Large Language Models Slow Coder Employment Growth
Leland D. Crane and Paul E. Soto (March 2026) analyze whether large language models have affected coder employment using O*NET linked to CPS data. They find aggregate coder employment growth sharply decelerated after ChatGPT's introduction and show, using a novel industry-level control variable, that the slowdown reflects an occupation-specific shock rather than industry composition effects. Coder employment continues to grow but much more slowly than pre-2022.
Key Points
- 1Find sharp deceleration in coder employment growth after ChatGPT's late‑2022 introduction.
- 2Use industry-level control to rule out industry composition as the primary cause, indicating occupation-specific shock.
- 3Suggest practitioners and policymakers reassess coder labor demand, training, and automation mitigation strategies.
Scoring Rationale
Strong empirical evidence from an official Fed research paper, but limited to coder occupations and observational inference.
Sources
Public references used for this report.
Practice with real Logistics & Shipping data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Logistics & Shipping problems
