In this paper, we analyze how skewness and heavy tails affect the estimated relationship between the real economy and the corporate bond-yield spread-a popular predictor of real activity. We use quarterly US data to estimate Bayesian VAR models with stochastic volatility and various distributional assumptions regarding the innovations. In-sample, we find that-after controlling for stochastic volatility-innovations in GDP growth can be well described by a Gaussian distribution. In contrast, the yield spread appears to benefit from being modeled using non-Gaussian innovations. When it comes to real-time forecasting performance, we find that the yield spread is a relevant predictor of GDP growth at the one-quarter horizon. Having controlled for stochastic volatility, gains in terms of forecasting performance from flexibly modeling the innovations appear to be limited and are mostly found for the yield spread.