Lassoed Boosting and Linear Prediction in the Equities Market

TL;DR

This paper introduces a two-stage linear regression method combining lasso screening and boosting re-estimation, demonstrating superior prediction accuracy and sparser models in equity return prediction compared to existing methods.

Contribution

It proposes a novel two-stage estimation approach that enhances variable selection and prediction accuracy in linear models, especially in financial applications.

Findings

Lassoed boosting performs as well as the relaxed lasso in simulations.

Under certain scenarios, it produces sparser models.

It achieves the lowest mean-squared prediction error in equity return prediction.

Abstract

We consider a two-stage estimation method for linear regression. First, it uses the lasso in Tibshirani (1996) to screen variables and, second, re-estimates the coefficients using the least-squares boosting method in Friedman (2001) on every set of selected variables. Based on the large-scale simulation experiment in Hastie et al. (2020), lassoed boosting performs as well as the relaxed lasso in Meinshausen (2007) and, under certain scenarios, can yield a sparser model. Applied to predicting equity returns, lassoed boosting gives the smallest mean-squared prediction error compared to several other methods.