A new coworker: better a male, a female or a friend?
In Uganda, the persistence of gender segregated labor markets result in the clustering of women within the least lucrative sectors. The proposed project examines the extent of gender bias perpetuated by employees through the referral system, which is a significant driver behind hiring decisions within informal labor markets. We present employees of businesses operating in highly segregated occupations with potential CVs to investigate their gender preferences as well as how these preferences interplay with personal connection to the candidate. On a small subsample we use monetary incentives to gain insight into the range of the value that respondents place upon upholding this form of discrimination. Measuring this barrier to gender integration is essential for designing interventions which address discrimination in hiring. A gender wage gap can arise because of within-occupation wage differentials as well as across-occupation gaps. In a country like Uganda, where labor markets are strongly segregated by gender, the extent to which gender wage gaps are explained by the occupational structures is likely to be relevant. As is similar in other contexts, women are clustered in low-paying careers such as cosmetology and tailoring, and men are clustered in more lucrative professions such as welding and automotive repair. This trend is persistent. On one hand, stereotypes and the threat of workplace harassment limit the labor supply in those sectors. On the other hand, firm owners of businesses in male dominated sectors are, by definition, more likely to be male (as are the rest of the employees). The affinity bias (i.e. the tendency to warm up to people like yourself) together with implicit biases, is likely to perpetuate gender segregation (Milkman et al., 2015). For instance, Booth and Leigh (2010) found that pro-female biases increase with the proportion of women in the occupation in Australia. In informal labor markets, where job network is key for landing a job through referrals, biases among employees, and not only firm owners (or human resources workers), act as an additional barrier to gender equality in access to certain occupations (Beaman et al., 2018). This grant would allow us to conduct the preliminary research necessary to best design a resume audit study aimed at exploring how gender discrimination varies according to the sector, the gender, and the relationship to the candidate of the person evaluating the resumes. Differently from previous studies that submit fake profiles to current openings (handled by either HR departments or firm owners), we aim to investigate the existence of a gender bias among employees, the channel through which firm owners often receive candidate referrals. Resume audit studies have been widely employed to study labor market discrimination against women (Riach and Rich, 1998, 2002, 2006; Carlsson, 2011) as well as race (Bertrand and Mullainathan, 2004), sexuality (Weichselbaumer, 2003; Drydakis, 2009), age (Carlsson and Eriksson, 2019) and obesity (Rooth, 2009). They allow for the manipulation of candidates' characteristics and for the assessment of employers' behavior in a real environment. Nonetheless, the literature of resume audit studies applied to gender discrimination is still limited. To the best of our knowledge, all correspondence studies disregard the gender of the employer and how this could interact with the assessment of candidates' profiles. Also, in most studies resumes were shown to employers or HR departments and have not been used to assess the gender bias of coworkers. Last, most of the evidence comes from rich countries and, to the best of our knowledge, no similar work has been done to address this dimension of gender discrimination in Africa. This proposal development grant is instrumental toward the research team's development of a strong proposal for future JPAL-GEA competitions. It permits the PIs to align their research design to maximally speak to themes A and C as outlined in the framing paper. Our future randomized evaluation will exploit a substantially bigger sample of workers and will be therefore powered to pin down the value attached to gender discrimination by employers and employees, how widespread gender biases are in specific sectors, to what extent financial incentives can promote gender integration, the relevance of across-occupation gaps in explaining wage gaps, and the value that workers give to their networks against unknown candidates, which we now only investigate descriptively. We will leverage this grant to carry out FGDs with firm owners and VTI instructors to understand the characteristics of Ugandan labor markets, namely the features that employers and coworkers value in candidates, the ways through which candidates successfully flag such desired characteristics, and, on a more practical note, how to build credible CVs and profiles. In this exploratory work, we will also have the chance to understand the impact that different amounts of money have in incentivizing workers to select candidates that have chances of being retained. We intend to use reasonable monetary incentives to create credible games, and this is an opportunity to fine-tune the amounts. In addition, we will be able to fine-tune the survey instruments which would be used and to seek qualitative evidence on gender discrimination in this setting through anecdotal and focus group discussions, as well as through the use of structured interviews with key informants. We also intend to carry out FGDs, structured and guided by BRAC staff, with teachers from VTIs and other workers from outside the sample with whom the team has previously worked with on other projects to assist with the development of the CVs. The bulk of our exploratory evidence will come from a cohort of 500 successful alums from vocational training institutes the research team has been following for more than 2 years as part of a different project. We will leverage a soon-to-be-launched survey and add a small extra module to explore how gender preferences in segregated occupations interplay with a referral-based hiring system. About 30% of the respondents are self-employed and 70% are wage-employed. We will focus on the latter group. We will show them pairs of CVs (one male-named and another female-named), randomizing the quality of the CV (high and low quality), and ask who they would refer to a subsidized one-month internship position at their company, to elicit any form of strong gender preference. Subsequently, we will ask for them to mention two of the best candidates they can think of in their network but of opposite genders. For each of the network mentioned we will ask gender, age, relationship as well as previous work experience to be able to compare all profiles. Finally, we will ask who they would pick for the position among all four candidates. To make the exercise somewhat credible (while meeting our budget constraints), we will run a lottery to subsidize the internship for 7% of the chosen candidates (500*70%*7%=24). To ensure wage-employed respondents have quality-seeking preferences, we will offer to all of them an incentive of 45 USD if the picked candidate is retained (assuming he/she is selected in the internship lottery), which is equivalent to half of our respondents' average monthly income (computed as part of different project). However, if the respondent picks a candidate whose gender matches the dominance of their sector, we will increase the incentive if they give up on their choice and pick a candidate whose gender does not match the dominance of their sector, randomizing the amount of the increase. With this, we will be able to gain some insight into how much respondents value their discrimination.