The Capitalist Peace Erik Gartzke Columbia University It is widely accepted that democracies are less conflict prone,if only with other democracies.Debate persists,however,about the causes underlying liberal peace.This article offers a contrarian account based on liberal political economy.Economic development,free markets,and similar interstate interests all anticipate a lessening of militarized disputes or wars.This "capitalist peace"also accounts for the effect commonly attributed to regime type in standard statistical tests of the democratic peace. Which Liberal Peace? First,the historic impetus to territorial expansion is tempered by the rising importance of intellectual and The discovery that democracies seldom fight each other financial capital,factors that are more expediently enticed has led,quite reasonably,to the conclusion that democ- than conquered.Land does little to increase the worth racy causes peace,at least within the community of liberal of the advanced economies while resource competition polities.Explanations abound,but a consensus account of is more cheaply pursued through markets than by means the dyadic democratic peace has been surprisingly slow to of military occupation.At the same time,development materialize.I offer a theory of liberal peace based on capi- actually increases the ability of states to project power talism and common interstate interests.Economic devel- when incompatible policy objectives exist.Development opment,capital market integration,and the compatibility affects who states fight (and what they fight over)more of foreign policy preferences supplant the effect of democ- than the overall frequency of warfare.Second,substantial racy in standard statistical tests of the democratic peace.In overlap in the foreign policy goals of developed nations in fact,after controlling for regional heterogeneity,any one the post-World War II period further limits the scope and of these three variables is sufficient to account for effects scale of conflict.Lacking territorial tensions,consensus previously attributed to regime type in standard samples about how to order the international system has allowed of wars,militarized interstate disputes(MIDs),and fatal liberal states to cooperate and to accommodate minor disputes. differences.Whether this affinity among liberal states If war is a product ofincompatible interests and failed will persist in the next century is a question open to or abortive bargaining,peace ensues when states lack dif- debate.Finally,the rise of global capital markets creates a ferences worthy of costly conflict,or when circumstances new mechanism for competition and communication for favor successful diplomacy.Realists and others argue that states that might otherwise be forced to fight.Separately, state interests are inherently incompatible,but this need these processes influence patterns of warfare in the be so only if state interests are narrowly defined or when modern world.Together,they explain the absence of war conquest promises tangible benefits.Peace can result from among states in the developed world and account for the at least three attributes of mature capitalist economies. dyadic observation of the democratic peace. Erik Gartzke is associate professor of political science and a member of the Saltzman Institute of War and Peace Studies.Email: gartzke@columbia.edu.Web:www.columbia.edu/"eg589/.Comments are welcome and appreciated. I thank Charles Boehmer,Bear Braumoeller,Michael Doyle,Monica Duffy Toft,Peter Furia,Kristian S.Gleditsch,Arman Grigorian, J.Joseph Hewitt,Robert Jervis,Stephanie Neuman,John Oneal,Jack Snyder,David Sobek,Kenneth Waltz,Erich Weede,and seminar participants at Notre Dame University,The Ohio State University,the Olin Institute for Strategic Studies,Harvard University,the Oslo Peace Research Institute,the University of Pittsburgh,Princeton University,the University of California,Berkeley,and Uppsala University for comments.Menzie D.Chinn,James Gwartney,and Dennis Quinn provided data.Richard Tucker shared BTSCS and DYADTSCS.An early draft of the study was presented at the Midwest Political Science Association Conference,Chicago,IL,15-28 April,2004.Any errors are my own. Additional tests ofkey variables,model specifications,and possible confounding factors appear in the appendix.A Stata"do"file replicating all aspects of the analysis is available from the author. American Journal of Political Science,Vol.51,No.1,January 2007,Pp.166-191 2007,Midwest Political Science Association ISSN0092-5853 166
The Capitalist Peace Erik Gartzke Columbia University It is widely accepted that democracies are less conflict prone, if only with other democracies. Debate persists, however, about the causes underlying liberal peace. This article offers a contrarian account based on liberal political economy. Economic development, free markets, and similar interstate interests all anticipate a lessening of militarized disputes or wars. This “capitalist peace” also accounts for the effect commonly attributed to regime type in standard statistical tests of the democratic peace. Which Liberal Peace? The discovery that democracies seldom fight each other has led, quite reasonably, to the conclusion that democracy causes peace, at least within the community of liberal polities. Explanations abound, but a consensus account of the dyadic democratic peace has been surprisingly slow to materialize. I offer a theory of liberal peace based on capitalism and common interstate interests. Economic development, capital market integration, and the compatibility of foreign policy preferences supplant the effect of democracy in standard statistical tests of the democratic peace. In fact, after controlling for regional heterogeneity, any one of these three variables is sufficient to account for effects previously attributed to regime type in standard samples of wars, militarized interstate disputes (MIDs), and fatal disputes.1 Ifwar is a product of incompatible interests and failed or abortive bargaining, peace ensues when states lack differences worthy of costly conflict, or when circumstances favor successful diplomacy. Realists and others argue that state interests are inherently incompatible, but this need be so only if state interests are narrowly defined or when conquest promises tangible benefits. Peace can result from at least three attributes of mature capitalist economies. Erik Gartzke is associate professor of political science and a member of the Saltzman Institute of War and Peace Studies. Email: gartzke@columbia.edu. Web: www.columbia.edu/˜eg589/. Comments are welcome and appreciated. I thank Charles Boehmer, Bear Braumoeller, Michael Doyle, Monica Duffy Toft, Peter Furia, Kristian S. Gleditsch, Arman Grigorian, J. Joseph Hewitt, Robert Jervis, Stephanie Neuman, John Oneal, Jack Snyder, David Sobek, Kenneth Waltz, Erich Weede, and seminar participants at Notre Dame University, The Ohio State University, the Olin Institute for Strategic Studies, Harvard University, the Oslo Peace Research