I can hardly believe it, but I’ve read that a famous old bit by Henny Youngman has been purged from Florida textbooks, apparently because it was deemed offensive to those who wrote, told, and laughed at the joke. I won’t tell it here, but you can look it up. And if you’re a reader of a certain age, you’ll know it as soon as I note that it’s at the heart of the U.S. Justice Department’s (DOJ) antitrust case against Google.
But first, a ray of sunshine. On Sept. 1, the Federal Trade Commission (FTC) announced that it had reached a proposed consent order settling its suit to block Amgen’s acquisition of Horizon Therapeutics, which I wrote about here. In that same post, I noted a certain softening in the litigate-to-the-bitter-end/take-no-settlements rhetoric we’d been hearing from Assistant U.S. Attorney General Jonathan Kanter and FTC Chair Lina Khan. As I noted there, DOJ had recently agreed to settle its case against Assa Abloy and, as I’d covered elsewhere, the commission had withdrawn its case seeking to block Meta’s acquisition of Within.
In the latter case, the FTC lost its bid for a preliminary injunction in federal district court, based on Judge Davila’s “[h]aving reviewed and considered the objective evidence of Meta’s capabilities and incentives, the Court is not persuaded that this evidence establishes that it was ‘reasonably probable’ Meta would enter the relevant market.” Withdrawing is not the same as settlement, but it’s hardly litigating a case to its conclusion.
The FTC’s case against the Amgen/Horizon merger was a weak one, as we argued in an amicus brief submitted to the district court by ICLE and scholars of law and economics (submitted, but not technically filed, given that the case settled before the judge formally ruled on our motion to file, which accompanied the amicus brief itself). A statement on the settlement—issued by Khan and joined by Commissioners Rebecca Kelly Slaughter and Alvaro Bedoya—paints this as a win for the commission and consumers, and, not incidentally, “an advance in our pharmaceutical merger program.”
Sort of. Settling the case saved the FTC considerable resources, while avoiding a loss on the merits. And recognizing the merits of settlement and, specifically, the behavioral remedy, also made good sense. That is, it all made good sense given that they’d brought the case and were about to approach the bench in the U.S. District Court for the Northern District of Illinois.
But it shouldn’t have gotten so far. The core of the consent agreement has Amgen agreeing to do what they’d offered to do all along; that is, agree not to enter into contracts with pharmacy benefit managers (PBMs) or health plans for drug-formulary designs that would preclude entry by competitor products to the drugs at issue (Tepezza and Krystexxa), should such products successfully complete Phase III clinical trials, gain marketing approval from the U.S. Food and Drug Administration, and seek to enter the market (Tepezza and Krystexxa both enjoy monopolies already, so this was all about a stack of contingencies). As my ICLE colleagues Geoff Manne and R.J. Lehmann pointed out, our amicus brief had said precisely this all along. The FTC’s theory of harm was both speculative and remote, and remediable in any case—ex post, should anticompetitive conduct resulting in cognizable harms come to pass, and ex ante, straightforwardly, as Amgen had offered to enter into such a settlement all along.
But better late than never. So, at least one and a half cheers for my old agency for seeing the light before coming to the very end of the tunnel. This isn’t exactly calling a loss a win, and I’m inclined to give them their puffery.
I had forgotten what I’d covered in which post, so I had to search them out. Search, with Google. Which brings us back to DOJ’s monopoly maintenance (exclusionary practices in violation of Section 2 of the Sherman Act) case against Google. The trial began this week.
Everyone with any chance at all of reading this post will know about the trial already. It’s in all the papers. Here’s one story from The New York Times, with links to quite a few more. Here’s the Wall Street Journal under the demure title, “U.S. v. Google: What to Know About the Biggest Antitrust Trial in 20 Years.” Here’s one of many in the Washington Post, another in the Los Angeles Times, and before we run out of timeses: The Times of London and the Financial Times.
