| Summary: | The paper explores the differences in the empirical determinants of crime using a spatial model which distinguishes resident and non-resident offenders. Using data for German municipalities, the model is estimated by means of a spatial GMM approach, where the local property value is instrumented by a couple of amenity variables. The results show that aside of the local property value several local population characteristics, such as income, poverty, inequality, unemployment, family disruption, and citizenship have the expected effects on crime committed by resident offenders. However, crime committed by non-resident offenders is shown to be significantly related to the conditions in adjacent municipalities as captured by spatial lags of population characteristics.
|
|---|
