The newsworthiness of an event is partly determined by how unusual it is and this paper investigates the business cycle implications of this fact. We present a tractable model that features an information structure in which some types of signals are more likely to be observed after unusual events. Counterintuitively, more signals may then increase uncertainty. When embedded in a simple business cycle model, the proposed information structure can help us understand why we observe (i) large changes in macro economic aggregate variables without a correspondingly large change in underlying fundamentals (ii) persistent periods of high macroeconomic volatility and (iii) a positive correlation between absolute changes in macro variables and the cross-sectional dispersion of expectations as measured by survey data. These results are consequences of optimal updating by agents when the availability of some signals is positively correlated with tail-events. The model is estimated by likelihood based methods using raw survey data and a quarterly time series of total factor productivity along with standard aggregate time series. The estimated model suggests that there have been episodes in recent US history when the impact on output of innovations to productivity of a given magnitude were up to twice as large compared to normal times.