Recent work by Assistant Professor of Economics Erik Lillethun shows important and interesting findings across behavioral economics, cybersecurity strategy, and information design. The paper “Long-Run Choice Anomalies in Reinforcement Learning with Bounded Memory,” coauthored with Erin Giffin (Colby College) and published in the Journal of Economic Behavior & Organization, demonstrates that even unbiased learning models can generate persistent behavioral decision-making biases when agents have bounded memory. The study shows how individuals may favor actions yielding small, frequent rewards despite large, infrequent risks — helping explain phenomena such as the chronic under-purchase of insurance.
Another new publication, “Strategic Cyberwarfare,” coauthored with our own Associate Professor Economics Rishi Sharma and appearing in Economic Inquiry, develops a game-theoretic framework to study cyber conflict between attacking and defending nations. The analysis highlights how network connections between computer systems shape cyberwar outcomes, influencing national welfare, optimal defensive expenditures, and the value of centralized versus decentralized cybersecurity decision making.
Finally, “Information Design and Reputations in the Frequent-Interaction Limit,” now accepted at the Journal of Mathematical Economics, examines how consumers may actually benefit when quality information is hidden or partially obscured — for example, by a review platform — because this can encourage firms to exert higher effort and deliver better quality in equilibrium. It further distinguishes the best information policies from the perspective of consumers when a platform is free to design any policy versus when it is restricted to simpler and more robust policies based only on publicly observable information.