Fundamental Analysis and the Modelling of Normal Returns in the UK

Fundamental analysis involves the use of accounting data to predict future stock returns, future stock prices, or both. One aspect of such research is whether accounting information can be used to generate profitable portfolio strategies. Central to this style of fundamental analysis, is the measurement of the benchmark of ‘normal' returns against which ‘abnormal' returns can be identified. One approach is to regress a time series of portfolio excess returns on variables thought to capture risk. The Fama-French three factor model has been used to capture risk in both the UK and the US and we evaluate its effectiveness for this role using UK data. Numerous versions of the model are found in the UK literature and we test nine such versions. None of them can be awarded a clean bill of health in terms of standard asset pricing tests. Therefore, we conclude that fundamental analysis studies using UK data must recognise that, when using the three factor model, different ways of applying the model can result in quite different estimates of abnormal returns. Investigation of the robustness of results to different applications of the model then seems desirable. By implication, it is also desirable that fundamental analysis researchers should investigate the robustness of their results to other available methods of estimating normal returns