Testing the Predictive Ability of the Dividend Valuation Model on Ordinary Shares

Abstract

The valuation of ordinary shares is much more difficult than that of preferred shares and bonds because of more uncertainty and instability surrounding the income from ordinary shares. The satisfying aspect of investing in bonds is the certainty of income. Bonds have a superior claim on assets and income in the event of liquidation, and as long as the interest and principal on a bond are adequately secured and the yield satisfactory, the decision to invest is not difficult. Greater risk is associated with preferred share investment because dividends are less certain and do not represent a fixed commitment of the company. However,. yields can easily be determined and whether the dividends and principal are secure can be readily established. The valuation of ordinary shares is more complicated and this has led to the development of a number of models, including the dividend valuation model and earning per share valuation model. This study sought to determine the predictive ability of the Dividend Valuation Model on the ordinary shares. Data collected in form of share prices, market indices and dividend per share from the Nairobi Stock Exchange (NSE) secretariat were used to predict share prices for each of the thirteen companies studied. The market model was used to provide a link between the expected values which are non-observable and the real values that were used in testing the model. The predicted share prices were compared with actual prices by computing the difference between them. The differences between the two prices were subjected to t-tests. The tests of significance showed t hat; out of the thirteen companies studied, only three showed that the differences were not significant. We therefore concluded that the dividend valuation model was a poor predictor of share prices in the NSE