How do armed actors respond to population movements during civil wars? I argue that displacement alters local balances of control between territorial rulers and challengers. Local territorial rulers have incentives to govern violently if displaced persons perceived as members of opposing loyalty groups move into their territories and challengers spoil local governance by inflicting harm on civilians if incoming supporters of a local ruler reinforce the governor’s control. To test these dynamics, I use a combination of manual coding and machine learning to create a novel monthly dataset of territorial control, one-sided violence against moving populations and displacement patterns disaggregated by ethno-religious groups in the war against the Islamic State (2014–2017) in Iraq. Using negative binomial count models, I find that territorial challengers and rulers distinctively respond to population movements in Iraq. The paper extends previous theories of civilian victimization and territorial control by conceptualizing local populations as a dynamic element that explains where fleeing civilians become moving targets.

中文翻译：

---

控制不断变化的世界：伊拉克的领土控制、流离失所和平民攻击的蔓延

武装行动者如何应对内战期间的人口流动？我认为流离失所改变了领土统治者和挑战者之间的局部控制平衡。如果被视为反对忠诚团体成员的流离失所者进入其领土，当地领土统治者就有动机进行暴力统治；如果当地统治者的新支持者加强了总督的控制，挑战者就会对平民造成伤害，从而破坏当地治理。为了测试这些动态，我结合使用手动编码和机器学习来创建一个新颖的每月数据集，其中包含领土控制、针对流动人口的单方面暴力以及针对伊斯兰国的战争中按民族宗教团体分类的流离失所模式（2014 年） –2017）在伊拉克。使用负二项式计数模型，我发现领土挑战者和统治者对伊拉克的人口流动有不同的反应。该论文扩展了先前关于平民受害和领土控制的理论，将当地居民概念化为一个动态因素，解释了逃离的平民在何处成为移动目标。