Portfolio optimization with hierarchical risk parity and multi-criteria decision making: BIST 100 application

To secure a comfortable future and effectively maximize their savings, investors seek to build optimal portfolios by carefully selecting the best investment options available in the market. For more than fifty years, investors have relied on the Mean Variance Model (MVM) proposed by Markowitz in 1952 to construct these portfolios. However, numerous studies have criticized the MVM from various angles. One important point of criticism is the concentration problem. Although the MVM is based on diversification, it often recommends investing in a limited number of assets. This limited approach can lead to financial losses for investors. To address this problem, Prado introduced the Hierarchical Risk Parity (HRP) algorithm in 2016, which integrates concepts from graph theory and machine learning. While HRP effectively solves the concentration problem, it introduces new challenges, such as the problem of concatenation and difficulties in determining the optimal number of clusters. To overcome these challenges, this study proposes an HRP approach to multi-criteria decision making (MCDM). The proposed approach incorporates the MEREC and WEDBA methods. The MEREC method, an objective approach to weighting criteria, was used to assign appropriate weights to the criteria. At the same time, the WEDBA method was used to evaluate the inventories. This approach was applied to the stocks listed in the BIST 100 Index. The results showed that the HRP-MCDM approach provided more successful results compared to the traditional HRP algorithm, thus providing a more robust solution to the problems previously encountered.