Estimating Spatially Disaggregated Data by Entropy Econometrics: An Exercise of Ecological Inference for Income in Spain

Abstract

The availability of geographically disaggregated data, especially referred to the urban and
metropolitan areas, is a growing need not only for academic studies in the field of economics
but also for policy makers. However, in many cases the degree of disaggregation of official
statistics does not allow to have information at a desirable level. In this paper a methodology to
approximate highly-disaggregated data for the Spanish economy using entropy econometrics is
proposed. The paper illustrates how the procedure works taking as empirical application the
estimation of income for the Spanish municipalities classified according to their size. An
evaluation of the estimates is presented by a simulation exercise and by comparing our results
with previous estimates obtained by statistical agencies using more information-intensive
estimation techniques. Our results suggest that entropy estimators could be considered as an
alternative for recovering disaggregated economic data from aggregate figures, given that the
errors seem relatively low.