[TOTM: The following is part of a digital symposium by TOTM guests and authors on the law, economics, and policy of the antitrust lawsuits against Google. The entire series of posts is available here.]
U.S. antitrust regulators have a history of narrowly defining relevant markets—often to the point of absurdity—in order to create market power out of thin air. The Federal Trade Commission (FTC) famously declared that Whole Foods and Wild Oats operated in the “premium natural and organic supermarkets market”—a narrowly defined market designed to exclude other supermarkets carrying premium natural and organic foods, such as Walmart and Kroger. Similarly, for the Staples-Office Depot merger, the FTC
narrowly defined the relevant market as “office superstore” chains, which excluded general merchandisers such as Walmart, K-Mart and Target, who at the time accounted for 80% of office supply sales.
Texas Attorney General Ken Paxton’s complaint against Google’s advertising business, joined by the attorneys general of nine other states, continues this tradition of narrowing market definition to shoehorn market dominance where it may not exist.
For example, one recent paper critical of Google’s advertising business narrows the relevant market first from media advertising to digital advertising, then to the “open” supply of display ads and, finally, even further to the intermediation of the open supply of display ads. Once the market has been sufficiently narrowed, the authors conclude Google’s market share is “perhaps sufficient to confer market power.”
While whittling down market definitions may achieve the authors’ purpose of providing a roadmap to prosecute Google, one byproduct is a mishmash of market definitions that generates as many as 16 relevant markets for digital display and video advertising, in many of which Google doesn’t have anything approaching market power (and in some of which, in fact, Facebook, and not Google, is the most dominant player).
The Texas complaint engages in similar relevant-market gerrymandering. It claims that, within digital advertising, there exist several relevant markets and that Google monopolizes four of them:
Publisher ad servers, which manage the inventory of a publisher’s (e.g., a newspaper’s website or a blog) space for ads;
Display ad exchanges, the “marketplace” in which auctions directly match publishers’ selling of ad space with advertisers’ buying of ad space;
Display ad networks, which are similar to exchanges, except a network acts as an intermediary that collects ad inventory from publishers and sells it to advertisers; and
Display ad-buying tools, which include demand-side platforms that collect bids for ad placement with publishers.
The complaint alleges, “For online publishers and advertisers alike, the different online advertising formats are not interchangeable.” But this glosses over a bigger challenge for the attorneys general: Is online advertising a separate relevant market from offline advertising?
Digital advertising, of which display advertising is a small part, is only one of many channels through which companies market their products. About half of today’s advertising spending in the United States goes to digital channels, up from about 10% a decade ago. Approximately 30% of ad spending goes to television, with the remainder going to radio, newspapers, magazines, billboards and other “offline” forms of media.
Physical newspapers now account for less than 10% of total advertising spending. Traditionally, newspapers obtained substantial advertising revenues from classified ads. As internet usage increased, newspaper classifieds have been replaced by less costly and more effective internet classifieds—such as those offered by Craigslist—or targeted ads on Google Maps or Facebook.
The price of advertising has fallen steadily over the past decade, while output has risen. Spending on digital advertising in the United States grew from $26 billion in 2010 to nearly $130 billion in 2019, an average increase of 20% a year. Over the same period, the producer price index (PPI) for internet advertising sales declined by nearly 40%. Rising spending in the face of falling prices indicates the number of ads bought and sold increased by approximately 27% a year.
Since 2000, advertising spending has been falling as a share of gross domestic product, with online advertising growing as a share of that. The combination of increasing quantity, decreasing cost and increasing total revenues are consistent with a growing and increasingly competitive market, rather than one of rising concentration and reduced competition.
There is little or no empirical data evaluating the extent to which online and offline advertising constitute distinct markets or the extent to which digital display is a distinct submarket of online advertising. As a result, analysis of adtech competition has relied on identifying several technical and technological factors—as well as the say-so of participants in the business—that the analysts assert distinguish online from offline and establish digital display (versus digital search) as a distinct submarket. This approach has been used and accepted, especially in cases in which pricing data has not been available.
But the pricing information that is available raises questions about the extent to which online advertising is a distinct market from offline advertising. For example, Avi Goldfarb and Catherine Tucker find that, when local regulations prohibit offline direct advertising, search advertising is more expensive, indicating that search and offline advertising are substitutes. In other research, they report that online display advertising circumvents, in part, local bans on offline billboard advertising for alcoholic beverages. In both studies, Goldfarb and Tucker conclude their results suggest online and offline advertising are substitutes. They also conclude this substitution suggests that online and offline markets should be considered together in the context of antitrust.
While this information is not sufficient to define a broader relevant market, it raises questions regarding solely relying on the technical or technological distinctions and the say-so of market participants.
In the United States, plaintiffs do not get to define the relevant market. That is up to the judge or the jury. Plaintiffs have the burden to convince the court that a proposed narrow market definition is the correct one. With strong evidence that online and offline ads are substitutes, the court should not blindly accept the gerrymandered market definitions posited by the attorneys general.
[TOTM: The following is part of a digital symposium by TOTM guests and authors on the law, economics, and policy of the antitrust lawsuits against Google. The entire series of posts is available here.]
Judges sometimes claim that they do not pick winners when they decide antitrust cases. Nothing could be further from the truth.
Competitive conduct by its nature harms competitors, and so if antitrust were merely to prohibit harm to competitors, antitrust would then destroy what it is meant to promote.
What antitrust prohibits, therefore, is not harm to competitors but rather harm to competitors that fails to improve products. Only in this way is antitrust able to distinguish between the good firm that harms competitors by making superior products that consumers love and that competitors cannot match and the bad firm that harms competitors by degrading their products without offering consumers anything better than what came before.
That means, however, that antitrust must pick winners: antitrust must decide what is an improvement and what not. And a more popular search engine is a clear winner.
But one should not take its winningness for granted. For once upon a time there was another winner that the courts always picked, blocking antitrust case after antitrust case. Until one day the courts stopped picking it.
That was the economy of scale.
The Structure of the Google Case
Like all antitrust cases that challenge the exercise of power, the government’s case against Google alleges denial of an input to competitors in some market. Here the input is default search status in smartphones, the competitors are rival search providers, and the market is search advertising. The basic structure of the case is depicted in the figure below.
Although brought as a monopolization case under Section 2 of the Sherman Act, this is at heart an exclusive dealing case of the sort normally brought under Section 1 of the Sherman Act: the government’s core argument is that Google uses contracts with smartphone makers, pursuant to which the smartphone makers promise to make Google, and not competitors, the search default, to harm competing search advertising providers and by extension competition in the search advertising market.
The government must show anticompetitive conduct, monopoly power, and consumer harm in order to prevail.
Let us assume that there is monopoly power. The company has more than 70% of the search advertising market, which is in the zone normally required to prove that element of a monopolization claim.
The problem of anticompetitive conduct is only slightly more difficult.
Anticompetitive conduct is only ever one thing in antitrust: denial of an essential input to a competitor. There is no other way to harm rivals.
(To be sure, antitrust prohibits harm to competition, not competitors, but that means only that harm to competitors necessary but insufficient for liability. The consumer harm requirement decides whether the requisite harm to competitors is also harm to competition.)
It is not entirely clear just how important default search status really is to running a successful search engine, but let us assume that it is essential, as the government suggests.
Then the question whether Google’s contracts are anticompetitive turns on how much of the default search input Google’s contracts foreclose to rival search engines. If a lot, then the rivals are badly harmed. If a little, then there may be no harm at all.
The answer here is that there is a lot of foreclosure, at least if the government’s complaint is to be believed. Through its contracts with Apple and makers of Android phones, Google has foreclosed default search status to rivals on virtually every single smartphone.
That leaves consumer harm. And here is where things get iffy.
Usage as a Product Improvement: A Very Convenient Argument
The inquiry into consumer harm evokes measurements of the difference between demand curves and price lines, or extrapolations of compensating and equivalent variation using indifference curves painstakingly pieced together based on the assumptions of revealed preference.
But while the parties may pay experts plenty to spin such yarns, and judges may pretend to listen to them, in the end, for the judges, it always comes down to one question only: did exclusive dealing improve the product?
If it did, then the judge assumes that the contracts made consumers better off and the defendant wins. And if it did not, then off with their heads.
So, does foreclosing all this default search space to competitors make Google search advertising more valuable to advertisers?
Those who leap to Google’s defense say yes, for default search status increases the number of people who use Google’s search engine. And the more people use Google’s search engine, the more Google learns about how best to answer search queries and which advertisements will most interest which searchers. And that ensures that even more people will use Google’s search engine, and that Google will do an even better job of targeting ads on its search engine.
And that in turn makes Google’s search advertising even better: able to reach more people and to target ads more effectively to them.
None of that would happen if defaults were set to other engines and users spurned Google, and so foreclosing default search space to rivals undoubtedly improves Google’s product.
This is a nice argument. Indeed, it is almost too nice, for it seems to suggest that almost anything Google might do to steer users away from competitors and to itself deserves antitrust immunity. Suppose Google were to brandish arms to induce you to run your next search on Google. That would be a crime, but, on this account, not an antitrust crime. For getting you to use Google does make Google better.
The argument that locking up users improves the product is of potential use not just to Google but to any of the many tech companies that run on advertising—Facebook being a notable example—so it potentially immunizes an entire business model from antitrust scrutiny.
It turns out that has happened before.
Economies of Scale as a Product Improvement: Once a Convenient Argument
Once upon a time, antitrust exempted another kind of business for which products improve the more people used them. The business was industrial production, and it differs from online advertising only in the irrelevant characteristic that the improvement that comes with expanding use is not in the quality of the product but in the cost per unit of producing it.
The hallmark of the industrial enterprise is high fixed costs and low marginal costs. The textile mill differs from pre-industrial piecework weaving in that once a $10 million investment in machinery has been made, the mill can churn out yard after yard of cloth for pennies. The pieceworker, by contrast, makes a relatively small up-front investment—the cost of raising up the hovel in which she labors and making her few tools—but spends the same large amount of time to produce each new yard of cloth.
Large fixed costs and low marginal costs lie at the heart of the bounty of the modern age: the more you produce, the lower the unit cost, and so the lower the price at which you can sell your product. This is a recipe for plenty.
But it also means that, so long as consumer demand in a given market is lower than the capacity of any particular plant, driving buyers to a particular seller and away from competitors always improves the product, in the sense that it enables the firm to increase volume and reduce unit cost, and therefore to sell the product at a lower price.
If the promise of the modern age is goods at low prices, then the implication is that antitrust should never punish firms for driving rivals from the market and taking over their customers. Indeed, efficiency requires that only one firm should ever produce in any given market, at least in any market for which a single plant is capable of serving all customers.
For antitrust in the late 19th and early 20th centuries, beguiled by this advantage to size, exclusive dealing, refusals to deal, even the knife in a competitor’s back: whether these ran afoul of other areas of law or not, it was all for the better because it allowed industrial enterprises to achieve economies of scale.
It is no accident that, a few notable triumphs aside, antitrust did not come into its own until the mid-1930s, 40 years after its inception, on the heels of an intellectual revolution that explained, for the first time, why it might actually be better for consumers to have more than one seller in a market.
These theories suggested that consumers might care as much about product quality as they do about product cost, and indeed would be willing to abandon a low-cost product for a higher-quality, albeit more expensive, one.
From this perspective, the world of economies of scale and monopoly production was the drab world of Soviet state-owned enterprises churning out one type of shoe, one brand of cleaning detergent, and so on.
The world of capitalism and technological advance, by contrast, was one in which numerous firms produced batches of differentiated products in amounts sometimes too small fully to realize all scale economies, but for which consumers were nevertheless willing to pay because the products better fit their preferences.
What is more, the striving of monopolistically competitive firms to lure away each other’s customers with products that better fit their tastes led to disruptive innovation— “creative destruction” was Schumpeter’s famous term for it—that brought about not just different flavors of the same basic concept but entirely new concepts. The competition to create a better flip phone, for example, would lead inevitably to a whole new paradigm, the smartphone.
This reasoning combined with work in the 1940s and 1950s on economic growth that quantified for the first time the key role played by technological change in the vigor of capitalist economies—the famous Solow residual—to suggest that product improvements, and not the cost reductions that come from capital accumulation and their associated economies of scale, create the lion’s share of consumer welfare. Innovation, not scale, was king.
Antitrust responded by, for the first time in its history, deciding between kinds of product improvements, rather than just in favor of improvements, casting economies of scale out of the category of improvements subject to antitrust immunity, while keeping quality improvements immune.
Casting economies of scale out of the protected product improvement category gave antitrust something to do for the first time. It meant that big firms had to plead more than just the cost advantages of being big in order to obtain license to push their rivals around. And government could now start reliably to win cases, rather than just the odd cause célèbre.
It is this intellectual watershed, and not Thurman Arnold’s tenacity, that was responsible for antitrust’s emergence as a force after World War Two.
Usage-Based Improvements Are Not Like Economies of Scale
The improvements in advertising that come from user growth fall squarely on the quality side of the ledger—the value they create is not due to the ability to average production costs over more ad buyers—and so they count as the kind of product improvements that antitrust continues to immunize today.
But given the pervasiveness of this mode of product improvement in the tech economy—the fact that virtually any tech firm that sells advertising can claim to be improving a product by driving users to itself and away from competitors—it is worth asking whether we have not reached a new stage in economic development in which this form of product improvement ought, like economies of scale, to be denied protection.
Shouldn’t the courts demand more and better innovation of big tech firms than just the same old big-data-driven improvements they serve up year after year?
Galling as it may be to those who, like myself, would like to see more vigorous antitrust enforcement in general, the answer would seem to be “no.” For what induced the courts to abandon antitrust immunity for economies of scale in the mid-20th century was not the mere fact that immunizing economies of scale paralyzed antitrust. Smashing big firms is not, after all, an end in itself.
Instead, monopolistic competition, creative destruction and the Solow residual induced the change, because they suggested both that other kinds of product improvement are more important than economies of scale and, crucially, that protecting economies of scale impedes development of those other kinds of improvements.
A big firm that excludes competitors in order to reach scale economies not only excludes competitors who might have produced an identical or near-identical product, but also excludes competitors who might have produced a better-quality product, one that consumers would have preferred to purchase even at a higher price.
To cast usage-based improvements out of the product improvement fold, a case must be made that excluding competitors in order to pursue such improvements will block a different kind of product improvement that contributes even more to consumer welfare.
If we could say, for example, that suppressing search competitors suppresses more-innovative search engines that ad buyers would prefer, even if those innovative search engines were to lack the advantages that come from having a large user base, then a case might be made that user growth should no longer count as a product improvement immune from antitrust scrutiny.
And even then, the case against usage-based improvements would need to be general enough to justify an epochal change in policy, rather than be limited to a particular technology in a particular lawsuit. For the courts hate to balance in individual cases, statements to the contrary in their published opinions notwithstanding.
But there is nothing in the Google complaint, much less the literature, to suggest that usage-based improvements are problematic in this way. Indeed, much of the value created by the information revolution seems to inhere precisely in its ability to centralize usage.
