Association for Public Policy Analysis & Management

This study estimates the prevalence of ninth grade repeating and examines how recent growth in ninth grade repeating and how grade repeating in general introduce bias in estimation of the returns to education and experience. Using the National Center for Education Statistics’ Common Core of Data, I construct estimates of ninth grade repeating rates for each state from 1965-2012. Since the late 1980s, about 10 percent of all public high school students repeated the ninth grade at least once. I then show that 10 percent of the growth in ninth grade repeating can be attributed to changes in Compulsory Schooling Laws (CSLs) over the past 50 years. The rise in ninth grade repeating has important implications for estimation of the returns to education and experience. Using the NLSY79 Children, I estimate the returns to education using CSLs and quarter of birth as instrumental variables, then correct for potential endogeneity with ninth grade repeating. Compulsory schooling instruments create bias of up to 38 percent when failing to account for endogenous ninth grade repeating. I then examine how grade repeating in any grade affects estimation of wage, hours, and employment differentials. Labor market experience is not typically reported in most datasets, and instead is replaced by a “Potential Experience” proxy – calculated as age minus educational attainment minus six. This experience measure faces measurement error due to grade repeating; each attempt at repeating a grade reduces the true value of potential experience by one. In the racial discrimination literature, this measurement error is endogenous, as blacks are twice as likely as whites to repeat grades. I show that the residual black-white wage gap for early-career workers is overstated by 10 percent when not properly accounting for repeated grades. I also show that the wage gap between GED recipients or high school dropouts without a GED and high school graduates is overstated by 15 percent, and the hours gap is overstated by 33 percent when failing to account for grade repeating. These findings suggest that failure to account for the massive increase in ninth grade repeating in the United States and the high level of grade repeating overall has the potential to substantially bias any contemporary estimates of the returns to education and experience, as well as estimates reliant on controlling for education and experience.