We study policy experimentation in organizations with endogenous membership. An organization initiates a policy experiment and then decides when to stop it based on its results. As information arrives, agents update their beliefs, and become more pessimistic whenever they observe bad outcomes. At the same time, the organization’s membership adjusts endogenously: unsuccessful experiments drive out conservative members, leaving the organization with a radical median voter. We show that there are conditions under which the latter effect dominates. As a result, policy experiments, once begun, continue for too long. In fact, the organization may experiment forever in the face of mounting negative evidence. This result provides a rationale for obstinate behavior by organizations, and contrasts with models of collective experimentation with fixed membership, in which under-experimentation is the typical outcome. Our results extend to settings with heterogeneous payoffs and settings with heterogeneous levels of membership (e.g., shareholders of a firm).

Link to paper: scholar.princeton.edu/sites/default/files/germang/files/experimentationpaper-1019_01.pdf

Please sign up for meetings here: docs.google.com/spreadsheets/d/1WTp83AxJaVGv6sMlwBer43q68i43dAe9BSeyb26ypa4/edit#gid=0