SMB, HML, and the empirical success of the Fama-French (1993) model
Fama and French (1993) turned the size and book-to-market anomalies into factors that explain the cross-section of equity returns. Their three-factor model includes the return on the market portfolio, SMB, and HML. SMB is the size-related factor, whereas HML is the book-to-market-related factor. Arguably, HML has greater ability to explain the cross-section than SMB.
My paper "Can size, book-to-market, and momentum be risk factors that predict future GDP growth?" (co-authored with Jimmy Liew) I provide the first evidence that SMB and HML can predict future GDP growth in the US and several other developed markets. This paper is the first to suggest that HML and SMB may contain business-cycle related information.
In my paper "News about future GDP growth as a risk factor in equity returns" I further explore the above idea. In particular, I show that one can replicate the performance of the Fama-French (1993) model by considering a model that includes the return on the market factor, in addition to a mimicking portfolio that captures news about future GDP growth. In the presence of this mimicking portfolio, SMB and HML lose their ability to explain the cross-section.
GDP is an aggregate variable. At a first-level, it can be decomposed into investment and consumption. In my paper "Sector investment growth rates and the cross-section of equity returns" (co-authored with Qing Li, and Yuhang Xing), we show that it is the investment side of GDP that is most important for explaining equity returns. A sector-investment growth asset pricing specification greatly outperforms the Fama-French model in explaining the cross-section of equity returns. It also completely eliminates the ability of SMB and HML to explain the cross-section.