Americans Keep Voting to Centralize the Internet
In the early days of the internet, theorists mistook its decentralized architecture for a feature, rather than a bug. But internet users have since shown, time and again, that they believe the opposite.
For example, the basic protocols governing email were engineered to allow every American to run his own personal email server.
But Americans hated the freedom that created—not least the spam—and opted instead to get their email from a single server: the one run by Google as Gmail.
The basic protocols governing web traffic were also designed to allow every American to run whatever other communications services he wished—chat, video chat, RSS, webpages—on his own private server in distributed fashion.
But Americans hated the freedom that created—not least having to build and rebuild friend networks across platforms–—and they voted instead overwhelmingly to get their social media from a single server: Facebook.
Indeed, the basic protocols governing internet traffic were designed to allow every business to store and share its own data from its own computers, in whatever form.
But American businesses hated that freedom—not least the cost of having to buy and service their own data storage machines—and instead 40% of the internet is now stored and served from Amazon Web Services.
Similarly, advertisers have the option of placing advertisements on the myriad independently-run websites that make up the internet—known in the business as the “open web”—by placing orders through competitive ad exchanges. But advertisers have instead voted mostly to place ads on the handful of highly centralized platforms known as “walled gardens,” including Facebook, Google’s YouTube and, of course, Google Search.
The communications revolution, they say, is all about “bringing people together.” It turns out that’s true.
And that Google should win on consumer harm.
Remember the Telephone
Indeed, the same mid-20th century antitrust that thought so little of economies of scale as a defense immunized usage-based improvements when it encountered them in that most important of internet precursors: the telephone.
The telephone, like most internet services, gets better as usage increases. The more people are on a particular telephone network, the more valuable the network becomes to subscribers.
Just as with today’s internet services, the advantage of a large user base drove centralization of telephone services a century ago into the hands of a single firm: AT&T. Aside from a few business executives who liked the look of a desk full of handsets, consumers wanted one phone line that they could use to call everyone.
Although the government came close to breaking AT&T up in the early 20th century, the government eventually backed off, because a phone system in which you must subscribe to the right carrier to reach a friend just doesn’t make sense.
Instead, Congress and state legislatures stepped in to take the edge off monopoly by regulating phone pricing. And when antitrust finally did break AT&T up in 1982, it did so in a distinctly regulatory fashion, requiring that AT&T’s parts connect each other’s phone calls, something that Congress reinforced in the Telecommunications Act of 1996.
The message was clear: the sort of usage-based improvements one finds in communications are real product improvements. And antitrust can only intervene if it has a way to preserve them.
The equivalent of interconnection in search, that the benefits of usage, in the form of data and attention, be shared among competing search providers, might be feasible. But it is hard to imagine the court in the Google case ordering interconnection without the benefit of decades of regulatory experience with the defendant’s operations that the district court in 1982 could draw upon in the AT&T case.
The solution for the tech giants today is the same as the solution for AT&T a century ago: to regulate rather than to antitrust.
Microsoft Not to the Contrary, Because Users Were in Common
Parallels to the government’s 1990s-era antitrust case against Microsoft are not to the contrary.
As Sam Weinstein has pointed out to me, Microsoft, like Google, was at heart an exclusive dealing case: Microsoft contracted with computer manufacturers to prevent Netscape Navigator, an early web browser, from serving as the default web browser on Windows PCs.
That prevented Netscape, the argument went, from growing to compete with Windows in the operating system market, much the way the Google’s Chrome browser has become a substitute for Windows on low-end notebook computers today.
The D.C. Circuit agreed that default status was an essential input for Netscape as it sought eventually to compete with Windows in the operating system market.
The court also accepted the argument that the exclusive dealing did not improve Microsoft’s operating system product.
This at first seems to contradict the notion that usage improves products, for, like search advertising, operating systems get better as their user bases increase. The more people use an operating system, the more application developers are willing to write for the system, and the better the system therefore becomes.
It seems to follow that keeping competitors off competing operating systems and on Windows made Windows better. If the court nevertheless held Microsoft liable, it must be because the court refused to extend antitrust immunity to usage-based improvements.
The trouble with this line of argument is that it ignores the peculiar thing about the Microsoft case: that while the government alleged that Netscape was a potential competitor of Windows, Netscape was also an application that ran on Windows.
That means that, unlike Google and rival search engines, Windows and Netscape shared users.
So, Microsoft’s exclusive dealing did not increase its user base and therefore could not have improved Windows, at least not by making Windows more appealing for applications developers. Driving Netscape from Windows did not enable developers to reach even one more user. Conversely, allowing Netscape to be the default browser on Windows would not have reduced the number of Windows users, because Netscape ran on Windows.
By contrast, a user who runs a search in Bing does not run the same search simultaneously in Google, and so Bing users are not Google users. Google’s exclusive dealing therefore increases its user base and improves Google’s product, whereas Microsoft’s exclusive dealing served only to reduce Netscape’s user base and degrade Netscape’s product.
Indeed, if letting Netscape be the default browser on Windows was a threat to Windows, it was not because it prevented Microsoft from improving its product, but because Netscape might eventually have become an operating system, and indeed a better operating system, than Windows, and consumers and developers, who could be on both at the same time if they wished, might have nevertheless chosen eventually to go with Netscape alone.
Though it does not help the government in the Google case, Microsoft still does offer a beacon of hope for those concerned about size, for Microsoft’s subsequent history reminds us that yesterday’s behemoth is often today’s also ran.
And the favorable settlement terms Microsoft ultimately used to escape real consequences for its conduct 20 years ago imply that, at least in high-tech markets, we don’t always need antitrust for that to be true.
This week the Senate will hold a hearing into potential anticompetitive conduct by Google in its display advertising business—the “stack” of products that it offers to advertisers seeking to place display ads on third-party websites. It is also widely reported that the Department of Justice is preparing a lawsuit against Google that will likely include allegations of anticompetitive behavior in this market, and is likely to be joined by a number of state attorneys general in that lawsuit. Meanwhile, several papers have been published detailing these allegations.
This aspect of digital advertising can be incredibly complex and difficult to understand. Here we explain how display advertising fits in the broader digital advertising market, describe how display advertising works, consider the main allegations against Google, and explain why Google’s critics are misguided to focus on antitrust as a solution to alleged problems in the market (even if those allegations turn out to be correct).
Display advertising in context
Over the past decade, the price of advertising has fallen steadily while output has risen. Spending on digital advertising in the US grew from $26 billion in 2010 to nearly $130 billion in 2019, an average increase of 20% a year. Over the same period the Producer Price Index for Internet advertising sales declined by nearly 40%. The rising spending in the face of falling prices indicates the number of ads bought and sold increased by approximately 27% a year. Since 2000, advertising spending has been falling as a share of GDP, with online advertising growing as a share of that. The combination of increasing quantity, decreasing cost, and increasing total revenues are consistent with a growing and increasingly competitive market.
Display advertising on third-party websites is only a small subsection of the digital advertising market, comprising approximately 15-20% of digital advertising spending in the US. The rest of the digital advertising market is made up of ads on search results pages on sites like Google, Amazon and Kayak, on people’s Instagram and Facebook feeds, listings on sites like Zillow (for houses) or Craigslist, referral fees paid to price comparison websites for things like health insurance, audio and visual ads on services like Spotify and Hulu, and sponsored content from influencers and bloggers who will promote products to their fans.
And digital advertising itself is only one of many channels through which companies can market their products. About 53% of total advertising spending in the United States goes on digital channels, with 30% going on TV advertising and the rest on things like radio ads, billboards and other more traditional forms of advertising. A few people still even read physical newspapers and the ads they contain, although physical newspapers’ bigger money makers have traditionally been classified ads, which have been replaced by less costly and more effective internet classifieds, such as those offered by Craigslist, or targeted ads on Google Maps or Facebook.
Indeed, it should be noted that advertising itself is only part of the larger marketing market of which non-advertising marketing communication—e.g., events, sales promotion, direct marketing, telemarketing, product placement—is as big a part as is advertising (each is roughly $500bn globally); it just hasn’t been as thoroughly disrupted by the Internet yet. But it is a mistake to assume that digital advertising is not a part of this broader market. And of that $1tr global market, Internet advertising in total occupies only about 18%—and thus display advertising only about 3%.
Ad placement is only one part of the cost of digital advertising. An advertiser trying to persuade people to buy its product must also do market research and analytics to find out who its target market is and what they want. Moreover, there are the costs of designing and managing a marketing campaign and additional costs to analyze and evaluate the effectiveness of the campaign.
Nevertheless, one of the most straightforward ways to earn money from a website is to show ads to readers alongside the publisher’s content. To satisfy publishers’ demand for advertising revenues, many services have arisen to automate and simplify the placement of and payment for ad space on publishers’ websites. Google plays a large role in providing these services—what is referred to as “open display” advertising. And it is Google’s substantial role in this space that has sparked speculation and concern among antitrust watchdogs and enforcement authorities.
Before delving into the open display advertising market, a quick note about terms. In these discussions, “advertisers” are businesses that are trying to sell people stuff. Advertisers include large firms such as Best Buy and Disney and small businesses like the local plumber or financial adviser. “Publishers” are websites that carry those ads, and publish content that users want to read. Note that the term “publisher” refers to all websites regardless of the things they’re carrying: a blog about the best way to clean stains out of household appliances is a “publisher” just as much as the New York Times is.
Under this broad definition, Facebook, Instagram, and YouTube are also considered publishers. In their role as publishers, they have a common goal: to provide content that attracts users to their pages who will act on the advertising displayed. “Users” are you and me—the people who want to read publishers’ content, and to whom advertisers want to show ads. Finally, “intermediaries” are the digital businesses, like Google, that sit in between the advertisers and the publishers, allowing them to do business with each other without ever meeting or speaking.
The display advertising market
If you’re an advertiser, display advertising works like this: your company—one that sells shoes, let’s say—wants to reach a certain kind of person and tell her about the company’s shoes. These shoes are comfortable, stylish, and inexpensive. You use a tool like Google Ads (or, if it’s a big company and you want a more expansive campaign over which you have more control, Google Marketing Platform) to design and upload an ad, and tell Google about the people you want to read—their age and location, say, and/or characterizations of their past browsing and searching habits (“interested in sports”).
Using that information, Google finds ad space on websites whose audiences match the people you want to target. This ad space is auctioned off to the highest bidder among the range of companies vying, with your shoe company, to reach users matching the characteristics of the website’s users. Thanks to tracking data, it doesn’t just have to be sports-relevant websites: as a user browses sports-related sites on the web, her browser picks up files (cookies) that will tag her as someone potentially interested in sports apparel for targeting later.
So a user might look at a sports website and then later go to a recipe blog, and there receive the shoes ad on the basis of her earlier browsing. You, the shoe seller, hope that she will either click through and buy (or at least consider buying) the shoes when she sees those ads, but one of the benefits of display advertising over search advertising is that—as with TV ads or billboard ads—just seeing the ad will make her aware of the product and potentially more likely to buy it later. Advertisers thus sometimes pay on the basis of clicks, sometimes on the basis of views, and sometimes on the basis of conversion (when a consumer takes an action of some sort, such as making a purchase or filling out a form).
That’s the advertiser’s perspective. From the publisher’s perspective—the owner of that recipe blog, let’s say—you want to auction ad space off to advertisers like that shoe company. In that case, you go to an ad server—Google’s product is called AdSense—give them a little bit of information about your site, and add some html code to your website. These ad servers gather information about your content (e.g., by looking at keywords you use) and your readers (e.g., by looking at what websites they’ve used in the past to make guesses about what they’ll be interested in) and places relevant ads next to and among your content. If they click, lucky you—you’ll get paid a few cents or dollars.
Apart from privacy concerns about the tracking of users, the really tricky and controversial part here concerns the way scarce advertising space is allocated. Most of the time, it’s done through auctions that happen in real time: each time a user loads a website, an auction is held in a fraction of a second to decide which advertiser gets to display an ad. The longer this process takes, the slower pages load and the more likely users are to get frustrated and go somewhere else.
As well as the service hosting the auction, there are lots of little functions that different companies perform that make the auction and placement process smoother. Some fear that by offering a very popular product integrated end to end, Google’s “stack” of advertising products can bias auctions in favour of its own products. There’s also speculation that Google’s product is so tightly integrated and so effective at using data to match users and advertisers that it is not viable for smaller rivals to compete.
We’ll discuss this speculation and fear in more detail below. But it’s worth bearing in mind that this kind of real-time bidding for ad placement was not always the norm, and is not the only way that websites display ads to their users even today. Big advertisers and websites often deal with each other directly. As with, say, TV advertising, large companies advertising often have a good idea about the people they want to reach. And big publishers (like popular news websites) often have a good idea about who their readers are. For example, big brands often want to push a message to a large number of people across different customer types as part of a broader ad campaign.
Of these kinds of direct sales, sometimes the space is bought outright, in advance, and reserved for those advertisers. In most cases, direct sales are run through limited, intermediated auction services that are not open to the general market. Put together, these kinds of direct ad buys account for close to 70% of total US display advertising spending. The remainder—the stuff that’s left over after these kinds of sales have been done—is typically sold through the real-time, open display auctions described above.
Different adtech products compete on their ability to target customers effectively, to serve ads quickly (since any delay in the auction and ad placement process slows down page load times for users), and to do so inexpensively. All else equal (including the effectiveness of the ad placement), advertisers want to pay the lowest possible price to place an ad. Similarly, publishers want to receive the highest possible price to display an ad. As a result, both advertisers and publishers have a keen interest in reducing the intermediary’s “take” of the ad spending.
This is all a simplification of how the market works. There is not one single auction house for ad space—in practice, many advertisers and publishers end up having to use lots of different auctions to find the best price. As the market evolved to reach this state from the early days of direct ad buys, new functions that added efficiency to the market emerged.
In the early years of ad display auctions, individual processes in the stack were performed by numerous competing companies. Through a process of “vertical integration” some companies, such as Google, brought these different processes under the same roof, with the expectation that integration would streamline the stack and make the selling and placement of ads more efficient and effective. The process of vertical integration in pursuit of efficiency has led to a more consolidated market in which Google is the largest player, offering simple, integrated ad buying products to advertisers and ad selling products to publishers.
Google is by no means the only integrated adtech service provider, however: Facebook, Amazon, Verizon, AT&T/Xandr, theTradeDesk, LumenAd, Taboola and others also provide end-to-end adtech services. But, in the market for open auction placement on third-party websites, Google is the biggest.
The cases against Google
The UK’s Competition and Markets Authority (CMA) carried out a formal study into the digital advertising market between 2019 and 2020, issuing its final report in July of this year. Although also encompassing Google’s Search advertising business and Facebook’s display advertising business (both of which relate to ads on those companies “owned and operated” websites and apps), the CMA study involved the most detailed independent review of Google’s open display advertising business to date.
That study did not lead to any competition enforcement proceedings, but it did conclude that Google’s vertically integrated products led to conflicts of interest that could lead it to behaving in ways that did not benefit the advertisers and publishers that use it. One example was Google’s withholding of certain data from publishers that would make it easier for them to use other ad selling products; another was the practice of setting price floors that allegedly led advertisers to pay more than they would otherwise.
