THE LIMITS OF THE LIBERAL PEACE. Håvard Hegre International Peace Research Institute, Oslo (PRIO)

Transcription

1 THE LIMITS OF THE LIBERAL PEACE Håvard Hegre International Peace Research Institute, Oslo (PRIO) Cand.polit. thesis Department of Political Science University of Oslo June 1999

2

3 iii Table of Contents Table of Contents...iii List of Figures and Tables...iv Preface... v Chapter 1 Introduction... 1 Chapter 2 Theoretical Background Three Schools of Thought: Realism, Liberalism, and Structuralism The Liberal Peace How May Trade Lead to Peace? Realist Objections to the Liberal Peace Structuralist Objections to the Liberal Peace Levels of Analysis Summary Chapter 3 Peace through Interdependence? Peace through Interdependence Modeling Peace through Interdependence as a Deterrence Game Objection: Causality Runs from Peace to Trade, not Vice Versa Chapter 4 Development and Asymmetry Development and the Liberal Peace Asymmetrical Relationships and the Liberal Peace Asymmetric Development Chapter 5 Research Design Statistical Problems and Methods Temporal-Spatial Domain Operationalizing Interdependence Operationalizing Development Operationalizing Regime Type Operationalizing Asymmetry The Dependent Variable: Fatal Dispute Control Variables Chapter 6 Testing the Hypotheses Hypothesis 1: Peace through Interdependence Hypotheses 2 and 3: The Direction of Causation Hypotheses 4 and 5: Development Hypotheses 6 and 7: Power Asymmetry Hypotheses 8 and 9: Asymmetry in the Gains from Trade Hypothesis 10: Development Asymmetry Summary of Results Chapter 7 Conclusion Appendices Appendix 1: Results of Gravity Model Regressions Appendix 2: Descriptive Statistics References

4 iv List of Figures and Tables Figures Figure 3-1. The Prisoners Dilemma Figure 3-2. The Threat Game Figure 3-3. Subgame-perfect Equilibrium Strategies and Payoffs in the Threat Game for Various Intervals of T i Relative to b D and b A, for Fixed Values for k i and r i Figure 3-4. Expected Utility in Equilibrium to Aggressia and Deterristan, as a Function of T i, the Symmetric Case, for Fixed b A, b D, p, and k i Figure 3-5. Payoff Matrix Incorporating Relative-gains Considerations Figure 4-1. Effects of Asymmetry for Equilibrium Outcomes for Various Intervals of T i Relative to b D and b A, Relevant Areas Only Figure 6-1 Relationship between Lowest GDP per capita, Gravity Model Measure of Interdependence, and Estimated Relative Risk of Fatal Dispute, Figure 6-2 Relationship between Energy Consumption per Capita, ln(salience) Measure of Interdependence, and Relative Risk of Fatal Dispute, Figure 6-3 Estimated Baseline Hazard of Fatal Disputes, Tables Table 6-1 Test of Hypothesis 1: Estimated Effect of Interdependence on the Risk of Fatal Dispute, Table 6-2 Test of Hypothesis 1: Estimated Effect of Interdependence on the Risk of Fatal Dispute, , Alternative Specifications Table 6-3 Test of Hypotheses 4 and 5: Estimated Effect of Interdependence and Development on the Risk of Fatal Dispute, Table 6-4 Test of Hypothesis 6: Estimated Effect of Power Asymmetry on the Risk of Fatal Dispute, Table 6-5 Test of Hypothesis 8: Estimated Effect of Trade Asymmetry on the Risk of Fatal Dispute, Table 6-6 Test of Hypothesis 10: Estimated Effect of Development Asymmetry on the Risk of Fatal Dispute,

5 v Preface Early drafts of material in this thesis were presented to the 32 nd Meeting of the Peace Science Society (International), East Brunswick, NJ, October 1998, to the 7 th Norwegian National Conference in Political Science, Røros, January 1999, and to the 40 th Convention of the International Studies Association, Washington, DC, February I have benefitted greatly from generous comments from the many people who have read various versions of the manuscript. Firstly, I wish to thank my thesis advisor, Jon Hovi. His careful advice and thoughtful and constructive comments have substantially improved the presentation and argument. The thesis was written at the International Peace Research Institute, Oslo (PRIO) as part of the project on methodological issues in the study of democracy and peace, and is an off-shoot of my work on democracy and peace. I am grateful to PRIO for providing a stimulating and pleasant social environment. I am especially thankful to Nils Petter Gleditsch for his continuous support and encouragement through several years, for arranging excellent working conditions at PRIO for me, and for giving me the opportunity to present my work at international conferences. The members of the Conditions of War and Peace Program at PRIO, in particular Gleditsch, Scott Gates, and Indra de Soysa, have provided very useful feedback. Finally, I wish to thank Katherine Barbieri, Kristian S. Gleditsch, Leif Helland, Soo Yeon Kim, Michael Mousseau, John Oneal, Arvid Raknerud, Bruce Russett, Gerald Schneider, Richard Tucker, and Michael D. Ward for comments on the conference papers. Thanks to Susan Høivik for language-editing parts of the manuscript. Finally, thanks to Toril, Maria, Aurora, Gabriel, and Kaspar for tolerating the several years of studies this thesis completes, which for them implied low income and frequent absences on my part.

