OPTIMIZING RETURNS IN A CHALLENGING MARKET: AN APPLICATION OF THE MEAN-VARIANCE MODEL ON THE INDONESIAN LQ45 INDEX

Portfolio Optimization Under Varying Market Regimes: A Comparative Study Using ANN Forecasts on MAI Stocks

This paper explores the effectiveness of integrating Artificial Neural Network (ANN) based return forecasts into three portfolio optimization frameworks: Equal-Weight (EW), Mean-Variance (MV), and Black-Litterman (BL), under varying market regimes, using 67 stocks listed on Thailand’s Market for Alternative Investment (MAI) as a case study. Portfolios are constructed using ANN predictions and tested across pre-COVID stable conditions (2019) and a volatile period (2020–2024), with an additional evaluation of rebalanced portfolios on 2024 data. Results indicate that BL achieves the highest risk-adjusted returns under stable conditions, while EW outperforms significantly during high volatility, driven largely by outlier stock performance during the COVID-19 recovery. Rebalancing with updated ANN forecasts did not guarantee improved performance, highlighting both the promise and limitations of machine learning-enhanced optimization in emerging markets.

Evaluation of individual stock portfolio performance on the Indonesia stock exchange using the Ortiz, Jorion, and Markowitz approaches

Background: Indonesia's capital market has experienced a sustained increase in investor participation, creating a stronger need for systematic and implementable portfolio construction methods. This study evaluates estimation risk in mean-variance optimization by comparing traditional Markowitz optimization with two mean-shrinkage approaches: the parameter-focused Bayes-Stein estimator of Jorion and the decision-focused optimal shrinkage of means proposed by Ortiz et al. Methods: The study uses monthly individual stock data from the Indonesia Stock Exchange over January 2006-December 2025. Excess returns are calculated relative to a monthly risk-free proxy, and portfolios are evaluated using a 120-month rolling-window out-of-sample backtest under long-only constraints. The sample is a balanced panel of 74 stocks with complete monthly data, and the monthly deposit insurance rate of the Indonesia Deposit Insurance Corporation (LPS) is used as the risk-free proxy. Portfolio performance is assessed using monthly and annualized Sharpe ratios, while weight stability is assessed using average weight volatility and turnover. Newey-West tests are used to evaluate whether differences across methods are statistically significant. Findings: The Ortiz approach consistently selects an optimal shrinkage intensity of zero, making its weights and performance effectively identical to the traditional Markowitz portfolio. Markowitz and Ortiz record an annualized out-of-sample Sharpe ratio of 0.0713, while Jorion records 0.0608. The statistical tests indicate that differences in out-of-sample performance and stability are not significant across the three approaches. In economic terms, the annualized Sharpe ratios of all three methods are very low (below 0.08), indicating that long-only optimization of individual Indonesian stocks delivered only marginal risk-adjusted excess returns over the sample period. Conclusion: In the Indonesian stock market setting, more complex mean-shrinkage methods do not automatically produce superior portfolio outcomes. Novelty/Originality of this article: This article provides Indonesian market evidence on the comparison between parameter-focused and decision-focused shrinkage approaches within a consistent rolling-window portfolio backtesting framework.

REASSESSING TECHNICAL ANALYSIS IN EMERGING MARKETS: A 20-YEAR STUDY OF MECHANICAL TRADING RULES AND MARKET EFFICIENCY ON THE LQ45 INDEX

This study conducts an empirical re-evaluation of the effectiveness of technical analysis within the framework of the Efficient Market Hypothesis (EMH) in an emerging market setting. Utilizing a comprehensive two-decade daily dataset (2006–2025) of the Indonesian LQ45 Index, the research simulates a mechanical trading strategy based on 20-day and 50-day Exponential Moving Average (EMA) crossovers, contrasting its performance against a passive Buy-and-Hold (B&H) benchmark. To eliminate subjective bias and ensure methodological rigor, the analysis employs continuous log-returns and assumes a frictionless market. Statistical inference, performed using Welch's independent samples t-test, indicates no significant difference in raw returns between the active EMA strategy and the passive benchmark (p-value = 0.959), thereby strongly supporting the weak-form EMH. Notably, the study reveals a sustained negative equity risk premium over the two decades, with both equity strategies yielding lower returns (6.07% and 6.42%, respectively) than the 7.99% risk-free rate of 10-year government bonds. Despite underperforming in absolute returns, the EMA strategy demonstrates exceptional risk mitigation capabilities, drastically reducing total portfolio volatility from 23.65% to 15.39% and compressing downside deviation from 18.72% to 15.05%. Evaluated through the Modified Sharpe and Sortino ratios, these findings redefine the utility of moving averages: while technical analysis fails to maximize profit or generate anomalous returns, it functions as a highly effective portfolio insurance mechanism, capable of mitigating downside risk during market downturns.

An Empirical Study of Robust Portfolios Based on Statistical Noise Reduction and Regularization Constraints

This research paper discusses the issue of out-of-sample robustness of portfolio optimization within a complex market environment. Within the mean-variance framework, we compare four methods for covariance estimation and weighting: a baseline model based on sample covariances, the Ledoit-Wolf shrinkage estimator, Garber's robust covariance estimator, and a minimum-variance model incorporating an L2 regularization term. This study uses data from 60 American stocks from 2019 to 2021, with four rebalancing cycles (5, 20, 60, and 100 days) were established, and a moving-window backtesting method was applied to evaluate the performance of the various methods in terms of returns, risk, and out-of-sample stability. The empirical results indicate that the L2 normalization model generates higher cumulative returns and Sharpe ratios during most rebalancing periods, demonstrating effective stabilization of the weights; Garbner's robust covariance technique demonstrates consistent resilience to downturns in high-volatility market environments; And Reduit-Wolf's contraction estimates also show a relatively stable improvement in sample skewness compared to the benchmark model. These results suggest that including a covariance noise or normalization term in the traditional expected value and variance-based approach to portfolio management helps reduce the impact of estimation error and market volatility on portfolio performance. The period of analysis for this paper is specified. Furthermore, it covers the observation period, highlights limitations related to market comprehensiveness and parameter selection, and provides guidance for future research by incorporating financial data from diverse sources and more advanced robust optimization techniques.

An Analysis of LQ45 Index Volatility Using the GARCH Model and Its Implications for Retail Investment Risk during the 2020-2025 Period

A Theoretical and Literature-Based Comparison of the Markowitz Model and the Index Model

This paper presents a comprehensive review of two foundational portfolio models in modern finance: the Markowitz mean-variance model and the Sharpe single-index model. The study systematically examines their theoretical frameworks, computational requirements, empirical performance, and practical applicability. The Markowitz model, as a model-free optimisation framework, aims to achieve theoretically optimal risk-return trade-offs by fully considering pairwise covariances among assets. However, its high sensitivity to estimation errors and intensive data requirements often limit its empirical usefulness, particularly for large portfolios. In contrast, the index model simplifies covariance estimation by attributing asset return variations to a common market factor, thereby reducing parameter dimensionality and enhancing computational efficiency. Multi-factor extensions further improve explanatory power while maintaining practical feasibility. By reviewing existing literature and empirical evidence, this paper highlights the strengths, limitations, and appropriate application scenarios of both models. The findings provide guidance for investors, portfolio managers, and researchers in selecting or integrating portfolio models to balance theoretical rigor with empirical robustness and operational efficiency.