Dhardiri, Hudan (2025) Peramalan Volatilitas Nilai Tukar Rupiah Terhadap USD dengan Pendekatan Markov switching GARCH MIDAS. Masters thesis, Institut Teknologi Sepuluh Nopember.
|
Text
6003232009-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy |
Abstract
Keterbukaan ekonomi membawa nilai tukar mata uang sebagai salah satu indikator penting bagi suatu negara. Volatilitas nilai tukar mata uang rupiah khususnya terhadap USD yang tinggi akan berdampak pada kinerja ekspor & impor, investasi, stabilitas harga domestik serta kemungkinan terjadinya krisis ekonomi bagi suatu negara. Penelitian ini bertujuan memodelkan dan peramalan volatilitas nilai tukar rupiah dengan pendekatan model Markov Switching GARCH-MIDAS berbasis teori Taylor Rule Fundamentals. Model GARCH-MIDAS digunakan untuk memodelkan volatilitas jangka pendek dan volatilitas jangka panjang dengan prediktor yaitu suku bunga, tingkat inflasi, dan output gap yang memiliki frekuensi berbeda. Pendekatan Markov Switching digunakan untuk mengatasi perubahan struktur pada data dengan memperhatikan peluang perpindahan regime. Selain itu, sebagai pembanding tambahan juga akan digunakan model hybrid dengan XGBoost. Data in-sample yang digunakan untuk pemodelan adalah data dari November 2005 hingga Juni 2025 serta data out-sample pada Juli 2025. Pemodelan dilakukan dengan dua spesifikasi prediktor yaitu nilai selisih secara langsung (Tipe I) dan nilai volatilitasnya (Tipe II), serta tiga spesifikasi lag pada model MIDAS yaitu lag 6 bulan, 12 bulan, dan 24 Bulan. Hasil Pemodelan menunjukkan model terbaik berdasarkan ukuran kebaikan AIC, BIC, dan HQIC pada model GARCH-MIDAS adalah model tipe I dengan lag 24 bulan dan model tipe 2 dengan lag 24 bulan. Sementara pada model MS-GARCH-MIDAS didapatkan hasil model terbaik dengan lag 12 bulan dengan prediktor tipe I dan lag 24 bulan dengan prediktor tipe II. Peramalan pada data out-sample menunjukkan pada model MS-GARCH-MIDAS tipe I dengan lag 24 bulan lebih baik dibandingkan model GARCH-MIDAS dan GARCH-MIDAS-XGBoost dengan indikator MAE dan RMSE.
==================================================================================================================================
Economic openness positions the exchange rate as a key macroeconomic indicator for a country. High volatility in the Indonesian rupiah, particularly against the USD, can affect export and import performance, investment activity, domestic price stability, and may increase the likelihood of an economic crisis. This study aims to model and forecast the volatility of the rupiah exchange rate using the Markov Switching GARCH-MIDAS framework based on Taylor Rule Fundamentals. The GARCH-MIDAS model is employed to separate short-term and long-term volatility components by incorporating interest rates, inflation, and the output gap, all with different data frequencies. The Markov Switching approach identifies changes in regimes, distinct periods with different statistical properties, by estimating the probability of shifts between these regimes. Additionally, an XGBoost model, a machine learning method, is included as a supplementary benchmark. The in-sample data covers the period from November 2005 to June 2025, while the out-of-sample data from July 2025. The modeling procedure considers two predictor specifications: direct value differences (Type I) and predictor volatility (Type II), along with three lag structures for the MIDAS component: 6 months, 12 months, and 24 months. The results indicate that the best-performing GARCH-MIDAS models based on AIC, BIC, and HQIC are the Type I model with a 24-month lag and the Type II model with a 24-month lag. For the MS-GARCH-MIDAS framework, the best-performing models are those with a 12-month lag for Type I predictors and a 24-month lag for Type II predictors. Out-of-sample forecasts show that the MS-GARCH-MIDAS Type I model with a 24-month lag outperforms both the GARCH-MIDAS and GARCH-MIDAS-XGBoost models based on MAE and RMSE indicators.
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | Nilai Tukar Rupiah, Volatilitas, GARCH, MIDAS, Markov switching, Exchange Rate, Volatility, GARCH, MIDAS, Markov switching |
| Subjects: | H Social Sciences > HA Statistics > HA30.3 Time-series analysis |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis |
| Depositing User: | Hudan Dhardiri |
| Date Deposited: | 19 Jan 2026 08:51 |
| Last Modified: | 19 Jan 2026 08:51 |
| URI: | http://repository.its.ac.id/id/eprint/129754 |
Actions (login required)
![]() |
View Item |
