G-Search's recent decline seems to be correlated with the availability of AI alternatives, or firms' differential success in incorporating its features into search. It will take a year or two to see effects of the remedial index share, which hasn't even occurred yet.

Changing market shares make coordinated results less likely. The important fact is that they are changing, as they are today.

In another indirect purchaser case, Mosaic Health, 2025 WL 2921807, the 2d Cir. held that indirect purchasers were not barred from damages for "lost access" caused by the alleged cartel's output reduction. Their claim did not depend on computing an overcharge.

There's truth on both sides: they agree that monopoly is bad for innovation; but aggressive protection for small firms is too. Much depends on where the peak of Aghion's ∩ lies. Further, outcomes are very heterogenous across industries.

I assume you mean "naked" price fixing. Yes, in theory anything that can be accomplished in a firm can be done by a joint venture (Coase), but there are still important practical differences (Hart): firms can solve many coordination problems better than agreements can.

A great deal of price discrimination -- probably most -- is output increasing. For example, firms expand into new territories or customers by offering lower prices; that's the way to reach marginal consumers. Forbidding it would be catastrophic.

yes, but the problem there is not price discrimination so much as just old fashioned collusion, particularly if the user of the algorithm agrees to employ its recommendations. YieldStar is a cartel-facilitating agent.

I'm reluctant to be too categorical, but in most situations that right. Output-increasing price discrimination often results in lower prices for both the high-paying and the low-paying buyers -- provided that both prices are above average variable cost.

The SCT already decided the RealPage case in 1923: American Linseed, 262 US 371, condemning the publisher of a newsletter who sold detailed price and output statistics on linseed, recommended prices, & agreed that subscribers would report back with any deviations from these.

court disagrees: "Client Defendants agree to set prices based on a pool of their horizontal competitors' proprietary data and reasonably believe that their competitors are using the same data and methods to price their properties" Realpage, 709 F.Supp3d at 494

RealPage's business, according to this case, appears to be cartel management services; you don't need to be a market participant to do that, but merely have knowledge about profit-maximizing output and a way of getting the real market participants to agree to it.

It's not all that uncommon that people -- particularly small business owners -- have no idea how the laws against price-fixing really work. What many don't understand is that agreement about output is about as common as agreeing about price -- and also illegal per se.

would you agree that it must refer to third degree price discrimination -- hence "surveillance"? Nevertheless, it was an effort to sully a generally neutral term.

I agree, most successful cartels and facilitators have used output rather than price as a metric; yet we always call it "price-fixing."

None of this is relevant. The cartel manager need not be accurate in order to create a profitable (and socially harmful) cartel. the smoking gun in the RealPage case was that RP's recipients agreed that they would abide by the recommendations.

legality likely turns on whether that agreement to adhere occurred. Products like Duetto GameTime yield similar results by letting algorithm "recommend" a price, but creating a default that charges that price unless the hotel operator changes it. That's simply naked price fixing.

People have been debating price discrim for a century, agreeing that if it reduces output it reduces welfare (measured by surplus, which you don't like); and that if it increases output welfare can go either way. Much of it is imposed by firms that have very little market power.

mainly good, although I wonder about making future patents available royalty free to these purchasers for five years. Knowing that your invention will immediately go uncompensated into a competitive market will largely undermine the incentive to invent.