A1023
Title: Down to earth: Estimating GDP components in small areas using satellite data
Authors: Claudio Lupi - University of Molise (Italy)
Luca Romagnoli - University of Molise (Italy) [presenting]
Abstract: Although the development of macroeconomic applications of data extracted from satellite imagery is relatively recent, satellite data have been regularly used to derive GDP estimates, especially in regions where official statistics are scarce or unavailable. Satellite nighttime light and land cover data are used here together with cross-validated geographically weighted regressions to estimate disposable income at the municipal level in Italy, two years ahead of official statistics. The geographically weighted regressions are used explicitly to account for the known spatial non-stationarity of the relationship between nighttime light and economic activity. In addition to disposable income at the municipal level, an estimate of value added with the same spatial breakdown is produced to supplement the official statistics, which are limited to the level of the Italian provinces and are only available with a considerable delay. For larger municipalities, finer estimates are possible, which map the spatial heterogeneity of income and value added even within the same municipality.
