Carnegie Mellon University - Department of Statistics
This paper examines the linkage between crack market activity and gun homicide suggested by Blumstein (1995), who argues that the arrival of crack stimulated an increased availability of guns among juveniles. This greater availability of guns, the argument continues, is responsible for the sharp upswing in juvenile homicide experienced in the United States in the mid-1980's. Using city-level data on crack arrests and gun-related juvenile homicide, we fit a change-point version of the Bass (1969) model of innovation diffusion. We find that, in most large American cities, the diffusion process for crack cocaine experienced an onset of dramatic growth that was followed by similar, slightly slower growth in gun homicides committed by juveniles. We further use cluster analysis to find that the spatial patterning of the two processes is similar, starting on the East and West Coasts and working their way toward other regions of the nation. Gun use in homicide among slightly older offenders (ages 18-24) also experienced a change at roughly the same time as the juveniles, but the rate of diffusion was considerably milder than for the younger group; offenders ages 25 or older generally show no growth in gun-related homicide whatsoever. In addition, there is no detectable surge in juvenile non-gun homicide activity. Based on these findings, we conclude that the crack cocaine markets-gun availability linkage is highly plausible, and we suggest directions for future research in clarifying the dynamics of the late-1980's surge in juvenile homicide.