Exchange rate volatility: A forecasting approach of using the ARCH family along with ARIMA SARIMA and Semi-Structural-SVAR in Turkey

Abstract

The ability to predict the volatility of exchange rate is an enormous challenge when it comes to economic and financial considerations. In this context, it is important to be able to predict the exchange rate volatility in financial markets and the world economy. This paper proposes a heightened approach to modelling and forecasting of exchange rate volatility in Turkey. For past recent years, Turkey experienced political turbulence that the possibility of effecting exchange rate,
thus create uncertainty volatility of exchange rate. The daily exchange rate data have been taken from 2005-2017 and applied autoregressive conditional heteroscedasticity ARCH and GARCH families (EGARCH, IGARCH, and PARCH) to forecast exchange rate volatility. The proposed methodology able to calculate the breakpoint by including dummy variables. The result is more confined after including dummy that EGARCH (1,1) is best performing to forecast exchange rate volatility and successfully overcome the leverage effect on the exchange rate. Moreover, this paper also investigates the monthly data forecasting by applying ARIMA SARIMA along with SVAR technique for next few months. The Exchange rate pass-through also encounter it, which indicates the pass-through is more pronounced in PPI than CPI. The forecast result of SARIMA and SVAR distribute the same direction of fluctuation in exchange rate that is declining of current exchange rate in the future. However, ARIMA’s forecast tends to increase and different from the two models.