Search Results

Back in September, the Senate Judiciary Committee’s Antitrust Subcommittee held a hearing on “The Power of Google: Serving Consumers or Threatening Competition?” Given the harsh questioning from the Subcommittee’s Chairman Herb Kohl (D-WI) and Ranking Member Mike Lee (R-UT), no one should have been surprised by the letter they sent yesterday to the Federal Trade Commission asking for a “thorough investigation” of the company. At least this time the danger is somewhat limited: by calling for the FTC to investigate Google, the senators are thus urging the agency to do . . . exactly what it’s already doing.

Read the full piece here.

In my series of three posts (here, here and here) drawn from my empirical study on search bias I have examined whether search bias exists, and, if so, how frequently it occurs.  This, the final post in the series, assesses the results of the study (as well as the Edelman & Lockwood (E&L) study to which it responds) to determine whether the own-content bias I’ve identified is in fact consistent with anticompetitive foreclosure or is otherwise sufficient to warrant antitrust intervention.

As I’ve repeatedly emphasized, while I refer to differences among search engines’ rankings of their own or affiliated content as “bias,” without more these differences do not imply anticompetitive conduct.  It is wholly unsurprising and indeed consistent with vigorous competition among engines that differentiation emerges with respect to algorithms.  However, it is especially important to note that the theories of anticompetitive foreclosure raised by Google’s rivals involve very specific claims about these differences.  Properly articulated vertical foreclosure theories proffer both that bias is (1) sufficient in magnitude to exclude Google’s rivals from achieving efficient scale, and (2) actually directed at Google’s rivals.  Unfortunately for search engine critics, their theories fail on both counts.  The observed own-content bias appears neither to be extensive enough to prevent rivals from gaining access to distribution nor does it appear to target Google’s rivals; rather, it seems to be a natural result of intense competition between search engines and of significant benefit to consumers.

Vertical foreclosure arguments are premised upon the notion that rivals are excluded with sufficient frequency and intensity as to render their efforts to compete for distribution uneconomical.  Yet the empirical results simply do not indicate that market conditions are in fact conducive to the types of harmful exclusion contemplated by application of the antitrust laws.  Rather, the evidence indicates that (1) the absolute level of search engine “bias” is extremely low, and (2) “bias” is not a function of market power, but an effective strategy that has arisen as a result of serious competition and innovation between and by search engines.  The first finding undermines competitive foreclosure arguments on their own terms, that is, even if there were no pro-consumer justifications for the integration of Google content with Google search results.  The second finding, even more importantly, reveals that the evolution of consumer preferences for more sophisticated and useful search results has driven rival search engines to satisfy that demand.  Both Bing and Google have shifted toward these results, rendering the complained-of conduct equivalent to satisfying the standard of care in the industry–not restraining competition.

A significant lack of search bias emerges in the representative sample of queries.  This result is entirely unsurprising, given that bias is relatively infrequent even in E&L’s sample of queries specifically designed to identify maximum bias.  In the representative sample, the total percentage of queries for which Google references its own content when rivals do not is even lower—only about 8%—meaning that Google favors its own content far less often than critics have suggested.  This fact is crucial and highly problematic for search engine critics, as their burden in articulating a cognizable antitrust harm includes not only demonstrating that bias exists, but further that it is actually competitively harmful.  As I’ve discussed, bias alone is simply not sufficient to demonstrate any prima facie anticompetitive harm as it is far more often procompetitive or competitively neutral than actively harmful.  Moreover, given that bias occurs in less than 10% of queries run on Google, anticompetitive exclusion arguments appear unsustainable.

Indeed, theories of vertical foreclosure find virtually zero empirical support in the data.  Moreover, it appears that, rather than being a function of monopolistic abuse of power, search bias has emerged as an efficient competitive strategy, allowing search engines to differentiate their products in ways that benefit consumers.  I find that when search engines do reference their own content on their search results pages, it is generally unlikely that another engine will reference this same content.  However, the fact that both this percentage and the absolute level of own content inclusion is similar across engines indicates that this practice is not a function of market power (or its abuse), but is rather an industry standard.  In fact, despite conducting a much smaller percentage of total consumer searches, Bing is consistently more biased than Google, illustrating that the benefits search engines enjoy from integrating their own content into results is not necessarily a function of search engine size or volume of queries.  These results are consistent with a business practice that is efficient and at significant tension with arguments that such integration is designed to facilitate competitive foreclosure.

