According to the seasonal random walk model, the expected earnings for the current quarter is the actual earnings from the same quarter in the previous year. However, the time-series of quarterly earnings are shown to follow a process that is more complicated than the seasonal random walk model indicates. Specifically, the seasonally differenced quarterly earnings are positively autocorrelated at the first three lags and negatively autocorrelated at the fourth lag Previous studies have shown that the stock market partially anchors its earnings expectations based on the seasonal random walk model, and underestimates the magnitudes of these autocorrelations by on average fifty percent for the first four lags. The stock market's underreaction to past earnings information is shown to be a contributing factor to the post-earnings-announcement drift phenomenon. Financial analysts have also been shown to underreact to various kinds of publicly available information. It is possible that the stock market's apparent inefficiency is due to the similar behavior by financial analysts This study investigates, first, whether financial analysts fully reflect information in past earnings in their quarterly earnings forecasts, and second, its implications for the post-earnings-announcement drift phenomenon. Specifically, I study whether analysts fully exploit the autocorrelations in seasonally differenced quarterly earnings. I find that the implied analysts' estimates for the lag one and lag two autocorrelations in seasonally differenced quarterly earnings are about 24% and 36%, respectively, of the corresponding time-series estimates, suggesting underreaction to past earnings information by financial analysts. The results are not driven by the existence of potentially stale forecasts or by firms with poor prior earnings performance. When relating analysts' underreaction to post-earnings-announcement drift, I find that the majority of the stock market's underreaction to past earnings information can be attributed to analysts' underreaction. Based on the above results, I construct a forecasting model to predict analysts' forecast error based on the past earnings information, and I show that profitable trading rules can be formed using the above earnings surprise forecasting model / acase@tulane.edu
| Identifer | oai:union.ndltd.org:TULANE/oai:http://digitallibrary.tulane.edu/:tulane_23953 |
| Date | January 1998 |
| Contributors | Wu, Shuang (Author), Jain, Prem C (Thesis advisor) |
| Publisher | Tulane University |
| Source Sets | Tulane University |
| Language | English |
| Detected Language | English |
| Rights | Access requires a license to the Dissertations and Theses (ProQuest) database., Copyright is in accordance with U.S. Copyright law |
Page generated in 0.0015 seconds