Toward a Multiprocess Model of Rivalry and the Democratic Peace

In this paper we examine recent efforts to combine quantitative research on the democratic peace with research on interstate rivalry. Using Monte Carlo simulations, we highlight problems with separately analyzing the processes associated with rivalry and the democratic peace. Specifically, we specify a multiprocess model and demonstrate that previous research on this topic may overestimate the pacifying effect of democracy on enduring rivalries. Since pairs of democracies are unlikely to experience interstate rivalry in the first place, the true effect of joint democracy is difficult to ascertain in a censored sample of interstate rivals. Our simulation results are consistent with historical data analysis that suggests that the pacifying effect of democracy is most pronounced in the enhanced probability of jointly democratic dyads averting the onset of rivalry. More generally, this article fits into a larger body of research that examines the confounding effect of selection bias on world politics research.