So opened one of the Review's first explicitly analytic studies of international relations, a study concerned to predict and influence the risk of international conflict. Russell and Wright articulated the mission of what they described as the scientific study of international affairs, offering data-based evidence to demonstrate the feasibility of that undertaking. In this essay, I discuss why, after an extended exploration of relations between states and international systems, studies of international conflict have turned toward Russell and Wright's proposal. Using developments in noncooperative game theory and political economy modeling, studies of war and peace increasingly look within states at how domestic interests and institutions help shape international affairs. This evolution seems like a natural progression from macro-level accounts that linked system structure to conflict outcomes to the specification of their micro-foundations, exposing new insights and contextualizing some earlier ones in unanticipated ways.

Russell and Wright's (1933) view lay dormant for most students of international conflict until well after the empirical demonstration by Babst (1964) that democracies rarely, if ever, fight wars with each other. That insight, originally articulated by Immanuel Kant in 1795, slowly percolated through the security studies community, eventually leading to a proliferation of theories and empirical studies that look within states to sort out why different types of regimes behave differently when it comes to decisions about war and peace. No other aspect of conflict studies since the advent of nuclear deterrence theory has had as large an impact on American foreign policy as the theorizing and empirical assessments of the democratic peace. Presidents Bill Clinton and George W. Bush embraced the central insight from this literature, with Bush making the promotion of democracy a theme of his foreign policy. Therefore, understanding the shift in research strategies that led to the insights of the democratic peace is important from both a theoretical and a practical standpoint.

When Russell and Wright (1933) wrote, Wilsonian idealism, the then dominant approach to international affairs, had begun to face intellectual and policy challenges from realists. During the decade of the 1930s, realists observed that the Russian Revolution, naval arms races both in Europe and between European powers and Japan, and regional conflicts seemed incompatible with idealist expectations. They noted that deliberation through the League of Nations was insufficient to sustain peace. By the end of World War II, structural realist approaches largely supplanted Wilsonian idealism. Thus, even as the United Nations replaced the League as a deliberative peace-keeping body, states began organizing into regional alliances for mutual defense and hard-nosed realist tactics replaced idealist expectations in the nascent cold war. Today, realist ideas, in their turn, are being modified as they face practical and theoretical challenges. These challenges reflect (1) the force of history that has uncovered the insufficiency of realist theories to anticipate or explain the implosion of the Soviet Union and the unraveling of its East European empire (Gaddis 1992); (2) the discovery that several of the propositions of structural realist theories contain internal inconsistencies when assessed from the perspective of their micro-foundations (Milner 1998; Powell 1994, 1999); and (3) the growing evidence from political economy studies that the causes of and solutions to international conflict can be better understood by looking within states (Bueno de Mesquita et al. 2003; Fearon 1994; Schultz 2001; Werner 1996).

Political economy was turned to as one means to try to overcome the limitations of realist approaches. It is not, of course, the only such alternative. Numerous other theoretical perspectives were also stimulated by the perception that realism faced limits on its ability to explain remaining puzzles concerning international conflict. Neo-liberalism, gender studies, constructivism, and computational models are among the alternative and sometimes complementary approaches explored as means to solve conundrums regarding war and peace. Like some of the alternative approaches, political economy models—the focus here—enjoy a history of success in other aspects of social inquiry. These include elucidating the democratic peace; clarifying issues in American and comparative politics, at least since the origins of the “Rochester” school in the 1960s; and providing micro-foundations for research in economics and politics about growth, the resource curse, trade, and banking policy. Political economy models, like some other modeling, also benefited from improvements in game theory that facilitated the construction of more sophisticated models than was possible prior to the late 1970s, as well as from the stimulus provided by the recognition that domestic factors contributed significantly to the demise of the USSR and the end of the cold war.

The political economy perspective treats leaders, not states, as the object of study. In doing so, it argues that relations among nations are produced by the normal pulls and tugs of domestic affairs, taking into account the domestic and international constraints under which leaders in contending states operate. That this is a significant conceptual departure from state-centric, realist accounts is exemplified by the inability of such languages as English, French, Chinese, Urdu, and perhaps all others to describe international relations without invoking the nation as the key unit of analysis.

Political economy models and empirical assessments evaluate policy choices as parts of equilibrium behavior induced by domestic institutions. They conceive of the choice of foreign policy interactions as incentive-compatible with the motives of national leaders. These leaders' motivations may not be—and often are not—compatible with their nation's interest. Domestic institutional structures (such as the competitiveness of leadership selection or the extent of government accountability and transparency) help shape the interplay of leaders, elites, and ordinary citizens, resulting in policies that create the contours of the international environment—contours that in realist approaches are taken as given rather than as factors to be explained.

