# Quasi Maximum-Likelihood Estimation of Dynamic Panel Data Models

## TL;DR

This paper proves the convergence and normality of quasi maximum-likelihood estimators for dynamic panel data models, demonstrating their robustness and superior finite sample performance over GMM estimators through Monte Carlo simulations.

## Contribution

It introduces a robust QML estimation method for dynamic panel data models, including an ECME algorithm and a comparison with GMM estimators.

## Key findings

- QML estimators have smaller bias and root mean squared errors than GMM estimators.
- The paper establishes the theoretical properties of QML estimators, including convergence and asymptotic normality.
- Monte Carlo experiments show QML estimators outperform GMM in finite samples.

## Abstract

This paper establishes the almost sure convergence and asymptotic normality of levels and differenced quasi maximum-likelihood (QML) estimators of dynamic panel data models. The QML estimators are robust with respect to initial conditions, conditional and time-series heteroskedasticity, and misspecification of the log-likelihood. The paper also provides an ECME algorithm for calculating levels QML estimates. Finally, it uses Monte Carlo experiments to compare the finite sample performance of levels and differenced QML estimators, the differenced GMM estimator, and the system GMM estimator. In these experiments the QML estimators usually have smaller --- typically substantially smaller --- bias and root mean squared errors than the panel data GMM estimators.