We use supervised learning to identify factors that predict the cross-section of returns and maximum drawdown for stocks in the US equity market. Our data run from January 1970 to December 2019 and our analysis includes ordinary least squares, penalized linear regressions, tree-based models, and...

We provide a data-driven adjustment for estimated betas that leads to material improvements in the accuracy of weights and risk forecasts for minimum variance portfolios. Like the widely used Blume 2/3 rule and Vasicek correction developed in the 1970s, our beta adjustment operates by shrinking...