Optimasi Portofolio Saham IDX30 Menggunakan Conditional Value at Risk dengan Estimasi Return Quantile Regression Forest dan Hierarchical Risk Parity

Setyaningtyas, Tri Ega (2026) Optimasi Portofolio Saham IDX30 Menggunakan Conditional Value at Risk dengan Estimasi Return Quantile Regression Forest dan Hierarchical Risk Parity. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5006221099-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (9MB) | Request a copy

Abstract

Indeks IDX30 merupakan salah satu indeks saham utama di Bursa Efek Indonesia yang terdiri atas saham-saham berkapitalisasi besar dan berlikuiditas tinggi. Karakteristik return saham yang tidak selalu simetris, volatil, dan memiliki ekor distribusi tebal menyebabkan pendekatan optimasi portofolio berbasis mean-variance kurang mampu menggambarkan risiko kerugian ekstrem secara menyeluruh. Penelitian ini bertujuan untuk menganalisis distribusi return saham IDX30, mengukur risiko ekstrem menggunakan Conditional Value-at-Risk (CVaR), menentukan komposisi bobot portofolio melalui Hierarchical Risk Parity (HRP), serta mengevaluasi kinerja portofolio berdasarkan Adjusted Sharpe Ratio dan Maximum Drawdown. Data yang digunakan berupa harga penutupan harian dan volume perdagangan 15 saham IDX30 selama periode Januari 2019 hingga Februari 2026. Data harga penutupan ditransformasikan menjadi log return. Estimasi kuantil return dilakukan menggunakan Quantile Regression Forest (QRF) pada kuantil Q(0,10), Q(0,05), dan Q(0,01) sebagai dasar pengukuran Value-at-Risk dan CVaR pada tingkat kepercayaan 90%, 95%, dan 99%. Hasil penelitian menunjukkan bahwa return saham IDX30 memiliki rata-rata di sekitar nol, nilai skewness yang menunjukkan ketidaksimetrian distribusi, serta nilai kurtosis lebih besar dari tiga yang mengindikasikan distribusi leptokurtik dan heavy-tailed. Nilai CVaR meningkat seiring kenaikan tingkat kepercayaan. CVaR tertinggi pada tingkat kepercayaan 90% terdapat pada ADRO sebesar 0,055722, sedangkan CVaR tertinggi pada tingkat kepercayaan 95% dan 99% terdapat pada BBCA masing-masing sebesar 0,073539 dan 0,121439. Hasil pembentukan portofolio seluruh saham menunjukkan bahwa HRP-CVaR 90% menghasilkan return tahunan tertinggi sebesar 9,1656%, Adjusted Sharpe Ratio sebesar 0,2083, dan Maximum Drawdown sebesar -25,5319%. Evaluasi kombinasi portofolio menunjukkan bahwa Kombinasi 4 dengan HRP-CVaR 99% menghasilkan return tahunan tertinggi sebesar 34,0358%, Adjusted Sharpe Ratio sebesar 0,9565, dan Maximum Drawdown sebesar -36,7440%.
================================================================================================================================================
The IDX30 index is one of the major stock indices in the Indonesian capital market, consisting of large-cap and highly liquid stocks that are widely used as a benchmark in portfolio construction. However, the return characteristics of IDX30 stocks, which are asymmetric, volatile, and heavy-tailed, make conventional mean-variance portfolio optimization less effective in capturing extreme downside risk. This study aims to optimize an IDX30 stock portfolio by integrating Quantile Regression Forest (QRF) as a quantile-based return estimation method, Conditional Value-at-Risk (CVaR) as a measure of extreme risk, and Hierarchical Risk Parity (HRP) as a risk-based portfolio weighting method. The data used consist of daily closing prices and trading volumes of 15 stocks that consistently belonged to the IDX30 index during the period from January 2019 to February 2026. The closing price data were transformed into logarithmic returns, while QRF was used to estimate return quantiles at Q(0.10), Q(0.05), and Q(0.01), which were then used as the basis for calculating Value-at-Risk and CVaR at the 90%, 95%, and 99% confidence levels. The results show that IDX30 stock returns have mean values close to zero, skewness values indicating distributional asymmetry, and kurtosis values greater than three, indicating leptokurtic and heavy-tailed distributions. The CVaR values increase as the confidence level rises. The highest CVaR at the 90% confidence level is found in ADRO at 0.055722, while the highest CVaR values at the 95% and 99% confidence levels are found in BBCA at 0.073539 and 0.121439, respectively. The full-stock portfolio evaluation shows that HRP-CVaR 90% produces the highest annual return of 9.1656%, with an Adjusted Sharpe Ratio of 0.2083 and a Maximum Drawdown of -25.5319%. In the portfolio combination evaluation, Combination 4 with HRP-CVaR 99% produces the highest annual return of 34.0358%, with an Adjusted Sharpe Ratio of 0.9565 and a Maximum Drawdown of -36.7440%.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Adjusted Sharpe Ratio, Conditional Value-at-Risk, Hierarchical Risk Parity, IDX30, Quantile Regression Forest
Subjects:	Q Science Q Science > Q Science (General) Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. T Technology > T Technology (General) T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T57.6 Operations research--Mathematics. Goal programming
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Actuaria > 94203-(S1) Undergraduate Thesis
Depositing User:	Tri Ega Setyaningtyas
Date Deposited:	16 Jul 2026 07:17
Last Modified:	16 Jul 2026 07:17
URI:	http://repository.its.ac.id/id/eprint/135187

Actions (login required)

View Item