Changes in the spatial pattern of net earnings: Evidence from Serbia

Abstract

Spatial autocorrelation analysis is an important method that can reveal the structure and patterns of economic spatial variables. It can be used to identify not only global spatial patterns in the country, but also characteristic locations at micro levels. In this research, we used spatial autocorrelation methodologies, including Global Moran’s I and Local Getis—Ord Gi statistics to identify the intensity of the spatial clustering of municipalities in Serbia by the level of average monthly net earnings from 2001 to 2010. We identified and mapped local clusters (hot and cold spots) by the level of average monthly net earnings for the same period. The results show that overall spatial segregation between municipalities with high and low average monthly net earnings was predominantly increasing during the investigated period. Local statistics illustrated that overall spatial segregation followed a broad north—south divide, with a concentration of municipalities with high net earnings in the north of Serbia, and low net earnings in the south. Closer inspection showed that at the beginning of the study period, there were three statistically significant hot spots in the north. As time passed, only one highly clustered hot spot remained — the Belgrade region. One cold spot retained a relatively stable position in the country’s southeast. This research shows that spatial changes of net earnings can be successfully studied with respect to statistically significant global and local spatial associations in the variables using spatial autocorrelation analysis.