Graduating Into Disruption: Labor Market Outcomes for AI-Exposed College Majors
September 2026
Working Paper Number:
CES-26-56
Abstract
How are new college graduates affected by the rise of artificial intelligence, and what can this tell us about the mechanisms behind AI's overall labor market effects? We use administrative records on college graduates to observe how economic outcomes among college majors with differing levels of labor market AI exposure have evolved since large language models became available. We find that post-graduation employment, earnings, and job switching patterns among the most AI-exposed college majors began to diverge immediately following the introduction of ChatGPT in late 2022. In regression-adjusted estimates, the most AI-exposed decile of college majors saw their likelihood of initial employment decline by five percentage points, while full-quarter initial earnings declined by thirteen percent. This earnings decline is comparable in magnitude to the earnings losses associated with graduating into a large recession. Roughly half of the decline in earnings is attributable to lower earnings within the industry sectors that employ these graduates, with the remainder resulting from a shift in the industry mix into lower-wage sectors such as restaurants and retail. The effects attenuate as graduates move further from labor market entry but remain substantial for the most exposed majors.
Document Tags and Keywords
Keywords
Keywords are automatically generated using KeyBERT, a powerful and innovative
keyword extraction tool that utilizes BERT embeddings to ensure high-quality and contextually relevant
keywords.
By analyzing the content of working papers, KeyBERT identifies terms and phrases that capture the essence of the
text, highlighting the most significant topics and trends. This approach not only enhances searchability but
provides connections that go beyond potentially domain-specific author-defined keywords.
:
employment effects,
employ,
job,
hiring,
occupation,
educated,
college,
graduate,
clerical,
enrollee,
recession employment,
career,
employment recession,
degree,
academic,
ai
Tags
Tags are automatically generated using a pretrained language model from spaCy, which excels at
several tasks, including entity tagging.
The model is able to label words and phrases by part-of-speech,
including "organizations." By filtering for frequent words and phrases labeled as "organizations", papers are
identified to contain references to specific institutions, datasets, and other organizations.
:
Bureau of Labor Statistics,
Employer Identification Numbers,
Department of Education,
Department of Labor,
North American Industry Classification System,
American Community Survey,
Alfred P Sloan Foundation,
Longitudinal Employer Household Dynamics,
Technical Services,
Quarterly Census of Employment and Wages,
Census Bureau Disclosure Review Board,
Accommodation and Food Services,
COVID-19,
COVID,
J23
Similar Working Papers
Similarity between working papers are determined by an unsupervised neural
network model
know as Doc2Vec.
Doc2Vec is a model that represents entire documents as fixed-length vectors, allowing for the
capture of semantic meaning in a way that relates to the context of words within the document. The model learns to
associate a unique vector with each document while simultaneously learning word vectors, enabling tasks such as
document classification, clustering, and similarity detection by preserving the order and structure of words. The
document vectors are compared using cosine similarity/distance to determine the most similar working papers.
Papers identified with 🔥 are in the top 20% of similarity.
The 10 most similar working papers to the working paper 'Graduating Into Disruption: Labor Market Outcomes for AI-Exposed College Majors' are listed below in order of similarity.
-
Working PaperYou're (not) Hired: Artificial Intelligence and Early Career Hiring in the Quarterly Workforce Indicators🔥
April 2026
Working Paper Number:
CES-26-27
Using detailed tabulations from matched employer-employee administrative data, I document evidence of an immediate, sizable, and persistent decrease in the level of early career (22-24 year old) hires following introduction of ChatGPT within the industry-state cells that are most exposed to AI. The decline in hires is the primary cause of large observed declines in employment over the subsequent period. Regressionadjusted employment of early career workers in the most AI-exposed quintile of industry-state cells declined by 12% over the 10 quarters following the introduction of ChatGPT, even as employment in lessexposed industries has remained stable. The rate of hiring largely recovered by early 2025, attributable to a smaller employment base. Earnings growth of early career workers in the most exposed industries slowed slightly relative to those in less exposed industries. Although the most AI-exposed quintile of detailed industries is dominated by a handful of industry sectors, I find that the association of higher AI exposure with reduced early career employment and fewer hires is observed across most sectors of the economy. Timing of effects in event studies is consistent with an immediate effect on hiring following introduction of ChatGPT. However, triple difference estimates provide some evidence of earlier trend shifts on employment, hiring, and separations around the onset of the COVID pandemic. I discuss potential explanations, including the increase in remote work and increased educational attainment among workers in AI-exposed occupations. Nonetheless, job gains to early career workers and backfill hires show evidence of discontinuous decline at the time of ChatGPT's release in comparison to older workers in the same industries. A local projections analysis at the NAICS industry group level shows that industries with high AI exposure are not particularly sensitive to unexpected fluctuations in monetary policy on average relative to other industries in employment, hiring, or separations. A historical decomposition suggests that up to one quarter of relative early career employment declines through 2025q2 may be attributable to monetary policy shocks through 2023, but the analysis does not find evidence that these shocks can explain the rapid decline in hires at the most AI-exposed firms in comparison to others.View Full Paper PDF
