Interests, Institutions, and Information

Domestic Politics and International Relations

Helen V. Milner

• Publisher: Princeton University Press
• Year: 2020
• Print ISBN: 9780691011769, 9780691011776
• eBook ISBN: 9780691214498
• Topic: Politics & International Relations
• Subtopic: Social Policy
• Index: Politics & International Relations

Increasingly scholars of international relations are rallying around the idea that "domestic politics matters." Few, however, have articulated precisely how or why it matters. In this significant book, Helen Milner lays out the first fully developed theory of domestic politics, showing exactly how domestic politics affects international outcomes. In developing this rational-choice theory, Milner argues that any explanation that treats states as unitary actors is ultimately misleading. She describes all states as polyarchic, where decision-making power is shared between two or more actors (such as a legislature and an executive). Milner constructs a new model based on two-level game theory, reflecting the political activity at both the domestic and international levels. She illustrates this model by taking up the critical question of cooperation among nations.


Milner examines the central factors that influence the strategic game of domestic politics. She shows that it is the outcome of this internal game--not fears of other countries' relative gains or the likelihood of cheating--that ultimately shapes how the international game is played out and therefore the extent of cooperative endeavors. The interaction of the domestic actors' preferences, given their political institutions and levels of information, defines when international cooperation is possible and what its terms will be. Several test cases examine how this argument explains the phases of a cooperative attempt: the initiation, the negotiations at the international level, and the eventual domestic ratification. The book reaches the surprising conclusion that theorists--neo-Institutionalists and Realists alike--have overestimated the likelihood of cooperation among states.

Part One

THE THEORY

Two

Actors’ Interests, Policy Preferences, and the Demand for International Cooperation

WHEN domestic actors share power over decision making and their policy preferences differ, treating the state as a unitary actor risks distorting our understanding of international relations. Instead polyarchy reigns, affecting how states act in international politics. This chapter lays the basis for my primary claim: the structure of domestic preferences holds a key to understanding international cooperation. Domestic actors’ preferences are primordial. This chapter sets forth the interests of three key groups of domestic actors: the executive (the prime minister or president), the legislature, and interest groups. It then derives their policy preferences from these basic interests. The central question addressed is why these groups would ever be interested in cooperating. The main concern is to show when these three sets of actors will have policy preferences that favor international cooperation and when important differences in their preferences will arise.

The policy preferences of actors in domestic politics derive from their basic interests. Actors are assumed to have certain fundamental interests, captured by their utility functions, which they attempt to maximize. For political actors, this means maximizing their ability to retain office; for social actors, maximizing their net income. For both, their most preferred policy—or their “ideal point”—is that policy choice in the issue area that maximizes their basic interests—that is, retaining political office or maximizing income. This chapter seeks to model how these actors’ policy preferences are shaped and to suggest the conditions under which various structures of domestic preferences are more likely. In particular, I am concerned with the case of divided government, where the preferences of the executive and the legislature diverge. As I will show in chapter 3, structures of preferences involving divided government have important consequences for international cooperation. This chapter shows the conditions under which (more) divided government is likely.

The Agents: Executives, Legislatures, and Interest Groups

Two sets of policy preferences are of importance; those of political actors, such as the executive and the legislature, and those of societal actors, such as interest groups. Each of these three groups is assumed to be unitary and rational. These assumptions are significant, and here I explain what they mean for each set of actors.

The executive refers to the executive branch of government or the cabinet either led by the prime minister in parliamentary systems or the president in presidential ones, as well as the departments supporting the cabinet in both systems. Obviously, to claim that the executive branch is unitary is to make a simplifying assumption. Politics within the executive branch may be as complex and consequential as politics between it and the other branches (King 1976). For heuristic purposes, however, the executive is considered to act as if she were a unitary entity. This can be justified in three ways. First, since the prime minister or president is primes inter pares among the cabinet, all decisions must have her backing to go forth. Therefore the prime minister or president is the actor referred to by the executive. Second, one could argue that on each policy issue the cabinet minister in charge of that issue is the most important (or only) decision maker, and hence the minister is the unitary actor meant by the executive (Laver and Shepsle 1995). In contrast, by the executive one might refer to the median cabinet member; that is, the minister who casts the deciding vote for a policy represents the executive. Whichever way one proceeds—and I use the first formulation– the consequence is to make the executive branch a unitary actor.

The executive is also treated as being rational. The executive wants to maximize her utility, which is assumed above all to depend on reelection. Reelection is not the only goal attributed to political actors; some have argued that political actors desire to implement their party program most of all (e.g., de Swaan 1973). In reality, some combination of these two motives is probably most accurate. Here, however, I make the simplifying assumption that staying in office is the main goal of executives. As Snyder and Diesing note, “Any politician who wishes to participate in public action is thus faced with the imperative of maintaining or increasing his power, authority or influence. As Sam Rayburn observed, “To be a statesman you have to get elected’” (1977:354).

