This paper extends the Bayesian proxy SVAR model (BP-SVAR) of Caldara and Herbst (2019) to examine changes in the transmission of structural shocks in the presence of regime shifts in an economy. I provide a Metropolis-within-Gibbs sampling algorithm to approximate the posterior distribution of model parameters. The model is then used to examine the role of credit spreads on the transmission of monetary policy shocks in the United States between 1994-2007, where identification is achieved using a proxy constructed from high-frequency financial data. The main finding is that the effect of credit spreads differs across regime. Credit spreads significantly change the transmission of monetary policy shocks from 2000-2007 supporting Caldara and Herbst (2019), although, their inclusion appears to only alter the response of industrial production in the short-term with no other significant changes to the rest of the economy during the mid to late 1990s. This result highlights the empirical relevance of accounting for regime changes when assessing the impact of economic shocks.