Multivariate GARCH modelling of volatility of Nigerian stock market and some economic indices

Abstract

Volatility and co-volatility modelling among financial series is an important aspect in financial time series. Multivariate generalized autoregressive conditional heteroscedasticity (MGARCH) models, model the variances and co-variances among financial data. The GARCH model has been applied in modelling volatility of various univariate time series data but limited work has been done in the application of multivariate GARCH models in modelling multivariate time series data. This study is aimed at applying the Multivariate GARCH model on the co-volatility of Nigerian stock market, USD/Naira exchange rate and inflation rate using the Baba-Engle-Kraft-Kroner (BEKK) model for estimation. The best fit model for the series is the BEKK (1, 2) model in terms of minimum model evaluation tools.The variance-covariance models shows that volatilities of exchange rate and inflation rate influence volatility of stock price returns. It also shows that shock in stock prices greatly affect the volatility of stock prices.

Keywords: Baba-Engle-Kraft-Kroner model, Exchange rate, Inflation rate, Stock prices, Univariate GARCH,