The Case Against Google Search?

An act of Congress authorized Judge Learned Hand of the Second Circuit Court of Appeals to speak for the Supreme Court when he wrote the Alcoa decision in 1945, after several Supreme Court Justices had been disqualified from hearing the appeal. The issue was whether the government’s monopolization case against the sole U.S. producer  of virgin aluminum ingot should have been dismissed. Alcoa had market dominance in part because of superior production techniques embodied and protected by a pool of patents. One allegation was that the firm had strategically expanded its fixed capacity so as to stymie the growth of its rivals’ market share. “The successful competitor,” Judge Hand said, “having been urged to compete, must not be turned upon when he wins.”

The folks at Google must be wondering whether this admonition is all but forgotten. The U.S. Federal Trade Commission, state Attorneys General in Texas, New York, California, Ohio, Mississippi and Oklahoma, competition authorities in Europe, South America, and Asia and a legion of prospective private plaintiffs are all investigating whether Google manipulates its search results to favor its own properties over its competitors.

While the FTC can try to seek relief under Section 5 of the Federal Trade Commission Act outlawing “unfair methods of competition,” the legal recourse for would-be private plaintiffs—shopping sites like Yelp, Twenga and Foundem, travel sites like Expedia and Trip Advisor, and other “specialty search engines”—is limited to Section 2, the anti-monopolization provision of the Sherman Act. Unfortunately for any would-be plaintiff, a Section 2 monopolization claim based on Google’s search results is likely to run into market definition problems and to stretch the limits of substantive monopolization law.

The Microsoft analogy

On first impression, the analogy to the Microsoft case would seem to fit quite well: both companies are high-tech “platform monopolists” that dominate markets for an essential function in information technology. Their policies affect competition in a host of complementary adjacent markets. As a monopolist in the market for PC operating systems, Microsoft violated Section 2 by maintaining its dominant position in the OS market through anticompetitive conduct aimed at firms in adjacent markets. The court ruled that anticompetitive conduct that foreclosed Netscape from the web browser market served to maintain Microsoft’s monopoly in the OS market because browsers posed a nascent but direct threat to the dominance of Windows OS. The D.C. Circuit Court of Appeals agreed that keeping Netscape down by anticompetitive conduct unlawfully maintained the OS monopoly in violation of Section 2.

Another Section 2 ruling against Microsoft for “leveraging” its monopoly in the OS market, on the other hand, failed on appeal. The appeals court analyzed the leveraging theory as attempted monopolization, to prevail on which, the court said, the “plaintiffs must as a threshold matter show that the browser market can be monopolized, i.e., that a hypothetical monopolist in that market could enjoy market power. This, in turn, requires plaintiffs (1) to define the relevant market and (2) to demonstrate that substantial barriers to entry protect that market.”

Thus, there were two Section 2 theories at work in Microsoft. Under the first theory, monopoly maintenance, a firm may be liable when its anticompetitive conduct serves to maintain the firm’s existing monopoly power. Under the second theory, attempted monopolization, a firm may be liable for anticompetitive conduct that creates a dangerous probability of monopolizing a defined market.

In a case against Google, either of these Section 2 theories potentially could apply. But, as the Microsoft court requires, the plaintiffs will first have to define a relevant product market in which Google can enjoy market power protected by substantial barriers to entry.

Is there a product market for search?

New criticism of the market definition exercise in antitrust litigation has emerged in recent years by antitrust scholars asking, essentially, why courts should engage in a questionable and potentially misleading practice (market definition) that plays no role whatsoever in the work of professional economists. Nevertheless, the market definition exercise persists as an organizing principle and a bridge—clumsy as it may be—between observable evidence and the unobservable parameters in the mathematical concepts of elasticity and market power. For better or worse, the court’s view of market definition reverberates throughout the economic reasoning of an antitrust litigation. It makes sense, therefore, to ask whether “search” defines a relevant market, whether Google can posses market power in the market, and, if so, whether that market power is protected by barriers to entry.

