A summary of my research portfolio and longer term agenda is here
Cognitive Endurance as Human Capital (available open access from QJE)
with Supreet Kaur, Geeta Kingdon and Heather Schofield, The Quarterly Journal of Economics, Volume 140, Issue 2, May 2025, Pages 943–1002
Abstract: Schooling may build human capital not only by teaching academic skills but by expanding the capacity for cognition. We focus specifically on cognitive endurance: the ability to sustain effortful mental activity over a continuous stretch of time. As motivation, we document that globally and in the United States, the poor exhibit cognitive fatigue more quickly than the rich do across field settings; they also attend schools that offer fewer opportunities to practice thinking for continuous stretches. Using a field experiment with 1,600 Indian primary school students, we randomly increase the amount of time students spend in sustained cognitive activity during the school day—using either math problems (mimicking good schooling) or nonacademic games (providing a pure test of our mechanism). Each approach markedly improves cognitive endurance: students show 22% less decline in performance over time when engaged in intellectual activities—listening comprehension, academic problems, or IQ tests. They also exhibit increased attentiveness in the classroom and score higher on psychological measures of sustained attention. Moreover, each treatment improves students’ school performance by 0.09 standard deviations. This indicates that the experience of effortful thinking itself—even when devoid of any subject content—improves general cognitive capacity. Finally, we complement these results with quasi-experimental variation indicating that an additional year of schooling improves cognitive endurance, but only in higher-quality schools. Our findings suggest that schooling disparities may further disadvantage poor children by hampering the development of a core mental capacity.
Pre-registration: https://www.socialscienceregistry.org/trials/2673
Featured on The Report Card Podcast, the Freakonomics Podcast (Part 1 and Part 2), and PNAS Magazine
Inducing Positive Sorting through Performance Pay: Experimental Evidence from Pakistani Schools
with Tahir Andrabi
Abstract: Attracting and retaining high-quality teachers has a large social benefit, but it is challenging for schools to identify good teachers ex-ante. This paper uses teachers' contract choices and a randomized controlled trial of performance pay with 7,000 teachers in 243 private schools in Pakistan to study whether performance pay affects the composition of teachers. Consistent with adverse selection models, we find that performance pay induces positive sorting: both among teachers with higher latent ability and among those with a more elastic effort response to incentives. Teachers also have better information about these dimensions of type than their principals. Using two additional treatments, we show effects are more pronounced among teachers with better information about their quality and teachers with lower switching costs. Accounting for these sorting effects, the total effect of performance pay on test scores is twice as large as the direct effect on the existing stock of teachers, suggesting that analyses that ignore sorting effects may substantially understate the effects of performance pay.
Featured on VoxDev and Econimate.
Pre-registration: https://www.socialscienceregistry.org/trials/4471
Subjective versus Objective Incentives and Employee Productivity
with Tahir Andrabi
Abstract: A central challenge facing firms is how to incentivize employees. While objective, output-based incentives can motivate effort, they may lead employees to reduce effort on non-incentivized outcomes and may fail in settings where effort is weakly tied to output. We study the effect of subjective incentives (manager evaluation) and objective incentives (test scores) for teachers using an RCT in 234 Pakistani schools. First, we show that subjective and objective incentives increase test scores equally, relative to no incentives. However, objective incentives decrease non-test score student outcomes. Second, we show that teachers' effort response is very different under each scheme, with pedagogy only improving under subjective incentives. Finally, we rationalize these effects through the lens of a moral hazard model with multi-tasking. We use additional variation to isolate the causal effect of contract noise and distortion and show that these channels explain our reduced form effects.
Pre-registration: https://www.socialscienceregistry.org/trials/3835
Financial Favoritism and the Gender Gap in Performance Evaluations
Abstract: While we have extensive evidence on the prevalence of gender bias in hiring, promotions and wages, we know less about which personnel policies may mitigate or exacerbate these biases. To test this, I conduct a large scale field experiment with 2,000 employees in 150 schools in Pakistan and randomly vary i). whether performance evaluations conducted by managers affect employee’s pay or are just used for feedback and ii). how often managers observe a given employee. When performance evaluations determine teachers' raises, men receive 7% higher raises, despite similar or lower performance to women on a rich set of productivity measures. When there are no financial stakes, women receive slightly higher evaluations than men. However, the effect of financial stakes on the gender gap disappears when teachers are observed more frequently. To understand mechanisms, I conduct a follow up vignette survey to test whether our results are due to differential manager expectations about employee reactions to low raises (e.g. differential turnover or effort) or differential perceived deservedness of scarce financial resources. The results favor this second explanation as managers favoring single-earner (lower income) households over dual-earner (higher income) households, which is highly correlated with employee gender in places with low female labor force participation. Combined, this suggests the underlying bias is financial favoritism toward bread-winners but that better information about employee performance reduces the extent that managers act on this preference.
Understanding the Social Network Premium in Hiring Low-Wage Workers
with Maryiam Haroon
In South Asia, three-quarters of ultra-poor households report casual labor (short-term work in agriculture, construction, and manual labor) as their primary form of income. Employers seeking to hire laborers rely almost exclusively on their social network, meaning less connected and marginalized groups are often severely disadvantaged. In this project, we seek to understand why employers rely so heavily on social networks to hire. We partner with a construction firm and randomly vary the information, incentives, and monitoring structure for hiring managers to tease apart the five main hypothesizes that have been proposed to explain the role of social networks in hiring: information asymmetries, reciprocity, match-specific productivity, homophily or simply that more connected works are also more productive. We then test whether employers change their hiring patterns and laborers change their behavior under these different conditions. By understanding why employers rely on social networks, we can better design policies to serve socially disconnected workers.
This project is currently in the field. More on this project here.
Back to Basics: Experimental Evidence on Targeted Instruction
with Tahir Andrabi, Asim Khwaja and Isabel MacDonald
Across low-income countries, students lag substantially behind grade-level expectations. Combined with significant within-classroom variation in performance, this poses an instructional challenge. We conduct a randomized controlled trial with 40,000 students in grades 2–4 mathematics across 560 public primary schools in Pakistan to understand what constraints prevent teachers from tailoring instruction to students' current level. Teachers were randomized into a control group or one of three treatment arms that progressively relaxed resource, information, and planning constraints: a Workbook arm provided varying levels of each learning objective; an Information arm that added granular information on student topic mastery for teachers; and a Guidance arm that layers on top teaching plans specifying which content and practice problems to assign each student based on their baseline mastery. After a 40-day intervention, all three arms raised student math scores by similar magnitudes—0.18, 0.16, and 0.19 SD, respectively, equivalent to 0.4 years of additional learning. We cannot reject equality of treatment effects across the arms and find positive effects across the entire student ability distribution. This is despite classroom observations and teacher surveys confirming that the treatments produced genuinely different pedagogical behaviors with greater targeted instruction in the guidance arm. The results indicate that even in contexts with high within-class learning variation, focusing on more basic and lower-level learning objectives benefits all students, regardless of their initial learning levels, with minimal additional benefit from further personalization across students. This suggests that appropriately leveled curriculum and textbook materials are the first-order lever for addressing the foundational learning crisis.
Increasing Early Childhood Vaccination in Pakistan with Rachel Glennerster, Maryiam Haroon, and Leah Rosenzweig