Roll (1988) observes low R2 statistics for common asset pricing models due to vigorousfirms-specific returns variation not associated with public information. He concludes (p. 56) that this implies acirc;not;Seither private information or else occasional frenzy unrelated to concrete...