Commitment Problems in Alliance Formation

The forthcoming article “Commitment Problems in Alliance Formation” by Brett V. Benson and Bradley C. Smith is summarized by the authors below. 

Why do adversaries react to anticipated alliance formation or expansion with aggression in some cases but not others? In some cases, such as the 1990 expansion of NATO that accompanied German unification, adversaries acquiesce to allow for peaceful implementation of a new alliance. In other cases such as the 2008 Russian invasion of NATO aspirant Georgia, adversaries respond to the prospect of a new or expanded alliance with military aggression.  

We provide a formal theory that explains this variation as the result of temporal dynamics. Our argument emphasizes that the benefits of a military alliance may not arrive immediately. Rather, prospective allies must coordinate efforts to fully implement the terms of an alliance before realizing its security benefits. In addition to the process of negotiating, signing, and ratifying an alliance, prospective allies must reconcile disparate military technologies and command structures before realizing the full military benefit of alliance membership. This creates a window of opportunity for adversaries to use military force to achieve their coercive goals or to block the alliance altogether before formation or expansion is complete.  

Our explanation rests on three findings. First, bargaining concessions can serve a pacifying role. Prospective allies may make concessions to compensate an enemy state for the future power shift that occurs once an alliance is implemented. This presents a novel explanation for why alliances rarely provoke preventive war: allies may “buy off” shared enemies prior to the implementation of otherwise provocative alliances.  

Our second finding details the conditions under which these peaceful bargains are possible. If the anticipated power shift is too large and the speed of implementation too fast, then allies may not be able to offer concessions sufficient to secure the target’s acquiescence. Consequently, preventive war results. Following this logic, we find that allies are often better off limiting the scope of an alliance’s military benefits or the speed of its implementation to avoid provoking an enemy. The costs of such limits are justified by the benefit of avoiding war.  

Our third finding is that the conditions favorable for preventive war also render it unlikely to occur. Preventive war is attractive when a large power shift from an alliance is expected to arrive rapidly due to quick implementation of an alliance. If an alliance is expected to be implemented rapidly, the window of opportunity for preventive war often closes before an attack can be carried out. Hence, while preventive war cannot be ruled out altogether, the conditions that make it most attractive also render it unlikely.  

Importantly, our findings extend to defensive alliances. Though they are designed to achieve peace through deterrence, anticipation of a new defensive alliance may provoke an enemy by foreclosing an enemy’s future coercive options. Though defensive alliances deter once they are implemented, anticipation of a new defensive alliance may be provocative. We find that this mechanism was present in a number of important historical cases, including instances of NATO expansion as well as the 1954 Taiwan Straits crisis. Overall, our analysis highlights temporal dynamics as a key and novel factor connecting alliance politics and war.  

About the Authors: Brett V. Benson is an Associate Professor of Political Science at Vanderbilt University and Bradley C. Smith is an Assistant Professor of Political Science at Vanderbilt University. Their research “Commitment Problems in Alliance Formation” is now available in Early View and will appear in a forthcoming issue of the American Journal of Political Science. 

Speak Your Mind

*

 

The American Journal of Political Science (AJPS) is the flagship journal of the Midwest Political Science Association and is published by Wiley.

%d bloggers like this: