We show that if they are allowed enough time to complete the learning, Q-learning algorithms can learn collude in an environment with imperfect monitoring adapted from Green and Porter (1984), without having been instructed do so, communicating one another. Collusion is sustained by punishments take form of “price wars” triggered observation low prices. The have a finite duration, being harsher...

Standard benefit-cost analysis often ignores distortions caused by taxation and the heterogeneity of taxpayers. In this paper, we theoretically numerically explore effect imperfect on public provision mortality risk reductions (or safety). We show that critically depends source imperfection as well individual utility survival probability functions. Our simulations based calibration distribution...