علی ارشدی

Stocks are fundamental instruments in capital markets, and understanding their interdependencies is crucial for effective investment and risk management. This study adopts a network-based framework to examine the structure and dynamics of the U.S. stock market by constructing correlation networks through analytical methods rather than purely data-driven approaches. Specifically, the Minimum Spanning Tree (MST) and Planar Maximally Filtered Graph (PMFG) techniques are employed to model the relationships among 371 constituent stocks of the S&P 500 index, using daily closing prices from January 11, 2013, to December 12, 2022. The key empirical findings indicate that MST and PMFG effectively capture the hierarchical organization of financial markets. These network structures facilitate the identification of stock clusters and central nodes, providing insights into systemic risk and potential contagion during periods of financial turbulence. Furthermore, the identified network topology supports enhanced portfolio diversification by enabling the selection of stocks across different clusters or based on centrality metrics. Temporal analysis of the network evolution also reveals shifts in market conditions and sectoral rotations. Additionally, the study integrates Markowitz’s portfolio optimization framework by applying the Sharpe ratio as the performance criterion. The resulting optimized portfolio achieves an expected annual return of 37.61%, with a volatility of 14.15%, yielding a Sharpe ratio of 2.6581 demonstrating robust risk-adjusted performance. In conclusion, MST and PMFG offer valuable tools for capturing market structure, informing portfolio construction, and enhancing risk assessment, thereby contributing to more resilient investment strategies.