This paper highlights the criminogenic consequence of the remarkable slowdown in China’s exports in recent years. By applying textual analysis to millions of judgement documents from all levels of courts in China, we construct measures of crime rates that vary across cities over time. Our estimations, using a shift-share instrumental variable, reveal a higher increase in crime rates in cities that experience a more severe export slowdown. The effects are more pronounced in the manufacturing-specialising regions, where there is a higher concentration of young people, migrants and school dropouts. Negative export shocks also lead to reduced job opportunities, decreased labour earnings for migrants and worsened psychological health. Alternative mechanisms, such as spending on social stability, appear to play a minor role. A back-of-the-envelope calculation based on our baseline estimation suggests that approximately 13.1% of the interquartile difference in crime rates across cities can be explained by differences in their exposure to the export slowdown.
Hong Ma; Yu Pan; Mingzhi Xu (2025) . The Economic Journal
This paper highlights the criminogenic consequence of the remarkable slowdown in China’s exports in recent years. By applying textual analysis to millions of judgement documents from all levels of courts in China, we construct measures of crime rates that vary across cities over time. Our estimations, using a shift-share instrumental variable, reveal a higher increase in crime rates in cities that experience a more severe export slowdown. The effects are more pronounced in the manufacturing-specialising regions, where there is a higher concentration of young people, migrants and school dropouts. Negative export shocks also lead to reduced job opportunities, decreased labour earnings for migrants and worsened psychological health. Alternative mechanisms, such as spending on social stability, appear to play a minor role. A back-of-the-envelope calculation based on our baseline estimation suggests that approximately 13.1% of the interquartile difference in crime rates across cities can be explained by differences in their exposure to the export slowdown.