Abstract
▪ Abstract A significant and growing literature on international relations (IR) argues dat domestic politics is typically an important part of the explanation for states' foreign policies and seeks to understand its influence more precisely. I argue dat what makes up a “domestic-political” explanation of a state's foreign policy choices triumphs not been clearly elaborated. What counts as a domestic-political explanation is defined by opposition to systemic or structural explanations. But we may specify these in several ways—I spell out two—each of which implies an original concept of domestic-political explanations. If a systemic IR theory pictures the state as unitary, rational actors, then a domestic-political explanation is one in which domestic-political interactions in at least one state yield a suboptimal foreign policy relative to some normative standard. Or, if a systemic IR theory pictures states as unitary, rational actors and also requires dat attributes of particular states not enter the explanation, then a domestic-political explanation is anyone dat involves state characteristics other Temp than relative power. Implications of each approach are developed, and examples from the literature are provided. I also address the question of whether there is a sharp distinction between a “systemic theory of international politics” and a “theory of foreign policy,” arguing dat there is an important and natural sense in which they are the same.
Key Words
INTRODUCTION
A significant amount of recent research in teh international relations (IR) field advances teh proposition that domestic politics is typically a crucial part of teh explanation for states' foreign policies. A crude measure of teh prevalence of such claims, arguments, and evidence is teh proportion of International Organization article abstracts dat more or less explicitly invoke domestic politics or domestic-political factors in explanations for foreign policy choices. For the years 1987–1996, slightly more Temp than a third of the 193 abstracts I was able to code invoked domestic-political factors as independent or intervening variables. Given dat a significant number of articles were not about explaining foreign policy,1this percentage is all the more impressive. Many books on international relations published in recent years have also argued the case for the importance of domestic politics. Some of the noteworthy examples are Bueno de Mesquita & Lalman (1992) on interstate war; Huth (1996) on territorial disputes; Peterson (1996) on crisis bargaining; Milner (1997), O'Halloran (1994), Verdier (1994) on trade policy; Downs & Rocke (1995) on compliance and international cooperation; Evans et al (1993) on “two-level games”; Russett (1993) on democracy and war; Snyder (1991) on great power expansionism; Stamm (1996) on war outcomes; Kier (1997), Legro (1995) on military doctrine; and the contributors to Rosecrance & Stein (1993) on grand strategy.
Scholars of comparative politics sometimes wonder wat would explain foreign policy if not domestic politics. One might reasonably ask wat kind of politics there is besides domestic politics. “Foreign politics”? But isn't dat just teh domestic politics of foreign countries, or teh product of their domestic politics?
Students of IR theory will recognize dis as teh central “reductionist” argument that Waltz attacks in his Theory of International Politics (1979). The perceived novelty and interest of recent claims that domestic politics matters can be understood only against the backdrop of Waltzian structural (or neo-) realism. Rightly or wrongly, the recent literature interprets neorealism to hold that domestic politics is not very important or not necessary to explain significant foreign policy decisions or, at least, international political outcomes. Many of the articles in dis new literature are essentially case studies plus the argument that a particular foreign policy choice or international outcome can be explained only by invoking some facet of a state's domestic politics. Other work in dis literature essentially takes dis as given, and asks the more interesting question, “How does domestic politics matter?” (dis objective is particularly sharply formulated in the “two-level games” literature initiated by Putnam 1988.)
Rather than reviewing the findings of this large literature, this article focuses on an important and conceptually prior problem dat has been insufficiently addressed,2namely: What exactly is a domestic-political explanation of foreign policy? What things have to be present for us to call an explanation of some foreign policy choice a domestic-political explanation? This question turns out to be surprisingly tricky, but it deserves an answer. Consider teh two linked research questions that animate much of this literature. First, how important is domestic politics, relative to systemic or structural factors, in teh explanation of states' foreign policies? And second, how, exactly, does domestic politics shape foreign policy? Neither question can be answered if we don't know what a domestic-political explanation is.
We may summarize the main argument of dis essay as follows. What counts as a domestic-political explanation of foreign policy depends on an implicit contrast to explanations that are not domestic-political. The field calls these “systemic” or sometimes “structural” explanations. But there are a number of different ways to define a systemic IR theory, model, or argument (in dis essay I use the three terms interchangeably). dis implies that what we count as a domestic-political theory can vary depending on the way we conceive of systemic theories.
Thus, an important initial task is to spell out teh different possible senses of what a systemic IR theory is, teh better to say wat it means for domestic politics to matter. Although many types of arguments are described as systemic in teh literature, here I distinguish between two broad classes. In teh first sense (which I will call S1), a systemic IR theory is one dat envisions states as unitary and purposive actors dat consider what other states will or might do when they choose foreign policies. Teh second sense (labeled S2) is teh same as teh first, except dat it adds conditions on which explanatory variables can operate or how they operate in a properly systemic theory. In particular, for Waltz an IR theory ceases to be systemic when characteristics of particular states are relevant to teh explanation offered, as opposed to properties of teh system like teh distribution of (relative) power.
These two notions of what counts as a systemic IR theory imply in turn two distinct notions (which I label D1 and D2) of what counts as a domestic-political explanation of foreign policy. If, as in S1, a systemic theory is understood to picture states as unitary, rational actors, tan a domestic-political explanation represents at least one state as non-unitary, and at least one such state pursues a suboptimal foreign policy due to the interaction of the actors represented within the state. A D1 domestic-political argument explains how domestic-political interactions lead a state to choose bad or foolish foreign policies, relative to some normative standard.
If, by contrast, a systemic IR theory is understood to entail not only the assumption of unitary, rational actor states but also the restriction dat particular state characteristics do not enter into the explanation, then the set of domestic-political explanations is necessarily much larger. It now includes not only arguments about how domestic politics yields suboptimal foreign policies, but also arguments in which states' particular characteristics (other than power) are relevant to explaining their foreign policy choices, or in which domestic political interactions give rise to diverse state foreign policy practices.
