Paper argues AI layoffs create a demand trap, with automation tax as the only fix

A theoretical economics paper posted on arXiv by University of Pennsylvania researcher Brett Hemenway Falk and Boston University professor Gerry Tsoukalas argues that AI-driven layoffs can push the broader economy into a destructive feedback loop. The paper, titled The AI Layoff Trap and published on March 21, 2026, says firms capture the full savings from replacing workers with AI, while the lost spending power of displaced workers is spread across the entire market. That setup, the authors argue, gives each company a strong incentive to automate even when collective over-automation hurts both labor and firms. The paper models a market with N identical companies and shows that each firm bears only 1/N of the demand loss caused by its own layoffs. It expresses the degree of over-automation as ℓ(1-1/N)/k, where ℓ is the demand loss caused by each displaced worker and k is the friction cost of adopting AI. According to the authors, the gap widens as competition increases and as AI becomes cheaper and easier to deploy. The paper also reviews policy tools often discussed in the AI labor debate, including UBI, retraining, employee profit-sharing, capital gains taxes, and negotiation, and concludes that none of them eliminate the problem at the margin. In the model, only a Pigouvian tax on automation changes firms’ incentives directly. The authors add that the paper is a theoretical exercise rather than an empirical measurement of the current labor market.