Fatikha, Ayunda (2026) Analisis Risiko Sistemik Saham Sektor Industri Menggunakan Conditional Value-at-Risk Berbasis AR-GARCH dan Quantile Regression Forest dengan Stochastic Search Variable Selection. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5003221023-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5003221023-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (7MB) | Request a copy |
Abstract
Sektor industri memiliki kontribusi terbesar terhadap PDB Indonesia dengan nilai 18,98% pada akhir tahun 2024 rentan terhadap risiko sistemik. Volatilitas return yang tinggi dan fenomena clustering volatility dapat menimbulkan kerugian ekstrem yang tidak hanya berdampak pada perusahaan individual, tetapi juga dapat menyebar ke seluruh sistem keuangan melalui spillover effect. Penelitian ini menganalisis risiko sistemik saham sektor industri di BEI menggunakan VaR berbasis AR-GARCH, CoVaR berbasis QRF, ΔCoVaR, dan SSVS pada sepuluh emiten sektor industri dengan kapitalisasi pasar tertinggi (ASII, IMPC, UNTR, SKRN, SINI, SMIL, ARNA, HEXA, TOTO, dan MARK) periode 3 Juli 2023 hingga 31 Desember 2025. Hasil penelitian menunjukkan bahwa seluruh saham memiliki distribusi return yang memiliki ekor tebal disertai pola asimetri, sehingga distribusi standardized skewed Student-t (sstd) digunakan dalam pemodelan AR-GARCH. Estimasi VaR menghasilkan model yang valid secara statistik pada sebagian besar emiten, kecuali SINI yang menunjukkan underestimation risiko pada kuantil 5%. Estimasi CoVaR QRF menunjukkan adanya keterkaitan risiko sistemik yang bervariasi antar emiten dan antar tingkat kuantil, di mana pada kondisi ekstrem batas kerugian kondisional berfluktuasi lebih dalam dibandingkan kondisi lebih moderat. Model CoVaR QRF terbukti valid secara statistik pada hampir seluruh saham, kecuali IMPC pada kuantil 1% akibat karakteristik distribusi return yang volatil dan lonjakan ekstrem pada akhir periode pengamatan. ΔCoVaR yang mengidentifikasi potensi kerugian kondisional yang semakin dalam pada kondisi ekstrem di beberapa saham, dengan TOTO mencatat kontribusi risiko sistemik paling terkendali di antara seluruh saham. Penerapan SSVS pada CoVaR menghasilkan model yang lebih parsimonious. Perbandingan akurasi ketiga model berdasarkan menunjukkan bahwa CoVaR QRF konvensional secara umum menghasilkan akurasi frekuensi pelanggaran yang lebih baik dibandingkan CoVaR SSVS, mengindikasikan bahwa seleksi variabel melalui SSVS menghasilkan model yang lebih parsimonious, namun tidak selalu meningkatkan akurasi frekuensi pelanggaran dibandingkan model yang menggunakan seluruh prediktor dalam mengidentifikasi sumber risiko sistemik sektor industri =====================================================================================================================================
The industrial sector contributes the largest share to Indonesia's GDP at 18.98% as of the end of 2024, yet remains vulnerable to systemic risk. High return volatility and volatility clustering phenomena in the industrial sector can generate extreme losses that not only affect individual companies but may also propagate throughout the financial system through spillover effects. This study analyzes the systemic risk of industrial sector stocks listed on the Indonesia Stock Exchange (IDX) using VaR based on AR-GARCH, CoVaR based on QRF, ΔCoVaR, and SSVS on ten industrial sector stocks with the highest market capitalization (ASII, IMPC, UNTR, SKRN, SINI, SMIL, ARNA, HEXA, TOTO, dan MARK) over the period of July 3, 2023 to December 31, 2025. The results indicate that all stocks exhibit leptokurtic return distributions with heavy tails, accompanied by varying degrees of asymmetry across stocks that leading to the adoption of the standardized skewed Student-t (sstd) distribution in AR-GARCH modeling. VaR estimation yields statistically valid models for most stocks, except for SINI which shows risk underestimation at the 5% quantile. QRF-based CoVaR estimation reveals varying systemic risk linkages across stocks and quantile levels, where conditional loss boundaries are deeper under extreme market conditions compared to more moderate conditions. The CoVaR QRF model proves statistically valid for almost all stocks, except for IMPC at the 1% quantile which exhibits a tendency toward risk underestimation due to its highly volatile return distribution and the presence of extreme spikes toward the end of the observation period. ΔCoVaR identifies increasingly deeper conditional loss potential under extreme conditions for several stocks, with TOTO recording the most controlled systemic risk contribution among all stocks. The application of SSVS to CoVaR yields a more parsimonious model. A comparison of the accuracy of the three models indicates that conventional CoVaR QRF generally produces better violation frequency accuracy than CoVaR SSVS, suggesting that variable selection through SSVS produces a more parsimonious model but does not consistently improve violation frequency accuracy compared to models using all predictors in identifying sources of systemic risk in the industrial sector.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | AR-GARCH, CoVaR, QRF, SSVS, Industrial Sector, AR-GARCH, CoVaR, QRF, SSVS, Sektor Industri |
| Subjects: | H Social Sciences > HA Statistics H Social Sciences > HA Statistics > HA30.3 Time-series analysis H Social Sciences > HA Statistics > HA31.7 Estimation H Social Sciences > HG Finance H Social Sciences > HG Finance > HG4529 Investment analysis H Social Sciences > HG Finance > HG4915 Stocks--Prices Q Science Q Science > QA Mathematics Q Science > QA Mathematics > QA274.2 Stochastic analysis Q Science > QA Mathematics > QA280 Box-Jenkins forecasting Q Science > QA Mathematics > QA402 System analysis. |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Ayunda Fatikha |
| Date Deposited: | 30 Jul 2026 01:36 |
| Last Modified: | 30 Jul 2026 01:36 |
| URI: | http://repository.its.ac.id/id/eprint/140034 |
Actions (login required)
 |
View Item |