Instead the CMA recommended the setting up of a “Digital Markets Unit” (DMU) that could regulate digital markets in general, and a code of conduct for Google and Facebook (and perhaps other large tech platforms) intended to govern their dealings with smaller customers.
The CMA’s analysis is flawed, however. For instance, it makes big assumptions about the dependency of advertisers on display advertising, largely assuming that they would not switch to other forms of advertising if prices rose, and it is light on economics. But factually it is the most comprehensively researched investigation into digital advertising yet published.
While the Scott Morton and Dinielli paper is extremely broad, it also suffers from a number of problems.
One, because it was released before the CMA’s final report, it is largely based on the interim report released months earlier by the CMA, halfway through the market study in December 2019. This means that several of its claims are out of date. For example, it makes much of the possibility raised by the CMA in its interim report that Google may take a larger cut of advertising spending than its competitors, and claims made in another report that Google introduces “hidden” fees that increases the overall cut it takes from ad auctions.
But in the final report, after further investigation, the CMA concludes that this is not the case. In the final report, the CMA describes its analysis of all Google Ad Manager open auctions related to UK web traffic during the period between 8–14 March 2020 (involving billions of auctions). This, according to the CMA, allowed it to observe any possible “hidden” fees as well. The CMA concludes:
Our analysis found that, in transactions where both Google Ads and Ad Manager (AdX) are used, Google’s overall take rate is approximately 30% of advertisers’ spend. This is broadly in line with (or slightly lower than) our aggregate market-wide fee estimate outlined above. We also calculated the margin between the winning bid and the second highest bid in AdX for Google and non-Google DSPs, to test whether Google was systematically able to win with a lower margin over the second highest bid (which might have indicated that they were able to use their data advantage to extract additional hidden fees). We found that Google’s average winning margin was similar to that of non-Google DSPs. Overall, this evidence does not indicate that Google is currently extracting significant hidden fees. As noted below, however, it retains the ability and incentive to do so. (p. 275, emphasis added)
Scott Morton and Dinielli also misquote and/or misunderstand important sections of the CMA interim report as relating to display advertising when, in fact, they relate to search. For example, Scott Morton and Dinielli write that the “CMA concluded that Google has nearly insurmountable advantages in access to location data, due to the location information [uniquely available to it from other sources].” (p. 15). The CMA never makes any claim of “insurmountable advantage,” however. Rather, to support the claim, Scott Morton and Dinielli cite to a portion of the CMA interim report recounting a suggestion made by Microsoft regarding the “critical” value of location data in providing relevant advertising.
But that portion of the report, as well as the suggestion made by Microsoft, is about search advertising. While location data may also be valuable for display advertising, it is not clear that the GPS-level data that is so valuable in providing mobile search ad listings (for a nearby cafe or restaurant, say) is particularly useful for display advertising, which may be just as well-targeted by less granular, city- or county-level location data, which is readily available from a number of sources. In any case, Scott Morton and Dinielli are simply wrong to use a suggestion offered by Microsoft relating to search advertising to demonstrate the veracity of an assertion about a conclusion drawn by the CMA regarding display advertising.
Scott Morton and Dinielli also confusingly word their own judgements about Google’s conduct in ways that could be misinterpreted as conclusions by the CMA:
The CMA reports that Google has implemented an anticompetitive sales strategy on the publisher ad server end of the intermediation chain. Specifically, after purchasing DoubleClick, which became its publisher ad server, Google apparently lowered its prices to publishers by a factor of ten, at least according to one publisher’s account related to the CMA. (p. 20)
In fact, the CMA does not conclude that Google lowering its prices was an “anticompetitive sales strategy”—it does not use these words at all—and what Scott Morton and Dinielli are referring to is a claim by a rival ad server business, Smart, that Google cutting its prices after acquiring Doubleclick led to Google expanding its market share. Apart from the misleading wording, it is unclear why a competition authority should consider it to be “anticompetitive” when prices are falling and kept low, and—as Smart reported to the CMA—its competitor’s response is to enhance its own offering.
The case that remains
Stripping away the elements of Scott Morton and Dinielli’s case that seem unsubstantiated by a more careful reading of the CMA reports, and with the benefit of the findings in the CMA’s final report, we are left with a case that argues that Google self-preferences to an unreasonable extent, giving itself a product that is as successful as it is in display advertising only because of Google’s unique ability to gain advantage from its other products that have little to do with display advertising. Because of this self-preferencing, they might argue, innovative new entrants cannot compete on an equal footing, so the market loses out on incremental competition because of the advantages Google gets from being the world’s biggest search company, owning YouTube, running Google Maps and Google Cloud, and so on.
The most significant examples of this are Google’s use of data from other products—like location data from Maps or viewing history from YouTube—to target ads more effectively; its ability to enable advertisers placing search ads to easily place display ads through the same interface; its introduction of faster and more efficient auction processes that sidestep the existing tools developed by other third-party ad exchanges; and its design of its own tool (“open bidding”) for aggregating auction bids for advertising space to compete with (rather than incorporate) an alternative tool (“header bidding”) that is arguably faster, but costs more money to use.
These allegations require detailed consideration, and in a future paper we will attempt to assess them in detail. But in thinking about them now it may be useful to consider the remedies that could be imposed to address them, assuming they do diminish the ability of rivals to compete with Google: what possible interventions we could make in order to make the market work better for advertisers, publishers, and users.
We can think of remedies as falling into two broad buckets: remedies that stop Google from doing things that improve the quality of its own offerings, thus making it harder for others to keep up; and remedies that require it to help rivals improve their products in ways otherwise accessible only to Google (e.g., by making Google’s products interoperable with third-party services) without inherently diminishing the quality of Google’s own products.
The first camp of these, what we might call “status quo minus,” includes rules banning Google from using data from its other products or offering single order forms for advertisers, or, in the extreme, a structural remedy that “breaks up” Google by either forcing it to sell off its display ad business altogether or to sell off elements of it.
What is striking about these kinds of interventions is that all of them “work” by making Google worse for those that use it. Restrictions on Google’s ability to use data from other products, for example, will make its service more expensive and less effective for those who use it. Ads will be less well-targeted and therefore less effective. This will lead to lower bids from advertisers. Lower ad prices will be transmitted through the auction process to produce lower payments for publishers. Reduced publisher revenues will mean some content providers exit. Users will thus be confronted with less available content and ads that are less relevant to them and thus, presumably, more annoying. In other words: No one will be better off, and most likely everyone will be worse off.
The reason a “single order form” helps Google is that it is useful to advertisers, the same way it’s useful to be able to buy all your groceries at one store instead of lots of different ones. Similarly, vertical integration in the “ad stack” allows for a faster, cheaper, and simpler product for users on all sides of the market. A different kind of integration that has been criticized by others, where third-party intermediaries can bid more quickly if they host on Google Cloud, benefits publishers and users because it speeds up auction time, allowing websites to load faster. So does Google’s unified alternative to “header bidding,” giving a speed boost that is apparently valuable enough to publishers that they will pay for it.
So who would benefit from stopping Google from doing these things, or even forcing Google to sell its operations in this area? Not advertisers or publishers. Maybe Google’s rival ad intermediaries would; presumably, artificially hamstringing Google’s products would make it easier for them to compete with Google. But if so, it’s difficult to see how this would be an overall improvement. It is even harder to see how this would improve the competitive process—the very goal of antitrust. Rather, any increase in the competitiveness of rivals would result not from making their products better, but from making Google’s product worse. That is a weakening of competition, not its promotion.
On the other hand, interventions that aim to make Google’s products more interoperable at least do not fall prey to this problem. Such “status quo plus” interventions would aim to take the benefits of Google’s products and innovations and allow more companies to use them to improve their own competing products. Not surprisingly, such interventions would be more in line with the conclusions the CMA came to than the divestitures and operating restrictions proposed by Scott Morton and Dinielli, as well as (reportedly) state attorneys general considering a case against Google.
But mandated interoperability raises a host of different concerns: extensive and uncertain rulemaking, ongoing regulatory oversight, and, likely, price controls, all of which would limit Google’s ability to experiment with and improve its products. The history of such mandated duties to deal or compulsory licenses is a troubled one, at best. But even if, for the sake of argument, we concluded that these kinds of remedies were desirable, they are difficult to impose via an antitrust lawsuit of the kind that the Department of Justice is expected to launch. Most importantly, if the conclusion of Google’s critics is that Google’s main offense is offering a product that is just too good to compete with without regulating it like a utility, with all the costs to innovation that that would entail, maybe we ought to think twice about whether an antitrust intervention is really worth it at all.
As the initial shock of the COVID quarantine wanes, the Techlash waxes again bringing with it a raft of renewed legislative proposals to take on Big Tech. Prominent among these is the EARN IT Act (the Act), a bipartisan proposal to create a new national commission responsible for proposing best practices designed to mitigate the proliferation of child sexual abuse material (CSAM) online. The Act’s proposal is seemingly simple, but its fallout would be anything but.
Section 230 of the Communications Decency Act currently provides online services like Facebook and Google with a robust protection from liability that could arise as a result of the behavior of their users. Under the Act, this liability immunity would be conditioned on compliance with “best practices” that are produced by the new commission and adopted by Congress.
Supporters of the Act believe that the best practices are necessary in order to ensure that platform companies effectively police CSAM. While critics of the Act assert that it is merely a backdoor for law enforcement to achieve its long-sought goal of defeating strong encryption.
The truth of EARN IT—and how best to police CSAM—is more complicated. Ultimately, Congress needs to be very careful not to exceed its institutional capabilities by allowing the new commission to venture into areas beyond its (and Congress’s) expertise.
More can be done about illegal conduct online
On its face, conditioning Section 230’s liability protections on certain platform conduct is not necessarily objectionable. There is undoubtedly some abuse of services online, and it is also entirely possible that the incentives for finding and policing CSAM are not perfectly aligned with other conflicting incentives private actors face. It is, of course, first the responsibility of the government to prevent crime, but it is also consistent with past practice to expect private actors to assist such policing when feasible.
By the same token, an immunity shield is necessary in some form to facilitate user generated communications and content at scale. Certainly in 1996 (when Section 230 was enacted), firms facing conflicting liability standards required some degree of immunity in order to launch their services. Today, the control of runaway liability remains important as billions of user interactions take place on platforms daily. Related, the liability shield also operates as a way to promote good samaritan self-policing—a measure that surely helps avoid actual censorship by governments, as opposed to the spurious claims made by those like Senator Hawley.
In this context, the Act is ambiguous. It creates a commission composed of a fairly wide cross-section of interested parties—from law enforcement, to victims, to platforms, to legal and technical experts—to recommend best practices. That hardly seems a bad thing, as more minds considering how to design a uniform approach to controlling CSAM would be beneficial—at least theoretically.
In practice, however, there are real pitfalls to imbuing any group of such thinkers—especially ones selected by political actors—with an actual or de facto final say over such practices. Much of this domain will continue to be mercurial, the rules necessary for one type of platform may not translate well into general principles, and it is possible that a public board will make recommendations that quickly tax Congress’s institutional limits. To the extent possible, Congress should be looking at ways to encourage private firms to work together to develop best practices in light of their unique knowledge about their products and their businesses.
In fact, Facebook has already begun experimenting with an analogous idea in its recently announced Oversight Board. There, Facebook is developing a governance structure by giving the Oversight Board the ability to review content moderation decisions on the Facebook platform.
So far as the commission created by the Act works to create best practices that align the incentives of firms with the removal of CSAM, it has a lot to offer. Yet, a better solution than the Act would be for Congress to establish policy that works with the private processes already in development.
Short of a more ideal solution, it is critical, however, that the Act establish the boundaries of the commission’s remit very clearly and keep it from venturing into technical areas outside of its expertise.
The complicated problem of encryption (and technology)
The Act has a major problem insofar as the commission has a fairly open ended remit to recommend best practices, and this liberality can ultimately result in dangerous unintended consequences.
The Act only calls for two out of nineteen members to have some form of computer science background. A panel of non-technical experts should not design any technology—encryption or otherwise.
To be sure, there are some interesting proposals to facilitate access to encrypted materials (notably, multi-key escrow systems and self-escrow). But such recommendations are beyond the scope of what the commission can responsibly proffer.
If Congress proceeds with the Act, it should put an explicit prohibition in the law preventing the new commission from recommending rules that would interfere with the design of complex technology, such as by recommending that encryption be weakened to provide access to law enforcement, mandating particular network architectures, or modifying the technical details of data storage.
Congress is right to consider if there is better policy to be had for aligning the incentives of the platforms with the deterrence of CSAM—including possible conditional access to Section 230’s liability shield.But just because there is a policy balance to be struck between policing CSAM and platform liability protection doesn’t mean that the new commission is suited to vetting, adopting and updating technical standards – it clearly isn’t. Conversely, to the extent that encryption and similarly complex technologies could be subject to broad policy change it should be through an explicit and considered democratic process, and not as a by-product of the Act.
This is the third in a series of TOTM blog posts discussing the Commission’s recently published Google Android decision (the first post can be found here, and the second here). It draws on research from a soon-to-be published ICLE white paper.
(Comparison of Google and Apple’s smartphone business models. Red $ symbols represent money invested; Green $ symbols represent sources of revenue; Black lines show the extent of Google and Apple’s control over their respective platforms)
For the third in my series of posts about the Google Android decision, I will delve into the theories of harm identified by the Commission.
The big picture is that the Commission’s analysis was particularly one-sided. The Commission failed to adequately account for the complex business challenges that Google faced – such as monetizing the Android platform and shielding it from fragmentation. To make matters worse, its decision rests on dubious factual conclusions and extrapolations. The result is a highly unbalanced assessment that could ultimately hamstring Google and prevent it from effectively competing with its smartphone rivals, Apple in particular.
1. Tying without foreclosure
The first theory of harm identified by the Commission concerned the tying of Google’s Search app with the Google Play app, and of Google’s Chrome app with both the Google Play and Google Search apps.
Oversimplifying, Google required its OEMs to choose between either pre-installing a bundle of Google applications, or forgoing some of the most important ones (notably Google Play). The Commission argued that this gave Google a competitive advantage that rivals could not emulate (even though Google’s terms did not preclude OEMs from simultaneously pre-installing rival web browsers and search apps).
To support this conclusion, the Commission notably asserted that no alternative distribution channel would enable rivals to offset the competitive advantage that Google obtained from tying. This finding is, at best, dubious.
For a start, the Commission claimed that user downloads were not a viable alternative distribution channel, even though roughly 250 million apps are downloaded on Google’s Play store every day.
The Commission sought to overcome this inconvenient statistic by arguing that Android users were unlikely to download apps that duplicated the functionalities of a pre-installed app – why download a new browser if there is already one on the user’s phone?
