In a future of widespread algorithmic pricing, inadvertent cooperation between algorithms is easier than ever, resulting in coordinated price rises. While governments have clear cut laws to assess and deter traditional collusion, automated pricing poses interesting questions as to what will or will not be allowed.
We build and demonstrate how a Q-learner algorithm, even under extremely innocent conditions, can reach a collusive outcome in a virtual marketplace. We also discuss which industries are in the most danger of finding themselves subject to greater restrictions or scrutiny, where “safe harbours” might exist, and what future digital regulation might look like.
The audience will understand:
You can view the slides below: