Research
Research
Working Papers
Skill Obsolescence and the Consequences of Job Loss (Job Market Paper)
Abstract: This paper studies how within-occupation skill obsolescence shapes the consequences of involuntary job loss. I combine the NLSY79 worker panel with occupation–year task demands derived from vacancy text to construct a worker-specific measure of exposure to obsolescence. Exposure arises when a worker’s strongest (primary) skills, or multiple skills, become less valued in their pre-displacement occupation. I distinguish workers who face primary-skill exposure, high exposure, and those who are non-exposed. Using a staggered difference-in-differences design, I estimate dynamic effects of displacement separately for each group. Displacement causes sizable and persistent earnings and employment losses, which are much larger for exposed workers. In the first year after displacement, exposed workers lose roughly 30% of their pre-displacement earnings, compared with about 20% for non-exposed workers, and their earnings remain depressed even eight years later. A decomposition shows that these persistent losses operate mainly through the extensive margin: exposed workers spend more weeks out of work rather than working fewer hours or at much lower wages. Following displacement, exposed workers move to occupations that are farther away in task space but better aligned with their underlying skills, indicating that restoring skill alignment requires longer and more costly search.
Economic Growth and Employment under Innovation Policies
Abstract: This paper studies the impact of innovation policies on productivity growth and unemployment when job matching is frictional. I build a DSGE model of Schumpeterian creative destruction with search-and-matching frictions, featuring two opposing channels: a capitalization effect that spurs vacancy creation and a creative-destruction effect that raises separations. Calibrated to U.S. data (2003–2019), the model evaluates three instruments—direct R&D wage subsidies, incremental R&D tax credits, and corporate profit-tax cuts. Stronger, targeted innovation support raises long-run TFP growth but increases equilibrium unemployment, revealing a transparent growth–unemployment trade-off. Doubling direct subsidies lifts TFP growth by 0.04 percentage points and delivers a 1.5% consumption-equivalent welfare gain; doubling tax credits yields smaller growth and welfare effects but superior cost-effectiveness per fiscal dollar. Broad corporate tax cuts reduce effective R&D incentives, lowering growth and welfare despite slightly lower unemployment.
Worker-Level Exposure to Within-Occupation Skill Obsolescence
Abstract: This paper introduces a worker-level measure of exposure to within-occupation skill obsolescence and uses it to examine how workers' wages and occupational mobility are associated with changing occupational skill demand. The measure combines monthly NLSY79 work histories from 1978 to 2000 with vacancy-text-based measures of occupational skill demand, allowing workers in the same occupation to differ in exposure depending on their baseline cognitive, manual, and interpersonal skill endowments. I distinguish between primary-skill exposure, where the worker's strongest skill becomes less central while other skill dimensions gain importance, and high exposure, where multiple skill dimensions decline in relative importance within the occupation. High exposure is associated with approximately 2 percent lower hourly wages and a higher probability of subsequent occupational switching, while primary-skill exposure is not systematically associated with either outcome. Conditional on switching, high-exposure and primary-exposure workers exhibit smaller wage losses than non-exposed switchers, consistent with selective reallocation. The findings indicate that the breadth of within-occupation skill-demand change, rather than the decline of any single skill dimension, is the relevant margin for understanding worker adjustment to technological change.