<![CDATA[Investment in the Indonesian capital market has grown significantly, surpassing 18 million investors as of August 2025. This study compares five portfolio optimization methods—Classical Mean–Variance, Fast Minimum Covariance Determinant (FMCD), Robust S-Estimator, Robust Constrained M (CM) Estimator, and Mean–Value at Risk (Mean–VaR)—using LQ100 constituent stocks. Daily closing data from January 2023 to December 2024, selecting ten stocks with the highest Sharpe ratios to construct the portfolio. Each model was optimized under various levels of risk aversion and evaluated through backtesting from January to August 2025. Using an initial capital of Rp 100 million, the results indicate that while robust estimators such as FMCD and CM provide greater stability during market volatility, the Classical Mean–Variance model with moderate risk aversion (γ = 25) yields the most profitable and well-diversified portfolio composition, with the largest allocations in AMMN.JK (20.36%), BSSR.JK (19.65%), and JPFA.JK (9.22%). The backtesting results in a total projected profit of approximately Rp 13.7 million over eight months. These findings confirm that the Classical Markowitz framework remains a reliable and efficient approach for portfolio allocation in the Indonesian stock market, especially for moderately risk-averse investors seeking a balance between diversification and return stability.]]>
Copyrights © 2026
