Does Mentor's Ability Matter for Compensation?

We use an agency model with moral hazard and adverse selection to study the effect of a mentor's ability on the compensation of his mentees. An agent who is trained by a mentor of higher ability receives valuable experience that increases not only his productivity but also his output sensitivity to effort. In our model's equilibrium, the greater productivity translates into higher total compensation, and the greater output sensitivity of effort leads to stronger incentives. We test these predictions by using data from college football coaches and we find strong empirical support for our hypotheses. Football coaches who have previously worked as assistants to head coaches of superior ability, are on average more productive. We find that, 1% increase in mentor's performance (our proxy for mentor's ability) increases head coach's total compensation by $7,800. Finally, head coaches who had better mentors received on average stronger incentives in the form of bonus payments, a result which is consistent with our model's predictions.