The purpose of this paper is threefold: first, we warn analysts against the use of structural coeffcient estimates alone for deducing quantitative inferences concerning many policy experiments. Second, we offer a potential solution to the problem by proposing a mutatis mutandis approach to...