Determining investorsʹ utility function to portfolio selec‎tion with novel approach in UTA-STAR method

2/21/2027 12:00:00 AM

This study aims to determine investors’ utility functions for portfolio construction by employing Expected Utility Theory through an innovative an‎d structured approach. The main objective is to develop a systematic framework for analyzing investor behavior according to different risk preferences an‎d for constructing optimal portfolios consistent with individual preferences. To this end, thirty stocks from the Dow Jones Industrial Average were selec‎ted based on expert opinions an‎d academic relevance. The selec‎ted stocks were clustered into three groups—risk-averse, risk-neutral, an‎d risk-seeking—based on the stan‎dard deviation of returns as a risk measure. This clustering process was carried out using machine learning techniques to identify groups of stocks with similar risk characteristics. Within each cluster, stock ranking was performed using the multi-criteria decision-making method ELECTRE III, considering five independent fundamental criteria: P/E ratio, Debt-to-Equity ratio, Asset Turnover, Current Ratio, an‎d Return on Equity (ROE). The weights of these criteria were determined using the DEMATEL method based on expert judgments. In the next stage, utility functions for three key criteria—return, risk, an‎d liquidity—were extracted for each category of investors using the UTASTAR method. This approach enables the derivation of group-specific utility functions an‎d integrates investor preferences with real market data. To construct optimal portfolios, a three-objective Markowitz optimization model was employed, simultaneously considering return maximization, risk minimization, an‎d liquidity maximization. The extracted utility values were incorporated into the model as parameters, an‎d the area under each utility function was used as the corresponding objective weight to transform the multi-objective problem into a single-objective formulation. Validation results indicate that portfolios constructed using the utility-based approach exhibit performance more consistent with investors’ risk preferences an‎d achieve superior realized returns compared to portfolios obtained from the classical Markowitz model. Overall, this research presents a practical an‎d theoretically grounded framework for investor behavior analysis, stock prioritization, an‎d portfolio optimization. The proposed approach can serve as an effective decision-support tool for financial advisors, portfolio managers, an‎d researchers, while also providing a foundation for future studies integrating expected utility theory with multi-criteria decision-making techniques.

تابع مطلوبيت , تئوري مطلوبيت انتظاري , خوشه‌بندي سهام , الكتره3 , مدل ماركويتز

Utility Function , Expected Utility Theory , Stock Clustering , ELECTRE III , Markowitz Model

Hatef Karimi Mandegi

Mohammadreza Mohammadaliha