation does not reflect the most critical aspect of free trade agreements in an anarchic international system: security externalities. Explicit consideration of these effects suggests that a bipolar international political system has an advantage relative to its multipolar counterpart with respect to the opening of markets among states. Less credible exit threats and stronger incentives to engage in altruism within its alliances explain the advantage of a two-power system. The real income gains that motivate free trade are also the source of the security externalities that impede it. Their ability to internalize these effects makes military alliances the natural basis of agreements to open inter￾national markets. The evolutionary pros￾pects of alliances vary, however: ones that are the products of bipolar systems are more likely to evolve into free trade coalitions than are their multipolar counterparts. I argue that less credible exit threats and stronger incentives to engage in alliance altruism explain the advantage of a two-power system. Several important limits of the argu￾ment should be made clear immediately. I argue at the systemic level: I attempt to isolate the political and economic incen￾tives for free trade that occur at the level of the international system. I do not, therefore, consider the impact of unit￾level factors on the pursuit of these incen￾tives. For example, neither the role of special interest groups nor the organiza￾tion of domestic exchange via hierarchies instead of markets is considered (Doyle 1986); Frieden 1988; Ruggie 1982). The conditions I assume are those of standard international trade theory (Ethier '1983), and illustrative rather than systematic em￾pirical referents are used. These restric￾tions are appropriate given my purpose: to demonstrate analytically the advantage of a bipolar international political system with respect to free trade. Open Markets, Prisoner's Dilemmas, and State Power The earliest variant of hegemonic sta￾bility theory bases its representation of in￾ternational trade as a Prisoner's Dilemma (PD) on the public good literature (Kindleberger 1973). It claims that a stable system of international free trade involves the supply of a public good. Such goods are joint in supply and nonexcludable, that is, (1) any individual's consumption of these goods does not preclude their consumption by others; and (2) no indi￾vidual can be excluded or prevented from consuming such goods, whether he or she has contributed to their production or not. As a result, the preferences of each member of a large, or of any short-lived, group facing a public good problem con￾form to a PD. The corresponding payoff AMERICAN POLITICAL SCIENCE REVIEW VOLUME 83 NO. 4 DECEMBER 1989