Research areas and output

Papers searching for home!

Building Bridges or Walls? The Dynamics of Cooperation in Diverse Societies

This paper examines the emergence of cooperation in heterogeneous societies, where individuals trade off coordination within their own group against cooperation across groups. I introduce a novel dynamic coordination model in which majority and minority group members pre-commit either to their group norm or to a costly bridging action enabling cross-group interaction. The model generates four self-enforcing social conventions, mutual cooperation, mutual non-cooperation, and two asymmetric conventions, and yields new predictions for when efficient cross-group cooperation is unstable. Asymmetries in group size and coordination incentives can make mutual cooperation unstable and generate persistent inequality even when cooperation is Pareto superior. A laboratory experiment with induced group identities supports these predictions: societies often converge to non-cooperation, and when cooperation arises, it is disproportionately undertaken by minorities, earning lower payoffs and unrealized efficiency gains. Together, the theory and experiment show that within-group coordination pressures can endogenously produce inefficient and unequal social norms, with persistent inter-group disparities in otherwise mutually beneficial environments.

Working Paper, Under review (last version: February 2026) 

Inequality of Opportunity and Discrimination (with Deepti Bhatia and Urs Fischbacher)

Identity-based discrimination continues to impact labor market outcomes across multiple contexts. While explicit forms of discrimination have been mitigated through policy and institutional interventions, subtle mechanisms persist. For instance, differential access to resources can lead to disparities in performance, potentially masking underlying biases. We investigate whether inequality of opportunity influences discrimination through a laboratory experiment in India, focusing on caste-based discrimination and belief updating. We introduce inequality of opportunity by exogenously varying the degree of advantage in a previously performed task. We test whether providing information about an advantage affects the likelihood of hiring potential employees across groups and whether employers update their beliefs over time. Our results indicate that, conditional on being informed about the advantages, employers initially prefer employees from their own caste. Over time, they update their beliefs about the true ability distribution, leading to statistical discrimination in favor of upper-caste employees.

Working paper available upon request, Resubmited at Journal of Economic Psychology 

Affirmative Action: Efficiency and inequality (with Urs Fischbacher, Jan Hausfeld, and Yvette Lambi)

Inequality in wealth, income, and opportunities persists across various identity groups. Affirmative action (AA) is commonly used to address historical disadvantage and encourage members of marginalized groups to participate in competitive environments. Existing evidence suggests that AA increases beneficiaries’ willingness to compete, but largely overlooks heterogeneity within disadvantaged groups. Using a laboratory experiment, we investigate how within-group and between-group inequality, as well as affirmative action based on either group identity or income, influence competitive behavior in Tullock contests, policy preferences, fairness perceptions, and group preferences. We find that when groups are internally homogeneous, affirmative action increases beneficiaries' investment without discouraging non-beneficiaries. In contrast, under within-group heterogeneity, beneficiaries remain encouraged, while non-beneficiaries are discouraged. As unaffected social planners, participants prefer group-based affirmative action over income-based affirmative action, even though group-based affirmative action is considered the least fair option. We further find that out-group members with high endowments receive the most deduction points from other participants. These results show that ignoring within-group heterogeneity can obscure both the behavioral and distributional consequences of affirmative action, highlighting its importance for policy design. 

Working Paper, Submitted

Helping the kids to catch up: What predicts foundational learning in urban India? 

Many children in developing countries lack fundamental numeracy and literacy skills. No-detention policies imply that children are automatically promoted to the next higher grade regardless of their learning. A potential risk is that children who lack foundational skills lag and cannot follow the content being taught in the classroom. NGOs try to address this by providing remedial programs, especially for children from poorer backgrounds. We collaborate with an NGO in urban India that implements a low-cost program following the Japanese Kumon model in which students complete worksheets that slowly increase in difficulty. We examine the predictors of learning progression and student engagement using administrative data to measure learning and engagement, and survey data of parents and NGO teachers. Our results show that students with mothers who have completed at least secondary education have a higher learning progression and more frequent engagement. In comparison, students with fathers who have completed at least secondary education have a lower learning progression and less frequent engagement. Further, the teacher employed by the NGO is the most important predictor for both outcomes. Our empirical findings highlight that it is crucial to actively engage fathers in outreach activities and invest in the teachers' training and monitoring to improve learning progression and student engagement for all students. 

Status: Addressing the reviewer's comments after a rejection 

Working paper available upon request