A substantial body of research on the UK's National Minimum Wage (NMW) has concluded that the the NMW has not had a detrimental effect on employment. This research has directly influenced, through the Low Pay Commission, the conduct of policy, including the subsequent introduction of the National Living Wage (NLW). We revisit this literature and offer a reassessment, motivated by two concerns. First, much of this literature employs difference-in-difference designs, even though there are significant challenges in conducting appropriate inference in such designs, and they can have very low power when inference is conducted appropriately. Second, the literature has focused on the binary outcome of statistical rejection of the null hypothesis, without attention to the range of (positive or negative) impacts on employment that are consistent with the data. In our re-analysis of the data, we conduct inference using recent suggestions for best practice and consider what magnitude of employment effects the data can and cannot rule out. We find that the data are consistent with both large negative and small positive impacts of the UK National Minimum Wage on employment. We conclude that the existing data, combined with difference-in-difference designs, in fact offered very little guidance to policy makers.
minimum wage difference-in-difference power minimum detectable effects