Asset pricing with data revisions

Abstract

We document two important asset pricing implications of the data release process of US consumption growth. First, initial releases are more suitable for asset pricing than final revised releases. This is because most revisions smooth out essential short-term consumption growth fluctuations. Second, first revisions incorporate novel information and their magnitude is strongly linked to consumption growth ambiguity. We formulate a novel consumption-based model, the Revised CCAPM, which incorporates these two effects using vintage data. It explains a striking 75% of the cross-sectional variation in average returns on 25 size-value portfolios. These results support the concept of state-dependent ambiguity attitudes.