Domestic Politics and International Relations

Peter Gourevitch. Handbook of International Relations. Editor: Walter Carlsnaes, Thomas Risse, Beth A Simmons. Sage Publication. 2002.

Did the Soviet Union and the United States struggle for fifty years because one was communist, the other capitalist? Or was it because of their situation in the world, as the two major powers of the global system, inescapably locked into a classic security dilemma? If all countries were democracies, would there be peace, as Woodrow Wilson predicted? Are democracies developing in Latin America and East Asia because of forces internal to each country, or are their political systems shaped by international forces? These questions show the centrality of domestic politics to any understanding of the strategic interactions in the world, that is, of international relations (Bueno de Mesquita, 2000; Lake and Powell, 1999).

That international relations and domestic politics interact quite profoundly no longer seems to be a controversial statement (Kahler, 2001; Keohane and Milner, 1996; Müller and Risse-Kappen, 1993; Zürn, 1993). Appreciation of domestic politics in understanding international relations has certainly risen over the past two decades. Putnam’s (1988) metaphor of two-level games is widely cited to inspire and legitimate work on domestic politics. A very substantial literature examines the role of interest groups, domestic institutions, ideas, nongovernmental organizations, civil society transnational relations and ‘the second image reversed’ (Gourevitch, 1978), a phrase used by this author to express the impact of international pressures on domestic politics, and the consequences that has back on international relations.

For many decades discourse has been structured around the notion of ‘levels,’ or, in Waltz’s (1959) language, ‘images’–system (third image, characterized by anarchy), state (second image, characterized by institutions and socio-economic structures) and individual (first image, characterized by individual psychology). These distinctions no longer capture intellectual life in our field, as indeed they have for many years not captured reality. Anarchy and institutionalization are important distinctions but they are endpoints on a continuum and they do not differentiate domestic from international politics (Lake, 1998). Various aspects of international life are highly institutionalized–the European Union, the postal and telephone regime, the rules of diplomatic exchange. Many domestic situations are anarchic–gangs in urban slums, warlords in countries torn by civil strife. International and domestic are both ‘politics.’ They can be understood by the same categories and concepts. The real question is not whether the two ‘levels’ are distinct, but how to study their unmistakable interaction.

Seeing the domestic and international dimensions as part of a whole does not ease the challenge of understanding. Indeed, it may make it harder, for it raises difficult problems of endogeneity. That issue has been handled in three ways. First, system-level theorizing holds domestic politics constant and explores variance in the international arena. This approach flattens the role of domestic politics to zero in order to see whether changes in the environment within which states operate alters their behavior. This is, of course, the central premise of realism: assuming a unitary, rational state in order to examine variance within the international system. It is as well the central premise of any system approach, including neoliberal institutionalism (NLI) and system constructivism, which retain assumptions of unitary actors while challenging realism’s account of their interactions. NLI (Keohane, 1984) says states can cooperate in anarchy if there are institutions; system constructivism (Wendt, 1992) says states operate according to a code derived from the international arena. Neither includes domestic politics.

The second level of theorizing holds the system relatively constant and looks at the aspects of domestic politics which shape how a country responds to its environment. The core of any argument about the importance of domestic politics lies with the degree of freedom countries have in how they interpret their situation in the world. The system may indeed have ‘imperatives’ but these are rarely wholly determinative of a country’s choices. The system has many dimensions, with multiple, often conflicting incentives, confusing signals, complex information. Disagreement within countries over policy leads to a politics of choice among alternatives.

If nations have choices, we need theories and research that explains how countries make these choices. An important branch of domestic politics theorizing looks within the country to find factors that shape its choices. It reverses system theorizing: instead of seeing how the system induces behavior in its units, it looks at aspects within the units that shape its behavior outward. This is the classic version of second-image reasoning.

A third level of analysis about domestic politics seeks to explore interactions: of levels (system and unit), of countries with each other and of transnational forces. This approach stresses strategic interaction of all elements. Countries evaluate each other’s domestic political situation; actors derive their preferences from a pattern of influences involving other actors and the internal politics of other countries. Institutions, interests and ideas within countries are constantly influencing the same factors in other countries. This third body of work, the interaction of domestic and international, is the least well developed, and the place that particularly requires further analysis.

This chapter will focus on the second and third programs of research. The first model, focusing on system, is examined in other chapters of this volume. Since theorizing arises out of the problem to be examined, the discussion is organized around issue areas.

The Politics of Trade Disputes: Interest Groups Institutions Ideas

The most highly articulated body of work exploring domestic/international interaction has arisen from international political economy. Foreign economic policy inherently links the two by seeking the causes behind a nation’s choices. Schattschneider (1935) developed one of the most widely accepted concepts in politics–that concentrated interests defeat diffuse ones–to explain the passage of the Smoot Hawley tariff. Olson (1982) applied his concept of collective action to economic policy with the idea of ‘all-encompassing’ coalitions or institutions, those that aggregate at a high level versus those that induce fragmentation to explain why the former produce more ‘general welfare enhancing’ policies than the latter. A number of significant writings helped advance the analysis of trade relationships: Gerschenkron’s (1962) study of the iron–rye coalition in Germany, Hirschman’s (1945) analysis of the use of state power to promote economic dependencies, Kindleberger’s (1951) look at interest groups in trade, Bauer, de Sola Pool and Dexter’s (1963) study of public opinion, interest groups and institutions, Gilpin’s (1975) and Vernon’s (1971) arguments about the role of multinational corporations.

By the 1960s and 1970s, these ideas had generated a sustained attack on unitary system models. Writers in the emerging field of international political economy disaggregated the state, exploring how its internal processes explained policy. They dislocated security issues from their primacy and challenged as well the interest in ideas or national culture as explanations of foreign policy (Hoffmann, 1960). A number of works emerged to examine interest groups, institutions and the patterns of their interaction (Gourevitch, 1977, 1986; Katzenstein, 1978; Keohane and Nye, 1977; Krasner, 1978; Kurth, 1979). That literature has strongly influenced the way the field deals with the interaction among preferences (or interests), institutions and ideas; the building blocks of most debates in international relations.

Arguing from Preferences

What explains a country’s choice between free trade and protectionism? The general logic of a preferences argument is familiar to everyone: interests within a country are differentially impacted by trade according to their competitive position in the world and the national economies; they lobby the government to get the policies that fit their preferences (Frieden, 1999). The controversies arise over specifying the interests and then linking these to decision-making processes.

By the 1980s, descriptive accounts of economic cleavages gave way to deductively derived ones, drawing on economic models, and such models are now required of any serious work in this field. The fights here have turned on class versus factor as the basis of cleavages. Rogowski’s (1989) quite important book uses Stoper-Samuelson models of trade to predict that those whose factor endowment was scarce within their country would prefer protectionism, while those whose factors were abundant would support free trade. In applying this logic to historical cases, Rogowski was careful not to say that the explanation of preference cleavages was sufficient to explain the choice of country policy–but his account is frequently taken as the standard bearer of that logic. The most extensive version of the preferences approach has become known as ‘endogenous’ macro theory, for which the work of Magee, Brock and Young (1989) is among the most important statements.

The sectors argument draws on Ricardo–Viner trade theory to stress factor specificity. Frieden (1991a) argues that as some force (technology, geography) inhibits the mobility of factors of production, these become tied to a particular industry. All the participants of that industry then have an interest in its preservation and growth, and will ally despite the difference in their class position.

Hiscox (1999, 2001) argues that the two theories should be read as empirical alternatives. Both are deductively valid, but apply to different historical moments. Factor specificity is not a constant, but an attribute that can change over time. Policy preferences shift as the degree and location of factor specificity shifts. Hiscox measures changes in asset specificity and finds that changes in that parameter do correlate with patterns of class (when asset specificity is low) and sectoral conflict (when it is high).

The logic of the preference approach invites researchers to investigate other principles of cleavage, or incentives for a particular policy position, and apply them to a range of policy areas. Milner (1997) argues that the concentration of interests is not exclusive to protectionists: there are groups who benefit strongly from exports and can lead a fight against the protectionists (see also Destler and Odell, 1987). Frieden (1991b) looks at capital markets, Bernhard and Leblang (1999) at banking policy, Henning (1994) at financial institutions.

Interest groups are not the only source of pressure on politicians. Public opinion can be an important factor, directly in democratic elections, indirectly in authoritarian regimes. Understanding mass preferences of individual voters deductively requires better economic models than we have (Frieden and Martin, 2000). Inductive approaches (surveys, interviews) provide important information on economic policy concerning European economic and monetary integration (Gabel, 2000) and on trade policy (Scheve, 1999; Scheve and Slaughter, 2001).

Preferences need not be economic, or based solely on economic interest. Voters and interest groups may have ideological preferences on economic issues arising from a value system about justice and equality or about nationalism. Goldstein (1993) has applied this reasoning to US trade policy. And many issues may influence foreign policy which are not economic: ethnic attachments; irridentist claims of territory from an adjacent country; religious difference–all of these can be seen as preferences by members of society which influence the decision-making of their leaders. Some constructivist arguments can be read this way: ideational foundations of preferences which put pressure on politicians.

What unifies all preferences arguments, then, is the logic of looking to the goals of political actors, to assume that institutions arise themselves from preferences, and that how institutions work turns on who uses them.


The institutionalist critique of the preferences approach is in its general form familiar; a statement of preferences by itself does not tell us how they are aggregated into an outcome. Outcomes can be varied by altering the procedures of aggregation. Policy choices are thus a function of institutional arrangements. The major quarrels within this part of the field lie in specifying which institutions matter and how.

The analysis of reciprocal trade legislation in the United States has been particularly fruitful for these debates. The shift in American policy from protectionism to free trade in the mid-twentieth century coincided with the growth of presidential power. One branch of institutionalists suggest the one caused the other (Haggard, 1988). With smaller constituencies, Congress reflects the particularism of its electors, while the President faces a larger more diverse constituency which he must integrate with a broader appeal. Congress is vulnerable to protectionist log-rolling while the Presidency is able to articulate the larger gains from free trade. Thus the growth of Presidential power produces a shift in policy away from protectionism.

A contrary line of reasoning reverses the causality. Congress supports reciprocal trade legislation because a majority there wants it, not because the President pushes them toward it. Free traders understood that foreign countries would not negotiate with the executive if any agreements reached could be amended on the floor of Congress (Lohmann and O’Halloran, 1994). To enable international negotiations, the majority protected its members from the temptations of protectionist logrolling by imposing a self-binding rule which authorizes the President to negotiate without possibility of amendment. Thus, the growth of presidential power on trade is not the cause of a shift in power, but a reflection of a change in policy preference, by a Congress that can revoke the authorization or fail to renew it, as happened after 1994 (Bailey et al., 1997; Epstein and O’Halloran, 1999; Gilligan, 1997; O’Halloran, 1994). Evidence from other periods of history and countries shows as well that trade policies have shifted without changes in institutions, and different systems have produced similar policies (Gourevitch, 1986).

While much of the thinking about institutions arose in the exploration of the relationship between the American Congress and Presidency, these ideas have been extended to a wider ranger of countries and situations, exploring the role of legislative–executive relations, the relationship between elected leaders and their constituencies, and voting rules and party systems (Cox and McCubbins, 1993, 2001; Rogowski, 1999). Proportional representation systems were adopted in the European countries, Rogowski (1987) argues, as a way of orchestrating compromises on policy questions those countries needed to deal with foreign trade pressures. The voting system in Japan has favored the representation of producer groups who feel threatened by trade, thereby exacerbating Japan’s conflict with the United Sates (Cox, 1997; Ramseyer and Rosenbluth, 1993). Cowhey (1995) explores the way differences in the political systems of the United States and Japan shape their policy disagreements. Andrew MacIntyre (2001) shows the ways difference in political systems influences East Asian responses to the Asian financial crisis of 1997. The choice between presidential and parliamentary systems can have significant consequences for the pattern of power (Laver and Shepsle, 1994; Shugart and Carey, 1992; Shugart and Haggard, 2001). Lijphart (1999) provides an impressive foundation for further research about institutional structures by carefully specifying the arrangements in many countries on a wide variety of dimensions, and he provides some application to policy.

This literature on institutional design derives largely from the rationalist model of institutions. While these approaches are important, significant work is done in other traditions (Hall and Taylor, 1996; Weber, 1997). Historical institutionalist work seeks to provide process tracing of major events and patterns in the formation of patterns of power (Steinmo et al., 1992). A concern with norms leads to an interest in the normative and cultural foundations of cooperation and conflict (Kratochwil, 1989). Sociological approaches explore structures of social reproduction and control (Mayer, 1959; Powell and DiMaggio, 1991). Eising (2002) explores the interaction of rationalist models with notions of policy learning. Levy (2001) shows the way in which the French state is constrained in policy options by the weakness of social institutions in French society, a theme also explored by Putnam, Leonardi and Nanetti (1993) and Zeigler (1997).

Much of the debate about institutions revolves around explaining and debating structures rather than the linkage between these and policy outcomes. IPE could benefit from further integration of comparative, international and institutionalist work with public policy outcomes in economic issue areas other than trade (for example, social welfare, education, health and safety, environment), and from research that explores the interactive effects between preferences and institutions.

