Forecasting the Volatility of the Rupiah Exchange Rate Against the United States Dollar Using GARCH-M

Source abstract

This study aims to analyze and forecast the volatility of the Indonesian Rupiah exchange rate against the United States Dollar (USD/IDR) using the Generalized Autoregressive Conditional Heteroscedasticity (GARCH) and GARCH in Mean (GARCH-M) models. The data used consist of daily USD/IDR exchange rates from March 2022 to August 2024, totaling 630 observations, which were transformed into returns. The analysis begins with a stationarity test using the Augmented Dickey-Fuller (ADF), indicating that the data become stationary after first-order differencing. The ARIMA(0,1,1) model is selected as the best mean model, and the ARCH effect test confirms the presence of conditional heteroscedasticity, justifying the use of GARCH modeling. The estimation results show that the GARCH(1,1) model is the most appropriate, exhibiting a high level of volatility persistence. This model effectively captures the volatility clustering phenomenon and produces independent residuals. Forecasting for the next 30 days indicates increasing uncertainty as the forecast horizon extends. The model evaluation yields an RMSE of 67.45, MAE of 52.18, and Theil’s U of 0.0022, indicating excellent forecasting performance. The findings of this study are expected to provide valuable insights for policymakers and market participants in managing exchange rate risk.