Modelowanie zależności między ekonomicznymi procesami przestrzennymi a poziom agregacji danych

Abstract

Statistic and econometric analyses of spatial phenomena use the data of different levels of spatial aggregation. It has particular consequences for the possibilities of discovering different properties and structures of the spatial processes. These properties are the basis for the proper specification of the spatial models of the dependence between the processes. In the paper it is discussed how the spatial aggregation affects the change of the properties and the structures of the spatial processes. In particular the differences as regards the character and the strength of the spatial aggregation influence on the separate components of the processes are shown. The main part of the paper is the discussion on the data aggregation influence on the model- ling of the dependence between the spatial processes. A special attention is paid to the econometric modelling of the phenomena observed at different levels of the data aggregation with the presence of the spatial autocorrelation. The important result of the discussion is to point at the problem of the “naive” spatial models interpretation. The considerations are based on the generated data and they are illustrated with an empirical example.