This site is undergoing review. Some annotations were human-generated, some AI-generated — all are being verified.
Back to papersAbstract Summary Main Finding Notes NBER WP 31222
[Claude classification]: NBER WP 31222. Revised January 2026. Uses GPT 3.5 Turbo for task classification following Eloundou et al. (2023) rubric. Novel contribution: distinguishes core vs. supplemental task exposure to differentiate substitution vs. complementarity effects. ChatGPT release date (Nov 30, 2022) treated as quasi-experimental shock. Event study shows 0.44-0.45% daily abnormal returns for high-exposure firms during two-week window. Effects robust to controlling for product-market AI exposure, data assets, firm characteristics, and industry fixed effects. Labor market results show 8% decline in job postings and 0.6% decline in wages for exposed occupations, concentrated in occupations with high core-task exposure. Firm-level profitability and analyst earnings forecasts increase for high-exposure firms. Authors argue results support labor-substitution channel over complementarity.
[Claude classification]: NBER WP 31222. Revised January 2026. Uses GPT 3.5 Turbo for task classification following Eloundou et al. (2023) rubric. Novel contribution: distinguishes core vs. supplemental task exposure to differentiate substitution vs. complementarity effects. ChatGPT release date (Nov 30, 2022) treated as quasi-experimental shock. Event study shows 0.44-0.45% daily abnormal returns for high-exposure firms during two-week window. Effects robust to controlling for product-market AI exposure, data assets, firm characteristics, and industry fixed effects. Labor market results show 8% decline in job postings and 0.6% decline in wages for exposed occupations, concentrated in occupations with high core-task exposure. Firm-level profitability and analyst earnings forecasts increase for high-exposure firms. Authors argue results support labor-substitution channel over complementarity.
[Claude classification]: NBER WP 31222. Revised January 2026. Uses GPT 3.5 Turbo for task classification following Eloundou et al. (2023) rubric. Novel contribution: distinguishes core vs. supplemental task exposure to differentiate substitution vs. complementarity effects. ChatGPT release date (Nov 30, 2022) treated as quasi-experimental shock. Event study shows 0.44-0.45% daily abnormal returns for high-exposure firms during two-week window. Effects robust to controlling for product-market AI exposure, data assets, firm characteristics, and industry fixed effects. Labor market results show 8% decline in job postings and 0.6% decline in wages for exposed occupations, concentrated in occupations with high core-task exposure. Firm-level profitability and analyst earnings forecasts increase for high-exposure firms. Authors argue results support labor-substitution channel over complementarity.
[Claude classification]: NBER WP 31222. Revised January 2026. Uses GPT 3.5 Turbo for task classification following Eloundou et al. (2023) rubric. Novel contribution: distinguishes core vs. supplemental task exposure to differentiate substitution vs. complementarity effects. ChatGPT release date (Nov 30, 2022) treated as quasi-experimental shock. Event study shows 0.44-0.45% daily abnormal returns for high-exposure firms during two-week window. Effects robust to controlling for product-market AI exposure, data assets, firm characteristics, and industry fixed effects. Labor market results show 8% decline in job postings and 0.6% decline in wages for exposed occupations, concentrated in occupations with high core-task exposure. Firm-level profitability and analyst earnings forecasts increase for high-exposure firms. Authors argue results support labor-substitution channel over complementarity.
[Claude classification]: NBER WP 31222. Revised January 2026. Uses GPT 3.5 Turbo for task classification following Eloundou et al. (2023) rubric. Novel contribution: distinguishes core vs. supplemental task exposure to differentiate substitution vs. complementarity effects. ChatGPT release date (Nov 30, 2022) treated as quasi-experimental shock. Event study shows 0.44-0.45% daily abnormal returns for high-exposure firms during two-week window. Effects robust to controlling for product-market AI exposure, data assets, firm characteristics, and industry fixed effects. Labor market results show 8% decline in job postings and 0.6% decline in wages for exposed occupations, concentrated in occupations with high core-task exposure. Firm-level profitability and analyst earnings forecasts increase for high-exposure firms. Authors argue results support labor-substitution channel over complementarity.
[Claude classification]: NBER WP 31222. Revised January 2026. Uses GPT 3.5 Turbo for task classification following Eloundou et al. (2023) rubric. Novel contribution: distinguishes core vs. supplemental task exposure to differentiate substitution vs. complementarity effects. ChatGPT release date (Nov 30, 2022) treated as quasi-experimental shock. Event study shows 0.44-0.45% daily abnormal returns for high-exposure firms during two-week window. Effects robust to controlling for product-market AI exposure, data assets, firm characteristics, and industry fixed effects. Labor market results show 8% decline in job postings and 0.6% decline in wages for exposed occupations, concentrated in occupations with high core-task exposure. Firm-level profitability and analyst earnings forecasts increase for high-exposure firms. Authors argue results support labor-substitution channel over complementarity.
[Claude classification]: NBER WP 31222. Revised January 2026. Uses GPT 3.5 Turbo for task classification following Eloundou et al. (2023) rubric. Novel contribution: distinguishes core vs. supplemental task exposure to differentiate substitution vs. complementarity effects. ChatGPT release date (Nov 30, 2022) treated as quasi-experimental shock. Event study shows 0.44-0.45% daily abnormal returns for high-exposure firms during two-week window. Effects robust to controlling for product-market AI exposure, data assets, firm characteristics, and industry fixed effects. Labor market results show 8% decline in job postings and 0.6% decline in wages for exposed occupations, concentrated in occupations with high core-task exposure. Firm-level profitability and analyst earnings forecasts increase for high-exposure firms. Authors argue results support labor-substitution channel over complementarity.
