Tuesday, February 1, 2022

A tepid take on the bipartisan drive to ban tech self-preferencing

A tepid take on the bipartisan drive to ban tech self-preferencing

By Matthew yglesias

In an outbreak of bipartisanship that didn’t get much attention, the Senate Judiciary Committee voted last week 16-6 to advance the American Innovation and Online Choice Act, a bill written by Amy Klobuchar and Chuck Grassley.

Committee Democrats unanimously supported the bill, along with Grassley, Lindsey Graham, Ted Cruz, Josh Hawley, and John Kennedy. That’s a lot of bipartisan support and would seem to indicate a real chance of securing 60 votes and passing. But it’s not clear whether everyone who voted for the bill actually supports it. The two senators from California, among others, have indicated serious reservations about the details, but have made the calculation that there will be further negotiations before it passes and want to be part of the team.

The White House, as I understand it, is being pulled in multiple directions on this legislation. Some senior Biden aides think the bill is a good idea, while others disagree. Some think it would be helpful for him to speak up in favor of it, but others don’t. And, again, almost everyone expects to see additional changes before it passes. But I’m not so sure — if everyone keeps voicing support for a flawed bill because they assume someone else will force changes, it just might pass.

So what is it?

The bill defines a class of “dominant online platforms” based on the size of their user base and their market capitalization. And then it tries to say they can’t preference their own products on the platforms they control. So, for example, Apple couldn’t preferentially steer you toward Apple’s first-party apps on the app store. Critics say it’s so broadly worded that Apple couldn’t even preload apps on your iPhone. That’s pretty clearly a dumb idea — and it’s something critics of the bill say is required by its wording rather than something proponents say they want — but it illustrates the problem with the bill.

The idea that there is a generalized problem with companies “self-preferencing” and/or that it needs to be specifically banned on “dominant online platforms” is very misguided.

The core concern of the coalition of small tech companies behind this bill was originally about Google and the idea that network effects give them an entrenched position that they can leverage into power in other markets.

Self-preferencing in general
Elizabeth Warren likes to use this analogy about tech platforms: “You can be the umpire — that is, you can run the platform, the marketplace — or you can have a team in the game — that is, you can be one of the competing businesses. But you don’t get to do both.”

The problem is that there’s nothing tech-specific about the critique. Costco owns a store. And in the store, they sell — among other things — Kirkland brand products. Target has store brands. So does Walmart. So do Safeway and Giant. If you go to a small town that doesn’t have a chain grocery store, the odds are good that the grocery store they do have will be a member of the Independent Grocers’ Association and will carry IGA store brand products.


Why have store brands?

The most basic reason is price. Some customers have a strong loyalty to a name-brand product like Coca-Cola. So the company that makes the product will want to charge a large markup relative to what it actually costs to produce their particular formula of sugary water. And the retailer also wants to earn a nice margin on that product. But the retailer can maintain the same profit margin at a much lower price by also carrying some Generic Cola.

In that case, you wouldn’t necessarily want to promote Generic Cola over Coca-Cola by giving it a better place on the shelf. But then again you might. This isn’t widely known, but everything in a large retail store is negotiable. If you have a new product that you want to put in front of customers, you may have to pay a slotting fee to retail chains to get it on the shelf. The space is finite, after all. And when you first introduce Flamin’ Hot Cheetos Mac ‘n Cheese, there’s no proven market for it. So if a new entrant wants the opportunity to gain a slice of the boxed mac ‘n cheese, they may need to pay for the privilege.

And as this Trax Retail piece explains, you can also find yourself paying for more favorable shelf positioning — especially for positioning in special displays at the ends of aisles. And a product that’s been on the shelves for a while might find itself being shaken down for a “pay to stay” fee if the retailer feels it’s no longer really delivering.

Store brands help retailers gain leverage in these cases. Generic milk sitting on the shelf next to name-brand milk helps a store understand how valuable the presence of that name-brand product is to the store.

