Loading...Loading...

Portfolio Selection via Principal Component Analysis

Piedboeuf, Clément

(2023)

Details

Supervisors

Lassance, Nathan

Faculty

Louvain School of Management

Degree label

Master [120] : ingénieur de gestion, à finalité spécialisée

Abstract

Portfolio Selection is a critical process for investors and portfolio managers seeking to optimise returns while managing their risks. We provide an in-depth analysis of four different portfolio optimisation strategies that uses principal components. The strategies are: the Bounded-Noise portfolio, the PC-Mean-Variance and PC-Minimum-Variance portfolios, the portfolios of Severini and finally the PC-Variance-Parity portfolio. In first instance, we compare the performance of the different strategies to each other. We have found that the Bounded-Noise portfolio performs significantly better than the other portfolios. But closely followed by the PC-Mean-Variance portfolio. We also saw that the second portfolio of Severini yield very poor performance. In second instance, we compare the PC-based strategies to the classical strategies. The Mean-variance portfolio, the Minimum-Variance portfolio, the equally weighted portfolio and the Asset-Variance-Parity portfolio. We saw that most of PC-based strategies outperformed the classical strategies at the exception of the second portfolio of Severini that performs significantly worse. Finally, we highlight the positive impact of an optimal selection of principal components on the performance of the PC-based strategies. Overall, the thesis contributes to the understanding of portfolio optimization strategies and provides recommendations for investors based on empirical analysis.