Perbandingan Mean-Value At Risk Dan Mean-Entropic Value At Risk Pada Optimasi Portofolio Utang Valuta Asing Pemerintah Indonesia

A`in, Zalfa Naila Qurrotul (2026) Perbandingan Mean-Value At Risk Dan Mean-Entropic Value At Risk Pada Optimasi Portofolio Utang Valuta Asing Pemerintah Indonesia. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5002221120-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Posisi pinjaman luar negeri pemerintah tahun 2024 sangat didominasi oleh mata uang USD, yaitu mencapai USD 104.551 miliar dari total USD 203.144 miliar. Dominasi yang tinggi ini meningkatkan kerentanan portofolio utang terhadap fluktuasi nilai tukar dan suku bunga valas, sehingga diperlukan strategi diversifikasi yang pruden. Untuk menjawab tantangan alokasi mata uang tersebut, penelitian ini mengadopsi model Mean-Variance dengan formulasi Value at Risk (VaR) dan Entropic Value at Risk (EVaR) sebagai ukuran risiko yang lebih superior. Simulasi optimasi portofolio dilakukan pada ketiga model tersebut dengan memanfaatkan data historis suku bunga dan nilai tukar dari DJPPR Kementerian Keuangan periode 24 November 2020 hingga 24 November 2025. Implementasi model dieksekusi melalui pemrograman konveks pada MATLAB menggunakan algoritma SQP untuk menghasilkan komposisi optimal pada dua skenario utama: Skenario 2 (tanpa batasan alokasi) dan Skenario 3 (rentang modifikasi proporsi existing). Hasil pengujian pada Skenario 3 menunjukkan bahwa seluruh model mengalami konvergensi alokasi instrumen yang didominasi oleh USD 5Y (47.90%), USD 30Y (16.80%), USD 10Y (13.40%), dan EUR 7Y (13.4006%). Struktur ini berhasil menekan ekspektasi nilai utang hingga 0.0000458823, jauh lebih efisien dibandingkan portofolio Existing 2024 (0.0000582900). Sementara itu, pada Skenario 2, ketiadaan batasan alokasi membuat seluruh model memusatkan 100% porsi pada instrumen USD 30Y dengan ekspektasi nilai utang terendah secara mutlak sebesar 0.0000170546. Ditinjau dari karakteristik risiko, model Mean-EVaR secara konsisten menghasilkan estimasi nilai risiko yang paling besar dibandingkan model Mean-VaR dan Mean-Variance pada seluruh skenario (mencapai 0.00267573 pada Skenario 3 dan 0.00255201 pada Skenario 2). Berdasarkan hasil yang didapat, Skenario 3 memberikan rekomendasi alokasi proporsional yang paling optimal dan realistis untuk diterapkan, sedangkan model Mean-EVaR terbukti menghasilkan estimasi risiko yang paling tinggi dibandingkan model Mean-VaR dan Mean-Variance.
======================================================================================================================================
The position of the government’s foreign debt in 2024 was heavily dominated by the USD, reaching USD 104.551 billion out of a total of USD 203.144 billion. This high dominance increases the debt portfolio’s vulnerability to exchange rate and foreign interest rate fluctuations, necessitating a prudent diversification strategy. To address this currency allocation challenge, this study adopts the Mean-Variance model with Value at Risk (VaR) and Entropic Value at Risk (EVaR) formulations as superior risk measures. Portfolio optimization simulations were conducted across the three models using historical interest rate and exchange rate data from the Directorate General of Budget Financing and Risk Management (DJPPR), Ministry of Finance, for the period of November 24, 2020, to November 24, 2025. Model implementation was executed via convex programming in MATLAB using the SQP algorithm to produce optimal compositions under two primary scenarios: Scenario 2 (unconstrained allocation) and Scenario 3 (modified range of existing proportions). The test results for Scenario 3 indicate that all models exhibited convergence in instrument allocation, dominated by USD 5Y (47.90%), USD 30Y (16.80%), USD 10Y (13.40%), and EUR 7Y (13.4006%). This structure successfully reduced expected costs to 0.0000458823, which is significantly more efficient than the 2024 Existing portfolio (0.0000582900). Meanwhile, in Scenario 2, the absence of allocation constraints caused all models to concentrate 100% of the allocation on the USD 30Y instrument, achieving the lowest absolute expected cost of 0.0000170546. In terms of risk characteristics, the Mean- EVaR model consistently generated the highest risk estimates compared to the Mean- VaR and Mean-Variance models across all scenarios (reaching 0.00267573 in Scenario 3 and 0.00255201 in Scenario 2). Based on the results obtained, Scenario 3 provides the most optimal and realistic proportional allocation recommendation to implement, whereas the Mean-EVaR model is proven to generate the highest risk estimation compared to the Mean-VaR and Mean-Variance models.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Optimasi Portofolio, Utang Valuta Asing, Mean-Variance Optimization, Value-at-Risk (VaR), Entropic Value-at-Risk (EVaR). Portfolio Optimization, Foreign Currency Debt, Mean-Variance Optimization, Value-at-Risk (VaR), Entropic Value-at-Risk (EVaR).
Subjects:	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) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Zalfa Naila Qurrotul A`in
Date Deposited:	30 Jul 2026 04:30
Last Modified:	30 Jul 2026 04:30
URI:	http://repository.its.ac.id/id/eprint/139888

Actions (login required)

View Item