This paper uses a two-step approach to characterize the evolution of US macroeconomic and financial variables during episodes of very high uncertainty. First, we identify episodes of very high uncertainty using a regime-switching model. Second, we assess the behavior of macroeconomic and financial variables during these episodes of very high uncertainty. This methodology is analogous to the approach followed by Baele et al. (2013), who study episodes of fights to safety in financial markets. We find that very high uncertainty episodes are associated with a weaker growth performance and sharp declines in stock prices. However, we find that this relation is non-linear in that uncertainty does not seem to matter during periods characterized by medium or low uncertainty.