We estimate Bayesian VAR models in order to investigate the relation between Treasury yields and the corporate bond yield spread in Australia. Recent developments in Bayesian model selection allow us to formally assess the relevance of stochastic volatility and drifting parameters. A model comparison indicates that a model with stochastic volatility and constant parameters is preferred. Our results imply that while previous studies may have relied on empirically flawed models, their main conclusion – namely that an increase in the risk free rate decreases the corporate bond yield spread – appears to be an empirically robust finding.