This thesis addresses the question as to why multilevel politics is becoming an integral part of politics in Europe. Multilevel politics is conceptualized as a system which functions through a complex web of political relations within and across levels of decision making. The thesis argues that the rise of multilevel politics can be explained by its institutionalization in terms of the emergence, the evolution and especially the effects of relevant institutions. Based on a mixed-method research project, the influence of European institutions on subnational actors and the alignment of actor motives with institutional characteristics are empirically shown.

The first chapter of the dissertation establishes the centrality of institutions for political transformation, examines the role of transnational and domestic institutions for multilevel politics, and contextualizes the research question in terms of institution-actor relations. The second chapter develops a new-institutionalist theoretical framework that explains the emergence, the evolution and the effects of the institutions, and formulates a series of hypotheses with regard to freestanding institutional influence, power distribution, material benefits and political identification. The third chapter outlines the mixed-method research design which addresses individual-level and institutional-level variations through a Europe-wide survey and a comparative case study.

The fourth chapter on survey results shows generally favourable views on multilevel politics, and strong associations of these views with the independent variables under scrutiny. The fifth chapter specifies a multivariate model which includes all posited variables and confirms the majority of the hypotheses. Therefore, the new-institutionalist argument is broadly confirmed, while there is relatively weak evidence to sustain sociological explanations. The final chapter compares the Committee of the Regions and the Congress of Local and Regional Authorities, and examines the institutional characteristics which correspond to the hypothesized variables. It is then concluded that the two institutions share several overarching similarities, and display complementarity in other aspects.