For informed background on the case (and other litigation against Google), I recommend the Truth on the Market symposium “The Future of American Antitrust: the Google Lawsuits,” with contributions from Geoff Manne, Sam Bowman, Eric Fruits, Gus Hurwitz, Nicolas Petit, Brian Albrecht, and Thom Lambert, as well as a very recent podcast by my ICLE colleague Geoff Manne and Corbin Barthold. And there’s a nice (if brief)discussion between Randy Picker and Spencer Webber Waller here.
There are complicated issues of law and fact, but at the heart of the matter are two sets of agreements under which Google is the default search engine on various browsers and cell phones. For example, Google is pre-loaded on Apple iPhones, and Google pays Apple (nowadays, via a revenue-sharing agreement) for that default status. DOJ argues that the default status unlawfully excludes rivals such as Bing and Yahoo, and thus maintains Google’s monopoly (or very large share) of general search. DOJ asks, in effect, why else would Google pay so much money for the default, if not to earn monopoly profits?
DOJ argues, in essence, that Google overpays for the default position, and that they do so to maintain their monopoly share of general search (or, very high market share: about 89% of general search in the United States, not that there aren’t other forms of search as well as other general search engines). One strand of the argument turns on the claim that the default position helps Google maintain access to a critical data advantage over its rivals, such as Microsoft’s Bing, such that those rivals (or would-be rivals) cannot effectively compete with Google, and hence cannot bid competitively for such default status themselves.
A few preliminary observations (made by many others):
For one, Google’s default placement agreements are not exclusive; that is, Google is paying to be the exclusive default, but they are not paying to prevent consumers from loading alternative search engines.
Second, it’s easy. Very easy. It took all of a handful of seconds for me to locate Bing on my iPhone’s App Store, and load it, and neither Apple nor Microsoft charged me a nickel to do it.
Third, not everyone is in a position to outbid Google for default status, but not nobody either. Microsoft pre-loads Bing as the default search engine with its Edge browser on Windows. True, Bing’s share of general search is a fraction of Google’s, but Microsoft has a market cap in excess of $2.5 trillion, well higher than that of Google’s parent company, Alphabet. Microsoft could bid for default status, if it wanted to (and thought it worthwhile).
Fourth, the agreements in question, in their current form, coincide with Google’s current share of general search (obviously), but Google’s antecedent default agreements date way back to 2005, when Google was not nearly the leading search engine, much less a monopolist.
Finally, DOJ says that Google has an insurmountable data advantage (see Thom Lambert on that argument here), but that’s not obviously true, and there are reasons to doubt it, absent the sort of rigorous demonstration that we’ve yet to see. Of course, Google does have access to data at a phenomenal scale and scope. And access to data is critical to running (and refining) a competitive search engine. But more data are more valuable only up to a point, and only in context. Diminishing returns and all that jazz.
How much data of what sort are needed? And how do we know that others cannot acquire what’s needed? Certainly, one needs enough—whatever that is—but one also needs the right computational resources (including applications resources—ways to make clever use of the data). Whether Google’s data edge (if it’s an edge, and if it’s an edge dependent on the default agreements) is decisive is a question of fact. Indeed, Bing’s incorporation of ChatGPT’s LLM is an interesting development, and what it will do to consumer perceptions of Bing’s search quality going forward is an interesting question.
Still, monopoly maintenance, they say. There are some odd things about the DOJ position, and today I’ll touch on just a few. But first, I have a question from left field—not my usual position—and it’s not that different from a question Judge Amit Mehta asked in a preliminary hearing: why isn’t this a monopsony-maintenance case brought against Apple?
After all, Apple—also larger than Google—has about 57% market share of mobile-phone sales in the United States, and about a 55% share of tablet sales. Apple’s installed base is huge, and they’re engaged in differentiated competition: I’d be willing to trade my iPhone (a terrific device, supplied by my employer) for a competitor’s phone (I’ve seen other terrific devices), but many millions of consumers would resist such a switch, even in the face of a significant price differential.
