Back to Blog

The Fed Just Put a Number on AI's Impact on Programmer Jobs: 500,000 Missing Roles

AIResearch

The question of whether AI is actually displacing coders has been fought mostly with vibes and anecdotes — until now. A new Federal Reserve working paper is the first institutional-level analysis to directly tie ChatGPT’s rise to a measurable drop in U.S. programmer employment, and the numbers are hard to dismiss.

What the Fed Found

Fed economists Leland D. Crane and Paul E. Soto published “AI and Coder Employment: Compiling the Evidence,” finding that growth in programming-intensive occupations fell roughly 50% after ChatGPT launched in November 2022. Before that release, those jobs were growing at about 5% annually — well above the broader labor market. Stretched over three years, the gap between actual and projected employment adds up to an estimated 500,000 positions that likely would have existed without the rise of large language models.

The paper is notable not just for its findings, but for what it controlled away. The post-2022 tech slowdown, rising interest rates, and the unwinding of pandemic-era hiring surges could all explain some of the decline. The authors tested for all of these, and programming employment still fell roughly 3% per year below the expected baseline. The culprit, their analysis suggests, is AI adoption itself.

The 18-Month Lag No One Expected

One of the more revealing details is the timing. The employment gap didn’t open until roughly mid-2024 — about 18 months after ChatGPT launched. That lag suggests companies didn’t immediately cut programmers; instead, they stopped hiring new ones. The effect shows up in hiring velocity, not in layoffs or wage compression.

Programmer wages, notably, haven’t declined. That distinction matters: this isn’t companies slashing headcount, it’s companies deciding one experienced developer with AI tools can cover work that previously required two hires.

Junior Developers Are Bearing the Brunt

The generational dimension is stark. Early-career workers in high-AI-exposure occupations — software developers among them — have seen employment decline relative to both experienced developers in the same field and early-career workers in less exposed industries. If you’re a junior dev entering the market right now, the economic pressure Gen Z is already feeling around AI displacement now has institutional confirmation from the Fed.

This isn’t surprising in retrospect. Senior developers have the judgment, system knowledge, and architectural experience that AI tools still struggle to replicate. Junior roles — ticket-driven feature work, boilerplate, test writing, basic API integration — are exactly the tasks where LLMs are most immediately useful.

What This Means Going Forward

The paper carries a preliminary designation and hasn’t completed the Fed’s full review process. The authors describe it as “only a first step” and caution that the 500,000 figure isn’t a literal count of eliminated jobs — some workers likely shifted into adjacent roles. But the directional finding is clear, and it arrives at a moment when the industry is still debating whether to take AI displacement seriously as a policy question.

History offers little comfort to those hoping this reverses. Technological transitions rarely stop once the productivity math tips decisively — they accelerate. What the Fed study does, more than anything, is move this conversation from speculation to evidence. For developers, the implication is straightforward: the tooling you learn to use well will matter more than the volume of code you can produce.

Further Reading

AI Disclosure

This document is drafted by an AI skill and is provided for informational and governance support purposes only. It does not constitute legal advice or a formal compliance determination. Do not publish or rely on this notice as a substitute for review by qualified legal counsel or a licensed compliance professional with jurisdiction-specific expertise.