WORKING PAPERS

"Worker Productivity Impacts of Climate Change Induced Salinization" (joint with Shonchoy, A. and Ravindran, S.) 

Climate change-induced salinization (CCIS) has aggravated the global safe drinking water crisis, particularly in vulnerable coastal regions of low-income countries. However, little is known about the economic impacts of high salt intake resulting from CCIS. Using 12 rounds of high-frequency data collected from 615 factory workers in Khulna, Bangladesh, we examine the effects of high salt exposure on worker productivity along both intensive (worker efficiency conditional on showing up for work) and extensive (worker absenteeism due to sickness) margins. Utilizing within-worker variation over time and exogenous variation over the hotter summer months, we find that high salt intake via drinking water is associated with a 1.69 percentage-point (3%) decrease in worker efficiency and a 0.5-day (96%) increase in absenteeism due to illness for workers with above-median age. There is significant heterogeneity by gender and type of work - efficiency losses and absenteeism are larger for male workers and those engaged in physically intensive work. We identify hypertension as the likely mechanism: above median-age workers exposed to high salt intake are 53% more likely to be hypertensive. While the strong negative link between high salt intake and worker efficiency, mediated by hypertension, underscores the significant economic losses and public health risks posed by CCIS, our findings on the short-term responsiveness of blood pressure to salt intake also suggest that targeted policy interventions could effectively reverse these adverse effects, safeguarding both health and livelihoods in affected communities.

“Mind the Data Gaps: An Examination of Women-Owned Enterprise Representation” (Joint with Hardy M. and Kagy G.)

We document large variations in women-owned enterprise representation and estimates of gender gaps in enterprise performance across commonly available data sources in Sub-Saharan Africa. We provide empirical evidence that these differences are driven by differences in gender-blind sampling protocols. Women-owned enterprises are less likely to meet the sampling criteria for most widely available enterprise data. Those that do meet the criteria are more positively selected on performance, relative to male-owned enterprises. We also document differences in implied policy and research priorities across these data sources, driven both by target sample characteristics and survey structure. Our findings caution the secondary use of commonly available enterprise data for gender equity and inclusion research and highlight a potential representation issue in insights drawn from such data.

“Lottery-Based School Admission: Effects on Classroom Diversity and Academic Achievement” 

I examine the impact of a lottery-based school admission system on socioeconomic diversity in the classroom and academic achievement in Bangladesh. In 2010, aiming to reduce the anxiety that competitive, exam-based school admission systems bring into preschoolers’ and parents’ lives, and to promote equal access to high-achieving schools, the government of Bangladesh introduced lotteries for allocating Grade 1 spots in schools in case of oversubscription. There is a variation in the timing of policy reform implementation. Exploiting the variations within schools and cohorts, and using the difference-in-differences regression model, I find that the reform enhanced the access of low socioeconomic status (SES) students to relatively high-achieving schools. One important concern is that by placing low-SES students in high-performing schools, the policy may harm high-SES students through negative peer effects, and low-SES students if they are ill-prepared for such an academic environment. I find no evidence of such adverse effects on academic outcomes at the end of elementary-level education.

"Human Capital and Transformative Entrepreneurship. Evidence from a Regression Discontinuity Design," (joint with Nikolov, P.) 

Using a “fuzzy” regression discontinuity design, we examine the short-run impacts of a large skills training program in Nepal on business ownership, business practices, and business performance. We find, approximately after twelve months of the start of the program, that the youth skill training generated a significant positive increase in business ownership of at least 8 percentage points (based on ITT estimates) and up to 30 percentage points for compliers (based on LATE estimates) from a baseline mean of 11 percent. There is an increase in sales record keeping practice among program participants compared to the control group. We also detect an average improvement of 7 to 31 percentage points (ITT and LATE estimates, respectively) in association with formal trade and business organizations. However, among the microenterprises, we find no significant positive impacts on business performance in terms of monthly business earnings and having hired employee. The overall positive impacts on business ownership and business practices seem to be driven by service and production-based businesses which are primarily operated by women – i.e., beautician, food producer, weaving and garments etc. We find that women, on average, benefit far more compared to men in terms of gaining skills, applying standard business accounting practices, and getting involved into home-based businesses.