Jobs of the Future
Forecasting the Economic Effects of AI
Ezra Karger et al.
Federal Reserve Working Paper, March 2026
Abstract:
We elicit forecasts of how AI will affect the U.S. economy, comparing the beliefs of five groups: academic economists, employees at AI companies, policy researchers focused on AI, highly accurate forecasters, and the general public. The median respondent in each group expects substantial advances in AI capabilities by 2030, small declines in labor force participation consistent with demographic shifts, and an annual GDP growth rate of 2.5%, which exceeds both the typical medium-run (2.0%) and long-run (1.7%) baseline forecasts from government agencies and private-sector forecasters. Conditional on a “rapid” AI progress scenario, in which AI systems surpass human performance on many cognitive and physical tasks, experts forecast substantial, though not historically unprecedented, economic shifts: annualized GDP growth rising to around 4% and the labor force participation rate falling from its current level of 62% to 55% by 2050, with roughly half of that decline -- equivalent to around 10 million lost jobs -- attributable to AI. A variance decomposition suggests that expert disagreement about these effects is driven primarily by different beliefs about the economic effects of highly capable AI systems rather than by disagreement about the pace of AI progress. These forecasts map onto notably different policy preferences across groups: experts strongly favor targeted measures such as worker retraining, whereas the general public supports both targeted programs and broader interventions, including a job guarantee and universal basic income.
The Effects of California's $20 Fast Food Minimum Wage on Prices
Jeffrey Clemens et al.
NBER Working Paper, March 2026
Abstract:
We analyze the effect of California's $20 fast food minimum wage (Assembly Bill 1228), enacted in September 2023 and implemented in April 2024, on consumer prices using the Bureau of Labor Statistics' Consumer Price Indices for food away from home across 21 metropolitan statistical areas. Food away from home prices in California's four in-sample MSAs increased by 3.3 to 3.6 percent relative to 17 control MSAs through December 2024. Our estimates are stable across a number of specifications. Placebo tests on price indices for goods and services that were not affected by the policy, including food at home, show no differential increases in California's MSAs. The price increases we estimate likely arise in part from spillovers to the full-service sector, as well as changes in the production functions and product quality choices of limited service restaurants.
The Impact of Non-Competes on Wages and Job Tenure: New Evidence from NLSY Data
Bart Hobijn, Tristan Potter & André Kurmann
Federal Reserve Working Paper, March 2026
Abstract:
Non-compete agreements (NCAs) are pervasive even in low-wage labor markets, yet most evidence relies on variation in enforceability rather than NCA incidence. Using longitudinal data from the NLSY97, we study how signing an NCA affects wage trajectories and job tenure. Exploiting complete work histories and applying a clean-controls local projections difference-in-difference design, we find a striking divergence: NCAs are associated with significantly slower wage growth for low-education workers over four years, but faster wage growth for high-education workers. Effects on job tenure are imprecisely estimated for both groups.
The Skill Premium in Times of Rapid Technological Change
Tarek Alexander Hassan, Aakash Kalyani & Pascual Restrepo
NBER Working Paper, March 2026
Abstract:
This paper shows that the pace of technology creation is a key driver of the skill premium. It develops a model in which skilled workers have a comparative advantage in learning new technologies. As technologies age, they become standardized and accessible to other workers. The skill premium is determined by the interplay between the pace of technology creation and standardization. A rapid pace of technology creation leads to a sustained increase in the skill premium. We calibrate the model using novel text-based data on new technologies and their changing demand for skills as they age. These data show that new technologies are initially skill intensive but become less so as they age. The data also point to an increased pace of new technology creation starting in the 1970s and tapering off in the 2000s. In response to this rapid pace of technology creation, the model generates a 32 percent increase in the college premium, which begins to reverse in the 2010s. Our framework also explains why the college premium is higher in dense cities, why its increase was mainly urban, and why it rose first for young workers and later for older workers.
The Macroeconomics of Asymmetric Automation: Frontier Technologies, Productivity, and the Labor Share
Carlos Chavez
University of Chicago Working Paper, March 2026
Abstract:
Automation technologies differ in their capacity to generate complementary human tasks. I build a two-frontier growth model in which physical and cognitive automation advance simultaneously but reinstate labor at different rates. The cognitive reinstatement rate is structurally estimated from U.S. commuting-zone wage data. Robot automation generates approximately 31 percent more aggregate productivity per unit of frontier advance than AI, driven almost entirely by differential task creation and reallocation rather than direct displacement. Three macro results follow. First, both frontiers lower the labor share, but robots are 42 percent as damaging as AI per unit of advance. Second, distributional consequences are supply-side dominant; aggregate demand channels are second-order. Third, the optimal R&D subsidy favors robots over AI at approximately 1.6-to-one, reflecting both efficiency and equity motives. A uniform automation tax is welfare-reducing.
