Losing the competition: A multilevel analysis of over-education in the United States, 1971-2006

Since 1970, the U.S. labor market has seen a dramatic increase in the occurrence of over-education, where a worker has more education than what is common for their occupation. Most study has focused on the consequences of over-qualification rather than the causes of or demographic differences in over-education. This study considers four previously unexamined questions. First, to what extent do workers' log odds of over-education depend on the specific occupation and year in which they work? Additionally, how have changes in the incentives for higher education and recent labor market changes contributed to the rise in over-education in the contemporary period? Finally, which of two sets of theories better explains the effects of race and gender on workers' log odds of over-education? The project explores these questions using a sample of black, white, and Hispanic workers aged 25-65 years collected from the Current Population Survey over the years 1971 to 2006 and a three level logistic model wherein workers are nested within occupations and years. The study finds non-significant effects for the education incentive and labor market changes on over-education. However, the research finds considerable occupation and year level variance in workers' log odds of over-qualification. Most surprisingly, the study finds that white women and minority workers have significantly and substantially lower odds of over-education than white men. Moreover, this research finds that occupational segregation plays a mediating role in the effects of race and gender on over-education. These results suggest that gender queue and occupational/social closure theories better explain the effects of race and gender on over-education than job competition theory.