Inclusion of own content accordingly appears to be just one dimension upon which search engines have endeavored to satisfy and anticipate heterogeneous and dynamic consumer preferences.  Consumers today likely make strategic decisions as to which engine to run their searches on, and certainly expect engines to return far more complex results than were available just a few years ago. For example, over the last few years, search engines have begun “personalizing” search results, tailoring results pages to individual searchers, and allowing users’ preferences to be reflected over time.  While the traditional “10 blue links” results page is simply not an effective competitive strategy today, it appears that own-content inclusion is.  By developing and offering their own products in search results, engines are better able to directly satisfy consumer desires.

Moreover, the purported bias does not involve attempts to prominently display Google’s own general or vertical search content over that of rivals.  Consider the few queries in Edelman & Lockwood’s small sample of terms for which Google returned Google content within the top three results but neither Bing nor Blekko referenced the same content anywhere on their first page of results.  For the query “voicemail,” for example, Google refers to both Google Voice and Google Talk; both instances appear unrelated to the grievances of general and vertical search rivals.  The query “movie” results in a OneBox with the next 3 organic results including movie.com, fandango.com, and yahoo.movies.com.  The single instance in Edelman & Lockwood’s sample for which Google ranks its own content in the Top 3 positions but this content is not referred to at all on Bing’s first page of results is a link to blogger.com in response to the query “blog.”  It is difficult to construct a story whereby this result impedes Bing’s competitive position.  In fact, none of these examples suggests that efforts to anticompetitively foreclose rivals are in play.  To the contrary, each seems to be a result of simple and expected procompetitive product differentiation.

Overall, the evidence reveals very little search engine bias, and no overwhelming or systematic biasing by Google against  search competitors.  Indeed, the data simply do not support claims that own-content bias is of the nature, quality, or magnitude to generate plausible antitrust concerns.  To the contrary, the results strongly suggest that own-content bias fosters natural and procompetitive product differentiation.  Accordingly, search bias is likely beneficial to consumers—and is clearly not indicative of harm to consumer welfare.

Antitrust regulators should proceed with caution when evaluating such claims given the overwhelmingly consistent economic learning concerning the competitive benefits generally of vertical integration for consumers.  Serious care must be taken in order not to deter vigorous competition between search engines and the natural competitive process between rivals constantly vying to best one another to serve consumers.

Filed under: advertising, antitrust, business, economics, exclusionary conduct, google, Internet search, law and economics, monopolization, technology Tagged: Bias, Bing, Blekko, Competition law, Edelman, google, microsoft, Web search engine

TOTM readers know that I’ve long been skeptical of claims that expansive use of Section 5 of the FTC Act will prove productive for consumers.  I’ve been critical of recent applications of Section 5 such as Intel and N-Data.  Now comes yet another FTC consent decree in PoolCorp.  I’m still skeptical.  Indeed, PoolCorp appears to provide ammunition for those (like me) who have criticized the Commission’s stance on expansive use of Section 5 precisely upon the grounds that it can and will be applied to conduct that is either competitively neutral or even procompetitive.

Commissioner Rosch’s dissent makes many of the key points.  Indeed his opening line gets straight to the point: “This case presents the novel situation of a company willing to enter into a consent decree notwithstanding a lack of evidence indicating that a violation has occurred.”

Before getting to specifics, the sharp disagreement between the majority and Commissioner Rosch on both the most basic of facts and economic principles is hard to miss, and gives the entire exchange a rather peculiar feel.  Here’s an example.  The majority describes the case as a standard application of a “Raising Rivals’ Costs” theory, citing Krattenmaker & Salop.  The allegation is that:

Specifically, the Complaint alleges that PoolCorp, which possesses monopoly power in many local distribution markets, threatened its suppliers (i.e., pool product manufacturers) that it would no longer distribute a manufacturer’s products on a nationwide basis if that manufacturer sold its products to a new distributor that was attempting to enter a local market.

