Political survival, policy distribution, and alliance formation Anessa L. Kimball De´partement de science politique, Universite´Laval Abstract Existing work cannot explain why countries form alliances when direct security threats are not a key political issue, though we know countries routinely do engage in that behavior. Countries form alliances to manage the essential problem that they must use finite budget resources to provide social policy and national security; the ‘guns versus butter’ dilemma. States ally to ‘contract out’ national security via the formation of alliance contracts so they can allocate more resources to domestic concerns. Alliances increase the effi- ciency of security policy by providing the same level of security with fewer resources, thus freeing those resources for use in other domains. Not only should alliances form when security threats do not dominate the political agenda, but also domestic political and economic demands will influence alliance decisions. In positing a domestic politics-based explanation for alliance formation, this article argues that increased demands for social policy goods increase the chances of alliance formation as leaders seek greater policy allocation efficiency. The use of a production possibilities frontier illustrates the central argument. Those claims are examined on a sample of all country-years from 1816–2000 using a probit model. Empirical results suggest changes in the demand for social policy goods, operationalized as changes in the infant mortality rate, are an important cause of alliance behavior. Keywords alliances, guns vs butter, political survival Introduction This article proposes a new explanation for why states form alliances. Existing theories of alliance formation fall into a narrowly defined realist perspective, based on the necessary existence of threats, or a large, poorly defined liberal perspec- tive, based on a vague concept of preference convergence (Stein, 1990). 1 This article suggests a domestic politics-based explanation for alliances by linking political survival (Bueno de Mesquita et al., 1999, 2003) and the distributional dilemma (Powell, 1999). Alliance explanations based on domestic political considerations are not new, as others 2 have integrated domestic elements into their research. However, most existing explanations lack a specific mechanism identify- ing what causes states to ally. I identify a causal mechanism arising from the link between a leader’s concern for domestic political survival (Bueno de Mesquita et al., 1999, 2003) and his/her need to efficiently man- age the distributional dilemma (i.e. the ‘guns versus butter’ trade- off) (Powell, 1999). Moreover, I present an empirical measure capturing that mechanism. Finally, I report an empirical analysis of all states between 1816 and 2000 that supports the argument presented here. This large-N analysis, as a first cut, has the benefit of looking broadly at the relationship between my proposed cau- sal mechanism and alliance formation. Therefore, the following research question is posed: how do domestic political demands affect a state’s decision to form a new alliance? Research linking domestic politics to foreign policy behavior 3 Briefly, the central claim is that states outsource some portion of their national security (by forming alliances) in order to allocate more resources to domestic political demands, thus satisfying their minimum winning coalition (MWC) 4 and maximizing potential votes. This is consistent with liberal claims that both domestic politics and efficiency/welfare concerns affect foreign policy. But it also suggests a solution Corresponding author: anessa.kimball@pol.ulaval.ca 1 Research following the realist perspective includes Walt (1987), David (1991), Christensen & Snyder (1990), and Sweeney & Fritz (2004). Research drawing upon the liberal approach includes Simon & Gartzke (1996), Siverson & Emmons (1991), and Lai & Reiter (2000), as well as the ‘security community’ concept (Deutsch, Burrell & Kann 1957). Gibler (2008) draws on both approaches and others but his interest is not in the domestic causes of alliances. 2 See Altfeld (1984), David (1991), Barnett & Levy (1991), Levy & Barnett (1992). 3 The literature linking the internal and external factors relevant in foreign policy decisionmaking is abundant though unexamined here. See Gourevitch (1978), Putnam (1988), Bueno de Mesquita & Lalman (1992), among others. 4 The MWC represents the smallest sized group of individuals whose approval is required so a leader may retain political office (Bueno de Mesquita et al., 1999). Journal of Peace Research 47(4) 407–419 ª The Author(s) 2010 Reprints and permission: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0022343310368346 jpr.sagepub.com