Essays on value, momentum and the preference for skewness

Abstract

This dissertation attempts to explain the cross-section of expected returns based on the valuation equation of Miller and Modigliani (1961) and the cumulative prospect theory of Tversky and Kahneman (1992). It consists of three essays on empirical asset pricing. The first essay proposes an alternative metric for value investing: the dividend-to-market ratio (DM). The book-to-market ratio (BM) which is currently used in academia and industry, is a noisy measure for value investing, because book value is a weak indicator of intrinsic value. Motivated by the valuation equation of Miller and Modigliani (1961), this paper suggests that DM is much more efficient in identifying undervalued stocks than BM, due to the strong link between expected dividends and intrinsic value. DM also provides a better estimation of expected stock returns compared to the linear combination of BM, profitability and investment used in the five-factor model of Fama and French (2015a), because it allows for non-linearities between expected returns and these variables. Results of cross-sectional regressions at the firm level and time-series regressions at the portfolio level consistently show that DM has a far stronger link with expected returns than BM, and it also outperforms a linear combination of BM, profitability and investment. The second essay examines the prediction of cumulative prospect theory whereby investors prefer lottery-like (or positively skewed) payoffs, resulting in overpricing and low expected returns to such assets. Given that earnings surprises are associated with lottery-like payoffs, investors should be willing to pay more for stocks with a high probability of generating positive earnings surprises. Empirical tests in this study consistently suggest that there is a strong negative correlation between the predicted profitability shocks (PPS) and expected stock returns. This essay contributes to the literature in asset pricing by revealing the link between skewness preference and prominent anomalies such as BM, profitability and price momentum. The explanatory power of BM and operating profitability disappears after controlling for PPS, which indicates that both could be noisy proxies for PPS in predicting average returns. Further, the price momentum effect cannot be driven out by earnings momentum once PPS is taken into account, which demonstrates that price momentum has incremental explanatory power for stock returns over that provided by earnings momentum. The third essay (co-authored with Takeshi Yamada and Tariq Haque) examines if crash-risk is systematically priced in momentum portfolio returns. A recent paper by Daniel and Moskowitz (2016) documents that crashes occur in the momentum strategy, and investors may take many years to recover from the resulting losses. Thus, this essay asks, if crash risk exists in momentum portfolios, is such risk priced in the market? To this end, we develop a measure, tail coskewness, that focuses exclusively on how tail events (low-probability events leading to large gains or losses) contribute to the systematic skewness of momentum portfolios. The results show that tail coskewness not only subsumes the risk premium associated with coskewness which may be a determinant of cross-sectional returns as shown by Harvey and Siddique (2000), but also that associated with firm size. The paper uses US data from 1927 as well as international data, where robust results are found across different time-periods and markets.