This assumption of an “office-seeking” motivation has a long tradition (Downs 1957). Practically, what this assumption means is that the policy preferences of executives need not follow their party platforms nor their campaign promises. Instead, executives can pick and choose among policies to best serve their immediate electoral interests. Moreover, it means that electoral considerations motivate policy choices. “Economic policy is chosen by political agents and political agents seek to win office through elections. An adequate theory of economic policy formation, therefore, will take account of the electoral incentives facing political decision-makers” (Austen-Smith 1991:73).

In order to maximize their chances of reelection, executives have to worry about two factors: the overall economy and the preferences of interest groups that support them (e.g., Grossman and Helpman 1994, 1995). Executives first must ensure that the general performance of the economy is good. A declining economy may lead voters to support those out of power in the hope that they can improve the situation. As the leading, recent survey of the relationship between economics and elections in Western democracies shows, voters use retrospective economic evaluations of their national situation both to punish and reward incumbents (Lewis-Beck 1990). This “performance constraint” forces executives to be concerned not just with policies that please their particular constituents but also with those that serve the general economy. It may also create a trade-off with the second element in executives’ utility function: their support for special interests.

Although ultimately voters elect political leaders (directly or indirectly), special interests can be an enormous help to leaders. They can produce contributions, votes, campaign organization, media attention, and so on, all of which may make the difference between a winning and a losing campaign. Leaders need the support of interest groups, and to gain it they must promote (retard) the policies that help (hurt) these groups. “Politicians seek office through elections; to run campaigns they need resources; to acquire such resources, they include redistributional policies in their platforms . . .; these [policies] induce the favored interest groups to provide such resources to support the relevant candidate” (Austen-Smith 1991:73). Hence executives will be concerned with maximizing their chances of reelection, which depend on both the state of the economy and the support of key interest groups. Executives will thus try to choose policies that optimize both the state of the national economy and the interests of their interest group supporters. Policies entailed by international cooperation will only be chosen if they fit this criteria.

The second actor is the legislature. Again it is assumed to be unitary and rational. Neither of these assumptions is unproblematic. The title of an article by Kenneth Shepsle (1992) summarizes the problems of the unitary actor assumption: “Congress Is a ‘They,’ Not an ‘If’: Legislative Intent as Oxymoron.” As with the executive, legislatures have important internal politics that affect how they operate and what policies they adopt (e.g., Fenno 1973; Krehbiel 1992). Moreover, bicameralism explicitly undercuts the assumption of legislative unity. Nevertheless it seems useful to abstract from these considerations. What we want to know is whether the legislature will vote in favor of an international agreement proposed by the executive. Hence the focus is on the median legislator. The member of the legislature who casts the deciding vote on the international agreement becomes the actor who represents the “unitary” legislature. His preferences are what we mean by the legislature’s preferences.

Like the executive, legislators are assumed to be rational. They seek to maximize their utility, which derives from maintaining their office. As political actors, both the legislature and the executive are assumed to have similar interests, although they may prefer different policies to maximize these interests. Hence legislators want reelection, which depends on the state of the economy and the support of their interest group constituents. Policies that enhance the economy and bring gains to their interest group supporters will be their preferred ones. These, however, may or may not be the same policies that the executive prefers.

Why, if the legislature and executive share common interests, might their policy preferences differ? Executives and legislators represent different constituencies. Both the type and importance of special interests in their constituencies may differ. In presidential systems, where the two are elected in separate elections this point is fairly obvious. Executives must worry about a national constituency, whereas legislators are concerned with their local district. Depending on the electoral laws, their district may represent a small or large part of the nation. Moreover, in multimember districts legislators may represent only part of their district, further narrowing their constituency and differentiating it from the executive’s. As Olson (1993) and others (e.g., Lohmann and O’Halloran 1994) have noted, politicians with more encompassing jurisdictions are more concerned with overall national outcomes and thus have different preferences than those with more narrowly defined jurisdictions.

In parliamentary systems differences in preferences between the executive and the legislature seem less likely. The executive is usually chosen by the legislature; thus the election for legislators may be considered the same as that for the prime minister. Nevertheless each legislator is concerned primarily with his district whereas the prime minister must be concerned with all the districts, especially the median one. As Shugart and Carey (1992:3–4) observe:

A major dilemma in democratic regimes concerns a divergence between what representative assemblies do best and what executives must do if democracy itself is to function well. Assemblies. . . are intended to be representative of the population. A typical democratic assembly is elected for the purpose of giving voice to the interests of localities or to the diversity of ideological or other partisan divisions in the polity or society. That is, assemblies are ordinarily expected to be parochial in nature. Executives, on the other hand, are charged with acting to address policy questions that affect the broader interests of the nation, as well as to articulate national goals.