6 vi The process of writing this thesis has been facilitated by a generous grant from the Ryoichi Sasakawa Young Leaders Fellowship Fund, for which I am grateful. Section 5.1 draws heavily on an article I wrote together with Arvid Raknerud (Raknerud & Hegre, 1997). The analysis in Chapter 6 applies the method developed in that article. The section describing the Cox regression model was mostly Raknerud s work, although I have commented on and influenced the presentation. The rest of that article was written jointly.

7 Chapter 1 Introduction For the past fifty years, the Western states have enjoyed a continuous peace among themselves. Never before in the history of the nation-state has a large group of countries managed to keep peace for such a long period (Mueller, 1989). How can this be explained? According to the proponents of the liberal peace, the Western peace rests on the pacifying effects of democratic regimes and commercial links between the states. In this thesis, I will investigate this claim theoretically and empirically, with emphasis on the trade promotes peace aspect of the liberal peace proposition. The liberal peace argument has been attacked from several sides. Scholars of the realist school emphasize the primacy of security to economic factors, and question the liberal assumption that causation goes from trade (and democracy) to peace, and not vice versa. Other scholars, among them many of radical or structuralist leanings, question whether the pacifying effects of trade are confined to symmetrical relationships. I also look into other theoretical arguments stressing the importance of socio-economic development for the liberal peace. To see whether there are limits to the liberal peace, I study four core questions in this thesis. Firstly, I investigate the trade promotes peace hypothesis itself: Is the extent of the trade bonds between two states positively correlated with peace between them? Secondly, if there is a relationship, are trade bonds causally prior to the peace they are supposed to explain? Thirdly, can we expect pacifying effects of highly asymmetrical trade relationships? And finally, does the liberal peace require a certain amount of socio-economic development in the states that take part in it? I discuss the theoretical arguments in the literature and seek to model them in a game-theoretical model. I derive a set of hypotheses from the theoretical discussion, and survey a set of large-n quantitative studies to see how these fare when confronted with historical data. Finally, I test some of the hypotheses

8 2 The Limits of the Liberal Peace using Cox regression, a method that in several aspects improves on the method employed in comparable earlier studies. As these studies do, I employ the data on militarized interstate disputes, alliances, military capabilities, and development made available by the Correlates of War Project at the University of Michigan. Data on trade, political regime type, and GNP were taken from the IMF, the Polity Project, and Penn World Tables, respectively. Representing the arguments in a formal model has certain advantages. Modeling helps ensuring logic consistency, enforces explicitness about assumptions, enables a more concentrated exposition of an idea than is possible with verbal arguments, and eases the extension of the argument. 1 Here, I am interested in the limits of the liberal peace. Game-theoretical models have shown themselves useful for clarifying the conditions under which a hypothesis is valid. For example, Powell (1991) demonstrated that some arguments concerning relative gains (e.g., Snidal 1991a; 1991b) do not apply if a state has the opportunity to alter the rules of the game by the use of force, e.g. by eradicating the opponent as an independent actor. I use the model to investigate whether my model of the peace through interdependence hypothesis yields the same conclusions in situations of asymmetry in the gains from trade, and whether we can expect the same conclusion for any level of socio-economic development. The formal modeling allows me to extend the argument, to investigate under which conditions the liberal peace hypothesis applies. The model also sheds some light on the question of direction of causation. Harsanyi has defined game theory as the theory of rational behavior by two or more interacting rational individuals, each of them determined to maximize his own interests, whether selfish or unselfish, as specified by his own utility function (1986: 89, emphasis in original). A common criticism of gametheoretical arguments in international relations literature is that these assumptions are unrealistic. First of all, states are not individual actors, and may therefore have inconsistent and intransitive preferences due to weaknesses in the mechanisms of preference aggregation. Decisions may be the unintended consequences of group interactions and bureaucratic politics. Moreover, even if 1 This discussion of advantages and disadvantages of game theory is based on Hovi & Rasch (1993: 28 33; 1996: 74 84, 96 99) and Hollis & Smith (1991: ).

