COMPARATIVE ANALYSES OF DISTRIBUTIONS IN ASSYMETRY GARCH MODELLING: A STUDY OF NGN/USD EXCHANGE RATE

Abstract

In an era of globalization characterized by flexible exchange rate systems, including Nigeria,
the examination of foreign exchange rate volatility has become critically significant in recent
decades, attracting the interest of both scholars and policymakers. Examining the dynamic
variability of exchange rate series with distributional assumptions is highly significant. This
paper examines the volatility of the Naira/US dollar exchange rate utilizing the EGARCH (1,
1) model from the GARCH family, with particular attention to generalized t, skewed student
t, and skewed normal distributions. Data on the secondary Naira/Dollar exchange rate was
obtained from the Central Bank of Nigeria's website, covering the period from January 2003
to April 2023. Monthly exchange rate returns were utilized to estimate the GARCH
parameters employing the previously described distributions. The findings demonstrated that
the skew student t (ST) distribution had superior predictive capability for N/$ exchange rate
volatility, as evidenced by its elevated log-likelihood, reduced AIC, and diminished BIC
within the chosen EGARCH (1, 1) model family. Furthermore, the results from the forecast
evaluation revealed the existence of generated conditional variance, suggesting that the
variance reverts to a long-term mean. The EGARCH model provided significant insights into
volatility dynamics; however, it is concluded that the selection of distribution is crucial for
improving its performance, with the skew Student's t distribution, due to its flexibility and
adaptability to varying market conditions, identified as the most effective estimator in this
study