To better understand the attributes of the earnings-related factors, EIU and ES, we examine the relation of these factors to some firm characteristics and macroeconomic variables. By examining the relation of these factors to firm characteristics, we ...
To better understand the attributes of the earnings-related factors, EIU and ES, we examine the relation of these factors to some firm characteristics and macroeconomic variables. By examining the relation of these factors to firm characteristics, we confirm that EIU is related to discount rate news and ES is related to cash flow news. We also find that these two factors have a significant relation to future activities of the macroeconomic variables and have different sensitivities to these macroeconomic variables across business cycles. In particular, ES has a strong positive relation with future growth of cash flow-related macroeconomic variables such as corporate net cash flow, net corporate dividends, inflation, and consumption, while EIU has no significant relation to these macroeconomic variables. Overall, EIU and ES contain preemptive forward-looking information about future activities of the macroeconomy.
Our three-factor model outperforms the models frequently mentioned in the literature in terms of pricing ability. To compare their pricing performance, we use various test portfolios sorted by firm characteristics that are related to such well-known anomalies as firm size, book-to-market momentum, post-earnings-announcement-drift, long-term reversal, and liquidity. Industry portfolios, which are somewhat free of data-snooping bias (Lo and MacKinlay, 1990), are also considered as test assets. We find that the factor related to earnings surprise is the main force in explaining anomalies related to price momentum and post-earnings-announcement drift as well as the industry effect, while the factor related to earnings information uncertainty is the main force in explaining anomalies related to size, liquidity, and long-term reversal. Either factor alone does not explain satisfactorily book-to-market. Book-to-market is explained by the collaboration of both these factors. The three-factor model also explains well the January effect, which is one of the most prominent intertemporal regularities in stock returns. The main force in explaining the January effect is a factor related to earnings information uncertainty. Based on these findings, we induce that price momentum and post-earnings-announcement drift are anomalies driven by cash flow news, while firm size, liquidity, long-term reversal, and the January effect are anomalies mainly driven by discount rate news. Book-to-market may be driven by both.
Even without the market factor, the two-factor model containing only the two earnings-related factors, strikingly, performs as well as the three-factor model in terms of statistical significance in explaining the anomalies related to the firm characteristics and the January effect. However, the two-factor model produces a slightly greater pricing error (measured by the Jensen alpha) than does the three-factor model. This indicates that the market factor simply plays a role in reducing the magnitude of the pricing error (economic significance); however, the inclusion of the market factor into the two-factor model does not significantly alleviate the statistical significance of the pricing errors. To examine whether the earnings-related factors are priced in stock returns, we also conduct cross-sectional regression (CSR) tests and find that the earnings-related factors are significantly priced in most cases, regardless of the inclusion of the beta variables of the other risk factors. Our findings on the excellent performance of the earnings-related factor model are compatible with Fama and French (1995) and Chordia and Shivakumar (2006) showing that common factors in returns are related to common factors in earnings.