This paper proposes a general way to conceive public policy when there is no consensual account of the situation of interest. The approach builds on an extension and dual formulation of the traditional theory of economic policy. It does not need a representative policymaker’s utility function (as in the literature on ambiguity), a reference model (as in robust control theory) or some prior probability distribution over the set of supplied scenarios (as in Bayesian model-averaging). The method requires instead that the willingness to accept a policy’s projected outcomes coincide with the willingness to pay to correct the current situation. Policies constructed in this manner are shown to be e ective, robust and simple in a precise and intuitive sense.