9 Chapter 1: Introduction 3 states may usefully be conceived of as unitary actors, the strategic situations they face are often so complex that it is impossible to expect them to have all the information required to act rationally. However, game theorists still defend these assumptions by pointing to the fact that in practice one single person will normally have decision-making authority when it comes to foreign policy. Furthermore, they argue that the advantages of game-theoretical modeling outweigh the disadvantages of such drastic simplifications. Moreover, even if decisions in individual situations may not be characterized as rational, rationality may still be seen as a regulative idea (Elster, quoted in Hovi & Rasch, 1996: 75). Rational models, then, predict how actors will act under ideal conditions. I formulate the hypotheses as probabilistic laws. These are tested using quantitative methods to analyze data for a large number of countries. Such hypothesis testing has several advantages relative to confronting the propositions with verbal information for a handful of case-studies. Using a large-n research design facilitates abstracting from particular instances, and avoids the accusation of selecting cases that fit the theory. With more observations of a phenomenon, we may be more certain that confirming (or disconfirming/invalidating) observations were not just coincidences. On the negative side, quantitative studies over-simplify complex phenomena. For example, in the analysis reported in Chapters 5 and 6 I will treat the militarized dispute between Greece and Turkey in 1986 as the same thing as the Iran-Iraq war they are both an instance of a fatal dispute. Still, I intend here to test general hypotheses, presumed to be valid for all countries at all times. For the present purpose, the advantages of a quantitative study outweigh the disadvantages. The thesis is organized as follows: In Chapter 2, I summarize the theoretical background for the liberal peace debate. In Chapter 3, I go more into detail with the trade through interdependence argument, and presents a formal model. I then go through the direction of causation objection to the argument, and discuss it in light of the model. The counter-arguments concerning asymmetry and development are presented and confronted with the formal model in Chapter 4. In Chapter 5, I defend my choice of statistical method, discuss the operationalization of the theoretical variables, and specify the data sources. The results of

10 4 The Limits of the Liberal Peace the analysis are reported in Chapter 6. Then, in Chapter 7, I try to connect the numerous threads. My overall conclusion is that interdependent pairs of states dyads really are more peaceful than non-interdependent pairs, but that this does not apply to dyads involving less-developed countries. For dyads of developed countries, on the other hand, trade is an important factor for peace. Furthermore, I argue that the direction of causation counter-argument applies only below a certain interdependence threshold. Finally, I find no support for the idea that the peace through interdependence hypothesis requires a symmetrical relationship.

11 Chapter 2 Theoretical Background The inquiry into the relationship between trade and conflict brings us into the heart of the debate between the three schools of thought on international relations: realism, liberalism, and structuralism Three Schools of Thought: Realism, Liberalism, and Structuralism The realist school may be traced back to Machiavelli and Hobbes some even count Thucydides among the realists. Although the term realists subsumes a wide variety of scholars, they share a set of assumptions of the conditions for international interaction (see Holsti, 1995: 36 37; Mearsheimer, 1995: 10): The most important feature of the international system is seen as the mode of organization: anarchy, or the absence of any authority above the individual, sovereign states. Without such an authority, enforcement of international laws and regulations is impossible. Given these structural conditions, the most basic interest of states becomes survival, since all other interests are dependent on the existence of the state. At the same time, all states have a potential to hurt or destroy each other there exists no means to ensure survival that cannot be used for attack. This is the security dilemma : if one state increases its security, the security of other states will decrease (see, e.g., Snyder, 1984). They, in turn, will arm to regain their relative loss, such that in the end the first state is as insecure as at the onset. States may never be sure of each other s intentions, since occupying another state is one way to increase security. War will always be a possibility in the relationship between states. For realists, non-state actors play only subordinate parts on the world stage. 2 See Viotti (1987) and Wight (1991) for a similar classification. Viotti uses the terms realism, pluralism, and globalism, whereas Wight calls the writer in these categories realists, rationalists, and revolutionists.

12 6 The Limits of the Liberal Peace Liberal theory counts numerous strands, and the tracing of its intellectual history is even more difficult than for realism. Locke and Grotius emphasize individual rights and the existence of natural law. Thus, right from the start, liberalism has challenged two realist assumptions: The nation-state is not the only important actor in international politics (cf. Holsti, 1995: 40), 3 and the question of war and peace does not dominate all other issues. According to Zacher & Matthews (1995: ), liberal thinkers share this set of assumptions: Liberals regard individual human beings as the primary international actors. Liberals view states as the most important collective actors in our present era, but they are seen as pluralistic actors whose interests and policies are determined by bargaining among groups and elections. Liberals believe that human and state interests are shaped by a wide variety of domestic and international conditions. Ultimately they are determined by bargaining power among interest groups, but these groups definition of their interests are affected by a host of factors. [Emphasis in original] Liberal economic and political theory have been closely related since the 18 th century. The liberal concern for the individual was emphasized by Adam Smith and David Ricardo in their work in economics. Kant, Paine, Bentham, James Mill, and John Stuart Mill all argued for free trade, liberty for individuals and for republican or democratic government. These ideas were linked up in the liberal opposition to mercantilism: Mercantilism saw the wealth in the world as constant. Trade, then, was a zero-sum game. Moreover, accumulating gold was seen by mercantilists as equivalent to increasing state power, since war was financed largely through the state s gold reserves and through loans. Given these assumptions, all economic and individual interests were necessarily subordinated to the pursuit of state power. The liberal opposition to the traditional political systems then automatically meant an opposition to its economic doctrine: Mercantilism was seen to arise from the nature of aristocratic states, and therefore the political priority of liberals was to topple the interventionist, powerseeking state structures that were the legacy of the eighteenth century (Buzan, 1984: 600). The assumption behind economic theories of states interests is that states will act so as to maximize the welfare of their citizens. However, many liberal 3 Due to its view on actors, this group of theories is also called pluralism (Hollis & Smith, 1991; Viotti & Kauppi, 1987).

