Replication Data for: Asymmetric Crime Dynamics In and Out of Lockdowns

This article studies the dynamic impact of a temporary policy restricting social encounters due to coronavirus disease 2019 (COVID-19) on criminal activity in Bihar, India. Using a regression discontinuity design in time and criminal case—level and arrest data, I document an immediate drop in crime of over 35% due to the lockdown. Analysis over a longer timespan shows asymmetric dynamics by crime type. The lockdown was more effective in preventing personal crimes such as murders but was less effective in preventing property crimes, which increased beyond pre-lockdown levels once the lockdown was lifted. The increase in property crimes seems to be driven by temporal crime displacement from “former offenders” and not by “new offenders.” These asymmetric dynamics across crime types provide new insights into criminals’ intertemporal decisions

本文研究了2019冠状病毒病（COVID-19）背景下，临时性社交接触限制政策对印度比哈尔邦（Bihar）刑事犯罪活动的动态影响。本文采用时间维度断点回归设计，结合刑事案件层级数据与逮捕数据开展实证分析，发现封锁政策实施后犯罪率立即下降超35%。更长时间跨度的分析则揭示了不同犯罪类型间的非对称动态特征：封锁政策在遏制杀人等侵犯人身权利犯罪方面成效显著，但对财产犯罪的管控效果有限；在封锁解除后，财产犯罪率甚至超过了封锁前的水平。进一步分析表明，财产犯罪率的上升似乎源于「既往犯罪者」的临时犯罪转移，而非「新犯罪者」的新增作案。不同犯罪类型间的这种非对称动态特征，为罪犯的跨期决策行为提供了全新的研究视角。