Causal Claims in Economics
We taught an AI pipeline to read 45,000 economics papers and map exactly what each one claims causes what. The share of causal claims has roughly quadrupled since 1990 — about a third of claims in recent papers are causal. Papers making novel, well-identified causal claims are more likely to land in top journals and gather citations.
Read the full abstract
As economics scales, a key bottleneck is representing what papers claim in a comparable, aggregable form. We introduce evidence-annotated claim graphs that map each paper into a directed network of standardized economic concepts (nodes) and stated relationships (edges), with each edge labeled by evidentiary basis, including whether it is supported by causal inference designs or by non-causal evidence. Using a structured multi-stage AI workflow, we construct claim graphs for 44,852 economics papers from 1980–2023. The share of causal edges rises from 7.7% in 1990 to 31.7% in 2020. Measures of causal narrative structure and causal novelty are positively associated with top-five publication and long-run citations, whereas non-causal counterparts are weakly related or negative.
Presented at
- 12 Mar 2026 — MIT FutureTech Seminar, Virtual
- 5–8 Dec 2025 — CEPR Paris Symposium — Growth Programme, Paris
- 30 Jun 2025 — Metascience 2025, London
- 2–6 Jun 2025 — Networks in Science of Science, Maastricht
- 29–31 May 2025 — EAYE Annual Meeting, London
- 20 May 2025 — ZBW Seminar, Hamburg
- 28–29 Apr 2025 — MPWZ–CEPR Text-as-Data, Virtual
- 7–8 Apr 2025 — PolMeth Europe, London
- 13 Mar 2025 — Text-as-Data (TaDa) Seminar, Virtual
- 25 Nov 2024 — Leibniz Open Science Day, Berlin
- 5–6 Nov 2024 — Causal Data Science Meeting, Virtual
- 10 Oct 2024 — Imperial Internal Seminar, London