Abstract
Our paper proposes an original angle to study the free-rider problem in the provision of public goods when the regulator has no information about agents' preferences. For a given outcome - specifically a Lindahl allocation - we ask what assumptions have to be imposed on simple mechanisms (in a precisely defined sense) that have the ability to Nash-implement it. Our answer lies in two main results: i) transfers necessarily belongs to a class of mechanisms that are linear in individual contributions to the public good, ii) there exists a subset of this class that fully implement Lindahl allocations. This subset encompasses, but does not reduce to, Walker (1981).