Problemtizing Interests and Institutions

Much of the research on preferences and institutions has sought to understand the importance of each by holding the other constant. This is a necessary research ploy, but needs also to be questioned. Where do interests come from and where do institutions come from? Rationalists tend to be take interests as given, as somehow primordial. Rationalists tend to be materialists, seeing interests as arising from some structural logic of situation, usually economic, but interests could also derive from beliefs, which shape a preference, thus an interest. Interests drive preferences, which in turn drive the creation of institutions: since everyone understands that institutional arrangements influence outcomes, everyone will work to get the institutional pattern that increases their chances of victory.

Once institutions are established, they themselves structure debates about changing institutions. In the case of reciprocal trade noted above, those who favor and oppose free trade will fight over the institution of delegation to the President, but this takes place in the framework of a stable set of rules (Congress) about how to resolve the disagreement. Where there are no institutions or weak ones (as in the formation of many international institutions such as the United Nations) there are no rules to govern an argument about them, so they are shaped by non-institutional processes–military power, bargaining, ideology etc. (Gourevitch, 1999).

Preferences and institutions cannot be modeled independent of politics. Either can change if the opportunity to do so exists. Verdier (1994) argues that in competitive capitalist democracies, preferences are shaped by the politically defined probability of success or failure in attaining policy goals, and not by a purely economic logic. In choosing among options, political actors consider not only what they want but the costs of getting each option and the probability of getting it. Economic position, the presumed driver of preferences in trade issues, cannot be ascertained outside of property relations and these in turn derive from politics: the definition of property, the rules of competition, the structure of markets can all be seen as derived from political authority, as politics can change them. Since institutions shape the probability of policy outcomes, and institutions can be changed by politics, then a shift in institutions can lead to a change in desired options.

This leads to a difficult analytic point. If we observe a shift in policy position, do we classify it as a change of preferences? It could be understood instead as a change not in core preferences (the desire to maximize profit, for example) but rather a change in tactics (modifying a regulation in order to get the profits). Altering institutions shifts power relationships and thus alters incentives, provides new information and a number of other things which can change policy position. This is different from a change in preference deriving from a change in identity. Operationalizing the difference is a challenge to constructivists and materialist rationalists alike (Eising, 2002; Frieden, 1999). If a worker shifts from free trade to protectionism because of decisions made by the European Union, this will most likely be agreed to be a change of tactics. If an individual who expressed him- or herself politically as a worker now does so as a Catalan, is this a change in identity or in tactics (the Catalan parties can provide more)?

The core of a constructivist critique of interest lies in the notion of constituted identity. Prior to interests and institutions lies the formation of an identity and a framework of discourse. Debates about interests, institutions and tactics can only take place in a framework of shared assumptions, understanding and ‘common language.’ These are the primordial elements, socially constructed as a cultural force, not a hard material fact. Other chapters of this volume explore constructivism. The relevant point here relates to research strategy concerning domestic politics. In challenging the rationalist versions of the interest and preferences approaches, the constructivists seek more work on the formation of ‘interest’ identities. To sharpen the debate, alternative accounts of classic political economy dependent variables are needed.

Globalization and Domestic Politics: The Convergence–Divergence Debate

Globalization has made international relations important to scholars of comparative politics, and issues about the internal workings of countries relevant to IR scholars. That trade influences countries is not a new idea. Wallerstein (1974) and Anderson (1974) take it back to the emergence of Holland in the sixteenth century. The claim for originality in current trends lies in the intensity, extent and reach of global markets: a profoundly deeper division of labor, no longer just trade of primary products for manufactured goods, but trade of components through international production networks; coverage of the whole globe in a world economic system; and penetration of the system into culture, the movement of peoples and social institutions. The virtues and vices of globalization are hotly contested, but these are not the primary concern here.

The central issue for our purposes concerns the implications of convergence–divergence arguments for the linkage between domestic and international relations. Does globalization constrain all countries to become alike, or at least significantly more similar, in their economic policies, institutions, political economy, culture and social structure (Friedman, 1999)? If so, globalization acts as a system, constraining the units in it; it works like the system modeled by realists, a force that shapes profoundly the actions of its members, even internally. Or is there a substantial amount of slack allowed by the international economic system, so that countries can integrate into the world economy but none the less differ substantially in a wide range of policies, institutions and practices? If so, we then again need theories that explain what options countries actually chose to take (Berger and Dore, 1996).

On the side of convergence lie many examples of governments abandoning programs under the pressure of world markets: Mitterrand’s socialist government of the early 1980s; the requirement by the European Monetary Union to reduce deficits and debts; World Bank and IMF demands on Mexico, East Asia, Russia, Ecuador and other countries in recent years. How do these pressures work? The answer may appear obvious from the preferences pressure model noted above. Equity investors move against a currency, withdraw assets, or fail to invest in a country whose economic policy they dislike. This is obvious in a crisis situation such as a falling currency, hyperinflation or political instability (Haggard and Kaufman, 1995).

In crisis or non-crisis situations, what is it exactly that investors look for in defining good from bad policy (Goldstein et al., 2000; Mosely, 2000)? Political variables can be quite important, Simmons (1994) shows, as indicators of policy; in the interwar years, markets feared governments with ties to left parties and trade unions. In recent years, analysis has come to understand the role of regulatory patterns that structure market relationships. For example, Shinn (2001) argues that investors demand a premium for buying shares in the ‘closed corporate’ governance model of Germany and Japan over shares in countries like the United States with more open models that protect external shareholders (Gourevitch, 2001; Roe, 1994).

Against the convergence hypothesis is a rapidly growing literature showing quite substantial country differences in a number of policy areas. Comparing OECD countries, Iversen and Wren (1998) demonstrate that despite substantial commonality of economic pressures from the European and global economies, countries do differ significantly in the policy choices they make, picking different combinations of trade-offs among equality, growth and stability: a Christian Democratic model, which stresses stability, a Social Democratic model, which stresses equality, and a neoliberal model, which stresses growth. Garrett and Lange (1991; also Garrett, 1998) agree that there has been policy divergence. They provide an explanation based on partisan conflict–left vs. right strategies for dealing with international trade. Esping-Andersen (1999) stresses path dependence in shaping divergent responses to common pressures: previous choices interact with new trends in labor markets and family patterns to shape current developments. Streeck (1997) calls attention to the role of worker training systems in accounting for divergence. Scharpf and Schmidt (2000) and their collaborators examine vulnerabilities and capabilities of countries and provide case studies on such issues as women in the labor market, retirement systems, tax competition, as well as employment and equality. Kitschelt et al. (1999), Hall and Soskice (2001) and Stephens and Huber-Stephens (2001) all explore the importance of variance in micro-institutions interacting with political and macro-economic variables. Some older work by Cameron (1978) and Katzenstein (1985), focusing on the small countries in Europe have higher levels of government spending than large ones, has been extended by Rodrik (1997) and Bates (1997).

