ABSTRACT
This paper modeled Nigeria exchange rate using the GARCH model with three distributions, the normal distribution, student-t distribution and the generalized error distribution with the aim of comparing their ability to appraise excess kurtosis in closeness to the theoretical value of kurtosis. The data applied in this study is Nigeria exchange rate of Naira against British Pound sterling from 2002 to 2020, comprising of 228 observations. The result from the analysis revealed that GARCH (2, 1)-std, was the best model for Naira to Pound exchange rate return series. The selection was based on minimum information criteria and its ability to appraise excess kurtosis in closeness to the theoretical value of kurtosis. The result of the fitted model also indicate that the parameters are significant and that volatility is persistent. Diagnostic analysis of the selected model revealed that the model was correctly specified with no ARCH effect and autocorrelation in the residuals of the models. Our study has also has proven that the GARCH model with respect to student distribution perform better in terms of appraising excess kurtosis in closeness to the theoretical value of kurtosis.
Support the magazine and subscribe to the content
This is premium stuff. Subscribe to read the entire article.
