Robust Estimation and Inference Methods for Econometric Analysis Under Data Contamination

Econometric analysis is often based on clean and well-behaved data assumptions but financial and economic data in the real world is often contaminated by outliers, measurement error, missing cases and misspecification of the model. The parameter estimates when classical econometric methods are used are severely biased, and such contamination can invalidate the statistical inference. The chapter makes a significant attempt to analyze the effective estimation and inference methods which are aimed at handling such problems and improve the credibility of econometric analysis in the circumstances when the data is contaminated. It presents the background principles of robustness, influence functionalities, and breakdown points, and also reviews various robust estimators and inference routines which can be used with cross-sectional, time-series, and panel data models.