The displacement of civilians is a frequent, yet understudied, outcome of armed groups’ and civilians’ behavior during civil wars. In particular, I find that displacement as a strategy of armed groups is an especially undeveloped area of study, and argue that a focus on mass displacements as the dependent variable can provide leverage for explaining variation in outcomes over time and across space. I suggest that three sets of factors explain the variation: armed groups’ goals, competition among armed groups, and community governance mechanisms. After outlining the theory and hypotheses, I consider the implications of the approach for appropriate units of analysis. With data on displacement in the Colombian civil war, I use both events of massive displacement and municipal population flows as indicators of the dependent variable to consider the plausibility of the framework’s empirical implications. Finally, I propose additional qualitative, micro-level research strategies to enable tests of the mechanisms underlying the theory