From Econometrics to Machine Learning: Transforming Empirical Asset Pricing
综述了实证资产定价从传统计量经济学向机器学习的转变，比较了两者的优势与挑战，并提出了一个基于随机贴现因子的统一框架，在保持经济可解释性的同时整合机器学习，为金融市场的深入洞察和稳健实证研究提供了新方向。
ABSTRACT Empirical asset pricing is undergoing a transformation with the advent of big data and machine learning. Traditional multifactor models offer simplicity and interpretability but struggle with high‐dimensional covariates and nonlinear relationships. Machine learning, with its predictive power and flexibility, provides a promising alternative. This paper surveys the transition from econometrics to machine learning, tracing the evolution of asset pricing models, addressing empirical challenges, and comparing the strengths and challenges of both approaches. A unified framework based on the stochastic discount factor is proposed, integrating machine learning while preserving economic interpretability. By emphasizing predictive accuracy and theoretical rigor, this paper highlights how machine learning can reshape empirical asset pricing, offering deeper insights into financial markets and new directions for robust empirical research.