Part 4. There's No Break Up. But Plenty of Questions About Remedies in US v. Google
Part 4 of 4 analyzing Judge Mehta's landmark US v. Google decision. There will be no break up. No choice screens. Just an injunction on default contracts. Who wins and loses? What happens in January?
Welcome back to Part 4, the finale of our deep dive into the DOJ and States’ landmark victory in US v. Google. We are done with the substantive merits – and now can focus on the remedy and what comes next. And next we are shifting gears into some new and interesting topics that have caught my eye over the past few weeks.
If you are not caught up with our first two entries in the US v. Google series – start here:
Part I. Causation Confusion: Why Judge Mehta Got the Causation Standard Wrong and Whether it Matters
Part II. Foreclosure Fun & Games in US v. Google
Part III. Procompetitive Predicament: Why Does Google Pay Billions for the Default If Not to Foreclose Rivals?
And as always, welcome to those of you who are joining us for the first time at Competition on the Merits! If you are enjoying this – please send to friends to subscribe, add yourself to the growing list of paid subscribers, or shoot me a note via the Substack app or on email (jwright@lodestarle.com) to let me know what you want to hear about. Or all of the above!
Let’s get to it. There has been a lot of ink spilled in a short period of time about Google remedies. It is safe to say the liability finding has captured the imagination of the long list of Google enemies on the left and right. And while much of that commentary scores an A+ for creativity, the reality is that most of the remedial ideas are far-fetched and unrealistic. There are places you can go to indulge in those thought experiments. But I won’t waste time on them here. So if you are looking for good news, or legal or economic support, for your favorite wild-eyed and far-fetched remedy idea … this column is not going to make you happy. But if you want to know what is likely to happen, you have come to the right place.
But since we are here – it is good to know exactly how imaginative and ambitious the remedy proposals have been. Adam Kovacevich cleverly described these as “wishcasting” remedies. The label fits. Pick your favorite:
Tim Wu wants to break up Google in the form of an Android and Chrome spin off and require “free and open access” to Google’s AI technologies;
Some have proposed to prevent Google from showing specialized search results;
The DOJ leaned into the calls to break up Google, eagerly leaking to Bloomberg (who was equally eager to report) they are considering the wildest of these options.
Before we get into the easy and hard questions, let us recall what the court actually found. Judge Mehta found that Google possessed monopoly power in the “general search services” antitrust market, and the “search text advertising” market. But monopolization findings are made of two components: a showing that the defendant possesses monopoly power and that the defendant has engaged in “exclusionary” conduct that either allows it to acquire or maintain that power. As much as Google defended the monopoly power element, I think it is fair to say nobody was surprised by Judge Mehta’s conclusion that Google was a monopolist. Indeed, a distinctive element of American antitrust (in a good way!) is that monopoly power alone is not only no antitrust offense.
The case was largely about the second element. The DOJ alleged Google’s default search agreements with Apple and Android OEMs met the requirements for “exclusionary conduct,” and were not competition on the merits. We’ve written about those portions of Judge Mehta’s analysis at length earlier in the series. So let’s keep it moving.
The key point to understand – as readers who have read Part I, Part II, and Part III, of our series already do – is that the court held that Google’s search distribution agreements violated Section 2 of the Sherman Act because they foreclosed Bing and other rivals from competing for scale sufficient to constrain Google in the general search market.
But there is more. I think it is really important to pin down exactly the scope of the unlawful conduct specified in Judge Mehta’s opinion. While Judge Mehta finds that the requisite harm to competition in the relevant markets was established and could be attributed to Google’s search agreements, it made some other findings about those contracts that are pretty interesting from a remedy perspective. Specifically, Judge Mehta’s record includes evidence – read in the light most favorable to the DOJ and the States – that but-for the unlawful conduct Google’s share of searches would be about 10% lower. That is, the estimated impact of Google’s unlawful conduct relative to a choice screen was (at most) a 10% bump in search share. That’s not a huge impact.
Think about it this way – if the Court bans the contracts that 10% share can be re-distributed among Bing and other search rivals. Perhaps Bing gets 5 or 6%? Duck Duck Go gets a few percentage points. Others split the rest? The main point is that we have some idea of the aggregate impact of the contracts found unlawful. The point of the remedy is to undo that impact. But for now – it is important to keep your eye on the fact that the estimate of what the unlawful contracts do is in the record.
OK. We are up to speed. We know the illegal conduct. And we know what is on the wishlist for Google’s rivals and the antitrust activist crowd. Now what? I think a fun way to organize the questions surrounding remedies is by level of difficulty: easy, difficult, and impossible. Let’s start with the easy ones.
