Integrasi Generalized Extreme Value Dan Regresi Kuantil Dalam Mengukur Kontribusi Risiko Sistemik Saham Grup Prajogo Pangestu

Rahmawati, Linda Dwi Rahmawati (2026) Integrasi Generalized Extreme Value Dan Regresi Kuantil Dalam Mengukur Kontribusi Risiko Sistemik Saham Grup Prajogo Pangestu. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221045_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (9MB) | Request a copy

Abstract

Fenomena volatilitas ekstrem pada awal tahun 2025 di BEI ditandai dengan kejatuhan signifikan pada saham-saham Grup Prajogo Pangestu hingga memicu mekanisme circuit breaker, hal ini menunjukkan kegagalan model statistik konvensional dalam mengantisipasi risiko sistemik. Distribusi normal sering kali meremehkan potensi kerugian pada kondisi ekor tebal (fat tails), sehingga diperlukan pendekatan yang lebih robust untuk memodelkan kejadian langka namun berdampak masif. Penelitian ini bertujuan untuk mengestimasi profil risiko ekstrem dan transmisi risiko sistemik pada saham PTRO, BRPT, TPIA, CUAN, dan BREN menggunakan kerangka Extreme Value Theory (EVT). Metodologi yang digunakan adalah distribusi Generalized Extreme Value (GEV) dengan pendekatan Block Minima (BM) melalui model statis dan dinamis. Risiko individual diukur menggunakan Value at Risk (VaR), sementara risiko sistemik dianalisis melalui Conditional Value at Risk (CoVaR) dan ΔCoVaR berbasis regresi kuantil. Hasil penelitian menunjukkan bahwa pada tingkat kepercayaan 99%, CUAN dan BREN memiliki risiko mandiri (VaR) tertinggi (-13,65% dan -13,38%), sementara PTRO mencatatkan dampak domino sistemik ekstrem (CoVaR) terdalam hingga -29,50%. CUAN teridentifikasi sebagai systemically important firm dengan kontribusi marjinal (ΔCoVaR) tertinggi sebesar -13,60%, sedangkan BRPT menjadi emiten paling stabil di seluruh level pengujian. Pendekatan dinamis terbukti sangat adaptif melalui pembagian blok (n=5). Uji Kupiec dan akurasi Expected Shortfall (ES) mengonfirmasi validitas model yang kokoh dengan selisih risiko di bawah 1%. Akurasi tertinggi ditemukan pada PTRO, BRPT, dan CUAN (selisih 0,19% - 0,60%) pada tingkat kepercayaan 99%. Temuan ini memvalidasi model GEV-BM sebagai instrumen mitigasi risiko yang efektif untuk menghadapi volatilitas ekstrem di sektor pertambangan, energi dan petrokimia.
====================================================================================================================================
The extreme volatility phenomenon in early 2025 on the IDX was marked by a significant drop in Prajogo Pangestu Group shares, triggering a circuit breaker mechanism. This demonstrates the failure of conventional statistical models in anticipating systemic risk. The normal distribution often underestimates potential losses in fat tail conditions, so a more robust approach is needed to model rare but massively impactful events. This study aims to estimate the extreme risk profile and systemic risk transmission in PTRO, BRPT, TPIA, CUAN, and BREN shares using the Extreme Value Theory (EVT) framework. The methodology used is the Generalized Extreme Value (GEV) distribution with a Block Minima (BM) approach through static and dynamic models. Individual risk is measured using Value at Risk (VaR), while systemic risk is analyzed using Conditional Value at Risk (CoVaR) and ΔCoVaR based on quantile regression. The results indicate that at a 99% confidence level, CUAN and BREN exhibit the highest individual risk (VaR) at -13.65% and -13.38%, respectively. In terms of systemic risk, PTRO recorded the most extreme domino impact (CoVaR) reaching -29.50%, while CUAN dominated as the systemically important firm with the highest marginal contribution (ΔCoVaR) of -13.60%. Conversely, BRPT was found to be the most resilient issuer with the lowest systemic impact across all testing levels. The dynamic approach proved highly adaptive in capturing weekly risk fluctuations through block partitioning (n=5). Kupiec’s test and Expected Shortfall (ES) accuracy confirmed the model's robust validity, with risk deviations (absolute error) predominantly below 1%. The highest precision was observed for PTRO, BRPT, and CUAN, showing minimal risk gaps (0.19% – 0.60%) at the 99% confidence level. These findings validate the GEV-BM model as an effective risk mitigation instrument against extreme volatility in the energy and petrochemical sectors.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Extreme Value Theory, GEV, VaR, CoVaR, Risiko Sistemik.
Subjects:	H Social Sciences > HG Finance > HG4529.5 Portfolio management
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Linda Dwi Rahmawati
Date Deposited:	02 Feb 2026 01:47
Last Modified:	02 Feb 2026 01:47
URI:	http://repository.its.ac.id/id/eprint/131168

Actions (login required)

View Item