M-quantile small area estimation for panel data

Abstract

Economic indicators need to be estimated at regional level. Small area estimation based on M-quantile regression has recently been introduced by Chambers and Tzavidis (Biometrika 93:255–268, 2006) and it has proved to provide a valid alternative to traditional methods. Thus far, this method has only been applied to cross-sectional data. However, it is well known that the use of panel data may provide significant gains in terms of efficiency of the estimators. This paper explores possible extensions of M-quantile-based small area estimators to the panel data context. A model-based simulation study is presented.