-
Working PaperAI Exposure and Adoption Among U.S. Firms🔥
September 2026
Working Paper Number:
CES-26-61
Measures of exposure to artificial intelligence (AI) are widely used to study where AI is likely to affect firm and worker outcomes, yet limited evidence exists on how closely exposure relates to realized AI adoption at the firm level. We examine this relationship by linking 14 firm-level exposure measures, constructed from occupational exposure estimates in the literature and occupational employment shares from the Bureau of Labor Statistics Occupational Employment and Wage Statistics program, to direct measures of firm AI adoption from the Census Bureau's Business Trends and Outlook Survey. Exposure is positively and statistically significantly associated with adoption, but explains only a modest share of its variation. The strength of the relationship varies considerably across standardized exposure measures: a one-standard-deviation increase in firm-level exposure is associated with a 4'11 percentage point higher firm adoption probability, falling to 1'8 percentage points after controlling for year and sub-sector fixed effects. The exposure'adoption relationship is also heterogeneous across firm-size classes and sectors. Exposure is thus an informative but incomplete signal of adoption. More broadly, the results have implications for the interpretation of exposure-based measures in studies of the economic effects of AI.View Full Paper PDF
-
Working PaperCollege Majors and Earnings Growth🔥
February 2026
Working Paper Number:
CES-26-14
We estimate major-specific earnings profiles using matched American Community Survey (ACS) and Longitudinal Employer-Household Dynamics (LEHD) data. Building on Deming and Noray (2020), we exploit a long earnings panel to overcome key limitations of cross-sectional approaches to lifecycle estimation. We find that engineering and computer science majors experience earnings growth that is comparable to or faster than that of other majors, a category including humanities, education, psychology, and similar fields. In contrast, Deming and Noray (2020) use a crosscohort approach and find that earnings for engineering and computer science majors decline relative to other fields over the lifecycle.View Full Paper PDF
-
Working PaperGender Differences in the Returns to Double Majoring: A Case of Signaling?
September 2026
Working Paper Number:
CES-26-60
Over 15 percent of college graduates in the United States graduate with more than one major. While much research exists on the returns to different individual majors, less is known about the causal effects of double majoring. This study provides novel estimates on the returns to those who choose to double major as an undergraduate. We improve on previous studies that rely on controls for observable characteristics by addressing selection concerns in two notable ways. First, we control for institution-specific differences that may influence both the decision to double major and subsequent earnings using proprietary data by including institution fixed effects. Second, we adopt a partial identification approach to address non-random selection into double majoring within institutions. Results broadly align with estimates from prior studies at the aggregate level, but reveal notable gender differences not previously detected. Women experience an earnings return to double majoring of approximately 5 percent, while the return for men is statistically negligible. Our analysis suggests that this discrepancy is consistent with double majors serving as a signaling effect in the labor market, which may help offset the gender pay gap.View Full Paper PDF
-
Working PaperThe Microstructure of AI Diffusion: Evidence From Firms, Business Functions, and Worker Tasks
April 2026
Working Paper Number:
CES-26-25
Using novel, nationally representative data from the 2026 AI supplement to the U.S. Census Bureau's Business Trends and Outlook Survey (BTOS), we characterize AI diffusion across three interconnected layers: overall firm use, deployment across business functions, and worker-task use. This multi-layered approach provides a nuanced picture of business AI adoption. During the supplement reference period (Nov 2025-Jan 2026), 18% of firms used AI in a business function, rising to 32% on an employment-weighted basis; adoption is expected to reach 22% within six months. AI use is substantially higher in large firms and knowledge-intensive sectors, with use rates reaching 50%-60% (60%-70%, employment-weighted) for very large firms in the Information, Professional Services, and Finance sectors. Among adopting firms, the scope of use remains limited: 57% of users integrate AI in three or fewer business functions, most commonly Sales and Marketing (52%), Strategy and Business Development (45%), and IT (41%). In 23% (41%, employment-weighted) of firms, workers use AI in work-related tasks. Writing, document analysis, and