The national focus of the executive and the more local concerns of legislators help explain why, although they may have the same interests, legislators and executives may have distinct policy preferences.

The third set of actors are interest groups. Each group is assumed to be unitary and rational. Politics within an interest group is important, as the literature on collective action suggests (Olson 1965). But I assume that each interest group behaves as a unit, reflecting the median member’s preferences. Moreover, as discussed in detail later, each interest group is rational; it tries to maximize its members’ income—for example, wages, profits, and so on. This assumption needs less justification than the others since it is fairly standard. Hence an interest group prefers policies that maximize its income, and it will support (oppose) policies entailed by international cooperation that promote (detract from) this interest.

In the model the policy preferences of these three domestic groups and of the foreign country are used to define the “structure of preferences.’’ This structure relates the relative policy preferences of these groups along a single dimension. It assumes, that is, that the actors have preferences for a policy that can be arrayed along a line; they desire more or less of the policy. Their preferences are reducible to a single dimension. This assumption is a substantial but frequently employed one (McKelvey 1979; Schofield 1983; Enelow and Hinich 1990; Romer and Rosenthal 1978, 1979; Ferejohn and Shipan 1990; Laver and Hunt 1992). Empirically, this assumption does have some support as well (Poole and Rosenthal 1991). Thus one can label the preferences of each group as more or less hawkish or dovish, given their relationship to those of the foreign country. Actors whose preferences are closer to those of the foreign country are termed dovish; actors farther away are considered hawkish. The terms hawk and dove are used to simplify categorization of the actors’ preferences; they are not intended as normative or pejorative terms. Rather than saying they are to the left or right of another actor, one can say a group’s policy preferences are more hawkish or dovish than another’s. The structure of preferences, then, varies according to the relative positions of the actors’ preferences.

The Executive’s Preferences versus the Legislature’s: Divided Government

This section discusses why policy preferences might vary among political leaders, particularly between the executive and the legislature. As noted above, political actors have the same basic interest—retaining political office—but may have different policy preferences. A divergence between the policy preferences of the executive and the median legislator creates divided government. The more divergent these preferences are, the more divided government is.

Here divided government is a continuous variable. Most often it is used as a dichotomous one—either the party in control of the executive is the same as the majority in the legislature or it is not. This dichotomous usage is not always helpful since intraparty differences can matter. If the majority legislative party also controls the executive but lacks party discipline and/or is internally divided, then even so-called unified government may appear divided. In addition, though, the use of divided government as a continuous variable allows one to appreciate how both parliamentary and presidential systems may be divided and to understand how degrees of internal division matter. As chapter 3 shows, divided government—and the degree of division—matter for international cooperation. Here I address when the executive’s ideal policy choice will differ from that of the median legislator and what factors will drive them further apart.

Divided government is a term usually reserved for presidential systems. In this context it occurs when the president’s party is not the one in control of the majority in the legislature. In these systems two agents are elected– usually in separate elections—to represent the public: the executive and the legislature. The potential for conflict between these two is elevated when they have different policy preferences. Used dichotomously, this term implies that the preferences of the executive and the legislature differ (e.g., O’Halloran 1994; Lohmann and O’Halloran 1994); party affiliation is employed as a proxy for preferences. As a continuous variable, this term describes how much these preferences differ. It also captures the divergences in preferences that may occur even when the same party controls both but party discipline is low and/or the two agents have divergent preferences because of their different constituencies. In two-party presidential systems, like that in the United States, divided government may be sporadic; however, in multiparty presidential systems, like many in Latin America, legislative majorities depend on coalitions of parties, thus making divided government fairly constant.

Divided government is also an obvious possibility in “semi-presidential” political systems. In these systems a blend of presidential and parliamentary procedures is employed (Lijphart 1984; Shugart and Carey 1992; Baylis 1996). Semi-presidential systems—like the Fifth Republic in France and the current systems in Finland, Russia, and some countries in Eastern Europe—combine a popularly elected president with a cabinet run by a prime minister and dependent on the legislature’s confidence, thus creating two agents to represent the electorate. When the legislative majority comes from the same party as does the president, these systems work much like unified presidential ones. When the executive comes from a different party than the one that controls the legislative majority, divided government emerges in which the conflict between prime minister and president can be intense.

Parliamentary systems are not usually associated with divided government; they should be...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Dedication
• Contents
• List of Tables and Figures
• Acknowledgments
• One: Introduction
• Part One: The Theory
• Part Two: The Case Studies
• Part Three: Conclusions
• Appendix
• Bibliography
• Index