Minimum wages and school enrollment: New evidence from commuting zones

This paper uses a linear probability model in a difference-in-difference framework to examine the effect of the minimum wage on school enrollment for aged 16-19 teenagers with educational attainment below high school diplomas in the United States. The sample in this paper is from individual-level data of the American Community Survey over the period 2000-2011. I first confirm that a traditional two-way fixed-effect model yields a negative effect of the minimum wage on school enrollment. However, my models, which allows year fixed effects to vary across 4 Census Regions, 9 Census Divisions and 741 Commuting Zones (CZs), suggest no adverse effect of the minimum wage on school enrollment. By comparing the traditional model and models with detailed spatial controls, my findings imply that the negative estimated relationship between the minimum wage and school enrollment is spurious and driven by insufficient controls for spatial heterogeneity in the traditional approach. I perform three robustness checks to test the validity of my identification strategies. I evaluate the sensitivity of results by using a restricted sample in cross-border CZs, the linear state-specific time trend and alternative sample periods. My results are robust to these specifications. In addition, I examine the effect of the minimum wage on part-time labor outcomes and college enrollment. For teenagers who are enrolled in school and have part-time jobs, I find that higher minimum wages do not reduce part-time working hours but increase part-time hourly wages. Moreover, I do not find any adverse effect of the minimum wage on college enrollment.