Institute, the University of Pittsburgh, Princeton University, the University of California, Berkeley, and Uppsala University for comments. Menzie D. Chinn, James Gwartney, and Dennis Quinn provided data. Richard Tucker shared BTSCS and DYADTSCS. An early draft of the study was presented at the Midwest Political Science Association Conference, Chicago, IL, 15–28 April, 2004. Any errors are my own. 1Additional tests of key variables, model specifications, and possible confounding factors appear in the appendix. AStata “do” file replicating all aspects of the analysis is available from the author. First, the historic impetus to territorial expansion is tempered by the rising importance of intellectual and financial capital, factors that are more expediently enticed than conquered. Land does little to increase the worth of the advanced economies while resource competition is more cheaply pursued through markets than by means of military occupation. At the same time, development actually increases the ability of states to project power when incompatible policy objectives exist. Development affects who states fight (and what they fight over) more than the overall frequency of warfare. Second, substantial overlap in the foreign policy goals of developed nations in the post–World War II period further limits the scope and scale of conflict. Lacking territorial tensions, consensus about how to order the international system has allowed liberal states to cooperate and to accommodate minor differences. Whether this affinity among liberal states will persist in the next century is a question open to debate. Finally, the rise of global capital markets creates a new mechanism for competition and communication for states that might otherwise be forced to fight. Separately, these processes influence patterns of warfare in the modern world. Together, they explain the absence of war among states in the developed world and account for the dyadic observation of the democratic peace. American Journal of Political Science, Vol. 51, No. 1, January 2007, Pp. 166–191 C 2007, Midwest Political Science Association ISSN 0092-5853 166
CAPITALIST PEACE 167 The notion of a capitalist peace is hardly new. claims of classical liberal political economists like Mon- Montesquieu,Paine,Bastiat,Mill,Cobden,Angell,and tesquieu,Richard Cobden,and Norman Angell.As with others saw in market forces the power to end war.Unfortu- previous research,this study finds support for a liberal nately,war continued,leading many to view as overly op- peace,though the key causal variables,and some major timistic classical conceptions of liberal peace.This study policy implications,are considerably changed. can be seen as part of an effort to reexamine capitalist peace theory,revising arguments in line with contempo- rary insights much as Kantian claims were reworked in Two Traditions of Liberal Peace response to evolving evidence of a democratic peace. Existing empirical research on the democratic peace, while addressing many possible alternatives,provides an Liberal scholarship details two paths to peace,one domi- incomplete and uneven treatment of liberal economic nated by democracy,the other guided by the philosophy processes.Most democratic peace research examines trade of market economics.This article briefly reviews each tra- in goods and services but ignores capital markets and of- dition,offering a few critical comments. fers only a cursory assessment of economic development (Maoz and Russett 1992).Several studies explore the im- pact of interests,though these have largely been dismissed The Political Tradition by democratic peace advocates(Oneal and Russett 1999a; Democratic peace research most often attributes its Russett and Oneal 2001).These omissions or oversights intellectual genesis to Kant's essay Perpetual Peace, help to determine the democratic peace result and thus though scholars like Abbe de Saint-Pierre,Rousseau,and shape subsequent research,thinking,and policy on the Bentham all provided similar arguments prior to Kant. subject of liberal peace.This study offers evidence that Early twentieth-century scholar-statesmen like Woodrow liberal economic processes do in fact lead to peace,even Wilson and Nicolas Murray Butler advanced the pacific ef- accounting for the well-documented role of liberal pol- fects of democracy in their writings,and to a lesser extent itics.Democracy cohabitates with peace.It does not,by in practice.After a cold war hiatus,contemporary politi- itself,lead nations to be less conflict prone,not even to- cians like Bill Clinton and George W.Bush have again ward other democracies. picked up the banner of liberal peace in an era of U.S. The argument and evidence provided here are bound hegemony. to draw criticism.Skepticism in the face of controversial Early statistical work questioned the liberal convic- claims is natural,reasonable,even essential for the cumu- tion that democracies are generally less warlike(cf.Wright lation of knowledge.The democratic peace observation 1942).Babst(1964,1972)was the first to identify the spe- is supported by an exceptionally large and sophisticated cial dyadic observation.5 Small and Singer(1976)drew at- body of research.2 At the same time,excessive deference to tention to the topic,paradoxically by seeking to establish previous conclusions privileges conventional wisdom.3 A that Kant was wrong.Rummel(1979,1983,1985)argued willingness to doubt that which we have come to believe for a libertarian peace,incorporating,amongother things, is a hallmark of scientific inquiry.Indeed,the weight of free markets:"The more freedom that individuals have existing evidence does not directly contradict this study in a state,the less the state engages in foreign violence" as previous research has typically failed to address the (1983,27).Doyle (1983a,1983b,1997)examines three traditions of liberalism exemplified by Kant,Machiavelli, 2Empirical regularity cannot be the only reason for broad inter- est in the democratic peace.As Cederman(2003)points out,the relationship between the frequency and intensity of wars is also 4See Jacob(1974)for a compilation of essays.Ceadel notes of the period that"The argument that 'republican'regimes were necessary "lawlike"(literally a power rule).This relationship has generated for peace,...,was already a near-commonplace of Anglo-American little interest and received almost no attention since its discovery radicalism"(2000,16). by Richardson (1960). 3Accumulation is not cumulation.Replication offers a limited form s"Ultimately,the best strategy to ensure our security and to build a durable peace is to support the advance of democracy elsewhere. of robustness.As one author puts it,"Is it surprising that repeat- Democracies don't attack each other"(Clinton 1994)."Democra- edly testing the same primary independent and dependent variables cies don't go to war with each other....I've got great faith in generally produces the same results?"(Van Belle 2006,14).Jervis democracies to promote peace"(Bush 2004). (1976)offers an entertaining parable based on the writings of A.A. Milne.While out hunting"woozles,"Piglet and Winnie-the-Pooh Even proponents appear to acknowledge that democratic pacifism mistake their own tracks in the snow for those of their elusive prey. is at best a considerably weaker phenomenon than the dyadic rela- As the two frightened characters circle back on their own trail,the tionship (Benoit 1996;Chan 1984;Ray 2001;Rousseau et al.1996; "evidence"of woozles mounts.... Rummel 1996;Weede 1984)
