Paul Bingley, Kaare Christensen, and Kristoffer Markwardt:

Does formal schooling contribute to individual labor market productivity or does it act as a signal to employers of predetermined labor market skills? We test for whether employers statistically discriminate between workers on the basis of their schooling, by assuming we can observe a proxy for worker productivity that the em- ployer cannot – father, brother and co-twin earnings. Using population-based Danish administrative data, we find that employers initially statistically discriminate be- tween workers on the basis of schooling, but schooling earnings differentials fall over time as employers learn about worker productivity. We further propose a novel test for job market signaling using differences in twin pair earnings growth, and find that signaling is important at the upper end of the schooling distribution – explaining a large proportion of the college wage premium.