Modelling the Naira Exchange Rate Dependence Using Static and Time-Varying Copula

Abstract

This paper examines the dependence structure of different currencies versus the Nigerian Naira using constant and time-varying copula. Daily Naira/USD, Naira/Yuan, Naira/Pound, and Naira/Euro exchange rates from 23 December 2011 to 12 May 2020 were utilised. We fitted eight constant and time-varying copula families using the exchange rate standardised residuals. The study finds that the Naira exchange rate may be estimated with student t-copula, Symmetrized Joe-Clayton (SJC), or Rotated Gumbel copula models and Autoregressive (AR)– Glosten Jagannathan Runkle Generalized Autoregressive Conditional Heteroscedastic (GJR-GARCH) (1,1) models with skewed t residuals for margins. The Naira exchange rate returns is time-varying, tail-dependent, and asymmetric. The study recommends that portfolio diversification, asset allocation of Central Bank of Nigeria foreign reserves, bank risk capital aggregation, and risk management decisions should not be based on linear correlation coefficient (Gaussian copula) but on copula models that can capture asymmetry and tail dependence, such as Student t, SJC, and Gumbel copulas.