Status
Under review, Journal of Applied Econometrics.
Download
- Paper (available upon request)
Abstract
Why did the credibility revolution spread so widely without converging on one method? Using full-text method classifications in the American Economic Review and American Economic Journals, 2011 to 2024, I build a multinomial discrete-choice model in which diffusion comes from fundamentals, slow adjustment, or a network externality that tips the field above a threshold. The data do not support tipping. Instrumenting coauthor exposure and isolating the timing of adoption, I bound the local peer effect, its upper edge an order of magnitude below the tipping threshold. The timing-isolating designs are null to negative, and the positive designs fail their pre-trend placebos and are consistent with sorting. The network correlations reflect community sorting and are near-uniform across a sevenfold barrier range. The aggregate dynamics are the persistence of fundamentals under slow adjustment. The most rewarded design, regression discontinuity, is not the most adopted, because its entry barrier is high.
Keywords
Coordination Games, Network Externalities, Discrete Choice, Social Interactions, Equilibrium Multiplicity, Technology Adoption, Lock-in
Citation
Chávez Padilla, Carlos César. 2026. “The Diffusion of Econometric Methods: A Structural Model of the Credibility Revolution.” Under review, Journal of Applied Econometrics.