An important theme in much of this literature is an interaction affect among variables. Some analysts argue that making central banks truly independent from political influence produces a ‘technology’ that prevents inflation through strict monetary policy (Alesina and Grilli, 1993; Grilli et al., 1991). Conversely, the effect of central bank independence, Hall and Franzese (1998) argue, is strongly mediated by other institutions in society: strong trade unions, industry-wide bargaining, agreement on wage increases tied to productivity are all institutional features of German economic life which condition the impact of bank policy. Soskice and Iversen (2000) show that the validity of rational expectations models turns on institutional features of the economy; the models’ predictions are correct with perfectly fluid factor markets, but wrong when there is stickiness in those markets, such as trade unions, oligopolies or corporatist practices. Iversen (1999) develops a model that shows the need to integrate the autonomy of the bank with the character of the labor market. Drawing on concepts of delegation, it can be argued central bank autonomy relies on political support for the policies and outcomes they generate. Autonomy given is autonomy that can be withdrawn.

Hall and Soskice (2001) extend this embeddedness concept to find strong connections among many dimensions of policy and the economy. The United States and other neoliberal economies have fluid labor markets, generalized systems of education and training, fluid mechanisms of price and wage determination, stockholder highly competitive mechanisms of corporate control, and quite varied welfare systems. Germany and other countries in a ‘continental model’ in contrast have highly structured labor markets, education and training that link workers to specific jobs, structured mechanisms of price and wage determination, interlocking ownership patterns with weak shareholder involvement and no market for control, and highly structured welfare systems. There is a logic to the bundling: the German system provides incentives for investment in specific assets, which it needs to protect by highly structured institutions; the United States model rewards generalized investments by individuals and investors. If tightly bundled, the pieces of the system do not change separately; they are in some kind of equilibrium; changes in one will either be contained, or be quite destabilizing. Each system has its own strengths and weaknesses, its own comparative advantage.

This argument has quite substantial implications for the globalization debate. If quite different systems can operate effectively, there is more than one way to be efficient. Economic competition will thus not produce convergence, but rewards to specialization (Gourevitch, 2001). The division of labor will intensify, rewarding the specific features of national production systems. The convergence logic requires the assumption that there really is only way to do things, and the market has the capacity to reward those countries which do that, whatever the political resistance to doing so.

The existence of divergence, in turn, highlights the importance of domestic politics. External pressures do not translate themselves into policy automatically. Some political actors within a country must decide that accepting the dictates of the pressure is better than resistance. Evans (1979) analyzed ‘compradors,’ the domestic allies of foreign capital. Allies or not, policies do require domestic support, which leads us right back to the issues of interest group power, institutional aggregation and ideology noted above. The notion of embeddedness suggests the need for research strategies that do not isolate variables, but explore their relationships.

The Domestic Politics of Security Issues

If the boundary between international and domestic seems to be fading, so is the one between security and political economy. These two branches of work have generated contrasting research and theoretical traditions. There is good reason to think that behavior varies with issue areas: for example, it may be easier to cooperate on trade issues because there are so many more interactions. At the same time, both economic and security concerns involve politics. Both entail the making of choices in a situation of strategic interaction, and both can be analyzed from a common theoretical framework–the example just noted can be explained from a single logic, the impact of frequent iterations on cooperation (Axelrod, 1984).

While security studies and realist views of system and unitary actors are quite intermingled, the security field has its own rich tradition of domestic politics explanations of policy choices. Military budgets and policy have long been explored by the same logic of interest groups and institutions noted above in discussing trade. Where realists explain Germany’s decision to build a navy before the First World War as part of systemic rivalry with Britain, others see it as a response to domestic lobbies seeking to link a steel industry facing weak demand to military groups wanting bigger budgets (Kehr, 1977; Kurth, 1971; Snyder, 1991).

Recent work extends this type of analysis by looking at the domestic elements of strategic doctrine. Papayoanou (1999) shows the British were hindered in sending clear signals to Germany in 1914 about their course of action in case of war because domestic economic interests were divided between those with strong ties to Germany and those with ties to the Empire and other countries. Brooks (2000) links the ability of a country to develop comprehensive strategic doctrine to the structural relationships between the military and civilian institutions.

An important locus of work on security deals with historical memory in shaping expectations about future behavior. Public opinion among Japan’s neighbors protests when Japanese textbooks minimize Japan’s role in the Second World War and wartime atrocities (Buruma, 1994). Germany faced more muted reactions to unification, but there was concern none the less. Are these countries safely peaceful? The system answer is of course to look at the balance of power arrayed to deter aggression, while an international institutionalist would point to Germany’s integration into the European Union and NATO, and Japan’s treaty commitments. Domestic politics approaches look inside the country. Japan prohibits war in its constitution, limits the size of the military, and subjects the military to extensive civilian control by placing representatives of other civilian ministries on the bureaucratic structures that supervise it (Katzenstein, 1996). Germany now makes clear that the armed forces are subject to democratic civilian control.

Emphasis on interest groups would call attention to profound social change in each country: land reform in Japan; the elimination of the Junkers as a social class in Germany. Emphasis on culture would stress pacifism and anti-militarist sentiment in each country. Berger (1993) shows how arguments over the causes of the war in Japan correlate with political positions: there is widespread agreement on the role of the military, but not on the other causes–zaibatsu interested in expanding investment in Korea, Manchuria, China; social groups aligned with the military, repression of democratic forces, Japanese culture.

Another line of research on security issues applies interest in institutions to civil–military relations. Where realists interpret military structure and behavior in terms of system threats, a domestic view sees the armed forces as having their own internal interests and political role. As comparativists study democratization processes, a vital element is the changing relationship of the military to the formal political system and to society. Studies of Latin America frequently portray a military active in internal politics far more than one defending national borders or projecting force outside the country. The military regards itself as an arbiter of domestic political struggles, a guarantor of goals it regards civilian forces as unable to attain or protect, and often demands institutional guarantees of that role (Loveman, 1993; Stepan, 1988). Thus Pinochet overturned the Chilean constitution to stop Allende; one example of many military coups. Argentina provides an example of an old idea in this field of study: that foreign policy adventures are used to cover up political weakness at home. The Argentine generals used the Malvinas/Falklands crisis for this purpose, but lost their bet to UK prime minister Thatcher, for whom it did provide a political rebound (Richardson, 1996).

