Template-Type: ReDIF-Paper 1.0
Series: Tinbergen Institute Discussion Papers
Creation-Date: 2018-06-21
Revision-Date: 2020-11-01
Number: 18-056/VII
Author-Name: Timo Klein
Author-Email: t.klein@uva.nl
Author-Workplace-Name: University of Amsterdam
Title: Autonomous Algorithmic Collusion: Q-Learning Under Sequantial Pricing
Abstract: Prices are increasingly set by algorithms. One concern is that intelligent algorithms may learn to collude on higher prices even in absence of the kind of communication or agreement necessary to establish an antitrust infringement. However, exactly how this may happen is an open question. I show in a simulated environment of sequential competition that competing reinforcement learning algorithms can indeed learn to converge to collusive equilibria. When the set of discrete prices increases, the algorithm considered increasingly con- verges to supra-competitive asymmetric cycles. I show that results are robust to various extensions and discuss practical limitations and policy implications.
Classification-JEL: L13, L41, D43, D83
Keywords: Algorithmic Collusion, Pricing Algorithms, Machine Learning, Reinforcement Learning, Q-Learning
File-URL: https://papers.tinbergen.nl/18056.pdf
File-Format: application/pdf
File-Size: 616393 bytes
Handle: RePEc:tin:wpaper:20180056