We propose an approach for Bayesian inference in \TV SVARs identified with sign restrictions. The linchpin of our approach is a class of rotation-invariant \TV SVARs in which the prior and posterior densities of any sequence of structural parameters belonging to the class are invariant to...