COMPSTAT 2022: Start Registration
View Submission - COMPSTAT2022
Title: Full bias correction approaches for M-quantile small area estimators: Application to Italian labour force survey Authors:  Francesco Schirripa Spagnolo - Università di Pisa - Dipartimento di Economia e Management (Italy) [presenting]
Gaia Bertarelli - Sant'Anna school of Advanced Studies (Italy)
Raymond Chambers - University of Wollongong (Australia)
David Haziza - Université de Montréal (Canada)
Nicola Salvati - University of Pisa (Italy)
Abstract: When representative outlier units are a concern for the estimation of population quantities, it is essential to pay attention to them in a small area estimation (SAE) context. Standard approaches use plug-in robust prediction replacing parameter estimates in optimal but outlier-sensitive predictors with outlier robust versions. These predictors are efficient under the correct model but may be sensitive to the presence of outliers because they use plug-in robust prediction which usually leads to a low prediction variance and a considerable prediction bias. We propose two new full bias correction methods to reduce the prediction bias of the robust M-quantile predictors in SAE for continuous, count and binary data. The first estimator is based on the concept of conditional bias. The second one is based on a full bias correction. The properties of the proposed estimators are empirically assessed in model-based and design-based simulations. These estimators correct for the bias and are more efficient than the robust-predictive estimators and robust-projective estimators in the presence of area and individual outliers. Two estimators of the prediction mean-squared error are described. The methodology proposed is applied to Italian annual Labour Force Survey data for estimating the proportion of the unemployed in local labour market areas.