Three essays on empirical asset pricing

• Main Author: Cai, Yu
• Other Authors: Lau Sie Ting
• Format: Theses and Dissertations
• Language: English
• Published: 2012
• Subjects: DRNTU::Business::Finance::Assets
• Online Access: https://hdl.handle.net/10356/50618
• Institution: Nanyang Technological University

The dissertation comprises three essays on empirical asset pricing. The essays connect to one another as the corporate earnings announcement event is the key to constructing the main variables in the three studies. Essay 1 examines relative profitability of stock recommendations associated with analyst pessimism. Analysts are by definition pessimistic if their prior earnings forecasts tend to fall below consensus. We find that recommendations or revisions are more profitable when pessimistic rather than optimistic analysts issue them in the U.S. from 1994 to 2009. The return differential is around 4.20 percent per year for the stock recommendation, and 7.08 percent for the recommendation revision (to buy or strong buy), between the top and bottom pessimism tercile of analysts. Our finding does not result from the conditional pessimism measure of Hugon and Muslu (2010) or another ex-ante measure of earnings forecast accuracy. Essay 2 studies stock estimation risk and its implication on the cross section of stock returns. When prior information is of low quality, price change is often large on the earnings announcement date. We use the standard deviation of historical earnings announcement return as a proxy for stock estimation risk. We find that there is no significant premium to estimation risk in the U.S. stock market from 1978 to 2009, a result at odd with the theoretical prediction of Kumar, Sorescu, Boehmer and Danielsen (2008). In addition, we find that price and earnings momentum as well as forecast-revision drift is stronger among stocks with higher estimation risk, a result consistent with the notion that investors suffer from bias in processing information. Essay 3 follows on the literature’s extensive use of earnings announcement effects to show informed trading of financial institutions. In this study, we use a measure of investment skills or informational superiority for mutual funds based on their buy trades. If a mutual fund buys into stocks that subsequently enjoy high earnings announcement returns, the fund is either skilled or informed according to this measure. From 1984 to 2009, this measure successfully predicts fund winners. The top performing decile of mutual funds beat the bottom by around 50 basis points per quarter. The measure shows persistence over time. We infer that the inherent investment talent or informational advantage of fund managers contributes to alpha.