Measuring regional economic disparities in Serbia: Multivariate statistical approach

Abstract

The identification of regional economic disparities and their extent is an important factor affecting regional development policy formulation. In this work we propose an alternative, multivariate statistical methodology for evaluation of level of economic development of districts in Serbia, and their classification in homogeneous groups, based on five economic indicators. First, the new composite indicator for measuring economic development level (IED) is created using factor analysis, and then the districts were classified according to the obtained IED values. The evaluation of structural quality of thus formed groups was conducted using the non-hierarchical clustering procedure. The approach presented in this paper takes account of statistical assumptions on which the valid application of multivariate methods is based, which makes it advantageous over the current approaches in the literature. The resulting categorization into three district groups clearly confirms the presence of very pronounced regional economic disparities between the less developed districts in southern and eastern part and more developed districts in northern part of Serbia. Districts with city of Belgrade and Novi Sad occupy the dominant positions compared to other districts.

Keywords

• multivariate statistical analysis
• composite indicator
• regional economic development
• regional disparities
• districts
• Serbia