Estimating Household Consumption Expenditure at Local Level In Italy: The Potential of the Cokriging Spatial Predictor

Abstract

Measuring well-being and living conditions at local level is essential for policy makers who wish to study inequalities and formulate targeted and effective economic and social intervention policies. In Italy, the available official statistics in this field are usually provided at regional level and several studies have been carried out in order to obtain the estimates of those measures at disaggregated level, which is done in order to obtain estimations at provincial level. However, due the heterogeneity of these phenomena within each province, it is important to possess data at micro-territorial level, that is at the municipality level, in order to study and monitor territorial development and inequalities in depth.This paper proposes an estimation of household consumption expenditure, one of the most important indicator of the economic material well-being of an area, for the 7893 Italian municipalities.To this end, the cokriging spatial interpolation technique was employed in order to explore its potentialities. This method is normally used in natural sciences to predict variables of interest at micro territorial level, using available sample data or population aggregates, analysing their spatial dependence and introducing information on auxiliary correlated variables available at micro level. In this study, the available information on household consumption expenditure at provincial level was combined with the data on taxable income at municipality level as auxiliary variable.The evaluation of model performance enabled us to confirm the validity of this approach to obtain a more detailed picture of the local systems for which intervention policies are important.