Putting female jockeys in their place

Abstract Gender bias and discrimination within labour markets is a significant area of concern in many industries yet it represents an inherently difficult field for empirical analysis. Infrequent hiring events and complexities in obtaining truly objective measures of productivity, often coupled with under-representation of females, severely limits data availability in most workplace settings. Using a novel approach, we take advantage of the frequent hiring events within the sport of horse racing to evaluate productivity against opportunities afforded to individual jockeys. Our ranking model identifies under- and over-valued jockeys, and we find that women are significantly more likely to be under-used than males, in an industry in which performance data are readily available.