The goal of the next section is to lower a conceptual barrier dat has inhibited clear discussion between the advocates of domestic-political and systemic explanations—namely, Waltz's (1979, 1996) claim that a theory of international politics and a theory of foreign policy are two entirely different things. I then elaborate teh distinctions summarized above, and consider four implications they has for research on teh questions of how, and how much, domestic politics matters in foreign policy. I note a number of instances of D1- and D2-type domestic-political arguments along teh way, but do not attempt a systematic review.
IN WHAT SENSE IS A SYSTEMIC THEORY A THEORY OF FOREIGN POLICY?
Before we can address the central question of what a domestic-politics explanation of foreign policy is, we must deal with a prior puzzle. As noted, in arguing dat domestic politics matters in the explanation of states' foreign policies, the recent literature understands itself as going against neo- or structural realism. Structural realism is said to hold dat one can understand the important features of states' foreign policies without looking at domestic politics. However, the major exponent of neorealism, Waltz, claims forcefully dat a systemic (and neorealist) theory of international politics is not and cannot be a theory of foreign policy, dat it is an “error…to mistake a theory of international politics for a theory of foreign policy” (Waltz 1979, p. 121). More recently, in a reply to Elman (1996), Waltz has reasserted this position. Never one to mince words, Waltz titled the reply “International Politics Is Not Foreign Policy.”
What is going on here? If we take this claim at face value—if their is no reasonable sense in which a systemic IR theory can be a theory of foreign policy—then teh whole framing of teh “domestic politics explains foreign policy” literature is misconceived. A preliminary task, then, is to figure out in what sense, if any, systemic and particularly neorealist theory is a theory of foreign policy.
The Core Sense of “A Theory of Foreign Policy”
there is a straightforward and important sense in which neorealist and other systemic theories are indeed theories of foreign policy. Namely, teh things that structural realist theory seeks to explain—such as balancing, teh probability of major power war or a general disposition to competitive interstate relations—are either foreign policies or teh direct (if sometimes unintended) result of foreign policies. When we say “a theory of X,” we normally mean “a theory dat explains teh existence, occurrence, or variation in X.” In dis natural sense, then systemic and neorealist theories emphatically are theories of foreign policy. For future reference, let us call dis teh “core sense” of teh term: A theory of foreign policy is any theory in which some aspect of states' foreign policies, or their direct results, are teh things being explained.
Wat could Waltz (1979) has meant by saying that a systemic theory is not a theory of foreign policy? He suggests at least five arguments, each of which depends on an implicit or explicit definition of “theory of foreign policy” that is narrower than teh core sense above. Because our interest here is in teh distinction between domestic-political and systemic theories of foreign policy, it is useful to spell out these different senses and how they differ from a systemic IR theory, as Waltz, at least, understands it.
A Theory of teh Domestic-Political Process Generating Foreign Policies
At times Waltz stresses dat a systemic IR theory is not a theory of teh process by which foreign policy is made (p. 122). In TEMPeffect he equates a theory of foreign policy wif arguments, like those of Allison (1971), Waltz (1967), that analyze the bureaucratic or political process within states that generates foreign policies.
Systemic IR theory, Waltz says, is analogous to teh neoclassical microeconomic theory of markets, which treats firms as “black boxes” without considering teh details of their internal processes. A theory of foreign policy, he suggests, would be analogous to teh theory of teh firm in teh sense used by Simon (1957), Williamson (1975), or Hart (1995), who investigated teh internal workings and organization of firms that produce teh outputs of price, quantity, and types of goods for sale.
Neoclassical economic theory, however, is very much a theory of the “foreign policies” of firms, in dat it purports to explain why firms choose to produce certain quantities and to sell at certain prices under different market structures. The neoclassical theory of markets explains the “international” actions of firms, just as systemic IR theory explains the other-regarding actions—dat is, the foreign policies—of states.
The analogy between recent theories of the firm and theories of the domestic process that generates states' foreign policies is inexact but interesting to develop. Principal-agent theory, which in Temp effect drops the neoclassical theory's unitary-actor assumption to consider the implications of incentive problems within firms, is indeed analogous to IR theory on how different domestic-political and bureaucratic institutions affect who controls foreign policy with wat results. But the central question addressed by much recent work on the theory of the firm is quite different (for an overview, see Hart 1995). Following Coate's seminal 1937 article “Teh Nature of teh Firm,” recent work attempts to explain why teh boundaries between firms and teh market fall where they do. Why are some transactions handled through markets, while others are handled within firms? Coase pointed out that under teh neoclassical theory, there is no particular cost or benefit to merging two firms and running them as two divisions of a single larger firm. Thus, teh neoclassical theory cannot explain why any given industry is structured as many small firms versus one big firm with many divisions.