But this reasoning is far from watertight. For instance, the 17th most-downloaded Android app, the “Super-Bright Led Flashlight” (with more than 587million downloads), mostly replicates a feature that is pre-installed on all Android devices. Moreover, the five most downloaded Android apps (Facebook, Facebook Messenger, Whatsapp, Instagram and Skype) provide functionalities that are, to some extent at least, offered by apps that have, at some point or another, been preinstalled on many Android devices (notably Google Hangouts, Google Photos and Google+).
The Commission countered that communications apps were not appropriate counterexamples, because they benefit from network effects. But this overlooks the fact that the most successful communications and social media apps benefited from very limited network effects when they were launched, and that they succeeded despite the presence of competing pre-installed apps. Direct user downloads are thus a far more powerful vector of competition than the Commission cared to admit.
Similarly concerning is the Commission’s contention that paying OEMs or Mobile Network Operators (“MNOs”) to pre-install their search apps was not a viable alternative for Google’s rivals. Some of the reasons cited by the Commission to support this finding are particularly troubling.
For instance, the Commission claimed that high transaction costs prevented parties from concluding these pre installation deals.
But pre-installation agreements are common in the smartphone industry. In recent years, Microsoft struck a deal with Samsung to pre-install some of its office apps on the Galaxy Note 10. It also paid Verizon to pre-install the Bing search app on a number of Samsung phones, in 2010. Likewise, a number of Russian internet companies have been in talks with Huawei to pre-install their apps on its devices. And Yahoo reached an agreement with Mozilla to make it the default search engine for its web browser. Transaction costs do not appear to have been an obstacle in any of these cases.
The Commission also claimed that duplicating too many apps would cause storage space issues on devices.
And yet, a back-of-the-envelope calculation suggests that storage space is unlikely to be a major issue. For instance, the Bing Search app has a download size of 24MB, whereas typical entry-level smartphones generally have an internal memory of at least 64GB (that can often be extended to more than 1TB with the addition of an SD card). The Bing Search app thus takes up less than one-thousandth of these devices’ internal storage. Granted, the Yahoo search app is slightly larger than Microsoft’s, weighing almost 100MB. But this is still insignificant compared to a modern device’s storage space.
Finally, the Commission claimed that rivals were contractually prevented from concluding exclusive pre-installation deals because Google’s own apps would also be pre-installed on devices.
However, while it is true that Google’s apps would still be present on a device, rivals could still pay for their applications to be set as default. Even Yandex – a plaintiff – recognized that this would be a valuable solution. In its own words (taken from the Commission’s decision):
Pre-installation alongside Google would be of some benefit to an alternative general search provider such as Yandex […] given the importance of default status and pre-installation on home screen, a level playing field will not be established unless there is a meaningful competition for default status instead of Google.
In short, the Commission failed to convincingly establish that Google’s contractual terms prevented as-efficient rivals from effectively distributing their applications on Android smartphones. The evidence it adduced was simply too thin to support anything close to that conclusion.
2. The threat of fragmentation
The Commission’s second theory of harm concerned the so-called “antifragmentation” agreements concluded between Google and OEMs. In a nutshell, Google only agreed to license the Google Search and Google Play apps to OEMs that sold “Android Compatible” devices (i.e. devices sold with a version of Android did not stray too far from Google’s most recent version).
According to Google, this requirement was necessary to limit the number of Android forks that were present on the market (as well as older versions of the standard Android). This, in turn, reduced development costs and prevented the Android platform from unraveling.
The Commission disagreed, arguing that Google’s anti-fragmentation provisions thwarted competition from potential Android forks (i.e. modified versions of the Android OS).
This conclusion raises at least two critical questions: The first is whether these agreements were necessary to ensure the survival and competitiveness of the Android platform, and the second is why “open” platforms should be precluded from partly replicating a feature that is essential to rival “closed” platforms, such as Apple’s iOS.
Let us start with the necessity, or not, of Google’s contractual terms. If fragmentation did indeed pose an existential threat to the Android ecosystem, and anti-fragmentation agreements averted this threat, then it is hard to make a case that they thwarted competition. The Android platform would simply not have been as viable without them.
The Commission dismissed this possibility, relying largely on statements made by Google’s rivals (many of whom likely stood to benefit from the suppression of these agreements). For instance, the Commission cited comments that it received from Yandex – one of the plaintiffs in the case:
(1166) The fact that fragmentation can bring significant benefits is also confirmed by third-party respondents to requests for information:
(2) Yandex, which stated: “Whilst the development of Android forks certainly has an impact on the fragmentation of the Android ecosystem in terms of additional development being required to adapt applications for various versions of the OS, the benefits of fragmentation outweigh the downsides…”
Ironically, the Commission relied on Yandex’s statements while, at the same time, it dismissed arguments made by Android app developers, on account that they were conflicted. In its own words:
Google attached to its Response to the Statement of Objections 36 letters from OEMs and app developers supporting Google’s views about the dangers of fragmentation […] It appears likely that the authors of the 36 letters were influenced by Google when drafting or signing those letters.
More fundamentally, the Commission’s claim that fragmentation was not a significant threat is at odds with an almost unanimous agreement among industry insiders.
For example, while it is not dispositive, a rapid search for the terms “Google Android fragmentation”, using the DuckDuckGo search engine, leads to results that cut strongly against the Commission’s conclusions. Of the ten first results, only one could remotely be construed as claiming that fragmentation was not an issue. The others paint a very different picture (below are some of the most salient excerpts):
“There’s a fairly universal perception that Android fragmentation is a barrier to a consistent user experience, a security risk, and a challenge for app developers.” (here)
“Android fragmentation, a problem with the operating system from its inception, has only become more acute an issue over time, as more users clamor for the latest and greatest software to arrive on their phones.” (here)
“Android Fragmentation a Huge Problem: Study.” (here)
“Google’s Android fragmentation fix still isn’t working at all.” (here)
“Does Google care about Android fragmentation? Not now—but it should.” (here).
“This is very frustrating to users and a major headache for Google… and a challenge for corporate IT,” Gold said, explaining that there are a large number of older, not fully compatible devices running various versions of Android.” (here)
Perhaps more importantly, one might question why Google should be treated differently than rivals that operate closed platforms, such as Apple, Microsoft and Blackberry (before the last two mostly exited the Mobile OS market). By definition, these platforms limit all potential forks (because they are based on proprietary software).
The Commission argued that Apple, Microsoft and Blackberry had opted to run “closed” platforms, which gave them the right to prevent rivals from copying their software.
While this answer has some superficial appeal, it is incomplete. Android may be an open source project, but this is not true of Google’s proprietary apps. Why should it be forced to offer them to rivals who would use them to undermine its platform? The Commission did not meaningfully consider this question.
And yet, industry insiders routinely compare the fragmentation of Apple’s iOS and Google’s Android OS, in order to gage the state of competition between both firms. For instance, one commentator noted:
[T]he gap between iOS and Android users running the latest major versions of their operating systems has never looked worse for Google.
Likewise, an article published in Forbes concluded that Google’s OEMs were slow at providing users with updates, and that this might drive users and developers away from the Android platform:
For many users the Android experience isn’t as up-to-date as Apple’s iOS. Users could buy the latest Android phone now and they may see one major OS update and nothing else. […] Apple users can be pretty sure that they’ll get at least two years of updates, although the company never states how long it intends to support devices.
However this problem, in general, makes it harder for developers and will almost certainly have some inherent security problems. Developers, for example, will need to keep pushing updates – particularly for security issues – to many different versions. This is likely a time-consuming and expensive process.
To recap, the Commission’s decision paints a world that is either black or white: either firms operate closed platforms, and they are then free to limit fragmentation as they see fit, or they create open platforms, in which case they are deemed to have accepted much higher levels of fragmentation.
This stands in stark contrast to industry coverage, which suggests that users and developers of both closed and open platforms care a great deal about fragmentation, and demand that measures be put in place to address it. If this is true, then the relative fragmentation of open and closed platforms has an important impact on their competitive performance, and the Commission was wrong to reject comparisons between Google and its closed ecosystem rivals.
3. Google’s revenue sharing agreements
The last part of the Commission’s case centered on revenue sharing agreements between Google and its OEMs/MNOs. Google paid these parties to exclusively place its search app on the homescreen of their devices. According to the Commission, these payments reduced OEMs and MNOs’ incentives to pre-install competing general search apps.
However, to reach this conclusion, the Commission had to make the critical (and highly dubious) assumption that rivals could not match Google’s payments.
To get to that point, it notably assumed that rival search engines would be unable to increase their share of mobile search results beyond their share of desktop search results. The underlying intuition appears to be that users who freely chose Google Search on desktop (Google Search & Chrome are not set as default on desktop PCs) could not be convinced to opt for a rival search engine on mobile.
But this ignores the possibility that rivals might offer an innovative app that swayed users away from their preferred desktop search engine.
More importantly, this reasoning cuts against the Commission’s own claim that pre-installation and default placement were critical. If most users, dismiss their device’s default search app and search engine in favor of their preferred ones, then pre-installation and default placement are largely immaterial, and Google’s revenue sharing agreements could not possibly have thwarted competition (because they did not prevent users from independently installing their preferred search app). On the other hand, if users are easily swayed by default placement, then there is no reason to believe that rivals could not exceed their desktop market share on mobile phones.
The Commission was also wrong when it claimed that rival search engines were at a disadvantage because of the structure of Google’s revenue sharing payments. OEMs and MNOs allegedly lost all of their payments from Google if they exclusively placed a rival’s search app on the home screen of a single line of handsets.
The key question is the following: could Google automatically tilt the scales to its advantage by structuring the revenue sharing payments in this way? The answer appears to be no.
For instance, it has been argued that exclusivity may intensify competition for distribution. Conversely, other scholars have claimed that exclusivity may deter entry in network industries. Unfortunately, the Commission did not examine whether Google’s revenue sharing agreements fell within this category.
It thus provided insufficient evidence to support its conclusion that the revenue sharing agreements reduced OEMs’ (and MNOs’) incentives to pre-install competing general search apps, rather than merely increasing competition “for the market”.
To summarize, the Commission overestimated the effect that Google’s behavior might have on its rivals. It almost entirely ignored the justifications that Google put forward and relied heavily on statements made by its rivals. The result is a one-sided decision that puts undue strain on the Android Business model, while providing few, if any, benefits in return.
This is the second in a series of TOTM blog posts discussing the Commission’s recently published Google Android decision (the first post can be found here). It draws on research from a soon-to-be published ICLE white paper.
This improper market definition might not be so problematic if the Commission had then proceeded to undertake a detailed (and balanced) assessment of the competitive conditions that existed in the markets where Google operates (including the competitive constraints imposed by Apple).
Unfortunately, this was not the case. The following paragraphs respond to some of the Commission’s most problematic arguments regarding the existence of barriers to entry, and the absence of competitive constraints on Google’s behavior.
The overarching theme is that the Commission failed to quantify its findings and repeatedly drew conclusions that did not follow from the facts cited. As a result, it was wrong to conclude that Google faced little competitive pressure from Apple and other rivals.
1. Significant investments and network effects ≠ barriers to entry
In its decision, the Commission notably argued that significant investments (millions of euros) are required to set up a mobile OS and App store. It also argued that market for licensable mobile operating systems gave rise to network effects.
But contrary to the Commission’s claims, neither of these two factors is, in and of itself, sufficient to establish the existence of barriers to entry (even under EU competition law’s loose definition of the term, rather than Stigler’s more technical definition)
Take the argument that significant investments are required to enter the mobile OS market.
The main problem is that virtually every market requires significant investments on the part of firms that seek to enter. Not all of these costs can be seen as barriers to entry, or the concept would lose all practical relevance.
For example, purchasing a Boeing 737 Max airplane reportedly costs at least $74 million. Does this mean that incumbents in the airline industry are necessarily shielded from competition? Of course not.
Instead, the relevant question is whether an entrant with a superior business model could access the capital required to purchase an airplane and challenge the industry’s incumbents.
Returning to the market for mobile OSs, the Commission should thus have questioned whether as-efficient rivals could find the funds required to produce a mobile OS. If the answer was yes, then the investments highlighted by the Commission were largely immaterial. As it happens, several firms have indeed produced competing OSs, including CyanogenMod, LineageOS and Tizen.
The same is true of Commission’s conclusion that network effects shielded Google from competitors. While network effects almost certainly play some role in the mobile OS and app store markets, it does not follow that they act as barriers to entry in competition law terms.
As Paul Belleflamme recently argued, it is a myth that network effects can never be overcome. And as I have written elsewhere, the most important question is whether users could effectively coordinate their behavior and switch towards a superior platform, if one arose (See also Dan Spulber’s excellent article on this point).
The Commission completely ignored this critical interrogation during its discussion of network effects.
2. The failure of competitors is not proof of barriers to entry
Just as problematically, the Commission wrongly concluded that the failure of previous attempts to enter the market was proof of barriers to entry.
This is the epitome of the Black Swan fallacy (i.e. inferring that all swans are white because you have never seen a relatively rare, but not irrelevant, black swan).
The failure of rivals is equally consistent with any number of propositions:
There were indeed barriers to entry;
Google’s products were extremely good (in ways that rivals and the Commission failed to grasp);
Google responded to intense competitive pressure by continuously improving its product (and rivals thus chose to stay out of the market);
The Commission did not demonstrate that its own inference was the right one, nor did it even demonstrate any awareness that other explanations were at least equally plausible.
3. First mover advantage?
Much of the same can be said about the Commission’s observation that Google enjoyed a first mover advantage.
The elephant in the room is that Google was not the first mover in the smartphone market (and even less so in the mobile phone industry). The Commission attempted to sidestep this uncomfortable truth by arguing that Google was the first mover in the Android app store market. It then concluded that Google had an advantage because users were familiar with Android’s app store.
To call this reasoning “naive” would be too kind. Maybe consumers are familiar with Google’s products today, but they certainly weren’t when Google entered the market.
Why would something that did not hinder Google (i.e. users’ lack of familiarity with its products, as opposed to those of incumbents such as Nokia or Blackberry) have the opposite effect on its future rivals?
Moreover, even if rivals had to replicate Android’s user experience (and that of its app store) to prove successful, the Commission did not show that there was anything that prevented them from doing so — a particularly glaring omission given the open-source nature of the Android OS.
The result is that, at best, the Commission identified a correlation but not causality. Google may arguably have been the first, and users might have been more familiar with its offerings, but this still does not prove that Android flourished (and rivals failed) because of this.
4. It does not matter that users “do not take the OS into account” when they purchase a device
The Commission also concluded that alternatives to Android (notably Apple’s iOS and App Store) exercised insufficient competitive constraints on Google. Among other things, it argued that this was because users do not take the OS into account when they purchase a smartphone (so Google could allegedly degrade Android without fear of losing users to Apple)..
In doing so, the Commission failed to grasp that buyers might base their purchases on a devices’ OS without knowing it.
Some consumers will simply follow the advice of a friend, family member or buyer’s guide. Acutely aware of their own shortcomings, they thus rely on someone else who does take the phone’s OS into account.
But even when they are acting independently, unsavvy consumers may still be driven by technical considerations. They might rely on a brand’s reputation for providing cutting edge devices (which, per the Commission, is the most important driver of purchase decisions), or on a device’s “feel” when they try it in a showroom. In both cases, consumers’ choices could indirectly be influenced by a phone’s OS.
