The Commitment Problem in Ad Hoc Humanitarian Intervention

• In Event: Poster Session: Conflict Processes II

Abstract

Some rebel groups, who expect third parties to militarily intervene on their behalf, commit provocative violence while others do not. If armed conflict is costly, an incumbent government that faces increasing demands from a rebel group that is backed by a third party should offer a concession that the rebel is likely to accept. Existing theories have been unable to explain why some rebel groups still provoke the government, despite the possibility of government concessions. We argue that the temporariness of the distribution of power that is produced by humanitarian intervention causes a commitment problem. It leads the rebel to commit provocative violence to induce the government retaliation followed by humanitarian intervention before the “window of opportunity” closes.

While rebel leaders sometimes (or at least in some cases) have the incentive to provoke the government in the shadow of intervention, scholarly debates on the specific mechanism that leads to this perverse incentive have yet to reach a consensus. Some argue that humanitarian intervention causes the “moral hazard” problem (Kuperman 2008). Drawing on the contract theory in economics, proponents explain the perverse incentive in humanitarian intervention as an intentional risky behavior by an insurance policyholder, i.e., the rebel, to get insurance coverage, i.e., intervention.

However, this logic does not apply to humanitarian intervention because the moral hazard explanation is dyadic. The issue of humanitarian intervention involves three players: a rebel group, the third party, and the incumbent government. If the government is a strategic player, it will make concessions that are large enough to dissuade the rebel; costly conflict should be avoided, even when the rebel is emboldened by a powerful third party. With the possibility of concessions by the government, the changing balance of power due to the prospect of humanitarian intervention alters the terms of the settlement but not the probability of a bargaining breakdown (Cetinyan 2002). Thus, the moral hazard perspective does not explain the logic of perverse incentives in humanitarian intervention.

In this research, we argue that the expectation of humanitarian intervention causes a commitment problem. The government’s concession does not “absorb” the rebel’s increasing demands because the former cannot promise not to exploit the latter. The root cause of the commitment problem is the temporariness of humanitarian intervention. By temporariness, we mean that intervention in a situation in which the change in power distribution between the government and the rebel due to outside intervention is not likely to last long after the dispute is settled and the third party withdraws. Humanitarian intervention tends to have a myopic perspective without an institutional design for the post-war period, and it lacks strategies for peacebuilding. As a result, the gap between the rebel’s expected payoff of war with support from the third party and that of the post-crisis period after the dispute has been settled as well as the third party's exit make it difficult for the incumbent government to commit to abiding by the terms of the settlement.

To demonstrate this, we developed a game-theoretic model in which the incumbent government, a rebel group, and the third party interact through three periods: the pre-crisis period, the crisis period, in which the third party chooses whether to commit to humanitarian intervention should war erupt, and the post-crisis period. In this model, the government and the rebel can either go into a costly conflict or refrain from attacking the opponent, and humanitarian intervention is assumed to increase the probability of the rebel’s victory in war. We present an equilibrium, where the rebel uses provocative violence if the gap between the expected levels of intervention of the crisis period and that of the post-crisis period is large. Empirically, NATO’s intervention in Libya illustrates this logic.

This study contributes to the literature in two ways. This research firstly provides a coherent alternative to the moral hazard explanation for paradoxical violence in the shadow of humanitarian intervention. While existing studies have not sufficiently considered the temporariness of the impact of intervention and the commitment problem, we shed light on a new aspect of the unintended effect of humanitarian intervention. In addition, this is the first game-theoretic analysis of perverse incentives in humanitarian intervention that presents a model with complete information. While some scholars have developed formal models to theoretically identify the conditions that determine the effects of intervention, to the best of our knowledge, war erupts only after incomplete information is introduced in their models. This research advances our understanding of how intervention sometimes leads to bargaining breakdowns, even if players have common knowledge about their preferences.

Author

• Hiroto Sawada, National Institute for Defense Studies