Abstract:
Much research has highlighted that sub-state entities (SSEs) - such as the German Länder, Spanish autonomous communities or French regions - mobilise at the European level. This literature, however, suffers from a research gap on the question of how this sub-state activity interacts with that of its own member state. In other words, while it is clear that SSEs do represent their interests in Brussels, it is not so clear whether this activity is carried out with their member state (cooperation), without their member state (non-interaction) or against their member state (conflict). This thesis fills such a research gap by 1) identifying what the pattern of interaction between state and sub-state EU interest representation corresponds to and by 2) identifying what the determinants of such a pattern are. To achieve this double task, quantitative and qualitative methods are employed. The quantitative section consists of regression analysis on data collected through a survey addressed to the Heads of regional offices in Brussels (n=104). It highlights that cooperation is the most frequent outcome, followed by noninteraction. Conflicting interest representation is the least frequent outcome. It also indicates that, contrary to expectations, devolution levels do not affect conflict but increase the frequency of cooperation and decrease that of non-interaction. Meanwhile, party political incongruence fails to affect conflict, decreases cooperation and increases non-interaction. Finally, preference intensity configurations affect all three outcomes. This quantitative work was complemented by a series of in-depth case study analyses of Scotland (UK), Salzburg (Austria), Rhône-Alpes and Alsace (both France). Based on over a hundred semi-structured interviews with state, sub-state and supra-state officials and politicians, these case studies confirmed the overall findings reached through quantitative means and further suggested that the effect of devolution overrides that of party political incongruence. Additional statistical testing confirmed this inductive finding. The concluding sections highlight this research’s overall theoretical and policy implications.