Testing The Performance Of Conditional Variance-Covariance In Diagonal Mgarch Models Using Exchange Rate And Nigeria Commercial Banks Interest Rates
Deebom Zorle Dum
Rivers State Universal Basic Education Board (RSUBEB) Rivers State, Nigeria
Bharat Kumar Meher
Assistant Professor,Department of Commerce, D.S.College, Katihar, Under Purnea University, India - 854105
Inamete Emem Ndah H
Department of Statistics Technology, Federal Polytechnic of Oil and Gas, Bonny
Keywords: Exchange Rate, Commercial Banks Interest Rates, MGARCH
Abstract
The study examines the performance of conditional variance-covariance in Diagonal MGARCH models using exchange rate and Nigerian commercial Banks interest rate on time deposits maturing in a monthly order. The data for the study was on monthly Exchange rate (Excr), commercial banks interest rate on time deposits maturing in one month (CBIRTD1), commercial banks interest rate on time deposits maturing in three months (CBIRTD3), commercial banks interest rate on time deposits maturing in six months (CBIRTD6), commercial banks interest rate on time deposits maturing in twelve months (CBIRTD12).The monthly exchange rates (Naira/Dollars). These data were extracted from the Central Bank of Nigeria (CBN) statistical database website (www.cbn.gov.ng). The data spanned from January 1991 – December 2019 and they were fitted to monthly compound returns with the aid of STATA and Eviews Software version 15 and 10 respectively. Multivariate Diagonal VECH (DVECH) and BEKK-GARCH models were used to determine how variance- covariance interact over time and these models estimated conditional variance-covariance matrix of multiple of five time series data.The result obtained from the diagonal multivariate VECH GARCH model show that the covariance between these variables does not cluster overtime. It was also found that shocks arising from exchange rate have an effect on commercial banks interest rate on time deposits maturing whose estimates have the same positive conditional variance. The diagonal multivariate VECH model shows ‘positive semi-definite’ characteristic i.e variance-covariance matrix has positive estimates on its leading diagonal, and this shows that the entries are symmetrical about their leading diagonal. Also, the result obtained from the diagonal multivariate BEKK GARCH model shows that the estimated conditional variance depends on its own past historical innovations. This simply means volatility of exchange rate responds significantly to past squared shocks in its own sector, as well as in the commercial Banks interest rates operation.The result also confirmed that there exist a strong evidence of a time-varying conditional covariance and interdependence between exchange rate and indicators of Nigerian commercial banks interest rate. Similarly, evidence of cross-volatility effects was confirmed; past innovations in exchange rate have greatest influence on future volatility of Nigeria commercial banks interest rate returns. Based on Model selection criteria using the Akaike information criteria (AIC) diagonal VECH model is better fitted than the diagonal multivariate BEKK GARCH model