Skills,Performance,and Earnings in the Tournament Compensation Model

Abstract

Previous research on professional golf has estimated reduced-form models in which earnings are a function of skills measured as year-long averages. Using individual, tournament-level data, this article makes three improvements. First, the use of tournament-level data removes measurement error in the skills by adjusting the data for tournament-level idiosyncrasies such as the effect of altitude on driving distance. Second, the collection of microdata allows one to examine variance and skewness of the skills distributions in addition to just the mean. And third, the article estimates a structural model in which golfers use their skills to perform well in competitions by shooting low scores at one level, while those score distributions underlie the tournament earnings of the golfers at the other level. Tighter estimates of the coefficients are resolved and the overall adjusted R² improves from 0.36 to more than 0.90.