This paper asks two questions. First, can we detect empirically whether the shocks recovered from the estimates of a structural vector autoregression are truly structural? Second, can the problem of non-fundamentalness be solved by considering additional information? The answer to the first...