What do we know about ‘machine collusion’
Some legal academics have claimed that ‘machine collusion’ – tacit collusion generated by self-learning pricing algorithms without human involvement – is a real threat that will go uncheck by current antitrust. Half a decade after these claims received wide publicity the only evidence are two experimental desk-top computer simulations with uncertain real-world relevance which at best suggest the theoretical possibility of machine collusion. The prospect of machine collusion is more limited and the law more adaptable than this literature assumed.
So far there is no evidence, no cases, a retreat by the academic lawyers from the more grandiose claims that evil algorithms are taking over the economy while at the same time many competition regulators ‘digitalizing’ themselves to ensure that the law and its enforcement can address the threat posed by algorithmic and machine collusion should it eventuate.
Published in Journal of European Competition Law & Practice 2021.
A pre-publication version is available at SSRN Electronic Journal.