Factor-augmented sparse MIDAS regressions with an application to nowcasting

TL;DR

This paper introduces a novel factor-augmented sparse MIDAS regression method for high-dimensional mixed-frequency data, demonstrating improved nowcasting of U.S. GDP growth, especially during the COVID-19 pandemic, by capturing both sparse and dense shocks.

Contribution

It develops a new estimation approach combining sparse and dense techniques for MIDAS regressions, with proven convergence rates under complex data conditions.

Findings

Outperforms standard methods in nowcasting GDP during COVID-19

Captures both idiosyncratic and common shocks effectively

Provides a practical implementation in the midasml R package

Abstract

This article investigates factor-augmented sparse MIDAS (Mixed Data Sampling) regressions for high-dimensional time series data, which may be observed at different frequencies. Our novel approach integrates sparse and dense dimensionality reduction techniques. We derive the convergence rate of our estimator under misspecification due to the MIDAS approximation error, $\tau$-mixing dependence, and polynomial tails. Our method's finite sample performance is assessed via Monte Carlo simulations. We apply the methodology to nowcasting U.S. GDP growth and demonstrate that it outperforms both sparse regression and standard factor-augmented regression during the COVID-19 pandemic. These findings indicate that the growth through this period was influenced by both idiosyncratic (sparse) and common (dense) shocks. The approach is implemented in the midasml R package, available on CRAN.