The conditions that must be satisfied for an exclusionary theory are well known.  Substantial foreclosure of a critical input is one such necessary, but not sufficient, condition for the possibility of competitive harm.  The majority argues that PoolCorp “foreclosed new entrants from obtaining pool products from manufacturers representing more than 70 percent of sales.”  But standard antitrust analysis tells us that such foreclosure is not enough to support an inference of harm to competition.  First, we must ask whether the threatened refusals to deal actually had any impact on the allegedly impaired rivals or whether they were able to easily realign supply contracts?  Second, and most fundamentally, we must ask whether the conduct at issue had any impact on competition itself, or upon consumers in the form of higher prices, reduced output, lower quality, etc.?

Here is where things get, well, weird.

Did PoolCorp’s actions actually disadvantage any rivals?  The majority concedes that “Some of PoolCorp’s targets were able to survive by purchasing pool products from other distributors rather than directly from the
manufacturers.” Well, that doesn’t sound too bad for the Commission.  If a few firms survived but others were excluded (surely the implication of the sentence), we should continue our analysis.  But was there actually any foreclosure?

Here’s Commissioner Rosch in dissent:

“The investigation revealed that PoolCorp’s demands were not honored by manufacturers.”

What about those potential entrants that were excluded — the ones that were not so lucky as the surviving targets the majority mentions?

“Another problem with this case is that no entrants were actually excluded.”

Yikes.  One gets the impression that the Commissioners are not talking about the same case.  The majority is full of broad generalizations and assertions but no real discussion of facts.  Commissioner Rosch’s dissent offers a bit more on the exclusion claim:

“The only claim to the contrary is in Paragraph 28 of the complaint, which alleges that in Baton Rouge, “the new entrant’s business ultimately failed in 2005” because of the lack of “direct access to the manufacturers’ pool products.” The complaint neglects to mention that this entrant was able to secure supplies from other sources and later sold itself to an established out-of-state distributor. Since then, that distributor, which has had full access to supplies, has been a highly effective rival to PoolCorp. Thus, to the extent PoolCorp’s threats had an effect in Baton Rouge, they may have led to more, not less, competition.”

Not good for the Commission majority.  But injury to rivals isn’t our primary concern.  What about injury to competition?  Here, things get even murkier.  The majority plainly asserts “the harm to consumers that occurred as a result was substantial” and “consumers had fewer choices and were forced to pay higher prices for pool products.”  Sounds relatively straightforward.  Once again, Commissioner Rosch’s dissent exposes disagreement over the most basic of antitrust-relevant facts (emphasis added):

“A third problem with this case is that there was no consumer injury. The investigation did not uncover price increases, service degradation, or other anticompetitive effects in any local markets.”

Rosch goes on:

The basis for the majority statement’s claim that there was “substantial” consumer harm resulting from the alleged conduct of Respondent is a mystery. The complaint contains no factual allegations of any harm to consumers, much less “substantial” harm. Likewise, there are no factual allegations in the complaint corroborating the majority’s claim that consumers “had fewer choices and were forced to pay higher prices for pool products.”

This is a real mess.  Proponents of an expanded application of Section 5 (including Commissioner Rosch) frequently argue that it is capable of being applied with certain limiting principles, including demonstration of consumer injury.  To his credit, Commissioner Rosch is sticking to his guns on consumer injury as a limiting principle here.  But the evidence that the Section 5 is too enticing a tool for the Commission in cases lacking consumer injury is mounting.  The public disagreement over basic facts — is there harm to consumers or not?  was there foreclosure or not?  if so, how much? — also does not inspire confidence that the Commission’s discretion in applying Section 5 in cases where the conduct lies outside the scope of the Sherman Act for technical reasons will be applied in a manner consistent with the consumer welfare goals of antitrust.