EASY REMEDIES QUESTIONS: WILL THE COURT BREAK UP GOOGLE?
No. Absolutely not. No break up. No AI duty to deal. No banning specialized search. Not a chance. I’m tempted to leave it at that. I could also go through the long list of cases in which a plaintiff prevails on a Section 2 exclusive dealing claim and evaluate the remedy (hint: it is an injunction prohibiting the unlawful contract and fencing-in relief to prevent obvious attempts to recreate that contract in different form). But let’s do a bit more work here. A few basic remedies principles will suffice to get us there.
Courts imposing Section 2 remedies generally try to achieve three goals: terminating the unlawful conduct, preventing it in the future, and re-establishing competition in the relevant market. The Court has described its task as "unfetter[ing] a market from anticompetitive conduct," rather than targeting and installing any particular form of competition. The D.C. Circuit itself has held that while antitrust remedies should deprive "an antitrust violator of the fruits of its violation,” such remedies do “not entail conferring a correlative benefit upon the particular competitor harmed by the violation," and should not proceed by “providing aid to a particular competitor.” Courts generally avoid remedies that place the court in the position of picking winners and losers or becoming a regulatory agency. In Verizon Communications Inc. v. Law Offices of Curtis V. Trinko, LLP, the Supreme Court warned against exactly that, warning that courts should not be placed in the position of "'assum[ing] the day-to-day controls characteristic of a regulatory agency.'"
The bottom line is that remedies are meant to undo the unlawful conduct. Courts generally enjoin the unlawful contract. And here we have record evidence that the actual impact of the contracts is relatively narrow – a 10% shift in search share. It is hard to imagine Judge Mehta entertaining broader remedies than a plain injunction and fencing-in relief. In fact, it is much easier to imagine more narrow remedies that allow Google to use less restrictive distribution contracts. And of course, the court cannot impose a choice screen remedy since Apple and Android OEMs are not parties to the lawsuit. More on that later.
Please try to ignore the breathless reporting on remedies that will never happen. Reporters need something to do as well. Please ignore the DOJ playing footsie with these radical remedy requests. None of that is going to happen. The contracts will be banned with some fencing relief, and perhaps even with some exceptions.
HARD REMEDIES QUESTIONS: WHO WINS AND LOSES FROM THE LIKELY REMEDY?
Let’s presume now the remedy is an injunction on Google search distribution contracts with Apple and Android OEMs. What should we make out of such an injunction? Does Apple want that? Do the Android OEMs? Does Mozilla?
I doubt it. You’ve surely heard that Google participates in a market without prices. How do we do antitrust when the price is zero has been a theme of antitrust discussion for about 25 years now. There are interesting questions here – and of course in multi-sided markets there is often a price somewhere! And, more importantly, there is output to measure everywhere! But the critical point about competition for the contract is that it provides an opportunity for price competition.
As readers know, the right way to think about the economics here is that Apple and Android OEMs create an auction for non-loyal consumers. That is, Apple and Android want to provide the best products for their users, and know that those with intense brand preferences for search engines will switch from the default. But Apple’s shelf space is the scarce and valuable asset here – to the tune of $20-something billion a year from Google. Apple (and Android OEMs) can create price competition from Google and Bing (and others) for its premium shelf space, i.e. access to the consumers who do not have intense search brand preferences, by holding an auction.
The first-order effect of the likely injunction will be to ban altogether or restrict competition in the auction for Apple’s shelf space. Who wins and loses? Well, Apple loses. Duh. Apple still cares about its own brand and, as Judge Mehta makes clear, Google is still the best search engine. Apple (and others) can still make Google the default. And they just might. But what Apple and Android OEMs cannot do is hold an auction for the shelf space. Google Loses Antitrust Lawsuit, Apple Hurt Most – is a reasonable headline.
Who else loses? Going down the line, the other beneficiaries of the Apple / Android OEM auctions are: consumers, Mozilla, and Google itself. Consumers lose because they are the indirect beneficiaries of the price competition that Apple’s auction creates. We discussed this issue at length in Part 3, but fairly conventional economic principles imply that the large payments from Google to Apple and Android benefit consumers. An interesting side note from the trial record is that Google’s witness did an excellent job at showing the large payments did not result in a large increase in Apple’s net margins. What is more difficult to prove is precisely how those large payments are passed on. Some margins are obvious: lower device prices, higher quality, etc. But at trial the “pass-through” benefits were left largely to implication from economic theory. I thought that Google’s counsel left some money on the table by failing to press DOJ’s witness (Whinston) on agreement on the basic economic principle that constant margins for Apple with large payments meant they were passed on somewhere. But the overall point is a basic one: users benefit from the large payments and those second-order benefits will be eliminated or reduced by the injunction.
