Essays in Behavioral Development Economics
Name
rajah-krajah-phd-economics-2026-thesis.pdf
Size
9.94 MB
Format
Adobe PDF
Checksum (MD5)
faba6586860c1dc3019181ee74c92f2d
Author(s)
Rajah, Kailash
Advisor(s)
Duflo, Esther
Schilbach, Frank
Banerjee, Abhijit
Date Issued
May 2026
Publisher
Massachusetts Institute of Technology
Abstract
This thesis comprises three chapters studying how social and psychological factors shape economic outcomes. The first two chapters examine the determinants of married women’s labor supply; the third explores learning and skill transfer among children working in informal markets. The first chapter studies how spousal jealousy constrains women’s employment in India. I conduct two field experiments in partnership with a private firm in Patna, Bihar. In the first experiment, I randomly offer identical business process outsourcing jobs in either a women-only and a mixed-gender office. I find that women are 46% (13 percentage points) more likely to apply for the women-only job and 31% (6 percentage points) more likely to attend the workplace. I present evidence that safety and social image concerns are not the primary mechanisms: informing women of the firm’s rigorous workplace safety measures improves safety perceptions but has no effect on job applications, and making women’s choices more public reduces rather than increases preferences for women-only work. I do, however, find that treatment effects are significantly larger among households where women report having more jealous husbands. To directly test for a jealousy mechanism, I conduct a second experiment. In this experiment, I offer women a job in the women-only call center, which this time comes with an online peer support program. In the peer support program, women speak with a more experienced peer working in another office for career and professional advice. Over half of households pay a fee, equivalent to a full day’s household income, to guarantee having a female peer rather than to risk having a male peer. Consistent with a spousal jealousy mechanism, introducing monitoring —by giving the husband the option of joining the calls—reduces this demand by over a third. Even after eliminating any possible differences in the quality of mentoring, by having the mentoring take place via prerecorded videos, one-third of households still pay for a female peer, suggesting that jealousy motives extend to even passive interactions. Together, the two experiments provide strong causal evidence that spousal jealousy is an economically important barrier to women’s employment. The second chapter, written with Ishaana Talesara, tests whether breadwinning norms— the expectation that husbands should out-earn their wives—constrain married women’s labor supply. We conduct a field experiment with 171 married women in Delhi, eliciting approximately 5,000 job application decisions with wages randomized above and below each husband’s stated income. In our pre-registered main specification, we find no evidence of a discontinuity in labor supply at the husband’s income. We can reject negative discontinuities as small as 1.5 percentage points—an order of magnitude smaller than the 30-percentage-point gap observed in cross-sectional distributions. A 1% increase in wages raises labor supply by approximately 2%, consistent with neoclassical models. These null results hold across subgroups with the strongest a priori exposure to breadwinning norms, including households with conservative gender attitudes, those practicing veiling, and those with salaried husbands. The results suggest that previously documented cross-sectional patterns may reflect selection or marriage market sorting rather than intra-household labor supply constraints. The third chapter, written with Abhijit V. Banerjee, Swati Bhattacharjee, Raghabendra Chattopadhyay, Esther Duflo, Alejandro J. Ganimian, and Elizabeth S. Spelke, studies whether arithmetic skills acquired by children working in informal markets transfer across domains. We study 1,436 children selling goods in markets in Kolkata and Delhi and 471 nonworking school children. Nearly all market children perform complex arithmetic correctly in real transactions, yet less than one-third pass basic abstract division and subtraction exercises. School children show the reverse: performing well on abstract problems but struggling on simulated market transactions, even with unlimited time and paper. We rule out stress, incentives, and unfamiliarity as explanations through embedded randomized experiments. These findings point to a fundamental failure of skill transfer in both directions and call for curricula that explicitly bridge formal and intuitive mathematical reasoning. JEL Codes: O12, I25, J16
MIT Department
Massachusetts Institute of Technology. Department of Economics
Terms of Use
In Copyright - Educational Use Permitted
Copyright retained by author(s)
Persistent DSpace Link