CAPITALIST PEACE 167 The notion of a capitalist peace is hardly new. Montesquieu, Paine, Bastiat, Mill, Cobden, Angell, and others saw in market forces the power to end war. Unfortunately, war continued, leading many to view as overly optimistic classical conceptions of liberal peace. This study can be seen as part of an effort to reexamine capitalist peace theory, revising arguments in line with contemporary insights much as Kantian claims were reworked in response to evolving evidence of a democratic peace. Existing empirical research on the democratic peace, while addressing many possible alternatives, provides an incomplete and uneven treatment of liberal economic processes. Most democratic peace research examines trade in goods and services but ignores capital markets and offers only a cursory assessment of economic development (Maoz and Russett 1992). Several studies explore the impact of interests, though these have largely been dismissed by democratic peace advocates (Oneal and Russett 1999a; Russett and Oneal 2001). These omissions or oversights help to determine the democratic peace result and thus shape subsequent research, thinking, and policy on the subject of liberal peace. This study offers evidence that liberal economic processes do in fact lead to peace, even accounting for the well-documented role of liberal politics. Democracy cohabitates with peace. It does not, by itself, lead nations to be less conflict prone, not even toward other democracies. The argument and evidence provided here are bound to draw criticism. Skepticism in the face of controversial claims is natural, reasonable, even essential for the cumulation of knowledge. The democratic peace observation is supported by an exceptionally large and sophisticated body of research.2 At the same time, excessive deference to previous conclusions privileges conventional wisdom.3 A willingness to doubt that which we have come to believe is a hallmark of scientific inquiry. Indeed, the weight of existing evidence does not directly contradict this study as previous research has typically failed to address the 2Empirical regularity cannot be the only reason for broad interest in the democratic peace. As Cederman (2003) points out, the relationship between the frequency and intensity of wars is also “lawlike” (literally a power rule). This relationship has generated little interest and received almost no attention since its discovery by Richardson (1960). 3Accumulation is not cumulation. Replication offers a limited form of robustness. As one author puts it, “Is it surprising that repeatedly testing the same primary independent and dependent variables generally produces the same results?” (Van Belle 2006, 14). Jervis (1976) offers an entertaining parable based on the writings of A. A. Milne. While out hunting “woozles,” Piglet and Winnie-the-Pooh mistake their own tracks in the snow for those of their elusive prey. As the two frightened characters circle back on their own trail, the “evidence” of woozles mounts. ... claims of classical liberal political economists like Montesquieu, Richard Cobden, and Norman Angell. As with previous research, this study finds support for a liberal peace, though the key causal variables, and some major policy implications, are considerably changed. Two Traditions of Liberal Peace Liberal scholarship details two paths to peace, one dominated by democracy, the other guided by the philosophy of market economics. This article briefly reviews each tradition, offering a few critical comments. The Political Tradition Democratic peace research most often attributes its intellectual genesis to Kant’s essay Perpetual Peace, though scholars like Abbe de Saint-Pierre, Rousseau, and Bentham all provided similar arguments prior to Kant.4 Early twentieth-century scholar-statesmen like Woodrow Wilson and Nicolas Murray Butler advanced the pacific effects of democracy in their writings, and to a lesser extent in practice. After a cold war hiatus, contemporary politicians like Bill Clinton and George W. Bush have again picked up the banner of liberal peace in an era of U.S. hegemony.5 Early statistical work questioned the liberal conviction that democracies are generally less warlike (cf. Wright 1942). Babst (1964, 1972) was the first to identify the special dyadic observation.6 Small and Singer (1976) drew attention to the topic, paradoxically by seeking to establish that Kant was wrong. Rummel (1979, 1983, 1985) argued for a libertarian peace, incorporating, among other things, free markets: “The more freedom that individuals have in a state, the less the state engages in foreign violence” (1983, 27). Doyle (1983a, 1983b, 1997) examines three traditions of liberalism exemplified by Kant, Machiavelli, 4See Jacob (1974) for a compilation of essays. Ceadel notes of the period that “The argument that ‘republican’ regimes were necessary for peace, . . ., was already a near-commonplace of Anglo-American radicalism” (2000, 16). 5“Ultimately, the best strategy to ensure our security and to build a durable peace is to support the advance of democracy elsewhere. Democracies don’t attack each other” (Clinton 1994). “Democracies don’t go to war with each other. . . . I’ve got great faith in democracies to promote peace” (Bush 2004). 6Even proponents appear to acknowledge that democratic pacifism is at best a considerably weaker phenomenon than the dyadic relationship (Benoit 1996; Chan 1984; Ray 2001; Rousseau et al. 1996; Rummel 1996; Weede 1984)