The study of culture, never wholly absent from security studies (Benedict, 1946; Leites, 1950; Osgood, 1953), has attracted increasing attention, in part from the theoretical advances of constructivism and in part from the explosion of ethnic conflicts around the world. Huntington’s The Clash of Civilizations (1996) unleashed a storm of criticism, and heightened interest in the issue. The main criticism is not so different from the ones levied by domestic politics experts against unitary actor models: countries and cultures are complex constructs of many elements, capable of supporting a wide range of actions. The ideas that triumph may indeed have an influence over policy, but one needs a political explanation for that triumph rooted in institutions, interest, politics and conflicting ideas.

European considerations of culture in the security field focus more on the structure of communication and discussion, of which Habermas is the most famous theoretician. This has influenced a substantial literature (Risse, 2000), much of it theoretical, which appears to be generating much needed application to empirical cases. It could be promising to link this line of work to research drawing on cognition, learning, information cost theory–a range of material dealing with how people think about policy issues and communicate with each other (Jervis, 1976).

The resurgence of ethnic conflicts in recent years provides ground for new work on culture. Within countries, these can lead to civil war–a breakdown of authority which swiftly erases the boundary between international and domestic politics. (See the section below on ‘Composing and Decomposing States.’)

Political Forms: Democracy and Authoritarianism

How do democratic institutions influence foreign policy and strategic interaction? If democratization continues to spread, this becomes an increasingly important question. The discussion of the democratic peace literature is treated elsewhere in this volume. Our concern here is with the way the fact of democratization influences international relations.

In a number of ways theorists argue democracies are able to make more credible commitments to other countries because of their institutions. Forced by democratic rules to operate publicly, leaders have to worry about ‘audience costs’ (Fearon, 1994), the price leaders pay to constituencies for going against their wishes. One audience is the foreign power, who judges intention, resolve and capacity. Another audience is a domestic one, those publics that have the capacity to remove leaders from office or prevent them from attaining desired objectives. Public accountability in a democracy limits the range of likely behaviors that happen in an autocracy, where the rulers have fewer immediate constraints. External observers can thus evaluate the future in connection to the cost they think leaders will avoid paying. It is harder in a democracy to shift policy quickly, hence they are more credible in commitments they make (Schultz, 1999). Democracies may take longer to make decisions, but these have more stability because they have required a broader engagement of society in the approval process (Martin, 2000).

It is often argued that the problem of audience costs makes democracies less effective in foreign policy (Crozier et al., 1975). Conversely, it can be argued that audience costs may strengthen a government in negotiating cooperation. Because open political processes reveal the domestic political game, leaders are able to show convincingly their costs to making concessions (Cowhey, 1995; Schultz, 1999). An agreement may be more difficult to reach but more credible once signed. Martin (1994) argues open processes of ratification in the European Union give greater strength to the agreements than was the case when they were done by executives behind closed doors.

Do democracies make ‘better’ decisions? The need for open debate may yield better information, since rulers are not able fully to limit the boundary of acceptable discussion. The extent of pollution and health hazards was for example quite hidden in the Soviet bloc since there was no ability to object. The ability to have voice and power shapes the considerations that enter a decision. Sen (1981) argues that since the eighteenth century famines have occurred not because there was insufficient food but because the poor lacked the voice or power to make their needs felt, thus to have food made available to them. Women, he argues, are poorly treated in many countries again because they are disenfranchised. Mass publics may be more averse to the loss of life in war than are leaders, not less belligerent or xenophobic, but more ready to abandon costly projects (Lake, 1992).

Can mass publics in a democracy really understand the issues? Information theory gives new answers to this old question. Foreign policy is by no means the only issue area that poses great requirements on individuals for information and analysis: health, environment, science policy, safety standards for products all involve complexity beyond the reach of most people, and, for the wide range of issue areas, beyond the scope of even the brightest well-educated person. Voters find ways of simplifying their information tasks by looking for shorthands, or heuristic indicators. These often consist of individuals or groups whose expertise they trust, whose values are similar to theirs: thus an environmentally concerned voter looks to the Sierra Club for advice, or a human rights voter to Amnesty International (Lupia and McCubbins, 1998). In this way, the public is able to monitor its agents without the kind of deep investment in every issue which no person can do. In this regard, foreign policy is no different than any other issue area of complexity.

Democracies do of course vary substantially in their institutions. In addition to voting rules, legislative executive relations and party systems, there is a rich tradition of examining bureaucratic structures and organizational processes. Allison’s (1971) famous account of the Cuban missile crisis remains quite appropriately a classic, both for its introduction of alternatives to unitary models of the state, and its strategy of presenting alternative explanations of the same event. Its weakness is that the power of the bureaucracies and organizations themselves cannot be understood outside the context of the political system itself. Having shown that a fight among bureaucracies occurs, the bureaucratic politics model does not tell us how the fight is resolved, why one ministry coalition defeats another. Having shown us the consequences of organizational process, that model does not tell us why that process was selected, or why the agency in question had the power to apply its procedures. As Bendor and Hammond (1992) argue, the presidency sits at the top of an internal hierarchy whose character influences the President’s power resources. A model of the political system must therefore lie at the core of any institutional analysis of foreign policy. During the Cuban missile crisis, it was in the end the President who made the final choice for moderation (Evans et al., 1993; Pastor, 1980; Welch, 1992).

Another famous example comes from the mobilization plans of the European powers in 1914 and Germany’s Schlieffen Plan, which together helped turn a crisis into full-scale war (Craig, 1955; Keegan, 1998; Miller et al., 1991). Posen (1984) argues the military prefer offense-oriented strategic doctrines, largely because this demands more resources and puts them centrally into decision-making, while the civilians are more cautious, fiscally prudent and thus inclined toward defense-oriented doctrines. Kier (1997), by contrast, argues the choice of doctrines turns on military culture and its relationship to domestic politics. In Japan, it was the army that favored expansion into Manchuria, then into China; the navy was more cautious about the ability to project force in competition with the navies of the other great powers (Sagan, 1989). Richardson (1996) shows how these processes influenced conflict between the United States and the UK in the Falklands, Suez, and the Skybolt missile cases.

Trade policy shows bureaucratic politics at work. The US Commerce Department and trade-related agencies push Japan for trade concessions, while the State and Defense Departments favor deference to Japan’s importance as an ally on Asian security considerations. The disagreement is mirrored in Japan’s internal debates on how to respond to trade pressures: Japan’s Ministry of Foreign Affairs often favors concessions to the United States while the Ministry of International Trade and Industry often takes a tougher stand (Destler, 1995; Schoppa, 1997).

