We propose and test a simple hedging hypothesis for prediction interval formation in stock price forecasting. In the presence of uncertainty, forecasters hedge their forecasts by adjusting the bounds of the prediction interval in a way that reflects their forecast of the average forecast of others. This hypothesis suggests a positive relationship between the belief wedge, defined as the difference between the subject's forecast of the average forecast of others and the subject's own point forecast, and the asymmetry of the prediction interval. Empirical support for the hedging hypothesis is drawn from two in‐class surveys, an experiment, and a large survey of professional analysts' forecasts of future stock prices.