Network analysis of regulatory regime performance

Network analysis is an increasingly popular approach to the study of international regulatory regimes. Network analysis methods have previously been developed for other contexts, from electricity networks to online social networks. Many of these methods translate well to the setting of regulatory regimes - especially for positive, qualitative questions such as identifying groups and influential actors. To the best of our knowledge, however, there has been little concerted effort to theoretically derive appropriate quantitative network measures for normative questions. Thus, authors who wish to ask questions about regime performance have been forced to borrow measures and methods developed for other contexts. The current paper addresses this gap by deriving quantitative measures of regime performance from first principles. We first develop a model of actor behaviour and interactions in a regime, including objectives and learning mechanisms. We then develop first-best measures of regime performance, derived from regulatory and governance theory. Noting that not all first-best measures are feasible to implement on real data, we also develop and identify a range of second-best measures. Coded simulations of our model regime allow us to directly observe and compare the ability of the different measures to accurately reflect regime performance. On this basis, we recommend a set of empirically feasible network measures of regime performance. We discuss the implications of our findings for analysis of the emerging regime of embedded emissions verification frameworks for products such as iron and steel.