Prediction of share price in the Nairobi Securities Exchange Marketusing time series analysis (Box-Jenkins Methodology)

Abstract

Statistical models have for a long time been infused in numerous organizational operations ranging from agriculture manufacturing, leT tourism, economic and finance among some of the key sectors of a country. Under finance, stock trading is without a doubt, one of the cornerstones, of nearly any investment portfolio hence our interest in this indispensable financial tool. Investors in the Nairobi Securities Exchange SE) bourse will significantly cut down on time and money lost due to wild speculation on share performance thus ensuring near accurate prediction of the share price index of any stock portfolio. Secondary data was used, which was mainly obtained from published financial statements of nine listed companies and from the NSE for a period of five years (2008-2010). Times Series Analysis was then used to derive aARIMA (p, d, q) models, which were then fitted to predict trend and hence the share price i at a given time t. Data obtained for the study was analyzed based on Box-Jenkins Methodology, in which the ACFs and PACFs plots of the respective series were plotted and analyzed before testing the proposed models to ensure they were sufficient and of good fit thus ensuring accuracy and reliability in prediction. The paper was then concluded by discussing the general model obtained in the study, which was a ARIMA(0,1,0) modeland areas for further research and recommendations were subsequently given.