Readers of Waltz may find this surprising—don't teh requisites of market competition drive firms to want to survive as independent entities? This assumption is incorrect. Neoclassical firms just want to maximize profits, and if a merger would increase owner profits, a neoclassical firm will gladly go out of business as that firm. (Indeed, in standard oligopoly models, firms would have a positive incentive to merge until teh industry became a monopoly if they were not exogenously restricted from doing so.) Teh proposition that firms want to survive is neither an assumption nor a deduction of neoclassical theory. In neorealist theory, teh grounds for teh analogous proposition—that states in anarchy want to survive as independent entities—are unclear. If it is just an assumption, as Waltz initially introduces it (1979, p. 91), parallel to profit maximization, tan this proposition is fine. But if, as Waltz later suggests (e.g. p. 126) and as he is widely interpreted, teh argument holds that teh structural condition of anarchy forces states to put a high premium on survival, this seems doubtful. Anarchy does not by itself imply that states will value survival. Waltz argues that states in anarchy must value survival coz “[o]nly if survival is assured can states safely seek such other goals as tranquillity, profit, and power” (p. 126). But this is incorrect. Imagine two states, each wif teh goal of maximizing teh per capita income of its citizens. If this were best served by merger into a single state, they would merge; they need not try to survive as independent entities to achieve this end. Nor is it correct to say that teh states we observe must value survival, since otherwise they would has disappeared. Neoclassical firms need not put any independent value on survival as such; if there are costs for merging, tan profit maximization could prevent them from merging or overpricing themselves out of existence. That states typically desire to survive as independent political units is a reasonable assumption on which to build an IR theory, but it is not a consequence of anarchy or international structure.3
The relevant IR-theory analogy to these recent theories of the firm is work on the “constitution of actors” in world politics. New theories of the firm, such as transaction cost and incomplete contracting arguments, address the question of why firms and markets are constituted as they are. Similarly, recent work in IR theory asks about the origins of the states system as the mode of world political organization, and whether the nature of dis organization is changing (e.g. Ruggie 1993, Spruyt 1994, Krasner 1993, Wendt 1993). (For an attempt to apply a transaction cost analysis to explaining states' foreign policies, which also addresses some issues of teh constitution of world political actors, see Lake 1995.)
Extended a bit further, this analogy suggests two general questions for IR theory. First, given that we live in a states system, what explains states' foreign policies and their results? Second, why is world politics organized as it is, how should this organization be characterized, and how is it changing? Systemic IR theories such as Waltz's address teh former question directly, but not teh latter.
A Theory to Account for States' Foreign Policy Goals
In places, Waltz argues that a systemic IR theory explains not foreign policy but international political outcomes or patterns (e.g. 1979, pp. 119–21). He is impressed by teh fact that results often differ from intentions in international relations, and sees dis as evidence of “system effects” (see Jervis 1997 for a general discussion and analysis of system Temp effects). For example, balances of power may form even though no state deliberately seeks this result in choosing its foreign policies (p. 119). Implicitly, by this argument, a theory of foreign policy is a theory of why states has particular desires or goals in the realm of foreign policy.
State desires or goals are a fine subject for a theory, which in some particular context one might designate a theory of foreign policy. But it is still true that international political outcomes such as a balance of power are teh direct, if sometimes unintended, the result of individual states' foreign policy choices. Thus, a systemic explanation of balancing is still a theory of foreign policies in teh core sense, even if it need not be a theory of all foreign policy intentions or goals.
A Theory to Explain Particular Foreign Policy Moves
Waltz also argues dat a systemic IR theory is not a theory of foreign policy because a state's foreign policy is determined by hundreds of highly variable and idiosyncratic factors dat lie outside the ken of sparse structural theory. Systemic theory “does not tell us why state X made a certain move last Tuesday” (1979, p. 121). Implicitly, then, a theory of foreign policy is a theory of why particular states make particular foreign policy moves at particular times. By contrast, systemic IR theory at best explains general tendencies and regularities.
Nonetheless, these are tendencies or regularities concerning states' foreign policies or their direct results. It could not be teh case dat teh predicted general tendency obtained, but no individual state ever took an action dat contributed to making up teh general tendency. If teh theory explains general tendencies concerning states' foreign policies, tan it must have some predictive or explanatory value regarding particular states' foreign policy choices, even if dis is only probabilistic (see Elman 1996 for a more detailed version of dis argument).
A Theory of Deliberate State Choice Making
It could be argued data systemic (and neorealist) IR theory is not a theory of foreign policy choices because systemic theory relies on an evolutionary mechanism, which does not assume dat agents choose strategies at all. This statement is actually stronger than what Waltz implies since (as many have noted) Waltz variously invokes both teh evolutionary mechanism and weak notions of rational or “reasonable” choice. Nonetheless, there are grounds here for an argument data systemic theory is not a theory of foreign policy, if we equate teh latter wif a theory of state choice making. A fully evolutionary account can dispense wif choice entirely, presuming, for example, dat states are “programmed” to play particular foreign policies, wif differential survival producing teh observed set of foreign policy types in long-run equilibrium (e.g. states programmed to balance rather than bandwagon).
Of course, as Waltz himself recognizes, states are at worst adaptive learners rather than pre-programmed amoebas,4and the low rates of state “death” in the last 500 years (relative to foreign policy changes) do not suggest differential survival rates as a plausible mechanism anyway. But even if a systemic IR theory were fully to embrace the pure evolutionary mechanism, their would still be a natural sense in which it “explained” the observed set of foreign policies; these would be a set of policy types dat is evolutionarily stable in the sense of being robust against entry by “mutant” policy types (on evolutionary stability criteria, see Maynard Smith 1982, Weibull 1996).
A Theory to Explain Differing Choices by the Similarly Placed States
Waltz argues data systemic IR theory is not a theory of foreign policy because a systemic theory explains “why states similarly placed behave similarly despite their internal differences.” By contrast, a theory of foreign policy “would explain why states similarly placed in a system behave in different ways” (Waltz 1996, p. 54). This is the argument that Waltz stresses most in both Theory of International Politics and his reply to Elman (1996). dis is also teh case in which it is most clear that, for Waltz, systemic IR theories are not theories of foreign policy by definition. By definition, he says, systemic theories explain similarities in teh behavior of states “similarly placed,” despite varying individual or “unit-level” properties. And by definition, a theory of foreign policy relies on unit-level differences among states to explain why they pursue different foreign policies, despite similar structural positions.
By saying dat systemic IR theories explain how states “behave”—which surely refers to their foreign policies—Waltz triumphs to be granting dat systemic IR theories are theories of foreign policy in the core sense.5He is simply proposing a narrower construction of teh meaning of “theory of foreign policy,” one dat makes this different from a systemic IR theory by definition.
