Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/185575 
Authors: 
Year of Publication: 
2018
Series/Report no.: 
Tinbergen Institute Discussion Paper No. TI 2018-056/VII
Publisher: 
Tinbergen Institute, Amsterdam and Rotterdam
Abstract: 
A novel debate within competition policy and regulation circles is whether autonomous machine learning algorithms may learn to collude on prices. We show that when firms face short-run price commitments, independent Q-learning (a simple but well-established self-learning algorithm) learns to profitably coordinate on either a fixed price or on asymmetric price cycles -- although convergence to rational and Pareto-optimal collusive behavior is not guaranteed. The general framework used can guide future research into the capacity of more advanced algorithms to collude, also in environments that are less stylized or more case-specific.
Subjects: 
pricing algorithms
algorithmic collusion
machine learning
reinforcement learning
Q-learning
sequential pricing
JEL: 
K21
L13
L49
Document Type: 
Working Paper

Files in This Item:
File
Size
319.7 kB





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.