Essays on Labor Market Frictions in Developing Countries

Abstract

This dissertation comprises three essays on economic development, examining how labor market power, outside options, and social structure shape worker outcomes — and how policy can either reinforce or disrupt these dynamics. The first chapter studies whether redirecting referrals toward underrepresented workers in segmented labor markets can simultaneously improve equity and productivity. In an Indian manufacturing firm, I experimentally increased the share of referral invitations allocated to lower-caste minority incumbents. The policy raised their employment share by 15 percentage points (62%) and increased team output by 0.09 standard deviations (5%), driven by a 4 percentage point (41%) reduction in monthly turnover. Contrary to common concerns about team diversity, treatment did not reduce worker cohesion. A lab-in-field experiment pins down the mechanism: it is recruitment mode — not entrant identity — that drives short-run costs, as lower-caste workers entering as outsiders rather than via referrals reduced cohesion by 9% and output by 22%. Supervisors exposed to the policy continued allocating referrals to lower-caste workers after the intervention ended, consistent with Bayesian updating. Lower-caste referral candidates saw large gains in job offers and employment with no detectable displacement of upper-caste candidates, showing that redirecting referrals can reassign jobs toward workers with low outside options while improving firm performance. The second chapter, coauthored with Steven Brownstone, studies how differences in village land concentration stemming from feudal land grants hundreds of years ago shape present-day labor markets in India. Implementing a regression discontinuity along feudal borders that no longer correspond with modern administrative boundaries, we find that large discontinuities in land concentration persist across these boundaries, with the smallest land parcels in previously feudal areas 20% larger than their non-feudal counterparts. These differences are associated with 8% lower agricultural wages for women — but not men, who are more able to travel and access outside options — despite no differences in yields, aggregate labor demand or supply, output prices, or other non-labor inputs. Village elected bodies in feudal areas scuttle the implementation of India's flagship workfare program, with 71% fewer person-days offered during peak agricultural months when large landowners need labor, and no difference in the lean season when the program is most active. This suppression operates through caste-based ties between large landowners and village elected representatives, highlighting land inequality as a key mechanism through which labor market inequities persist and emphasizing the value of outside options where employers hold market power. The third chapter, coauthored with Naveen Kumar, studies how targeted schooling policies affect both human capital and social networks, with consequences for labor market mobility. Exploiting oversubscribed Grade 5 admissions lotteries to a system of high-quality residential schools in India designed for disadvantaged castes, we follow a cohort of students to ages 21 and 25. Attendance raises completed schooling by 0.38 years, Grade 12 scores by 0.26 standard deviations, and college enrollment by 19% — closing nearly one-third of the caste gap in educational attainment. However, these human capital gains come at a cost to social networks: treated students form peer ties that are 9% smaller and 89% more caste-homogeneous, resulting in narrower job-search networks and a lower probability of finding employment through referrals. Labor market participation initially rises but later converges as unemployment falls. A simple search model illustrates how higher ability raises the returns to search while homophilous networks slow job-offer arrivals. The findings show that targeted schooling can close learning gaps without closing mobility gaps when opportunity flows through socially segregated networks. Together, these chapters show that labor market inequality in developing countries is reproduced not only through differences in skill or effort, but through the social structures that govern access to work, information, and opportunity. Referral networks, concentrated land ownership, and segregated schooling each shape workers’ outside options in ways that can entrench advantage across settings and over time. The broader lesson is that improving productivity or expanding formal access is often not enough: when opportunity is mediated by unequal social and institutional arrangements, the returns to those gains remain uneven. Reducing inequality therefore requires not only raising workers’ capabilities, but also reshaping the structures through which labor markets allocate opportunity.