One recent ComScore estimate of Google’s market share of “core search” showed this:

Core Search Entity

Explicit Core Search Share (%)

Jan-12 Feb-12 Point Change
Total Explicit Core Search 100.0% 100.0% N/A
1  Google Sites 66.2% 66.4% 0.2
2  Microsoft Sites 15.2% 15.3% 0.1
3  Yahoo! Sites 14.1% 13.8% -0.3
4  Ask Network 3.0% 3.0% 0.0
5  AOL, Inc. 1.6% 1.5% -0.1

There are 17.6 billion explicit core searches annually. ComScore explains that “explicit core searches” exclude contextually driven searches that do not reflect specific user intent to interact with the search results. If the table accurately captures the U.S. product market for search (in the sense that no further market redefinition is necessary to insure that all the firms whose product is substitutable in the relevant market are included), then Google’s share is 66.4%, at or close to the level at which antitrust traditionally presumes monopoly power (about 70%).

But, as the DOJ and FTC go about defining a market, an initial proposed market definition may have to be broadened to include successively larger numbers of firms offering a greater variety of potential substitutes until a market is defined in which a hypothetical monopolist can exercise market power by raising prices above marginal cost for a sustained period of time. The redefinition process hinges market definition on the percentage of purchases foregone in response to a small but non-transitory percentage increase in price by a hypothetical monopolist. A price of zero—search engines obviously compete on some basis other than price—undermines the underlying intellectual basis for the conventional understanding of market definition.

The zero price level wreaks even more havoc on the concept of market power, but note first that the process of defining a market in search is confounded further by the nature of the “product” itself and the internet “platform” on which it depends. We need to ask whether Google and the websites of the other firms listed by ComScore as participants in a core search market really reflect substitutability. It is far more likely that the list reflects an a priori selection of superficially similar websites offering generic search rather than a set of internet destinations that users consider substitutable.

Perhaps more than any other “product,” the purpose of a user’s search greatly influences the set of acceptable substitute websites. The ComScore list only describes a product market if searches are limited to some notion of “generic search” calling for “generic results” from a “generic search engine.” But many (if not most) of the searches performed by Google users would probably not fit this description. For example, a Google user out to purchase a digital camera might never think of substituting another generic search engine for Google, although Amazon or EBay, which are not listed in the “core search” market, might be perfectly acceptable. A user looking for a map might choose Google but substitute with MapQuest and not Bing, and so on.

Users navigate to web pages driven entirely by preferences and habits, so there is no observable means for classifying websites into sets of substitutable products. Unless we are prepared to classify “products” on the basis of what is in the user’s mind when the search is performed market definition in search is likely to be hopelessly ad hoc.

Can Google possess market power?

Even if one accepts that “core search” defines a sensible antitrust product market, can Google enjoy market power in it? Google collects revenue from advertisers through multiple channels, but its search results are free, so market power in the traditional sense of raising price above marginal cost can only be exercised (or attempt to be exercised) in the advertising market and not in the search market. Indeed, with a zero price the common definitions of market power, such as the Lerner index, L = P – MC / P , completely break down. For search results we have P = 0 and MC ≈ 0, so it is nearly impossible for Google to accrete any appreciable market power using this kind of definition.

This comports with intuition about the evils of monopoly. Monopolists possess market power if they can raise prices above a reasonably competitive level and consumers have nowhere else to turn. In so doing consumer surplus is appropriated by the higher price and lower output, the famous “deadweight loss triangle.” If price can never rise above zero, in what sense can Google possess market power that permits it to appropriate consumer surplus?

