The application of local indicators for categorical data (LICD) in the spatial analysis of economic development

The objective of this paper is to identify classes of regions presenting different economic situations and apply a join-count test to examine spatial dependences between these classes. The test examines spatial autocorrelation on the basis of qualitative data. The global join-count test indicates general interactions occurring between regions, while the local join-count test examines a tendency to form the spatial clusters (e.g. metropolitan areas). The study covers the situations of 66 Polish NUTS 3 regions in 2011. Regions were divided into two classes presenting relatively low and high levels of economic development. Taxonomic methods of multivariate data analysis were applied in the research. The global test proved spatial clustering of economically poor regions but was statistically insignificant as regards well-developed regions. Thus the joincount local join-count test was additionally applied. The test indicated the occurrence of five spatial clusters of NUTS 3 regions. Three of them include economically well-developed regions, while two of them present poor economic situations. Furthermore three spatial outliers (local growth centres), which deteriorate the economic situation of eastern Poland, were also recognized.