Comparative Regional GDP Analysis – Case Study of Croatia

This paper as a part of an extensive research about the Croatian economic challenges within global recession environment has focused on the estimation of integral Croatian counties GDP per capita model. The main intention is to quantify each partial influence of a set of independent regional variables on regional GDP per capita. Stepwise variable selection method points out that regional GDP per capita is well explained only by two regional variables: Exports and Foreign tourist arrival. Namely, correlation analysis indicated existence of strong multicolinearity effects between Exports and all other regressors variables. Regional Croatian counties GDP per capita model is estimated by classical multiple regression approach and then by multiple Monte Carlo simulations. The comparative analysis of the results confirms the advantage of Monte Carlo simulation in evaluation of interdependence especially regarding Gauss-Markov conditions on the stochastic part of econometric model. For this regional analysis multiple Monte Carlo simulation of regression model is used for the first time to measure individual average influence of each regressor variable on the GDP per capita. The comparative analysis is enriched by Cluster analysis of the same set of Croatian counties according to appropriate indicators. Complete set of the comparative analysis results confirms the general conclusion that EU regional Croatian NUTS 2 spacing has been done according to a set of administrative geographic and demographic criteria neglecting the socio-economic one. That is why the legal Croatian counties administrative spacing into 21 counties is currently the best reflection of the Croatian socio-economic and geographic status by relevant development and economic indicators. That is why the new methodology for spatial-economic entities' classification is a more useful characterisation of the territory for policy-making purposes.