The growing political science literature on multi-level governance and wicked policy problems typically studies challenges related to democratic accountability and coordination. Without disputing the importance of these phenomena, I will argue in this proposed paper that the wicked issues literature should pay more attention to the information aspect of policymaking: knowledge production, knowledge dissemination and experience-based policy learning. And vice versa: The literature on learning in organizations and policy subsystems should pay more attention to learning in situations with wicked issues characteristics.
The limited dialogue between these literatures is surprising given that the need for information and learning must be assumed to extraordinarily high in awkwardly structured policy subsystems and issues with “wicked” characteristics: responsibility exceeding authority, absence of shared problem frames within the policy subsystem, underdeveloped means-ends understandings, incongruence between short and long-term goals, and problem severity that fluctuates over time.
In this paper I propose a framework for analyzing knowledge production and learning in loosely coupled policy subsystems dealing with wicked issues.
The paper draws on results from a comparison of three policy areas in Norway: refugee settlement, city planning and regional research and development. These cases share important governance characteristics, but the emphasis on market-type and network-type steering instruments vary. Importantly, there is also variation regarding the political salience of the issues and the presence of a vigilant public. These similarities and differences enable a discussion of various possibilities and limitations for learning in relation to wicked issues, notably the impact on learning from preexisting governance instruments and political salience. Does incentive-driven governance undermine the potential for shared policy learning within the policy subsystem? Does high and low political salience render learning irrelevant, leaving only medium-salient issues as relevant for systematic knowledge production and learning?
