Minimum wage effects on employment, substitution, and the quality of the teenage labor supply: Evidence from personal data

Using personnel data from a large U.S. retail firm with more than 700 stores nationwide, this study examines how the 1996 federal minimum wage increase affected both the level of employment at each store and also the fraction of teenagers employed. Because the minimum wage caused the relative wages of teenagers to increase at many stores, examining the relative employment of teenagers permits me to test the hypothesis of substitution toward more skilled (i.e., relatively high-wage) employees. The basic results indicate, first, that while the legislation had a much greater impact on average wages in some stores than in others, there is no significant difference in post-legislation employment growth at high-impact and low-impact stores. Second, I find that in stores where the legislation caused larger increases in the relative wage of teenagers, the employment of teenagers rose faster after the law was implemented. These results appear inconsistent with the traditional view that minimum wages lead employers to reduce employment, and to substitute relatively high-skilled, high-wage labor for the labor of the less-skilled workers whose relative wages go up. However, additional analysis suggests that the apparent contradiction can be explained in the context of a model (Drazen, 1986) that incorporates imperfect information about the quality of teenage applicants into the traditional, competitive framework. Specifically, the evidence suggests that high-quality teenagers often were not paid wages proportional to their output, and were drawn into the labor market when the minimum wage forced stores to pay teenagers more like adults. In these stores, the legislation appears to have increased the demand for teenage labor by raising the average quality of the teenage applicant pool. In contrast, where the legislation increased the wages of both teenagers and adults without compressing the wage distribution, the result was a decline in employment, as predicted by the conventional model.