Why do inefficient civil wars happen when Coasian bargaining is possible? One theory for that is that players have limited commitment to the implementation of the terms of such bargains. This paper provides conditions under which, given the assumption that Coasian bargains are possible, such commitment levels are identifiable in the data. More explicitly, if there is variation in control of resources under peace and under war in the data, one can identify commitment to Coasian bargains. When applying the model to panel data for 133 countries between 1975-2004, the results indicate that a model of limited commitment is a poor explanation for the country with the average observables in the data. For about 90 percent of my sample, commitment constraints do not seem to be binding, and these countries face a probability of war of around 2 percent. For the rest of my sample, commitment constraints are binding and these country-years face a probability of war of around 80 percent. These conclusions have the important caveat that they might be sensitive to the type of heterogeneity considered.