Deep residual models improve asset pricing predictions and returns

Residual Learning in Empirical Asset Pricing

Machine Learning

Summary

Predicting asset prices well is important but hard. The authors found that making neural networks deeper helps, but only if the deeper model builds on simpler ones instead of replacing them. They show this method, called residual learning, improves financial return predictions more than shallow or traditional deep models. This approach can make complex asset pricing models more practical and valuable.

What this means in practice

  • For quantitative finance teams: Enhance portfolio performance predictions by using deep residual neural network models that improve on simpler baseline models.$Commercial implications: This approach enables financial technology firms to offer more accurate asset pricing tools that improve portfolio returns for investors.
  • For machine learning engineers: Integrate residual learning to deepen neural network models in financial applications, improving their effectiveness over standard feedforward networks.

Authors

Dexin Peng, Xiaoyu Wang

Abstract

Shallow models are special cases of deep models, and deep models theoretically have the potential to outperform the shallow ones. However, the existing empirical asset pricing literature provides strong benchmarks for shallow models. Residual learning allows neural network models in asset pricing to go deeper by preserving and refining their shallow counterparts. The out-of-sample Sharpe ratio for value-weighted long-short portfolios of deep residual models (2.07) is higher than that for the corresponding shallow ones (1.92) and more than twice that of the deep feedforward models (0.89). We show that model depth is a source of additional economic value in asset pricing. Residual learning can be used to deepen other neural-network-based asset pricing models if they contain intermediate layers. Our design also provides one way to scale asset pricing models, making native "large asset pricing models" more feasible.