13 Chapter 2: Theoretical Background 7 thinkers (among them Kant, 1795/1991; Keohane, 1984) also recognize structural factors: Liberal states are not merely aggregations of individual preferences, they are also systems and guarantors for such aggregation. To the extent that this very system is threatened, the main interest of a state will be survival, as realism predicts. Finally, the eclectic set of ideas labeled structuralism emphasizes the importance of global structures. Many of these, but not all, draw on the writings of Karl Marx. Structural theories typically disagree with the realist focus on states as the primary actors: Marxist strands see states only as tools for the capitalist class (although like other organizations, they may have their own dynamics, cf. Wallerstein, 1974: 402). The structure and dynamics of the entire world system are seen as more fundamental driving forces than the individual states. Structural theories conflict with liberal theories as to whether all economic exchange is mutually beneficial. The structure of the system, structuralists argue, ensures that this exchange is unequal, such that some parts of the system will exploit other parts. Marxist versions of this argument also see the exchange in itself as nothing but exploitation of the proletariat by the capitalist class. 2.2 The Liberal Peace With the liberal peace, I refer to the two ideas that states with democratic government will keep peace with each other, and that states that trade extensively will do likewise. These ideas are rooted in the liberal focus on individuals as the primary actors. To place the modern exponents into a broader historical perspective, I will look more closely into Immanuel Kant s vision of a perpetual peace based on these two mechanisms. Kant s thinking on this point was reintroduced to the field of international relations by Michael W. Doyle (1983a; 1983b; 1986), and has since been the primary philosophical reference for the democratic peace literature. 4 Then, I will define interdependence more precisely, and proceed to show how modern liberal reasoning on the relationship 4 Kant was by no means the first to forward these ideas, though. Enlightenment theorists such as Rousseau, Montesquieu, Paine, and Godwin all precede him in arguing that states founded on democratic principles must also be against war (Flessen, 1999: 13 17; Gates, Knutsen & Moses, 1996: 6 7).

14 8 The Limits of the Liberal Peace between interdependence and peace may be divided into four categories, with partly overlapping explanations for why trade should promote peace. The first two identify causal processes between two interacting states (at the dyadic level), whereas the two other concentrate on processes within the interacting states Kant s Perpetual Peace Kant s vision of a perpetual peace is found in Zum Ewigen Frieden (Kant, 1795/1991). It rests on three definitive articles of peace: First Definitive Article of a Perpetual Peace: The Civil Constitution of Every State shall be Republican. Kant defines the republican constitution as founded upon three principles: firstly, the principle of freedom for all members of a society (as men); secondly, the principle of the dependence of everyone upon a single common legislation (as subjects); and thirdly, the principle of legal equality for everyone (as citizens) (p. 99, emphasis in original). Republics are peaceful since the consent of the citizens is required to decide whether or not war is to be declared (p. 100). 6 Kant s explanation of the pacifying mechanism deserves being quoted in full: [I]t is very natural that they will have great hesitation in embarking on so dangerous an enterprise. For this would mean calling down on themselves all the miseries of war, such as doing the fighting themselves, supplying the costs of the war from their own resources, painfully making good the ensuing devastation, and, as the crowning evil, having to take upon themselves a burden of debt which will embitter peace itself and which can never be paid off on account of the constant threat of new wars. But under a constitution where the subject is not a citizen, and which is therefore not republican, it is the simplest thing in the world to go to war. For the head of state is not a fellow citizen, but the owner of the state, and a war will not force him to make the slightest sacrifice so far as his banquets, hunts, pleasure palaces and court festivals are concerned. He can thus decide on war, without any 5 See McMillan (1997) for an excellent introduction to the literature on interdependence and conflict. 6 Kant takes pain to distinguish his republic constitution from the democratic one (Kant, 1795/1991: ). This has been used to argue that Kant s peace was not democratic at all (Gates, Knutsen & Moses, 1996: 6). However, Kant s classification of regime types follows Aristotle s. Here, democracy means direct democracy. This is necessarily a despotism, because it establishes an executive power through which all the citizens may make decisions about (and indeed against) the single individual against his consent (Kant, 1795/1991: 101, emphasis in original). What is required for the perpetual peace is a constitution where the executive power is separated from the legislative power. To ensure this, the government necessarily must be representative, he argues (p. 101). This definition of a republic is not inconsistent with modern, representative democracies.

