Research

Job seekers in India are randomly shown information about local labor market conditions drawn from a popular job portal — job postings, applicants, or both, for their preferred city and occupation. Treated respondents are more likely to be employed seven months later, though the effect varies considerably by baseline employment status.

Using data on 1.5 million Sri Lankan migrants to the Gulf region, this paper tests whether a regulator's public rating of recruitment intermediaries improves placement quality. Agencies eligible for the program, especially previously under-performing ones, invest in the rating criteria and are less likely to exit the market; eligible agencies see more and higher-quality foreign demand once ratings are revealed, and migrants placed through them receive higher salaries and encounter less abusive employers.

Exploiting umpiring reforms in international cricket, where two umpires make independent decisions in each other's presence, this paper shows that home-team umpires are less biased when working with a neutral colleague — one who is a national of neither team. The temporary debiasing is driven by social pressure to appear impartial in front of a neutral peer.

A 2013 Sri Lankan policy restricted women's migration for foreign employment based on their own age and their children's ages. The policy sharply reduced female migration to the Gulf region and household remittance income, lowering consumption. Affected mothers respond by reducing fertility, while their children experience worse outcomes, including reduced school attendance.

Using the India-Pakistan cricket rivalry as a shock to national identity, this paper finds that when India loses to Pakistan, Indian districts with a significant Muslim minority see a seven-fold increase in rioting — a pattern absent after India loses to other countries. The result points to the fragility of national identity for religious minorities in the aftermath of an intergroup loss.

A cluster-randomized trial of mobile agricultural extension across villages with existing farmer learning groups finds that untreated group members are no more likely to gain knowledge or change practices than farmers in villages without the service, suggesting that information spillovers from treated to untreated peers may be small.

In a randomized experiment with 1,500 Indian firms, providing job portal advertising together with the ability to verify applicant identity increases portal-based hiring by 68 percent and the likelihood of filling a vacancy by 11 percent. Advertising attracts more skilled applicants, while verification lets firms screen unfamiliar candidates; neither treatment alone has much effect.

Studying land inheritance in rural India, this paper finds that while larger inheritances raise future household consumption on average, first-born sons do not share in these gains. Inheriting land reduces migration and entry into non-agricultural work for first-borns, consistent with cultural norms that place filial responsibility on the eldest son; later-born sons' occupational choices and migration are unaffected.

Higher wages in Gulf countries are often paired with poor working conditions and employer malpractice. Using administrative data on complaints Sri Lankan migrants filed with their consulates, this paper documents that exposure to fraud and abuse is systematically linked to characteristics of the local recruitment agency that placed the migrant.

Randomizing access to a mobile-phone agricultural extension service in India, this paper shows that treated farmers are more likely to recommend inputs to their peers, and that peers of treated farmers change their own practices. ICT-based advisory services can encourage peer interaction and help technology diffuse beyond direct recipients.

A two-year evaluation of a voice-based ICT advisory service for cotton and cumin farmers in India finds that farmers substantially change their sources of information and consistently adopt inputs recommended by the service, raising cumin yields by 28 percent and cotton yields by 8.6 percent for a sub-group that received reminders, though overall yield gains remain unclear.

This paper evaluates the Becker-DeGroot-Marschak (BDM) mechanism for incentive-compatible demand elicitation outside the lab. In two field experiments — for a new agricultural information service and for rainfall index insurance — demand curves estimated from BDM align closely with those from individually randomized fixed prices for the information service, with more mixed results for insurance. Overall, the paper finds no evidence of systematic bias, supporting the use of discrete-price BDM in low-literacy field settings.