Disentangling Exploration from Exploitation

A key tension in the study of experimentation revolves around the exploration of new possibilities and the exploitation of prior discoveries. Starting from Robbins (1952), a large literature in economics and statistics has married the two: Agents experiment by selecting potentially risky options and observing their resulting payoffs. This framework has been used in many applications, ranging from pricing decisions to labor market search. Nonetheless, in many applications, agents' exploration and exploitation need not be intertwined. An investor may study stocks she is not invested in, an employee may explore alternative jobs while working, etc. The current paper focuses on the consequences of disentangling exploration from exploitation. This talk will cover some insights generated from work joint with Alessandro Lizzeri (Princeton University) and Eran Shmaya (Stony Brook University). We consider the classical Poisson bandit problem that has served as the canonical model for experimentation. We fully characterize the solution when exploration and exploitation are disentangled, both for the "good news" and "bad news" settings. We illustrate the stark differences the optimal exploration policy exhibits compared to the standard setting. In particular, we show that agents optimally utilize the option to observe projects different than the ones they act on. In the good news case, the optimal policy entails the continued exploration of a singular arm-no matter how pessimistic the decision-maker becomes about that arm-until news arrives. In contrast, in the bad news, exploration can involve the use of more than a single arm, but entails at most one switch. In all settings, the separation of exploration from exploitation guarantees asymptotic efficiency. BIO: Leeat Yariv is the Uwe E. Reinhardt Professor of Economics at Princeton University. She is also the director of the Princeton Experimental Laboratory for the Social Sciences (PExL), which she opened. She is the lead editor of AEJ: Micro and has served on the editorial boards of multiple journals. She is a member of the American Academy of Arts and Sciences and a fellow of the Econometric Society and the Society for the Advancement of Economic Theory. She is also a research associate of the National Bureau of Economic Research (NBER) and a research fellow of the Center for Economic and Policy Research (CEPR). Yariv's work focuses on market design, social networks, and political economy. She uses theory, lab experiments, and field studies to understand how individuals connect to one another and how they make decisions, on their own and collectively.