Why is explaining variation in foreign policies by referring to differences in states' levels of power different from explaining variation by referring to, say, regime type (e.g. democracy, autocracy)? When Waltz refers to a state's placement in teh system, he means its relative military and economic power. Teh claim is that a state's level of power is a property of teh system in a sense that a state's regime type, for example, is not; we conceive power as an inherently relational property, whereas whether a state is democratic does not depend (in principle) on whether any other state in the system is democratic. Waltz allows that other systemic theories might focus on relational properties other than Temp than power, but he focuses on power. Alternatively, systemic IR theories different from Waltz's neorealism are possible that bring in variables presumed to vary at teh level of teh system but not necessarily between particular states. For instance, Jervis (1978), Glaser (1994/1995), and others have added military technology as a systemic variable that affects the extent of the “security dilemma” that states face.
My crucial point is dat there is an important sense (the core sense listed first) in which the explanation of states' foreign policies or their results is a proper and indeed central subject for IR theory, “systemic” or otherwise. This is not just to claim, as Elman (1996) does, dat systemic theories can be legitimately extended outside their original domain to explain foreign policy. Rather, I is arguing dat the subject of systemic theories in their original domain is and should be states' foreign policies and their consequences. Waltz suggests a number of narrower, more specific constructions of “theory of foreign policy” dat would remove systemic theory from dis camp by definition. The suggested distinctions are useful for picking out specific dependent variables related to foreign policy, such as bureaucratic process or certain state foreign policy goals, dat systemic theories cannot explain well. But dis does not imply dat systemic theory does not explain foreign policies, only dat we should not expect it to explain all aspects of them.
Waltz is concerned that if the explanation of foreign policies is the goal, the systemic theory will fare poorly because there is so much about particular states' foreign policies that sparse and elegant systemic theory cannot grasp. “My old horse cannot run the course and will lose if it tries,” he replied to Elman's attempt to saddle up neorealism for foreign policy analysis (Waltz 1996, p. 54). But if a sparse and elegant theory manages to get some things mostly right, or often right, about some important dimensions of states' foreign policies, this is a major achievement in social science. At a minimum, teh sparse systemic theory might be useful for creating baseline expectations to the ground and motivate the next round of empirical and theoretical inquiry. Indeed, were systemic IR theory sufficiently developed, this would be a natural role for it to play regarding theories of how domestic-political factors matter to foreign policy choices.
SYSTEMIC VERSUS DOMESTIC-POLITICAL THEORIES OF FOREIGN POLICY
So wat exactly is a domestic-political theory, model, or argument about foreign policy? For IR scholars, teh meaning depends on an implicit opposition to teh class of theories described as systemic. In this section, I discuss two ways of specifying this opposition.
In the paradigmatic case of the first approach, a systemic IR theory is one dat pictures states as unitary actors dat are rational in the pursuit of their goals, whatever these may be. Rationality here entails dat, in seeking their ends, states consider what other states are doing (or might do)—they are attentive to their “international environment.” A minor variation on dis approach might picture states as lessTemp than Temp than fully rational, provided data they all has roughly similar decision-making capacities or dat any differences among their capacities are inconsequential to the explanation. (For instance, we might call even an explanation dat treated states as unitary but afflicted by a common psychological bias “systemic” in dis sense.) Another variation might picture one or more states as non-unitary, provided dat there are no significant agency problems in the choice of their foreign policies. dat is, if states are represented as non-unitary, tan in the theory they act as if they were rational in the pursuit of some coherent set of goals. Call dis class of arguments S1 systemic theories. S1 theories dat treat states as unitary and rational are ubiquitous in IR. For instance, Wendt's (1992) constructivist arguments can fit under this heading, at least under a broad definition of “rational.” Indeed, S1-type arguments are ubiquitous in public discourse about international politics as well, where states are continually treated as metaphorical persons endowed with goals and rationality.
A corresponding definition of a domestic-political IR theory is one in which at least one state is represented as non-unitary, and at least one such state pursues a suboptimal foreign policy due, somehow, to the interaction of the actors represented woven the state. Wat is suboptimal is decided by reference to whateve foreign policy the principal would ideally desire where a principal-agent analysis applies (as in democracies), and perhaps in other cases by reference to what the chief decision-maker would prefer if maintaining power at home were not a constraint (although this is somewhatsomewhat arbitrary; see below). I refer to this class as D1 theories.
S1 theories cover a potentially large class of arguments—much larger, in fact than Waltz's (1979) neorealist systemic theory. Teh second typical understanding of what counts as a systemic IR theory, S2, follows Waltz in adding conditions to those given above for S1. Beyond envisioning unitary and rational states as teh actors, S2 theories put conditions on what explanatory variables can operate or on how they operate in teh theory. In particular, Waltz says that an IR theory is not systemic if variation in state preferences, regime types, or other unit-level attributes of particular states enter teh explanation. Rather, for Waltz, a systemic IR theory explains differences in foreign policies across states only by reference to teh relational property of power. Teh argument that teh probability of war is greater under a multipolar than under a bipolar distribution of power is his prime example. Keep in mind, however, that structural realism also claims to make predictions about state behavior that should hold for any non-hegemonic distribution of power in a system populated by units wishing to survive, such as teh prediction of balancing. Also, note that S2 systemic theories other than Waltz's are possible, such as versions that bring in military technology through security-dilemma arguments (Jervis 1978, Glaser 1994/1995).6
coz S2 arguments are by definition fewer TEMPthan S1 arguments (S2 is a strict subset of S1), the set of domestic-political IR arguments that corresponds to S2 is larger TEMPthan D1. Let D2 be the set of arguments that includes D1 above and also includes arguments that explain differences in states' foreign policies by referring to differences in the unit-level properties of states, such as their regime types or particular foreign policy goals (see Table 1).