168 ERIK GARTZKE and Schumpeter,favoring Kantian republicanism.?Levy saw constitutional constraints as inhibiting the sovereign's (1988)characterized the democratic peace as"lawlike",s innate proclivity toward warfare.Norms explanations as- anticipating the bulk of quantitative research.Numerous sign an analogous role to informal processes said to evolve studies have now reported a negative statistical association in mature democracies(Dixon 1993,1994;Ember,Ember, between dyadic democracy and disputes,crises,or wars, and Russett 1992;Mintzand Geva 1993;Owen 1994,1997; especially in the post-World War II period(Beck,Katz, Russett 1993).12 Constructivists argue that warfare is be- and Tucker 1998;Bremer 1992,1993;Gelpi and Griesdorf coming socially unacceptable(Cederman 2001a,2001b; 1997;Gleditsch 1995;Gleditsch and Hegre 1997;Gleditsch Cederman and Rao 2001;Mueller 1989;Risse-Kappen and Ward 1997;Hensel,Goertz,and Diehl2000;Hermann 1995,1997;Wendt 1999).13 Some see the evolution of and Kegley 1995,1996;Hewitt and Wilkenfeld 1996; a common identity (Deutsch 1978;Flynn and Farrell Huth and Allee 2002,2003;Maoz and Abdoladi 1989; 1999).14 Others claim that mature democracies do not Maoz and Russett 1992,1993;Morgan and Campbell go to war with states that they perceive as democratic,but 1991;Morgan and Schwebach 1992;Oneal and Russett may fight with unrecognized democracies(Weart 1994, 1997,1999a,1999c;Oneal,Russett,and Berbaum 2003; 1998).15 Oneal et al.1996;Raknerud and Hegre 1997;Ray 1993, Constraint theories have been criticized as ad hoc and 1995;Russett 1993;Russett and Oneal 2001;Russett, deductively flawed(Bueno de Mesquita et al.1999;Gates, Oneal,and Davis 1998;Senese 1997;Van Belle 1997;Ward Knutsen,and Moses 1996;Layne 2003;Rosato 2003).Note and Gleditsch 1998;Weede 1992).10 that the statement"democracies behave differently toward Theories of democratic peace have also proliferated, each other than toward nondemocracies"characterizes, in no small part because of the difficulty in accounting in broad strokes,both the observation and many of the for the special dyadic nature of the observation.1 Ini- theories designed to explain the observation.Efforts to tial accounts focused on linking domestic liberal norms avoid a circularity between theory and evidence benefit or institutions to constraints on the use of force.Institu- most from new empirical content(Huth and Allee 2003). tional explanations assert that elements of the apparatus Work by Mousseau(2000),Hegre(2000),and Mousseau, of liberal government interfere with the exercise of mili- Hegre,and Oneal(2003),for example,restricts the do- tary foreign policy(Bueno de Mesquita and Lalman 1992; main of the democratic peace to states with advanced Maoz and Russett 1993;Russett 1993).Kant (17951957) industrialized economies.It is not obvious from existing explanations for the democratic peace why norms,insti- 7Doyle chooses Schumpeter(1950,1955)because his views differ tutions,or other factors would inhibit conflict among rich from the tradition of liberal political economy and because he "saw the interaction of capitalism and democracy as the foundation of democracies but fail to do so for poor democratic states.16 liberal pacifism"(Doyle 1986,1152).Thus,even Doyle's attempt Bueno de Mesquita et al.(1999,2003)offer an elab- at synthesis assumes away an independent effect of capitalism on orate and carefully constructed explanation in which peace. 8Waltz asserts that "theories explain laws"(1979,6),suggesting a need in international relations for more laws or fewer theories. Part of the intellectual appeal of the democratic peace is that it is 2Old democratic dyads appear about as dispute prone as newer dyads(Enterline 1998;Ward and Gleditsch 1998) something to explain. For reviews,seeChan(1993,1997),Gleditsch(1992),Maoz(1997), 13Kaysen(1990)offers a critique of Mueller's theory and an under- appreciated perspective on the liberal peace. Morgan(1993),and Ray(1997,1998,2000). Several authors claim that the democratic peace relationship ex- The argument potentially applies to autocratic regimes with a common identity (Peceny,Beer,and Sanchez-Terry 2002). isted in the nineteenth century(Gochman 1997;Oneal and Russett 1999c;Tucker 1997).Others find evidence in the Italian Renaissance 15Liberal leaders(or populations)can also intentionally downplay (Sobek 2003)or classical Greece (Weart 1994,1998).Russett and the“democraticness”of another regime(Oren 1995).“Simply be- Antholis (1992)and Russett (1993,chapter 3)identify the "frag- cause it won the votes of a desperate people is no reason to grant ile emergence"(1993,43)of norms of democratic cooperation, even the slightest scrap of legitimacy...to Hamas"(Mortimer B. though Bachteler(1997)views Delian cooperation as a product of Zukerman,U.S.News and World Report,February 13,2006,page Athenian hegemony (see also Russett 1997).Dixon,Mullenbach, 63). and Carbetta (2000)suggest that democratic peace arises in the twentieth century.Hewitt and Young(2001)date the origins of the 16Mousseau(2003)argues that capitalism creates a culture of con- democratic peace at somewhere between 1924(for wars)and 1930 tracts,which then conditions democratic peace.Strangely,his anal- (for MIDs).For additional debate about the ancient evidence,see ysis focuses on an interaction term between democracy and eco- Robinson(2001a,2001b)and Weart(2001). nomic development,rather than examining free market activity, laissez-faire policies,or the availability or enforceability of con- "The growing consensus that democracies rarely ifever fight each tracts.Trading democracies should presumably also be affected, other is not matched by any agreement as to how best to explain this since trade involves extensive contracting.However,Mousseau, strong empirical regularity(Levy 2002,359).Lipson quips about Hegre,and Oneal report that the interaction term between democ- the democratic peace that "We know it works in practice.Now we racy and trade dependence is statistically insignificant(2003,Table have to see if it works in theory!"(2005,1). 2,296)
168 ERIK GARTZKE and Schumpeter, favoring Kantian republicanism.7 Levy (1988) characterized the democratic peace as “lawlike”,8 anticipating the bulk of quantitative research.9 Numerous studies have now reported a negative statistical association between dyadic democracy and disputes, crises, or wars, especially in the post–World War II period (Beck, Katz, and Tucker 1998; Bremer 1992, 1993; Gelpi and Griesdorf 1997; Gleditsch 1995; Gleditsch and Hegre 1997; Gleditsch and Ward 1997; Hensel, Goertz, and Diehl 2000; Hermann and Kegley 1995, 1996; Hewitt and Wilkenfeld 1996; Huth and Allee 2002, 2003; Maoz and Abdoladi 1989; Maoz and Russett 1992, 1993; Morgan and Campbell 1991; Morgan and Schwebach 1992; Oneal and Russett 1997, 1999a, 1999c; Oneal, Russett, and Berbaum 2003; Oneal et al. 1996; Raknerud and Hegre 1997; Ray 1993, 1995; Russett 1993; Russett and Oneal 2001; Russett, Oneal, and Davis 1998; Senese 1997; Van Belle 1997; Ward and Gleditsch 1998; Weede 1992).10 Theories of democratic peace have also proliferated, in no small part because of the difficulty in accounting for the special dyadic nature of the observation.11 Initial accounts focused on linking domestic liberal norms or institutions to constraints on the use of force. Institutional explanations assert that elements of the apparatus of liberal government interfere with the exercise of military foreign policy (Bueno de Mesquita and Lalman 1992; Maoz and Russett 1993; Russett 1993). Kant ([1795]1957) 7Doyle chooses Schumpeter (1950, 1955) because his views differ from the tradition of liberal political economy and because he “saw the interaction of capitalism and democracy as the foundation of liberal pacifism” (Doyle 1986, 1152). Thus, even Doyle’s attempt at synthesis assumes away an independent effect of capitalism on peace. 8Waltz asserts that “theories explain laws” (1979, 6), suggesting a need in international relations for more laws or fewer theories. Part of the intellectual appeal of the democratic peace is that it is something to explain. 