Zahra Siddique (University of Bristol) on "Words Matter: Gender, Jobs and Applicant Behavior in India" (joint work with Sugat Chaturvedi and Kanika Mahajan)

We examine employer preferences for hiring men vs women using 0.16 million job advertisements posted between July 2018 and February 2020 on an online job portal in India which caters primarily to young urban job seekers. We also use data on all applications made to these ads (from 1.04 million active job seekers) to study consequent search behaviours. We find that gender preferences are more likely to be exhibited for low-skill jobs and that these preferences operate to maintain existing occupational gender stereotypes. Using the applications data we find that an employer’s explicit preference for women significantly reduces the total number of applications to a job ad and changes the gender mix of the applicant pool in favor of women. We also exploit textual information contained in the title and description of a job ad to construct gender stereotypes and exploit these measures to address two questions. First, do stereotypes matter for gender wage gaps? Second, do such stereotypes affect applicant behaviour? Lastly, we uncover what lies beneath the constructed stereotypes, focusing on the importance of skills and personality. Our analysis highlights the important role played by explicit gender preferences and associated stereotypes in explaining observed gender disparities in the labor market.