dicted users engage in economically motivated violent crime (e.g., robbery) in order to support an expensive habit of drug use. Finally, systemic violence occurs as a consequence of the nature of illegal drug trafficking—territorial disputes among dealers, conflicts with disgruntled customers and angry neighbors, intimidation to promote debt collection, and so on. Goldstein and colleagues (1992) estimate that in New York City in 1988, 74 percent of all drug- and alcohol-related homicides fell into the systemic category; of those, 61 percent involved crack cocaine and 27 percent involved powder cocaine. Another 4 percent of all drug- and alcohol-related homicides were primarily economically compulsive. Only 14 percent were primarily psychopharmacological in nature, and 68 percent of those involved alcohol—a legal drug. Two open questions are whether the relative frequencies of each type of homicide have changed over time, and whether nonhomicide drug crimes fit a similar profile.
But as several authors have argued in detail (Kleiman, 1992; MacCoun et al., 1996; MacCoun and Reuter, 2001), the premise that drug prohibition causes some drug-related harm does not necessarily imply that ending prohibition would, on net, reduce total drug-related harm. It is likely that many of the harmful consequences of drug use would be significantly reduced under a regulated policy of legal access to drugs—but not all harms. And even if average harm declined overall, there is no guarantee that net or total harm would decline as a result. Recall that total harm= average harm per incident of use×total use. If average harm declined under legalization, the effect on total harm would then depend on whether total use increased. If the average harmfulness of an incident of use (on any given dimension of harm) dropped but total use increased, the net effect on total harm is uncertain; it could fall, remain constant, or increase significantly (MacCoun and Reuter, 2001). Thus, legalization is an unproven and potentially risky strategy for reducing drug-related harms.
For these reasons, the observation that drug illegality contributes to drug harmfulness is by no means tantamount to endorsing drug legalization. In discussing this point, we acknowledge concerns about the fear of sending the wrong message about the harms of drugs. However, inadequate analysis of the consequences of drug use gives a rhetorical advantage to advocates of legalization without allowing a serious examination of the merits and weaknesses of their arguments. Moreover, the failure to address the full consequences of drug prohibition results in lost opportunities for reducing drug-related harm under a policy of prohibition.