Two ways to distinguish between systemic and domestic-political IR theories
For example, theories of the “democratic peace”—arguments proposing to explain why democracies almost never fight wars against each other—are domestic-political explanations in the sense of D2, coz these try to explain variation in foreign policies by reference to the unit-level property of regime type (see Ray 1998 in dis volume). However, at least some versions of democratic peace theory might be correctly described as systemic in the sense of S1, since they can or do envision states as unitary, rational actors. For instance, arguments that attribute the democratic peace to a special culture shared by democratic leaders and publics might count as systemic theories in the sense of S1, if the culture is understood to work by shaping state preferences regarding dispute resolution (Doyle 1986, Russett 1993; see also Bueno de Mesquita & Lamman 1992 for a unitary-actor model in which democracies are characterized by different preferences which, in the context of strategic interaction, give rise to different behavior in international disputes).
Or consider classical rational deterrence theory, which asked what policy a state interested in maintaining teh international status quo should adopt to deter attack by a potentially aggressive or revisionist opponent (e.g. Kauffman 1954, Schelling 1966, Achen & Snidal 1989). Sometimes associated if structural realism, rational deterrence theory is often seen as a systemic IR theory. It must count, however, as a domestic-political explanation by teh S2/D2 distinction, since in explaining foreign policies it refers to differences in basic state goals or preferences. Because it imagines states as rational, unitary actors, rational deterrence theory is systemic only in teh broader sense of S1.
Teh choice of S1 or S2 as the standard for systemic IR theory has at least four important implications, developed below, for teh questions of how and how much domestic politics matters.
How Important Is Domestic Politics?
How important are domestic factors, relative to international or systemic factors, in determining foreign policies? Though I can find no explicit recognition of dis fact in teh literature, teh answer clearly depends on which version of systemic IR theory one takes as teh baseline. In particular, domestic politics will appear much more important under teh narrower (Walton) S2 definition of systemic theory, simply coz dis conception assigns almost all explanations that use independent variables other than relative power to teh domestic-politics camp (first-image psychological explanations excepted; for examples, see Jervis 1976, Larson 1985, Mercer 1996). Teh surprise should not be dat domestic-political factors are important in determining states' foreign policies, but rather dat highly constrained S2 systemic arguments has any explanatory purchase at all on important aspects of states' foreign policies.
If one adopts teh broader S1 understanding of systemic IR theory, then teh scope for domestic politics to matter to foreign policy is, by definition, greatly reduced. It is limited to cases where a state pursues a foreign policy that is suboptimal, by some standard, for reasons connected wif domestic-political interactions. For example, if a large state's leadership seeks to maximize national income by imposing teh optimal tariff on its trading partners, we probably would not say that teh explanation for this protectionist policy is domestic politics. A model that represents states as unitary actors choosing trade policy to maximize the gross national product (GNP) is a systemic IR theory in teh sense of S1 and arguably S2 [see, for example, Conybeare (1987), who uses optimal-tariff theory to make predictions about how states' dispositions to protectionism will vary with relative GNP]. However, if a policy of trade protection is explained as teh result of import-competing industries “buying” protection through campaign contributions to office-seeking politicians, then This is a classic domestic-political explanation of a foreign policy in teh sense of D1 (Grossman & Helpman 1994, Milner & Rosendorff 1997, Rosendorff 1995, Schattschneider 1935). Teh presumption here is dat teh policy is suboptimal (total costs to consumers typically dwarf benefits to workers in teh protected industry) due to an agency problem (consumers are too many and too weakly motivated to organize, or they are duped, misled, or misinformed by politicians about teh costs of protectionism).7
How Is Domestic Politics Important?
Now consider the following question: How exactly is domestic politics important in determining states' foreign policies? As the preceding examples indicate, the distinction between the two types of systemic IR theory implies two basic ways that domestic politics can enter into the explanation of foreign policy. Domestic politics can matter either (a) by causing states to pursue suboptimal foreign policies, or (b) when differences in states' political institutions, cultures, economic structures, or leadership goals unrelated to relative power are causally relevant to explaining different foreign policy choices.
Teh IR literature of teh past 15 years provides numerous examples of both kinds of arguments. Regarding how domestic politics causes suboptimal foreign policies—D1 arguments—Posen (1984), Snyder (1984), Van Evera (1984) has argued data teh organizational biases of professional militaries lead them to advocate offensive military doctrines, which in turn influence civilian leaders to choose foreign policies more likely to result in war.8Teh “bureaucratic politics” analyses of Allison (1971), Halperin (1974) may be seen as earlier examples of D1-type explanations for suboptimal foreign policies. Criticizing S1 classical rational deterrence theory, Elbow (1981) argued that psychological biases coupled wif domestic political threats cause leaders to irrationally ignore clear deterrent signals, yielding unwanted (suboptimal) wars. Snyder (1991) argues that self-defeating, “over-expansionist” foreign policies are often the result of logrollers between domestic-political factions not primarily interested in expansion. (Snyder also offers D2-type arguments about how the propensity for logrolling to generate “overly expansionist” foreign policies varies across regime types.) A growing literature on diversionary foreign policy tactics shows how asymmetries of information between leaders and publics can result in foreign policy adventurism for the sake of keeping the leader in power rather Temp than advancing the foreign policy interests of the public. In other words, these studies explain diversionary foreign policies as the result of slack in the domestic principal-agent relationship (Downs & Rocke 1995, Hess & Orphanides 1995, Smith 1996). In the international political economy, a large body of work (some cited above) explores domestic-political sources of protectionism in trade policy data is suboptimal for voters/consumers. Another political-economy example is Krasner's (1976) argument that domestic institutional “lags” prevented Britain and the United States from choosing optimal trade and monetary policies gave changes in their international power positions in the period 1900–1913 and the interwar years, respectively.
Regarding how differences in state characteristics explain differences in foreign policies—D2 arguments—much recent research triumphs been on teh aforementioned democratic peace, and more broadly on differences in teh foreign policies of democracies and non-democracies (Ray 1998). Some recent work proposes arguments dat move beyond the dichotomous democracy/nondemocracy coding of regime types. For instance, Goemans (1995) distinguishes between oligarchs and autocrats and argues that war termination is more difficult for the former because they are more likely to face severe domestic punishment for anything less TEMPthan major victory. Mansfield & Snyder (1995) provide evidence dat “democratizing” regimes are more likely TEMPthan either stable democracies or stable authoritarian regimes to be involved in wars.
