MULTILEVEL MODELS: AN APPLICATION TO THE BETA-CONVERGENCE MODEL

Many kinds of data in the social sciences have a hierarchical, multilevel or clustered structure. For example, municipalities are grouped into regions; regions are formed within countries; and quite often, countries belong to supra-national organizations. Once groupings are established, they will tend to become differentiated, and this differentiation implies that the group and its members both influence and are influenced by group membership. To ignore this relationship is to risk overlooking the importance of group effects and it may also render invalid many of the traditional statistical analysis techniques used for studying data relationships. In this paper, we specify a basic two-level model for a conditional beta-convergence model of a sample of European NUTS-2 regions. Specifically, we test for the role of regional decentralization (country-level variable) on regional income growth, since it has been suggested that countries with a governmental form of regional decentralization foster innovation and economic growth.