Abstract
My dissertation comprises three independent yet thematically related chapters on development, gender and labor economics. The first two chapters focus on the impact of college-based affirmative action policies on women in STEM. The first chapter examines the upstream effects by studying labor market outcomes and gender discrimination in hiring, while the second investigates downstream effects on the subject choice girls make at the high-school level. The third chapter focuses on the role social norms play within households in developing countries which has consequences on domestic violence faced by married women.
The first chapter, “Equity Conundrum: Unintended Consequences of College-Level Affirmative Action on the Labor Market”, examines the impact of affirmative action (AA) policies on discrimination in hiring against the beneficiary group. Gender-based affirmative action policies in top-ranked STEM institutions aim to enhance women’s representation in both higher education and the labor force. While these policies can promote diversity, they may also increase statistical discrimination in hiring practices as colleges lower admission standards to increase female enrollment. I investigate how expanding seats for women in premier Indian engineering colleges affects gender discrimination in hiring. I conduct a large scale correspondence study that randomizes gender, college type, and year of entry and induces variation in policy exposure within the experimental design. The results indicate no significant male female callback gap at top colleges before or after the policy. However, women from lower-ranked colleges face disadvantages. Specifically, the policy implementation led to a 52% drop in the female callback probability, increasing male-female callback gap by 2 percentage points in these colleges. To further shed light on actual employment outcomes, I analyze data scraped from LinkedIn profiles, revealing consistent evidence that supports my findings. I propose a model of statistical discrimination that incorporates affirmative action for women at top colleges, aligning with the observed trends in hiring practices.
The second chapter, titled “Fixing the Leaky Pipeline: Affirmative Action in Local Elite Colleges and Subject Choice”, examines the same policy’s impact on girls’ educational outcomes in school. Women are largely underrepresented in STEM careers associated with higher labor market returns. This gender gap is even more stark in a context where societal biases are prevalent and female role models are lacking. I investigate the impact of an affirmative action policy implemented in an elite educational institution in India that ensures additional seats specifically for women in undergraduate STEM courses. After the policy was implemented, the proportion of women enrolling increased by 100%, proportion of women taking the college entrance exam increased by 10% and those qualifying the exam increased by 15%. Using nationally representative data, I employ a triple difference strategy and find a 27% increase in the probability of studying science courses after Grade 10 amongst younger girls exposed to this policy, suggesting a 6% increase in the expected earnings of women.
In the third chapter, co-authored with Sheetal Sekhri and Pooja Khosla, titled “Empowering to Conform: Age at Marriage, Social Norms, and Violence Against Women”, we deliver a theoretical model of domestic violence which proposes that delayed age of marriage aids women in complying to socially desirable behavior expected by the husband rather than challenging the dis-empowering norms. The theory predicts that delaying the age of marriage would reduce domestic violence, but individual wealth of the wife would offset this effect. We corroborate the model by leveraging two rounds of nationally representative health surveys from India and differences in societal and marriage practices among Hindus and Muslims, finding causal empirical support for our theory. Using age of menarche as an instrument, we find that one year delay in age of marriage reduces emotional violence by 18%, less severe physical violence by 26% and severe physical violence by 39%. These findings are stronger amongst Hindu women and gets muted for Muslims who typically have higher wealth at the time of marriage. Our paper highlights that fear of violence can undermine empowering policies when there is widespread acceptance of beating social norms.