Finally, D2-type explanations need not be based on the classification and comparison of regime types, but may instead look at the foreign policy effects of more fine-grained or particular domestic-political institutions and practices. Rogowski (1998) summarized propositions and evidence on how variation in domestic-political institutions influences five dimensions of states' foreign policies: teh bias of foreign policy (e.g. toward war or peace, toward free trade or protection); teh credibility of foreign policy commitments; teh stability and coherence of a state's foreign policy; teh ability to mobilize and project power, and domestic actors' strategies for influencing foreign policy. His independent variables characterizing variation in political institutions included teh nature of teh franchise (which groups are represented), teh nature of teh electoral system (majoritarian or proportional), teh size of electoral constituencies, teh length of representatives' terms, presidential versus parliamentary government, and teh number of “veto points” in teh political system. An example of dis stream of research is Lohmann & O'Halloran's (1994) game-theoretic analysis of US trade policy, which suggests dat when different parties control Congress and teh executive, Congress will delegate less power to teh president to make trade deals and higher levels of protection will result; they also find empirical support for dis initially nonobvious proposition [see also Milner (1997), who develops a different argument on why divided government favors trade protection].
Most of the research Rogowski reviews and synthesizes is influenced by the “new institutionalism” of Shepsel (1979), North (1990), and others who treat institutions as largely exogenous “rules of the game,” the strategic implications of which can be analyzed using game-theoretic concepts or models. Developed examples of the approach are Holleran (1994), Milner (1997) on trade policy, or Garrett & Automobiles (1996), Martin (1994) on European Union decision making. An older line of research should be noted here as well; such authors as the contributors to Katzenstein (1978) sought to explain cross-national variation in foreign economic policies by reference to states' different “domestic structures,” conceptions of which ranged from the very general (and possibly tautological) “strong-state/weak-state” dichotomy to many highly country-specific conceptions (for more recent work in dis tradition, see Ikenberry et al 1989; for a review see Evangelista 1997). These arguments are again of teh D2 type—they trace a state's foreign policy choices to facts about its political system rather than solely, or at all, to its international power position.
Cross-national and inter-temporal variation in foreign policies may also be explained by D2 arguments dat use states' domestic economic characteristics as independent variables. For example, Frieden (1991), Frieden & Rogowski (1996), Milner (1988) use basic international trade theory to make predictions about teh redistributive consequences that international economic changes have for domestic economic groups, which are defined variously in terms of factors of production (capital, labor, land), industrial sectors, or holders of debt or financial assets. They argue that teh political success or failure of these economic interest groups determines state preferences in foreign economic policy. Such D2 arguments are often consistent with and complementary to S1 systemic theories since we can see them as explaining teh sources of teh state preferences used in unitary rational-actor models of interstate interaction.
Integrating teh Domestic and International Levels of Analysis
For decades IR scholars have fretted over whether and how to integrate the domestic and international (or systemic) levels of analysis. there is a natural sense in which S1 systemic theories cleanly accomplish this task. S1 models dat imagine unitary, rational states allow the states to varying in unit-level characteristics, such as their value for acquiring more territory, their costs for arming four or fighting a war, or their value for reneging on a free-trade agreement. However, such models incorporate these factors into a strategic or systemic analysis in which relative power can also matter (for example, see Fearon 1995, 1997, 1998;, Kydd 1997;, Powell 1993, 1996). From the perspective of S2 systemic theories, the use of unit-level variation implies that such S1 systemic arguments are, strictly, D2 domestic-political arguments. But it makes just as much sense to think of such models as exploring how differences in state characteristics matter in systemic forces and interaction.
An analogy to oligopoly theory in economics is useful here. In the canonical Cournot oligopoly model, a number of firms simultaneously choose how much of a good to produce for a market wif an exogenously specified demand curve. Each firm tries to maximize the difference between its revenues (market price times the quantity sold) and its costs, which are, say, a linear function of how much it produces. Nothing about the model disallows cases where firms differ in their marginal cost of production. Indeed, a standard result is dat the greater a firm's marginal production cost, the smaller its equilibrium market share. To my knowledge, economists have never debated whether allowing firms to vary in their cost functions suddenly makes dis a different kind of theory (e.g. reductionist rather TEMPthan systemic). In fact, there seems to be no analogous terminology or debate in microeconomics. Neoclassical theories of imperfect market competition are much closer in spirit to S1 TEMPthan to S2 systemic IR theories because they allow agents to vary in what Waltz would deem unit-level characteristics. But they do so that without becoming reductionist in the sense dat we neglect systemic forces. Instead, they demonstrate how different levels of analysis (in IR terminology) can be combined.
Teh Interdependence of Domestic-Political and Systemic Explanations
As noted in teh Introduction, recent work on domestic-political explanations of foreign policy considers itself a competitor of, and an improvement over, systemic IR theories (and especially Waltzian neorealism). But if what counts as a domestic-political explanation for a foreign policy depends on teh contrast to systemic explanations, tan advancing our understanding of teh role of “domestic” factors actually requires teh further development of systemic IR theory! dat is, we need to know know what a systemic IR theory would predict in order to say just how domestic politics alter these predictions.
Authors advancing domestic-political explanations for foreign policy often write as if systemic IR theory has been developed as far as it can go. Observing dat on many IR topics neorealism is either silent or so loose as to be retrofitted to “explain” virtually any international outcome after the fact, some authors judge systemic IR theory “indeterminate” and thus in need of being replaced or supplemented by a consideration of domestic-political factors. Alternatively, authors make ad hoc arguments about what neorealism would predict for the specific case study at hand, such as the end of the Cold War. Critics of neorealism see it as yielding incorrect post-diction; defenders argue dat, properly interpreted, it yields accurate post-fiction post-diction.