The plaintiffs’ side at this point might counter that the Supreme Court in American Tobacco and elsewhere has recognized market power as “the power to raise prices or exclude competition.” The trouble is that Google does not appear to have the kind of power that would enable it to exclude competition in something resembling the core search market. Any exclusionary power it could exercise occurs within Google search results, in the rivalry between links that appears only after the user already has chosen Google to perform the initial search. The fact that Google’s exclusionary market power arises in an adjacent complementary “market” will affect the plaintiff’s choice of legal theory, but first note that even if core search makes sense as an antitrust market in which Google (somehow) enjoys market power, that power would still need to be durable.

Can market power in search be durable?

Accepting the proposition that Google’s share of the core search market really does reflect market power (without looking too deeply into its precise nature), it remains to be determined whether it can be durable. In a conventional Section 2 case the durability of the monopolist’s market power is usually evident from observable barriers to entry that keep competition at bay. Alcoa had patents; AT&T, a government charter; Microsoft, a popular OS platform with a tremendous installed base of hardware, applications and data impossible ever to duplicate.

Antitrust insists that unlawful monopolization involve durable market power because it tolerates (and even promotes) non-durable, “first mover advantage” and other “quasi-monopoly rents” that accrue to an innovator or a maverick. Quasi-monopoly rents accrue in the period before competitive firms can enter and are the benefits of monopoly to which Justice Scalia was referring in his opinion in Verizon v. Trinko without, apparently, realizing it. Because antitrust is concerned with cases of monopoly not susceptible to cure by the entry of a new competitor, only durable monopoly draws antitrust scrutiny. If intervention is needed it is precisely because there is anticompetitive conduct affecting a market into which competitors cannot enter.

If Google has an anticompetitive mechanism by which to maintain its high market share or lock-in its users, it is not immediately apparent. Billions of internet web pages can be accessed by a standardized web address, so switching between competing destinations is effectively costless. The open architecture of the internet allows any entrepreneur to put up a web page in competition with Google and have virtually as much access as Google to the pool of end users and supply of inputs. The market is attractive enough for the likes of Microsoft and Yahoo! and Google can do little or nothing beyond being a formidable competitor to forestall their entry or expansion.

Critics point to Google’s impossible-to-match inventory of web pages and the scale of its web searching operation as a barrier to entry. But these features do not exclude new entrants. The whole litany of Google’s advantages over its rivals and potential rivals still leaves an observer wondering how Google maintains market share when the price of its product and the cost of switching are both zero. Where consumer behavior is solely a matter of preference and habit, users will persist in navigating to Google only if they derive greater utility from going there than going elsewhere. This kind of success, Judge Hand might remind us, should not be turned against the winner.

Scylla and Carbides

Even if a plaintiff can clear the hurdles of defining a sensible market in which Google possesses durable market power, a case challenging alleged manipulation of search results will still require a legally sufficient theory of liability. For specialty search websites that allege that Google favors its own competing sites in its search results the choice between monopoly maintenance and attempted monopolization is not a happy one.

The injury to competition in a monopoly maintenance claim is that the incumbent’s monopoly position is prolonged beyond that which would have resulted from normal market forces acting in the absence of the anticompetitive restraint. Given that the antitrust standing doctrine requires the conduct that causes the alleged antitrust injury to the private plaintiff be the same conduct that injures competition, a monopoly maintenance claim against Google would require proof that by manipulating its search results Google thereby extended its monopoly in core search, putting the plaintiff in the untenable position of having to demonstrate how conduct that hurt or crippled its position in a specialty search market materially contributed to the durability of Google’s monopoly in the core search market. A monopoly maintenance claim, therefore, is not likely to be a good fit.

An attempted monopolization theory, on the other hand, presumably would put the focus where it more properly belongs, on the competitive effects of Google’s conduct in one or more distinct specialty search markets. Moreover, if Google’s market power resides in its power to exclude competition it would appear to reside in such an adjacent market. But, the plaintiff will then have  to define such a market before the legal inquiry can move on to the question of whether there is a dangerous probability that Google will monopolize it. It may be instructive that the attempted monopolization and tying claims in the Microsoft case ultimately came to naught because of the obstacles to proving a separate browser market, the price of which was zero.

