This article provides an introduction to difference and system GMM in Stata, focusing on their application in dynamic panel data models. It explains the theoretical foundations of GMM, including the linear GMM estimator, and discusses the design of difference and system GMM estimators for "small T, large N" panels. The article outlines the use of the xtabond2 command in Stata to implement these estimators, including the Windmeijer correction for standard errors and the Arellano–Bond test for autocorrelation. It also describes the abar command for performing the Arellano–Bond test after other Stata commands. The article emphasizes the importance of proper use of these estimators, including the need for robust standard errors and careful selection of instruments. It highlights the efficiency of GMM in asymptotic settings and the challenges of finite samples, including the potential for overfitting and the need for feasible GMM estimators. The article concludes with practical tips for using these estimators effectively in empirical research.This article provides an introduction to difference and system GMM in Stata, focusing on their application in dynamic panel data models. It explains the theoretical foundations of GMM, including the linear GMM estimator, and discusses the design of difference and system GMM estimators for "small T, large N" panels. The article outlines the use of the xtabond2 command in Stata to implement these estimators, including the Windmeijer correction for standard errors and the Arellano–Bond test for autocorrelation. It also describes the abar command for performing the Arellano–Bond test after other Stata commands. The article emphasizes the importance of proper use of these estimators, including the need for robust standard errors and careful selection of instruments. It highlights the efficiency of GMM in asymptotic settings and the challenges of finite samples, including the potential for overfitting and the need for feasible GMM estimators. The article concludes with practical tips for using these estimators effectively in empirical research.