The purpose of this study was to describe the differences in illegal adolescent substance use among representative samples of secondary school students in Philadelphia (N = 712), Toronto (N = 983), and Montreal (N = 824) and apply a novel theoretical perspective. Collective efficacy is a widely cited criminological theory that has not been applied to minor forms of delinquency such as underage alcohol use and cannabis use. We examined the influence of four measures of collective efficacy on substance use among students aged 14–17 years in these three cities. We found that adolescents in Montreal consistently reported higher prevalence of alcohol and cannabis use than those in either Toronto or Philadelphia, and these differences remained significant after controlling for compositional differences. We also found that collective efficacy has some theoretical and empirical merit in all three cities surveyed, as the component of neighborhood closure was associated consistently with adolescent substance use involving alcohol and cannabis. However, the other collective efficacy measures of social cohesion, parental closure, and child control were not significantly related to these forms of adolescent substance use and showed some interaction effects by site. Future research is needed to better understand the value of collective efficacy in explaining adolescent substance use, and should be applied to more serious forms of drug use and drug selling.