Those are general problems with Section 5.  As applied here, the majority opinion is also analytically incoherent.   The Commission majority must deal with the fact that there appears to be no real foreclosure as a result of PoolCorp’s conduct — recall that what the majority described as a few successful surviving firms turns out to be no actual exclusion whatsoever.  Despite the fact that absence of foreclosure or injury to rivals in a case like this is typically the end of the line for the plaintiff, the Commission doesn’t appear to be bothered at all by the lack of evidence of harm to rivals or consumers.  Responding to the fact of no foreclosure, the Commission writes:

“However, we assess consumer harm relative to market conditions that would have existed but for the respondent’s allegedly unlawful conduct. Here, PoolCorp’s strategy significantly increased a new entrant’s costs of obtaining pool products. Conduct by a monopolist that raises rivals’ costs can harm competition by creating an artificial price floor or deterring investments in quality, service and innovation.”

This doesn’t make any sense.  If there is no foreclosure, there is no risk of consumer harm.  Period.  Indeed, while the majority asserts it, there appears to be no actual evidence of consumer harm.  At a minimum, its up for serious debate.  If it were true that PoolCorp’s strategy “increased a few entrant’s cost of obtaining pool products” in practice, and that there were sufficient exclusion to create additional market power, two things would be true: (1) one would observe harm to the rival, and (2) there would be harm to competition in the form of higher prices or reduced output.  Apparently, the Commission could must neither — even when challenged by Commissioner Rosch’s dissent to do so.

One last observation.  Commissioner Rosch’s dissent hints that economic analysis in the case demonstrated that “even if” PoolCorp fully foreclosed its rivals the harm to consumers would be minimal and a waste of Commission resources.   Query: what role are agency economists playing in the Commission’s Section5 agenda?  Unfortunately, it does not appear to be a significant one.

Filed under: antitrust, economics, exclusionary conduct, federal trade commission, monopolization

In my last post, I discussed Edelman & Lockwood’s (E&L’s) attempt to catch search engines in the act of biasing their results—as well as their failure to actually do so.  In this post, I present my own results from replicating their study.  Unlike E&L, I find that Bing is consistently more biased than Google, for reasons discussed further below, although neither engine references its own content as frequently as E&L suggest.

I ran searches for E&L’s original 32 non-random queries using three different search engines—Google, Bing, and Blekko—between June 23 and July 5 of this year.  This replication is useful, as search technology has changed dramatically since E&L recorded their results in August 2010.  Bing now powers Yahoo, and Blekko has had more time to mature and enhance its results.  Blekko serves as a helpful “control” engine in my study, as it is totally independent of Google and Microsoft, and so has no incentive to refer to Google or Microsoft content unless it is actually relevant to users.  In addition, because Blekko’s model is significantly different than Google and Microsoft’s, if results on all three engines agree that specific content is highly relevant to the user query, it lends significant credibility to the notion that the content places well on the merits rather than being attributable to bias or other factors.

How Do Search Engines Rank Their Own Content?

Focusing solely upon the first position, Google refers to its own products or services when no other search engine does in 21.9% of queries; in another 21.9% of queries, both Google and at least one other search engine rival (i.e. Bing or Blekko) refer to the same Google content with their first links.

But restricting focus upon the first position is too narrow.  Assuming that all instances in which Google or Bing rank their own content first and rivals do not amounts to bias would be a mistake; such a restrictive definition would include cases in which all three search engines rank the same content prominently—agreeing that it is highly relevant—although not all in the first position. 

The entire first page of results provides a more informative comparison.  I find that Google and at least one other engine return Google content on the first page of results in 7% of the queries.  Google refers to its own content on the first page of results without agreement from either rival search engine in only 7.9% of the queries.  Meanwhile, Bing and at least one other engine refer to Microsoft content in 3.2% of the queries.  Bing references Microsoft content without agreement from either Google or Blekko in 13.2% of the queries:

This evidence indicates that Google’s ranking of its own content differs significantly from its rivals in only 7.9% of queries, and that when Google ranks its own content prominently it is generally perceived as relevant.  Further, these results suggest that Bing’s organic search results are significantly more biased in favor of Microsoft content than Google’s search results are in favor of Google’s content.

Examining Search Engine “Bias” on Google

The following table presents the percentages of queries for which Google’s ranking of its own content differs significantly from its rivals’ ranking of that same content.