15 Chapter 2: Theoretical Background 9 significant reason, as a kind of amusement, and unconcernedly leave it to the diplomatic corps (who are always ready for such purposes) to justify the war for the sake of propriety. (p. 100) Second Definitive Article of a Perpetual Peace: The Right of Nations shall be based on a Federation of Free States: Each nation, for the sake of its own security, can and ought to demand of the others that they should enter along with it into a constitution, similar to the civil one, within which the rights of each could be secured (p. 102). With this Article, Kant appears to have in mind a mutual nonaggression pact, perhaps a collective security agreement, and the cosmopolitan law set forth in the Third Definitive Article (Doyle, 1986: 1158). Third Definitive Article of a Perpetual Peace: Cosmopolitan Right shall be limited to Conditions of Universal Hospitality. This article is fundamental to the perpetual peace since it is necessary for continents distant from each other to enter into peaceful mutual relations which may eventually be regulated by public laws, thus bringing the human race nearer and nearer to a cosmopolitan constitution (p. 102). This natural right of hospitality does not extend beyond those conditions which make it possible for them to attempt to enter into relations with the native inhabitants (p. 106, emphasis in original). Thus, Kant uses this Article to denounce the imperial powers conquest of the peoples of Africa and America. Kant argues that the perpetual peace is guaranteed by nature, and even by war itself (pp ). Nature has made it possible for people to make a living all over the world. War is nature s means of scattering the peoples of the world to everywhere on earth. Moreover, the threat of war with a neighboring people has forced each people to form itself internally into a state in order to encounter the other as an armed power (p. 112, emphasis in original). Kant further argues that republics will emerge from this state formation because the republic constitution is the only one which does complete justice to the rights of man (p. 112). Although nature separates the nations, it also unites them. And this is where the economic aspect of the liberal peace enters Kant s argument: 7 7 In fact, this quotation is the only reference to the trade promotes peace thesis in Perpetual Peace.

16 10 The Limits of the Liberal Peace On the other hand, nature also unites nations which the concept of cosmopolitan right would not have protected from violence and war, and does so by means of their mutual self-interest. For the spirit of commerce sooner or later takes hold of every people, and it cannot exist side by side with war. And of all the powers (or means) at the disposal of the power of the state, financial power can probably be relied on most. Thus states find themselves compelled to promote the noble cause of peace, though not exactly from motives of morality. And wherever in the world there is a threat of war breaking out, they will try to prevent it by mediation, just as if they had entered into a permanent league for this purpose (p. 114, emphasis in original) It is important to note that Kant s argument rests on individual self-interest, not on idealistic moral concepts: the problem of setting up a [republican] state can be solved even by a nation of devils (so long they possess understanding) (p. 112). For Kant s state of nature is a state of war, just as is Hobbes (Kant, 1975/1991: 98). Kant s idea of a democratic peace has been the focus for a large number of studies in the past fifteen years, studies with theoretical as well as empirical focus. 8 Key empirical works are Doyle (1986), Bremer (1992), and Maoz & Russett (1992; 1993). Much of the present work on the liberal peace build on the framework laid out in these articles, this thesis being no exception. Moreover, some of the same writers are now working with questions of trade and conflict (Barbieri & Bremer, 1999; Oneal et al., 1996; Oneal & Russett, 1997; 1999). Through its foundation in individual self-interest, Kant s vision of a peaceful federation of free republics is fully consistent with the Enlightenment idea of peace through trade. These ideas were well established by the time of Kant: in the early 17 th century Emeric Crucé and Hugo Grotius had argued that the interaction of enlightened and rational economic actors in different nations would stimulate mutual dependence and consequently peace (Thomassen, 1998). 2.3 How May Trade Lead to Peace? The modern literature has proposed several mechanisms for trade to promote peace. Below is a brief review of these. In the remainder of the thesis, I will concentrate on the first and the last of these mechanisms. 8 See Gleditsch (1992) and Chan (1997) for reviews of the literature.

17 Chapter 2: Theoretical Background Trade leads to interdependence, which inhibits war Modern exponents of the first mechanism have hardly changed Montesquieu s 250-year old wording: The natural effect of commerce is to bring about peace. Two nations which trade together, render themselves reciprocally dependent: if the one has an interest in buying the other has an interest in selling; and all unions are based upon mutual needs (De l esprit des lois, Book XX, ch. II, 1748, quoted in Hirschman, 1945/1980: 10). Mutual dependence acts as a form of economic deterrence. This reciprocal dependence is usually called interdependence. Interdependence, according to Keohane & Nye (1977: 8 12), is mutual dependence between states, meaning that situations and events in one state affect other states, and vice versa. Interdependence may be cultural, technological, political, or economic. The more costs and benefits the relationship entails, the more interdependent will the states be. Such relations may also have varying degrees of symmetry. If a relation between two states is entirely asymmetric, it is a relation of dependence. Moreover, Keohane & Nye (pp ) distinguish between sensitivity and vulnerability: Sensitivity involves degrees of responsiveness within a policy framework how quickly do changes in one country bring costly changes in another, and how great are the costly effects? For example, most oilimporting countries are sensitive to an oil embargo, since this will entail higher oil prices, costs of reallocation, etc. Country a s vulnerability concerns the extent to which it may counter the costs (in the long run) by political measures. Vulnerability rests on the relative availability and costliness of the alternatives that various actors face. If state a cuts off its oil exports to country b, b is more vulnerable the more costly it is to replace this oil import with domestic production or imports from other countries, or to replace oil with other sources of energy. The greater the mutual dependence, the less the risk of war. For the deterrence to work, it is required that both states are vulnerable and not only sensitive, and that the relationship is symmetrical Trade stimulates the formation of international regimes, which inhibit war Another mechanism between trade and peace is that trade leads to cooperation on mutual elimination of trade restrictions (Keohane, 1984: 75 78). Such coop-

