The Google Search remedies decision just dropped. The court rejected breakups, payment bans, and choice screens. But it still ordered data sharing and syndication duties. This reasoning will shape antitrust for years. My thoughts https://x.com/nytimes/status/1962982232013906263…

First, what the court actually ordered: - No exclusive contracts for Search/Chrome/Assistant/Gemini - Must share some search index metadata (one-time) - Must offer syndication to rivals - 6-year term What it DIDN'T order matters just as much.

No Chrome divestiture. No Android split. It seems like a real shocker to some corners of antitrust. Those were never going to happen if the court followed the law at all. People (including the DOJ) deluded themselves. That's straight from Microsoft.

The court also rejected a payment ban, finding it would cause "substantial—in some cases, crippling—downstream harms." Think about Mozilla: most of its revenue comes from Google search deals. A ban would've killed Firefox. The court recognized this collateral damage.

But here's where it gets economically interesting. The court explicitly said plaintiffs don't need to show but-for causation for remedies. Google argued you can't impose remedies unless you prove we wouldn't have monopoly "but for" the exclusive defaults. Makes sense to me.

Court says: wrong standard. The key distinction is "significant causal connection" ≠ "but-for causation". Google said they were the same. Mehta here disagreed, arguing the D.C. Circuit never required but-for. That's Google misreading precedent.

By distinguishing the two, Mehta creates a sliding scale. The more drastic the remedy, the stronger your causal showing needs to be. Breakup? Need clear "significant causal connection" Behavioral limits? Inference of contribution is enough

This is one of those legal word plays that drives me nuts. Any reasonable understanding of "But for" already bakes a quantitative element. Causation is quantitative! The standard clearly couldn't be that "but for" means Google would have some tiny share "but for".

Why this matters: traditional antitrust wants clean counterfactuals. What would market share be without the bad conduct? This court says: in dynamic tech markets, that's impossible to reconstruct. So we'll use proportional remedies based on inference strength.

The court's logic: - Google's defaults "significantly contributed" to maintaining monopoly (inference) - Can't prove they were THE cause or measure the exact impact - So calibrate remedy intensity to evidence strength Middle-ground behavioral remedies fit this framework.

But again, causation is not a yes/no thing. There is no ONE cause of anything. More importantly, how do you know your remedy will work? We should be forward looking. If exclusivity didn't (partially) cause the monopoly, ending exclusivity probably won't fix it.

I take it that the court is acknowledging that antitrust's traditional causal framework breaks down in platform markets. Too many variables, network effects, quality differences. Can't isolate impact of any single practice. I think that's wrong. This holds everywhere.

So instead of giving up (no remedies) or going nuclear (breakup), the court creates this proportionality doctrine. Strong evidence of some contribution → moderate remedies. Would need smoking gun for structural relief. That seems like the right approach.

But this is we argued that the "but for" standard requires thisthe "but for" standard requires. Again, this is why is seems like word play to me. https://laweconcenter.org/resources/law-economics-scholars-amicus-to-us-district-court-in-united-states-v-google/…

You want strong causal evidence for ANY remedy that imposes costs. Whether it's breakup or data sharing, if it won't increase competition, it's deadweight loss. The form of remedy doesn't change that basic welfare calculation.

I take that to mean that plaintiffs no longer need to model the counterfactual world carefully. Just show exclusionary conduct + some inference it helped maintain power. For economists who build but-for models... your work just became less relevant in remedy proceedings.

I've seen some people say this looks like relitigating the case at remedy. I think that's exactly what's happening—and it's good! The liability decision had real problems with causation. The remedy phase is Judge Mehta's chance to course-correct without reversing himself.

Enough on but for. Let's talk about the forced index sharing. This is a bigger deal than it looks. Google must provide rivals with search index metadata - URLs, crawl dates, spam scores. One-time transfer at marginal cost. That's years of infrastructure investment handed over.

This is exactly what Trinko warned against (which the court rejects). Forcing successful firms to share the fruits of their investment reduces everyone's incentive to build. Why spend billions on crawling infrastructure if courts might just order you to hand over the output?

The one-time nature is interesting though. Court recognizes ongoing sharing would be too regulatory. But one snapshot? That ages quickly. Web changes daily. This index will be stale in months. So maybe it's symbolic more than substantive.

Lots more to dig into. It's a long opinion. Overall assessment: a reasonable ruling grounded in the law and economics. Some subtle confusing points around causation that will keep lawyers busy but don't seem to add much to our understanding.