Incorporating cognitive support in investment DSS

Abstract

<p>The financial well-being of individuals depends on the appropriateness of their investment decisions. As the world is moving from defined benefit (DB) pension plans to defined contribution (DC) pension plans, the burden of making the right investment decisions has now shifted to employees. One implicit assumption about this global trend is that plan participants are capable of making sound decisions themselves. However, several studies have demonstrated that people often fall prey to various psychological biases and make flawed investment decisions.</p> <p>The use of computers in support of decision-making is not new. For several decades, the finance and investment sectors have extensively used different computing methods and technologies collectively known as decision support systems (DSS) to support investors. However, the type of support provided by such DSS has primarily been quantitative in nature. As such, existing investment DSS do not assist decision makers in overcoming the impact of their psychological biases, which have been shown to play a critical role in investment decision-making. The overall objective of this research is to investigate the potential for building a human-centered investment DSS that can provide qualitative support to investors.</p> <p>In this research, a theoretical framework of investment decision-making is proposed by using the psychological concepts of beliefs, preferences and attitudes. Major investment-related biases are identified and a taxonomy is suggested to classify them as cognitive, affective, and conative.</p> <p>An empirical study involving 119 subjects was conducted to verify the impact of cognitive biases in investment decision-making and to assess the effectiveness of decision aids in lowering the impact of such biases on the ability of investors to make sound investment decisions. In this study, feedback and graphical aids were provided as cognitive support in six investment decision-making tasks involving framing, representativeness, and ambiguity biases. A large majority of subjects exhibited the influence of these biases in their investment decision-making. Cognitive support was found to significantly improve asset allocation decisions for most subjects. Demographic variables collected during the experiment enabled several analyses leading to some additional interesting observations. Findings from this study indicate the usefulness of personalization in investment DSS.</p> <p>This research culminated with a vision toward the development of a human-centered investment DSS that may provide qualitative support to its users. Different philosophical inquiring systems were described as potential debiasing strategies for the proposed DSS. An architecture was suggested for implementing such a DSS with a detailed example illustrating the feasibility of the proposed system. The dissertation concludes with an outline of potential contributions of this study and directions for future research in this area.</p>