Utility Maximization with Behavioral Constraints in E-Commerce

The boom of e-commerce and live-stream shopping show that the classical utility models become inadequate due to the breakdown of rationality and constant preferences. In practice, online consumers follow known-to-be irrational purchasing behaviors that are fast, emotional, time-pressured, and influenced by social influence. In this paper, we develop a behavioral utility maximization model that captures three behavioral constraints: loss aversion, social comparison, and dynamic pricing. The model generalizes the classical Lagrangian model by considering time-varying elasticity and psychological constraints of loss aversion and mental accounting. We estimate the model using both survey data and simulated shopping sessions and compare it against the Cobb-Douglas baseline. The model has a better fit to the data in terms of decision-making, i.e., explaining quicker responses when facing countdown timers and explaining higher willingness to pay when there are peer activities/influencer endorsements. These results provide strong evidence that psychology and social influence significantly impact consumer decisions. The contributions of this research are dual. On the theoretical side, this paper extends utility theory by introducing behavioral realism. On the applied side, our research reveals the decisions’ designs of platforms and raise ethical issues that whether it is right to manipulate consumers. This research shows that researching digital consumer behavior requires inputs from both economics, psychology and computation.