Fabian Lange, McGill University

Earnings Dynamics: the Role of Learning, Human Capital, and Performance Incentives
Friday, 10 November 2017 - 2:30 pm to 4:00 pm
Location
Contact information
Contact person: 
Jason Garred
Email: 
jgarred@uottawa.ca
Phone: 
613-562-5800
Extension: 
1750
Registration
Registration required: 
No
Cost to attend: 
Free of charge
Event language: 
Intended audience: 
Absract:    The investment motive is central to two large literatures on life-cycle earnings.  First, the literature on career concerns highlights how informal incentives from workers' desire to invest in their reputation – the market perception of their abilities – can substitute for explicit incentives in compensation contracts to induce workers to exert effort. Over the life cycle, career concerns weaken and thus compensation becomes increasingly responsive to performance in order to provide incentives for effort. Second, the literature on human capital investments considers how workers invest in their productive capacity and how average compensation varies over the life cycle in response to these investments. Its life-cycle predictions derive from the declining incentive to invest in human capital as individuals approach retirement. In this paper, we analyze how these two investment problems interact in shaping the life-cycle profile of earnings and pay-for-performance. We formulate a fully dynamic, yet tractable equilibrium model of learning-by-doing with symmetrically unobserved ability and privately observed effort. We characterize the optimal compensation contract and decompose the implied sensitivity of pay to performance into distinct components capturing the influence of the dynamic hedging of risk, career concerns, and learning-by-doing. We show that the model is identified using the covariance-structure of earnings over the life-cycle. Using data from the PSID as well as firm level data on compensation, we investigate how the strength of pay-for-performance varies over a worker's career and estimate the parameters of the model.