In more technical terms, a phone’s hardware and software are complementary goods. In these settings, it is extremely difficult to attribute overall improvements to just one of the two complements. For instance, a powerful OS and chipset are both equally necessary to deliver a responsive phone. The fact that consumers may misattribute a device’s performance to one of these two complements says nothing about their underlying contribution to a strong end-product (which, in turn, drives purchase decisions). Likewise, battery life is reportedly one of the most important features for users, yet few realize that a phone’s OS has a large impact on it.
Finally, if consumers were really indifferent to the phone’s operating system, then the Commission should have dropped at least part of its case against Google. The Commission’s claim that Google’s anti-fragmentation agreements harmed consumers (by reducing OS competition) has no purchase if Android is provided free of charge and consumers are indifferent to non-price parameters, such as the quality of a phone’s OS.
5. Google’s users were not “captured”
Finally, the Commission claimed that consumers are loyal to their smartphone brand and that competition for first time buyers was insufficient to constrain Google’s behavior against its “captured” installed base.
It notably found that 82% of Android users stick with Android when they change phones (compared to 78% for Apple), and that 75% of new smartphones are sold to existing users.
The Commission asserted, without further evidence, that these numbers proved there was little competition between Android and iOS.
But is this really so? In almost all markets consumers likely exhibit at least some loyalty to their preferred brand. At what point does this become an obstacle to interbrand competition? The Commission offered no benchmark mark against which to assess its claims.
And although inter-industry comparisons of churn rates should be taken with a pinch of salt, it is worth noting that the Commission’s implied 18% churn rate for Android is nothing out of the ordinary (see, e.g., here, here, and here), including for industries that could not remotely be called anticompetitive.
To make matters worse, the Commission’s own claimed figures suggest that a large share of sales remained contestable (roughly 39%).
Imagine that, every year, 100 devices are sold in Europe (75 to existing users and 25 to new users, according to the Commission’s figures). Imagine further that the installed base of users is split 76–24 in favor of Android. Under the figures cited by the Commission, it follows that at least 39% of these sales are contestable.
According to the Commission’s figures, there would be 57 existing Android users (76% of 75) and 18 Apple users (24% of 75), of which roughly 10 (18%) and 4 (22%), respectively, switch brands in any given year. There would also be 25 new users who, even according to the Commission, do not display brand loyalty. The result is that out of 100 purchasers, 25 show no brand loyalty and 14 switch brands. And even this completely ignores the number of consumers who consider switching but choose not to after assessing the competitive options.
In short, the preceding paragraphs argue that the Commission did not meet the requisite burden of proof to establish Google’s dominance. Of course, it is one thing to show that the Commission’s reasoning was unsound (it is) and another to establish that its overall conclusion was wrong.
At the very least, I hope these paragraphs will convey a sense that the Commission loaded the dice, so to speak. Throughout the first half of its lengthy decision, it interpreted every piece of evidence against Google, drew significant inferences from benign pieces of information, and often resorted to circular reasoning.
The following post in this blog series argues that these errors also permeate the Commission’s analysis of Google’s allegedly anticompetitive behavior.
Yesterday was President Trump’s big “Social Media Summit” where he got together with a number of right-wing firebrands to decry the power of Big Tech to censor conservatives online. According to the Wall Street Journal:
Mr. Trump attacked social-media companies he says are trying to silence individuals and groups with right-leaning views, without presenting specific evidence. He said he was directing his administration to “explore all legislative and regulatory solutions to protect free speech and the free speech of all Americans.”
“Big Tech must not censor the voices of the American people,” Mr. Trump told a crowd of more than 100 allies who cheered him on. “This new technology is so important and it has to be used fairly.”
Despite the simplistic narrative tying President Trump’s vision of the world to conservatism, there is nothing conservative about his views on the First Amendment and how it applies to social media companies.
I have noted in severalplaces before that there is a conflict of visions when it comes to whether the First Amendment protects a negative or positive conception of free speech. For those unfamiliar with the distinction: it comes from philosopher Isaiah Berlin, who identified negative liberty as freedom from external interference, and positive liberty as freedom to do something, including having the power and resources necessary to do that thing. Discussions of the First Amendment’s protection of free speech often elide over this distinction.
With respect to speech, the negative conception of liberty recognizes that individual property owners can control what is said on their property, for example. To force property owners to allow speakers/speech on their property that they don’t desire would actually be a violation of their liberty — what the Supreme Court calls “compelled speech.” The First Amendment, consistent with this view, generally protects speech from government interference (with very few, narrow exceptions), while allowing private regulation of speech (again, with very few, narrow exceptions).
Contrary to the original meaning of the First Amendment and the weight of Supreme Court precedent, President Trump’s view of the First Amendment is that it protects a positive conception of liberty — one under which the government, in order to facilitate its conception of “free speech,” has the right and even the duty to impose restrictions on how private actors regulate speech on their property (in this case, social media companies).
But if Trump’s view were adopted, discretion as to what is necessary to facilitate free speech would be left to future presidents and congresses, undermining the bedrock conservative principle of the Constitution as a shield against government regulation, all falsely in the name of protecting speech. This is counter to the general approach of modern conservatism (but not, of course, necessarily Republicanism) in the United States, including that of many of President Trump’s own judicial and agency appointees. Indeed, it is actually more consistent with the views of modern progressives — especially within the FCC.
For instance, the current conservative bloc on the Supreme Court (over the dissent of the four liberal Justices) recently reaffirmed the view that the First Amendment applies only to state action in Manhattan Community Access Corp. v. Halleck. The opinion, written by Trump-appointee, Justice Brett Kavanaugh, states plainly that:
Ratified in 1791, the First Amendment provides in relevant part that “Congress shall make no law . . . abridging the freedom of speech.” Ratified in 1868, the Fourteenth Amendment makes the First Amendment’s Free Speech Clause applicable against the States: “No State shall make or enforce any law which shall abridge the privileges or immunities of citizens of the United States; nor shall any State deprive any person of life, liberty, or property, without due process of law . . . .” §1. The text and original meaning of those Amendments, as well as this Court’s longstanding precedents, establish that the Free Speech Clause prohibits only governmental abridgment of speech. The Free Speech Clause does not prohibit private abridgment of speech… In accord with the text and structure of the Constitution, this Court’s state-action doctrine distinguishes the government from individuals and private entities. By enforcing that constitutional boundary between the governmental and the private, the state-action doctrine protects a robust sphere of individual liberty. (Emphasis added).
Former Stanford Law dean and First Amendment scholar, Kathleen Sullivan, has summed up the very different approaches to free speech pursued by conservatives and progressives (insofar as they are represented by the “conservative” and “liberal” blocs on the Supreme Court):
In the first vision…, free speech rights serve an overarching interest in political equality. Free speech as equality embraces first an antidiscrimination principle: in upholding the speech rights of anarchists, syndicalists, communists, civil rights marchers, Maoist flag burners, and other marginal, dissident, or unorthodox speakers, the Court protects members of ideological minorities who are likely to be the target of the majority’s animus or selective indifference…. By invalidating conditions on speakers’ use of public land, facilities, and funds, a long line of speech cases in the free-speech-as-equality tradition ensures public subvention of speech expressing “the poorly financed causes of little people.” On the equality-based view of free speech, it follows that the well-financed causes of big people (or big corporations) do not merit special judicial protection from political regulation. And because, in this view, the value of equality is prior to the value of speech, politically disadvantaged speech prevails over regulation but regulation promoting political equality prevails over speech.
The second vision of free speech, by contrast, sees free speech as serving the interest of political liberty. On this view…, the First Amendment is a negative check on government tyranny, and treats with skepticism all government efforts at speech suppression that might skew the private ordering of ideas. And on this view, members of the public are trusted to make their own individual evaluations of speech, and government is forbidden to intervene for paternalistic or redistributive reasons. Government intervention might be warranted to correct certain allocative inefficiencies in the way that speech transactions take place, but otherwise, ideas are best left to a freely competitive ideological market.
The outcome of Citizens United is best explained as representing a triumph of the libertarian over the egalitarian vision of free speech. Justice Kennedy’s opinion for the Court, joined by Chief Justice Roberts and Justices Scalia, Thomas, and Alito, articulates a robust vision of free speech as serving political liberty; the dissenting opinion by Justice Stevens, joined by Justices Ginsburg, Breyer, and Sotomayor, sets forth in depth the countervailing egalitarian view. (Emphasis added).
President Trump’s views on the regulation of private speech are alarmingly consistent with those embraced by the Court’s progressives to “protect members of ideological minorities who are likely to be the target of the majority’s animus or selective indifference” — exactly the sort of conservative “victimhood” that Trump and his online supporters have somehow concocted to describe themselves.
The First Amendment does more than protect the interests of corporations. As courts have long recognized, it is a force to support individual interest in self-expression and the right of the public to receive information and ideas. As Justice Black so eloquently put it, “the widest possible dissemination of information from diverse and antagonistic sources is essential to the welfare of the public.” Our leased access rules provide opportunity for civic participation. They enhance the marketplace of ideas by increasing the number of speakers and the variety of viewpoints. They help preserve the possibility of a diverse, pluralistic medium—just as Congress called for the Cable Communications Policy Act… The proper inquiry then, is not simply whether corporations providing channel capacity have First Amendment rights, but whether this law abridges expression that the First Amendment was meant to protect. Here, our leased access rules are not content-based and their purpose and effect is to promote free speech. Moreover, they accomplish this in a narrowly-tailored way that does not substantially burden more speech than is necessary to further important interests. In other words, they are not at odds with the First Amendment, but instead help effectuate its purpose for all of us. (Emphasis added).
Consistent with the progressive approach, this leaves discretion in the hands of “experts” (like Rosenworcel) to determine what needs to be done in order to protect the underlying value of free speech in the First Amendment through government regulation, even if it means compelling speech upon private actors.
Trump’s view of what the First Amendment’s free speech protections entail when it comes to social media companies is inconsistent with the conception of the Constitution-as-guarantor-of-negative-liberty that conservatives have long embraced.
Principle #2: Any new intermediary liability law must not target constitutionally protected speech.
The government shouldn’t require—or coerce—intermediaries to remove constitutionally protected speech that the government cannot prohibit directly. Such demands violate the First Amendment. Also, imposing broad liability for user speech incentivizes services to err on the side of taking down speech, resulting in overbroad censorship—or even avoid offering speech forums altogether.
Principle #4: Section 230 does not, and should not, require “neutrality.”
Publishing third-party content online never can be “neutral.” Indeed, every publication decision will necessarily prioritize some content at the expense of other content. Even an “objective” approach, such as presenting content in reverse chronological order, isn’t neutral because it prioritizes recency over other values. By protecting the prioritization, de-prioritization, and removal of content, Section 230 provides Internet services with the legal certainty they need to do the socially beneficial work of minimizing harmful content.
The idea that social media should be subject to a nondiscrimination requirement — for which President Trump and others like Senator Josh Hawley have been arguing lately — is flatly contrary to Section 230 — as well as to the First Amendment.
Conservatives upset about “social media discrimination” need to think hard about whether they really want to adopt this sort of position out of convenience, when the tradition with which they align rejects it — rightly — in nearly all other venues. Even if you believe that Facebook, Google, and Twitter are trying to make it harder for conservative voices to be heard (despite all evidence to the contrary), it is imprudent to reject constitutional first principles for a temporary policy victory. In fact, there’s nothing at all “conservative” about an abdication of the traditional principle linking freedom to property for the sake of political expediency.
First, [my administration would restore competition to the tech sector] by passing legislation that requires large tech platforms to be designated as “Platform Utilities” and broken apart from any participant on that platform.
* * *
For smaller companies…, their platform utilities would be required to meet the same standard of fair, reasonable, and nondiscriminatory dealing with users, but would not be required to structurally separate….
* * * Second, my administration would appoint regulators committed to reversing illegal and anti-competitive tech mergers…. I will appoint regulators who are committed to… unwind[ing] anti-competitive mergers, including:
Let’s consider for a moment what this brave new world will look like — not the nirvana imagined by regulators and legislators who believe that decimating a company’s business model will deter only the “bad” aspects of the model while preserving the “good,” as if by magic, but the inevitable reality of antitrust populism.
Utilities? Are you kidding? For an overview of what the future of tech would look like under Warren’s “Platform Utility” policy, take a look at your water, electricity, and sewage service. Have you noticed any improvement (or reduction in cost) in those services over the past 10 or 15 years? How about the roads? Amtrak? Platform businesses operating under a similar regulatory regime would also similarly stagnate. Enforcing platform “neutrality” necessarily requires meddling in the most minute of business decisions, inevitably creating unintended and costly consequences along the way.
Network companies, like all businesses, differentiate themselves by offering unique bundles of services to customers. By definition, this means vertically integrating with some product markets and not others. Why are digital assistants like Siri bundled into mobile operating systems? Why aren’t the vast majority of third-party apps also bundled into the OS? If you want utilities regulators instead of Google or Apple engineers and designers making these decisions on the margin, then Warren’s “Platform Utility” policy is the way to go.
Grocery Stores. To take one specific case cited by Warren, how much innovation was there in the grocery store industry before Amazon bought Whole Foods? Since the acquisition, large grocery retailers, like Walmart and Kroger, have increased their investment in online services to better compete with the e-commerce champion. Many industry analysts expect grocery stores to use computer vision technology and artificial intelligence to improve the efficiency of check-out in the near future.
Smartphones. Imagine how forced neutrality would play out in the context of iPhones. If Apple can’t sell its own apps, it also can’t pre-install its own apps. A brand new iPhone with no apps — and even more importantly, no App Store — would be, well, just a phone, out of the box. How would users even access a site or app store from which to download independent apps? Would Apple be allowed to pre-install someone else’s apps? That’s discriminatory, too. Maybe it will be forced to offer a menu of all available apps in all categories (like the famously useless browser ballot screen demanded by the European Commission in its Microsoft antitrust case)? It’s hard to see how that benefits consumers — or even app developers.
Internet Search. Or take search. Calls for “search neutrality” have been bandied about for years. But most proponents of search neutrality fail to recognize that all Google’s search results entail bias in favor of its own offerings. As Geoff Manne and Josh Wright noted in 2011 at the height of the search neutrality debate:
[S]earch engines offer up results in the form not only of typical text results, but also maps, travel information, product pages, books, social media and more. To the extent that alleged bias turns on a search engine favoring its own maps, for example, over another firm’s, the allegation fails to appreciate that text results and maps are variants of the same thing, and efforts to restrain a search engine from offering its own maps is no different than preventing it from offering its own search results.
Nevermind that Google with forced non-discrimination likely means Google offering only the antiquated “ten blue links” search results page it started with in 1998 instead of the far more useful “rich” results it offers today; logically it would also mean Google somehow offering the set of links produced by any and all other search engines’ algorithms, in lieu of its own. If you think Google will continue to invest in and maintain the wealth of services it offers today on the strength of the profits derived from those search results, well, Elizabeth Warren is probably already your favorite politician.