Some retail chains, notably Walmart and Costco, like to brag that they don’t charge slotting fees. But what they do instead is demand discount pricing to be sold in their stores at all. Before he was a tech-oriented anti-monopolist, Barry Lynn’s big thing was complaining that Walmart used market power to drive supplier prices down too low. The point is that regardless of the details, retail stores own platforms and they manipulate what happens on that platform — including by selling store brand goods. Umpires play the game all the time.

“Tech” does not seem special to me
Klobuchar and Grassley seem to recognize that the Warren Doctrine doesn’t make sense and have written the legislation such that it wouldn’t apply to their home state companies like Target and Hy-Vee.

Instead, the bill is written to target “dominant online platforms,” which are really just five companies (Alphabet, Amazon, Apple, Meta, and Microsoft) to bar them from certain kinds of self-preferencing and self-dealing. The presumption, in other words, is that there’s something fundamentally different about being a big internet company.

But what is that?

I’ve often debated this specifically with regard to Amazon, which seems in a very literal sense to be in the same business as all these other retailers. Someone will inevitably claim that Amazon is a monopoly and that’s why it’s different. And as a leading Amazon apologist, I always want to ask what market does Amazon monopolize? You can go, right now, in your car and drive to a Walmart or a Target and buy all kinds of stuff. Retail is a competitive industry! The takes game massively over-indexes on carless New Yorkers who may underestimate how easy it is for the median American to get a big box store. But these stores also have websites.

For a long time, Amazon was the dominant option for online shopping because Amazon’s inventory and delivery were so much better than everyone else’s. But that’s good old-fashioned competition, and rivals have gotten better at e-commerce.

There’s also a burgeoning industry sector of companies that use Facebook, Instagram, or podcast advertising to do direct consumer response and sell through their own websites. And there are big companies like Shopify whose whole thing is helping people build that kind of business.

The other major relevant markets here I will admit are less competitive. In tablets, you have four players that are all Big Five tech companies. Apple and Microsoft compete to sell high-end computer operating systems, and to an extent, they both compete with Google Chromebooks on the low-end. Apple competes with basically just Google in the smartphone operating market, though they do compete with a bunch of vendors in smartphone hardware. Facebook in a sense competes with the whole universe for your attention, but in another sense has a dominant network-driven position on a certain kind of social networking. But Facebook also does not seem to be facing any accusations in this regard and isn’t a big piece of the lobbying puzzle.

In general, to describe these companies as “dominant tech platforms” is question-begging. Dominate what? How so?

Google does seem special to me
As I understand it, the original movers in D.C. around the antitrust critique of big tech were Yelp, who had pretty specific beef with Google.

To oversimplify a bit, once upon a time a restaurant-related Google search would yield a lot of prominent Yelp results. But now if I search for a new local restaurant, I get Google reviews in the box and Yelp is down the page. Naturally, Yelp did not like this turn of events, and it’s in their interest to develop theories as to why Google should not be allowed to do things that are bad for Yelp.


I know people who think Yelp’s position on this is absurd. To the extent that people like using Yelp as a tool, there is nothing stopping them from navigating to Yelp.com or using the Yelp app on their phone. And, indeed, precisely because I find Yelp to be a valuable service if I’m traveling to an unfamiliar place and want to check out some restaurant reviews, that’s normally what I’ll do. So one view is that Yelp is basically asking the government to force Google to do Yelp’s own customer acquisition work for them for free.

But here’s what gives me pause: Google really does have monopoly power.

Not because there’s no non-Google way to find things, but because Google web search benefits from network effects. User data is an input to the search algorithm, such that more users give Google more data which give them a better search engine. So they win on the merits against rival engines rather than winning through nefarious tactics. But they win today in part because they won yesterday. And they will win tomorrow in part because they are winning today. That network effect means that they can deliberately make the product worse and still have the best product. So even if showing a Yelp link would be best for users, they can get away with showing a Google reviews page instead.