18 12 The Limits of the Liberal Peace eration may be formalized into an international regime. According to liberal theorists, these regimes dampen conflicts in themselves. They serve as the fora for negotiations, highlight the states common interests, broaden the involved states repertoire of non-military means of force through issue-linking, and ease the inclusion of third-party mediators to conflicts. Thus, trade helps to put into practice Kant s second and third definitive articles of a perpetual peace: the federation of free states, and the conditions of unlimited hospitality (Kant, 1795/1991: ) Trade increases wealth, which paves the way for democracy, which inhibits war According to Erich Weede (1995), international trade leads to peace through changes within the states: Free trade increases the wealth of countries. Greater wealth, in turn, tends to reduce class conflict and to invite domestic compromises, and consequently leads to democracy (cf. Lipset, See also section below). Democracies, in turn, do not wage war with each other, according to the democratic peace thesis. This forms a strong causal chain, where trade primarily affects the monadic (nation) level, but reinforces a dyadic effect through wealth and democracy Territory and trade: Antithetical routes to wealth As Kant noted in the passage quoted (p. 10 above), liberals assume that trade cannot exist side by side with war. In The Great Illusion (1910; 1938), Norman Angell depicts territorial expansion and expansion through trade as contrasting objectives for nations. As formulated by Richard Rosecrance (1986), states are forced to make a choice between expanding territory or increasing trade as a basis for increasing wealth, power, and welfare. Naturally, all states are concerned with territory, since [ ] nations are themselves territorial organizations. Unchecked expansion by one state will impinge upon the territory controlled by others. Second, power, an objective of state policy, was historically defined in territorial terms. The state with the greatest land mass would have the largest population, the greatest stock of natural resources, and presumably as well the largest wealth (1986: 6 7).

19 Chapter 2: Theoretical Background 13 Consequently, making war is a means to increasing territory and wealth. One alternative way to wealth is international trade. But war and trade are antithetical routes to wealth: If national policies of economic growth depend upon an expanding world market, one country can hardly expect to rely primarily upon territorial aggression and aggrandizement. To attack one s best customers is to undermine the commercial faith and reciprocity in which exchange takes place. Thus, while the territorial and military-political means to national improvement causes inevitable conflict with other nations, the trading method is consistent with international cooperation. (Rosecrance, 1986: 13 14) This view is not a contradiction or opposition to the classical peace through interdependence hypothesis, but an extension of it: While trading states try to improve their position and their own domestic allocation of resources, they do so within a context of accepted interdependence. They recognize that the attempt to provide every service and fulfill every function of statehood on an independent and autonomous basis is extremely inefficient, and they prefer a situation which provides for specialization and division of labor among nations. One nation s attempt to improve its own access to products and resources, therefore, does not conflict with another state s attempt to do the same (p. 24). Changing their orientation from the military-political world to the trading world does not imply that trading states relieve themselves of security concerns. On the contrary, economic interdependence has to be accompanied by military interdependence: Trading states will also normally form alliances as a precaution against sudden intrusion by military-political nations (p. 24). 2.4 Realist Objections to the Liberal Peace Realists focus on the state as actor, stressing the primacy of security issues. Consequently, they question the direction of causation assumed in many expositions of the liberal argument. Moreover, since power has always been highly concentrated in a small number of nations, realists have tended to focus on these powers. Many realists therefore argues for studying these questions at the systemic level, in terms of polarity and hegemony Direction of causation Realists stress the dominance of security issues over economic issues. Not seeing this, they claim, is to violate the assumption that the international system is anarchical. As a logical consequence, then, the most important realist counter-

20 14 The Limits of the Liberal Peace arguments question the direction of causation in the liberal reasoning. Anticipating the costs of broken trade ties in wartime, a state will have an incentive to limit its trade with other states if it perceives the probability of war with them in the near future to be high. This is a classic realistic argument, found in Waltz (1986) and perhaps most explicitly in Copeland (1996). Realists argue that interdependence is a double-edged sword. If a country is dependent on resources in another country, it may be tempted to secure access to the resources by occupying the other country, thereby unilaterally solving its dependency problem (Copeland, 1996: 10; Liberman, 1996: 148; Mearsheimer, 1990: 45). A rupture of international trade may also create losses beyond the loss of the gains from trade. The economy has to readjust, it will lose productivity, and social problems may emerge from the ensuing unemployment. All in all, the country may be worse off than if the trade ties never had existed (see also Buzan, 1984: ; Hirschman, 1945/1980: 26 29). This argument is especially valid if the trade relation is asymmetrical (see section below). Another aspect of this point is the relative-gains argument. According to realists, states care more about relative gains than about absolute gains: Economic gains may be converted to military force. This is what Hirschman (1945/1980: 14) refers to as the supply effect of foreign trade: By providing a more plentiful supply of goods or by replacing goods wanted less by goods wanted more (from the power standpoint), foreign trade enhances the potential military force of a country. The security dilemma therefore dictates that states should care more about relative gains and losses than absolute gains Hegemony and the Cold War Another realist objection is that a liberal zone of peace requires a hegemon to blossom. According to hegemonic stability theory (see Keohane, 1984), order in world politics is typically created by a single dominant power a hegemon. Without this hegemon, the order will collapse. The liberal peace requires adhesion to the rules of a liberal international economic power. The only way this can be enforced, the argument goes, is through an economically and militarily superior power.

