Investigating crime harm index in the low and downward crime contexts: A spatio-temporal analysis of the Japanese Crime Harm Index

Abstract

Since Louis Thurstone, there have been nearly 100 years of attempts to compare and evaluate the severities of different crime types. Further research has been recently conducted, following methods such as the Cambridge Crime Harm Index (CCHI), which evaluates the harmful impact of crime on the society and identifies the areas in a city with truly high crime harm by weighting the more serious crimes. Although this index facilitates informed urban policy, its effectiveness in settings such as that of Japan, in which the rate of violent crimes, including gun crime, is low, and where crime occurrence is generally infrequent, remains unclear. Here, the Japanese CHI was analyzed by examining the criminal offense data of 77 jurisdictions in central Tokyo in 2010–2019. The pre-weighted counts and post-weighted harm were compared to examine the correlations and time-series changes between them, as well as their differences via different indicators, such as the Gini coefficient and Moran's I statistic. Additionally, space-time scans and principal component regressions were employed to determine the specific characteristics of the jurisdictions where harm accumulates. The result indicated that although the temporal and spatial variability of harm was higher than that of its counts, the tendency of high-harm jurisdictions to cluster was lower. Further, the implications of our findings, particularly regarding the socioeconomic factors that were revealed by the regression analysis, to urban policymakers were also discussed.