By drawing attention to leaders, recent scholarship illuminates empirical regularities not previously discerned. Gaubatz (1991), Fordham (1998a, 1998b), and Smith (2004), for instance, show that war-timing by democratic leaders depends on the election cycle, electoral rules, and domestic political circumstances. Fearon (1994) and Smith (1996) suggest that democratic leaders are more constrained than autocrats to carry out the threats they make because of domestic political audience costs. Schultz (1998, 2001) shows that the existence of a domestic political opposition limits democratic foreign policy adventurism in ways not experienced by non-democrats. Bueno de Mesquita et al. (1999, 2003) develop a “selectorate” theory of politics, showing how variations in the size of a polity's political institutions, that is, its selectorate and its winning coalition, help explain the empirical regularities that make up the democratic peace.¹

They also suggest an explanation for the preparedness of democracies to fight wars of imperial and colonial expansion and the preparedness of democracies to overthrow foreign rivals. Bueno de Mesquita and Downs (2006) build on the selectorate account to explain the willingness of democratic military interveners to restrict the spread of democracy even when these interveners declare that they are motivated to promote democratic reform elsewhere.

Each of these studies shares a conviction that policy decisions are strategic, taking into account expected responses by foreign and domestic adversaries and supporters, and so use noncooperative game theory as their foundational analytic structure. Many then go on to use statistical and case study methods to probe the generality and the verisimilitude of their proposed explanations of international conflict. These studies also suggest critical empirical tests to help sort out the explanatory power of alternative perspectives.

A game-theoretic focus on strategic interaction that assumes that states are rational unitary actors shows that war, being costly, is always ex post inefficient (Fearon 1995). Political economy, game-theoretic models agree but add that war, although ex post inefficient in terms of citizen welfare, can be beneficial for leaders (Chiozza and Goemans 2004). Just consider Margaret Thatcher's poor prospects of reelection as Britain's prime minister before the Falklands/Malvinas War. Her popularity soared following the UK's victory, which may have been instrumental to her reelection in 1983. We can only conjecture on what the electoral consequences would have been for Thatcher had she—cost-effectively—bought off Argentina's generals and the Falklands' shepherds rather than fight to defend Britain's territorial claims.

The political economy approach's micro-foundations separate the interests of leaders from those of the broader populace, or the national interest. This may be its most significant evolutionary step beyond realism's focus on the state as a unitary entity. As many have noted, the policies that leaders adopt to enhance their hold on power often make their subjects worse off (Acemoglu and Robinson 2001; Bueno de Mesquita et al. 2003; McGuire and Olson 1996; Wintrobe 1990). Consider, for instance, the personally beneficial and nationally debilitating actions of North Korea's Kim Jong-il, Zimbabwe's Robert Mugabe, Iraq's Saddam Hussein, or the Philippine leader Ferdinand Marcos. Each managed to retain his position of national leadership for many more years—often decades longer—than democratic counterparts can reasonably aspire to, and yet each contributed to the impoverishment of his nation's ordinary citizens even as he and his inner circle of cronies ensured their own wealth. In such self-aggrandizing circumstances it is difficult to square their international interactions with a state-centric approach's attentiveness to the “national interest” (Bueno de Mesquita et al. 2003; Chiozza and Goemans 2003, 2004; Morrow et al. 2006). Yet such policies are understandable when one expands upon earlier macro-level theories by adding within-state micro-foundations.

When domestic institutions constrain a leader to require a broad base of support—as is true in most democracies—then private rewards are an inefficient way to retain power (Bueno de Mesquita et al. 2003; de Tocqueville 2000; Lake and Baum 2001). Democratic leaders would have to spread these rewards across so many people that each would receive too little for the benefits to influence their loyalty to the incumbent. In such a situation, it is more efficient for leaders to rely on public goods as their best means to retain office. When political institutions compel a leader to depend on many supporters, so that a bundle of public goods is the reward for retaining the incumbent, the institutions of governance induce weak loyalty to the incumbent. After all, everyone benefits from public goods, whether they support the incumbent or not. Conversely, when a leader needs backing from only a few people to stay in power, the few are expected to be loyal both because they are well rewarded with private benefits and because they face a high risk of losing those privileges if a challenger topples the incumbent regime. Thus it is that resource allocations—including provisions for national defense—are induced by domestic political institutions rather than by international compulsions, culture, or the luck of the draw in leaders.