Working Paper: NBER ID: w24070
Authors: Serhiy Kozak; Stefan Nagel; Shrihari Santosh
Abstract: We construct a robust stochastic discount factor (SDF) that summarizes the joint explanatory power of a large number of cross-sectional stock return predictors. Our method achieves robust out-of-sample performance in this high-dimensional setting by imposing an economically motivated prior on SDF coefficients that shrinks the contributions of low-variance principal components of the candidate factors. While empirical asset pricing research has focused on SDFs with a small number of characteristics-based factors—e.g., the four- or five-factor models discussed in the recent literature—we find that such a characteristics-sparse SDF cannot adequately summarize the cross-section of expected stock returns. However, a relatively small number of principal components of the universe of potential characteristics-based factors can approximate the SDF quite well.
Keywords: Stochastic Discount Factor; Asset Pricing; Bayesian Approach; High-Dimensional Data
JEL Codes: C38; G12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| stochastic discount factor (SDF) (D15) | expected stock returns (G17) |
| characteristic-sparse SDF (C69) | inadequate summary of expected stock returns (G17) |
| redundancy among characteristics (C52) | insufficient for predicting returns (G17) |
| Bayesian approach (C11) | better out-of-sample performance (C52) |
| SDF coefficients penalty (C69) | avoidance of overfitting (C52) |
| stochastic discount factor (SDF) (D15) | significant abnormal returns (G14) |