Fighting Abroad, Fighting at Home (and vice versa)

Abstract:

There has been increasing scholarly attention paid to the relationship between civil war and international disputes. While this literature includes a rich set of theoretical expectation, the empirical evidence offered to support them thus far has included several important shortcoming. Most crucially, previous influential models of the effect of civil war on interstate disputes assume that civil war initiation and duration is exogenous from underlying international hostilities. Here, I show that this assumption neither matches the theoretical mechanisms being analyzed, nor is it necessary to bring quantitative evidence to bear on interstices of domestic and interstate conflict. I use special regressor methods, suggested by Lewbel (2000) to account for the cross-level, monadic-to-dyadic, relationship, as well as the potential for endogeneity. I illustrate why conventional single-equation approaches, as well as parametric bivariate probit models produce biased inferences on the effect of civil war on interstate disputes. Using the negative of the log of inter-capital distance as the special regressor, I show that there is an absence of clear evidence for an exogenous effect of civil war on interstate conflict. Instead, more research should explore the role of dynamic international hostility in causing both conflict processes.