Book Description

This thesis comprises three chapters studying labor markets in developing countries. The first two chapters examine two sources of gender gaps in the labor market -- gender differences in employers' monopsony power over their workers, and the possibility that the decision-makers who design workplaces do not prioritize women's needs when doing so. The final chapter focuses on a different population, of the poorest Indian households, and studies whether a "big-push" program providing these households with a large asset transfer can durably lift them out of poverty. The first chapter examines the extent and sources of gender differences in employers' monopsony power over their workers in Brazil. I exploit establishment-level demand shocks induced by the end of the Multi-Fiber Arrangement to show that women are substantially less likely than men to separate from an employer that lowers their wage. The implied gender difference in monopsony power would generate an 18pp gender wage gap among equally productive workers, explaining over half the raw gender wage gap. To study the source of this gender difference in monopsony power, I build and estimate a discrete choice model wherein employers can have more monopsony over women either because women strongly prefer their current employer, or because they have fewer good employers than men. Of the 18pp monopsony gender gap, I find that 10 points are attributable to women's stronger preference for their specific employer, and 8 points to the fact that good jobs for women are highly concentrated in the textile sector. Surprisingly, I show that this concentration is itself largely a product of amenities/disamenities present in different sectors, rather than gender-specific comparative advantage. My findings demonstrate that although the textile industry provides women desirable jobs, this desirability confers its employers with higher monopsony power. By contrast, desirable jobs for men are not similarly concentrated. The second chapter (joint with Viola Corradini and Lorenzo Lagos) investigates why workplaces are not better designed for women. In particular, we show that changing the priorities of those who set workplace policies can create female-friendly jobs. Starting in 2015, Brazil's largest trade union federation, the Central Única dos Trabalhadores (CUT) made women central to its bargaining agenda. We use a difference-in-differences design to compare establishments negotiating with CUT-affiliated unions to those negotiating with non-CUT unions. We find that "bargaining for women" increases female-centric amenities in collective bargaining agreements as well as in practice. These changes cause women to queue for jobs at treated establishments and separate from them less--both of which are revealed preference measures of firm value. We find no evidence that the gain in amenities comes at the expense of either men or women's employment or wages, or of firm profits. Our results thus suggest that changing institutional priorities can narrow the gender compensation gap. The final chapter (joint with Abhijit Banerjee and Esther Duflo) studies the long-run effects of a "big-push" program that provides a large asset transfer to the poorest Indian households. The program is premised on the idea that the poor are stuck in a poverty trap, which implies that a one-time capital grant that makes very poor households substantially less poor ("big push") can set off a virtuous cycle that takes them out of poverty. In a randomized controlled trial that follows these households over ten years, we find that the program improves poor households' well-being over the long run, increasing their consumption by 0.6 standard deviations (SD), food security by 0.1 SD, income by 0.3 SD, and health by 0.2 SD. These effects grow for the first seven years following the transfer and persist until year ten. One main channel for persistence is that treated households take greater advantage of opportunities for income gains that arise naturally over time, such as by diversifying into lucrative wage employment and migration.