Despite a great deal of empirical investigation in the last 30 years, little systematic support has been produced for the diversionary hypothesis, or the notion that national leaders facing domestic discontent seek to generate popular rallies by using military force abroad (Ostrom and Job 1986). Scholars have advanced several potential explanations for this state of affairs: that armed diversion simply generates too many political risks and costs (e.g., Meernik 2000); that it fails to produce politically useful rallies with regularity (e.g., Lian and Oneal 1993); and that more direct ameliorative measures hold out greater hopes of success (Clark 2003). Two recent advances, however, suggest that the dearth of findings is as attributable to variability in interstate conditions and individual leadership attributes as to general diversionary ineffectiveness. First, the "strategic conflict avoidance" (SCA) perspective (Smith 1996; Leeds and Davis 1999; Fordham 2005; Foster 2006b; 2008) demonstrates that potential international targets increase cooperation toward and avoid initiating conflict against would-be antagonists when the latter's diversionary incentives are strong. Second, work based in leadership psychology shows that leadership traits that are associated with greater risk taking, conceptual simplicity, and aggressive responses to constraints are crucial intervening variables that increase the likelihood of diversionary behavior by particular leaders (Keller and Foster 2012; Foster and Keller 2013). These two promising lines of new research have important theoretical implications for each other, but have not yet been brought together.
The primary purpose of the current project, simply put, is to integrate the expectations of the SCA and "first-image" psychological approaches to the study of diversionary conflict. Because leadership traits are largely observable in political rhetoric and are likely to mold the perceptions people form of leaders, foreign leaderships should be able to develop expectations about the willingness of certain counterparts to challenge political obstacles and constraints and, by extension, the likelihood that they will utilize diversionary conflict strategies. In particular, following from Keller's (2005) framework, we hypothesize that American presidents whose spontaneous rhetoric indicates high task emphasis, a need for power, a distrust of others, and unabashed nationalistic proclivities should be more likely to engage in diversionary conflict, should be seen as more likely to do so by potential targets, and should thus be treated in a more conciliatory fashion when diversionary incentives are operative.
Time-series analysis of the American foreign policy experience for the period 1953-2000 largely bears out these expectations. The primary independent variable, which is an interaction of the economic misery index (unemployment plus inflation) and an "at-a-distance" scale measure of constraint challenging proclivity revealed in presidential rhetoric, is found to be negatively and significantly associated with the initiation of escalatory militarized incidents (as measured in the militarized interstate dispute [MID] data set) against the United States (1953-2000). Moreover, this interaction term is positively and significantly associated with increasingly cooperative diplomatic behavior by enduring interstate rivals (using the World Event/Interaction Survey [WEIS] events data) in the period 1966-1992. Marginal-effects analyses reveal that presidents whose aggregate previous rhetoric has revealed strong constraint-challenging tendencies are targeted for fewer militarized incidents and more rival cooperation as the misery index rises, while strong constraint respecters are on the receiving end of more conflictual diplomatic overtures by enduring rivals as the misery index rises.
The article proceeds as follows. First, the considerable literature on diversionary force usage is addressed, focusing particularly on the SCA and first-image perspectives' explanations for broadly inconsistent findings. Next, an effort is developed to theoretically integrate these two perspectives. Discussions of the research design and analysis follow, and the article concludes with suggestions for future work.
The Diversionary Hypothesis
The diversionary hypothesis rests upon the assumption, widely held among political analysts for centuries, that national leaders are able to enhance internal unity during periods of domestic discord by fighting external enemies (Macchiavelli 1532 ; Bodin 1576 ; Simmel 1896 ; Coser 1956). As such, when indicators of poor leadership performance (or domestic dissatisfaction with leadership performance) are strong, the hypothesis predicts that the likelihood of international conflict initiation by that state's leadership increases (Waltz 1967; Mueller 1970; Stohl 1980). Beginning with a seminal study by Ostrom and Job (1986), contemporary quantitative scholarship has conducted dozens of inquiries into the influence upon international conflict propensity of such domestic factors as civil war, rebellion, and terrorism (e.g., Gelpi 1997; Miller 1995; Davies 2002; Gleditsch, Saleyhan, and Schultz 2008; Foster 2015), economic weakness (e.g., Russett 1990; Wang 1996; Fordham 1998a; Mitchell and Moore 2002), and low public approval (e.g., Morgan and Bickers 1992; DeRouen 1995; Morgan and Anderson 1999).
