This study extends the existing research on the use of financial statement variables to predict one-year-ahead earnings changes. A principal component analysis was conducted on 61 financial statement variables in an attempt to describe the dimensionality of the variables and facilitate the development of parsimonious earnings prediction models. This study finds that the 61 variables embody a much richer array of information than suggested by previous research. The variables could not be described by a small number of principal components. / The effect on the predictive ability of different earnings prediction model specifications was assessed by examining 36 different models. It was found that (1) models using a dichotomous earnings change variable as the dependent variable performed as well as models using a trichotomous earnings change variable; (2) models with a one-year drift term achieved greater predictive ability than similar models using a four-year drift term; (3) models with the strongest fit in the estimation period did not necessarily dominate in the predictive ability tests; and (4) the accuracy of the predictions of many of the models in this study was greater than the results obtained in previous studies. / This study also provides additional evidence on the extent to which the information regarding one-year-ahead earnings contained in current financial statements is fully reflected in stock prices. It was found that a simulated trading strategy did not perform well in the period subsequent to 1983. Evidence is also provided that the trading strategy generates abnormal returns in periods extending beyond 36 months. This provides further support that the probabilistic measure of one-year-ahead earnings changes is proxying for differences in expected returns rather than exploiting the underutilized information contained in financial statements. / Lastly, three stratifications of the sample firms did not increase the performance of the trading strategy on a consistent basis. Although one stratification did increase the effectiveness of the strategy, it is likely it did so by further sorting firms according to determinants of expected returns. / Source: Dissertation Abstracts International, Volume: 56-10, Section: A, page: 4032. / Major Professor: Kenneth S. Lorek. / Thesis (Ph.D.)--The Florida State University, 1995.
Identifer | oai:union.ndltd.org:fsu.edu/oai:fsu.digital.flvc.org:fsu_77573 |
Contributors | Ortega, William Roland., Florida State University |
Source Sets | Florida State University |
Language | English |
Detected Language | English |
Type | Text |
Format | 301 p. |
Rights | On campus use only. |
Relation | Dissertation Abstracts International |
Page generated in 0.0021 seconds