![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
Abstract
Recent literature suggests that well-intentioned third-party intervention in military conflict can lead to moral hazard by acting as a subsidy to rebellion. In this paper, the following are suggested: failure to intervene can also lead to conflict, which will be called the ‘insufficient deterrence’ problem; and a classification scheme for potential conflicts in the world, which can be used to derive the optimal intervention policy on a case-by-case basis. The optimal policy may be to intervene at random because certainty in some cases will encourage parties who benefit from intervention or parties who lose to attack. The model also suggests that post-conflict aid and a siding-with-the-winner approach may promote the incidence of conflict, even if such policies lower the loss from unavoidable conflicts.
Notes
As discussed in the special issue, there is debate on whether the term ‘moral hazard’ is appropriate to describe the problem. I think that it is indeed, and will follow the precedent. As mentioned in Kuperman Citation(2005), moral hazard has long been used in economics to describe excessive risk taking by countries who expect IMF rescue packages in the case of financial crises (Lane & Phillips, Citation2000, and references therein). Though in some cases the rebel groups discussed in the special issue may not be taking on excessive risk themselves, they make the civilian population take on excessive risk because they benefit from the insurance ‘paid out’ to civilians when there is intervention.
Recently, for example, US intervention stopped Serbian aggression in Bosnia, British intervention helped to end the conflict in Sierra Leone, and failed interventions in Rwanda and Sudan may have allowed the genocides occurring there.
Following Doyle and Sambanis Citation(2006), I use peacekeeping to denote efforts to sustain an established peace. Peacemaking is non-violent efforts to establish a peace. Peace enforcement is establishing a peace by some degree of force, when at least one party to the conflict would otherwise obstruct the peace.
I am very grateful to an anonymous referee for this point.
I do not consider the direct cost of the intervention to include political costs for the government of the intervening country, due, for example, to the risk of becoming unpopular when the lives of domestic soldiers are lost (as happened for the US government in 1993 during the intervention in Somalia): this is a cost for the political leadership, not for the intervening society as a whole. Including such costs would also make it hard to interpret the model: if the intervening government is very afraid of becoming unpopular, this hardly means that the country should not intervene. However, the value of the expected loss of lives should of course be included in the cost.
While I treat both the two groups and the intervener M as unified rational actors, the ability of M to act freely may be questioned in some cases. For example, UN decisions may be mainly the product of member state preferences and actions mediated by an institutional framework. However, as mentioned, the intervener's payoff follows from the normative focus of the paper plus the resource maximization criterion. For discussion of the actorness of intergovernmental institutions, see, for example, Groenleer and van Schaik Citation(2005), Peterson and Smith Citation(2003), Bretherton and Vogler Citation(1999) and Chatzopoulos and Lambridis Citation(2007). On non-governmental organizations, see Martens Citation(2002).
By a ‘realistic expectation’, I do not mean perfect foresight, that is, I do not assume that groups can perfectly foresee the future. Rather, if a group knows that with probability 1/3 each its payoff will be 2, 3, or 4, respectively, then its realistic expectation of its payoff should be 3. Thus, in statistics terminology, each group expects a payoff equal to the objective statistical expectation of that payoff.
However, to resolve this moral hazard problem, condition Equation(1) must not be satisfied, so y must not be too large, unless M can commit not to intervene.
I thank an anonymous referee for this point.
However, to resolve this insufficient deterrence problem, equation Equation(1) must be satisfied, so x must not be too large, unless M can commit to intervene.
Underconfidence, if it occurs (White, Citation1995), has implications opposite to those of overconfidence.