Financial markets are betting that artificial intelligence will permanently increase software engineering productivity by 32.6 percent, according to a new analysis published by economists at the University of California, Berkeley and the London School of Economics. The researchers examined stock market movements between November 2022 and December 2025 to gauge investor expectations about AI's impact on developer efficiency. Rather than measuring actual output, the team converted market behavior into an estimate of anticipated productivity gains.
The study tracked how company stock returns responded to news about AI, and whether that response varied based on the share of each firm's payroll dedicated to software engineers. "We empirically measure whether firms with larger software engineering payroll shares experience larger stock-price increases when the AI stock index rises," Chen Lian, assistant professor of finance at UC Berkeley, explained. The research found that between ChatGPT's launch in November 2022 and December 2025, AI developments shifted the market's expected present value of software engineering productivity by an amount equal to a permanent 32.6 percent increase. When fed through the team's economic model, this productivity estimate corresponds to a GDP boost equivalent to a permanent 3.61 percent level increase.
The authors note that their market-derived productivity figure of 32.6 percent aligns with the 21 to 56 percent acceleration on individual tasks reported in other studies. However, the researchers caution that task-level improvements can be constrained by bottlenecks that cap overall productivity gains—in software development, code reviews that can't match the pace of increased commits might serve as one such limitation. "Our estimates capture the market's assessment of current and future productivity gains, and markets can be overly optimistic or pessimistic," Lian acknowledged. The advantage of this approach, he said, is a forward-looking measure available in real time when many of AI's effects haven't yet materialized.
The methodology translates investor behavior into quantified expectations by examining the relationship between AI-related information reflected in stock prices and company exposure to software engineering costs. The researchers combined empirical measurements with an economic model to infer how AI has altered investor expectations about developer productivity and the resulting GDP consequences. By "news about AI," the team means new information captured in stock prices—not specific announcements or product launches. A preliminary review of their data suggests total software engineering employment among covered firms actually grew over the past few years, though the productivity estimate doesn't depend on employment trends.
Lian indicated the research methods can be applied to study AI's economic impact through other channels, which the team plans to explore in subsequent work. The paper, titled "The Macroeconomic Effect of AI: Sizing the Software Engineering Channel," was published through the National Bureau of Economics Research. While the analysis captures what investors currently believe about AI's transformative potential, the authors stress that market expectations don't always match reality—financial markets can swing between excessive optimism and unwarranted pessimism. The research offers a real-time snapshot of collective market judgment rather than a guarantee of actual outcomes. For companies weighing AI investments and policymakers planning for economic shifts, the distinction between anticipated transformation and delivered results will determine whether today's market enthusiasm proves prescient or premature.