Job Assignment
Artificial Intelligence and the Labor Market
Menaka Hampole et al.
NBER Working Paper, February 2025
Abstract:
We leverage recent advances in NLP to construct measures of workers' task exposure to AI and machine learning technologies over the 2010 to 2023 period that vary across firms and time. Using a theoretical framework that allows for a labor-saving technology to affect worker productivity both directly and indirectly, we show that the impact on wage earnings and employment can be summarized by two statistics. First, labor demand decreases in the average exposure of workers' tasks to AI technologies; second, holding the average exposure constant, labor demand increases in the dispersion of task exposures to AI, as workers shift effort to tasks that are not displaced by AI. Exploiting exogenous variation in our measures based on pre-existing hiring practices across firms, we find empirical support for these predictions, together with a lower demand for skills affected by AI. Overall, we find muted effects of AI on employment due to offsetting effects: highly-exposed occupations experience relatively lower demand compared to less exposed occupations, but the resulting increase in firm productivity increases overall employment across all occupations.
Automation and the rise of superstar firms
Hamid Firooz, Zheng Liu & Yajie Wang
Journal of Monetary Economics, April 2025
Abstract:
We provide empirical evidence suggesting that the rise of superstar firms is linked to automation. We explain this empirical link in a general equilibrium framework with heterogeneous firms and variable markups. Firms can operate a labor-only technology or, by paying a per-period fixed cost, an automation technology that uses both workers and robots. The fixed costs lead to an economy-of-scale effect of automation, such that larger and more productive firms are more likely to automate. Automation boosts labor productivity, allowing those large firms to expand further, raising industry concentration. Since robots substitute for workers, increased automation raises sales concentration more than employment concentration, consistent with empirical evidence. Under our calibration, a modest robot subsidy mitigates markup distortions and improves welfare by stimulating automation investment, bringing aggregate output closer to the efficient level.
Where Have All the Good Jobs Gone? Changes in the Geography of Work in the US, 1980-2021
Gordon Hanson & Enrico Moretti
NBER Working Paper, March 2025
Abstract:
We examine changes in the spatial distribution of good jobs across US commuting zones over 1980-2000 and 2000-2021. We define good jobs as those in industries in which full-time workers attain high wages, accounting for individual and regional characteristics. The share of good jobs in manufacturing has plummeted; for college graduates, good jobs have shifted to (mostly tradable) business, professional, and IT services, while for those without a BA they have shifted to (non-tradable) construction. There is strong persistence in where good jobs are located. Over the last four decades, places with larger concentrations of good job industries have tended to hold onto them, consistent with a model of proportional growth. Turning to regional specialization in good job industries, we find evidence of mean reversion. Commuting zones with larger initial concentrations of good jobs have thus seen even faster growth in lower-wage (and mostly non-tradable) services. Changing regional employment patterns are most pronounced among racial minorities and the foreign-born, who are relatively concentrated in fast growing cities of the South and West. Therefore, good job regions today look vastly different than in 1980: they are more centered around human-capital-intensive tradable services, are surrounded by larger concentrations of low-wage, non-tradable industries, and are more demographically diverse.
Measuring Work from Home
Shelby Buckman et al.
NBER Working Paper, February 2025
Abstract:
Headline estimates for the extent of work from home (WFH) differ widely across U.S. surveys. The differences shrink greatly when we harmonize with respect to the WFH concept, target population, and question design. As of 2025, our preferred estimates say that WFH accounts for a quarter of paid workdays among Americans aged 20-64. The WFH rate is seven percentage points higher for workers with children under eight in the household and about two percentage points higher for women than men. Desired WFH rates exceed actual rates in every major demographic group -- more so for women, workers with young children, and less educated workers.
The New Geography of Labor Markets
Mert Akan et al.
NBER Working Paper, March 2025
Abstract:
We use matched employer-employee data to study where Americans live in relation to employer worksites. Mean distance from employee home to employer worksite rose from 15 miles in 2019 to 26 miles in 2023. Twelve percent of employees hired after March 2020 live at least fifty miles from their employers in 2023, triple the pre-pandemic share. Distance from employer rose more for persons in their 30s and 40s, in highly paid employees, and in Finance, Information, and Professional Services. Among persons who stay with the same employer from one year to the next, we find net migration to states with lower top tax rates and areas with cheaper housing. These migration patterns greatly intensify after the pandemic and are much stronger for high earners. Top tax rates fell 5.2 percentage points for high earners who stayed with the same employer but switched states in 2020. Finally, we show that employers treat distant employees as a more flexible margin of adjustment.
The Effect of Franchise No-Poaching Restrictions on Worker Earnings
Brian Callaci et al.
Review of Economics and Statistics, forthcoming
Abstract:
We evaluate the nationwide impact of the Washington State Attorney General's 2018-20 enforcement campaign against no-poach clauses in franchising contracts, which prohibited worker movement across locations within a chain. Implementing a staggered difference-in-differences research design using Burning Glass Technologies job vacancies and Glassdoor salary reports from numerous industries, we estimate a 6% increase in posted annual earnings from the job vacancy data and a 4% increase in worker-reported earnings.
Right-to-work laws and venture capital investment
Helena Sarkodie et al.
Journal of Banking & Finance, March 2025
Abstract:
Using state-level data from the United States covering the period 1980 to 2020, we explore the effect of right-to-work (RTW) laws on venture capital (VC) investment. Employing a difference-in-differences strategy, we find that the passage of right-to-work laws increases venture capital investment. The results are robust to omitted variable bias, reverse causality and unobservable local economic conditions. We find that the positive effect of RTW laws on VC investments remains significant in states that are highly unionized and technological.
Working More to Pay the Mortgage: Household Debt, Interest Rates, and Family Labor Supply
Michał Zator
Journal of Finance, April 2025, Pages 1171-1207
Abstract:
I show that households work and earn more (less) when their floating-rate mortgage payments quasi-exogenously increase (decrease). The response is sizable and asymmetric: on average, households adjust their income by 35% of the change in their mortgage payment, but the response is significantly stronger following an increase in payments. While men in dual-earner, childless households respond the most on average, the asymmetry is most pronounced for women and young workers, who respond particularly strongly to payment increases. The asymmetry of the labor supply elasticity may help explain the wide range of elasticities found in previous research.