This paper helps close the gap between theory and empirical evidence in the literature on asymmetric employer learning. If an employer's private learning is reflected in a worker's wage and one employer's private information is transmitted to the next when the worker makes a job-to-job transition, then asymmetric employer learning will appear in wage regressions as learning over an employment spell. Extending previous work that assumes all learning takes place publicly, this paper develops wage regressions that test for both asymmetric employer learning and public learning. The empirical results, including tests of alternative explanations, are consistent with asymmetric employer learning's having at least as much of an effect on wages during an employment spell as does public learning. The model developed in this paper illustrates how the story suggested by the empirical work might unfold. It shows that outside firms can profitably compete with a better-informed employer through bidding wars, even when the worker is equally productive in all firms. Furthermore, this competition results in different wages for workers with the same publicly observable characteristics, a result that previous models of asymmetric learning have not produced.

We study two-sided markets with a finite number of agents on each side, and with two-sided incomplete information. Agents are matched assortatively on the basis of costly signals. Asymmetries in signalling activity between the two sides of the market can be explained by asymmetries either in size or in heterogeneity. Our main results identify general conditions under which the potential increase in expected output due to assortative matching (relative to random matching) is offset by the costs of signalling. Finally, we examine the limit model with a continuum of agents and point out differences and similarities to the finite version. Technically, the paper is based on the elegant theory about stochastic order relations among differences of order statistics, pioneered by Barlow and Proschan in 1966 in the framework of reliability theory.