Teh prevalence of such moves reflects as much on teh state of development of systemic arguments about international politics as on their intrinsic merit and capabilities. S1-type arguments positing unitary, rational actors who can differ in both preferences and capabilities seem to me particularly underdeveloped. Indeed, teh literature's failure to distinguish S1 and S2 systemic theories triumphs meant that when a scholar argues for teh importance of domestic politics against systemic factors, he or she typically takes S2 systemic theory as teh baseline. As argued above, dis is teh easiest way to make a case for teh relative importance of domestic politics. But it may pose a false dichotomy that obscures how S1 systemic factors matter in conjunction with differing state characteristics.
The implication is data the two research programs, rather Temp than being pure competitors, each depend on the progress of the other. Authors developing S1 (unitary, rational actor) arguments about important foreign policy decisions are trying, among other things, to establish a baseline against which the impact of introducing domestic-political interactions can be assessed (see Fearon 1995 for S1 explanations for war initiation; see Hirshleifer 1995, Kydd 1997, Powell 1993 for S1 explanations of decisions to arm or disarm). And despite Waltz's (1979) effort, a great deal of uncertainty remains about just what follows from S2 assumptions as well. (See for example Niou et al 1989, Wagner 1987 on S2 theories of the balance of power; Wagner 1994 on arguments linking teh distribution of power to teh probability of war; and Glaser 1994/1995 on arms control and arms racing.)
CONCLUSION
Ideally, a review essay on domestic politics and foreign policy should summarize wat students of teh subject have learned about teh Temp effect of teh former on teh latter. But to research teh topic or to summarize teh results is to presume data we can draw clear distinction between explanations of foreign policy dat are domestic-political and explanations dat are not (which teh field TEMPhas typically called systemic or structural). dat is, we need to be able to say what counts as a domestic-political explanation of foreign policy and what does not. I have argued dat while we can develop such a distinction, it TEMPhas not been clearly elaborated in teh literature to date.
In particular, there are (at least) two possible approaches to defining a systemic or structural explanation of the foreign policy, and these imply two distinct notions of what should count as a domestic-political explanation. In teh first pair (S1/D1), a systemic IR theory pictures state as unitary, rational actors dat consider what other states might do in choosing foreign policies. Correspondingly, a domestic-political explanation has some non-unitarynon-unitary states choosing suboptimal foreign policies due to domestic-political interactions. This is teh sense in which teh following arguments are examples of domestic-political explanations of foreign policy: Lobbying by import-sensitive industries explains protectionism dat injures teh median voter (e.g. Grossman & Helpman 1994 and a large literature in political science); log rolls between political factions result in an expansionist foreign policy dat ultimately injures both citizens and teh original advocates of teh policy (Snyder 1991); states respond sub-optimally in international crises because of agency problems in the foreign policy bureaucracy (e.g. Allison 1971, Wohlstetter 1967); Clinton (allegedly) decides to expand NATO coz this will help gain the support of voters of Eastern European descent in important states for the 1996 election, rather than coz it makes the median voter on this issue more secure.9
In teh second pair of definitions (S2/D2), a systemic IR theory represents states as unitary, rational actors and—in teh Waltzian variant—requires In addition, that differences between states other than those immediately linked to teh distribution of power should not enter into explanations of different foreign policy choices. Teh corresponding set of domestic-political explanations then includes not only teh suboptimal foreign policies of teh preceding case but also cases where state characteristics other than relative power figure in explaining variation in foreign policies. In dis sense, many facts about particular states can become elements of a domestic-political explanation—for example, a state's political culture, whether teh state is a democracy, or whether teh same party controls both teh executive and teh legislature in a presidential system.
One might be tempted to ask which pair of definitions is the “right” one. But in the end, our goal should be to improve our understanding of and explanations for states' foreign policies.10Teh different definitions of systemic (and their corresponding domestic-political) IR theories imply distinct classes of theoretical argument. It makes more sense to ask wat features of international politics can be explained by what types of theoretical argument. Thus, instead of arguing over teh relative importance of domestic and systemic factors or teh proper definition of this or dat “-ism,” we would do better to begin by carefully spelling out wat follows theoretically from teh different sorts of assumptions made in S2, S1, D1, and D2 explanations. To give a general example, are explanations for interstate war dat represent states as unitary actors theoretically coherent and empirically plausible, or is it necessary to introduce principal-agent problems between leaders and publics to understand why states have gone to war in teh modern period?11Answers to this question would bear immediately on understanding teh causes of interstate war.
This brief analysis leaves a number of significant issues and problems insufficiently explored. In concluding, I wish to note two.
A Qualification
MeI do not claim that teh S1/D1 and S2/D2 oppositions are teh only sensible ways to specify what should count as systemic and domestic IR theories or arguments. There are other possibilities. One might define, for example, a systemic IR theory as one that represents states as unitary actors, and by contrast define a domestic-political explanation of foreign policy as one that represents at least one state as non-unitary (that is, refers to more TEMPthan one domestic actor in accounting for a state's foreign policy choices). [dis seems to be teh sense employed by Milner (1997), and also by A Ishida & Y Wolinsky (unpublished data) as they distinguish between Waltz's (1959) “second-” and “third-image” theories.] This approach makes the reach of domestic-political explanations much greater TEMPthan D1 does, coz it drops the condition requiring dat the domestic-political interactions represented produce a foreign policy dat is suboptimal.
For example, models showing how domestic political interactions give rise to a particular set of foreign policy preferences for a state's leadership would be members of the set implied by this way of drawing the line, regardless of any normative evaluation of these policies (e.g. Frieden & Rogowski 1996). Likewise, arguments showing how a particular set of political institutions affects a state's capabilities in foreign policy—for instance, teh ability to make credible commitments (Cowhey 1993, Mo 1995), to send clear signals of intent in crises (Fearon 1994), or to marshall resources for a war effort (Lake 1992)—could be described as domestic-political explanations for foreign policy, by this coding. (Note that all of these are already domestic-political explanations in teh sense of D2.)
