Complex coalitions: Analysing climate change policy subsystems as multilayer networks

Policy actors who work alone are rarely successful in the policy process. Therefore, actors join forces in advocacy coalitions to pool critical resources and improve their chances of success in transforming their policy beliefs into public policy. Many policy processes are structured around competing coalitions. This often creates an equilibrium that works against major policy change. Understanding the role of competing coalitions in climate change policy subsystems is, thus, crucial for understanding how policy changes that are required to curb climate change might take place. Despite more than three decades of scholarship, conceptualising and measuring coalitions has, nonetheless, proved challenging. Scholars have traditionally relied on surveys to measure advocacy coalitions. However, surveys increasingly suffer from declining response rates, social desirability bias, and the difficulty of drawing network boundaries beforehand. It is also difficult to repeat surveys, which leads to an excessive reliance on cross-sectional analyses. Recently, scholars have resorted to more accessible secondary data sources that lend themselves readily to longitudinal analyses of belief similarity, including public statements in legacy media and public behaviour in social media. An issue is that media data does not include information about coordination of action, which is a decisive feature of advocacy coalitions. In addition, we argue that the different data generation methods capture different strategic considerations by the actors. In this paper, we present an approach to overcome some of these challenges by bundling data on Finnish climate politics from one survey, two prestige newspapers, and Twitter. More specifically, we depict subsystem politics as a multilayer network with two dimensions: different types of relationships and different points in time. From this network, we infer coalitions through multilayer community detection. An advantage of this approach is that it accounts for dependencies not only within, but also between layers. Another advantage is that it allows for examining how consistently actors operate across the different layers of subsystem politics. We find that in the case of Finnish climate politics, there are considerable differences in actors’ participation in the different layers. This suggests that different layers serve different strategic goals. What we also find is that although actors generally stay within the broader coalition lines in every layer (i.e., economy versus ecology), legacy media and social media in particular allow for the identification of more issue-specific and time-bound factions that sometimes cross those lines. We also note that Twitter reveals clearer divides between actors than prestige newspapers. Implications for the study of advocacy coalitions and subsystem politics more generally are derived.