Comment on “COVID‐19’s Lockdown and Crime Victimization: The State of Emergency under the Abe Administration”

Abstract

Shen et al. (2021) investigate the impact of the COVID-19 lockdown on criminal incidents in Japan between April and May, 2020, employing two empirical approaches, the difference-in-differences (DID) and event study approaches, where they set the start of the incidence (treatment) to April. They find that the COVID-19 lockdown reduced violent and property crime by 12.7% and 20.9%, respectively. Shen et al. provide a timely analysis of an important research question, and has at least two important contributions in the literature. First, this is the first study on the impact of COVID-19 on criminal incidents in the Japanese context. COVID-19 and the associated mitigation policies had an enormous influence on society, affecting a wide range of areas, including the economy, politics, culture, and crime. Timely assessments of the impact on crime have an important implication for designing further public policy responses to infectious disease pandemics. The impact of COVID-19 differs across countries, and each government adopted different policy measures. It is thus essential to evaluate the impact in various countries. Second, there is an extensive literature on the socio-economic determinants of crime, and Shen et al. add an important insight to this prior research. Prior studies show that various economic variables affect one's decision to commit a crime, including risk preference, income, opportunity costs, and unemployment (Brosnan, 2018). Shen et al. argue that the voluntary lockdown in Japan positively affected the demand side of crime through various direct and indirect channels. This indicates that COVID-19 would affect different types of crimes differently. Consistent with the theoretical explanations, Shen et al.'s empirical results show that the magnitude and statistical significance of the impact of the COVID-19 lockdown on crime differs according to the type of crime; the largest impact is on breaking-and-entering and there is no significant change in homicide. One of their interesting findings is the heterogeneous impact of COVID-19 by age group on criminal incidents, showing young individuals are the most effected. There would be two other potentially interesting differences in the impact: heterogeneity by region and by gender. The data used in the analysis is Hanzai Tokei (Crime Statistics), which reports the number of criminal incidents by prefecture. I hypothesize that the crime impacts could differ by the number of COVID-19 cases, with a greater impact in prefectures with more cases, as people are more likely to engage in precautionary behavior, including social distancing. Analyzing the impact using prefecture-level panel data sets would provide useful insights to understand the mechanisms and to obtain policy implications, particularly, given the differences in the timing of the state of emergency declarations across prefectures. In addition, utilizing human mobility data from mobile phones is becoming common in COVID-19-related analyses. Using such data would provide a detailed picture of the relationship between social distancing policies and crimes. Furthermore, it is well documented that the effect of COVID-19 differs by gender (Alon et al., 2020), which may result in a difference in crime rates. Given the finding of a decrease in sexual assaults, exploring the impact by gender may provide additional useful evidence to inform public policy to strengthen targeted interventions. One important remark is that Shen et al. assess the short-term impact rather than the medium- to long-term impact of COVID-19 on crime. The impact could differ in the medium- to long-term, as the prolonged impact of the COVID-19 pandemic affected individuals and society differently than it did in the initial phase of the pandemic. In particular, one of the indirect channels of the reduction in crime incidents, the mental health channel, may be different given the reported increase in the suicide rate in the second wave of the pandemic (Tanaka & Okamoto, 2021). Finally, we should note that the analysis is based on police data, which are subject to under- or non-reporting of specific types of crimes, such as serious assaults occurring at home (Boman & Gallupe, 2020; Leslie & Wilson, 2020).