This paper examines whether the expected equity premium constructed from survey forecasts is consistent with the implications of the rational belief approach or the distorted belief approach to explaining the equity premium puzzle. To address this question, a panel data model with a fixed individual effect and a business cycle effect is analyzed. The results appear more favorable to the distorted belief approach. The average expected equity premium is lower than the average realized equity premium during the sample period. The average bias across economists is significant and varies over the business cycle, which is consistent with distorted beliefs that are excessively pessimistic over expansions and excessively optimistic over contractions.