Research

Labor market trends and unemployment insurance generosity during the pandemic (with Lucas Finamor)

Economics Letters 199 (2021), 109722.

Selected press coverage: The Wall Street Journal, The Washington Post, Bloomberg, CNBC, Vox, NPR's All Things Considered, Marketplace.

"Who Gets a 'Good' Job? Firmwide Amenity Setting and the Structure of Labor Market Inequality" (Job Market Paper)

"Flexibility for Both Parents: Remote Work and the Evolution of Child Penalties” (with Elin Sundberg) – draft available upon request.

Abstract: Child penalties--the reductions in employment and earnings experienced by women after becoming parents--are a major driver of gender gaps in the labor market. This paper investigates the potential for remote work to mitigate these penalties. We leverage rich administrative data from Sweden to study this question in the context of the rise of remote work during and after the COVID-19 pandemic. We develop a collective model of household labor supply in which each agent can also choose whether or not to work flexibly. In the model, flexibility imposes a trade-off: on the one hand, it reduces commuting frictions; on the other, it may reduce workers' productivity. The model predicts that flexibility reduces motherhood penalties when primarily accessible to women -- i.e., if it shifts the relative opportunity cost of work for the two parents -- but has limited effects when both partners gain flexibility. Consistent with these predictions, we find that remote work has had limited impacts on motherhood penalties in earnings, employment, hours, and wages. Although we find precise but economically small reductions in employment and hours penalties for mothers, the effects on earnings and wage rates are more ambiguous. Furthermore, because remote work exposure is inversely correlated with ex ante motherhood penalties, the women who would stand to gain the most from flexibility are also the least likely to have actually gained access to it.

"Is the Grass Always Greener? Imperfect Information about Amenities and Job-to-Job Transitions"

Abstract: Non-wage amenities vary widely across firms, shaping worker preferences and sorting patterns in the labor market. Some amenities—such as paid vacation or remote work policies—are observable to job seekers during the search process, while others—such as job stress, autonomy, or respect from colleagues—are only revealed after a worker accepts the job. Similarly, information about pay may be partially observed during search. This paper studies how these information gaps shape worker decision-making by examining employment-to-employment (E:E) transitions, where workers have full information about the job they are leaving but only partial information about the job they are accepting. I quantify the amenity content and pay transparency of jobs using two complementary sources: (1) a novel AI-based labeling of publicly posted French job ads to infer advertised amenities and compensation information, and (2) large-scale worker surveys matched to administrative data capturing realized pay and ex post job characteristics. I develop a simple model of job choice with asymmetric information to interpret patterns of revealed preference and identify the role of informational frictions in shaping job mobility.

“The Disappearing Rung: AI and Moral Hazard in Human Capital Formation”

Abstract: This paper examines how AI-driven task automation alters the incentives for human capital investment in settings where skill acquisition is sequential and experience-based. Individuals often build expertise by progressing from simple, repetitive tasks to more complex, cognitively demanding work. AI tools can now substitute for many entry-level tasks, raising concerns that such substitution may disrupt learning-by-doing and reduce opportunities to acquire foundational skills. At the same time, AI may complement higher-skill workers by allowing them to offload routine components and focus more intensively on complex tasks. I develop a model that captures the static productivity gains and dynamic incentive distortions introduced by AI. The framework highlights how moral hazard can arise when short-run performance incentives encourage early reliance on automation, potentially crowding out long-run human capital accumulation. I derive conditions under which AI adoption enhances or erodes skill development, and discuss implications for organizational design, labor market mobility, and the regulation of AI usage in education and workplace settings.

"Race and Consumer Bankruptcy" (with Paul Goldsmith-Pinkham and Jialan Wang)

Employment Effects of Unemployment Insurance Generosity During the Pandemic (with Joseph Altonji, Zara Contractor, Lucas Finamor, Ryan Haygood, Ilse Lindenlaub, Costas Meghir, Cormac O’Dea, Dana Scott, Liana Wang, and Ebonya Washington)

Policy Report, Yale University Tobin Center for Economic Policy, 2020.