This dissertation consists of three essays in labor economics. The first chapter tackles a classic problem in labor economics: estimating the returns to schooling. Compulsory schooling laws have been used extensively as an instrument for years of completed schooling in estimating the causal effect of education on a number of outcomes. But pre-existing state-level trends in educational attainment induce a spurious positive relationship between educational attainment and compulsory schooling laws. An event study model reveals that the laws have no effect on the distribution of educational attainment, thus making them inappropriate as an instrument. The laws' ineffectiveness can be explained by non-compliance and by measurement error in the pr...