| Summary: | The main objective of this dissertation is to test rational expectation hypothesis on
earnings forecast in United Kingdom market. Ordinary Least Squares (OLS) and Least
Absolute Deviation (LAD) are introduced and presented. The study examines the
efficiency of the analysts earnings forecast by testing efficiency regarding information
variables including past earnings levels and changes, extreme past earnings changes, past
returns, past forecast errors and past forecast revisions. Actual and forecasted earning per
share (EPS) of 550 selected companies in UK market over the period from 1988 to 2003 is
obtained. Three statistical regression and test applied to test the rational expectation
include pooled regression and Fama-MacBeth regression and Wald tests which have
shown a consistent result of rejecting the efficiency of analysts forecast in the UK market.
We also conduct and compare rational expectation tests using OLS and LAD. Least
Absolute Deviation is taken into consideration of heteroscedasticity and non-normal
distribution, so consistent with previous study we conclude that LAD is more
advantageous than OLS regarding to expectation hypothesis tests.
|
|---|