USING PCA LOADINGS AS PORTFOLIO WEIGHTS: EVIDENCE FROM THE SAUDI EQUITY MARKET (TADAWUL)

Abstract

Portfolio construction lies at the core of modern investment management, providing a systematic framework for allocating capital across assets to optimize returns relative to risk. Traditional portfolio construction approaches, such as equal weighting or capitalization weighting, fail to adequately account for these complex interrelationships. Even conventional optimization techniques may struggle when the covariance structure among sectors is unstable or when the number of sectors approaches the number of historical observations, leading to estimation errors and suboptimal outcomes (Michaud, 1989). To address these challenges, researchers and practitioners increasingly employ data-driven techniques such as Principal Component Analysis (PCA) to extract latent factors that explain the co-movement of asset returns (Connor & Korajczyk, 1986; Pástor & Stambaugh, 2003). The primary contributions of this study are threefold. First, it enhances the understanding of sectoral dynamics within the Saudi equity market by identifying latent economic factors driving sector co-movements. Second, it demonstrates how PCA loadings can be transformed into actionable portfolio weights, offering a transparent and economically grounded alternative to conventional optimization methods. Third, it provides empirical evidence on the risk-return trade-offs of PCA-weighted portfolios relative to traditional strategies, highlighting the practical benefits and limitations of this approach.