Abstract:Modeling financial data often relies on assumptions that may prove insufficient or unrealistic in practice. The Geometric Brownian Motion (GBM) model is frequently employed to represent stock price processes. This study investigates whether the behavior of weekly and monthly returns of selected equities listed on the Ghana Stock Exchange conforms to the GBM model. Parameters of the GBM model were estimated for five equities, and forecasts were generated for three months. Evaluation of estimation accuracy was conducted using mean square error (MSE). Results indicate that the expected prices from the modeled equities closely align with actual stock prices observed on the Exchange. Furthermore, while some deviations were observed, the actual prices consistently fell within the estimated confidence intervals.
What problem does this paper attempt to address?
This paper explores the modeling problem of stock price dynamics in the Ghana Stock Exchange (GSE), mainly using the Geometric Brownian Motion (GBM) model. The aim of the study is to test whether the GBM model is applicable in describing the behavior of weekly and monthly returns of selected stocks on the GSE and to make predictions based on this model.
The author first estimates the parameters of the GBM model and makes three-month predictions for five stocks. The accuracy of the predictions is evaluated using the Mean Square Error (MSE). The results show a close correlation between the estimated prices of the model and the observed stock prices, with most actual prices falling within the estimated confidence intervals.
The paper reviews existing research in the financial field on stock price modeling, including extensions and improvements to the GBM model. The research methods include data exploration, ADF unit root test, normality tests (such as histograms, Q-Q plots, and Shapiro-Wilk test), Ljung-Box autocorrelation test, and Hurst index estimation, to examine the independence and randomness of the time series.
The study found that the GBM model assumption holds for the monthly returns of five stocks, including GCB. However, the weekly returns of most stocks do not conform to the model's assumptions of normal distribution and random walk. Therefore, the paper focuses on analyzing the monthly returns data that meet the GBM assumption.
The GBM model parameters are estimated using the maximum likelihood estimation method, and six months of predictions are generated based on the closing prices in September 2015. The prediction results show that despite some biases, the model predicts stock prices close to the actual trading prices, with actual prices falling within the estimated 95% confidence intervals.
The conclusion of the paper suggests that investors can use the GBM model to guide decision-making and predict stock prices on the GSE, especially for stocks that meet the GBM assumption, such as GCB and TLW. Future research is recommended to focus more on monthly returns data as they are more consistent with the behavior of the GBM model.