Forecasting Stock Prices Using Reported Earnings and Past Prices

Cesar G. Saldana, Jose S. Victoria

Abstract


This study explores the use of corporate income and past prices as bases for predicting stock prices, leading to well-informed investment decisions.  The regression model is used to show a significant relationship between reported earnings and the stock price (using San Miguel Corporation as example).  This result provides support for the use of Price-Earnings multiples by practitioners in the Philippine stock market in estimating the long-term or intrinsic value of a security.

 Likewise, the use of time series modelling as a prediction tool for SMC stock price is analysed.  The study intends to find out whether the price of San Miguel shares would have a long term influence on its present price.  Upon examination of a few autoregressive models fitted on weekly SMC price data, only the auto regressive model of order one is found to be a valid predictor.  This means that only the most immediate past value of share price is useful in predicting the present share price.


Full Text:

PDF