With Colin Sullivan and Carl Meyer

Project summary, AEA RCT Registration

We combine two field experiments in Bangladesh with a structural labor model to define and test for paternalistic discrimination, the differential treatment of two groups to protect one group---even against its will---from harmful or unpleasant situations. We observe real hiring and application decisions for a night-shift job that provides safe worker transport home at the end of the shift. In the first experiment, we vary employers' perceptions of job costs to female workers by experimentally varying information about the transport but holding taste-based and statistical discrimination constant. Not informing employers about the transport decreases demand for female labor by 22%. However, employers respond significantly less to a cash payment to female workers that would allow them to purchase safe transport themselves. This suggests that employers paternalistically prevent women from making their own choices. In the second experiment, not informing applicants about the transport reduces female labor supply by 15%. In structural simulations that combine the results of both experiments, eliminating paternalistic discrimination reduces the gender employment gap by 24% and increases female wages by 21%.


With Erica Field, Rachel Glennerster, Shahana Nazneen, and Xiao Yu Wang


American Economic Review, 2023, Online appendix

Child marriage remains common even where female schooling and employment opportunities have grown. We experimentally evaluate a financial incentive to delay marriage alongside a girls’ empowerment program in Bangladesh. While girls eligible for two years of incentive are 19 percent less likely to marry underage, the empowerment program failed to decrease adolescent marriage. We show that these results are consistent with a signaling model in which bride type is imperfectly observed but preferred types (socially conservative girls) have lower returns to delaying marriage. Consistent with our theoretical prediction, we observe substantial spillovers of the incentive on untreated nonpreferred types. 


With Pascaline Dupas, and Roberta Ziparo 

Revision requested at the American Economic Review, Online appendix

We study reputation dynamics within the household in a setting where women regularly receive transfers from their husbands for household purchases. We propose a signaling model in which wives try to maintain a good reputation in the eyes of their husbands to receive high transfers. This leads them to (a) avoid risky purchases (goods with unknown returns); and (b) knowingly over-use low-return goods to hide bad purchase decisions - we call this the intra-household sunk cost effect. We present supportive evidence for the model from a series of experiments with married couples in rural Malawi.

With Erica Field, Rachel Glennerster, and Reshmaan Hussam


Media: The Economist, Vox, VoxDev, CATO Institute

We document the consequences of a public health campaign that led to the sudden abandonment of local water infrastructure by one-fifth of Bangladesh’s population. Households who experienced quasi-randomly distributed arsenic contamination, and thus were likely to abandon their shallow tubewells, saw 28% greater child and 47% greater elderly mortality post-campaign than those not motivated to shift. Verbal autopsy data reveal that the sudden mortality increases are driven by diarrheal disease. Mortality changes depend on the distance to alternative clean water infrastructure: those with an (arsenic and pathogen-free) deep tubewell within 300 meters of their home experience no increase in mortality, but mortality rises as households are forced to walk further for arsenic-free water. Our results quantify the mortality benefits of water infrastructure and underscore the importance of physical proximity, rather than mere access, to pathogen-free water sources.

With B. Douglas Bernheim, Zach Freitas-Groff, and Sebastián Otero

Media: Barack Obama, Saturday Night Live, Stephen Colbert, The New York Times, HuffPost, Los Angeles Times, Politico, Vox

We investigate the effects of large group meetings on the spread of COVID-19 by studying the impact of eighteen Trump campaign rallies. Our analysis encompasses up to ten post-rally weeks for each event to capture the effects of subsequent contagion within the pertinent communities. Our method is based on a collection of regression models, one for each event, that capture the relationships between post-event outcomes and pre-event characteristics, including demographics and the trajectory of COVID-19 cases, in similar counties. We explore a total of 24 procedures for identifying sets of matched counties. For the vast majority of these variants, our estimate of the average treatment effect across the eighteen events implies that they increased subsequent confirmed cases of COVID-19 by more than 250 per 100,000 residents. Extrapolating this figure to the entire sample, we conclude that these eighteen rallies ultimately resulted in more than 30,000 incremental confirmed cases of COVID-19. Applying county-specific post-event death rates, we conclude that the rallies likely led to more than 700 deaths (not necessarily among attendees).

