Regulation of Algorithmic Collusion, Refined: Testing Pessimistic Calibrated Regret

TL;DR

This paper introduces a new auditing method for regulating algorithmic collusion that tests sellers' pessimistic calibrated regret using observed data, relaxing previous assumptions and strengthening non-collusion criteria.

Contribution

It develops a more permissive auditing approach that relaxes distribution support requirements and justifies using vanishing calibrated regret as a non-collusion measure.

Findings

The new method relaxes fully-supported price distribution assumptions.

Algorithms satisfying weaker regret conditions can be manipulated into supra-competitive pricing.

Auditing can be bypassed by misrepresenting costs, suggesting a need for a rule of reason.

Abstract

We study the regulation of algorithmic (non-)collusion amongst sellers in dynamic imperfect price competition by auditing their data as introduced by Hartline et al. [2024]. We develop an auditing method that tests whether a seller's pessimistic calibrated regret is low. The pessimistic calibrated regret is the highest calibrated regret of outcomes compatible with the observed data. This method relaxes the previous requirement that a pricing algorithm must use fully-supported price distributions to be auditable. This method is at least as permissive as any auditing method that has a high probability of failing algorithmic outcomes with non-vanishing calibrated regret. Additionally, we strengthen the justification for using vanishing calibrated regret, versus vanishing best-in-hindsight regret, as the non-collusion definition, by showing that even without any side information, the pricing algorithms that only satisfy weaker vanishing best-in-hindsight regret allow an opponent to manipulate them into posting supra-competitive prices. This manipulation cannot be excluded with a non-collusion definition of vanishing best-in-hindsight regret. We motivate and interpret the approach of auditing algorithms from their data as suggesting a per se rule. However, we demonstrate that it is possible for algorithms to pass the audit by pretending to have higher costs than they actually do. For such scenarios, the rule of reason can be applied to bound the range of costs to those that are reasonable for the domain.