Research

This paper studies how internet adoption reshapes firm production, internal organization, and management practices. We first develop a model where adoption simultaneously lowers multiple managerial, production, and input frictions within a firm. While adoption improves performance, the precise channels through which these gains materialize remain ambiguous. Using detailed data on production, novel data on ICT usage, and a long panel of Colombian manufacturing firms, we then empirically study the impact of broadband adoption spanning over a decade. Firms digitalize along multiple dimensions: externally --- through websites, online orders, exports, and e-commerce; and internally --- through digital communication, online worker hiring, and scaling-up investments in capital and frontline production labor. Quantitative estimates suggest that while better market access, ICT productivity, and lower labor hiring costs are important drivers of adoption, they each explain only 6-8% of the growth in real income over this period. These findings demonstrate the need to incorporate multiple, interdependent dimensions to assess the impact of technology adoption on firm performance, productivity, and welfare.

How does institutional design and governance architecture impact the implementation of industrial policies? We study this using the universe of applications and approvals for deforestation, a policy reform, and a structural model of applications, decisions, and government preferences for development vs conservation.

Women seeking to re-enter the workforce after career breaks often face substantial barriers to employment. We partner with a leading Indian organization to evaluate targeted, supply-side interventions to support their re-entry into the labor market.

We use nationally representative data from nearly 100 countries between 1960-2020 to document distinct gender patterns in employment across both sectors and occupations. Using a model-based approach, we find that declining gender barriers were a key driver of rising female labor force participation, the expansion of the service sector, and growth in real GDP per capita over this time period.

When competition for a job increases, individuals are less likely to share information—especially with high-ability peers—leading to lower-quality applicants and hires. Firms relying on social networks to recruit may therefore attract weaker candidates for competitive roles. While higher wages improve applicant quality, they don’t fully offset the disincentives to share information.

Expansion of 3G mobile internet in developing countries significantly increases employment, especially among women. While both men and women shift into service sector wage jobs, men move out of unpaid farm work into small agricultural enterprises, whereas women take on more unpaid roles and operate more small businesses. The results highlight a strong gendered dimension in how digital access reshapes labor markets.

We show that employment protection laws (EPL) that penalize formal firms for hiring workers on contract can have unintended consequences in economies with high informality. While such EPL reduce informality on the "intensive margin"—by discouraging contract employment within formal firms—it increases informality on the "extensive margin", as some firms exit the formal sector entirely. This raises overall informality and leads to declines in aggregate productivity, wages, and welfare.

Low female LFP in India explains much of the gender gap in self-employment, but not in entrepreneurship. Interpreted through the lens of a model, we find large, persistent barriers to firm growth for women, even in richer states—highlighting the need for multi-dimensional, state-specific policy solutions.

We evaluate an innovative remedial education program targeted at secondary school students in India and find substantial gains in basic literacy and numeracy skills. However, these improvements do not translate into better test scores on a high-stakes national exam. In fact, even students with severe learning deficits perform reasonably well on the exam, raising questions about its ability to meaningfully assess student proficiency. These findings highlight the promise of remedial programs, implemented even in advanced grades, but suggest a misalignment between learning and assessment. As education systems scale promising interventions, reliable assessments of student ability through national exams are a critical challenge, reflecting the need for an assessment reform.

We develop a new machine learning-based method to determine firms’ political connections in India and combine it with long-run firm-level financial data to see how they responded differentially after a large macroeconomic crisis.

We develop a framework to measure barriers to women’s labor force participation and entrepreneurship in India. We find that while women face high costs in growing firms as opposed to starting them, they are more successful in hiring other women—a pattern not explained by sectoral sorting. Removing these barriers would significantly boost female LFP, earnings, and aggregate productivity through reallocation from less productive male-owned firms to more productive female-owned firms.

We document the presence of matching frictions in the Indian labor market. Job-seekers have widely varying preferences over the same jobs, yet placement officers lack accurate knowledge of these preferences. Providing officers with this information improves short-term matching outcomes, though the effects fade by six months.

Some stylized facts and theories about dowries in India, using data on 74,000 marriages between 1930-2000.

We study how political institutions shape government responses to national crises, focusing on COVID-19 policy actions across 125 countries. We find that non-democracies implement stricter measures earlier, but democracies soon match or surpass them—especially in health-related policies. Democracies with greater electoral strength and longer time to the next election respond more aggressively. Media freedom slows containment, but enhances health interventions.