With Erica Field, Rachel Glennerster, Shahana Nazneen, Iman Sen, and Svetlana Pimkina 

Online appendix Data

A clustered randomized trial in Bangladesh examines alternative strategies to reduce child marriage and teenage childbearing and increase girls’ education. From 2008, girls in treatment communities received either i) a six-month empowerment program, ii) a financial incentive to delay marriage, or iii) empowerment plus incentive. Data from 4.5 years after program completion show that girls eligible for the incentive for at least two years were 24% less likely to be married under 18, 15%  less likely to have given birth under 20, and 25% more likely to be in school at age 22. Girls eligible for the empowerment program were 11% more likely to be in-school at age 22. We also find significant and large effects of the empowerment program on income-generating activities (IGAs): an increase in an IGA index by 0.5SDs.

With Erica Field, and Rachel Glennerster

Data from a carefully crafted hypothetical survey of 750 professional matchmakers in rural Bangladesh enable a novel analysis of how demographic characteristics of brides and grooms, including wealth, education and age, influence marriage market prices (dowry and denmeher). While a standard Hedonic approach to decomposing marriage market prices is likely to be biased by selection on unobservables, hypothetical price questions circumvent this issue by varying key demographic features of the bride while holding constant all other characteristics of both spouses. This is akin to a stated preference approach to demand elicitation, but because we survey third-party negotiators rather than potential buyers, our approach is less prone to standard concerns of bias in stated preference techniques. Analysis of survey responses reveals education, labor market participation, and youth are valued in the marriage market and that parents of educated, working, and young brides can negotiate a higher denmeher/dowry ratio. These findings are in contrast to observational data, which finds no consistent return to education or labor market participation in the marriage market. 

Cost-Benefit Analysis of Strategies to Reduce Child Marriage in Bangladesh

With Erica Field, Rachel Glennerster, and Kyle Murphy

In much of the developing world, early female marriage—defined as marriage before the age of 18—remains widespread despite age of consent laws banning the practice, government and NGO efforts to curtail it, increasing education levels, and economic growth. Bangladesh suffers from the fourth highest rate of early female marriage in the world, which has remained very high despite large-scale efforts to combat the problem. As marriage and education are closely linked in many contexts, we calculate the education benefits of delayed marriage. We perform comprehensive comparative cost-benefit analyses of six interventions from South Asia, Latin America, and Sub Saharan Africa which have demonstrated significant impacts on marriage age and/or early marriage. For each study, we estimate the educational benefits by converting delayed marriage effects into additional schooling using a conversion factor from Bangladesh. We find that a program which provides girls with financial incentives conditional on marriage status is both the most cost-effective way to avert child marriages and has the highest benefit-cost ratio and net present value per $1,000 spent. 


How Changing the School System Affected Women's Labor Market Outcomes

With Muriel Niederle, and Nicola Fuchs-Schuendeln

Data analysis.

The Economics of Domestic Violence - Evidence from Bangladesh

Project summary; AEA RCT Registration

In the field. 

Intimate Partner Violence and the Emergence of Social Norms  - A Randomized Control Trial

With Paula Lopez Pena, Sakib Mahmood, and Atonu Rabbani

AEA RCT Registration; Winner of the 2019 World Bank/SVRI Development Marketplace: Innovations to address gender-based violence award

In the field. 

Long-Term Impacts of Adolescent Girls' Empowerment Programs in Bangladesh

With Erica Field, Rachel Glennerster, and Sakib Mahmood

In the field.

Champions of Change - Changing Gender Attitudes and Behaviors through Social Targeting

With Yasmine Bekkouche, Sarah Deschênes, and Rozenn Hotte

Funding received.

The Effect of Reducing Schooling on Education and Labor Market Outcomes in Germany

With Nicola Fuchs-Schuendeln

Polishing the draft.