Research
2025
- Why Remote Work Stuck: A Structural Decomposition of the Post-Pandemic Equilibrium2025
The post-pandemic labor market featured a persistent increase in remote work arrangements. We ask whether this re-valuation of work is mostly due to a shift in worker preferences for flexibility or advances in technology that make remote work more feasible. We develop and estimate a general equilibrium search model featuring heterogeneity in worker skill and idiosyncratic worker tastes. A key innovation is our novel, continuous measure of occupational teleworkability, which allows the model to capture rich heterogeneity in remote-work potential across firms and occupations. We estimate the model’s deep parameters via the Simulated Method of Moments (SMM), disciplined by rich microdata from two distinct periods: 2019 (pre-pandemic) and 2024 (post-pandemic). Our analysis reveals that the shift is overwhelmingly driven by a profound revaluation of in-office time by workers. A structural decomposition shows that this preference shock accounts for 56.8% of the total increase in the average share of remote work, while concurrent shocks to remote work technology account for 33.0% of the shift.
- AI’s Dual Impact on Labor Markets: Automation, Augmentation, and Human Capital2025
This paper examines whether AI’s dual effects on labor markets—automation and augmentation—operate through distinct human capital channels. Using Microsoft’s "Working with AI" telemetry data matched to O*NET occupations and BLS employment statistics (N=107,901 observations), I test whether automation primarily substitutes for codified knowledge (formal education) while augmentation complements tacit knowledge (on-the-job experience). Preliminary results suggest education strongly predicts both automation and augmentation exposure, with surprising patterns for experience-based knowledge. I develop an overlapping generations model integrating Acemoglu-Restrepo task-based production with Ben-Porath human capital accumulation to explain age-heterogeneous impacts observed post-ChatGPT.
- Industry Workforce Heterogeneity and Wage Inequality2025
I reexamine the capital-skill complementarity hypothesis at the industry level as the driver of the increase in wage inequality between skilled and unskilled workers. Using the model proposed by Krusell, Ohanian Ríos-Rull, and Violante (2000), I decompose the growth of the skill premium into counteracting effects: (i) the negative effect in the relative price of skilled labor due to its relative increase in supply, and (ii) the positive effect of the increase in the marginal productivity of skilled workers relative to unskilled due to technological change (capital-skill complementarity). I show that for 78.6% of the industries in the sample (22 out of 28 industries) the demand effect due to capital-skill complementarity dominates the supply effect. The analysis includes a comprehensive decomposition showing that equipment investment complementary to skilled labor more than offset the downward pressure from increased college graduate supply, driving up skill premiums across most industries.