Why do armed groups uproot civilians in wartime? This dissertation identifies variation across civil wars in three population displacement strategies - cleansing, depopulation, and forced relocation - and tests different explanations for their use. I develop a new "assortative" theory of displacement, which argues that while some strategies (cleansing) aim to expel undesirable or disloyal populations, others (forced relocation, and perhaps depopulation) seek to identify the undesirables or the disloyal in the first place. When combatants lack information about opponents' identities and civilians' loyalties, they can use human mobility to infer wartime sympathies through "guilt by location." Triggering displacement forces people to send costly and visible signals of loyalty based on whether, and where, they flee. This makes communities more "legible," enabling combatants to use people's movements as a continuous indicator of affiliation and allegiance; and to extract rents and recruits from the population. I therefore show that in some cases, displacement is attractive because it offers unique solutions to information and resource problems in civil wars by acting as a sorting mechanism and a force multiplier.
To evaluate my theory, I adopt a multi-method, multi-level research design that focuses on displacement by state actors, the primary perpetrators of these methods. I introduce a new dataset of population displacement strategies in 160 civil wars (1945-2008), disaggregated by type, and conduct a series of quantitative tests. The data show that strategic displacement has been much more common in wartime than previously thought. I also find that, consistent with my expectations, different displacement strategies occur in different contexts and appear to follow different logics. Cleansing is more likely in conventional wars, where territorial conquest takes primacy, and when counterinsurgents have access to group-level identifiers that link civilians to an armed group. Forced relocation - the most common displacement strategy - is more likely in irregular wars, where identification problems are most acute, and when counterinsurgents lack access to other group-level heuristics for inferring civilian loyalties. The evidence indicates that cleansing follows a logic of punishment. The results for relocation, however, are consistent with the implications of my assortative logic: it is more likely to be employed by resource-constrained incumbents fighting insurgencies in "illegible" areas - rural, peripheral territories - and it correlates strongly with state efforts to mobilize the population into civilian defense forces.
While the cross-national analysis lends indirect support for my arguments, a series of case studies provide direct evidence for the assortative logic. Two in-depth case studies, of civil wars in Uganda and Syria, are based on extensive field research conducted between 2016 and 2019, including hundreds of interviews with perpetrators and victims of displacement. An additional chapter features three shorter case studies of conflicts in Burundi, Vietnam, and Indonesia.
As the most comprehensive study of strategic wartime displacement to date, this dissertation challenges some core assumptions about a devastating feature of modern conflict and an increasingly salient issue in world politics. My findings have important implications for research on forced displacement, civil war, and political violence, and can inform policy efforts to prevent, mitigate, and better respond to wartime migrations and their myriad consequences.
