The History of the Cross-Section of Stock Returns

Abstract

and Western Finance Association 2016 meetings for valuable comments. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research. NBER working papers are circulated for discussion and comment purposes. They have not been peer-reviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.

Sparse Signals in the Cross‐Section of Returns

“…In other words, the LASSO ignores any predictor weaker than 2.5% per month when making its one-minute-ahead return forecasts. This lower bound is twice as large as well-known predictors at the weekly and monthly horizon documented in the academic literature (McLean and Pontiff (2016), Harvey, Liu, and Zhu (2016), Linnainmaa and Roberts (2017)). As a result, there is little relation between the predictors identified by the LASSO each minute and existing predictors documented in the academic literature, which makes the LASSO's choice of predictors seem unexpected.…”

“…In other words, the LASSO typically ignores all predictors weaker than λ = 2.5% per month when making its one-minuteahead return forecasts. This lower bound is more than twice as large as the size of well-known predictors at the weekly and monthly horizon (McLean and Pontiff (2016), Harvey, Liu, and Zhu (2016), Linnainmaa and Roberts (2017)). This is why our implementation of the LASSO at the one-minute horizon is not identifying the existing steady long-lived predictors documented in the academic literature.…”

“…This paper builds on several strands of the asset-pricing and statistics literature. To start with, our paper relates to work on return predictability (McLean and Pontiff (2016), Harvey, Liu, and Zhu (2016), Linnainmaa and Roberts (2017)). Campbell and Thompson (2008) show that many of the predictors documented in the academic literature work much better insample.…”

Publication Bias and the Cross-Section of Stock Returns

“…Harvey, Liu, and Zhu (2015), Linnainmaa andRoberts (2016), andHou, Xue, andZhang (2017) find that most published results are false, while Green, Hand, and Zhang (2014), McLean and Pontiff (2016), Jacobs and Müller (2016) come to the opposite conclusion.…”

“…Some argue that, subject to this much questioning, the data will tell you whatever you want to hear. Indeed, the data have informed us of more than one hundred portfolios with high returns and low market risk, leading many to be suspicious of information obtained in this manner (for example, Lo and MacKinlay (1990), Sullivan, Timmermann, and White (1999), Harvey, Liu, and Zhu (2015), Linnainmaa and Roberts (2016), Chordia, Goyal, and Saretto (2017)). 1 Our interrogation of the data is subject to controls, however.…”