A study into the relationship between accounting information and share prices
This research reports on an investigation into the utility of published accounting information to the most sophisticated user the investor. The approach adopted is to paramorphically represent the investor's decision making process by seeking to establish a set of causal relationships between accounting number based inputs to this process, and the outputs, viz relative share prices. Unlike previous studies in this area, this research explicitly recognises the configural nature of human information processing activity. Application of an appropriate methodology to uncover and explore these configural relationships, Automatic Interaction Detector (AID), offers original insights into the share price fixing mechanism. More specifically the results obtained provide evidence on: 1) the degree of association between published accounting data and share prices and thus a measure of the value of accounting information to investors, 2) the complexity of the investor's decision making process and 3) the validity of certain established theories in finance. Whilst analysis highlights the strong relationship between historic accounting information and relative share values the evidence presented is consistent with the thesis that only a limited set of measureable accounts-based variables may be used in the assessment process, viz earnings, dividends, short-term liquidity and marketability. Moreover, the complex set of interactions identified between these variables confirms a priori expectations on the configural nature of the investor's decision making process. A close examination of these interactions reveals that although earnings and dividends may dominate relative share values, the extent of their influence varies with the underlying quality of earnings. These findings have implications for the theory of share valuation, the dividends versus earnings controversy and the role of investment risk in the U.K. stock market. Empirical evidence relating to to the capital structure debate is also provided. Other areas of investigation encompass an examination of the most appropriate measures of relative share valuation for this type of research, a comparison of the merits of linear additive and configural analytical techniques, and the dimensionality and normality of financial ratios. The operational utility of the 'decision-usefulness' criterion in the evaluation of accounting numbers and the use of pragmatic empiricism in the application of this criterion are also critically appraised.