Growth, cross-sectional returns and risks in the equity market

This dissertation consists of two chapters that study the risk-return relationships in the equity market, and provides new empirical evidence for understanding the sources of cross-sectional stock return patterns.The first chapter explains cross-sectional stock return patterns using a consumption-based intertemporal asset-pricing model with two differently priced shocks in the consumption dynamics. Two key features in the model are predictable consumption growth and recursive utility. Empirically, the two shocks are identified through a VAR specification of the consumption growth and a set of information variables commonly used for forecasting business cycles. Under this specification, strong evidence is found for long-term consumption growth predictability. The model does a superb job in fitting cross-sectional return patterns. In particular, the model explains the size and value effects better than standard asset-pricing models including the Fama-French three-factor model, which suggests that the size and book-to-market effects are consistent with risks associated with long-run consumption growth opportunities.The second chapter studies another intriguing cross-sectional return puzzle---the momentum effect, which refers to the persistence of relative return performance. This chapter provides new empirical evidence on the relationships between firm fundamentals, momentum portfolio returns and risks. Using COMPUSTAT/CRSP merged database, quarterly portfolios are constructed based on various past-return and expected earning growth signals. Evidence shows that the momentum ranking coincides significantly with dynamics of firm fundamentals. However, expected earning growth effect does not fully subsume the momentum effect. In particular, it is found that the momentum effect stands out more strongly in the subgroup of stocks with high expected earning growths. I then go further to test whether the connection between momentum and expected earning growth rankings implies a risk-based explanation for the momentum effect. A time-varying beta model is tested where the beta for a momentum portfolio is a function of the portfolio's expected earning growth. The results support the time-varying beta hypothesis. The risk-based explanation for the momentum effect, however, remains tenuous.