If democracies are less warlike and make better decisions, should promoting democracy be a goal in foreign policy? The peace–efficiency arguments add an instrumental dimension to the moral one, that democracy is a value in itself. The United States and other democracies may find it preferable to deal with other democracies–would a democratic China be a more cooperative one? China may not agree with the obverse: that a democratic United States is a better ally than an authoritarian one. Preferring democracy is not the same as expending costs to get one.

The question of democracy, long associated with issues of security, has increasingly entered the political economy field. The older argument comparing democracy and authoritarianism toward prospects for economic growth continues to stimulate research (Haggard and Kaufman, 1995). It has shifted now toward concerns with institutional design and effective governance, in which democratization is one component.

North and Weingast (1989) argued in a well-known paper that constitutionalism gave the British an economic advantage over their absolutist continental counterparts. Because Parliament could monitor the Crown, loans to the government were less likely to be renounced, reassuring lenders, giving seventeenth-century England lower interest rates than the Sun King’s France and the other absolute monarchies of Europe. The limited state was thus a stronger state (Brewer and Echhart, 1999). This line of reasoning turns on its head an earlier discussion about strong and weak states, which saw the centralized states as strong, and highly constitutionalized and pluralist ones like the United States as weak (Katzenstein, 1985). Constitutional governments may work more slowly, but because they must mobilize consent, they are more consistent and reliable.

This debate has taken a new turn in the past decade, with a widespread interest in transparency and effective governance. The rapid growth of East Asia and Japan in the post-war years undermined quite substantially the dependencia notion of world systems theorizing about growth: the idea that countries were confined to a particular location in the global division of labor from which they could not move. Clearly countries could move, if they made certain choices. The countries of rapid growth were frequently not democratic or if so, as in Japan, held to be state-centered in development model, and bureaucracy-dominated in policy formation (Haggard, 1990; Wade, 1990).

The Asian Financial Crisis of 1997, the slow growth of the Japanese economy, and gradual rejection of the import substituting and stateled policy model in Latin America altered the debate quite substantially. The developmental state was now called corrupt crony capitalism, vulnerable to moral hazard. The solution, in the so-called Washington consensus, was transparency which required greater democratization. International and domestic economic policy-making institutions developed a strong interest in the institutional foundations of policy-making, long absent from these circles. Policy-makers now speak of the problem of good governance, of how to design institutions in developing countries capable of making effective decisions, capable of providing public goods rather than particularistic pork barrel (Haggard and McCubbins, 2001).

As policy-makers seek advice from institutionalists on the design of good governance institutions, they find, MacIntyre (2001) argues, conflicting advice. One branch of institutionalists thinks democracy, with its multiple ‘veto-gates,’ inhibits decisive action, so that authoritarian governments have an advantage; the other approach thinks the veto-gates build consensus and commitment, so that democratic governments have the ability to be consistent (Lijphart, 1999; Shugart and Haggard, 2001; Tsebelis and Money, 1997). MacIntyre (2001) integrates the two lines of reasoning into a U-shaped curve; countries having one extreme or another on each dimension behave less well than governments having a blend of the two. International pressures and domestic political-processes and institutions blend well in this issue area. Haggard and Kaufman explore the ways democracy influences the adjustment countries make to international economic pressures in developing countries (Haggard and Kaufman, 1995). Further progress on the role of institutions requires more work integrating institutional analysis with ideas about interests and ideas for countries and regions where institutions have not received much attention, particularly the developing ones.

A particularly great challenge lies in modeling authoritarian regimes. These are no longer modeled as unitary states, expressing the will of a dominant leader. Experts on the USSR and Nazi Germany long ago introduced ideas of competing forces within these regimes, struggles for dominance involving a kind of politics. It has been none the less difficult to import all of the tools of institutionalist analysis because of the need to find a clear utility function for leaders and a stable account of their constituencies and incentives. In democracies, leaders can be assumed to seek re-election. For authoritarian leaders, the functional equivalent of the electorate and open voting systems must be found in order to define the leaders’ objective utility function. Roeder (1993) and Shirk (1993) have used the concept of ‘selectorate’ to solve this problem and apply it to the Soviet Union and China, respectively. Authoritarian leaders do not have to win public elections, but they do need to maintain the support of key élites and figures in a society to remain in office (Bueno de Mesquita, et al., 1999a, 1999b).

It remains none the less difficult to provide fully satisfying accounts of authoritarian regimes from an institutionalist’s point of view. Leaders appear to have substantial discretion, despite the constraints of a selectorate. They can shift institutions and their selectorate with greater ease than in a democracy, which leaves a more fluid game. None the less, this appears to be a rich area for further work, linking together insights from institutionalism, advances in the modeling of interests in the open world economy, and the evolution of ideas about appropriate models of institutions and development models.

Constructivism, Culture, and Domestic Politics

Constructivists have helped bring about a revival of interest in ideas. The portion of the debate relevant for this chapter is to explore the implications of constructivist approaches for research on domestic politics. Ideas have been taken seriously before in the study of both domestic politics and international relations, but it is fair to say that the major literatures of political economy and institutionalism do not give them much scope. Discussions of norms, discourses and constitutive understandings all require considerable research into the actions of individuals and groups within society and their dialogue with counterparts elsewhere.

Several lines of investigation can be observed. One approach engages in careful historical reconstruction of the emergence of norms by examining the statements or speech acts of leading figures and comparing these to patterns of behavior. In this way, Tannenwald (1999) looks at the norm against the use of nuclear weapons, and Lynch (1999) examines the writings of peace movements prior to the Second World War to challenge the realist treatment of them as ineffectual and naïve. Another approach looks at law as a generator and distributor of norms and the institutions that could enforce them (Finnemore and Toope, 2001); this leads to examination of courts and ‘law-like’ institutions, such as international regulatory bodies (Alter, 1998; Mattli and Slaughter, 1998). Discourse analysis looks very carefully at speech acts. Other work looks carefully at survey research on public opinion (Herrmann and Shannon, 2001). Tarrow’s (1998) work on the general phenomena of movements provides a basis for understanding those that advocate various value concerns in politics.

How are norms taken up in domestic politics? Norms shape preferences. As with any other variable involving preferences such as economic interest, the study of norms requires an analysis of why one or another normative orientation predominates. Ideas, understandings, discourse have a political sociology to them: groups who advocate or oppose them, institutions which favor or hinder them, prior cultural commitments that encourage or oppose their adoption. The political sociology of understandings can be non-ideational (looking at interest groups and institutions, in which case see the discussion above) or it can itself be culturalist in its causal mechanism. Integrative work on norms explores these dimensions: Kaufmann and Pape (1999) examine the social characteristics of the anti-slavery activists in Britain, how they organized for political action, how they made alliances with other groups, and how they were able to operate in the political institutions of the day. Keck and Sikkink (1998) show how NGO groups put pressures on governments to assist groups in other countries seeking to defend human rights; they need to mobilize voters, make allies, influence elections and work through political institutions. In studying Japan, Katzenstein (1996) seeks to locate norms in specific institutions; the purveyors of norms act instrumentally to design an institution that will favor a particular outcome (in this case to limit the autonomy of the Japanese military).

