Do Bold Prophecies Lead to Higher Profits?

Abstract: We find that non-herding ("bold") recommendations - defined as those deviating from last day's consensus by more than one level - had a greater investment value than herding recommendations prior to Regulation Fair Disclosure, which became effective in October, 2000. Indeed, a value-weighted portfolio that purchased S&P 100 companies with the most favourable consensus, made up of bold (all) recommendations, yielded a statistically significant (insignificant) annual abnormal gross return of 7.8% (3.8%), when rebalanced daily, over the period May 1994-Oct 2000. Unlike previous studies, we also find an economically, although not statistically, significant positive alpha net of transaction costs of 4.5% for the bold consensus. Surprisingly, the less sophisticated approach of restricting the consensus to upgrades and downgrades yielded not too distant results, with a statistically significant (insignificant) gross (net) alpha of 5.5% (1.8%). Our results raise the question whether the market was imperfect in the pre-regulation period, because investors could seemingly be rewarded for filtering public information, i.e. recommendations. All results remain robust to tests of investment delays, lower rebalancing frequencies, and business cycles.