| Summary: | The aim of this dissertation is to test whether analysts can efficiently use the publicly available information to make earning forecasts rational or not. Rational expectation tests are based on United Kingdom stock market data. It indicates the performance of the analyst's earning forecast during 31st, Jan, 1987 to 30th, June, 2003.This paper firstly introduce the fundamental theories for this study, such as Rational Expectation Hypothesis and Loss functions. Quadratic Loss function (OLS) and Linear Loss function (LAD) are chosen to evaluate the performance of analysts' earnings forecasts. The sample in this study includes 541 companies which are listed in the U.K stock market.
The test result is: Through analyzing 541 U.K listed companies' monthly-data on actual earnings, earnings forecasts, share prices and return index for 17 years (1987-2003) by using OLS and LAD estimations, I conclude that financial analysts do not use previous earnings, extreme earnings changes, previous forecast revisions, previous forecast errors and past stock returns information efficiently. This study reexamine previous researches by using U.K data.
|
|---|