A new report from the Federal Reserve Bank of New York challenges the popular narrative that AI is driving job losses among young graduates.
Instead, the study identifies the widespread shift to remote work as the primary culprit, citing a breakdown in on-the-job training and mentorship for junior employees.
Key Findings
The study analyzed federal employment data and internal company records, documenting a significant shift in labor market outcomes for young workers.
- Unemployment Rate Increase: The unemployment rate for college graduates under 29 rose by 20% when comparing the post-pandemic period (2022-2024) with the pre-pandemic period (2017-2019). In March 2026, the rate for young college graduates stood at 5.6%, up from 3.6% in March 2019. In contrast, older graduates saw a slight decrease in unemployment over the same period.
- Remote Work as a Primary Factor: The New York Fed economists estimate that the expansion of remote work accounts for 64% of the increase in youth unemployment. The report states that remote work has "weakened incentives to hire young workers by impeding on-the-job training."
- Limited Role of AI: The study asserts that the rise in youth unemployment predates the rapid diffusion of AI, and that the unemployment gap between younger and older workers persists even when controlling for exposure to AI.
Methodology and Supporting Data
Researchers used a combination of public data and proprietary information to reach their conclusions.
- Comparative Analysis: The study compared unemployment rates between inexperienced and experienced employees across "remotable" jobs (e.g., software engineers, financial analysts) and "non-remotable" jobs (e.g., nurses, funeral home managers). The unemployment gap between recent graduates and older graduates was found to be larger in occupations that can be performed remotely.
- Internal Company Data: Researchers analyzed proprietary data from an undisclosed Fortune 500 tech company. The data showed that software engineers received approximately 20% more feedback when working in proximity to colleagues. The company reduced its hiring of new graduates as remote work increased, but reversed this policy after implementing a return-to-office (RTO) mandate.
- Other Research: The New York Fed analysis is supported by a working paper by Fed economist Natalia Emanuel and professors Emma Harrington and Amanda Pallais. This paper found that while experienced software developers showed little change in code quality when working remotely, junior workers produced lower-quality output due to reduced feedback. A similar study of customer assistance employees found that remote work led to longer call resolution times. Researchers at the London School of Economics have reported similar findings in a study covering the U.S., U.K., Canada, and Australia.
Broader Context
The study examines a labor market where remote work is now a permanent fixture, but its effects are unevenly distributed across experience levels.
- Worker Preferences: A Gallup survey from May 2025 found that only 6% of Gen Z workers prefer fully on-site work, while 71% prefer a hybrid arrangement.
- Productivity: Research from the U.S. Department of Labor in 2024 indicated that industries with higher rates of remote work showed larger productivity increases.
- Impact on Hiring: The report suggests that companies may be hesitant to hire inexperienced workers for distributed teams, as remote work makes it more difficult to provide training and mentorship. Return-to-office mandates are often justified by employers based on the perceived value of in-person mentorship and collaboration.
Implications
The researchers note that high unemployment early in a worker's career can have lasting negative effects on lifetime earnings and career progression. The study concludes that the expansion of remote work, not artificial intelligence, has been the most significant factor driving this trend in the current labor market.