Assessing the Dyadic Nature of the Democratic Peace, 1918–88

The literature on the democratic peace has emerged from two empirical claims: (1) Democracies are unlikely to conflict with one another, and (2) democracies are as prone to conflict with nondemocracies as nondemocracies are with one another. Together these assertions imply that the democratic peace is a dyadic phenomenon. There is strong support for the first observation, but much recent scholarship contravenes the second. This paper assesses whether the democratic peace is a purely dyadic, a monadic, or perhaps a mixed dyadic and monadic effect. Our analysis offers two important advances. First, our model directly compares the dyadic and monadic explanations by using the state as the unit of analysis rather than the potentially problematic dyad. Second, our model controls for an important but overlooked confounding variable: satisfaction with the status quo. Our results indicate that the initiation of violence within crises is predominantly a dyadic phenomenon, but we also find evidence suggesting a strong monadic effect regarding the emergence of crises.