Empirical Asset Pricing via Machine Learning

We synthesize the field of machine learning with the canonical problem of empirical asset pricing: Measuring asset risk premia. We use the widely understood empirical setting of predicting the time series and cross section of... [ view full abstract ]

We synthesize the field of machine learning with the canonical problem of empirical asset pricing: Measuring asset risk premia. We use the widely understood empirical setting of predicting the time series and cross section of stock (and portfolio) returns to perform a comparative analysis of methods in the machine learning repertoire. At the broadest level, we find that machine learning has great promise for describing asset price behavior. We identify the best performing methods and trace their predictive gains to allowance of non-linear predictor interactions that are missed by other methods.