And regulatory oversight of algorithmic content won’t just result in an impoverished digital experience; it will inevitably lead to an authoritarian one, as well:
Any agency granted a mandate to undertake such algorithmic oversight, and override or reconfigure the product of online services, thereby controls the content consumers may access…. This sort of control is deeply problematic… [because it saddles users] with a pervasive set of speech controls promulgated by the government. The history of such state censorship is one which has demonstrated strong harms to both social welfare and rule of law, and should not be emulated.
Digital Assistants. Consider also the veritable cage match among the tech giants to offer “digital assistants” and “smart home” devices with ever-more features at ever-lower prices. Today the allegedly non-existent competition among these companies is played out most visibly in this multi-featured market, comprising advanced devices tightly integrated with artificial intelligence, voice recognition, advanced algorithms, and a host of services. Under Warren’s nondiscrimination principle this market disappears. Each device can offer only a connectivity platform (if such a service is even permitted to be bundled with a physical device…) — and nothing more.
But such a world entails not only the end of an entire, promising avenue of consumer-benefiting innovation, it also entails the end of a promising avenue of consumer-benefiting competition. It beggars belief that anyone thinks consumers would benefit by forcing technology companies into their own silos, ensuring that the most powerful sources of competition for each other are confined to their own fiefdoms by order of law.
Breaking business models
Beyond the product-feature dimension, Sen. Warren’s proposal would be devastating for innovative business models. Why is Amazon Prime Video bundled with free shipping? Because the marginal cost of distribution for video is close to zero and bundling it with Amazon Prime increases the value proposition for customers. Why is almost every Google service free to users? Because Google’s business model is supported by ads, not monthly subscription fees. Each of the tech giants has carefully constructed an ecosystem in which every component reinforces the others. Sen. Warren’s plan would not only break up the companies, it would prohibit their business models — the ones that both created and continue to sustain these products. Such an outcome would manifestly harm consumers.
Both of Warren’s policy “solutions” are misguided and will lead to higher prices and less innovation. Her cause for alarm is built on a multitude of mistaken assumptions, but let’s address just a few (Warren in bold):
“Nearly half of all e-commerce goes through Amazon.” Yes, but it has only 5% of total retail in the United States. As my colleague Kristian Stout says, “the Internet is not a market; it’s a distribution channel.”
“Amazon has used its immense market power to force smaller competitors like Diapers.com to sell at a discounted rate.” The real story, as the founders of Diapers.com freely admitted, is that they sold diapers as what they hoped would be a loss leader, intending to build out sales of other products once they had a base of loyal customers:
And so we started with selling the loss leader product to basically build a relationship with mom. And once they had the passion for the brand and they were shopping with us on a weekly or a monthly basis that they’d start to fall in love with that brand. We were losing money on every box of diapers that we sold. We weren’t able to buy direct from the manufacturers.
Like all entrepreneurs, Diapers.com’s founders took a calculated risk that didn’t pay off as hoped. Amazon subsequently acquired the company (after it had declined a similar buyout offer from Walmart). (Antitrust laws protect consumers, not inefficient competitors). And no, this was not a case of predatory pricing. After many years of trying to make the business profitable as a subsidiary, Amazon shut it down in 2017.
“In the 1990s, Microsoft — the tech giant of its time — was trying to parlay its dominance in computer operating systems into dominance in the new area of web browsing. The federal government sued Microsoft for violating anti-monopoly laws and eventually reached a settlement. The government’s antitrust case against Microsoft helped clear a path for Internet companies like Google and Facebook to emerge.” The government’s settlement with Microsoft is not the reason Google and Facebook were able to emerge. Neither company entered the browser market at launch. Instead, they leapfrogged the browser entirely and created new platforms for the web (only later did Google create Chrome).
Furthermore, if the Microsoft case is responsible for “clearing a path” for Google is it not also responsible for clearing a path for Google’s alleged depredations? If the answer is that antitrust enforcement should be consistently more aggressive in order to rein in Google, too, when it gets out of line, then how can we be sure that that same more-aggressive enforcement standard wouldn’t have curtailed the extent of the Microsoft ecosystem in which it was profitable for Google to become Google? Warren implicitly assumes that only the enforcement decision in Microsoft was relevant to Google’s rise. But Microsoft doesn’t exist in a vacuum. If Microsoft cleared a path for Google, so did every decision not to intervene, which, all combined, created the legal, business, and economic environment in which Google operates.
Warren characterizes Big Tech as a weight on the American economy. In fact, nothing could be further from the truth. These superstar companies are the drivers of productivity growth, all ranking at or near the top for most spending on research and development. And while data may not be the new oil, extracting value from it may require similar levels of capital expenditure. Last year, Big Tech spent as much or more on capex as the world’s largest oil companies:
The exact causes of the decline in business dynamism are still uncertain, but recent research points to a much more mundane explanation: demographics. Labor force growth has been declining, which has led to an increase in average firm age, nudging fewer workers to start their own businesses.
Furthermore, it’s not at all clear whether this is actually a decline in business dynamism, or merely a change in business model. We would expect to see the same pattern, for example, if would-be startup founders were designing their software for acquisition and further development within larger, better-funded enterprises.
Will Rinehart recently looked at the literature to determine whether there is indeed a “kill zone” for startups around Big Tech incumbents. One paper finds that “an increase in fixed costs explains most of the decline in the aggregate entrepreneurship rate.” Another shows an inverse correlation across 50 countries between GDP and entrepreneurship rates. Robert Lucas predicted these trends back in 1978, pointing out that productivity increases would lead to wage increases, pushing marginal entrepreneurs out of startups and into big companies.
It’s notable that many in the venture capital community would rather not have Sen. Warren’s “help”:
just to sustain constant growth in GDP per person, the U.S. must double the amount of research effort searching for new ideas every 13 years to offset the increased difﬁculty of ﬁnding new ideas.
If this assessment is correct, it may well be that coming up with productive and profitable innovations is simply becoming more expensive, and thus, at the margin, each dollar of venture capital can fund less of it. Ironically, this also implies that larger firms, which can better afford the additional resources required to sustain exponential growth, are a crucial part of the solution, not the problem.
Warren believes that Big Tech is the cause of our social ills. But Americans have more trust in Amazon, Facebook, and Google than in the political institutions that would break them up. It would be wise for her to reflect on why that might be the case. By punishing our most valuable companies for past successes, Warren would chill competition and decrease returns to innovation.
Finally, in what can only be described as tragic irony, the most prominent political figure who shares Warren’s feelings on Big Tech is President Trump. Confirming the horseshoe theory of politics, far-left populism and far-right populism seem less distinguishable by the day. As our colleague Gus Hurwitz put it, with this proposal Warren is explicitly endorsing the unitary executive theory and implicitly endorsing Trump’s authority to direct his DOJ to “investigate specific cases and reach specific outcomes.” Which cases will he want to have investigated and what outcomes will he be seeking? More good questions that Senator Warren should be asking. The notion that competition, consumer welfare, and growth are likely to increase in such an environment is farcical.
Last week, I objected to Senator Warner relying on the flawed AOL/Time Warner merger conditions as a template for tech regulatory policy, but there is a much deeper problem contained in his proposals. Although he does not explicitly say “big is bad” when discussing competition issues, the thrust of much of what he recommends would serve to erode the power of larger firms in favor of smaller firms without offering a justification for why this would result in a superior state of affairs. And he makes these recommendations without respect to whether those firms actually engage in conduct that is harmful to consumers.
In the Data Portability section, Warner says that “As platforms grow in size and scope, network effects and lock-in effects increase; consumers face diminished incentives to contract with new providers, particularly if they have to once again provide a full set of data to access desired functions.“ Thus, he recommends a data portability mandate, which would theoretically serve to benefit startups by providing them with the data that large firms possess. The necessary implication here is that it is a per se good that small firms be benefited and large firms diminished, as the proposal is not grounded in any evaluation of the competitive behavior of the firms to which such a mandate would apply.
Warner also proposes an “interoperability” requirement on “dominant platforms” (which I criticized previously) in situations where, “data portability alone will not produce procompetitive outcomes.” Again, the necessary implication is that it is a per se good that established platforms share their services with start ups without respect to any competitive analysis of how those firms are behaving. The goal is preemptively to “blunt their ability to leverage their dominance over one market or feature into complementary or adjacent markets or products.”
Perhaps most perniciously, Warner recommends treating large platforms as essential facilities in some circumstances. To this end he states that:
Legislation could define thresholds – for instance, user base size, market share, or level of dependence of wider ecosystems – beyond which certain core functions/platforms/apps would constitute ‘essential facilities’, requiring a platform to provide third party access on fair, reasonable and non-discriminatory (FRAND) terms and preventing platforms from engaging in self-dealing or preferential conduct.
But, as i’ve previously noted with respect to imposing “essential facilities” requirements on tech platforms,
[T]he essential facilities doctrine is widely criticized, by pretty much everyone. In their respected treatise, Antitrust Law, Herbert Hovenkamp and Philip Areeda have said that “the essential facility doctrine is both harmful and unnecessary and should be abandoned”; Michael Boudin has noted that the doctrine is full of “embarrassing weaknesses”; and Gregory Werden has opined that “Courts should reject the doctrine.”
Indeed, as I also noted, “the Supreme Court declined to recognize the essential facilities doctrine as a distinct rule in Trinko, where it instead characterized the exclusionary conduct in Aspen Skiing as ‘at or near the outer boundary’ of Sherman Act § 2 liability.”
In short, it’s very difficult to know when access to a firm’s internal functions might be critical to the facilitation of a market. It simply cannot be true that a firm becomes bound under onerous essential facilities requirements (or classification as a public utility) simply because other firms find it more convenient to use its services than to develop their own.
The truth of what is actually happening in these cases, however, is that third-party firms are choosing to anchor their business to the processes of another firm which generates an “asset specificity” problem that they then seek the government to remedy:
A content provider that makes itself dependent upon another company for distribution (or vice versa, of course) takes a significant risk. Although it may benefit from greater access to users, it places itself at the mercy of the other — or at least faces great difficulty (and great cost) adapting to unanticipated, crucial changes in distribution over which it has no control.
This is naturally a calculated risk that a firm may choose to make, but it is a risk. To pry open Google or Facebook for the benefit of competitors that choose to play to Google and Facebook’s user base, rather than opening markets of their own, punishes the large players for being successful while also rewarding behavior that shies away from innovation. Further, such a policy would punish the large platforms whenever they innovate with their services in any way that might frustrate third-party “integrators” (see, e.g., Foundem’s claims that Google’s algorithm updates meant to improve search quality for users harmed Foundem’s search rankings).
Rather than encouraging innovation, blessing this form of asset specificity would have the perverse result of entrenching the status quo.
In all of these recommendations from Senator Warner, there is no claim that any of the targeted firms will have behaved anticompetitively, but merely that they are above a certain size. This is to say that, in some cases, big is bad.
Senator Warner’s policies would harm competition and innovation
As Geoffrey Manne and Gus Hurwitz have recently noted these views run completely counter to the last half-century or more of economic and legal learning that has occurred in antitrust law. From its murky, politically-motivated origins through the early 60’s when the Structure-Conduct-Performance (“SCP”) interpretive framework was ascendant, antitrust law was more or less guided by the gut feeling of regulators that big business necessarily harmed the competitive process.
Thus, at its height with SCP, “big is bad” antitrust relied on presumptions that large firms over a certain arbitrary threshold were harmful and should be subjected to more searching judicial scrutiny when merging or conducting business.
A paradigmatic example of this approach can be found in Von’s Grocery where the Supreme Court prevented the merger of two relatively small grocery chains. Combined, the two chains would have constitutes a mere 9 percent of the market, yet the Supreme Court, relying on the SCP aversion to concentration in itself, prevented the merger despite any procompetitive justifications that would have allowed the combined entity to compete more effectively in a market that was coming to be dominated by large supermarkets.
As Manne and Hurwitz observe: “this decision meant breaking up a merger that did not harm consumers, on the one hand, while preventing firms from remaining competitive in an evolving market by achieving efficient scale, on the other.” And this gets to the central defect of Senator Warner’s proposals. He ties his decisions to interfere in the operations of large tech firms to their size without respect to any demonstrable harm to consumers.
To approach antitrust this way — that is, to roll the clock back to a period before there was a well-defined and administrable standard for antitrust — is to open the door for regulation by political whim. But the value of the contemporary consumer welfare test is that it provides knowable guidance that limits both the undemocratic conduct of politically motivated enforcers as well as the opportunities for private firms to engage in regulatory capture. As Manne and Hurwitz observe:
Perhaps the greatest virtue of the consumer welfare standard is not that it is the best antitrust standard (although it is) — it’s simply that it is a standard. The story of antitrust law for most of the 20th century was one of standard-less enforcement for political ends. It was a tool by which any entrenched industry could harness the force of the state to maintain power or stifle competition.
While it is unlikely that Senator Warner intends to entrench politically powerful incumbents, or enable regulation by whim, those are the likely effects of his proposals.
Antitrust law has a rich set of tools for dealing with competitive harm. Introducing legislation to define arbitrary thresholds for limiting the potential power of firms will ultimately undermine the power of those tools and erode the welfare of consumers.
What to make of Wednesday’s decision by the European Commission alleging that Google has engaged in anticompetitive behavior? In this post, I contrast the European Commission’s (EC) approach to competition policy with US antitrust, briefly explore the history of smartphones and then discuss the ruling.
Asked about the EC’s decision the day it was announced, FTC Chairman Joseph Simons noted that, while the market is concentrated, Apple and Google “compete pretty heavily against each other” with their mobile operating systems, in stark contrast to the way the EC defined the market. Simons also stressed that for the FTC what matters is not the structure of the market per se but whether or not there is harm to the consumer. This again contrasts with the European Commission’s approach, which does not require harm to consumers. As Simons put it:
Once they [the European Commission] find that a company is dominant… that imposes upon the company kind of like a fairness obligation irrespective of what the effect is on the consumer. Our regulatory… our antitrust regime requires that there be a harm to consumer welfare — so the consumer has to be injured — so the two tests are a little bit different.
Indeed, and as the history below shows, the popularity of Apple’s iOS and Google’s Android operating systems arose because they were superior products — not because of anticompetitive conduct on the part of either Apple or Google. On the face of it, the conduct of both Apple and Google has led to consumer benefits, not harms. So, from the perspective of U.S. antitrust authorities, there is no reason to take action.
Moreover, there is a danger that by taking action as the EU has done, competition and innovation will be undermined — which would be a perverse outcome indeed. These concerns were reflected in astatement by Senator Mike Lee (R-UT):
Today’s decision by the European Commission to fine Google over $5 billion and require significant changes to its business model to satisfy EC bureaucrats has the potential to undermine competition and innovation in the United States,” Sen. Lee said. “Moreover, the decision further demonstrates the different approaches to competition policy between U.S. and EC antitrust enforcers. As discussed at the hearing held last December before the Senate’s Subcommittee on Antitrust, Competition Policy & Consumer Rights, U.S. antitrust agencies analyze business practices based on the consumer welfare standard. This analytical framework seeks to protect consumers rather than competitors. A competitive marketplace requires strong antitrust enforcement. However, appropriate competition policy should serve the interests of consumers and not be used as a vehicle by competitors to punish their successful rivals.
