The effect of noncognitive ability on the earnings of young men: A distributional analysis with measurement error correction

Utilizing the National Educational Longitudinal Study data, this paper examines the role of pre-market cognitive and noncognitive ability, as measured in tenth grade, on the earnings of young men. In addition to the conditional mean, we estimate the impact over the earnings distribution using recently developed (instrumental) quantile regression method. Our results show that noncognitive ability is an important determinant of earnings, but the effects are not uniform across the distribution. We find noncognitive ability to be the most important at lower quantiles. The impact of cognitive ability, on the other hand, shows a more homogenous pattern. Several robustness checks support these results.