Research

How do information frictions distort the allocation between individuals and industries? Using confidential Canadian administrative data linking education and employment histories, I find that information frictions regarding individual industry skills generate persistent mismatch. I establish three key facts. Firstly,  mismatched individuals switch industries more. Secondly, skill uncertainty is partially resolved with tenure. Thirdly, using a Difference-in-Differences design that leverages LinkedIn’s market entry, I confirm that information improves labor market outcomes. To quantify the welfare cost, I develop and calibrate a life-cycle directed search model with Bayesian learning. Multidimensional skill workers choose majors and industries, and climb the job ladder within an industry. Imperfect information steers graduates to suboptimal majors, industries, and job rungs. Unresolved uncertainty about outside options, combined with search frictions, makes mismatch persistent. Average output per employed worker under perfect information 25% higher at entry, conditional on baseline major choices. This difference narrows to 12% after 30 years of on-the-job learning.  A counterfactual that shuts down this learning reveals a reallocation towards majors with lower uncertainty, especially for high-ability individuals. 

This paper develops an equilibrium theory of matching between buyers and sellers in the real estate market, especially investigating how partially committed asking prices respond to the pool of prospective buyers associated with each good. Buyers with heterogeneous financial abilities visit based on expected gain, suggesting that the pool of prospective buyers faced by the sellers depends on expected competition induced by the asking price. In a search market with asking prices, I show analytically that sellers optimally post lower asking prices when the targeted market is more competitive. I also show that the model-predicted sale-over-asking ratio is consistent with the empirically observed evidence from the Toronto real estate market.

Artificial Intelligence (AI) is reshaping returns to human capital. This paper examines how AI affects the value of work experience in entrepreneurship. Using employment histories from public LinkedIn profiles (2007–2019), we exploit industry level variation in AI exposure following the diffusion of neural networks and ImageNet after 2012. We find that both the share of founders and researchers increased, but entry gains were concentrated among more-experienced workers, especially those with research backgrounds. To understand the mechanism behind AI’s impact on the labor market, we develop a directed search model with occupational choice, multidimensional skills, and stochastic human capital investment. The model shows that AI shocks increase the productivity premium for researchers, shifting entrepreneurship toward more experienced individuals.

The informal sector is often viewed as a buffer during economic downturns, absorbing workers displaced from the formal sector and mitigating unemployment spikes. Using panel data from Continuous National Household Sample Survey (PNADC) between 2012 to 2018, we examine the short‐ and long‐term consequences of informal employment in Brazil across the business cycle and establish several new empirical facts. We observe that informal sector expands during Recession, consistent with the literature, indicating that informal sector acts as a buffer for workers. Our new finding is that a brief spell in the informal sector, lasting at most one quarter, increased the probability of formal re‐entry relative to unemployment. However, prolonged informal employment sharply reduced re‐entry probabilities into formal sector, with this scarring effect persisting after controlling for individual characteristics.

To interpret these patterns, we develop a directed search model with human capital depreciation, where depreciation depends on employment type and spell length. The framework captures the observed dual role of the informal sector as both a short‐term safety net and a long‐term trap. When designing labor market policies, our findings show that “when” to act is as important as “what” to do. Preserving the short‐term benefits of the informal sector requires timing as well as targeting, a dimension the literature has largely overlooked.