Description
Abstract: To reconcile the disconnect between survey expectations of stock returns and rational expectations, researchers have hypothesized that survey participants may confound beliefs and preferences by (i) reporting risk-neutral forecasts of future returns; or (ii) reporting pessimistically-tilted forecasts reflecting ambiguity aversion or robustness concerns. We find that these hypotheses are strongly rejected by the data, albeit for different reasons: Inconsistent with hypothesis (i), survey return forecasts are reliably much higher than risk-free interest rates and survey expected excess returns are predictably time-varying. Inconsistent with (ii), agents are not always pessimistic about future returns, but often predictably optimistic and unconditionally unbiased.