Asymmetric ex post winsorization in asset pricing involves discarding extreme returns below a certain one-sided percentile, i.e., kick out any return < 0.50th percentile (ex post). This does not constitute "data cleaning". Data mining? Maybe. The result: spurious anomaly

Alex Dickerson Well played 😊. Return winsorization is unjustifiable. If measurement error is a concern, one has to model it and write a proper likelihood (eg, price measurement error generates an MA(1) with negative coefficient). Winsorizing is either lazy or disingenuous (or both).

New chart coming soon (w/ Alejandro Lopez-Lira + Tom Zimmermann): The relationship between modeling rigor and post-sample performance is slightly negative 🤔

How credible was the "credibility revolution"? How robust is empirical research in economics? We just replicated a year's worth of the American Economic Review & had economists predict robustness. Here's what we learned. econstor.eu/handle/10419/2…

Bruno Rodrigues (@[email protected]) I usually just laugh at Python users who spend hours a day resolving circular imports, and taunt them with the meme below (they hate me)👇

Happy to announce that my paper, “Priced risk in corporate bonds” was awarded the Fama-DFA best paper award by the Journal of Financial Economics (JFE). I would like to extend a sincere thank you to my coauthors, the editorial team at the JFE and everyone else who brought this

Corporate bond issuer CUSIP to Compustat (GVKEY) and CRSP (PERMNO) monthly linker file now on Open Source Bond Asset Pricing (OSBAP), openbondassetpricing.com/bond-compustat… courtesy of Chuck Fang (Drexel), Chuck Fang. Chuck's associated paper, 'Monetary Policy Amplification through Bond Fund

The submission deadline for NBER/RFS big data conference this year is September 12, 2024. nber.org/calls-papers-a…

What have been the underlying drivers of a potential "replication crisis" in corporate bond asset pricing research over the last decade? Our updated working paper, papers.ssrn.com/sol3/papers.cf…, documents some root causes: (1) Not adjusting signals for potential microstructure noise