A time series analysis of the performance of seven stocks listed on the Zimbabwe Stock Exchange (2009-2015)
MetadataShow full item record
An attempt to fit a time series model that can be used to forecast patterns, trends and volatility of five stocks listed on the Zimbabwe Stock Exchange (ZSE) for the period 2009 to 2015 was made. Daily closing stock price data for selected companies trading on the ZSE was used. Using Python 3 and Compendium of statistics, tentative Generalized Additive Model (GAM) and Auto Regressive Integrated Moving Average (ARIMA) models were fitted. In order to assess the investment value at risk stock price volatility was modelled. The results show that the GAM model can be used for forecasting, since the forecasts exhibited excellent properties of being best linear unbiased estimates (blue) with least Mean Squared Error (MSE) compared to the traditional ARIMA (p, d,q) models. The most appropriate volatility model was found to be the GARCH (1,1) model which exhibited statistical significance based on its MSE at 0.05 significance level. In general, the GAM models outperformed their ARIMA counter parts and they are highly recommended in this research. It was also found that there is no one single GAM model for all the stocks, different stocks may require different GAM models. Further studies in this area should include more counters listed on the ZSE. Keywords: Generalised Additive Models; Best Linear Unbiased Estimates; Value at Risk; Volatility; Forecasts.