Ironically, the fundamental basis for the Commission’s decision is an analytical framework developed by economists at Harvard in the 1950s, which presumes that the structure of a market determines the conduct of the participants, which in turn presumptively affects outcomes for consumers. This “structure-conduct-performance” paradigm has been challenged both theoretically and empirically (and by “challenged,” I mean “demolished”).
Maintaining, as EC Commissioner Vestager has, that “What would serve competition is to have more players,” is to adopt a presumption regarding competition rooted in the structure of the market, without sufficient attention to the facts on the ground. As French economist Jean Tirole noted in his Nobel Prize lecture:
Economists accordingly have advocated a case-by-case or “rule of reason” approach to antitrust, away from rigid “per se” rules (which mechanically either allow or prohibit certain behaviors, ranging from price-fixing agreements to resale price maintenance). The economists’ pragmatic message however comes with a double social responsibility. First, economists must offer a rigorous analysis of how markets work, taking into account both the specificities of particular industries and what regulators do and do not know….
Second, economists must participate in the policy debate…. But of course, the responsibility here goes both ways. Policymakers and the media must also be willing to listen to economists.
In good Tirolean fashion, we begin with an analysis of how the market for smartphones developed. What quickly emerges is that the structure of the market is a function of intense competition, not its absence. And, by extension, mandating a different structure will likely impede competition, or, at the very least, will not likely contribute to it.
A brief history of smartphone competition
In 2006, Nokia’s N70 became the first smartphone to sell more than a million units. It was a beautiful device, with a simple touch screen interface and real push buttons for numbers. The following year, Apple released its first iPhone. It sold 7 million units — about the same as Nokia’s N95 and slightly less than LG’s Shine. Not bad, but paltry compared to the sales of Nokia’s 1200 series phones, which had combined sales of over 250 million that year — about twice the total of all smartphone sales in 2007.
By 2017, smartphones had come to dominate the market, with total sales of over1.5 billion. At the same time, the structure of the market has changed dramatically. In the first quarter of 2018, Apple’s iPhone X and iPhone 8 were thetwo best-selling smartphones in the world. In total, Apple shipped just over52 million phones, accounting for 14.5% of the global market. Samsung, which has a wider range of devices, sold even more: 78 million phones, or 21.7% of the market. At third and fourth place were Huawei (11%) and Xiaomi (7.5%). Nokia and LG didn’t even make it into the top 10, with market shares of only 3% and1% respectively.
Several factors have driven this highly dynamic market. Dramatic improvements in cellular data networks have played a role. But arguably of greater importance has been the development of software that offers consumers an intuitive and rewarding experience.
Apple’s iOS and Google’s Android operating systems have proven to be enormously popular among both users and app developers. This has generated synergies — or what economists call network externalities — as more apps have been developed, so more people are attracted to the ecosystem and vice versa, leading to a virtuous circle that benefits both users and app developers.
By contrast, Nokia’s early smartphones, including the N70 and N95, ran Symbian, the operating system developed for Psion’s handheld devices, which had a clunkier user interface and wasmore difficult to code — so it was less attractive to both users and developers. In addition, Symbian lacked an effective means of solving the problem of fragmentation of the operating system across different devices, which made it difficult for developers to create apps that ran across the ecosystem — something both Apple (through its closed system) and Google (through agreements with carriers) were able to address. Meanwhile, Java’s MIDP used in LG’s Shine, and its successor J2ME imposed restrictions on developers (such as prohibiting access to files, hardware, and network connections) that seem to have made it less attractive than Android.
The relative superiority of their operating systems enabled Apple and the manufacturers of Android-based phones to steal a march on the early leaders in the smartphone revolution.
The fact that Google allows smartphone manufacturers to install Android for free, distributes Google Play and other apps in a free bundle, and pays such manufacturers for preferential treatment for Google Search, has also kept the cost of Android-based smartphones down. As a result, Android phones are the cheapest on the market, providing a powerful experience for as little as $50. It is reasonable to conclude from this that innovation, driven by fierce competition, has led to devices, operating systems, and apps that provide enormous benefits to consumers.
The Commission decision would harm device manufacturers, app developers and consumers
The EC’s decision seems to disregard the history of smartphone innovation and competition and their ongoing consequences. As Dirk Auer explains, the Open Handset Alliance (OHA) was created specifically to offer an effective alternative to Apple’s iPhone — and it worked. Indeed, it worked so spectacularly that Android is installed on about 80% of all new phones. This success was the result of several factors that the Commission now seeks to undermine:
First, in order to maintain order within the Android universe, and thereby ensure that apps developed for Android would function on the vast majority of Android devices, Google and the OHA sought to limit the extent to which Android “forks” could be created. (Apple didn’t face this problem because its source code is proprietary, so cannot be modified by third-party developers.) One way Google does this is by imposing restrictions on the licensing of its proprietary apps, such as the Google Play store (a repository of apps, similar to Apple’s App Store).
Device manufacturers that don’t conform to these restrictions may still build devices with their forked version of Android — but without those Google apps. Indeed, Amazon chooses to develop a non-conforming version of Android and built its own app repository for its Fire devices (though it is still possible to add the Google Play Store). That strategy seems to be working for Amazon in the tablet market; in 2017 it rose past Samsung to become the second biggest manufacturer of tablets worldwide, after Apple.
Second, in order to be able to offer Android for free to smartphone manufacturers, Google sought to develop unique revenue streams (because, although the software is offered for free, it turns out that software developers generally don’t work for free). The main way Google did this was by requiring manufacturers that choose to install Google Play also to install its browser (Chrome) and search tools, which generate revenue from advertising. At the same time, Google kept its platform open by permitting preloads of rivals’ apps and creating a marketplace where rivals can also reach scale. Mozilla’s Firefox browser, for example, has been downloaded over 100 million times on Android.
The importance of these factors to the success of Android is acknowledged by the EC. But instead of treating them as legitimate business practices that enabled the development of high-quality, low-cost smartphones and a universe of apps that benefits billions of people, the Commission simply asserts that they are harmful, anticompetitive practices.
For example, the Commission asserts that
In order to be able to pre-install on their devices Google’s proprietary apps, including the Play Store and Google Search, manufacturers had to commit not to develop or sell even a single device running on an Android fork. The Commission found that this conduct was abusive as of 2011, which is the date Google became dominant in the market for app stores for the Android mobile operating system.
This is simply absurd, to say nothing of ahistorical. As noted, the restrictions on Android forks plays an important role in maintaining the coherency of the Android ecosystem. If device manufacturers were able to freely install Google apps (and other apps via the Play Store) on devices running problematic Android forks that were unable to run the apps properly, consumers — and app developers — would be frustrated, Google’s brand would suffer, and the value of the ecosystem would be diminished. Extending this restriction to all devices produced by a specific manufacturer, regardless of whether they come with Google apps preinstalled, reinforces the importance of the prohibition to maintaining the coherency of the ecosystem.
It is ridiculous to say that something (efforts to rein in Android forking) that made perfect sense until 2011 and that was central to the eventual success of Android suddenly becomes “abusive” precisely because of that success — particularly when the pre-2011 efforts were often viewed as insufficient and unsuccessful (a January 2012 Guardian Technology Blog post, “How Google has lost control of Android,” sums it up nicely).
Meanwhile, if Google is unable to tie pre-installation of its search and browser apps to the installation of its app store, then it will have less financial incentive to continue to maintain the Android ecosystem. Or, more likely, it will have to find other ways to generate revenue from the sale of devices in the EU — such as charging device manufacturers for Android or Google Play. The result is that consumers will be harmed, either because the ecosystem will be degraded, or because smartphones will become more expensive.
The troubling absence of Apple from the Commission’s decision
In addition, the EC’s decision is troublesome in other ways. First, for its definition of the market. The ruling asserts that “Through its control over Android, Google is dominant in the worldwide market (excluding China) for licensable smart mobile operating systems, with a market share of more than 95%.” But “licensable smart mobile operating systems” is a very narrow definition, as it necessarily precludes operating systems that are not licensable — such as Apple’s iOS and RIM’s Blackberry OS. Since Apple has nearly 25% of the market share of smartphones in Europe, the European Commission has — through its definition of the market — presumed away the primary source of effective competition. As Pinar Akmanhas noted:
How can Apple compete with Google in the market as defined by the Commission when Apple allows only itself to use its operating system only on devices that Apple itself manufactures?
The EU then invents a series of claims regarding the lack of competition with Apple:
end user purchasing decisions are influenced by a variety of factors (such as hardware features or device brand), which are independent from the mobile operating system;
It is not obvious that this is evidence of a lack of competition. A better explanation is that the EU’s narrow definition of the market is defective. In fact, one could easily draw the opposite conclusion of that drawn by the Commission: the fact that purchasing decisions are driven by various factors suggests that there is substantial competition, with phone manufacturers seeking to design phones that offer a range of features, on a number of dimensions, to best capture diverse consumer preferences. They are able to do this in large part precisely because consumers are able to rely upon a generally similar operating system and continued access to the apps that they have downloaded. As Tim Cook likes to remind his investors, Apple is quite successful at targeting “Android switchers” to switch to iOS.
Apple devices are typically priced higher than Android devices and may therefore not be accessible to a large part of the Android device user base;
And yet, in the first quarter of 2018, Apple phones accounted for five of the top ten selling smartphones worldwide. Meanwhile, several competing phones, including thefifth and sixth best-sellers, Samsung’s Galaxy S9 and S9+, sell forsimilar prices to themostexpensive iPhones. And a refurbished iPhone 6 can be had for less than $150.
Android device users face switching costs when switching to Apple devices, such as losing their apps, data and contacts, and having to learn how to use a new operating system;
This is, of course, true for any system switch. And yet the growing market share of Apple phones suggests that some users are willing to part with those sunk costs. Moreover, the increasing predominance of cloud-based and cross-platform apps, as well as Apple’s own “Move to iOS” Android app (which facilitates the transfer of users’ data from Android to iOS), means that the costs of switching border on trivial. As mentioned above, Tim Cook certainly believes in “Android switchers.”
even if end users were to switch from Android to Apple devices, this would have limited impact on Google’s core business. That’s because Google Search is set as the default search engine on Apple devices and Apple users are therefore likely to continue using Google Search for their queries.
This is perhaps the most bizarre objection of them all. The fact that Apple chooses to install Google search as the default demonstrates that consumers prefer that system over others. Indeed, this highlights a fundamental problem with the Commission’s own rationale, As Akman notes:
It is interesting that the case appears to concern a dominant undertaking leveraging its dominance from a market in which it is dominant (Google Play Store) into another market in which it is also dominant (internet search). As far as this author is aware, most (if not all?) cases of tying in the EU to date concerned tying where the dominant undertaking leveraged its dominance in one market to distort or eliminate competition in an otherwise competitive market.
As the foregoing demonstrates, the EC’s decision is based on a fundamental misunderstanding of the nature and evolution of the market for smartphones and associated applications. The statement by Commissioner Vestager quoted above — that “What would serve competition is to have more players” — belies this misunderstanding and highlights the erroneous assumptions underpinning the Commission’s analysis, which is wedded to a theory of market competition that was long ago thrown out by economists.
And, thankfully, it appears that the FTC Chairman is aware of at least some of the flaws in the EC’s conclusions.
Google will undoubtedly appeal the Commission’s decision. For the sakes of the millions of European consumers who rely on Android-based phones and the millions of software developers who provide Android apps, let’s hope that they succeed.
By Pinar Akman, Professor of Law, University of Leeds*
The European Commission’s decision in Google Android cuts a fine line between punishing a company for its success and punishing a company for falling afoul of the rules of the game. Which side of the line it actually falls on cannot be fully understood until the Commission publishes its full decision. Much depends on the intricate facts of the case. As the full decision may take months to come, this post offers merely the author’s initial thoughts on the decision on the basis of the publicly available information.
The eye-watering fine of $5.1 billion — which together with the fine of $2.7 billion in the Google Shopping decision from last year would (according to one estimate) suffice to fund for almost one year the additional yearly public spending necessary to eradicate world hunger by 2030 — will not be further discussed in this post. This is because the fine is assumed to have been duly calculated on the basis of the Commission’s relevant Guidelines, and, from a legal and commercial point of view, the absolute size of the fine is not as important as the infringing conduct and the remedy Google will need to adopt to comply with the decision.
First things first. This post proceeds on the premise that the aim of competition law is to prevent the exclusion of competitors that are (at least) as efficient as the dominant incumbent, whose exclusion would ultimately harm consumers.
Next, it needs to be noted that the Google Android case is a more conventional antitrust case than Google Shopping in the sense that one can at least envisage a potentially robust antitrust theory of harm in the former case. If a dominant undertaking ties its products together to exclude effective competition in some of these markets or if it pays off customers to exclude access by its efficient competitors to consumers, competition law intervention may be justified.
The central question in Google Android is whether on the available facts this appears to have happened.
What we know and market definition
The premise of the case is that Google used its dominance in the Google Play Store (which enables users to download apps onto their Android phones) to “cement Google’s dominant position in general internet search.”
It is interesting that the case appears to concern a dominant undertaking leveraging its dominance from a market in which it is dominant (Google Play Store) into another market in which it is also dominant (internet search). As far as this author is aware, most (if not all?) cases of tying in the EU to date concerned tying where the dominant undertaking leveraged its dominance in one market to distort or eliminate competition in an otherwise competitive market.
Thus, for example, in Microsoft (Windows Operating System —> media players), Hilti (patented cartridge strips —> nails), and Tetra Pak II (packaging machines —> non-aseptic cartons), the tied market was actually or potentially competitive, and this was why the tying was alleged to have eliminated competition. It will be interesting to see which case the Commission uses as precedent in its decision — more on that later.
Also noteworthy is that the Commission does not appear to have defined a separate mobile search market that would have been competitive but for Google’s alleged leveraging. The market has been defined as the general internet search market. So, according to the Commission, the Google Search App and Google Search engine appear to be one and the same thing, and desktop and mobile devices are equivalent (or substitutable).
Finding mobile and desktop devices to be equivalent to one another may have implications for other cases including the ongoing appeal in Google Shopping where, for example, the Commission found that “[m]obile [apps] are not a viable alternative for replacing generic search traffic from Google’s general search results pages” for comparison shopping services. The argument that mobile apps and mobile traffic are fundamental in Google Android but trivial in Google Shopping may not play out favourably for the Commission before the Court of Justice of the EU.
Another interesting market definition point is that the Commission has found Apple not to be a competitor to Google in the relevant market defined by the Commission: the market for “licensable smart mobile operating systems.” Apple does not fall within that market because Apple does not license its mobile operating system to anyone: Apple’s model eliminates all possibility of competition from the start and is by definition exclusive.
