NBER researchers measured AI’s macroeconomic impact through software engineering productivity. They used stock market data to build a real-time, forward-looking measure. Firms with higher software engineering payroll showed greater stock return sensitivity to an AI index. From November 2022 to December 2025, AI raised expected software engineering productivity by a permanent 32.6%. This translated to a 3.6% GDP increase in the baseline scenario. With R&D productivity gains, the GDP effect reached 6.5%. By mid-2026, amid rapid coding agent progress, the productivity and GDP effects had more than doubled. The study maps stock return sensitivity to productivity gains using a model. The figures come from an NBER working paper.

• 32.6% permanent software engineering productivity increase
• 3.6% GDP effect (baseline)
• 6.5% GDP effect (with R&D productivity gains)
• AI effect on productivity and GDP more than doubled by mid-2026