Structural Deep Learning in Conditional Asset Pricing

90 Pages Posted: 2 Jun 2022 Last revised: 4 Jun 2023

See all articles by Jianqing Fan

Jianqing Fan

Princeton University - Bendheim Center for Finance

Zheng Tracy Ke

Harvard University

Yuan Liao

Rutgers University, New Brunswick

Andreas Neuhierl

Washington University in St. Louis - John M. Olin Business School

Date Written: May 23, 2022

Abstract

We develop a theoretical analysis of deep learning predictions in cross-sectional asset pricing. Our method is guided by economic theory and employs time-varying conditioning information on alphas and betas carried by firm-specific characteristics. We derive formal asymptotic theory for neural network estimators, and show that deep learning regressions of return on characteristics can be decomposed into mispricing (alpha) and risk related components -- thereby opening the ``black box'' of machine learning predictions. The analysis also shows that agnostic plugin estimators (pre-training a network and then plugging in new data) can only recover the sum of the mispricing and risk-related components.

Keywords: factor pricing model, neural network, double descent, alphas, characteristics, risk premium

JEL Classification: G11, G12, C14, C45, C58

Suggested Citation

Fan, Jianqing and Ke, Zheng and Liao, Yuan and Neuhierl, Andreas, Structural Deep Learning in Conditional Asset Pricing (May 23, 2022). Available at SSRN: https://ssrn.com/abstract=4117882 or http://dx.doi.org/10.2139/ssrn.4117882

Jianqing Fan

Princeton University - Bendheim Center for Finance ( email )

26 Prospect Avenue
Princeton, NJ 08540
United States
609-258-7924 (Phone)
609-258-8551 (Fax)

HOME PAGE: http://orfe.princeton.edu/~jqfan/

Zheng Ke

Harvard University ( email )

1875 Cambridge Street
Cambridge, MA 02138
United States

Yuan Liao

Rutgers University, New Brunswick ( email )

New Brunswick, NJ
United States

Andreas Neuhierl (Contact Author)

Washington University in St. Louis - John M. Olin Business School ( email )

St. Louis, MO
United States

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
1,860
Abstract Views
3,913
Rank
16,934
PlumX Metrics