The effect of crime on healthcare utilization : a panel data approach

Abstract

This paper utilizes a behavioral model proposed by Andersen (1968) as motivation and a unique dataset compiled from three sources to test the effect that non-contemporaneous crime has on health care utilization. Using the Arellano-Bond estimation technique, results suggest that crime from the previous year and three years prior significantly decreases healthcare utilization. However, further investigation suggests that the effect is stronger in more urban counties with the third lag of crime being statistically significant and there is no evidence of an effect in smaller counties. Policy prescriptions are put forth that include safety vans for the ill in larger counties.