Optimal Job Design and Career Dynamics in the Presence of Uncertainty

The paper studies a learning model in which information about a worker's ability can be acquired symmetrically by the worker and a firm in any period by observing the worker's performance on a given task. Productivity at different tasks is assumed to be differentially sensitive to a worker's intrinsic talent: potentially more profitable tasks entail the risk of greater output destruction if the worker assigned to them is not of the ability required. We characterize the (essentially unique) optimal retention, task assignment and promotion policy for the class of sequential equilibria of this game, by showing that the equilibria of interest are strategically equivalent to the solution of an experimentation problem (a discounted multi-armed bandit with independent and dependent arms). These equilibria are all ex ante efficient but involve ex post inefficient task allocation and separation. While the ex post inefficiency of separations persists even as the time horizon becomes arbitrarily large, in the limit task assignment is efficient. When ability consists of multiple skills, low performing promoted workers are fired rather than demoted, if outcomes at lower level tasks, compared to those at higher level tasks, provide a sufficiently accurate measure of ability. We then examine the strategic effects of the dynamics of learning on a worker's career profile. We prove, in particular, that price competition among firms causes ex ante inefficient turnover and task assignment, independently of the degree of transferability of human capital. In a class of equilibria of interest it generates a wage dynamics consistent with properties observed in the data