These arguments differ from more system-oriented constructivist research, which focus on international interactions. Wendt (1992, 1999) is a major example. He speaks of states constructing understandings of the world, how they are socialized into accepting rules. This sort of reasoning implies a unitary way of thinking which downplays the arena of domestic politics. It draws us away from examining the processes within a country that lead to the absorption (or opposition) of these international norms. The cutting edge of constructivist research has got away from proving the importance of norms by positing two alternatives, realism or constructivism, and then proving the latter by showing flaws in the former. Instead, it has moved toward the integrative task of showing how ideas merge with material and institutional power through domestic political processes (Checkel, 1997; Katzenstein, 1996; Kaufmann and Pape, 1999; Keck and Sikkink, 1998).

The theoretical and research issues raised by constructivist concerns are quite significant for analyzing the range of issues arising over values in international relations: human rights, democratization, gender and related normative concerns on areas like the environment, abortion, the death penalty and cultural autonomy. Are such values universal or local/regional? This issue poses some considerable analytic problems in the treatment of domestic politics. How are we to know whether ideas are ‘alien imports’ ‘imposed?’ In particular, how are we to know this in political systems which are themselves not democratic? The ‘alien import’ argument is frequently made by authoritarian leaders in regimes where there is no possibility of autonomous debate; how are we to know if they speak for their people (Neier, 1993; Sen, 1981)?

Debates on norms in international relations lead therefore straight to domestic politics, to an analysis of why norms are supported or opposed in any given country, which in turn links this discussion to all the debates on how to explain domestic outcomes.

Non-State Processes and Domestic Politics–the Internationalization of NGOs

In the challenge to unitary state/system theories, the study of transnational relations and international civil society were in the forefront. Both decentered the state as the only relevant unit of international relations, and both disaggregated the state, opening it up to political analysis of contending views of the national interest. Recent work on NGOs (non-governmental organizations) builds on this earlier work in advancing arguments about new types of transnational groups, the role of norms and new forms of civil society. NGOs act across national borders to push for various goals: the environment, human rights, equality for women, working conditions for children (Evangelista, 1999; Keck and Sikkink, 1998).

Theorists in the English School situate NGOs in the framework of transnational forces as comprising a kind of ‘civil society’ at the international level (Bull, 1977). These ideas resonate with a revival of interest in civil society by comparativists interested in democratization (Pérez-Díaz, 1993; Putnam et al., 1993). In terms of domestic politics, this leads methodologically to a shift away from the centrality of formal processes of the state in shaping policy outcomes, and toward the study of social institutions. Consumer boycotts are an interesting example: consumers force countries to comply with child labor or environmental rules without formal government legislation by refusing to buy products that have not received approval by the NGO (Spar, 1998).

Alternatively, NGOs, along with other aspects of civil society and transnational forces, can be located in domestic political processes. Risse-Kappen (1995) argues that variance in the impact of transnational forces turns on differences in domestic structures. Keck and Sikkink (1998) show this in ‘boomerang’ effect. Protesters in country A are blocked by authoritarian institutions from the capacity effectively to pressure their government. They link up with sympathetic NGOs in country B, who are able to pressure their government to put pressure on the government of country A. If country B has leverage over A, that pressure can be effective. The power of the NGOs turns on their ability to persuade citizens in the democratic country to sanction their government unless it follows their suggestions.

In this perspective, NGOs can be analyzed as interest groups, operating like others, to influence public policy by mobilizing pressures on governments. What makes them distinctive is their transnational scope of activity and the use of new instruments of communication. When NGOs mobilize international pressure, we need to know who it is within the state that ‘feels’ this pressure, and acts to demand a policy response from their own government. In that regard, the argument is similar to that used in political economy. Japan is more likely to comply with US pressure (‘ghaitsu’) concerning trade imbalances, Schoppa (1997) argues, if there are domestic forces who agree with the complaint made by the Americans, forces that are sympathetic to the changes being proposed; where there are no such sympathetic domestic forces, Japan resists. Thus the impact of international pressure, be it from civil society, NGOs, or open state diplomacy, requires some understanding of the domestic debate about compliance–a model of what Kahler (2000b) calls the ‘compliance constituencies’ and the political process around them.

State Formation and Decomposition: Nations, Federations, Civil Wars, Evolution

States may not be the only players, but they are important ones. Yet international relations theory has generally ignored the problem of where states come from or why they fall apart. The emergence of the European Union, the decomposition of the USSR and the spread of ethnic conflict has brought these issues forward.

New Entities

The European Union has profoundly altered the study of the domestic–international interaction. By joining a number of countries, many historic enemies, into a highly developed international institution it raises all the theoretical issues of the domestic politics interaction being explored here–the role of institutions, interest groups, ideas; the importance of democracy, civil society and the formation of identity; concepts of delegation, agency, international institutions and system. The very definition of the EU shows the degree of thematic interaction: is the EU an international institution or a country, or something else altogether? That question leads to a number of others.

How does the European Union obtain compliance from its members? The system-centric approach of neoliberal institutionalist theory locates the causal mechanism to the utility of an institution to lower transaction costs among its participants (Keohane and Martin, 1995). By itself, that view says nothing about domestic politics; it only requires that countries have an interest in cooperation, but allows these to be unitary. The power of the EU lies in the desire of countries to continue the cooperation, the costs to them of leaving or having the union break up.

If we assume countries have internal debates about cooperation, then it is necessary to examine the politics that causes one or the other side to prevail. Cooperation turns on the success or failure of ‘compliance coalitions.’ The interesting debate there lies in modeling the impact on domestic politics of involvement in the EU. Moravcsik’s (1998) intergovernmentalism model has two steps to it: first, a domestic politics produces a definition of national interest; then, the agents of the nation bargain at the European level. The impact of the EU itself lies in the way its institutions influence the bargaining process between agents of the national governments, a classic institutionalist model. Cowles et al. (2001) seek to reverse the causal flow, from the European level downward: being in Europe alters interests, preferences and identities (Banchoff, 1999). Political actors reach different conclusions about what they want from being in Europe than they would if it did not exist. Those opinions do get refracted through national governments, but they are still different (Evangelista, 1997; Risse-Kappen, 1991). This interactive model is surely correct; its problems lie in the difficulties of operationalizing the differences between interest, preferences and identities, as these can easily be confused.

