Illicit drug use is thought to pose a public health problem for several possible reasons. Here I discuss one such reason: the supposed causal relation between illicit drug use (especially so-called "hard" illicit drugs) and violent crime. The hypothesis that drugs are causally connected to crime is also a favorite basis on which to argue that drug use should be criminalized. This hypothesis is difficult to test empirically. I build on some resent criminological findings of Frank Zimring fron New York City to suggest that many theorists have tended to overstate the drugs-crime connection. In New York City, violent crime has decreased greatly while "hard" illicit drug use has remained constant. As Zimring concludes, it is possible to make enormous progress in the war on crime without making any headway in the war on drugs. I examine the implications of these recent findings for debates about drug criminalization.