information search are the leading Generative AI use in tasks, though 65% of firms limit use to three or fewer tasks. The evidence points to both top-down and bottom-up diffusion channels: worker task use sometimes occurs without formal firm-level adoption, and firm-level adoption sometimes occurs without worker task use. Most users (66%) rely on AI solely to augment tasks, while AI-related employment decreases are rare, occurring in only 2% of firms. Regression analysis shows a robust positive correlation between firm commercial performance and the breadth of AI integration, including functional deployment, task-level use, and operational investment. A distinct divergence emerges, however, with respect to labor outcomes. Functional breadth and operational investment are positively associated with employment decreases, whereas worker-task integration shows no significant link to headcount reduction once functional integration and operational investment are taken into account.View Full Paper PDF
-
Working PaperWho Scars the Easiest? College Quality and the Effects of Graduating into a Recession
September 2024
Working Paper Number:
CES-24-47
Graduating from college into a recession is associated with earnings losses, but less is known about how these effects vary across colleges. Using restricted-use data from the National Survey of College Graduates, we study how the effects of graduating into worse economic conditions vary over college quality in the context of the Great Recession. We find that earnings losses are concentrated among graduates from relatively high-quality colleges. Key mechanisms include substitution out of the labor force and into graduate school, decreased graduate degree completion, and differences in the economic stability of fields of study between graduates of high- and low-quality colleges.View Full Paper PDF
-
Working PaperScreening Out the Needy: the Effects of SNAP Work Requirements
July 2026
Working Paper Number:
CES-26-46
We examine the effectiveness of work requirements as a screening device in the Supplemental Nutrition Assistance Program (SNAP). Work requirements for 'able-bodied adults without dependents' were suspended after the Great Recession and gradually reinstated across counties and states in the 2010s. Using linked administrative SNAP and employment data from five states and a triple-differences design, we find that work requirements reduce SNAP participation by seven percent without increasing labor supply and disproportionately screen out low-income individuals. We develop a welfare framework to interpret these results and find that the social costs of work requirements exceed budget savings.View Full Paper PDF
-
Working PaperEstimating the Graduate Coverage of Post-Secondary Employment Outcomes
September 2025
Working Paper Number:
CES-25-61
This paper proposes a new methodology for estimating the coverage rate of the Post-Secondary Employment Outcomes data product (PSEO), both as a share of new graduates and as a share of total working-age degree holders in the United States. This paper also assesses how representative PSEO is of the broader population of college graduates across an array of institutional and individual characteristics.View Full Paper PDF
-
Working PaperLabor Market Concentration, Earnings Inequality, and Earnings Mobility
September 2018
Working Paper Number:
carra-2018-10
Using data from the Longitudinal Business Database and Form W-2, I document trends in local industrial concentration from 1976 through 2015 and estimate the effects of that concentration on earnings outcomes within and across demographic groups. Local industrial concentration has generally been declining throughout its distribution over that period, unlike national industrial concentration, which declined sharply in the early 1980s before increasing steadily to nearly its original level beginning around 1990. Estimates indicate that increased local concentration reduces earnings and increases inequality, but observed changes in concentration have been in the opposite direction, and the magnitude of these effects has been modest relative to broader trends; back-of-the-envelope calculations suggest that the 90/10 earnings ratio was about six percent lower and earnings were about one percent higher in 2015 than they would have been if local concentration were at its 1976 level. Within demographic subgroups, most experience mean earnings reductions and all experience increases in inequality. Estimates of the effects of concentration on earnings mobility are sensitive to specification.View Full Paper PDF
-
Working PaperUnemployment Insurance Extensions, Labor Market Concentration, and Match Quality
April 2026
Working Paper Number:
CES-26-24
I investigate whether the effects of UI extensions are different for workers exposed to higher levels of local labor market concentration, a potential source of employer market power. I exploit measurement error in state unemployment rates that led to quasi-random assignment of UI durations in the U.S. during the Great Recession. Using matched employer-employee data from the Longitudinal Employer-Household Dynamics program, I find that UI extensions lengthen nonemployment durations by one week and cause economically meaningful but not statistically significant increases in earnings. The UI-earnings effect is significantly lower at higher levels of concentration, while there is no difference in the UI-duration effect. The lower UI-earnings effect is driven by the extremes of the distribution of concentration. My results suggest that match improvements from UI are attenuated at higher levels of concentration.View Full Paper PDF