Note that percentages below 50 in this table indicate that rival search engines generally see the referenced Google content as relevant and independently believe that it should be ranked similarly.

So when Google ranks its own content highly, at least one rival engine typically agrees with this ranking; for example, when Google places its own content in its Top 3 results, at least one rival agrees with this ranking in over 70% of queries.  Bing especially agrees with Google’s rankings of Google content within its Top 3 and 5 results, failing to include Google content that Google ranks similarly in only a little more than a third of queries.

Examining Search Engine “Bias” on Bing

Bing refers to Microsoft content in its search results far more frequently than its rivals reference the same Microsoft content.  For example, Bing’s top result references Microsoft content for 5 queries, while neither Google nor Blekko ever rank Microsoft content in the first position:

This table illustrates the significant discrepancies between Bing’s treatment of its own Microsoft content relative to Google and Blekko.  Neither rival engine refers to Microsoft content Bing ranks within its Top 3 results; Google and Blekko do not include any Microsoft content Bing refers to on the first page of results in nearly 80% of queries.

Moreover, Bing frequently ranks Microsoft content highly even when rival engines do not refer to the same content at all in the first page of results.  For example, of the 5 queries for which Bing ranks Microsoft content in its top result, Google refers to only one of these 5 within its first page of results, while Blekko refers to none.  Even when comparing results across each engine’s full page of results, Google and Blekko only agree with Bing’s referral of Microsoft content in 20.4% of queries.

Although there are not enough Bing data to test results in the first position in E&L’s sample, Microsoft content appears as results on the first page of a Bing search about 7 times more often than Microsoft content appears on the first page of rival engines.  Also, Google is much more likely to refer to Microsoft content than Blekko, though both refer to significantly less Microsoft content than Bing.

A Closer Look at Google v. Bing

On E&L’s own terms, Bing results are more biased than Google results; rivals are more likely to agree with Google’s algorithmic assessment (than with Bing’s) that its own content is relevant to user queries.  Bing refers to Microsoft content other engines do not rank at all more often than Google refers its own content without any agreement from rivals.  Figures 1 and 2 display the same data presented above in order to facilitate direct comparisons between Google and Bing.

As Figures 1 and 2 illustrate, Bing search results for these 32 queries are more frequently “biased” in favor of its own content than are Google’s.  The bias is greatest for the Top 1 and Top 3 search results.

My study finds that Bing exhibits far more “bias” than E&L identify in their earlier analysis.  For example, in E&L’s study, Bing does not refer to Microsoft content at all in its Top 1 or Top 3 results; moreover, Bing refers to Microsoft content within its entire first page 11 times, while Google and Yahoo refer to Microsoft content 8 and 9 times, respectively.  Most likely, the significant increase in Bing’s “bias” differential is largely a function of Bing’s introduction of localized and personalized search results and represents serious competitive efforts on Bing’s behalf.

Again, it’s important to stress E&L’s limited and non-random sample, and to emphasize the danger of making strong inferences about the general nature or magnitude of search bias based upon these data alone.  However, the data indicate that Google’s own-content bias is relatively small even in a sample collected precisely to focus upon the queries most likely to generate it.  In fact—as I’ll discuss in my next post—own-content bias occurs even less often in a more representative sample of queries, strongly suggesting that such bias does not raise the competitive concerns attributed to it.

Filed under: antitrust, business, economics, google, Internet search, law and economics, monopolization, technology Tagged: antitrust, Bias, Bing, Blekko, google, microsoft, search, Web search engine, Yahoo

Last week I linked to my new study on “search bias.”  At the time I noted I would have a few blog posts in the coming days discussing the study.  This is the first of those posts.

A lot of the frenzy around Google turns on “search bias,” that is, instances when Google references its own links or its own content (such as Google Maps or YouTube) in its search results pages.  Some search engine critics condemn such references as inherently suspect and almost by their very nature harmful to consumers.  Yet these allegations suffer from several crucial shortcomings.  As I’ve noted (see, e.g., here and here), these naked assertions of discrimination are insufficient to state a cognizable antitrust claim, divorced as they are from consumer welfare analysis.  Indeed, such “discrimination” (some would call it “vertical integration”) has a well-recognized propensity to yield either pro-competitive or competitively neutral outcomes, rather than concrete consumer welfare losses.  Moreover, because search engines exist in an incredibly dynamic environment, marked by constant innovation and fierce competition, we would expect different engines, utilizing different algorithms and appealing to different consumer preferences, to emerge.  So when search engines engage in product differentiation of this sort, there is no reason to be immediately suspicious of these business decisions.

