In two papers, Meyer and Wise (1983a,b) present an ingenious method for estimating the effect of minimum wage rates on wages and employment using data based only on the observed cross-sectional distribution of wages. They, and others who have used this method, have generally found that the minimum wage causes substantial losses in employment. In this paper we evaluate the robustness of this technique. We argue that the estimates, at least for the UK, are very sensitive to the functional form assumed for the distribution of wages and to the assumption made about how far up the wage distribution the minimum wage has spillover effects.