The standard rebuttal from Google is that “competition is only a click away.” People use Google search because Google search is good, and the proof is in the pudding that lots of people like to use Google even though Bing and Duck Duck Go are out there. But the network effects give me pause.

Infrastructure separation
This is an issue that, structurally, the country has faced before using the existing antitrust statutes rather than a special law.

In the movie business, for example, we had a decades-long rule known as the Paramount Decree, which said that a movie studio could not also own a movie theater chain. The basic thinking was that lots of geographic areas might have only one or two movie theaters. And so to promote competition between movies, that kind of vertical integration would be blocked. More recently, thanks to streaming and other technological changes, we’ve let that go away and decided that movie distribution is now a more competitive industry.

Then there is the slightly odd case of Comcast, a cable and ISP company, buying NBC Universal.

Owning telecom wires is a low-competition “natural monopoly” type of business. So letting wire-owning monopolists vertically integrate into content seems dangerous. The Justice Department reached a settlement with the merging companies that allowed the merger to go forward, but Comcast had to promise not to favor NBC Universal channels or content on its cable systems. AT&T buying Time Warner posed a structurally similar question, and the Trump administration tried to sue to block (in part, seemingly, to punish CNN) even though AT&T preemptively agreed to follow the Comcast rules. Trump lost.

In the late 1990s Microsoft litigation, the federal government argued that Windows had monopoly power thanks to network effects and that Microsoft was using it to muscle into a dominant position in the web browser field and other places. The government argued for breaking up the company, but they ultimately agreed to a settlement where Microsoft agreed to stop various tying practices.

I know one view in Silicon Valley is that this lawsuit was an example of government cluelessness about technical matters. After all, just a few years after the suit came the rise of the web, the rise of mobile, the rise of Google and Apple, and suddenly Microsoft’s future was in doubt until Satya Nadella engineered an impressive turnaround. I’m skeptical about this story. I think, absent the antitrust litigation, Microsoft doesn’t make the commitments to Apple (both financial and, more importantly, the continued development of Office and Internet Explorer) that kept the company in business during its worst years. And if Microsoft had carte blanche to lock things down, they could have controlled the development of the early web in a way that made it impossible for Google to take off.

So I see the Microsoft case as a basically successful example of the general principle: when a given sector of the economy is inclined toward monopolization, we restrain the monopolist’s ability to play in other sandboxes.

Why my take is tepid
This is all a long-winded way of saying that I am open to the possibility that Google — not “dominant internet platforms” or “tech giants,” but specifically Google — should be barred from self-preferencing.

You could even convince me that conduct remedies of that sort are insufficient and we should split the company up. One company, Google Web Search, would run a search engine and sells ads on it. It would be an incredibly high-margin business because those ads generate a lot of revenue and it would have very limited capacity to invest the revenue in new things. Due to poor growth prospects, it would trade at a low multiple to earnings but would throw off tremendous dividends — a classic “income stock” like an electrical utility.

Then the other company, Alphabet, would be a conglomerate with roots in the same corporate culture that brought us Google Web Search, a culture that is oriented —and even compared to other tech companies — toward engineering and finding solutions to difficult computer science problems. That company would also be profitable but much riskier, with a higher upside due to a lack of legal constraints on its ability to enter new markets, but also more downside risk due to competition.

You could talk me into it.

But my take on this is pretty tepid because I find Google’s basic argument plausible that the ability to do tie-ins has been good for consumers. The classic Google “list of links” was great. But it’s also really is great that Google can often answer your question right there on the results page. That’s why normally you have a whole process where the enforcement agencies investigate, bring a case that’s full of facts and analysis, and do some lawyering. On its face, both the cases for and against Google have some surface plausibility. What’s really needed at this point is to do the work of trying to figure out which side is right, not blow it up into much broader legislation that would make some new principle and apply it to markets that are much more competitive.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.