Robots Reinstate, AI Doesn't: Asymmetric Task Creation Across Automation Frontiers
Carlos Chavez
University of Chicago Working Paper, February 2026
Abstract:
Robot-exposed occupations follow V-shaped wage trajectories -- decline then recovery -- while AI-exposed occupations follow L-shapes with no recovery. I develop a two-frontier task model in which the direction of automation and the rate of complementary task creation are jointly determined. Physical automation produces spatially bundled complements -- maintenance, monitoring, quality control -- because automated and human tasks are co-located; cognitive automation produces fewer complements because its outputs are digital and non-spatial. Calibrating the model to structurally estimated reinstatement rates identifies a complementarity ratio of 4.88: each unit of robot frontier advancement generates nearly five times as many feasible complements as AI advancement. The model also produces a directed automation trap, but the calibrated feedback is modest (7% amplification) and the economy admits a unique equilibrium. Four tests probe the predictions: (i) a 20-year ACS panel documents the V/L-shape divergence among non-college workers, predating COVID, non-college wages in robot-exposed occupations follow a V-shaped trajectory, dipping during 2010–2014 and recovering by 2018, with positive wage gains extending to all workers during the recovery phase (βˆR = 0.015, p = 0.006), (ii) O*NET data show robot-exposed occupations create 1.5–2 times more complementary tasks, with AI-complementary creation insignificant post-ChatGPT, (iii) robot-exposed firms expand non-AI employment while AI-adopting firms contract it, (iv) a minimum wage triple-difference finds no factor-price effect, suggesting the asymmetry is technological rather than wage-driven.
The Automation Paradox: Why Robots Reduce Unemployment
Yosef Bonaparte
University of Colorado Working Paper, February 2026
Abstract:
Household and macroeconomic concerns about automation often emphasize job displacement and rising unemployment. This paper examines the labor market effects of robot adoption using cross-country panel data and a Bartik-style identification strategy. We show that higher robot adoption, particularly in service-oriented automation, is associated with lower unemployment rates. To address endogeneity and adoption censoring, we employ a Tobit-based control-function approach combined with fixed effects. We further identify the underlying mechanism driving these results. Robot adoption significantly increases labor productivity, measured by output per worker, and higher productivity is in turn strongly associated with lower unemployment. The economic magnitudes are substantial: a one-standard-deviation increase in robot adoption raises productivity by 5-6 percent and reduces unemployment by nearly 10 percent relative to the mean. These findings suggest that productivity gains, rather than labor displacement, dominate the aggregate labor market effects of automation, offering new insights for policy debates on robot taxation and technological change.
Occupation-Specific Education Requirements and Occupational Silos: Evidence from CPA Licensing Rules
Anthony Le & Parth Shah
University of Chicago Working Paper, January 2026
Abstract:
We study the effect of licensing-induced, occupation-specific education requirements on workers' occupational mobility and earnings. We study this question in the context of Certified Public Accountants' (CPAs) licensing rules, exploiting the staggered introduction of a change in the number and composition of CPAs' educational requirements across states. We find that an increase in mandatory accounting-specific credit hours leads to more time spent in accounting jobs, less cross-occupation job switching, and a reduction in the licensing earnings premium. Supplemental analyses indicate that the effects represent a specialization of worker skills rather than a general decline in CPAs' accounting performance. The collective findings suggest that by imposing occupation-specific course requirements, licensing regimes can create less portable human capital, reducing both occupational mobility and the licensing earnings premium.
Wage growth and labor market tightness
Sebastian Heise, Jeremy Pearce & Jacob Weber
Journal of Monetary Economics, March 2026
Abstract:
Good measures of labor market tightness are essential to predict wage inflation and to calibrate monetary policy. This paper highlights the importance of two measures of labor market tightness in determining wage growth: the quits rate and vacancies per effective searcher (V/ES) -- where searchers include both employed and non-employed job seekers. Among a broad set of indicators of labor market tightness, we find that these two measures are independently the most strongly correlated with wage inflation both in aggregate time series data and in industry-level panel data, and also predict wage growth best out of sample. These results are consistent with the predictions of a New Keynesian DSGE model where firms have the power to set wages and workers search on the job. We develop a new composite indicator of labor market tightness that can be used by policymakers to predict wage pressures in real time.
Labor supply responses to tax credit disbursements: Evidence from the EITC schedule
Andrew Bibler
Economic Inquiry, forthcoming
Abstract:
The Earned Income Tax Credit (EITC) schedule and lump-sum disbursement can create significant labor supply responses. I estimate labor supply responses to tax credit disbursements using a regression kink design. Among single workers, credits increase labor supply around the time that tax credits are disbursed at the first and second kinks in the EITC schedule but reduce labor supply on the intensive margin at the third kink. There is some evidence of heterogeneous responses among married women, including an increase in labor supply near the third kink, although findings in the sample of married women appear less robust.