Nowcasting Pertumbuhan Produk Domestik Bruto Di Asean Dengan Midas-Quantile Regression

Latifah, Turfah (2025) Nowcasting Pertumbuhan Produk Domestik Bruto Di Asean Dengan Midas-Quantile Regression. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003232027-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Prediksi pertumbuhan ekonomi yang akurat dan tepat waktu sangat penting untuk mendukung pengambilan kebijakan fiskal dan moneter di negara-negara ASEAN, khususnya Indonesia, Malaysia, dan Singapura. Namun, keterbatasan frekuensi rilis pertumbuhan Produk Domestik Bruto (PDB) sebagai indikator utama perekonomian menghambat pemantauan kondisi ekonomi secara real-time. Penelitian ini bertujuan memodelkan nowcasting pertumbuhan PDB dengan model Mixed Data Sampling-Quantile Regression (MIDAS-QR) untuk memprediksi nilai Growth-at-Risk (GaR) pada kuantil tertentu. Model ini mengintegrasikan data multi-frekuensi yang dirangkum dalam Financial Conditions Index (FCI), External Financial Environment Index (EFEI), dan Macroeconomic Prosperity Leading Index (MPLI) yang diperoleh melalui Principal Component Analysis (PCA). Evaluasi model dilakukan menggunakan Quantile Mean Absolute Error (QMAE), Quantile Root Mean Squared Error (QRMSE), dan uji Clark-West untuk mengukur akurasi dan keandalan prediksi di berbagai kuantil risiko, serta memvalidasi model menggunakan Expected Shortfall (ES) dan Kupiec Test. Hasil penelitian menunjukkan bahwa model MIDAS-QR secara signifikan meningkatkan ketepatan prediksi GaR pada kuantil ekstrem (95% untuk Indonesia dan Malaysia, serta 5% untuk Singapura) dibandingkan regresi kuantil konvensional. Model ini efektif mengikuti pola data aktual dan mampu mendeteksi risiko pertumbuhan ekonomi secara dini. Selain itu, hasil nowcasting pertumbuhan PDB bulanan dan kuartalan membuktikan kemampuan model MIDAS-QR dalam memantau dinamika ekonomi secara real-time serta memberikan peringatan dini atas potensi perlambatan ekonomi di ketiga negara tersebut. Temuan ini menegaskan pentingnya pemanfaatan data frekuensi tinggi dan pendekatan risiko ekor (tail-risk) dalam meningkatkan ketepatan prediksi pertumbuhan PDB dan stabilitas ekonomi di ASEAN. Oleh karena itu, MIDAS-QR direkomendasikan sebagai alat bantu pengambilan keputusan kebijakan makroekonomi.
==================================================================================================================================
Accurate and timely economic growth prediction is essential to support fiscal and monetary policy making in ASEAN countries, especially Indonesia, Malaysia, and Singapore. However, the limited frequency of Gross Domestic Product (GDP) growth releases as the main economic indicator hampers real-time monitoring of economic conditions. This study aims to model GDP growth nowcasting with the Mixed Data Sampling-Quantile Regression (MIDAS-QR) model to predict Growth-at-Risk (GaR) values at certain quantiles. This model integrates multi-frequency data summarized in the Financial Conditions Index (FCI), External Financial Environment Index (EFEI), and Macroeconomic Prosperity Leading Index (MPLI) obtained through Principal Component Analysis (PCA). Model evaluation is carried out using Quantile Mean Absolute Error (QMAE), Quantile Root Mean Squared Error (QRMSE), and the Clark-West test to measure the accuracy and reliability of predictions at various risk quantiles, and validated using Expected Shortfall (ES) and the Kupiec Test. The results show that the MIDAS-QR model significantly improves the accuracy of GaR predictions at extreme quantiles (95% for Indonesia and Malaysia, and 5% for Singapore) compared to conventional quantile regression. This model effectively follows the actual data pattern and is able to detect economic growth risks early. In addition, the results of monthly and quarterly GDP growth nowcasting prove the ability of the MIDAS-QR model to monitor economic dynamics in real-time and provide early warnings of potential economic slowdowns in the three countries. These findings emphasize the importance of utilizing high-frequency data and the tail-risk approach in improving the accuracy of GDP growth predictions and economic stability in ASEAN. Therefore, MIDAS-QR is recommended as a tool for macroeconomic policy decision-making.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Evaluasi Model, Growth-at-Risk (GaR), Mixed Data Sampling-Quantile Regression (MIDAS-QR), Pertumbuhan Produk Domestik Bruto (PDB), Principal Component Analysis (PCA). Model Evaluation, Growth-at-Risk (GaR), Mixed Data Sampling-Quantile Regression (MIDAS-QR), Gross Domestic Product (GDP) Growth, Principal Component Analysis (PCA).
Subjects:	Q Science Q Science > QA Mathematics Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Turfah Latifah
Date Deposited:	31 Jul 2025 08:38
Last Modified:	31 Jul 2025 08:38
URI:	http://repository.its.ac.id/id/eprint/124876

Actions (login required)

View Item