Regularized Quantile Regression with Interactive Fixed Effects

TL;DR

This paper introduces a nuclear norm penalized estimator for large panel data quantile regression models with interactive fixed effects, achieving uniform consistency and nearly optimal error bounds without pre-estimating fixed effects.

Contribution

It proposes a convex optimization approach that handles high-dimensional covariates and low-rank fixed effects simultaneously, with new theoretical guarantees and a method to estimate the number of fixed effects.

Findings

Estimator performs well in Monte Carlo simulations.

Provides uniform error bounds valid across quantile levels.

Achieves nearly optimal convergence rates for fixed effects estimation.

Abstract

This paper studies large $N$ and large $T$ conditional quantile panel data models with interactive fixed effects. We propose a nuclear norm penalized estimator of the coefficients on the covariates and the low-rank matrix formed by the fixed effects. The estimator solves a convex minimization problem, not requiring pre-estimation of the (number of the) fixed effects. It also allows the number of covariates to grow slowly with $N$ and $T$. We derive an error bound on the estimator that holds uniformly in quantile level. The order of the bound implies uniform consistency of the estimator and is nearly optimal for the low-rank component. Given the error bound, we also propose a consistent estimator of the number of fixed effects at any quantile level. To derive the error bound, we develop new theoretical arguments under primitive assumptions and new results on random matrices that may be of independent interest. We demonstrate the performance of the estimator via Monte Carlo simulations.