Perbandingan Model Hybrid LASSO-QRNN dan Embedded LASSO-QRNN pada Pemodelan Conditional Value at Risk Return Saham Sub-Sektor Minyak dan Gas Bumi di Indonesia

Syalsabila, Annisa (2024) Perbandingan Model Hybrid LASSO-QRNN dan Embedded LASSO-QRNN pada Pemodelan Conditional Value at Risk Return Saham Sub-Sektor Minyak dan Gas Bumi di Indonesia. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003221020-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2026. Download (5MB) | Request a copy

Abstract

Sub-sektor saham minyak dan gas bumi (migas) memegang peran kunci dalam dinamika ekonomi global dan investasi finansial saat ini. Analisis risiko terhadap sub-sektor migas menjadi lebih penting dilakukan mengingat dampak signifikan yang dimilikinya terhadap stabilitas ekonomi dan geopolitik. Oleh karena itu, analisis risiko memungkinkan investor dan pengambil keputusan untuk memahami dengan lebih baik dinamika dan potensi risiko yang terkait dengan sektor migas. Analisis risiko yang dilakukan dalam penelitian ini adalah analisis risiko sistemik. Penelitian ini bertujuan untuk menganalisis risiko sistemik pada perusahaan publik sub-sektor migas di Indonesia dengan menggunakan metode Conditional Value at Risk (CoVaR) yang mampu mendeteksi kemungkinan risiko sistemik antara dua pasar dengan memberikan informasi mengenai nilai risiko suatu pasar tergantung pada pasar lain yang dianggap berada dalam kondisi kesulitan keuangan. Pada penelitian ini proses seleksi variabel dipisah menjadi dua bagian analisis yang berbeda dan akan dibandingkan, yaitu proses hybrid dan proses embedded. Metode analisis risiko menggunakan model LASSO-QRNN digunakan karena memiliki kelebihan pada kemampuannya yang lebih banyak dalam menangkap krisis. Pemodelan Hybrid LASSO-QRNN dan Embedded LASSO-QRNN diaplikasikan pada data log return perusahaan-perusahaan sub-sektor minyak dan gas bumi di Indonesia dan variabel makroekonomi serta variabel komoditas energi dunia dengan perode data 3 September 2018 - 1 September 2023. Pemodelan dengan model hybrid LASSO-QRNN dan embedded LASSO-QRNN menunjukkan terjadinya peningkatan risiko sitemik pada periode terjadinya COVID-19 dan pada masa invasi Rusia ke Ukraina. Selain itu didapatkan bahwa model embedded LASSO-QRNN lebih baik dibandingkan dengan model hybrid LASSO-QRNN berdasarkan hasil backtesting
=============================================================================================================================
The oil and The sub-sector of oil and natural gas stocks plays a key role in the dynamics of the current global economy and financial investment. Risk analysis of the oil and gas sub-sector has become more crucial given its significant impact on economic and geopolitical stability. Therefore, risk analysis allows investors and decision-makers to better understand the dynamics and potential risks associated with the oil and gas sector. The risk analysis conducted in this study is systemic risk analysis. This research aims to analyze systemic risk in public companies within the oil and gas sub-sector in Indonesia using the Conditional Value at Risk (CoVaR) method. CoVaR can detect possible systemic risks between two markets by providing information about the risk value of one market depending on another market considered to be in financial distress. In this study, the variable selection process is divided into two different and comparable analyses: the hybrid process and the embedded process. The risk analysis method uses the LASSO-QRNN model due to its advantages in capturing crises more effectively. Hybrid LASSO-QRNN and Embedded LASSO-QRNN modeling are applied to the log return data of oil and gas sub-sector companies in Indonesia, as well as macroeconomic and world energy commodity variables, with a data period from September 3, 2018, to September 1, 2023. Modeling with the Hybrid LASSO-QRNN and Embedded LASSO-QRNN models indicates an increase in systemic risk during the COVID-19 period and the Russian invasion of Ukraine. Additionally, it is found that the Embedded LASSO-QRNN model is superior to the Hybrid LASSO-QRNN model based on the results of the backtesting.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	CoVaR, LASSO, risiko sistemik, quantile regression, systemic risk
Subjects:	H Social Sciences > HA Statistics > HA30.3 Time-series analysis H Social Sciences > HG Finance > HG4529 Investment analysis H Social Sciences > HG Finance > HG4910 Investments H Social Sciences > HG Finance > HG4915 Stocks--Prices
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Annisa Syalsabila
Date Deposited:	13 Feb 2024 23:29
Last Modified:	13 Feb 2024 23:29
URI:	http://repository.its.ac.id/id/eprint/106955

Actions (login required)

View Item