Obviously, Google loses too. Sort of. And probably not as much as Apple. Apple was able to benefit from intense price competition between Google and Bing and reap the benefits of that competition in the form of those large payments. In exchange, Google got a 10% bump in search share – which is certainly worth something! But I would rank Apple’s losses as more significant than Google’s. And as a share of overall revenue, perhaps Mozilla is the biggest loser.
Who wins? Bing is an unequivocal winner. Other search rivals as well. With the injunction they will compete against a hampered rival. And they will share in the shift in share from the 10% or so of searches that will be up for grabs now and were previously covered by the contracts. But if you are expecting a big jump in quality from Bing or Duck Duck Go from sharing those additional searches, prepare to be disappointed. I don’t really see that as a remedy issue. The gap between DOJ’s theory of harm and the evidence at trial – the default agreements shifted about 10% of search share to Google relative to a choice screen world – is significant. But I see that much more as a defect in the causal nexus between the conduct and effect than a remedy issue.
A fundamental tension that runs through the entire remedy discussion – and I put this in the “hard” section for a reason because I’m not sure Judge Mehta can answer it in a manner consistent with his findings – remains this logical gap between three findings:
That Google’s unlawful default contracts only impacted a small share of searches, something like 10% or so; and
The same contracts are found to have impacted the overall conditions of the general search market; and
The court’s conclusion that Google’s quality was such that Bing was not a real competitor to Google “at any price.”
The remedy is going to leave a lot of parties unsatisfied given these findings. Perhaps resolving this tension should have been listed in our next section.
IMPOSSIBLE (BUT FUN) QUESTIONS: WHAT COMES NEXT?
Done with the easy and difficult, and on to the impossible. Or at least impossible to predict with any certainty. I think there are a few fun issues I will simply highlight here and leave for future discussion. It is too early to talk about either of them in any great detail.
The first is the implications of the Google decision for the future development of AI in the search market. On the one hand, Google introduced evidence that the market was changing so quickly that the fight over payments for search access points might be yesterday’s battle. Tomorrow’s battle might well be understanding where entry and competition in the search space is coming from. Maybe that is right. And antitrust has been on the receiving end — from me too – of the criticism that it is often too reactionary in dynamic markets. But I think the specter of AI competition in search was handled well by Judge Mehta as a matter of law and judgment. That’s neither here nor there for how important this decision will be in five years given changes in technology. It might be relatively meaningless. On the other hand, if you were developing technology in the United States and cared deeply about broad distribution – would the Google decision and rhetoric coming out of the DOJ be cause for optimism or concern? This is an issue we’ve discussed at length when it comes to what competition policy can do when it comes to the so-called “Little Tech Agenda.”
Second, what happens with the Google decision after the election? The appellate proceedings will be pending and not completed in November and January. Antitrusters will recall the frequent conversation about what the Bush administration might do with the DOJ’s US v. Microsoft trial court win if they took the White House in 2000. Many presumed the Republicans would toss the DOJ suit. They did not. While the district court’s extreme break up remedy was out the window – for a variety of reasons including Judge Jackson’s various judicial ethics violations – Charles James’s DOJ pursued the appeal and liability was upheld at least on some grounds. I do not think there is any particular lesson to draw from the specifics of that history. But the timing is really interesting for political observers. Many conservatives, if you have not noticed, are pretty upset with Google’s behavior in and around political speech. There is a long list of grievances – but the general gist is that Google has been an active player in using its platform to censor speech from conservatives and put the thumb on the scale in favor of the Democratic party. Just yesterday, news hit that one of Google’s lawyers is helping prep Vice President Harris for her debates with former President Trump. Regardless of where one comes out on such speech, it is protected by the First Amendment. But holding First Amendment issues aside, the timing here might end up fascinating if President Trump returns to office for another term.
Republicans have a long list of grievances and desired behavioral changes that likely matter to it a lot more than default search status or a 10% shift in search share. Google might well wake up on January 20th with a pending appeal and a headache borne of the prospect of an antitrust remedy and continued antitrust scrutiny. Political gains from trade have been found from less.
That wraps up our US v. Google series. I hope you enjoyed it. On to new topics later this week. Don’t forget to subscribe and send to friends.