9For reviews, see Chan (1993, 1997), Gleditsch (1992), Maoz (1997), Morgan (1993), and Ray (1997, 1998, 2000). 10Several authors claim that the democratic peace relationship existed in the nineteenth century (Gochman 1997; Oneal and Russett 1999c; Tucker 1997). Others find evidence in the Italian Renaissance (Sobek 2003) or classical Greece (Weart 1994, 1998). Russett and Antholis (1992) and Russett (1993, chapter 3) identify the “fragile emergence” (1993, 43) of norms of democratic cooperation, though Bachteler (1997) views Delian cooperation as a product of Athenian hegemony (see also Russett 1997). Dixon, Mullenbach, and Carbetta (2000) suggest that democratic peace arises in the twentieth century. Hewitt and Young (2001) date the origins of the democratic peace at somewhere between 1924 (for wars) and 1930 (for MIDs). For additional debate about the ancient evidence, see Robinson (2001a, 2001b) and Weart (2001). 11“The growing consensus that democracies rarely if ever fight each other is not matched by any agreement as to how best to explain this strong empirical regularity” (Levy 2002, 359). Lipson quips about the democratic peace that “We know it works in practice. Now we have to see if it works in theory!” (2005, 1). saw constitutional constraints as inhibiting the sovereign’s innate proclivity toward warfare. Norms explanations assign an analogous role to informal processes said to evolve in mature democracies (Dixon 1993, 1994; Ember, Ember, and Russett 1992; Mintz and Geva 1993; Owen 1994, 1997; Russett 1993).12 Constructivists argue that warfare is becoming socially unacceptable (Cederman 2001a, 2001b; Cederman and Rao 2001; Mueller 1989; Risse-Kappen 1995, 1997; Wendt 1999).13 Some see the evolution of a common identity (Deutsch 1978; Flynn and Farrell 1999).14 Others claim that mature democracies do not go to war with states that they perceive as democratic, but may fight with unrecognized democracies (Weart 1994, 1998).15 Constraint theories have been criticized as ad hoc and deductively flawed (Bueno de Mesquita et al. 1999; Gates, Knutsen, and Moses 1996; Layne 2003; Rosato 2003). Note that the statement “democracies behave differently toward each other than toward nondemocracies” characterizes, in broad strokes, both the observation and many of the theories designed to explain the observation. Efforts to avoid a circularity between theory and evidence benefit most from new empirical content (Huth and Allee 2003). Work by Mousseau (2000), Hegre (2000), and Mousseau, Hegre, and Oneal (2003), for example, restricts the domain of the democratic peace to states with advanced industrialized economies. It is not obvious from existing explanations for the democratic peace why norms, institutions, or other factors would inhibit conflict among rich democracies but fail to do so for poor democratic states.16 Bueno de Mesquita et al. (1999, 2003) offer an elaborate and carefully constructed explanation in which 12Old democratic dyads appear about as dispute prone as newer dyads (Enterline 1998; Ward and Gleditsch 1998). 13Kaysen (1990) offers a critique of Mueller’s theory and an underappreciated perspective on the liberal peace. 14The argument potentially applies to autocratic regimes with a common identity (Peceny, Beer, and Sanchez-Terry 2002). 15Liberal leaders (or populations) can also intentionally downplay the “democraticness” of another regime (Oren 1995). “Simply because it won the votes of a desperate people is no reason to grant even the slightest scrap of legitimacy ... to Hamas” (Mortimer B. Zukerman, U.S. News and World Report, February 13, 2006, page 63). 16Mousseau (2003) argues that capitalism creates a culture of contracts, which then conditions democratic peace. Strangely, his analysis focuses on an interaction term between democracy and economic development, rather than examining free market activity, laissez-faire policies, or the availability or enforceability of contracts. Trading democracies should presumably also be affected, since trade involves extensive contracting. However, Mousseau, Hegre, and Oneal report that the interaction term between democracy and trade dependence is statistically insignificant (2003, Table 2, 296)
CAPITALIST PEACE 169 leaders intent on remaining in office seek to retain the Which explanations for the democratic peace are support of a winning coalition drawn from the group of most nearly"right?"Givenso manyaccounts,it is a safe bet politically relevant citizens (the selectorate).Size matters that at least some theories must be in error.Still,whittling in this theory,as leaders in societies with small winning away at the panoply of plausible arguments has proven coalitions (autocracies)can efficiently target benefits to difficult.Textbook social science begins with deductive key supporters,while leaders facing big winning coali- theories,implications of which are then tested empiri- tions(democracies)are better off providing public goods cally.The evolution of democratic peace research has been to stay in power.Democracies fight harder because lead- messier,with the bulk of explanations coming in response ers with large winning coalitions are more likely to be to the observation.Several authors seek to critically eval- replaced in the event of defeat.Two democracies,taken uate democratic peace theory (Gates,Knutsen,and Moses together,promise particularly expensive contests,leading 1996;Layne 1994;Rosato 2003),but culling theories de- democrats more often to prefer negotiated settlements.17 ductively is problematic.A poorly crafted argument could Comparisons of the costs or risks of war often tell us still be fundamentally sound.Conversely,explanations more about who gets what than about whether force is that are seen to be flawed are often revised,escaping in- needed to get there.Since disputes typically end in some tellectual death.Other theories might be deductively co- division of the stakes,the democratic peace can be de- herent,but remain products,rather than predictors,of the scribed in terms of the timing of bargains,rather than observation.It is not clear,for example,how to reconcile about tactics.Several authors view democracies as more the persuasive theory of Bueno de Mesquita et al.(2003) transparent(Mitchell 1998;Small 1996;Van Belle 1997).18 about large selectorates,with the persuasive evidence of Others argue that"audience costs"or opposition groups Mousseau,Hegre,and Oneal(2003)that only some large allow democracies better to signal resolve (Fearon 1994; selectorates matter.Given the malleability ofassumptions, Schultz 1998,1999;Smith 1998).Properly understood, one should be able to develop numerous logically coherent however,these explanations anticipate monadic demo- explanations for almost any given empirical relationship. cratic pacifism,not the dyadic democratic peace relation- Indeed,given the flexibility of assumptions,and the num- ship.Contests should be less likely in all dyads possess- ber of bright minds engaged,it is rather a riddle that the ing at least one democratic state,regardless of the regime democratic peace has yet to produce numerous logically type of the dyadic partner.19 As Schultzacknowledges,do- coherent theoretical explanations. mestic signaling arguments"...are fundamentally claims Statistical testing is necessary to substantiate any the- about democratic states,rather than democratic dyads" oretical claim that is at variance with the established (1999,243).20 democratic peace observation.21 This same evidence is also sufficient to challenge existing theories of the demo- 17To my knowledge,Bueno de Mesquita et al.