Working Papers
Peer-reviewed theory does not help predict the cross-section of stock returns
(with Alejandro Lopez-Lira and Tom Zimmermann)
Predictors that peer review attributes to risk underperform out-of-sample
Publication Bias in Asset Pricing Research (with Tom Zimmermann)
This review was requested by the Oxford Research Encyclopedia of Economics and Finance
We review the evidence on publication bias in anomalies and clarify common misinterpretations ("insignificant findings" are not "false findings").
Missing Data and the Dimensionality of Expected Returns (with Jack McCoy)
R&R, Journal of Financial Economics
Imputed data for 125 cross-sectional predictors (3.2 GB zipped csv), Same data without imputed values (Box-Cox transformed, 2 GB zip), Data readme
When applying ML methods to anomalies, you need to impute missing values. It turns out imputing with means is fine because the anomalies are, surprisingly, largely independent.
Do t-stat Hurdles Need to be Raised?
No, not really, because (1) t-hurdles are not identified due to publication bias and (2) strongly identified statistics make asset pricing look pretty good.
Paper (Another major revision, April 2022), Code
The revision addresses the points made in this referee report.
Old Draft (February 2020), Old Code
This old draft modeled standard errors, which could in principle help with identification (Andrews and Kasy 2019). But because there are so few observations of insignificant t-stats it doesn't change the main results.
Most claimed statistical findings in cross-sectional return predictability are likely true
Harvey, Liu, and Zhu (2016) is commonly interpreted as saying most findings are false. Strangely, their FDR estimates imply the opposite. I show their FDR numbers are robust and explain them with simple formulas.
An Irrelevance Theorem for Risk Aversion and Time-Varying Risk (with Francisco Palomino)
A general theorem explains why modern theories of risk tell us little about business cycles.
Permanent Working Papers
Semi-Parametric Restrictions on Production-Based Asset Pricing Models
Matching the data on asset prices requires either extremely volatile IST shocks or huge capital adjustment costs.
The best parts of this paper are extended and can be found in "An Irrelevance Theorem for Risk Aversion and Time-Varying Risk" with Francisco Palomino.