Update on algorithmic pricing in competition law - What you need to know

What happened?

The case centres on allegations that multiple property managers and landlords delegated rent-setting to a common algorithm: RealPage’s revenue management software. This software allegedly used non-public, competitively sensitive data from participating landlords to generate rent recommendations, which were widely adopted and led to coordinated rent increases across competing apartment complexes. The plaintiffs claim that this amounted to a hub-and-spoke conspiracy:

• the “hub” being RealPage’s pricing software,
• the “spokes” being competing landlords who shared data and aligned rent prices via that software.

When algorithms become antitrust accomplices – Lessons learnt from RealPage

The RealPage case illustrates how software can be used to coordinate prices. Key takeaways:

• Beware of a hub-and-spoke constellation. There is no need for direct communication, and no need for an explicit agreement between competitors. Using an independent service provider does not remove antitrust risk (see, e.g., VM Remonts, CJEU, Case C 542/14 CURIA - Documents).
• Avoid sharing non-public data. In RealPage, the central allegation is that RealPage used and exchanged sensitive information that would not normally be accessible to the defendants. As a consequence, the settlement reportedly prohibits the defendants from using non-public data and requires RealPage to use contractual restrictions on future data use.
• Expect more cases. Most importantly, the European Commission has announced that it is looking into antitrust collusion via algorithms. This includes the work of pricing consultants. For the US, look at, for example, Duffy v. Yardi Systems, Inc. that raises similar allegations of rent inflation through algorithmic coordination.
• AI is no shield. Keep in mind that neither using software nor an algorithm shields you from antitrust liability.

Companies should act proactively to ensure compliance.

At the moment, it is uncertain how the European Commission will apply Article 101 TFEU to algorithmic pricing. It is of yet unclear where the exact limits will be and, very importantly, who can be held liable.¹ What is clear is that there will be enforcement. Companies should therefore act proactively to ensure compliance. Some practical steps:

• Demand transparency. Companies should demand transparency regarding a software’s data sources. Require developers and data scientists to disclose which data is used to feed the algorithm. Is the data non-public? Sufficiently aggregated and anonymized? Or even public? The European Commission has detailed guidance on sharing of sensitive data which it is very likely going to apply equally here.
• Review contracts with software providers. Contracts with software providers should be reviewed for the above-mentioned antitrust risks. Examine whether providers can access sensitive data, and whether limitations on e.g. data use, audit rights, and cooperation clauses in the event of regulatory inquiries are included.
• Assess market adoption. Be vigilant where a software solution is promoted as “used by your competitors”. This is no red flag, but it definitely means that you should look into the software and it means more carefully. The risk is high where many market participants use the same algorithm to calculate prices or, e.g., capacity. Keep in mind that there is no need for an explicit agreement.
• Strengthen internal governance. Establish policies for approving and monitoring pricing or capacity software. Ensure developers can explain the algorithm and train your legal and business teams on legal limits.

Authored by Dr. Elena Wiese and Dr. Julian Urban.