Factor-based index tracking through the identification of leading stocks

Abstract: This paper explores innovative approaches to financial index tracking through the identification of leading stocks using factor models. The primary objective is to introduce a novel, cost-effective method for partial index replication by selecting stocks that exhibit behavior closely aligned with the overall market index. To this end, the proposed model employs Principal Component Analysis (PCA) to estimate latent factors driving asset returns and to identify top-performing stocks. Additionally, a hybrid trading strategy combining momentum and contrarian elements is implemented to enhance portfolio returns relative to the benchmark. The leading-stock-based tracking method not only achieves high accuracy in index replication but also significantly reduces transaction costs and liquidity risk. Simulation results demonstrate that the proposed model outperforms conventional index tracking techniques in terms of minimizing tracking error and generating excess returns. The methodology is applied to data from the Tehran Stock Exchange, encompassing 318 listed companies from the beginning to the end of the year 1400 (2021–2022), and the findings confirm the model’s effectiveness in replicating the overall market index. This approach offers portfolio managers a practical tool for constructing low-cost, low-risk portfolios with index-like performance.

Keywords: Index Tracking, Factor Models, Leading Stocks, Momentum and Contrarian Strategies, Principal Component Analysis (PCA)