Again, assuming that the market definition hurdles could be overcome, the plaintiffs may have a difficult time proving that Google is dangerously close to monopoly. Here are some figures from Experian Hitwise on “travel agency search:”

Rank

Website

Domain

Percentage of Visits

Previous Position

1 Expedia http://www.expedia.com 13.19% 1
2 Yahoo! Travel travel.yahoo.com 9.71% 3
3 priceline.com http://www.priceline.com 9.67% 2
4 Orbitz http://www.orbitz.com 6.49% 4
5 Travelocity http://www.travelocity.com 5.56% 5
6 Hotwire http://www.hotwire.com 5.14% 6
7 CheapOair.com http://www.cheapoair.com 4.85% 8
8 Kayak http://www.kayak.com 3.71% 9
9 Cheap Tickets http://www.cheaptickets.com 3.52% 10
10 bookingbuddy http://www.bookingbuddy.com 3.00%

Here is a similar ranking of comparison shopping sites:

Rank

Site

eBizMBA Rank

Unique Monthly Visitors

Alexa Rank

Quantcast Rank

1 BizRate

298

15,000,000

795

58

2 NexTag

436

13,500,000

595

662

3 shopLocal

617

6,200,000

1,382

302

4 slickdeals

630

6,000,000

690

366

5 woot

642

5,900,000

717

337

6 coupons

762

4,700,000

1,714

505

7 pronto

768

4,650,000

1,709

490

8 shopzilla

788

4,500,000

2,025

185

9 ShopAtHome

825

4,450,000

2,018

283

10 smarter

829

4,400,000

2,003

357

Such “top 10” lists say nothing about consumer behavior so they do not define an antitrust market. But they do illustrate how difficult it is to define a “specialty search market” and to establish that anyone, even Google, is dangerously close to monopolizing it.

Guilt by analogy

Bypassing these technical niceties, advocates for Google’s victims are pursuing a campaign of guilt by analogy. Monopoly leveraging claims arise when a firm controls a platform which inspires complementary goods or services. For example, in the 1980s IBM attempted to claim copyright protection for software application interfaces for its mainframes in order to control the application markets; AT&T used its monopoly control over local telephone to foreclose the long distance market to MCI; Conditions placed on the Comcast-NBCU merger are intended to prevent Comcast from using its market power in distribution to foreclose rivals of NBCU from the content market.

These and other examples of denial of access by platform monopolists were trotted out in support of a case against Google by attorney Gary L. Reback speaking on behalf of travel search sites at the American Antitrust Institute’s annual conference last month in Washington. As if to anticipate the criticism that these analogies were inapposite, Mr. Reback insisted that search results were no mere “lists of sites,” but rather a collection of links that take you someplace. Thus, they are “transmission interconnection protocols” that direct traffic “in an order of priority.” If you look at the internet as a “set of telecommunications pipes,” Mr. Reback said, “we can see that Google is a switch and Google’s search results are transmission interconnection priorities.”

Instead of making the analogy more apt, Mr. Reback’s comparison of search results to telecommunications protocols or network switches only reinforces why search results are not like network access: the links displayed by Google are not bottlenecks. Beyond its influence over users’ preferences, Google has no means of control over its users’ access to any website on the web. To characterize Google as a “gatekeeper” is to attribute to it power that it does not possesses.

Antitrust cases are notoriously complex and difficult to mount successfully, particularly against firms acting independently, where great care must be taken to distinguish between unilateral conduct that is legitimate, pro-competitive, or benign and unilateral conduct that is anticompetitive and harmful to consumers. While there is no doubt that Google is a company of tremendous power and influence, there may be less to Google’s antitrust transgressions, at least when it comes to search results, than meets the eye.

-jr

Advertisements
This entry was posted in Antitrust Current Events and tagged , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s