Domestic turmoil and diversionary hypothesis: A linear relationship?

Why do individual case studies reveal evidence in favor of diversionary behavior by states, yet quantitative studies often produce many mixed results? In this paper, I help to explain this puzzle by arguing that there is a curvilinear relationship between internal conflict and external diversionary behavior as opposed to a linear one. By looking at contentious issues mainly in the Western Hemisphere from 1962 to 2001 using data from the Issue Correlates of War (ICOW) project, I find that countries with moderate levels of domestic unrest are most likely to initiate a militarized interstate conflict. Furthermore, I find that in times of elite unrest, states are more likely to have diversionary behaviors. In contrast, mass unrest is unlikely to incentivize states to use foreign policy adventures to rally domestic support. Such a finding helps deepen our understanding of the complex conditions leading to diversionary actions.