Asking the right questions about substitution and complementarity: a comment on Moore's paper.

In this issue, the paper by Moore (2010) raises an important question regarding the wider effects of alcohol policies with respect to the use of other (illicit) drugs. From what the author identifies as sparse evidence of the impact of alcohol policy on other substance use, he draws a rather strong conclusion that ‘the value of alcohol-specific policy remains unknown’. But is this conclusion supported, even by the evidence available? In principle, the evaluation of policy interventions, whether before or after implementation, should consider all potential effects (positive and negative) and the possibility of unintended consequences of the intervention. One eminent health economist has used the comparison with medical treatments, which inevitably carry a risk of side-effects (Evans, 1984). More pragmatically, however, the resources needed for a comprehensive and wide-ranging analysis lead to a focus on those effects that are most important in scale and scope and a reliance on sensitivity analysis to identify whether conclusions are robust when factors in the evaluation are varied. Thus, we need to consider whether there is sufficient a priori evidence that policies to control the consumption of alcohol will have significant effects on the use of other drugs, the direction of these effects, …