The Causal Relationship between Managers’ and Investors’ Cognitive Biases and the Performance of the Iranian Stock Market: A Mixed-Methods Approach

Abstract

This study aimed to develop and test a causal model explaining the relationships among managers’ and investors’ cognitive biases, decision-making quality, artificial intelligence-based bias-reduction mechanisms, and the performance of the Iranian stock market. This developmental study used a qualitative-dominant sequential exploratory mixed-methods design. In the qualitative phase, 12 university professors, behavioral-finance researchers, market analysts, managers, and experienced investors were selected purposively and interviewed through in-depth semi-structured interviews until thematic saturation. The interview data were analyzed through thematic analysis using MAXQDA, and the extracted dimensions and indicators were refined through a Delphi process. In the quantitative phase, 200 managers, investors, financial analysts, consultants, and decision-makers active in the Iranian capital market were selected through random sampling. Data were collected using a researcher-developed questionnaire derived from the qualitative findings. Reliability and validity were examined using Cronbach’s alpha, composite reliability, and average variance extracted. Structural relationships were tested using structural equation modeling and bootstrapping. Fundamental cognitive biases significantly reduced decision-making quality (β = −0.42, t = 5.25, p < 0.001). Investment and financial biases (β = −0.40, t = 4.44, p < 0.001), informational and analytical biases (β = −0.31, t = 3.88, p < 0.001), narrative biases (β = −0.28, t = 3.11, p = 0.002), organizational and cultural biases (β = −0.25, t = 3.57, p < 0.001), and bias-related consequences (β = −0.29, t = 3.63, p < 0.001) also had significant negative effects. Artificial intelligence applications had the strongest positive effect on decision-making quality (β = 0.46, t = 5.75, p < 0.001), followed by advanced artificial-intelligence mechanisms (β = 0.38, t = 5.43, p < 0.001), bias-reduction strategies (β = 0.36, t = 5.14, p < 0.001), institutional strategies (β = 0.32, t = 4.57, p < 0.001), and organizational learning (β = 0.27, t = 3.38, p = 0.001). Decision-making quality significantly improved stock-market performance (β = 0.64, t = 9.21, p < 0.001). Cognitive biases impair managerial and investment decision-making and indirectly weaken stock-market performance, whereas artificial intelligence, systematic debiasing, institutional controls, and organizational learning improve decision quality and market outcomes.