Gender Bias in Job Assignment: Evidence from Retail Frontline Managers
SPEAKER
Dr. Susan Feng Lu
Associate Professor
Gerald Lyles Rising Star Professor of Management
The Krannert School of Management, Purdue University
ABSTRACT
Anecdotal evidence suggests that gender disparity in job assignment might be a main driver behind the gender pay gap. However, existing gender studies have focused on individual workers or C-suite executives, while systematic empirical studies on low-level managers are rare. In this study, we empirically investigate gender disparity in the job assignment of frontline managers, specifically the effect of gender on store-manager assignment, using personnel, sales, and operational data from a large sportswear retail chain. Our estimation strategy is a two-step fixed effects framework: 1) Separately identify manager fixed effects and store fixed effects by exploring manager switching across stores; 2) Estimate the effect of gender on store-manager assignment while controlling for the estimated manager fixed effects from the first step. Applying this estimation framework to the retail chain’s data, we uncover strong evidence for a gender bias in store-manager assignment—male mangers are more likely than their female counterparts to be assigned to stores with high sales potential, i.e., core stores, stores with a large number of sales clerks, or stores with a high sales target. To support our finding on gender bias in store-manager assignment, we test three alternative hypotheses. First, we find that male managers achieve similar sales when replacing their female counterparts in the same stores, and male managers are not better at managing the turnover of sales clerks, neither of which supports an alternative hypothesis of differential managerial abilities; Second, we find that male sales clerks on average achieved similar or lower sales individually than their female counterparts, thereby not supporting a career selection bias hypothesis that males seeking the sales clerk positions at the retail chain tend to have high sales skills. Third, we conduct a maximum likelihood estimation of a discrete-time Markov chain that models the processes of store-manager assignment and manager turnover, and do not find evidence to support the hypothesis that inequitable store-manager assignment is driven by different gender preferences to leave stores with low sales potential. Translating our findings into an actionable insight, we develop a gender inequity index (GII) to help organizations identify potential gender bias in job assignment. Applying this index to a simulation study using the retail chain’s store manager compensation data, we demonstrate the implications of gender-biased job assignment on the gender pay gap.