DAILY CHINESE YUAN – NIGERIAN NAIRA EXCHANGE RATES SARIMA MODELLING

Ette Harrison Etuk

Department of Mathematics,

Rivers State University of Science and Technology, Port Harcourt

Email: ettetuk@yahoo.com

ABSTRACT

A 179-point realization of daily exchange rates of the Chinese Yuan and the Nigerian Naira spanning from 18^th October 2015 and 13^th April 2016 is analyzed by seasonal autoregressive integrated moving average (SARIMA) methods.  The time plot shows an initial downward trend up to the middle of December 2015 and then an upward trend from then onwards. This means that prior to the middle of December 2015 the Naira was relatively appreciating before it started depreciating relatively.  The series is adjudged as non-stationary by the Augmented Dickey Fuller Test. A seven-point (i.e. seasonal) differencing of the series yields a series which, though adjudged stationary, still exhibits seasonality and therefore could not said to be stationary. A further (non-seasonal) differencing is done to achieve stationarity.  The autocorrelation structure of the resultant series suggests the possibility of some SARIMA models. These include a SARIMA(0,1,1)x(0,1,1)₁₂ and a SARIMA(0,1,0)x(0,1,1)₁₂. Comparison on the basis of the information criteria AIC, Schwarz criterion and Hannan-Quinn criterion shows that the former model is the superior. It is as well observed that its residuals are white noise.  Forecasting and simulation of these rates may be done on its basis.