| Summary: | The central concern in this thesis is to investigate the incremental information content of earnings and non-earnings accounting data, and their ability to predict stock returns. It investigates the information content of discretionary signals (i.e. dividends, issuing new equities, capital loans, capital expenditures, and extraordinary and exceptional items), and non-discretionary signals ( i.e. sales, earnings, cash flow from operations, and fund flow from operations), and their incremental information content relative to earnings, against both the level of earnings and the change in earnings. Further, the thesis extends the evidence on market efficiency in the UK by addressing the value of financial statement information in predicting future earnings and future abnormal returns. Apparent findings on informational efficiency might reflect inadequate risk adjustment by use of the CAPM. Therefore, this thesis investigates whether return patterns are better explained by incorporating security beta to define abnormal return or whether the same amount of security return variance is explained by computed market return (equivalent to assuming J3 =1, for all securities). Finally, the thesis re-examines the level versus the change specification of the returns-earnings relation in the UK. Using an event study methodology we examine whether share betas provide any explanation of security return patterns over and above that provided by general market movement. We find no evidence that security beta as a measure of risk provides any explanation over and above that provided by market movement. Also, we find that beta appeared to capture some of the specific risk of securities as a results by altering the required return for securities then created abnormal returns. The examination of the ability of financial statement information to predict future earnings suggests that the most important ratios in assessing firm's profitability future returns are pre-tax profit, percentage change in long-term debt, return on total assets, operating profit to total assets, times interest earned, and percentage change in capital expenditure to total assets lagged one year. The results suggest that publicly available current financial statement information about future earnings may be useful in predicting abnormal returns if alternative procedures (i.e., hedge portfolio) based on non-conventional techniques such as a two stage discriminant analysis procedure are used. This will provide a way of testing market efficiency which is free of the risk adjustment procedure and therefore avoid the ambiguities of the joint hypothesis test. The results of the incremental information content of discretionary and non-discretionary signals over earnings, using the specification of the earnings change and earnings level suggest that both the non-discretionary and discretionary signals have information content over earnings, in particular the change in sales, the change in dividends, the change in equities, the change in capital expenditures, and the change in extraordinary and exceptional items. In addition, the explanatory power (adjusted R2) of the models which include the level of earnings alongside the non-discretionary and discretionary signals is higher than those which include only discretionary or non-discretionary signals. The results also suggest that the level of earnings is more important than the change in earnings in explaining the variation of abnormal returns.
|
|---|