(2003)are unique cratic peace.Rather than seek to show that every con- in providing a formal logic consistent with the democratic peace ceivable attribute of democracy cannot possibly influence observation.To achieve this result,however,the authors appear to the propensity toward interstate violence-an impossible essentially turn Kant's original intuition on its head.Where Kant argues that war"does not require of the ruler...the least sacrifice" task-this article focuses on comparing democratic and and is“a poor game”for citizens“decreeing for themselves all the capitalist variants of the liberal peace empirically. calamities of war"including,among other things,"having to pay the costs of war from their own resources"([1795]1957,11),Bueno de Mesquita et al.assume elites bear the burden of improving the The Economic Tradition war effort:"By trying harder,B's leader reduces the amount of resources available to reward her supporters through private goods" What else but democracy could account for liberal (2003,232).Where Kant sees sovereigns and their key followers peace?One answer might be capitalism.The association as the principal beneficiaries of victory,Bueno de Mesquita et al. treat the spoils of war as public goods:"A military victory benefits everyone in nation B"(2003,233).Kant sees war as redistributing 2While some quantitative critiques of the democratic peace chal- welfare from the population to political elites.Bueno de Mesquita lenge its statistical validity (Spiro 1994),cultural bias(Henderson et al.argue that war is costly to elites and victory profitable to 1998),or generalizability (Henderson 2002),other studies treat populations. democracy as a product of peace(James,Solberg,and Wolfson 1999;Thompson 1996).Critics also offer a variety of alterna- isFinel and Lord(1999)argue,and provide some evidence,that tives,including alliance structures(Gowa 1994,1995),the cold war transparency can lead to greater noise and confusion. (Farber and Gowa 1995;Gowa 1999),satisfaction with the global 9Rational actors with common priors,and the same information, status quo (Kacowicz 1995;Lemke and Reed 1996),and com- should have identical beliefs(Myerson 1991,67-83). mon interstate interests (Farber and Gowa 1997;Gartzke 1998, 2000).Mansfield and Snyder (1995a,1995b,2002a,2002b)and 20Domesticopposition groups can revealinformation,orpool,con- Braumoeller (1997)argue that new democracies are more war- fusing observers,depending on electoral conditions(Ramsay 2003; like,while accepting that mature republics are peaceful.Enterline Trager 2004).Increased credibility also provides new incentives for (1996)and Thompson and Tucker (1997)attempt to counter this leaders to bluff (Nalebuff 1991). argument
CAPITALIST PEACE 169 leaders intent on remaining in office seek to retain the support of a winning coalition drawn from the group of politically relevant citizens (the selectorate). Size matters in this theory, as leaders in societies with small winning coalitions (autocracies) can efficiently target benefits to key supporters, while leaders facing big winning coalitions (democracies) are better off providing public goods to stay in power. Democracies fight harder because leaders with large winning coalitions are more likely to be replaced in the event of defeat. Two democracies, taken together, promise particularly expensive contests, leading democrats more often to prefer negotiated settlements.17 Comparisons of the costs or risks of war often tell us more about who gets what than about whether force is needed to get there. Since disputes typically end in some division of the stakes, the democratic peace can be described in terms of the timing of bargains, rather than about tactics. Several authors view democracies as more transparent (Mitchell 1998; Small 1996; Van Belle 1997).18 Others argue that “audience costs” or opposition groups allow democracies better to signal resolve (Fearon 1994; Schultz 1998, 1999; Smith 1998). Properly understood, however, these explanations anticipate monadic democratic pacifism, not the dyadic democratic peace relationship. Contests should be less likely in all dyads possessing at least one democratic state, regardless of the regime type of the dyadic partner.19 As Schultz acknowledges, domestic signaling arguments “... are fundamentally claims about democratic states, rather than democratic dyads” (1999, 243).20 17To my knowledge, Bueno de Mesquita et al. (2003) are unique in providing a formal logic consistent with the democratic peace observation. To achieve this result, however, the authors appear to essentially turn Kant’s original intuition on its head. Where Kant argues that war “does not require of the ruler . . . the least sacrifice” and is “a poor game” for citizens “decreeing for themselves all the calamities of war” including, among other things, “having to pay the costs of war from their own resources” ([1795] 1957, 11), Bueno de Mesquita et al. assume elites bear the burden of improving the war effort: “By trying harder, B’s leader reduces the amount of resources available to reward her supporters through private goods” (2003, 232). Where Kant sees sovereigns and their key followers as the principal beneficiaries of victory, Bueno de Mesquita et al. treat the spoils of war as public goods: “A military victory benefits everyone in nation B” (2003, 233). Kant sees war as redistributing welfare from the population to political elites. Bueno de Mesquita et al. argue that war is costly to elites and victory profitable to populations. 18Finel and Lord (1999) argue, and provide some evidence, that transparency can lead to greater noise and confusion. 19Rational actors with common priors, and the same information, should have identical beliefs (Myerson 1991, 67–83). 