The Macroeconomic Effect of AI: Measuring the Software Engineering Channel

Summary

This NBER Working Paper by Alex Blumenfeld, Jonathon Hazell, Chen Lian, and Andreas Schaab estimates how AI affects the economy through software engineering productivity. The authors use financial-market information to create a real-time, forward-looking measure, estimating how firms’ stock returns respond to an AI stock-market index and how that relationship varies with the share of payroll devoted to software engineering. They map this cross-firm relationship into productivity gains. From November 2022 through December 2025, the market’s expected present value of software engineering productivity rose by the equivalent of a permanent 32.6% increase. In the baseline estimate, this translates into a 3.6% increase in the level of GDP, or 6.5% when higher software productivity also raises R&D productivity. The paper reports that by mid-2026, amid rapid progress in coding agents, the estimated productivity and GDP effects had more than doubled relative to the end of 2025. The findings are estimates based on market expectations and a model, and the authors note that the views are theirs rather than necessarily those of the NBER.