Abstract
Volatility modelling in Nigeria and sub-Sahara Africa lacks the use of non–
Gaussian models for estimating stock market returns series; researchers
have been focusing on the Gaussian process. To overcome this
deficiency, we used non-Gaussian assumptions alongside normal
assumptions. Our major focus here is to investigate the presence of
leverage effect and volatility persistence in Nigeria Exchange non-Normal
specifications. GARCH models and their asymmetric extensions viz.
EGARCH, TGARCH and PGARCH were employed to conduct the analysis
with each estimated in Normal, Student’s-t and Generalized error
distributions. Student-t specification outperformed the other error
distribution models. Important outcomes of the study include that the
Exchange did not have an evidence of leverage effect; EGARCH, TGARCH
and PGARCH models all showed empirical results contrary to the
theoretical a priori signs of asymmetry. Secondary, there is high volatility
persistence in the market, EGARCH model even indicated an explosive
volatility persistence. We conclude that Nigeria Exchange is dominated
by uninformed investors with a short-term investment strategy.
Additionally, the market is very volatile and has the implication of
weakening the investors’ confidence.