20Domestic opposition groups can reveal information, or pool, confusing observers, depending on electoral conditions (Ramsay 2003; Trager 2004). Increased credibility also provides new incentives for leaders to bluff (Nalebuff 1991). Which explanations for the democratic peace are most nearly “right?” Given so many accounts, it is a safe bet that at least some theories must be in error. Still, whittling away at the panoply of plausible arguments has proven difficult. Textbook social science begins with deductive theories, implications of which are then tested empirically. The evolution of democratic peace research has been messier, with the bulk of explanations coming in response to the observation. Several authors seek to critically evaluate democratic peace theory (Gates, Knutsen, and Moses 1996; Layne 1994; Rosato 2003), but culling theories deductively is problematic. A poorly crafted argument could still be fundamentally sound. Conversely, explanations that are seen to be flawed are often revised, escaping intellectual death. Other theories might be deductively coherent, but remain products, rather than predictors, of the observation. It is not clear, for example, how to reconcile the persuasive theory of Bueno de Mesquita et al. (2003) about large selectorates, with the persuasive evidence of Mousseau, Hegre, and Oneal (2003) that only some large selectorates matter. Given the malleability of assumptions, one should be able to develop numerous logically coherent explanations for almost any given empirical relationship. Indeed, given the flexibility of assumptions, and the number of bright minds engaged, it is rather a riddle that the democratic peace has yet to produce numerous logically coherent theoretical explanations. Statistical testing is necessary to substantiate any theoretical claim that is at variance with the established democratic peace observation.21 This same evidence is also sufficient to challenge existing theories of the democratic peace. Rather than seek to show that every conceivable attribute of democracy cannot possibly influence the propensity toward interstate violence—an impossible task—this article focuses on comparing democratic and capitalist variants of the liberal peace empirically. The Economic Tradition What else but democracy could account for liberal peace? One answer might be capitalism. The association 21While some quantitative critiques of the democratic peace challenge its statistical validity (Spiro 1994), cultural bias (Henderson 1998), or generalizability (Henderson 2002), other studies treat democracy as a product of peace (James, Solberg, and Wolfson 1999; Thompson 1996). Critics also offer a variety of alternatives, including alliance structures (Gowa 1994, 1995), the cold war (Farber and Gowa 1995; Gowa 1999), satisfaction with the global status quo (Kacowicz 1995; Lemke and Reed 1996), and common interstate interests (Farber and Gowa 1997; Gartzke 1998, 2000). Mansfield and Snyder (1995a, 1995b, 2002a, 2002b) and Braumoeller (1997) argue that new democracies are more warlike, while accepting that mature republics are peaceful. Enterline (1996) and Thompson and Tucker (1997) attempt to counter this argument
170 ERIK GARTZKE between economic freedom and interstate peace has deep The chief challenge to the arguments of Angell and intellectual roots,though the liberal political economy other political economists is that they turned out to be tradition has received little attention in recent decades.22 wrong(Carr 1939;Morgenthau 1948).Two world wars Enlightenment figures like Montesquieu and Smith ar- and associated economic upheaval reversed the trend to- gued that market interests abominate war.Paine wrote ward globalization and dissolved optimism about a capi- that"commerce diminishes the spirit,both of patriotism talist peace.24 Cold war tensions ensured that scholarship and military defense"(cited in Walker 2000,59).Cob- was preoccupied with balancing and deterrence (Jervis den called trade "the grand panacea"([1867]1903,36). 1978;Richardson 1960;Snyder 1961;Waltz 1959,1979), Mill saw market forces as"rapidly rendering war obsolete" and that subsequent generations of researchers remained (1902,390).Angell argued that it had become "impossi- skepticalabout the prospects for liberal peace(Waltz 1970, ble for one nation to seize by force the wealth or trade 1999,2000).These same events led to the long hiatus of another...war,even when victorious,can no longer in democratic peace research.However,when interest in achieve those aims for which peoples strive"(1933,60). liberal peace returned,attention centered on democracy. Angell (1933)serves as a useful point of departure Kantian theory was given a thorough rewrite in an attempt in attempting to identify how capitalism contributes to to conform to the evolving evidence,while the capitalist interstate peace.Angell highlights two processes thought peace received little attention. to diminish the appeal of conquest among countries with Of the factors emphasized by liberal political modern industrial economies.First,changes in the na- economists,trade has been by far the most closely evalu- ture of production make it difficult to cheaply subdue and ated in contemporary scholarship(Bliss and Russett 1998; to profitably manage modern economies through force. Keohane and Nye 1989;Oneal and Ray 1997;Oneal et al. Industrial economies are increasingly dependent on in- 1996;Onealand Russett 1997,1999a;Polachek 1980,1997; puts that are more easily and cheaply obtained through Polachek,Robst,and Chang 1999).25 Yet,of the elements commerce than through coercion.Relating tales of Viking of global capitalism,trade is arguably the least impor- raids on the English countryside,Angell asks why,now tant in terms of mitigating warfare.Classical political that the tables have turned,he did not see "our navy economists had yet to consider the strategic nature of con- loading up a goodly part of our mercantile marine with flict(Schelling 1966).Iftrade makes one partner more pli- the agricultural and industrial wealth of the Scandina- ant,it should allow other states to become more aggressive vian peninsular"(1933,103).Governments,like individ- (Morrow 1999;Wagner 1988),so that the overall decline uals,choose between trade and theft in obtaining needed in warfare is small or nonexistent(Beck,Katz,and Tucker goods and services.Modernity made it easier to profit 1998;Gartzke,Li,and Boehmer 2001).Economic devel- from production and trade,and harder to draw wealth opment,financial markets,and monetary policy coordi- from conquered lands or confiscated loot.23 nation all arguably play a more critical role in promoting The second process Angell outlines involves eco- peace(Gartzke and Li 2003).Much of the impact of free nomic globalization.The integration of world markets markets on peace will be missed if much of what com- not only facilitates commerce,but also creates new inter- prises capitalism is omitted or ignored. ests inimical to war.Financial interdependence ensures What are the "aims for which peoples strive,"which that damage inflicted on one economy travels through Angell mentions?Much like realists,classical political the global system,afflicting even aggressors.Angell imag- economists assumed that warfare results from resource ines a Teutonic army descending on London:"the German competition.If there are other reasons why nations fight, General,while trying to sack the Bank of England,might then some wars will occur,despite the basic validity of capi- find his own balance in the Bank of Germany had van- talist peace arguments.It is then necessary to revise,rather ished,and the value of