FACTOR-BASED STOCK SELECTION AND MEAN-VARIANCE PORTFOLIO OPTIMIZATION: AN EMPIRICAL STUDY OF BISNIS-27

Abstract

Despite the significant growth of exchange-traded funds (ETFs) in Indonesia, the ETF market share remains very low, around 1% of total mutual fund industry assets under management, as most ETF products simply replicate the index and offer limited differentiated risk-adjusted returns. This paper proposes and empirically tests a factor investing framework combined with mean-variance optimization for portfolio construction on the Bisnis-27 Index constituents. First, all 27 constituent stocks are scored and ranked using a composite multi-factor model comprising growth, quality, value, momentum, and volatility factors, standardized with z-scores and equally weighted, from which the top 15 stocks are selected. Mean-variance optimization is then applied using Solver to determine the weights that maximize the portfolio's risk-adjusted return. The portfolio is evaluated in-sample over the observation period of November 2021 to April 2025 and compared with equal-weight, factor-score, and minimum-variance weighting, as well as the Bisnis-27 Index, using the Sharpe ratio, Treynor ratio, Jensen's alpha, and information ratio. The results show that factor-selected stocks outperform the benchmark and mean-variance optimization delivers the best risk-adjusted performance with a Sharpe ratio of 1.45, an annualized return of 25.57%, and an alpha of 23.96%. The findings indicate that a factor-based smart beta ETF strategy can be implemented in the Indonesian capital market