The ‘learning assumption’ revisited: Bureaucrats, politicians, and experts in policy making

Abstract This article examines the conditions under which policy learning occurs, distinguishing between instrumental learning (evidence-based, problem-solving oriented) and political learning (strategic, shaped by institutional position and incentives). Drawing on three cases presented by Claudio Radaelli – examining bureaucrats, politicians, and experts – the paper shows that much of what occurs in policy practice is best understood as political learning: actors acquire and deploy knowledge strategically, filtered through professional hierarchies, electoral pressures, and reputational concerns. While the literature has long recognised these dynamics, the paper argues that existing accounts of political learning understate its relational and contextual character. Political learning emerges through positioned interactions within institutional and epistemic communities, rather than from individual strategic calculation alone, and from alignment with shifting background political conditions that determine what knowledge becomes actionable. Recognising this relational dimension has implications for institutional design aimed at fostering democratic accountability and integrating expertise more effectively into policy making.