Wars of Succession

ABSTRACT

I analyze a model of bargaining, war, and endogenous leadership turnover in which (1) leader attributes affect war outcomes and (2) war can insulate settlements from renegotiation. Shifts in bargaining power caused by leadership turnover are endogenous and discontinuous, but sufficiently decisive war outcomes can solve the associated commitment problem. In contrast to other models where the shadow cast by a hawkish successor encourages moderation toward a dovish incumbent, the foreign state attacks instead—despite a dovish incumbent’s known preference for peace—using war to lock in a settlement that would otherwise be lost to future leadership turnover. I discuss the theory’s implications for widening the empirical scope of the commitment problem explanation for war to limited wars over relatively lower stakes and for integrating the politics of leadership turnover with the study of strategic rivalries.

KEYWORDS:

• Bargaining
• commitment
• leaders
• war

Acknowledgments

Thanks to Phil Arena, Kyle Beardsley, Cliff Carrubba, Terry Chapman, Amy Liu, Pat McDonald, Dan Reiter, Toby Rider, Emily Ritter, Harrison Wagner, Amy Yuen, three anonymous reviewers, and the editor of II for helpful comments and suggestions.

Notes

¹ Note that a “hawk” here is not a “chicken hawk” that makes frequent cheap-talk statements a willingness to use force. Hawkishness merely indicates a set of preferences or characteristics that leads some national leaders to enjoy better odds of victory against the same opponent than more dovish leaders.

² I treat ’s effect on war outcomes in reduced form, but I have also solved a more complicated version of the model with endogenous war efforts, where leaders differ in the marginal costs of effort, in which the main result is substantively identical.

³ In contrast to Chadefaux’s (2011) model, the terms of the first-period bargain are unrelated to future military strength, but it would be possible to add this feature to the model. Suppose that yielding more of the pie both helps stay in office and transfers military resources to her state. This occurs in a peaceful exchange, so there remains a chance that still takes office to make use of those resources. As long as the more hawkish leader will make different use of these new resources in the ways posited by the theory, then still prefers to face the dove over the hawk in the second period. Therefore, war would still occur in equilibrium and for the same reasons it does in the present model.

⁴ I have analyzed a version of the model in which players value the future at rate , and the results remain substantively identical, requiring in addition only that be sufficiently high.

⁵ This is not the only way that war might solve the commitment problem, but focusing on the costs of a future war is analytically useful. Suppose, as an alternative, that defeat in a war reduces a state’s future chances of winning. If this were the case, then the results of the next section would be similar (if more complicated to present), but it would violate parsimony by adding a dimension. In the current formulation, separating the effect of leader characteristics from others is easy, because leader characteristics are the only driver of war outcomes. The alternative would make it harder to identify the leader’s effect without changing why war happens in equilibrium.

⁶ Leventoglu and Slantchev (2007) show that war may also solve commitment problems by destroying enough of the goods being fought over, which by decreasing the value of fighting is similar to the decisiveness mechanism proposed here.

⁷ This model can be extended to an infinitely repeated setting, where the relevant state variables are the identity of state ’s leader, whether a war occurred in the previous period, and which side won the war. After solving the problem of how long war locks in settlements, one could enrich the political survival function by adding a potential switch between hawkish and dovish factions in state . This is an interesting avenue of inquiry that I leave for future research.

⁸ This assumes that ’s best possible deal is , or that preemptive appeasement as characterized by Wolford (2012:523–525) is not possible. A similar inequality in which , adjusted for the assumptions of this model, can also show the existence of preventive war, but in the interests of avoiding redundancies I present the simpler of the two inequalities here.

⁹ Chadefaux (2011:242) shows that when rising states are insufficiently patient, they may refuse even large concessions that would stave off attack if Powell’s (2004) inefficiency condition is not otherwise satisfied. In the present model, this means that war might occur even when the necessary conditions in Proposition 1 are not satisfied. But such a war would occur due to ’s political survival incentives, not because of shifting power.

¹⁰ Plataea was a former Athenian colony sacked by Thebes and Sparta two years earlier.