This is the eleventh in this year's series of posts by PhD students on the job market
Why do women’s labor market outcomes lag behind men’s in many developing countries despite the fact that the gender gap in high school and college enrollment in many countries has closed? Particularly in South Asia, informal conversations between policymakers, researchers, or just colleagues in the office underscore that women’s families are the constraint. Parents don’t want their daughters to work. But why is that? Dig a little deeper and, in some contexts, people describe that parents particularly don’t want their daughters to work with men.
But what do daughters want for themselves?
For my job market paper, I conducted an experiment on a formal job search platform in Pakistan to answer this question. My results show that women’s interest in a job is influenced by the gender of their future supervisor, but not so much the gender of who they will be working with. However, women believe their families do care about who they will work with. These preferences that women and their families have about the type of environment that women should work in have large impacts on women’s job search, which in turn can affect their eventual employment status. Here are the details.
About the experiment
About 4,000 women with at least a high school diploma from the city of Lahore (population 11 million) enrolled onto a free job search platform. Three quarters were in their final year of either high school or college and beginning to look for jobs. Consistent with social norms in many countries, about half of these women stated that they were most interested in being a teacher. The next most preferred occupation was manager, suggesting career ambition.
Jobseekers received a weekly set of matches to open job postings in their desired occupations based on their education and work experience. Furthermore, fitting with the normal hiring procedure in Lahore, female jobseekers were only matched to jobs which were willing to hire female applicants. About a quarter of those who initially enrolled were actively using the platform before my experiment began, meaning that they were applying for jobs or were picking up phone calls about jobs from the platform. In the remaining analysis, I focus attention on these jobseekers who actively used the platform, though results are similar for all jobseekers.
The SMS for each job posting included the job title, firm name, neighborhood, salary, and whether the position had flexible working hours -- a job characteristic that women prefer across settings. Jobseekers could call the service to send their resume to the position. A portion of the jobseekers also received calls each week from the platform asking them which (if any) jobs to which they were interested in applying.
I conducted two related experiments on this platform to understand women’s job search. First, in addition to the other job characteristics listed above, I randomly selected some individuals to receive extra information about each job posting: gender of supervisor, gender of coworkers (employees at the firm), or both gender of supervisor and gender of coworkers. The control group received no additional information. Second, among the portion of jobseekers who received calls to apply for jobs, I randomly selected half to be asked whether they had discussed their job search with their family in the last week before they were asked to which jobs they want to apply.
And the results
All results hold fixed observable job characteristics, the treatment indicator for the other information treatment, and other relevant control variables.
Receiving information about gender of the supervisor nearly doubled women’s job application rates compared to those who were not randomly selected to receive information. This demonstrates that jobseekers care about the gender of their potential supervisor. Furthermore, women were more likely to apply to a job with a female supervisor than a job with a male supervisor. There is a difference of about 2pp -- large compared to the control mean application rate of about 6.7%.
Information about the gender of coworkers did not have a significant impact on women’s job application rates, and women did not show a preference for female rather than male coworkers, in terms of which jobs they chose to apply to. While there exists a narrative that women don’t want to work with men, the experimental results don’t show that.
Being asked if they discussed their job search with their family in the last week decreased the application rate by about 30%. This is consistent with women knowing that their families might not be supportive of their job search.
Finally, receiving information about gender of coworkers caused women who were asked this question about family job search to be more likely to apply to the job. This indicates, consistent with the hypothesis and qualitative evidence about social norms, that women’s families care whether they will be working with men or women.
Discussion
Thus, women themselves want to know who their supervisor will be, but not necessarily who they will be working with. Women’s families, however, care about who they will work with.
What does this mean for policy?
My results show that women’s labor market outcomes could be improved even within the constraints of social norms. These results are encouraging and suggest that a low-cost information intervention providing accurate information about gender of the supervisor can increase educated women’s job application rates on a formal job search platform. This can be an important step in translating women’s educational achievements to the labor market. Suppose that firms began to publicize their female supervisors and/or that they have mostly or all female employees to potential job applicants. My results suggest that if women looking for jobs see women in supervisory roles, they are more likely to apply to these jobs. Furthermore, even the women most constrained by their families who want them to work with other women will be more likely to apply to jobs that they can see have other women in the workplace. Simply publicizing this information could increase the volume of job applications from women, which is a first step in improving employment rates at least for more educated young women with access to a phone.
Nivedhitha Subramanian is a Ph.D. candidate in Public Policy (Economics Concentration) at Duke University
Join the Conversation