Regional estimates of poverty indicators based on a calibration technique : 2015 edition

To answer to DG Regio's request, Eurostat wishes to carry out regional estimations of the AROPE indicator (At-Risk-Of-Poverty-or-Exclusion) and its components, derived from the EU-SILC survey. For that purpose, we suggest to use estimators which are named 'synthetic estimators' by the small area estimation theory. Their calculation would be made with a calibration method on regional margins supplying 22 sets of weights implying all the households of the national SILC sample (one set for each region). This method is adapted to the production of regional indicators built from variables of interest which appear well correlated with the calibration variables. We thus use external sources in order to produce regional margins from variables well correlated with the variables making up the AROPE indicator, in particular at the moment the source 'Revenus disponibles localisés'. The method is based essentially on: 1) the existence of 'enough explanatory' variables of the phenomena of interest and for which we can have regional margins; 2) the hypothesis that conditionally in these explanatory variables, the geography does not have impact anymore on the phenomenon of interest. Subject to these points, the production of the regional indicators using the SILC source seems satisfactory.