An empirical assessment of symmetric and asymmetric jump-diffusion models for the Nigerian stock market indices

We examine empirically, the suitability of three stock price models viz: geometric Brownian motion, symmetric and asymmetric jump-diffusion models, on the empirical log-returns of the Nigerian All-Share Index. 5334 daily observed data from January 2, 1998, to February 21, 2020, were utilized. Using a non-parametric jump-test method, our results show that jumps are present in the empirical log-returns of the stock market price. The results obtained for the optimal parameters in the models indicate high jump intensity, more upward jumps, and a positively skewed jump process. However, the parameters in the asymmetric jump-diffusion model were found to be more sensitive to the varied threshold of jumps in the log-returns than the symmetric jump-diffusion model. The suitability analysis results show that the symmetric jump-diffusion model fits the market indices better. Therefore, it can be used for future predictions of the market price.