Abstract
<p>Workers have good days and bad days, but human capital management research overwhelmingly focuses on mean performance. This paper examines whether organizations recognize and price worker volatility when they can measure it, and whether they should. In two studies using data from professional basketball, I document both wage and normative penalties for volatile workers. After separating persistent volatility from playing-time noise, more volatile players earn less on the margin where contracts are negotiable. A difference-in-differences design then exploits the quasi-random timing of absences as an identifying shock. Holding mean performance fixed, losing a reliable player inflicts a larger penalty on team point differential and win probability than losing a volatile one. I show that the steadiest players are nearly as costly to lose as high-mean stars, at forty percent of the salary. Taken together, these results suggest that performance management is a portfolio problem and worker volatility is an important locus for strategic action.</p>