Questions of identity in relation to institutions arise in analysis of the ‘democratic deficit.’ The more the reach of the EU grows, the greater the problem of popular commitment to its goals (Scharpf, 1999). Since national governments are represented, not people, the public feels some disconnection from the EU’s decisions. To increase the role of directly elected representatives undermines the national foundations of the EU. Can such a system of representation be created without a cultural foundation of identity with the Union rather than with nations? Haas doubted spillover could create a nation (Haas, 1958; Deutsch, 1953). It remains a serious question as to whether one can move from customs union to country without powerful nation-building experiences that transform identity as well as interests and institutions. From a methodological point of view, it is hard to define and test the movement from economic union to nation. If the EU is more than the average international institution, but not a nation, what is it? Perhaps we need new categories.

The study of the EU’s institutions shows how concepts drawn from domestic politics are being applied to an international process: the application of institutional design concepts from formal modeling on the decisionrules of the Union, the debates about reform and expansion (Tsebelis and Money, 1997); the role of the European Court of Justice, its ability to strengthen the Union through its rulings and the internalization of these rulings in each country (Alter, 1998; Mattli and Slaughter, 1998); the negotiation of Union-wide standards and regulations governing commerce, finance, safety and health are some examples (Scharpf and Schmidt, 2000); the spread of norms and cultural elements of cooperation (Sandholtz, 1993).

The European Union provokes comparisons of great theoretical interest. It is surely the most advanced contemporary case of integration. Can the same happen in other regions? Or do they lack some critical ingredient, and if so, what? An external threat, economic interdependence, a common culture or identity, geographical proximity (Kahler, 2000a)?


While new states emerge, others fall apart, changing the constituent elements of the international system. Civil wars push analytic boundaries quite forcefully. The breakdown of national authority often draws international actors into the domestic politics of the warring country. Morally, should outsiders enter a domestic quarrel? Sovereignty has been, as Krasner (1999) shows, an inconsistently applied principle, but it can be used to defend local autonomy, independence and culture. Domestic issues influence the probability of internal conflict, but international forces are usually an influence as well and almost always a player in the struggle (Walter, 1997). Governments face the same challenges as social scientists: what are effective means of nation-building?

Civil wars often arise from the assertion of cultural distinctions, such as religion, language, ethnicity. How do peoples living peacefully for many decades suddenly begin killing each other? A vigorous literature has arisen on this topic, blending game theory with careful ethnographic description seeking to understand the incentive situation that leads to rapid ‘tipping,’ where a new piece of information about possible dangers induces rapid change by a large number of people in their strategic evaluation of the ‘rewards’ to peace and violence (Fearon and Laitin, 1996; Laitin, 1998). A neighbor seen as friendly is now suddenly seen to be dangerous. A similar line of reasoning has been applied to the collapse of the regimes of Eastern Europe (Hirschman, 1993; Lohmann, 1994).

All of the concepts discussed so far about the creation or disintegration of states assumes some degree of conscious agency at work. A relatively new and interesting way of thinking about these issues, an evolutionary approach, draws on biological models, to analyze states as unconscious respondents to forces that push them one way or another (Kahler, 1999). The mechanism is adaptation to pressures from other units, themselves responding to overriding forces such as technology, war and markets (Downing, 1992; Ertman, 1997; Spruyt, 1994).

Most of these discussions blur old analytic boundaries. The emergence of new structures such as the European Union blurs the distinction between anarchy and institutionalization. Civil wars show that domestic politics can include anarchy. Evolutionary models blur the issues of choice and agency, and play down the role of domestic politics. Lake argues we should place states and anarchy as end points on a continuum of political forms, ranging empire (direct control from the center of sub-units), to federations, to voluntary confederations, regimes, alliances and finally anarchy; in his book, this allows him to examine the forces that shape one form of relationship over another (Lake, 1998).


The great challenge confronting the domestic politics research agenda is to model the interaction of countries with each other and with the system. We have developed strong research traditions that hold either system or country constant. We do not have very good theories to handle what happens when both are in play, when each influences the other, when the domestic politics of one country interacts with the domestic politics of another, an interaction which itself helps define a system that reverberates back on the parts.

We have good metaphors, but not clear research programs. Putnam’s (1988) ‘two-level games’ paper struck a very strong, and welcome, chord. It has been widely used to inspire and legitimize an exploration of domestic politics. Yet most writers evoke it rather than follow it as a formulated research program. Milner (1997) is the most careful application, but one sees there the difficulty predicting, hence testing, clear interactions. It is hard to specify the objectives of all the players and the results produce multiple equilibria, rather than a clear resolution. The general formulation does not, moreover, actually model interaction of the domestic politics of different countries. It asks what the win set in one country does to the likelihood of cooperating with another, but does not include the domestic politics of the second country.

Gourevitch’s (1978) ‘second image reversed’ has also been more of a metaphor than a guide to a research program. It helped introduce the idea that country institutions and internal interests were influenced by international forces, rather than being only the shaper of such forces. Often cited, the phrase evokes the idea of interaction between levels, more than a research strategy of how to study that interaction.

To develop interactive models, we need to avoid the reappearance of unitary actor assumptions, which often occurs. Endogenous macro theory appears to say that interests are shaped by a position in the global division of labor, but does not include the role of domestic politics in specifying world markets. In contrast to Rogowski’s account of governments as the relatively passive registrants of societal pressure, Simmons and Elkins (2001) develop a model that has governments as self-aware strategic actors. This makes much sense, but it does risk being unitary in its logic. NLI writing speaks of the state’s interests in the institutions, not of the support constituencies in each country that shape the determination of state interests. Constructivists speak of discourses or understandings between countries constituting their interests, but frequently do not explore why one or another discourse prevails in a particular country.

Waltz’s (1959) three images remain powerful and useful tools for organizing our thinking, but the thrust of work in recent years has been to break down these boundaries, to integrate anarchy and domestic politics, to integrate individualist perspectives with theorizing about states and institutions. Domestic politics has become central to most discussions of international relations.

If, as Lake and Powell (1999) write, strategic interaction lies at the core of the study of politics, the elements involved in shaping that interaction are indeed rather vast. They include the many branches of domestic politics interacting across boundaries through many different mechanisms (markets, culture, force). The integration of domestic politics with international, and the integration of international with a generalized study of politics gives us a language. It does not make our task easier as the integration of concepts is indeed difficult. It remains to be seen whether the field accepts this reworking of its categories–perhaps we will know more at the next edition of the Handbook.