Policy Coherence Within a Policy Subsystem: The Case of Norwegian Higher Education

Public sector policies are multi-faceted by nature. Individual policies are often a result of political compromises of different actors (e.g. national and local governments, state and non-state actors), concern multiple ambiguous policy goals, and consist of policy instruments with different underlying logics (Capano & Howlett, 2020; Howlett et al., 2015). Moreover, policy processes do not unfold in a vacuum. New policies ‘land’ in contexts marked by previous initiatives whose implementation may not have been completed, leading to tensions, reform fatigue and resistance (Wynen, Kleizen, et al., 2019; Wynen, Verhoest, et al., 2019). Finally, policies come in different shapes and sizes, ranging from comprehensive overarching reforms to minor adjustments (Hall, 1993). Therefore, at any given time, a particular policy subsystem (Sabatier & Weible, 2007) is likely to be subjected to multiple policies, which may not be coherent with each other, or internally (Howlett & Rayner, 2007). In this study, we focus on coherence (or lack thereof) of policy concerning higher education in Norway. We distinguish between two outcomes of interest: (1) symbolic and (2) substantive policy coherence. For symbolic coherence, we identify two sub-types: (1a) alignment between formulated policy goals and (1b) extent to which policies explicitly cross-reference each other. For substantive coherence, we analyse alignment between what kind of behaviour is being incentivized through (2a) policy instruments of the same type (e.g. funding) and (2b) policy instruments of different types (e.g. funding and regulation). The dual approach to conceptualizing policy (in)coherence is grounded in the understanding that different stages of the policy process – specifically policy formation and policy implementation – have different underlying logics. While symbolic policy incoherence may be beneficial for reaching consensus and moving things forwards during policy formation, substantive policy incoherence may be harmful for policy implementation. We treat the two types of policy (in)coherence as a set of ordered categories, ranging from significant incoherence to significant coherence. The main element in our empirical material are 23 white papers that the Norwegian government sent to the parliament in the 2000-2024 period. Through a combination of deductive and inductive coding, we identify stated policy goals and intended policy instruments in each white paper. These are then paired into dyads (253 in total), and each dyad is given four scores ranging from -1 to +1 for each subtype of coherence identified (1a, 1b, 2a, 2b, see above). With regards to conditions of interests, we focus on the role of policy learning (Dunlop & Radaelli, 2018), policy layering (Mahoney & Thelen, 2010), and stability of policy entrepreneurs on the bureaucratic and political side (Mintrom & Norman, 2009). Data related to conditions of interest are obtained from publicly available sources.