Factor-Augmented Machine Learning Panel Regressions

2026-07-07

AI Summary

This study addresses the challenge of cross-sectional dependence in high-dimensional panel data regression induced by common shocks. The authors propose a factor-augmented sparse group LASSO estimator that integrates MIDAS aggregation with a latent factor model. This approach uniquely combines factor structure with sparse group LASSO, leveraging the group sparsity inherent in mixed-frequency time series for effective modeling. Theoretical analysis demonstrates that, under cross-sectional dependence, the proposed method achieves substantially improved parameter estimation consistency and predictive accuracy compared to standard LASSO. Consequently, it offers a robust estimation and inference framework tailored for high-dimensional mixed-frequency panel data.

Abstract

This paper develops the asymptotic theory for high-dimensional panel data regressions in settings with cross-sectionally dependent errors driven by common shocks. We consider a factor-augmented sparse-group LASSO estimator that combines MIDAS aggregation with latent factors. The estimator can take advantage of the mixed-frequency group structure in the time-series dimension. Theory shows that it can outperform the standard LASSO estimator both for prediction and estimation while allowing for cross-sectional dependence.

Problem

• high-dimensional panel data
• cross-sectional dependence
• common shocks
• factor-augmented regression
• mixed-frequency data

Innovation

• factor-augmented
• sparse-group LASSO
• MIDAS aggregation
• cross-sectional dependence
• high-dimensional panel data