Predictive Analytics in Portfolio Management: A Fusion of AI and Investment Economics for Optimal Risk-Return Trade-Offs

Abstract

Portfolio management has become an essential area of study because of the introduction of artificial intelligence in managing portfolios in a given financial market where risks and returns are dynamic. This study aims to identify the extent of the adoption of AI in portfolio management and investigate the performance of AI models over conventional economic models. The study utilizes both quantitative survey questionnaires and qualitative interviews. The conceptual frameworks refer to the Modern Portfolio Theory, Behavioral Finance, Technology Acceptance Model, and Digital Governance framework. The qualitative results also indicate that portfolio managers understand the importance of using AI to improve portfolio optimization, risk management data, and predictive statistics. However, they also have critical explicit aspects about the data quality, interpretability/understanding of the model, and ethical issues. The quantitative analysis aims to investigate the influence of AI integration, behavioral biases, and perceived ease of use on the effectiveness of AI in portfolio management. AI implementation has a large influence on portfolio performance where the behavioral biases moderate adoption (Coefficients = 0.0075 and direct impact = 0.1235). Digital governance alters the interaction, enhancing the influence of AI (Effect = 0.5215; P < 0.001). Overall, AI enhances decision-making and portfolio selection, but concerns remain about data quality and technology application. Digital governance and human input are crucial for optimizing AI use in finance, despite concerns about data quality.