by Enrico Marelli
Start page: 149 - End page: 178
Keywords: regional economies; European regions; growth and convergence; specialisation
Jel code: O52; P52; R11
The purpose of this paper was to analyze specialization and convergence of European countries and regions, within the framework of integration in the EU. This is important not only for long-term real convergence processes, but also for a proper functioning of the monetary union (in the line of research on the OCA's criteria, asymmetry of shocks and synchronization of business cycles). The position of new member states is particularly delicate, also considering the forthcoming adoption of the euro by some of them. As indicated by the EU Treaty, economic growth should be balanced with economic and social cohesion that includes a careful consideration of regional disparities. Our empirical investigation focuses on the regions of EU25, further broken up into other relevant groupings (EU15, EMU, and the new members' EU10 group), over the period from 1980 (or 1990 for EU10) to 2005. This paper considers a rather fine regional disaggregation (NUTS-2 level), counting 250 regions. The analysis of different indices of specialisation point to a prevalent increase of homogeneity of sector structures across European regions, although in some cases (especially in the industrial sector and in some services) specialisation has increased. For convergence, a sigma convergence's analysis confirms a reduction of disparities, both at a country and regional level. However, a trade-off between fast national growth and internal distribution has emerged in the early stages of development, as in the case of new members. Moreover, beta convergence has also been established - regarding per capita income, employment and productivity - for almost all territorial aggregates (excluding the new members since 1999). The addition of structural variables, following a beta-conditional approach, indicates a positive role for services and a negative impact of agriculture. Finally, some preliminary results have been obtained by the innovative inclusion of specialisation indices within convergence regressions.