On the Comparison of VECH and BEKK in Modeling of Oil Prices, Stock Exchange, Exchange and Inflation Rates Volatility in Nigeria

Abstract

A crucial component of financial time series is the modeling of volatility and co-volatility. The variances and covariances among financial data are modeled by multivariate generalized autoregressive conditional heteroscedasticity (MGARCH) models. The generalized autoregressive conditional heteroscedasticity (GARCH) model has been used to describe the volatility of a variety of univariate time series data, but there has not been much research done on
using multivariate GARCH models to model multivariate time series data. Thus, this study aimed at comparing the performance of vector error conditional heteroscedasticity (VECH) and Baba Engle, Kraft, and Kroner (BEKK) MGARCH in modeling of oil prices, stock exchange, inflation and exchange rates volatility in Nigeria. The data for the study were collected from Central Bank of Nigeria Website and World Bank Data base. The data collected were analyzed using Augmented Dickey Fuller (ADF) test, diagonal VECH and diagonal BEKK. The results of the analysis revealed that diagonal BEKK performed better than the diagonal VECH in terms of model selection criterion. Based on the conditional-covariance results, it was concluded that the volatility spillover effects were strong and significant for all the variables except for the shocks of the returns of inflation rate and persistence shocks for the returns of exchange rate. Also, the
magnitude of the estimate is not homogeneous across the variables but remains within a relatively tight range. The study recommended that further research should consider comparing diagonal BEKK with other MGARCH models.