Modelling and Forecasting Estimation Exchange Rate Volatility in the Sudan

Abstract

The exchange rate is one of the macro-economic variables that have an impact on macroeconomic with its different sectors and the Exchange rate policy is one the most important policies that adopted by some countries to solve some of the economic problems. Therefore we must stand on timeline impact of the exchange rate in Sudan and build a predictive model for predicting exchange rates in Sudan. It requires finding suitable models to the nature of commercial time-series data. We must judge these models that they can represent the data, in this research we apply the Autoregressive conditional models conditioned by non- Homogenization on the exchange rates in Sudan to provide a predictive model Data collection was based on the monthly readings of exchange rates in Sudan in the period from 1/1/1999 to 31/12/2013 Issued by the Central System of Statistics and the Bank of Sudan Where he used a form of GARCH symmetric and asymmetric models to predict the best model in addition to the ARIMA and Autoregressive conditional Heteroskedasticity models conditioned by non- Homogenization Using the normal distribution and distribution of (t-student). 1- The summary statistics indicate that the returns series have monthly positive mean (0.0051) while the volatility is (0.013) without loss of generality the mean grows at linear rate while the volatility grows approximately at square root rate. 2- The returns series of the exchange rate shows positive skewness this implies that the series of exchange rate is flatter to the right 3- The kurtosis value is the higher than the normal and this suggest that the kurtosis curve of the exchange rate return series is leptokurtic. 4- The coefficient in the condition variance equation GARCH(1,1) the α significant and β not significant and the (α+β) is greater than one suggesting that the condition variance process is explosive. 5- The coefficient (risk premium) of in the mean equation is positive of the market which indicate the mean of the return sequence depend on past innovation and the past conditional variance. 6- The estimation of EGARCH(1,1) model for return series of exchange rate the γ is negative and significant meaning that return series have asymmetry and has greater impact of negative shocks indicate that the conditional variance has leverage effect and asymmetry of negative shocks. 7- The result indicate that the forecasting performance of the GJR-GARCH(1,1) and DGE-GARCH(1,1) models especially when fat-tailed asymmetric conditional distribution are taken into account in the conditional volatility is better than the GARCH(1,1) model. 8- However the comparison between the models with normal and student-t distribution shows that according to the different measures used for evaluating the performance of volatility forecasts the DGE-GARCH(1,1) model provides the best forecasts. 9- It is a found that the student-t distribution is more appropriates for modeling and forecasting exchange rate return volatility.