21 Chapter 2: Theoretical Background 15 Hegemony is defined as preponderance of material resources Hegemonic powers must have control over raw materials, control over sources of capital, control of markets, and competitive advantages in the production of highly valued goods (Keohane, 1984: 32). Of course, the hegemon must also be willing to use its power to enforce the rules. Although the USA had superior power in the interwar years, its isolationist policies kept it from using it. Military power is crucial to the hegemonic stability theory, since economic issues may become military-security issues if they are crucial enough to basic national interests. A hegemonic power must possess enough military power to protect the international political economy it dominates from incursion by hostile adversaries (Keohane, 1984: 39). In fact, the hegemonic stability theory is the systemic variant of the argument in the previous section. Just as bilateral trade requires the expectation of stable, peaceful relations between the two states, a liberal economy requires a stable, regulated system a liberal international regime, as it is often labeled (Keohane, 1984: 49ff.) Just as domestic economic activity will be restrained if private property rights are not protected by the state, economic activity between countries will be difficult if there is no hegemon with economic and military power to enforce the rules. A liberal economy is dependent on, as a pre-existing condition, the peace and stability it is supposed to explain (Buzan 1984: 607). Related to this is the argument that the liberal peace is an artifact of the Cold War. The Western states have had high levels of trade and an unprecedented period of peace in the 50 years following World War II, but this cannot be seen independently of the fact that the same states were on the same side in the global contest with the Soviet Union (Farber & Gowa, 1995). 2.5 Structuralist Objections to the Liberal Peace The fact that the liberal peace as of today applies mostly to the Western world makes it also vulnerable to critique from structuralists. The Western countries enjoy a preponderance of military superiority and economic power. Against the rosy liberal vision of world peace and mutually beneficial relations everywhere is the argument that the powerful nations have established the structure of the

22 16 The Limits of the Liberal Peace system in order to benefit from it. Although the use of military power is seldom necessary, it is the final guarantee for maintenance of the status quo. The works of Baran, Bornschier, Cardoso, Chase-Dunn, Galtung, Prebisch, Wallerstein and the dependencia school are important modern contributions to this general class of arguments. This school also stresses the inequality in the exchange between the developed and the developing world. In an unequal exchange, trade and investment are not necessarily mutually beneficial, as implied by classic economic theory. On the contrary, trade and investment may remove capital from the South. Recently, these views have been restated more formally by economists, with Paul Krugman as the prime exponent. Some of Krugman s models (1979; 1981) refer explicitly to the dependencia theorists, and may also be seen as consistent with the ideas of Galtung (1971). 2.6 Levels of Analysis Whether interdependence affects the probability of militarized interstate conflict may be studied at three different levels: systemic, nation, or dyadic (see Gleditsch & Hegre, 1997). At the systemic level, we may ask whether changes in the level of world trade affect the calculations of individual states. Rosecrance s trading world and realist theories of hegemony are located at this level. At the nation level, we may ask whether a state s level of economic openness affects its international behavior. At the dyadic level, we may ask whether the level of trade between two specific countries affects their mutual relations. I will restrict this analysis to the dyadic level in this thesis. Although all three levels are interesting for the study of the liberal peace, the dyadic seems to provides the closest focus on the core concept: interdependence, defined as mutual dependence between two countries. 2.7 Summary The liberal argument that trade and democracy promotes peace has been contested by realists and structuralists alike. Realists question the direction of causation, and the tenability of seeing dyadic trade relationships as independent of global structures of power, defined in terms of hegemony and polarity. Struc-

23 Chapter 2: Theoretical Background 17 turalists also emphasize the role of global structures, but have the division between developed and non-developed countries in mind. Even liberal scholars (e.g., Angell and Rosecrance) stress the importance of socio-economic development for the liberal peace proposition. In the next two chapters I discuss the theoretical debate on the peace through interdependence hypothesis. The survey in this chapter of the theoretical background indicates it should be confronted with three main objections: The question of the direction of causation, the importance of socio-economic development, and the role of asymmetries within the dyad.

