Dissecting Characteristics Nonparametrically

Abstract

We propose a nonparametric method to study which characteristics provide incremental information for the cross-section of expected returns. We use the adaptive group LASSO to select characteristics and to estimate how selected characteristics affect expected returns nonparametrically. Our method can handle a large number of characteristics and allows for a flexible functional form. Our implementation is insensitive to outliers. Many of the previously identified return predictors don’t provide incremental information for expected returns, and nonlinearities are important. We study our method’s properties in simulations and find large improvements in both model selection and prediction compared to alternative selection methods. Authors have furnished an Internet Appendix, which is available on the Oxford University Press Web site next to the link to the final published paper online.

Anomalies and the Expected Market Return

“…We consider three dimension-reduction techniques: The first takes the cross-sectional average of the individual predictors, the second extracts the first principal component from the set of predictors Ng (2007, 2009)), and the third, an implementation of partial least squares (PLS; Wold (1966)), extracts the first target-relevant factor from the set of predictors Pruitt (2013, 2015), Huang et al (2015)). As indicated previously, the forecasts based on strategies 3 Our study complements recent studies that employ machine-learning techniques to predict stock returns using alternative predictor variables in high-dimensional settings, including Rapach, Strauss, and Zhou (2013); Chinco, Clark-Joseph, and Ye (2019); Rapach et al (2019);Freyberger, Neuhierl, and Weber (2020); Gu, Kelly, and Xiu (2020); Kozak, Nagel, and Santosh (2020); Avramov, Cheng, and Metzker (2021); Chen, Pelger, and Zhu (2021); Cong et al (2021); Han et al (2021);and Liu, Zhou, and Zhu (2021).…”

Low‐Risk Anomalies?

“…Nonetheless, extending our work to alternative SDF specifications and applying it to shed light on other cross-sectional and possibly also time-series patterns in asset prices is likely to be a fruitful avenue for future research. It would be particularly interesting to study whether accounting for coskewness can contribute to recent efforts to organize the factor "zoo" (e.g., Cochrane (2011), Freyberger, Neuhierl, and Weber (2020), Feng, Giglio, and Xiu (2020)).…”

Forecast combinations for benchmarks of long-term stock returns using machine learning methods

“…However, sparsity could be an issue with our data set and a careful imposition of structure in the statistical modelling process is helpful. Note further that the use of nonlinear functions f in (1) has shown evidence of much stronger stock return predictability when compared to their linear counterparts (Lettau & Van Nieuwerburgh, 2008;Chen & Hong, 2010;Yang et al, 2010;Cheng et al, 2019;Caldeira et al, 2020;Freyberger et al, 2020). Thus, the local-linear smoother based on the standard L 2 -loss function is ideally suited.…”