> (i.e. in practical terms, there's no way regulators can police what algorithms sellers use - I can't think of exceptions to this, but perhaps there are some special cases)
Regulators can already police the data used as inputs in decision-making in industries like insurance, so policing the algorithms that operate on that data doesn't seem like too much of a reach.
> Regulators can already police the data used as inputs in decision-making in industries like insurance
How enforceable is policing which data can be used as inputs though?
It's common for insurance companies to price based on age and sex (e.g. teenage boys will typically pay higher car insurance premiums than similar aged girls). Presumably insurers are not allowed to price on a factor such as race. Unlike collusion, overt use of a variable like 'race' in a pricing model could be detected and enforced via a company whistleblower.
But how would a regulator find/prove algorithmic collusion?
In an extreme case, regulators could ban all use of competitors' data in a sellers' pricing models. But that seems extreme and unproductive since it could stop price wars (downward prices), as well as muting good effects of the 'invisible hand' (higher prices attracting more market entrants and greater investment)
I work in insurance, but not specifically in-depth on regulated insurance rates like personal auto.
That being said, I can add some insight. Most state insurance regulators require a company to justify the rate they're charging based on actual claims data (i.e. you wouldn't be allowed to use a competitor's pricing as a justification). Insurance companies would basically never share their claims data with their competitors, so there's functionally a ban on using competitor's data.
Any rate changes have to be justified (based on claims frequency and experience) to the state regulator. I don't think it's a perfect system by any means; insurance commissions aren't completely unbiased, and there's some flexibility in what data the insurer uses. But in my experience it's pretty effective at regulating the data you can and can't use.
The ultimate outcome is that most insurers in these markets run combined loss ratios of greater than 90% (so on an underwriting basis, more than 90% of the premium they earn goes to paying claims and overhead associated with managing those claims).
I think the model of "here's a regulatory body, justify what you're charging based on this set of allowed data" is a decent framework, even if it doesn't work in every market.
If you're curious, the SERFF website [0] has rate filings for a lot of states. So you can see when a rating factor changes and what it changed to. Most of the detailed claims data isn't available for data privacy reasons, but depending on the state you choose, there will be summary figures available.
[0]: https://www.serff.com/serff_filing_access.htm
> But how would a regulator find/prove algorithmic collusion?
They don't need to. At least in the US, courts look at the outcome and if the outcome is discriminatory that's the important part. This is under the idea of disparate impact. Beyond that, the realpage cases offer an example of modern day prosecution of algorithmic collusion.
Seems like an optimistic read on things. This is the kind of common-sense approach you would expect in a world without lawyers, just observing that collusion is bad because the effects are bad, and digging into the details of the causes are completely irrelevant for the public/plaintiff because it's really just on the company to fix the undesirable result.
IANAL but if realpages outcomes were definitive or reasonably generalized results dealing with the core issue, then similar arguments against e.g. Amazon would be a slam dunk. AFAIK, actual case outcome just hinges on details about "nonpublic data" and similar. Not remotely on bad effects for consumers or anything like that. Since printing realpages database in the newspaper would not actually help apartment-hunters, then this just tells landlords and third party markets how to do price-fixing legally next time? Most likely algorithmic pricing, surveillance pricing, etc is still coming to your grocery store after the issue is "settled" for property rental, or at least settled for realpages, in certain jurisdictions, for now.
> AFAIK, actual case outcome just hinges on details about "nonpublic data" and similar.
that sounds like insider trading. price fixing would need not involve nonpublic information (beyond the actual conspiracy to fix the prices as it helps to keep that part secret normally)
> “Settling Defendants have agreed not to provide nonpublic data to RealPage for use in competitor pricing recommendations and to refrain from using RealPage’s RMS that relies on non-public competitor data to make pricing recommendations,” attorneys wrote in the settlement filing.
https://www.multifamilydive.com/news/realpage-class-action-l...
I agree that "nonpublic" is barely related to the problem so how it's related to a solution is unclear. But it seems like this is the only general aspect of the outcome. Otherwise the outcome is just to stop doing this specific bad thing this specific time, and fines that are less than the profit made from bad behaviour.