Examining the role of recognition-based heuristics in sustainable investment decision-making using a structural equation modeling–artificial neural network-based approach

Understanding the factors influencing sustainable investment decision-making (SIDM) is important for promoting environmentally and socially responsible investment. This study examines the effect of recognition-based heuristics—specifically alphabetical order (ALPBORD), name fluency (NAMFLN), and name memorability (NAMMEM)—on the sustainable investment decision-making (SIDM) of individual investors trading on the Hong Kong Stock Exchange. Data were obtained through a survey of 369 individual investors using a combination of convenience purposive and snowball sampling approaches. A dual-stage methodological approach combining structural equation modeling (SEM) and an artificial neural network (ANN) was used to capture both linear connections and complex hidden nonlinear patterns in the relationship between variables. The SEM results indicate that the heuristics NAMFLN, ALPBORD, and NAMMEM all have a significant negative impact on SIDM, suggesting that reliance on these heuristics can distort investor judgment and lead to suboptimal investment decisions in environmental, social, and governance terms. ANN analysis further demonstrates the relative contribution of each heuristic to undermining sustainable investment behavior. Specifically, NAMFLN has the highest predictive power, followed ALPBORD and NAMMEM. These findings suggest that investors tend to favor firms with names that are easily recognized or memorable, or those appearing early in alphabetical listings, which may divert capital away from genuinely sustainable firms and result in less optimal ESG-aligned portfolios. Overall, this study advances the literature by demonstrating the vital role of recognition-based heuristics in shaping SIDM, providing empirical evidence of the behavioral mechanisms affecting sustainability-aligned investment, and pioneering the application of a hybrid SEM–ANN approach in this context.