Global Warming and U.S. Crime Rates

Abstract

Two archival analyses were performed to examine the association between annual temperatures and U.S. crime rates. The first was based on area-averaged temperatures in the United States as a whole for the years 1950 through 1999. Box-Jenkins time-series analyses indicated that annual temperatures were associated with assault but not murder rates in analyses that controlled for yearly population, ethnicity, and three economic variables. The second analysis was based on state-centered crime rates from 1960 through 1998 and included the same controls. Contrary to the general aggression model, cross-sectional time-series analyses indicated that annual temperatures were associated with rates for assault, rape, robbery, burglary, and larceny, but not murder or motor vehicle theft. The results are consistent with a routine activity theory interpretation of everyday and criminal behavior.