Anchoring Bias in Consensus Forecasts and Its Effect on Market Prices

Previous empirical studies on the “rationality” of economic and financial forecasts generally test for generic properties such as bias or autocorrelated errors but provide only limited insight into the behavior behind inefficient forecasts. This paper tests for a specific form of forecast bias. In particular, we examine whether expert consensus forecasts of monthly economic releases are systematically biased toward the value of previous months’ releases. Such a bias would be consistent with the anchoring and adjustment heuristic described by Tversky and Kahneman (1974) or could arise from professional forecasters’ strategic incentives. We find broad-based and significant evidence for this form of bias, which in some cases results in sizable predictable forecast errors. To investigate whether market participants’ expectations are influenced by this bias, we examine interest rate reactions to economic news. We find that bond yields react only to the residual, or unpredictable, component of the forecast error and not to the component induced by anchoring, suggesting that expectations of market participants anticipate this bias embedded in expert forecasts.