Abstraction and Intermediation: The Implications of Behavioral Decision Making for Three Applications in Finance

Modern financial theory and practice make extensive use of the concepts of abstraction and intermediation. Securities are designed, valued, and allocated within portfolios based on characteristics that contain little of the richness of the actual assets underlying their claims. Corporate financial decisions and policy analyses evaluate choices based on comparing abstract metrics that ignore or minimize the myriad attributes belonging to such choices. In many cases, these evaluations and choices are made by intermediaries (whether human experts or heuristic rules) that either offer specialized information or serve as a proxy for competitive markets. While economic theory has addressed issues in abstraction and intermediation that arise from incentive problems and information asymmetries directly, most approaches have maintained the traditional assumption of rational agents free of cognitive constraints and biases. This dissertation illustrates, through three applications, that abstraction and intermediation are at significant risk of suboptimal operation when faced with certain limitations prevalent in the behavioral decision-making literature.

The three diverse applications include the design of financial securities, the ratemaking practices of utility regulators, and the use of metrics and heuristics by intermediaries in corporate finance and policy analysis. Well-established areas in the behavioral decisionmaking literature, such as risk perception, communication about uncertainty, and heuristic choice and expert judgment are shown to challenge the application of these core principles of economics in practice. This dissertation develops a methodology for evaluation and demonstrates the practical consequences of behavioral decision-making in each of the three applications and examines possible responses and interventions to mitigate their effect.