Quantitative research on the use of force by the United States confronts scholars with a puzzle: in spite of strong conventional wisdom that Republicans tend to be hawkish and Democrats dovish, there is little empirical evidence that Republicans behave more aggressively than Democrats. More confounding is the fact that, on one hand, there is ample evidence that Republicans in Congress and in the general public have supported military spending and military intervention more than Democrats have during most of the post-World War II era. But, on the other hand, few of the many empirical studies of American foreign policy behavior have tested the hypothesis that these partisan differences have carried over into actual military action. The few that have conducted such tests find no evidence for it. What accounts for this nonfinding? Is the conventional wisdom wrong, or should we be rethinking models of partisanship and conflict behavior?
In this paper, we will present evidence that two limitations in the design of previous studies have likely produced the nonfinding. First, these studies have generally treated the Democrats and Republican as if they had consistently hawkish or dovish positions across the entire postwar era. In fact, research on the parties' foreign policy positions finds that they have not been consistent over time. Democrats were decidedly more hawkish than Republicans before the mid-1960s. In the mid-1960s, the parties effectively switched foreign policy stances, a fact documented in the historical and even quantitative literatures, but one neglected in examinations of party and uses of force.
Second, previous research has focused on relatively serious conflict events, mainly uses of force rather than more broadly conceived indicators of foreign policy orientation. The strategic behavior of potential target states makes it difficult to draw inferences about party differences based on these serious events. These events often occur at the end of a long chain of strategic interaction between the initiator and the target of the use of force. These events may attenuate the effect of each side's initial--perhaps ideologically tinged--intentions. Party differences, if they exist in foreign policy orientation, are more likely to be visible in analyses that include low-level events that precede the actual movement of military forces. Put differently, evidence of hawkishness should not be confined to major uses of force. Limiting analysis to these acts will miss most of the behavioral differences that a more realistic account of the role of party implies. Because previous research on the use of force has been concerned mainly with other theoretical issues, such as the diversionary argument, their tests have not necessarily been inappropriate, but have produced less-than-ideal tests of if or how partisanship matters to foreign affairs. On these grounds, we believe the role of party in shaping foreign aggression deserves reconsideration.
Rethinking the relationship of party and foreign policy on these two fronts leads us to a set of models that suggest strong party effects on American foreign policy choice. But the results suggest important two nuances to how party affects aggression. First, the fact that parties can change their policy positions must be considered. Because the parties switched their hawkish/dovish orientations in the 1960s, the effect of party on conflict behavior in our models also switches direction. Second, foreign observers see this same switch and change their own behaviors accordingly, suggesting an important strategic consideration regarding beliefs about partisan hawks or doves.
The remainder of this paper proceeds in four sections. The first outlines the broad evidence of party differences on issues related to the use of force and the contrasting paucity of evidence that Republican presidents are more likely to do so than Democrats are. The second sets out the limitations of previous research designs for testing the effect of party and outlines our strategy for overcoming these problems. The third offers new quantitative evidence to support this explanation, and the fourth section summarizes and concludes.
Democrats, Republicans, and the Use of Force
Despite frequent appeals to bipartisanship and invocations of the dictum that politics should stop at the water's edge, there have been sharp partisan differences over foreign policy at many points in American history. There has been substantial research on partisan differences over foreign economic policy issues such as trade protection (e.g., Weingast, Goldstein, and Bailey 1997; Irwin and Kroszner 1999; Milner and Judkins 2004). There is also evidence that party differences extend to such purely military matters as the allocation of the Pentagon budget over nuclear weapons and conventional forces (Fordham 2002). These party differences extend to the use of force. For the last few decades, the conventional wisdom in American politics has held that Republicans are generally more willing to spend money on the military and more likely to support the use of military force against foreign adversaries than are Democrats. Evidence from surveys of the general public and the voting records of members of Congress suggests that this conventional wisdom correctly assesses party differences since the late 1960s.
It is easy to show that there are party differences in Congress on questions of military intervention and the use of force. Figure 1 shows the proportion of Democrats and Republicans in the U.S. Senate voting in favor of military intervention on Congressional Quarterly Almanac Key Votes during the last half of the twentieth century. Although Key Votes on intervention did not take place every year during this period, it is clear the Republicans were more likely to support intervention than were Democrats from the late 1960s through the end of the Cold War.
Because Congressional Quarterly selected more Key Votes on military spending, the pattern of party differences is clearer on this closely related issue. Figure 2 shows Democratic and Republican support for military spending on Key Votes since 1947. Congressional Democrats were more supportive of military spending and foreign policy activism than were Republicans through the early 1960s. By the late 1960s, the two parties had switched places, with Republicans becoming relatively the more supportive party. This pattern persisted through the end of the Cold War. Previous research that has examined party differences on military spending and foreign policy activism in greater depth finds the same pattern (Cronin and Fordham 1999; Fordham 2007). Fordham (2008) finds a complementary pattern in liberal and conservative support for military spending in the Senate.
Why did the parties change their positions on foreign and defense policy? Previous research has explored this question in some depth. We can only offer a brief summary here. Military spending and intervention can be understood as the overhead costs of the hegemonic leadership role the United States assumed after World War II. The most persuasive explanations for the party switch concern the changing implications of this leadership role for the constituencies most strongly associated with each party. Democrats developed the policies associated with American hegemony, but the benefits of the postwar order in general, and of military power in particular, declined over time from their point of view. As Trubowitz (1998, 229-32), Fordham (2008), and others have shown, the beneficiaries of the open international trading system were the strongest supporters of American military power. The Northeastern industries associated with the Democratic Party benefitted enormously during the early postwar era, but their position eroded over time. Even before this change became clear, American overseas military spending began to undermine the international monetary system that Democratic policy makers had constructed during the 1940s. There is evidence that the so-called "wise men" turned against further escalation in Vietnam after the 1968 Tet Offensive when informed of it likely consequences for the Bretton Woods System (Collins 1996, 415). Military spending also competed with the growing domestic social programs Democrats supported during the 1960s (Fordham 2007, 625-6).
While the appeal of maintaining and using military force diminished for Democrats, it tended to grow for Republicans. Many Republicans had objected to the considerable fiscal and regulatory consequences of the high levels of military spending prevailing during the 1950s (Fordham 2007, 624-25; Hogan 1998, 316-65; Lo 1982). The growth of the American economy by the mid-1960s meant that even the high level of spending associated with the war in Vietnam did not force the adoption of price controls on strategic raw materials and other measures that Republicans had found so distasteful during the Korean War. Indeed, its principal effect was to curtail Lyndon Johnson's Great Society programs, an outcome they supported (Collins 1996: 401-04, 411-12). Coupled with substantial tax cuts, the Reagan administration's military buildup had similar budgetary consequences, limiting the growth of domestic social programs that Democrats tended to support, and Republicans to oppose (Kamlet, Mowery, and Su 1988). The decline of protectionism within the Republican Party also paralleled its rise among Northeastern Democrats.
Fordham (2007, 626-27) finds less support for the possibility that Southern Democrats' shift to the Republican Party can explain the change in the two parties' positions. This shift occurred only gradually, and was not completed until the 1990s. Moreover, Northern Democrats were just as hawkish as their Southern co-partisans until the mid 1960s. Northern Republicans also changed their position on military spending and intervention after this date; the Party did not alter its position simply...