[Claude classification]: NBER WP 31222. Revised January 2026. Uses GPT 3.5 Turbo for task classification following Eloundou et al. (2023) rubric. Novel contribution: distinguishes core vs. supplemental task exposure to differentiate substitution vs. complementarity effects. ChatGPT release date (Nov 30, 2022) treated as quasi-experimental shock. Event study shows 0.44-0.45% daily abnormal returns for high-exposure firms during two-week window. Effects robust to controlling for product-market AI exposure, data assets, firm characteristics, and industry fixed effects. Labor market results show 8% decline in job postings and 0.6% decline in wages for exposed occupations, concentrated in occupations with high core-task exposure. Firm-level profitability and analyst earnings forecasts increase for high-exposure firms. Authors argue results support labor-substitution channel over complementarity.
[Claude classification]: NBER WP 31222. Revised January 2026. Uses GPT 3.5 Turbo for task classification following Eloundou et al. (2023) rubric. Novel contribution: distinguishes core vs. supplemental task exposure to differentiate substitution vs. complementarity effects. ChatGPT release date (Nov 30, 2022) treated as quasi-experimental shock. Event study shows 0.44-0.45% daily abnormal returns for high-exposure firms during two-week window. Effects robust to controlling for product-market AI exposure, data assets, firm characteristics, and industry fixed effects. Labor market results show 8% decline in job postings and 0.6% decline in wages for exposed occupations, concentrated in occupations with high core-task exposure. Firm-level profitability and analyst earnings forecasts increase for high-exposure firms. Authors argue results support labor-substitution channel over complementarity.
[Claude classification]: NBER WP 31222. Revised January 2026. Uses GPT 3.5 Turbo for task classification following Eloundou et al. (2023) rubric. Novel contribution: distinguishes core vs. supplemental task exposure to differentiate substitution vs. complementarity effects. ChatGPT release date (Nov 30, 2022) treated as quasi-experimental shock. Event study shows 0.44-0.45% daily abnormal returns for high-exposure firms during two-week window. Effects robust to controlling for product-market AI exposure, data assets, firm characteristics, and industry fixed effects. Labor market results show 8% decline in job postings and 0.6% decline in wages for exposed occupations, concentrated in occupations with high core-task exposure. Firm-level profitability and analyst earnings forecasts increase for high-exposure firms. Authors argue results support labor-substitution channel over complementarity.
[Claude classification]: NBER WP 31222. Revised January 2026. Uses GPT 3.5 Turbo for task classification following Eloundou et al. (2023) rubric. Novel contribution: distinguishes core vs. supplemental task exposure to differentiate substitution vs. complementarity effects. ChatGPT release date (Nov 30, 2022) treated as quasi-experimental shock. Event study shows 0.44-0.45% daily abnormal returns for high-exposure firms during two-week window. Effects robust to controlling for product-market AI exposure, data assets, firm characteristics, and industry fixed effects. Labor market results show 8% decline in job postings and 0.6% decline in wages for exposed occupations, concentrated in occupations with high core-task exposure. Firm-level profitability and analyst earnings forecasts increase for high-exposure firms. Authors argue results support labor-substitution channel over complementarity.
Generative AI and Firm Values
Eisfeldt, Schubert, Zhang
2023NBER Working Paper109 citations
Exposure / measurementCausal
LLM / Generative AIFinanceAugmentation vs. substitutionHuman-AI collaboration
What are the effects of recent advances in Generative AI on the value of firms?Our study offers a quantitative answer to this question for U.S. publicly traded companies based on the exposures of their workforce to Generative AI.Our novel firm-level measure of workforce exposure to Generative AI is validated by data from earnings calls, and has intuitive relationships with firm and industry-level characteristics.Using Artificial Minus Human portfolios that are long firms with higher exposures and short firms with lower exposures, we show that higher-exposure firms earned excess returns that are 0.4% higher on a daily basis than returns of firms with lower exposures following the release of ChatGPT.Although this release was generally received by investors as good news for more exposed firms, there is wide variation across and within industries, consistent with the substantive disruptive potential of Generative AI technologies.
Eisfeldt, Schubert, and Zhang construct firm-level measures of workforce exposure to Generative AI by aggregating task-level LLM classifications to occupations and firms, and use an event study around ChatGPT's release to study effects on firm valuation, labor demand, wages, and profitability.
Firms with high workforce exposure to Generative AI experienced 5% higher cumulative abnormal returns in the two weeks following ChatGPT's release, driven primarily by labor substitution of core tasks, with firms reducing job postings and wages for exposed occupations while experiencing increased profitability.
Primary Datasets
Revelio Labs; Lightcast
- Key Methods
- Event study analysis of stock returns around ChatGPT release; difference-in-differences estimation of labor market outcomes; firm-level exposure measurement via task-based LLM classification aggregated through occupational employment shares.
- Sample Period
- 2022-2023
- Geographic Coverage
- US
- Sample Size
- 2,518 publicly traded firms; 678 SOC 6-digit occupations; 19,265 tasks; ~271,000 individual-month wage observations; 36,900 firm-month job posting observations
- Level of Analysis
- Individual, Firm, Occupation, Industry, Task
- Occupation Classification
- O*NET-SOC
- Industry Classification
- NAICS 3-digit; Hoberg-Phillips FIC50