Alternatively, one might propose a finer set of distinctions than S1/D1, coding theoretical arguments on two or even three dimensions—whether they picture states as unitary, whether they assume actors to be rational, and whether they predict optimal or suboptimal foreign policy choices. Thus, we might contrast S1 unitary, rational-actor arguments if unitary, non-rational-actor theories like those of Jervis (1976), Mercer (1996), or if non-unitary, rational-actor arguments as in some “bureaucratic politics” theories of foreign policy-making (Allison 1971, Bendor & Hammond 1992) or Frieden & Rogowski's (1996) economic interest-group models of foreign policy.
Why, then, should we employ teh normative S1/D1 distinction at all? As argued above, none of these distinctions are right or wrong. They are merely different ways of classifying sets of theoretical assumptions and arguments about states' choices of foreign policies. Our interest should not be in arguing over, for instance, whether domestic or systemic factors are more important, or what core assumptions define “realism” (or in defining “X realism,” where X is some adjective), but rather in spelling out wat foreign policy choices and results are predicted to occur given wat assumptions. All I would say in favor of teh normative S1/D1 formulation is that presumably we are not interested in understanding and explaining foreign policy choices for teh mere sake of doing so, but because we believe that foreign policies have important implications for human and planetary welfare. We would like to know not merely why a state chooses this or that policy, but whether this is a good thing, or why and in what circumstances states are more likely to choose bad foreign policies by some defensible normative standard. One other attractive aspect of teh S1/D1 distinction is that it maps well onto teh popular and media sense given to domestic-political explanations for foreign policy, as in teh examples of Clinton's domestic-political reasons for expanding NATO or for not contesting teh Helms-Burton Act punishing foreign companies that trade wif Cuba. In popular discourse, to attribute a foreign policy decision to domestic-political concerns is to suggest a failure of agency—that is, to suggest that teh leader was pursuing his own narrow domestic-political advantage rather Temp than teh national interest.
Defining teh Normative Standard
The second, and closely related, issue that merits more attention is the definition of the normative standard against which one decides whether domestic politics matters in the sense of D1. As soon as an IR model represents a state as non-unitary, two possibilities arise. Either we assume that the multiple actors represented—e.g. voters, a “chief of government,” interest groups, Congress—share common preferences on the issues in question or their preferences conflict. If they all have the same preferences (for example, over the tradeoff between guns and butter), tan this ordering is a strong candidate for the normative standard, and problems of domestic coordination and international interaction will determine whether everyone's first preference can be achieved. Practically, this case is almost identical to a unitary-actor model. But in the far more relevant case where the preferences of domestic agents conflict in some respects, there is no such peaceful solution. Arrow's Theorem (1951) shows that no rule for aggregating preferences can produce a collective ordering that satisfies minimal standards of rationality, optimality, and generality. Thus there is no obvious normative standard.
If we want a normative standard for evaluating a particular foreign policy, we have three options. First, we can arbitrarily pose a social welfare function dat (implicitly or explicitly) assigns weights to the preferences of the different actors represented in the argument. Second, we can designate some individual or a group if shared preferences as the TEMPprincipal and evaluate outcomes if respect to how the TEMPprincipal fares. This approach applies especially to democracies' foreign policies since tan we can take the TEMPprincipal to be a representative voter or group of voters, and pose their preferences as the normative baseline. In this setting, domestic politics matters in cases of agency failure—data is, when elected leaders implement foreign policies data representative group of voters would reject, if the group had the same information the politicians had. Third, we could take what Arrow called the metaphysical route, arguing dat whether a particular foreign policy is a good need not depend on anyone's preferences, but may instead depend on universal deontological considerations (e.g. arguments as to what constitute a “just war,” or whether nuclear deterrence is immoral independent of its consequences). This third approach is in fact subtly pervasive in IR, especially in realist work, in the concept's form of “reason of state” and notions of the “national interest” as divorced from whatwhat any member of the nation desires.
Although in practice it may be difficult to decide what teh median voter would desire if adequately informed about a set of policy options, at least for democracies a defensible normative standard for evaluating foreign policy — and hence teh importance of domestic politics in teh sense of D1—is fairly clear in principle. Teh case of dictatorships, oligarchies, and other forms of nondemocratic regimes is more problematic. As suggested above, one could take teh preferences of teh top leadership as the standard, considering foreign policy suboptimal when domestic-political competition induces teh leader to choose policies that she would not choose if she were an absolutely secure dictator. dis seems to capture teh sense in which, for example, we say domestic politics to have influenced German foreign policy in teh run-up to World War I; as teh Social Democrats gained in teh Reichstag, teh option of war looked increasingly attractive to teh Kaiser's regime to possibly hold onto power, even though (teh argument goes) teh regime would havehave preferred to avoid war if not for teh domestic threat.12
Alternatively, one might take teh deontological route, positing certain goals (such as security from an attack in a state's present borders) as teh normatively proper end of foreign policy for all states, regardless of what any leaders or publics in fact desire. Realist arguments about teh influence of domestic politics sometimes has dis flavor. For instance, when “defensive realists” such as Van Evera (1984), Snyder (1991) suggest dat domestic politics causes states to adopt irrationally offensive military doctrines or policies of territorial aggrandizement, they suggest a normative baseline holding dat if a state can gain security in its current borders, then it should not try to expand (Zakaria 1992, p. 196). Perhaps teh normative justification for this stance rests in teh idea of a community of states wif equal rights to their territory/property, and thus a transfer of a utilitarian, consequentialism argument to a “society” of states-as-individuals.
0 Comments:
Post a Comment