If Google is paying “too much” for its default status (or too much, absent the expectation of monopoly rents they wouldn’t otherwise realize), then how do we know it’s their fault that they are paying too much? If the default position is valuable, but not $10 billion worth of valuable, how do we know that the delta is not a signal of monopsony rents? And if the firms are splitting the rents, how do we calculate who is getting what share?
So, ok, why might Google pay so much money? Well, offhand, I’ve really no pocket calculation of how much they ought to pay. But suppose it’s marketing. Brian Albrecht and Thom Lambert liken Google’s payments to shelf-space promotions or “slotting fees,” as when, e.g., Coca Cola (which markets soft drinks very successfully) pays for prominent placement in grocery stores (and Brian again, here). As Brian says:
In the case of retail trade promotions, a promotional space given to Coca-Cola makes it marginally easier for consumers to pick Coke. But it does not reduce any consumer’s choice. The store will still have both items.
And, as noted already, it won’t be that hard to find Pepsi (if the store has it, and people look for it) or any other competing product.
The analogy is plain enough. Default search status places Google front and center on the relevant browser or cell phone. Finding an alternative is not hard or otherwise costly. But for some marginal consumers the default position is sticky. So, it’s worth something. How much?
Google’s annual revenues in 2022 were reported at more than $280 billion, an order of magnitude more than they pay for defaults, and still an order of magnitude below their market capitalization. If we think of the $10 billion as marketing expenditures, keeping Google front of mind for those who prefer it, and for those not-yet-committed, and for the marginal consumers who don’t care, have they spent too much?
It might well be the case that Google outbids Microsoft simply because the default wouldn’t be worth that much to Microsoft, and that the delta is due to the extent to which consumers prefer Google–that is, in the sense that matters, that Google is better. In the wake of the European Commission’s 2018 Android decision, Google had to implement a choice screen on Android devices in Europe (which had confined the market to Android devices). No more Google default: users of new devices with the choice screens are presented a half-dozen choices—including Google and obvious alternatives—upon startup. Placement is shuffled at random. Yet Google’s share of general search on mobile phones in Europe is reported to be greater than 96%.
Microsoft preloads Edge and Bing as defaults on Windows, but Bing trails way behind Google for general search in the United States on desktops, as well as mobile devices. It would seem that many people prefer Google, and prefer it enough to switch if they have to. And that suggests that it wouldn’t be rational for Microsoft to outbid Google for default status, given present demand. Some marginal consumers would stick with Bing, but apparently not enough to justify the expenditure. Which may partly explain DOJ’s “bonkers” (and unsuccessful) in limine motion to bar Google from offering evidence that its challenged conduct was not anti-competitive because it led to qualitative improvements in search.
Tricky spot there, otherwise–alleging that the deal leads to an insurmountable data advantage, but not conceding a qualitative superiority.
And it also calls to mind the question what would work as an adequate remedy (without wreaking havoc with consumers). DOJ hasn’t said. Geoff Manne’s podcast discusses that, too. What difference would it make to impose a European-style choice screen? Or to mandate a default position for some plucky little start-up firm, like Microsoft? Or … what? Any alternative would make some difference to Google and to others, but if the large majority will gravitate to Google in any case, what would the DOJ win? And if not for competition or consumers, for whom?
But maybe it’s not all due to marketing, even if the DOJ has it all wrong. Remember the old Microsoft case, and impediments to switching? Apple need not sell the default spot to anyone; and they need not sell it to Google. But what else might they or a rival do? In the but-for world, absent Google’s default agreements, would it be as easy for consumers to switch to Google as it is for them under current conditions to switch to Bing? Or anything else? And if Apple’s forays into search-engine development have not produced a rival to Google, have they been adequate to support a credible threat of a but-for world in which Google is at an undue disadvantage?
It’s all very complicated, but back to the Henny Youngman joke. Why does Google pay so much for default status? Maybe because it’s worth it?