Skip to content

When will the research paper disappear in economics?

Key Ideas

  • AI will soon allow anyone to take a published paper and tweak or improve it at will, making the static research paper an obsolete format.
  • Tools like Refine already enable AI-based evaluation of paper quality; the entire history of economics publications can be reassessed with token investment.
  • The future may involve publishing datasets and code rather than papers -- users could query them to generate the paper they want.
  • Economics may become a branch of software engineering, where advances are embodied in systems and "boxes" rather than static documents.
  • Tenure and career evaluation may need to shift toward rewarding people who build capabilities and evaluative systems rather than people who publish many papers.
  • Cowen provocatively asks whether Anthropic and OpenAI deserve an economics Nobel Prize.

Summary

Tyler Cowen offers a series of provocative questions about the future of the research paper in economics. His central argument is that AI makes the static, published research paper obsolete as a format. When anyone can use AI to tweak, improve, or re-evaluate any paper, the paper itself becomes a living, mutable object rather than a fixed contribution. Tools like Refine can already evaluate paper quality at scale, raising the question of whether journals are needed to certify which papers are "best."

Cowen pushes further: what if researchers published datasets and code instead of papers, and readers issued commands to generate whatever analysis they wanted? What if the important contributions in economics became evaluative systems and simulation frameworks ("boxes") rather than individual papers? He draws an analogy to software engineering, where the most important advances are often embodied in software rather than papers. He suggests that tenure should reward building capabilities and systems, and asks whether highly productive researchers who publish many papers will become less valuable as the individual paper becomes less scarce.

Relevance to Economics Research

This is a thought-provoking piece from one of the most influential economics bloggers, raising questions that every economist should consider. The idea that the research paper itself may become obsolete -- replaced by interactive datasets, code repositories, and AI-powered evaluation systems -- challenges fundamental assumptions about how economic knowledge is produced, validated, and rewarded. The implications for career incentives, tenure evaluation, and the role of journals are immediate and practical.