A uniform Berry--Esseen theorem on $M$-estimators for geometrically ergodic Markov chains

TL;DR

This paper establishes a uniform Berry--Esseen theorem for M-estimators derived from geometrically ergodic Markov chains, extending classical results to dependent data scenarios.

Contribution

It provides the first uniform Berry--Esseen bound for M-estimators in the context of geometrically ergodic Markov chains, under standard regularity conditions.

Findings

Uniform convergence rate established for M-estimators

Results extend classical i.i.d. Berry--Esseen bounds to Markov chains

Applicable under standard regularity and moment assumptions

Abstract

Let $\{X_n\}_{n\ge0}$ be a $V$-geometrically ergodic Markov chain. Given some real-valued functional $F$, define $M_n(\alpha):=n^{-1}\sum_{k=1}^nF(\alpha,X_{k-1},X_k)$, $\alpha\in\mathcal{A}\subset \mathbb {R}$. Consider an $M$ estimator $\hat{\alpha}_n$, that is, a measurable function of the observations satisfying $M_n(\hat{\alpha}_n)\leq \min_{\alpha\in\mathcal{A}}M_n(\alpha)+c_n$ with $\{c_n\}_{n\geq1}$ some sequence of real numbers going to zero. Under some standard regularity and moment assumptions, close to those of the i.i.d. case, the estimator $\hat{\alpha}_n$ satisfies a Berry--Esseen theorem uniformly with respect to the underlying probability distribution of the Markov chain.