Deep Learning for Predicting Asset Returns

TL;DR

This paper applies deep learning techniques like ReLU and LSTM to asset return prediction, uncovering nonlinear factors that improve predictability, especially at extreme characteristic values, using advanced algorithms and revisiting established datasets.

Contribution

It introduces a deep learning framework for asset return prediction, emphasizing nonlinear factors and demonstrating their significance with empirical analysis.

Findings

Nonlinear factors explain return predictability.

Deep learning models outperform traditional linear models.

Extreme characteristic values are key to prediction accuracy.

Abstract

Deep learning searches for nonlinear factors for predicting asset returns. Predictability is achieved via multiple layers of composite factors as opposed to additive ones. Viewed in this way, asset pricing studies can be revisited using multi-layer deep learners, such as rectified linear units (ReLU) or long-short-term-memory (LSTM) for time-series effects. State-of-the-art algorithms including stochastic gradient descent (SGD), TensorFlow and dropout design provide imple- mentation and efficient factor exploration. To illustrate our methodology, we revisit the equity market risk premium dataset of Welch and Goyal (2008). We find the existence of nonlinear factors which explain predictability of returns, in particular at the extremes of the characteristic space. Finally, we conclude with directions for future research.