Invited Commentary: Methods for estimating effects of minimum wages on health.

Economists have been researching effects of minimum wages on unemployment, poverty, income inequality and educational attainment for over 60 years. Epidemiologists have only recently begun researching minimum wages even though unemployment through education are central topics within social epidemiology. Buszkiewicz et al.'s study (this issue of the journal) is a welcome addition to this nascent literature. A commanding advantage of Buszkiewicz et al.'s study over others is its distinction between a "likely affected" group comprised of workers with <= 12 years of schooling versus "not likely affected" groups with 13+ years of schooling. But there are disadvantages, common to other studies. Buszkiewicz et al. use cross-sectional data; they include the self-employed as well as part-time and part-year workers in their treatment groups. Their definitions of affected groups based on education create samples with 75% or more of workers who earn significantly above minimum wages; definitions are not based on wages. Inclusion of workers not subject to (e.g self-employed) or affected by minimum wages bias estimates toward the null. Finally, minimum wage increases occur annually but the development of chronic diseases may take decades. Buszkiewicz et al.'s and most "minimum wage and health" studies do not analyze more than eight years.