Replication data for: Territorial Peace and Democratic Clustering. Journal of Politics. 2013

A consistent and robust finding in the democratic peace literature is that democracies tend to cluster together. The explanations for clustering rely on several factors, including democratic demonstration effects and aid from democracies to nascent opposition groups in non-democratic countries. This paper questions the logic of the clustering approach, both theoretically and empirically. Further, we develop an argument predicting democratic transitions based on the level of territorial threat targeting the state: high levels of threat cause political centralization and inhibit democratization; low levels of threat allow for decentralization and democratization. This approach explains how democratic transitions are linked to international borders and imply geographic clustering. Analyses of the post-World War II period are supportive of our arguments even when controlling for clustering-based predictors.