Abstract
In this paper a categorization of 1,212 European NUTS 3 regions is conducted by means of factor and cluster analysis according to economic structure and spatial characteristics. Subsequently the hypothesis is tested that these region types do explain differences in level and growth of regional income. The resulting 14 region types (10 non-urban and 4 urban) do show differences in regional income per capita and some of them are expected to converge to different steady state levels. In particular, region types with low employment rates obtain lower per capita income on average than others, while those with productivity differentials in favour of industry obtain higher GDP p.c. when judged against region types comparable in their degree of accessibility. In estimating regressions on conditional beta-convergence, the inclusion of national dummies shows a reduced speed of convergence (in the total and the western sample) and even divergence in the sample of (former communist) new member states, while the additional inclusion of region type dummies points again to a still low but higher speed of convergence. The estimates indicate significant lower steady state incomes in the peripheral agrarian regions, peripheral industrial regions with a lower productivity differential in favour of industry, peripheral tertiary regions and both types of central regions with low employment rates. A higher steady state income is estimated for metropolitan areas and big agglomerations.