Interestingly, this broad inquiry has produced little in the way of consistent evidence, either in support or contravention of the hypothesis (a trend noted as early as Levy ). Scholars have reached opposite conclusions about the hypothesis' veracity in both transnational studies (Miller 1995; Mitchell and Prins 2004) and even in studies of the same country (e.g., the United States; Gowa 1998; DeRouen 2000). Several works uncover potential diversion from economic distress (e.g., Fordham 2002), but certainly not all do (Meernik 2000); the ledger appears to be about even in regards to domestic violence (Trumbore 2003; Oakes 2012); and several works raise serious doubts about whether low public approval spurs diversionary incentives (e.g., Foster and Palmer 2006). Even those cases seemingly representing exemplars of diversion have been the subject of further scrutiny by scholars who raise plausible questions about the solidity of the evidence favoring a domestic politics explanation for interstate war (Oakes 2006; Fravel 2010).
What explains this state of affairs? Foster and Keller (2013), in their review of the literature, demonstrate that the vast majority of reasons advanced for the inconsistency of findings pertain to the array of domestic-level (or "second image"; Waltz 1959) impediments to and risks of diversionary strategies. Several scholars note that rallies associated with the use of force are too unpredictable for leaders to rely on them to alleviate domestic problems (Brody and Shapiro 1989; Norrander and Wilcox 1993; Oneal and Bryan 1995; Parker 1995) and that the types of (severe) international crises that engender meaningful rallies are rare and result from processes other than leaders' domestic political needs (Meernik 1994; Mitchell and Prins 2004). Political ideology could also dampen diversionary payoffs, with leaders whose core constituents are ideologically averse to military action or for whom direct "fixes" to domestic problems would resonate ideologically, identifying fewer diversionary incentives (Fordham 1998b; Palmer, London, and Regan 2004; Foster and Palmer 2006). Indeed, the barriers to realizing meaningful political payoffs to diversion have led some scholars to doubt whether diversion is ever a more fruitful strategy than directly addressing even the most difficult domestic problems (Clark 2001; 2003). In terms of political risks, most obviously, diversion can directly threaten a leader's credibility and tenure if their strategy is uncovered by the public (Meernik and Waterman 1996; Colaresi 2007). Moreover, like all military ventures, armed diversion involves the risk of unforeseen consequences: losing a conflict that one is expected to win or losing a conflict that leaders chose explicitly to revive perceptions of their managerial competence are both very real possibilities that would ultimately make any diversionary "cure worse than the disease" in the domestic setting (Richards et al. 1993; Bennett and Nordstrom 2000).
While the caveats raised by each of these lines of reasoning are welcome contributions to the study of diversion, we focus here specifically on two other approaches that focus on factors beyond domestic conditions and institutions. The first of these, the SCA perspective, stipulates that diversionary incentives are noted both by leaders facing domestic difficulties and by their potential international targets. Smith's (1996) game theoretic treatment stresses that the presence of domestic problems in a given state should induce potential targets to avoid engaging in behavior that could be portrayed as provocative. Indeed, it is possible that this interstate process does not simply scuttle the execution of diversionary strategies; the presence of diversionary incentives may lead other states to actively avoid conflict, perhaps by decreasing conflictual behavior and/or increasing cooperative behavior toward the would-be diverter. Leeds and Davis (1997) provide cross-national empirical support for this proposition, in that economic decline in Western industrialized democracies is negatively correlated with the placement of international demands by other states. Likewise, Fordham (2005) finds that increased unemployment and decreased gross domestic product growth in the United States is correlated with increased cooperation by the United States' historical rivals in the international arena, and Foster (2006b; 2008) finds that militarized incident initiation by all states against the United States...