Why do combatants intentionally uproot civilians? The forcible relocation of families and communities to concentration camps, “protected villages” and other special settlements is a regular feature of irregular war, occurring in almost a third of all counterinsurgency campaigns since 1816. Despite the historical regularity of such practices, most research has focused on individual decisions to flee and relocate, rather than the brute-force displacement of civilians by combatants. Using a dynamic model of popular support and micro-level counterinsurgency data from Soviet secret police archives, I show that civilian resettlement is not simply a by-product of war, but is the consequence of rational decision-making by combatants facing acute informational asymmetries. The identification problem of irregular war – the inability to selectively reward cooperators and punish defectors – compels the side with an informational disadvantage to interdict support for the opponent rather than attempt to deter it. Where coercive leverage is limited, resettlement offers a way to dampen the intensity of rebel activity without having to win hearts and minds. Word count: 8,175 (main text & appendix) + 1,581 (notes & tables) + 1,104 (references) = 10,860 ∗Draft. Please do not cite or circulate. The author is grateful to Robert Bates, Volha Charnysh, Timothy Colton, Rich Nielsen and workshop participants at Harvard for helpful comments on earlier drafts. A previous version of this research was presented at the 2013 Peace Science Society International meeting in Savannah, GA. Comments, questions welcome: zhukov-at-fas.harvard.edu.