24 18 The Limits of the Liberal Peace

25 Chapter 3 Peace through Interdependence? The liberal peace proposition has been heavily contested and a focus for a lively debate in recent years. However, this discussion easily degenerates into an unresolvable debate between realism and liberalism (and, to a lesser degree, structuralism). The arguments have been uncompromising and seemingly founded on totally divergent assumptions. Criticizing the interdependence hypothesis very often means a complete rejection of the liberal assumptions. Many researchers have lamented this situation, 9 and have sought to build bridges between the two. One effort (perhaps the most fruitful) proceeds by means of a rational-actor formulation of the realist and liberal positions. In this chapter, I investigate the arguments for and against the liberal peace in far greater detail, focusing on the first and the last mechanisms introduced above (sections and 2.3.4). I will discuss some formal models of the relationship that have been proposed, and introduce my own game-theoretical model of the mechanism. The various positions will be formulated in terms of probabilistic hypotheses. Translating from game-theoretical results to probabilistic hypotheses is a difficult task, however, and this part of the argument is more intuitive than formally precise. 3.1 Peace through Interdependence Trade theory According to Ethier (1995: 3), there are three fundamental reasons why nations trade. The classical explanation is that nations trade to benefit from comparative 9 Let s halt the phony realist-liberal debate (or, can there be more to IR theory than mutual assured destruction?) (Snidal, 1993: 740).

26 20 The Limits of the Liberal Peace advantage. 10 In addition, more recent economic theory describes how economies of scale and imperfect competition give incentives to trade. Classic economic theory shows that the free-trade solution is an efficient allocation of the production resources of any two given countries. If one of the countries levies a tariff, this will cause the domestic relative price of imports in terms of exports to exceed the foreign relative price. This will give the domestic firm an incentive to reallocate its production to a mix which is worth less at international prices than the country is capable of producing. If the tariff does not change the international price, the production and consumption loss is always larger than the income from the tariff itself. The country s trading partner will also lose from the tariff, since decreases in the demand for good 2 and in the supply of good 1 force it to reallocate its production and consumption. The tariff causes a reduction of trade between the states. Generally, tariffs always harm the world as a whole. On the other hand, a tariff may improve the terms of trade of the levying country if that country is large enough in world markets. Thus, with a moderate tariff, the favorable terms-of-trade effect could outweigh the unfavorable costs of consumption and production (Ethier, 1995: 222). Such a tariff one that maximizes this gain is called an optimum tariff Polachek s model Solomon Polachek s (1980) model of the relationship between trade and conflict is based on the standard trade model. He assumes that conflict makes trade more difficult through retaliatory tariffs, embargoes, and other barriers to trade. Conflict thus reduces trade, and the implicit price of being hostile is the diminution of welfare associated with potential trade losses (Polachek, 1980: 60). In his mathematical model (pp ), conflict is related to trade through inclusion of conflict in the welfare function w=w(c,z), where w is welfare, c is consumption, and z is net conflict, and through the assumption that conflict in- 10 A state has a comparative advantage over another in good 1 relative to good 2 if, in autarky, the relative price of good 1 in terms of good 2 would be lower in one state than in the other. Under free trade, each state will increase the production of the good in which it has a comparative advantage at the expense of the other to the point where the relative prices are equalized, and both states would be better off than under autarky.

27 Chapter 3: Peace through Interdependence? 21 creases import prices and decreases export prices. Any increase in welfare with increasing conflict is offset by the increased economic costs associated with the conflict. The model also shows that the marginal utility of conflict decreases with increasing trade, just as classic interdependence theory claims Cooperation in repeated Prisoners Dilemma games The previous section demonstrated that in some cases countries may have an incentive to impose a tariff. Given this, the other party may prefer to retaliate with a counter-tariff. Both states will benefit, however, if they cooperate on eliminating the tariffs. Such a situation may be modeled as a Prisoners Dilemma game. Also the security dilemma may be formulated as a Prisoners Dilemma. If a country increases its military spending, its rival will feel compelled to do the same to re-establish a balance. Both countries will then be worse off: The level of security remains unchanged, but military spending is higher. The security dilemma is also relevant for Rosecrance s depiction of a choice between a trading world and a military-political world: In order to rely on trade as an alternative to military means for increasing welfare, a state needs to be confident that the trade flow will not be cut off. All in all, this makes the work on cooperation in iterated Prisoners Dilemma games highly relevant for the liberal peace. Figure 3-1. The Prisoners Dilemma Cooperate (No tariff) Defect (Impose tariff) Cooperate M F (No tariff) M U Defect U P (Impose tariff) F P M denotes mutual cooperation, U unilateral cooperation, F free-riding, and P no cooperation. The game is a Prisoners Dilemma if F > M > P > U. (Adapted from Snidal, 1991b: 706) The equilibrium outcome of the one-shot Prisoners Dilemma is mutual defection: that both sides impose a tariff is dominant strategy for both players. This outcome is Pareto sub-optimal: Both players would be better off if they had cooperated. Cooperation seems impossible in such if a body that can enforce cooperation is lacking. However, Shubik (1970), Taylor (1976), and Axelrod (1981; 1984) have shown that cooperation may be an equilibrium outcome if the