The relationship between cocaine prices and crime has critical implications
for US. drug policy, but is theoretically indeterminate because cocaine pr
ice changes affect crime through changes in both cocaine consumption and ex
penditures. This paper investigates this relationship in annual data from 1
981-95 on 29 large US. cities, accounting for simultaneity by using two-sta
ge least squares with measures of wholesale supply factors and retail enfor
cement intensity as instruments for cocaine prices. Controlling for prices
of other drugs, deterrence, socioeconomic factors, and city and year-specif
ic effects, a strong negative relationship exists between cocaine prices an
d six of seven FBI index crimes.