Economic and Social Determinants of Internal Migration in Poland in the Light of the Classification Trees Method

This paper analyses socio-economic determinants of migration flows between Polish sub-regions in the years 2008-2010 by using the classification trees (CART) method. Six explanatory variables are selected to determine migration, including GDP per capita, number of economic entities (firms) per 100 inhabitants, investment assets and tangible assets per capita, registered unemployment rate, as well as average monthly salary. The CART method is then used to build models explaining the classification of migration flows into four quartile-based categories. The results confirm that the classification of internal migration flows is strongly determined by socio-economic features, in particular the number of economic entities, investment assets per capita and the unemployment rate. The suburbanisation from cities to neighbouring sub-regions is clearly demonstrated. Better developed regions, especially the largest Polish cities, have the highest migration outflows as well as inflows, yet retain positive net migration. We argue that the proposed analytical approach enables to determine the multidimensional relationships between explanatory social-economic variables and the migration coefficients under study