Neighborhood Income, Alcohol Availability, and Crime Rates

Author:
  • KGKwabena Gyimah-Brempong
    Kwabena Gyimah-Brempong

Abstract
This paper uses census tract data and instrumental variables (IV) estimation to investigate whether alcohol availability has differential crime effects in low and high income neighborhoods. I estimate a crime generating equation in the Becker tradition (Becker: 1968) that includes alcohol availability for low and high income neighborhoods and then compare the estimates from the two sets of equations to see if the two estimates are the same. The basic argument is that the political economy of alcohol outlet location leads to over concentration of strong intoxicating alcohol outlets in low income neighborhoods. This overconcentration of alcohol outlets leads to loss of social control hence excessively high crime rates than would be predicted by "crime fundamentals". The reverse ma be true for high income neighborhoods. Crime rates vary" enormously among jurisdictions with similar socio-economic characteristics within and across cities. In addition, these county-wide (citywide) averages masks a large heterogeneity of crime rates within jurisdictions, Crime rates are generally higher in poor neighborhoods than in affluent ones in urban America; so is the concentration of alcohol outlets. The high rates of crime in low income neighborhoods creates the impression that people in poor neighborhoods have higher crime propensities than their counterparts in high income neighborhoods. This study is informed by two complimentary observations, First, the correlation between alcohol availability and income in the sample I use is nonlinear. It is positive at low income levels but becomes negative at high income levels. Second, the relationship between crime rates and alcohol availability shows considerable discontinuity: Gyimah-Brempong and Racine (2006) find that the effect of alcohol availability on crime rate doubles when alcohol availability, as measured by the number of liquor licenses, reaches 10 in a census tract. The combination of these two effects is likely to have a nonlinear effect on the relation between alcohol availability and crime that cannot be captured by the usual

This publication has 36 references indexed in Scilit: