In this paper, we examine empirically whether credit plays a role as a nonlinear propagator of shocks. This propagation takes the form of a threshold vector autoregression in which a regime change occurs if credit conditions cross a critical threshold. Using nonlinear impulseresponse functions, we evaluate the dynamics implied by the threshold model. These suggest that shocks have a larger effect on output in the ''tight'' credit regime than is normally the case, and that contractionary monetary shocks typically have a larger effect than expansionary shocks. Finally, using a nonlinear version of historical decompositions, we attempt to determine the relative contribution to output growth of shocks and the nonlinear structure.