Assessing the Efficacy of the Laws of War

Regents Of The University Of Michigan - Ann Arbor, Ann Arbor MI

Investigators

Abstract

This project tests a formal model of the laws of war-formal treaties that seek to limit violence during wartime. These international laws are of particular interest because they shape behavior to the extent that the parties can enforce them on one another or respect their limits on their own. They form the extreme case of the self-enforcing nature of political institutions. The project is a formal treatment of war where there is a strategic competition between two states. They contend for a prize in a war of attrition model where the costs they pay each round depend on the current military balance between the two. In each round, they fight a "battle" where a fixed strategic-form, zero-sum represents the contest of each battle. The outcome of the "battle" game depends on their chosen strategies This determines the shift in the military balance in this round. In equilibrium, a side quits if the military balance moves against them beyond its "break point", which is endogenous to the game. A variant of this game has the costs in a given round depend on the strategies played in the battle game as well as the current military balance. In this model the laws of war are represented as a prewar agreement by both players not to play a set of strategies of the battle game. The question is when such agreements hold in equilibrium. A side considering defecting from such an agreement weighs the short-run benefits of using a banned strategy when: a) the other side cannot retaliate; b) the long-run consequences of both sides being free to use any strategy covered by the agreement; and c) any audience costs it suffers for being the first violator of the agreement. The incentive to defect is assessed by how the "break points" of the two sides shift as a consequence of a defection. The project tests these conclusions by collecting a data set on violations of the laws of war in the 20th century. These data include the major treaties in effect, addressing issues like prisoners of war and chemical weapons, and candidates for agreements that were not legalized in a treaty, such as aerial bombardment. Coding rules for extent and degree of violations are created and compliance with these agreements by warring states in the set of Singer-Small wars during the 20`h century are assessed through consultation with historical sources, primarily secondary sources. This data collection leads to two empirical tests of the models: first, a comparison of whether enforceable agreements as judged by the model are actually observed more often than limits judged unenforceable, and second, a test of the timing of violations, whether they tend to fall early during a war and whether the winning side is more likely to be the violator when they occur when it appears that one side is about to lose.