Abstract

DIAGONAL MULTIVARIATE GENERALISED AUTOREGRESSIVE CONDITIONAL HETEROSKEDASTICITY MODELS AND THEIR APPLICATIONS TO REAL-TIME VOLATILITY SERIES.

Usoro, Anthony E. and Ekong, Nsisong P.

---

The goal of this work was to create new multivariate time series models for the volatility series. Existing multivariate time series models for volatility series, such as Multivariate Generalized Autoregressive Conditional Heteroskedasticity (MGARCH) models, are used to identify new classes of models under certain conditions. The parameters for the UDMGARCH and LDMGARCH models are limited to the upper and lower diagonals of the coefficient matrices, respectively. Using empirical evidence from Nigerian crude oil quantity and price volatility series, the novel models are found to be adequate and have the same comparative advantage as the existing general MGARCH. As a result, UDMGARCH and LDMGARCH are established as new MGARCH model classes. Keywords: UDMGARCH, LDMGARCH, MGARCH, Crude Oil Quantity Volatility and Crude Oil Price Volatility.

PDF