Autoencoder asset pricing models

We propose a new latent factor conditional asset pricing model. Like Kelly, Pruitt, and Su (KPS, 2019), our model allows for latent factors and factor exposures that depend on covariates such as asset characteristics. But, unlike the linearity assumption of KPS, we model factor exposures as a flexib...

Full description

Saved in:
Bibliographic Details
Published in:Journal of econometrics 2021-05, Vol.222 (1), p.429-450
Main Authors: Gu, Shihao, Kelly, Bryan, Xiu, Dacheng
Format: Article
Language:English
Subjects:
Analysis
Asset management
Asset pricing
Autoencoder
Big data
Business schools
Conditional asset pricing model
Machine learning
Neural networks
Nonlinear factor model
Nonlinear systems
Prices
Pricing
Stock returns
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We propose a new latent factor conditional asset pricing model. Like Kelly, Pruitt, and Su (KPS, 2019), our model allows for latent factors and factor exposures that depend on covariates such as asset characteristics. But, unlike the linearity assumption of KPS, we model factor exposures as a flexible nonlinear function of covariates. Our model retrofits the workhorse unsupervised dimension reduction device from the machine learning literature – autoencoder neural networks – to incorporate information from covariates along with returns themselves. This delivers estimates of nonlinear conditional exposures and the associated latent factors. Furthermore, our machine learning framework imposes the economic restriction of no-arbitrage. Our autoencoder asset pricing model delivers out-of-sample pricing errors that are far smaller (and generally insignificant) compared to other leading factor models.
ISSN:0304-4076
1872-6895
DOI:10.1016/j.jeconom.2020.07.009