Title: Economic Stress and Crime in Japan
Name: Aki Roberts

Research Assistant Professor
Institute for Social Research / Sociology
2808 Central SE
Albuquerque, NM 87111

Year: 2002
Type: Research Grant

This research had three main parts. First, I examined the effect of levels of economic stress (income inequality and unemployment) on Japanese crime trends in the 47 Japanese prefectures (geographical units similar to provinces or states), with data from ten time points between 1955 and 2000. For the second part, using the prefectural-level time-series data, I examined both short- (immediate) and long-term (lasting) effects of economic stress variables on violent crimes, through "error correction" models. Economic stress, measured by unemployment and within- and between- prefecture income inequality, is significantly and positively associated with robbery rates. In the "error correction" analysis, unemployment and both measures of income inequality also had significant and positive short- and long-run effects on robbery. This suggests that high levels of those economic stress variables increase robbery rates immediately, and also have lasting effects. Unemployment also had significant and positive effects, including short- and long-run, on homicide, but income inequality measures did not. It may be that the nonsignificant results for income inequality reflect the more directly economic nature of robbery compared to homicide.

The most common explanation for Japanese postwar crime rates links unique cultural characteristics to a system of exceptionally effective informal social controls that, at the macro level, suggest low levels of social disorganization. However, in general measures of social disorganization (divorce, female labor participation, and urbanization) were not significantly associated with violent crime rates. None of the measures of social disorganization had significant effects in the robbery models. The significant and expected positive effects of divorce on homicide were the only evidence consistent with the social disorganization account.

The third part of the research was intended to foster a more general understanding of the relationship between economic stress and crime trends. I examined whether differences in levels of economic stress can account for the changing discrepancy in the United States and Japanese violent crime rates through time-series analysis of data from both countries between 1951 and 2000. The results of the pooled time-series analysis of 47 Japanese prefectures found that economic stress variables were better predictors than social disorganization variables of violent crime trends in postwar Japan. However, according to the results of time-series analysis of differences in violent crime rates between United States and Japan, differences in both economic stress and social disorganization measures helped explain the differences in violent crime rates between two countries. These differences were associated with differences in the level of divorce, female labor participation, income inequality, and poverty.

A paper on the results of the prefectural-level time-series is under review by the journal Criminology. I am waiting to submit the results from the "error correction" models and time-series analysis on United States and Japan differences to academic journals because I would like to cite the prefectural-level time-series paper if it is accepted for publication in Criminology.