A Network Analysis Method for Analyzing Complex Governance Structures

Scholars and practitioners alike increasingly recognize that governance often occurs not via hierarchies – in which one all-powerful government agency makes important decisions – but instead via networks, in which multiple actors play different roles to achieve outcomes that could not be achieved by one single agency acting alone. In recent years, scholarship on network governance has burgeoned, along with significant advances in our ability to map and analyze relationships between multiple actors and organizations in a given policy system. While network analysis methods are invaluable for understanding relationships among actors, network analysts to date have paid only limited attention to the legislative, regulatory, and other policy settings in which network relationships occur, to the horizontal and vertical relationships between those policy settings, and to the way that laws, regulations, and procedural rules create governance structures that shape policy actors’ actions and interactions. In this paper, we develop and explore the use of network analysis methods to understand relationships between policy decision-making venues within a given policy subsystem. Using renewable energy policy in Connecticut as our policy subsystem, we develop a method for systematically exploring the networked relationships between multiple decision-making venues at multiple governance levels, including the governor’s strategic energy planning committee, the Connecticut legislature, the Public Utilities Regulatory Authority, the Department of Energy and Environmental Protection, and regional policy venues associated with the Regional Greenhouse Gas Initiative and the Independent System Operator-New England. To do so, we build on methods for analyzing “networks of adjacent action situations” developed by McGinnis (2011), Kimmich (2013), and Kimmich & Villamayaor-Tomas (2017). Following Kimmich & Villamayor-Tomas (2017), we identify a set of relevant policy decision-making venues (“action situations”) related to renewable energy deployment in Connecticut and disaggregate each of those venues into “working components” including starting conditions, actors, rules, information, and outputs and outcomes. Next, we identify dependencies between working components of action situations, as well as the nature and direction of these links. More specifically, in this paper we treat each “working component” of a policy venue as a potential node in a network of policy venues. We also analyze not only the presence but the nature of links between policy venues, distinguishing between direct and indirect links among working components and analyzing the direction of influence implied by these links. We then apply network analysis tools—including multilevel graph analysis methods drawn from the social-ecological system literature (e.g., Bodin et al. 2019) and Bayesian decision networks (e.g., Nielson and Jenson 2009) to identify system structural properties, understand how they jointly produce outcomes of interest – in this case, clean energy policy, and simulate system interventions. This paper extends the use of network analysis beyond understanding relationships between actors to include understanding and analyzing relationships between policy venues. Second, it advances the methods used to examine networks of adjacent action situations, and shows how this approach method can be used to systematically characterize policy subsystems and explore relationships between multiple policy venues.