The global Great Recession has sparked renewed interest in the relationships between financial conditions and real activity. This paper considers the Swiss experience, studying the impact of credit market conditions and housing prices on real activity over the last three decades through the lens of a medium-scale structural Bayesian vector autoregressive model (BVAR). From a methodological point of view, the analysis is challenging for two reasons. First, we must cope with a large number of variables which leads to a high-dimensional parameter space in our model. Second, the identification of economically interpretable shocks is complicated by the interaction among many different relevant factors. As to the first challenge, we use Bayesian shrinkage techniques to make the estimation of a large number of parameters tractable. Specifically, we combine a Minnesota prior with information from training observations to form an informative prior for our parameter space. The second challenge, the identification of shocks, is overcome by combining zero and sign restrictions to narrow the plausible range of responses of observed variables to the shocks. Our empirical analysis indicates that while credit demand and, in particular, credit supply shocks explain a large fraction of housing price and credit fluctuation, they have a limited impact on real activity.