Are good-news firms riskier than bad-news firms? Byoung-Kyu Min a , Tong Suk Kim b,⇑ a Institute of Financial Analysis, University of Neuchatel, Neuchatel, Switzerland b Graduate School of Finance, Korea Advanced Institute of Science and Technology (KAIST), Seoul, Republic of Korea article info Article history: Received 5 April 2011 Accepted 28 December 2011 Available online 11 January 2012 JEL classifications: G12 G14 Keywords: Post-earnings-announcement drift Stochastic discount factor Market efficiency abstract This paper examines the relative risk of good-news firms, i.e., those with high standardized unexpected earnings (SUE), and bad-news (low SUE) firms using a stochastic discount factor approach. We find that a stochastic discount factor constructed from a set of basis assets helps explain post-earnings-announce- ment drift (PEAD). The risk exposures on the pricing kernel increase monotonically from the lowest to highest SUE sorted portfolios. Specifically, good-news firms always have higher risk exposures than bad-news firms in both 10 SUE sorted portfolios and 25 size and SUE sorted portfolios. However, the esti- mated expected risk premium is too small to explain the observed magnitude of returns on the PEAD strategy. Our risk adjustment can explain only about one-fourth of the total magnitude of the average realized return to the PEAD strategy. As a result, the average risk-adjusted returns of earnings momentum strategies are mostly positive and significant. Overall, our results support the view that at least some por- tion of the returns to the earnings momentum strategies examined represent compensation for bearing increased risk. Ó 2012 Elsevier B.V. All rights reserved. 1. Introduction Post-earnings-announcement drift (PEAD), or earnings momen- tum, refers to the fact that stock returns continue to drift in the direction of earnings surprises for several months after earnings are announced. It is well documented in the literature that a simple trading strategy that is long in stocks with the highest earnings surprises and short in stocks with the lowest earnings surprises generates significantly positive abnormal returns (e.g., Chordia and Shivakumar, 2005). PEAD is still robust after its initial discov- ery by Ball and Brown (1968). Subsequent studies have shown the robustness of earnings momentum using different samples and methods or using evidence on an international scale. 1 While PEAD has been robust over the four decades, interpreta- tions of this phenomenon are less clear. In a (semi-strong form) efficient market, all public information released into the market should be instantly reflected in stock prices, because investors immediately adjust their expectations about future earnings. Thus, returns on trading strategies that use only public information should display no abnormal patterns. Evidence of such abnormal returns on earnings momentum may therefore be interpreted as going against the efficient market hypothesis. For instance, Bernard and Thomas (1990) suggest that earnings momentum is a manifes- tation of a delayed response to the information in earnings announcements. Tests of market efficiency, however, are always necessarily joint tests of market efficiency and the asset pricing model used to determine the expected return (Fama, 1970; Roll, 1977). All studies that show earnings momentum rely on specific pricing models. For example, Bernard and Thomas (1990) adopt the five factor model of Chen et al. (1986). Chordia and Shivakumar (2005) use a factor related to news about future inflation. However, if the pricing models used in such studies are mis-specified, the abnormal returns inferred from their use are incorrect as well. That is, such studies may suffer from a ‘‘bad model’’ problem, as Fama (1998) points out. In this paper, we revisit the relative risk of good-news firms, i.e., those with high standardized unexpected earnings (SUE), and bad- news (low SUE) firms. Rather than choosing a particular pricing model, we take an alternative approach using a stochastic discount factor, first proposed by Chen and Knez (1996). 2 This approach ini- tially extracts the stochastic discount factor, or pricing kernel, from a set of basis assets, then uses these to price other assets. In contrast to 0378-4266/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.jbankfin.2011.12.017 ⇑ Corresponding author. Tel.: +82 2 958 3018. E-mail addresses: byoungkyu.min@unine.ch (B.-K. Min), tskim@business. kaist.ac.kr (T.S. Kim). 1 Foster et al. (1984), Bernard and Thomas (1990) and Livnat and Mendenhall (2006) confirm the robustness of Ball and Brown (1968) using recent samples and differing methods. Hew et al. (1996) and Booth et al. (1996) provide an international evidence for earnings momentum. 2 This approach has been used in various contexts in the literature. Ahn et al. (2003a) apply this approach to examine the profitability of momentum strategies. Ahn et al. (2003b) study the long-term returns of seasoned equity issues using this approach. Journal of Banking & Finance 36 (2012) 1528–1535 Contents lists available at SciVerse ScienceDirect Journal of Banking & Finance journal homepage: www.elsevier.com/locate/jbf