Reputational Information Cycles
Reputational incentives can narrow disagreement among professional forecasters and make their consensus less accurate. In our model, skilled forecasters extract more information from a signal that contains a shared error. When evaluated relative to their peers, forecasters overweight that signal to demonstrate skill, amplifying the shared error in the consensus. In Bloomberg’s U.S. inflation survey, forecast errors carry larger ranking penalties and forecasts cluster more tightly when the inflation risk premium is high. We interpret high-premium periods as times when reputational stakes are greater. As those stakes rise and fall, the distortion in what forecasters report changes, generating cycles in how much information the consensus conveys.