even the best of his investments than reject out of hand,economic explanations for lib- reduced"(1933,106-7).As wealth becomes less tangible, eral peace.This article next offers the outlines of a revised more mobile,distributed,and more dependent on the theory of capitalist peace. good will of investors,it also becomes more difficult to coerce (Brooks 1999;Rosecrance 1985) 24Kant himselfis pessimistic about human nature:"the natural state 2Notable exceptions include,but are not limited to,Mousseau is one of war."Citizens curb the warlike tendencies of leaders,the (2000),Tures(2004),and Weede(2003,2004,2005). same citizens that are in turn constrained by government.Kant 2The first edition appears in 1909 under the title Europe's Optical explicitly rejects the notion that individuals or nations are able to cooperate spontaneously."A state of peace,therefore,must be Illusion.Subsequent printings appeared in 1910/1912 as The Great established"(Kant [1795]1957,10). Illusion.The world wars are widely viewed as having repudiated Angell's capitalist peace thesis,along with the claims of Kant and 25See McMillan(1997)and Mansfield and Pollins(2001)for reviews Wilson (see Gartzke 2007). of the literature on interdependence and conflict
170 ERIK GARTZKE between economic freedom and interstate peace has deep intellectual roots, though the liberal political economy tradition has received little attention in recent decades.22 Enlightenment figures like Montesquieu and Smith argued that market interests abominate war. Paine wrote that “commerce diminishes the spirit, both of patriotism and military defense” (cited in Walker 2000, 59). Cobden called trade “the grand panacea” ([1867] 1903, 36). Mill saw market forces as “rapidly rendering war obsolete” (1902, 390). Angell argued that it had become “impossible for one nation to seize by force the wealth or trade of another ... war, even when victorious, can no longer achieve those aims for which peoples strive” (1933, 60). Angell (1933) serves as a useful point of departure in attempting to identify how capitalism contributes to interstate peace. Angell highlights two processes thought to diminish the appeal of conquest among countries with modern industrial economies. First, changes in the nature of production make it difficult to cheaply subdue and to profitably manage modern economies through force. Industrial economies are increasingly dependent on inputs that are more easily and cheaply obtained through commerce than through coercion. Relating tales of Viking raids on the English countryside, Angell asks why, now that the tables have turned, he did not see “our navy loading up a goodly part of our mercantile marine with the agricultural and industrial wealth of the Scandinavian peninsular” (1933, 103). Governments, like individuals, choose between trade and theft in obtaining needed goods and services. Modernity made it easier to profit from production and trade, and harder to draw wealth from conquered lands or confiscated loot.23 The second process Angell outlines involves economic globalization. The integration of world markets not only facilitates commerce, but also creates new interests inimical to war. Financial interdependence ensures that damage inflicted on one economy travels through the global system, afflicting even aggressors. Angell imagines a Teutonic army descending on London: “the German General, while trying to sack the Bank of England, might find his own balance in the Bank of Germany had vanished, and the value of even the best of his investments reduced” (1933, 106–7). As wealth becomes less tangible, more mobile, distributed, and more dependent on the good will of investors, it also becomes more difficult to coerce (Brooks 1999; Rosecrance 1985). 22Notable exceptions include, but are not limited to, Mousseau (2000), Tures (2004), and Weede (2003, 2004, 2005). 23The first edition appears in 1909 under the title Europe’s Optical Illusion. Subsequent printings appeared in 1910/1912 as The Great Illusion. The world wars are widely viewed as having repudiated Angell’s capitalist peace thesis, along with the claims of Kant and Wilson (see Gartzke 2007). The chief challenge to the arguments of Angell and other political economists is that they turned out to be wrong (Carr 1939; Morgenthau 1948). Two world wars and associated economic upheaval reversed the trend toward globalization and dissolved optimism about a capitalist peace.24 Cold war tensions ensured that scholarship was preoccupied with balancing and deterrence (Jervis 1978; Richardson 1960; Snyder 1961; Waltz 1959, 1979), and that subsequent generations of researchers remained skeptical about the prospects for liberal peace (Waltz 1970, 1999, 2000). These same events led to the long hiatus in democratic peace research. However, when interest in liberal peace returned, attention centered on democracy. Kantian theory was given a thorough rewrite in an attempt to conform to the evolving evidence, while the capitalist peace received little attention. Of the factors emphasized by liberal political economists, trade has been by far the most closely evaluated in contemporary scholarship (Bliss and Russett 1998; Keohane and Nye 1989; Oneal and Ray 1997; Oneal et al. 1996; Oneal and Russett 1997, 1999a; Polachek 1980, 1997; Polachek, Robst, and Chang 1999).25 Yet, of the elements of global capitalism, trade is arguably the least important in terms of mitigating warfare. Classical political economists had yet to consider the strategic nature of conflict (Schelling 1966). If trade makes one partner more pliant, it should allow other states to become more aggressive (Morrow 1999; Wagner 1988), so that the overall decline in warfare is small or nonexistent (Beck, Katz, and Tucker 1998; Gartzke, Li, and Boehmer 2001). Economic development, financial markets, and monetary policy coordination all arguably play a more critical role in promoting peace (Gartzke and Li 2003). Much of the impact of free markets on peace will be missed if much of what comprises capitalism is omitted or ignored. What are the “aims for which peoples strive,” which Angell mentions? Much like realists, classical political economists assumed that warfare results from resource competition. If there are other reasons why nations fight, then some wars will occur, despite the basic validity of capitalist peace arguments. It is then necessary to revise, rather than reject out of hand, economic explanations for liberal peace. This article next offers the outlines of a revised theory of capitalist peace. 24Kant himself is pessimistic about human nature: “the natural state is one of war.” Citizens curb the warlike tendencies of leaders, the same citizens that are in turn constrained by government. Kant explicitly rejects the notion that individuals or nations are able to cooperate spontaneously. “A state of peace, therefore, must be established” (Kant [1795] 1957, 10). 25See McMillan (1997) and Mansfield and Pollins (2001) for reviews of the literature on interdependence and conflict