Biased Surveys

We find evidence suggesting that surveys of professional forecasters are biased by strategic incentives. First, we find that individual forecasts overreact to idiosyncratic information but underreact to common information. Second, we show that this bias is not present in forecast data that is not subject to strategic incentives. We show that our evidence is consistent with a model of strategic diversification incentives in forecast reporting. Our results caution against the use of surveys of forecasts as a direct measure of expectations, as this would overestimate the likely deviations from rational expectations, the information precision, and the degree of disagreement. Lastly, we show how to correct for this.