In this paper, we develop a framework to analyze the relationship between evidence and policy. Postulating a normative criterion based on cost–benefit analysis and the value of a piece of information, as well as a topology of the policy space defined by three characteristics (epistemic uncertainty, interests, and the degree of value conflicts), we identify the (Nash) equilibria of an interaction between experts and citizens in providing information to a decision maker. In this setup, we study three institutional arrangements (evidence-based policy, deliberative governance, and negotiated conflict) that differ in terms of reliance on experts and citizens for providing information. We show that different degrees of uncertainty, interests, and value-relevance surrounding the issue at stake result in vastly different arrangement performances; hence, to foster efficiency, rules should be contingent.