Forecasting KRW-USD Exchange Rate Volatility and Analyzing the Risk Premium

Abstract

Recently, due to heightened uncertainty surrounding U.S. monetary policy, the volatility of the KRW-USD exchange rate has increased significantly. Therefore, this study transforms the KRW-USD exchange rate data into log return data to forecast exchange rate volatility and analyze the risk premium. The models used to predict volatility are the AR-GARCH model and the AR-GARCH-M model. The log return data were found to follow an AR(2)-GARCH(1,1) process under the AR-GARCH model, and an AR(1)-GARCH(1,1)-M process under the GARCH(p,q)-M model. The goodness-of-fit for the models was tested using the Portmanteau Q-test for autocorrelation, and the results indicated that the models were appropriate. Based on these fitted models, the predicted exchange rate volatility is expected to remain relatively low over a certain period. Furthermore, analysis of the risk premium using the AR(1)-GARCH(1,1)-M model showed that there is no risk premium present in exchange rate returns.