A FUNDAMENTAL VARIABLE-BASED PREDICTIVE MODEL : EXPLORATORY INVESTIGATION ON SELECTED STOCK MARKETS

Abstract

The aims of this dissertation are: 1. to examine the impact of some fundamental variables on the stock prices in Singapore and other selected share markets. 2. to build a general forecasting model for stock prices based upon an index that incorporates the differential yields of cash and equity in the markets selected. No finance theory specifies exactly what fundamental variables affect stock prices. This study attempts to discover what these variables are. A few key variables are selected by reviewing the literature on asset pricing anomalies since it is not possible to get data on many variables. Regression analysis is used to examine this relationship. The results show that there is a significant relationship between stock prices and some fundamental variables such as earnings yield, dividend yields, interest rates and money supply. The explanatory power of the model is moderate: R2 range from 50% to 75% and model fit is satisfactory. A predictive model for stock prices is developed using selected fundamental variables. The model is tested using data from stock markets of Germany, Japan, Singapore, United Kingdom and the United States over the period January 1978 to December 1992. The performance of the indicator in each market is evaluated. The accuracy of the buy/sell signals given by the indicator is tested by a dummy variable procedure. It is found that the model works with varying success for all the stock markets and best in the Singapore stock market: the dummy variables were highly significant. Further refinements of this promising model are in the direction of predictive accuracy of the model. More work is suggested .