Deep Learning for Conditional Asset Pricing Models

ArXiv ID: 2509.04812 “View on arXiv”

Authors: Hongyi Liu

Abstract

We propose a new pseudo-Siamese Network for Asset Pricing (SNAP) model, based on deep learning approaches, for conditional asset pricing. Our model allows for the deep alpha, deep beta and deep factor risk premia conditional on high dimensional observable information of financial characteristics and macroeconomic states, while storing the long-term dependency of the informative features through long short-term memory network. We apply this method to monthly U.S. stock returns from 1970-2019 and find that our pseudo-SNAP model outperforms the benchmark approaches in terms of out-of-sample prediction and out-of-sample Sharpe ratio. In addition, we also apply our method to calculate deep mispricing errors which we use to construct an arbitrage portfolio K-Means clustering. We find that the arbitrage portfolio has significant alphas.

Keywords: Asset Pricing, Deep Learning, Long Short-Term Memory (LSTM), Conditional Factor Risk Premia, Arbitrage Portfolios, Equities (US Stocks)

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 7.0/10
  • Quadrant: Holy Grail
  • Why: The paper employs advanced deep learning architectures (pseudo-Siamese networks with LSTM) and derives conditional asset pricing models with nonlinear stochastic discount factors, indicating high mathematical complexity. It demonstrates empirical rigor through out-of-sample backtesting on a long historical dataset (1970-2019), reporting performance metrics like Sharpe ratios and significant alphas on constructed portfolios.
  flowchart TD
    Goal["Research Goal: Develop a deep learning conditional asset pricing model & quantify mispricing"]
    Data["Data: U.S. Stocks (1970-2019)<br>Financial & Macro Characteristics"]
    Method["Methodology: Pseudo-Siamese Network for Asset Pricing (SNAP)<br>Deep Alpha/Beta + LSTM"]
    Proc["Computational Process:<br>1. Deep Factor Estimation<br>2. Mispricing Error Calculation<br>3. K-Means Clustering"]
    Out["Key Findings: High Out-of-Sample Sharpe Ratio<br>Significant Alphas in Arbitrage Portfolio"]