Although there is some internal logic in the Commission’s exclusion of Apple from the upstream market that it has defined, is this not a bit of a definitional stop? How can Apple compete with Google in the market as defined by the Commission when Apple allows only itself to use its operating system only on devices that Apple itself manufactures?
To be fair, the Commission does consider there to be some competition between Apple and Android devices at the level of consumers — just not sufficient to constrain Google at the upstream, manufacturer level.
Nevertheless, the implication of the Commission’s assessment that separates the upstream and downstream in this way is akin to saying that the world’s two largest corn producers that produce the corn used to make corn flakes do not compete with one another in the market for corn flakes because one of them uses its corn exclusively in its own-brand cereal.
Supply-side substitutability may also be taken into account when defining markets in those situations in which its effects are equivalent to those of demand substitution in terms of effectiveness and immediacy. This means that suppliers are able to switch production to the relevant products and market them in the short term….
Apple could — presumably — rather immediately and at minimal cost produce and market a version of iOS for use on third-party device makers’ devices. By the Commission’s own definition, it would seem to make sense to include Apple in the relevant market. Nevertheless, it has apparently not done so here.
The message that the Commission sends with the finding is that if Android had not been open source and freely available, and if Google competed with Apple with its own version of a walled-garden built around exclusivity, it is possible that none of its practices would have raised any concerns. Or, should Apple be expecting a Statement of Objections next from the EU Commission?
Is Microsoft really the relevant precedent?
Given that Google Android appears to revolve around the idea of tying and leveraging, the EU Commission’s infringement decision against Microsoft, which found an abusive tie in Microsoft’s tying of Windows Operating System with Windows Media Player, appears to be the most obvious precedent, at least for the tying part of the case.
There are, however, potentially important factual differences between the two cases. To take just a few examples:
Microsoft charged for the Windows Operating System, whereas Google does not;
Microsoft tied the setting of Windows Media Player as the default to OEMs’ licensing of the operating system (Windows), whereas Google ties the setting of Search as the default to device makers’ use of other Google apps, while allowing them to use the operating system (Android) without any Google apps; and
Downloading competing media players was difficult due to download speeds and lack of user familiarity, whereas it is trivial and commonplace for users to download apps that compete with Google’s.
Moreover, there are also some conceptual hurdles in finding the conduct to be that of tying.
First, the difference between “pre-installed,” “default,” and “exclusive” matters a lot in establishing whether effective competition has been foreclosed. The Commission’s Press Release notes that to pre-install Google Play, manufacturers have to also pre-install Google Search App and Google Chrome. It also states that Google Search is the default search engine on Google Chrome. The Press Release does not indicate that Google Search App has to be the exclusive or default search app. (It is worth noting, however, that the Statement of Objections in Google Android did allege that Google violated EU competition rules by requiring Search to be installed as the default. We will have to await the decision itself to see if this was dropped from the case or simply not mentioned in the Press Release).
In fact, the fact that the other infringement found is that of Google’s making payments to manufacturers in return for exclusively pre-installing the Google Search App indirectly suggests that not every manufacturer pre-installs Google Search App as the exclusive, pre-installed search app. This means that any other search app (provider) can also (request to) be pre-installed on these devices. The same goes for the browser app.
Of course, regardless, even if the manufacturer does not pre-install competing apps, the consumer is free to download any other app — for search or browsing — as they wish, and can do so in seconds.
In short, pre-installation on its own does not necessarily foreclose competition, and thus may not constitute an illegal tie under EU competition law. This is particularly so when download speeds are fast (unlike the case at the time of Microsoft) and consumers regularly do download numerous apps.
What may, however, potentially foreclose effective competition is where a dominant undertaking makes payments to stop its customers, as a practical matter, from selling its rivals’ products. Intel, for example, was found to have abused its dominant position through payments to a computer retailer in return for its not selling computers with its competitor AMD’s chips, and to computer manufacturers in return for delaying the launch of computers with AMD chips.
In Google Android, the exclusivity provision that would require manufacturers to pre-install Google Search App exclusively in return for financial incentives may be deemed to be similar to this.
Having said that, unlike in Intel where a given computer can have a CPU from only one given manufacturer, even the exclusive pre-installation of the Google Search App would not have prevented consumers from downloading competing apps. So, again, in theory effective competition from other search apps need not have been foreclosed.
It must also be noted that just because a Google app is pre-installed does not mean that it generates any revenue to Google — consumers have to actually choose to use that app as opposed to another one that they might prefer in order for Google to earn any revenue from it. The Commission seems to place substantial weight on pre-installation which it alleges to create “a status quo bias.”
The concern with this approach is that it is not possible to know whether those consumers who do not download competing apps do so out of a preference for Google’s apps or, instead, for other reasons that might indicate competition not to be working. Indeed, one hurdle as regards conceptualising the infringement as tying is that it would require establishing that a significant number of phone users would actually prefer to use Google Play Store (the tying product) without Google Search App (the tied product).
This is because, according to the Commission’s Guidance Paper, establishing tying starts with identifying two distinct products, and
[t]wo products are distinct if, in the absence of tying or bundling, a substantial number of customers would purchase or would have purchased the tying product without also buying the tied product from the same supplier.
Thus, if a substantial number of customers would not want to use Google Play Store without also preferring to use Google Search App, this would cause a conceptual problem for making out a tying claim.
In fact, the conduct at issue in Google Android may be closer to a refusal to supply type of abuse.
Refusal to supply also seems to make more sense regarding the prevention of the development of Android forks being found to be an abuse. In this context, it will be interesting to see how the Commission overcomes the argument that Android forks can be developed freely and Google may have legitimate business reasons in wanting to associate its own, proprietary apps only with a certain, standardised-quality version of the operating system.
More importantly, the possible underlying theory in this part of the case is that the Google apps — and perhaps even the licensed version of Android — are a “must-have,” which is close to an argument that they are an essential facility in the context of Android phones. But that would indeed require a refusal to supply type of abuse to be established, which does not appear to be the case.
What will happen next?
To answer the question raised in the title of this post — whether the Google Android decision will benefit consumers — one needs to consider what Google may do in order to terminate the infringing conduct as required by the Commission, whilst also still generating revenue from Android.
This is because unbundling Google Play Store, Google Search App and Google Chrome (to allow manufacturers to pre-install Google Play Store without the latter two) will disrupt Google’s main revenue stream (i.e., ad revenue generated through the use of Google Search App or Google Search within the Chrome app) which funds the free operating system. This could lead Google to start charging for the operating system, and limiting to whom it licenses the operating system under the Commission’s required, less-restrictive terms.
As the Commission does not seem to think that Apple constrains Google when it comes to dealings with device manufacturers, in theory, Google should be able to charge up to the monopoly level licensing fee to device manufacturers. If that happens, the price of Android smartphones may go up. It is possible that there is a new competitor lurking in the woods that will grow and constrain that exercise of market power, but how this will all play out for consumers — as well as app developers who may face increasing costs due to the forking of Android — really remains to be seen.
* Pinar Akman is Professor of Law, Director of Centre for Business Law and Practice, University of Leeds, UK. This piece has not been commissioned or funded by any entity. The author has not been involved in the Google Android case in any capacity. In the past, the author wrote a piece on the Commission’s Google Shopping case, ‘The Theory of Abuse in Google Search: A Positive and Normative Assessment under EU Competition Law,’ supported by a research grant from Google. The author would like to thank Peter Whelan, Konstantinos Stylianou, and Geoffrey Manne for helpful comments. All errors remain her own. The author can be contacted here.
Today the European Commission launched its latest salvo against Google, issuing a decision in its three-year antitrust investigation into the company’s agreements for distribution of the Android mobile operating system. The massive fine levied by the Commission will dominate the headlines, but the underlying legal theory and proposed remedies are just as notable — and just as problematic.
The nirvana fallacy
It is sometimes said that the most important question in all of economics is “compared to what?” UCLA economist Harold Demsetz — one of the most important regulatory economists of the past century — coined the term “nirvana fallacy” to critique would-be regulators’ tendency to compare messy, real-world economic circumstances to idealized alternatives, and to justify policies on the basis of the discrepancy between them. Wishful thinking, in other words.
The Commission’s Android decision falls prey to the nirvana fallacy. It conjures a world in which Google offers its Android operating system on unrealistic terms, prohibits it from doing otherwise, and neglects the actual consequences of such a demand.
The idea at the core of the Commission’s decision is that by making its own services (especially Google Search and Google Play Store) easier to access than competing services on Android devices, Google has effectively foreclosed rivals from effective competition. In order to correct that claimed defect, the Commission demands that Google refrain from engaging in practices that favor its own products in its Android licensing agreements:
At a minimum, Google has to stop and to not re-engage in any of the three types of practices. The decision also requires Google to refrain from any measure that has the same or an equivalent object or effect as these practices.
The basic theory is straightforward enough, but its application here reflects a troubling departure from the underlying economics and a romanticized embrace of industrial policy that is unsupported by the realities of the market.
In a recent interview, European Commission competition chief, Margrethe Vestager, offered a revealing insight into her thinking about her oversight of digital platforms, and perhaps the economy in general: “My concern is more about whether we get the right choices,” she said. Asked about Facebook, for example, she specified exactly what she thinks the “right” choice looks like: “I would like to have a Facebook in which I pay a fee each month, but I would have no tracking and advertising and the full benefits of privacy.”
Some consumers may well be sympathetic with her preference (and even share her specific vision of what Facebook should offer them). But what if competition doesn’t result in our — or, more to the point, Margrethe Vestager’s — prefered outcomes? Should competition policy nevertheless enact the idiosyncratic consumer preferences of a particular regulator? What if offering consumers the “right” choices comes at the expense of other things they value, like innovation, product quality, or price? And, if so, can antitrust enforcers actually engineer a better world built around these preferences?
Android’s alleged foreclosure… that doesn’t really foreclose anything
The Commission’s primary concern is with the terms of Google’s deal: In exchange for royalty-free access to Android and a set of core, Android-specific applications and services (like Google Search and Google Maps) Google imposes a few contractual conditions.
Google allows manufacturers to use the Android platform — in which the company has invested (and continues to invest) billions of dollars — for free. It does not require device makers to include any of its core, Google-branded features. But if a manufacturer does decide to use any of them, it must include all of them, and make Google Search the device default. In another (much smaller) set of agreements, Google also offers device makers a small share of its revenue from Search if they agree to pre-install only Google Search on their devices (although users remain free to download and install any competing services they wish).
Essentially, that’s it. Google doesn’t allow device makers to pick and choose between parts of the ecosystem of Google products, free-riding on Google’s brand and investments. But manufacturers are free to use the Android platform and to develop their own competing brand built upon Google’s technology.
Other apps may be installed in addition to Google’s core apps. Google Search need not be the exclusive search service, but it must be offered out of the box as the default. Google Play and Chrome must be made available to users, but other app stores and browsers may be pre-installed and even offered as the default. And device makers who choose to do so may share in Search revenue by pre-installing Google Search exclusively — but users can and do install a different search service.
Alternatives to all of Google’s services (including Search) abound on the Android platform. It’s trivial both to install them and to set them as the default. Meanwhile, device makers regularly choose to offer these apps alongside Google’s services, and some, like Samsung, have developed entire customized app suites of their own. Still others, like Amazon, pre-install no Google apps and use Android without any of these constraints (and whose Google-free tablets are regularly ranked as the best-rated and most popular in Europe).
By contrast, Apple bundles its operating system with its devices, bypasses third-party device makers entirely, and offers consumers access to its operating system only if they pay (lavishly) for one of the very limited number of devices the company offers, as well. It is perhaps not surprising — although it is enlightening — that Apple earns more revenue in an average quarter from iPhone sales than Google is reported to have earnedin total from Android since it began offering it in 2008.
Reality — and the limits it imposes on efforts to manufacture nirvana
The logic behind Google’s approach to Android is obvious: It is the extension of Google’s “advertisers pay” platform strategy to mobile. Rather than charging device makers (and thus consumers) directly for its services, Google earns its revenue by charging advertisers for targeted access to users via Search. Remove Search from mobile devices and you remove the mechanism by which Google gets paid.
It’s true that most device makers opt to offer Google’s suite of services to European users, and that most users opt to keep Google Search as the default on their devices — that is, indeed, the hoped-for effect, and necessary to ensure that Google earns a return on its investment.
That users often choose to keep using Google services instead of installing alternatives, and that device makers typically choose to engineer their products around the Google ecosystem, isn’t primarily the result of a Google-imposed mandate; it’s the result of consumer preferences for Google’s offerings in lieu of readily available alternatives.
The EU decision against Google appears to imagine a world in which Google will continue to develop Android and allow device makers to use the platform and Google’s services for free, even if the likelihood of recouping its investment is diminished.
The Commission also assessed in detail Google’s arguments that the tying of the Google Search app and Chrome browser were necessary, in particular to allow Google to monetise its investment in Android, and concluded that these arguments were not well founded. Google achieves billions of dollars in annual revenues with the Google Play Store alone, it collects a lot of data that is valuable to Google’s search and advertising business from Android devices, and it would still have benefitted from a significant stream of revenue from search advertising without the restrictions.
But that world in which Google won’t alter its investment decisions based on a government-mandated reduction in its allowable return on investment doesn’t exist; it’s a fanciful Nirvana.
Google’s real alternatives to the status quo are charging for the use of Android, closing the Android platform and distributing it (like Apple) only on a fully integrated basis, or discontinuing Android.
In reality, and compared to these actual alternatives, Google’s restrictions are trivial. Remember, Google doesn’t insist that Google Search be exclusive, only that it benefit from a “leg up” by being pre-installed as the default. And on this thin reed Google finances the development and maintenance of the (free) Android operating system and all of the other (free) apps from which Google otherwise earns little or no revenue.
It’s hard to see how consumers, device makers, or app developers would be made better off without Google’s restrictions, but in the real world in which the alternative is one of the three manifestly less desirable options mentioned above.
Missing the real competition for the trees
What’s more, while ostensibly aimed at increasing competition, the Commission’s proposed remedy — like the conduct it addresses — doesn’t relate to Google’s most significant competitors at all.
Facebook, Instagram, Firefox, Amazon, Spotify, Yelp, and Yahoo, among many others, are some of the most popular apps on Android phones, including in Europe. They aren’t foreclosed by Google’s Android distribution terms, and it’s even hard to imagine that they would be more popular if only Android phones didn’t come with, say, Google Search pre-installed.
It’s a strange anticompetitive story that has Google allegedly foreclosing insignificant competitors while apparently ignoring its most substantial threats.
The primary challenges Google now faces are from Facebook drawing away the most valuable advertising and Amazon drawing away the most valuable product searches (and increasingly advertising, as well). The fact that Google’s challenged conduct has never shifted in order to target these competitors as their threat emerged, and has had no apparent effect on these competitive dynamics, says all one needs to know about the merits of the Commission’s decision and the value of its proposed remedy.
In reality, as Demsetz suggested, Nirvana cannot be designed by politicians, especially in complex, modern technology markets. Consumers’ best hope for something close — continued innovation, low prices, and voluminous choice — lies in the evolution of markets spurred by consumer demand, not regulators’ efforts to engineer them.