No reason to be immediately suspicious – but there could, conceivably, be a problem.  If there is, we would want to see empirical evidence of it—of both the existence of bias, as well as the consumer harm emanating from it.  But one of the most notable features of this debate is the striking lack of empirical data.  Surprisingly little research has been done in this area, despite frequent assertions that own-content bias is commonly practiced and poses a significant threat to consumers (see, e.g., here).

My paper is an attempt to rectify this.  In the paper, I investigate the available data to determine whether and to what extent own-content bias actually occurs, by analyzing and replicating a study by Ben Edelman and Ben Lockwood (E&L) and conducting my own study of a larger, randomized set of search queries.

In this post I discuss my analysis and critique of E&L; in future posts I’ll present my own replication of their study, as well as the results of my larger study of 1,000 random search queries.  Finally, I’ll analyze whether any of these findings support anticompetitive foreclosure theories or are otherwise sufficient to warrant antitrust intervention.

E&L “investigate . . . [w]hether search engines’ algorithmic results favor their own services, and if so, which search engines do most, to what extent, and in what substantive areas.”  Their approach is to measure the difference in how frequently search engines refer to their own content relative to how often their rivals do so.

One note at the outset:  While this approach provides useful descriptive facts about the differences between how search engines link to their own content, it does little to inform antitrust analysis because Edelman and Lockwood begin with the rather odd claim that competition among differentiated search engines for consumers is a puzzle that creates an air of suspicion around the practice—in fact, they claim that “it is hard to see why results would vary . . . across search engines.”  This assertion, of course, is simply absurd.  Indeed, Danny Sullivan provides a nice critique of this claim:

It’s not hard to see why search engine result differ at all.  Search engines each use their own “algorithm” to cull through the pages they’ve collected from across the web, to decide which pages to rank first . . . . Google has a different algorithm than Bing.  In short, Google will have a different opinion than Bing.  Opinions in the search world, as with the real world, don’t always agree.

Moreover, this assertion completely discounts both the vigorous competitive product differentiation that occurs in nearly all modern product markets as well as the obvious selection effects at work in own-content bias (Google users likely prefer Google content).  This combination detaches E&L’s analysis from the consumer welfare perspective, and thus antitrust policy relevance, despite their claims to the contrary (and the fact that their results actually exhibit very little bias).

Several methodological issues undermine the policy relevance of E&L’s analysis.  First, they hand select 32 search queries and execute searches on Google, Bing, Yahoo, AOL and Ask.  This hand-selected non-random sample of 32 search queries cannot generate reliable inferences regarding the frequency of bias—a critical ingredient to understanding its potential competitive effects.  Indeed, E&L acknowledge their queries are chosen precisely because they are likely to return results including Google content (e.g., email, images, maps, video, etc.).

E&L analyze the top three organic search results for each query on each engine.  They find that 19% of all results across all five search engines refer to content affiliated with one of them.  They focus upon the first three organic results and report that Google refers to its own content in the first (“top”) position about twice as often as Yahoo and Bing refer to Google content in this position.  Additionally, they note that Yahoo is more biased than Google when evaluating the first page rather than only the first organic search result.

E&L also offer a strained attempt to deal with the possibility of competitive product differentiation among search engines.  They examine differences among search engines’ references to their own content by “compar[ing] the frequency with which a search engine links to its own pages, relative to the frequency with which other search engines link to that search engine’s pages.”  However, their evidence undermines claims that Google’s own-content bias is significant and systematic relative to its rivals’.  In fact, almost zero evidence of statistically significant own-content bias by Google emerges.

E&L find, in general, Google is no more likely to refer to its own content than other search engines are to refer to that same content, and across the vast majority of their results, E&L find Google search results are not statistically more likely to refer to Google content than rivals’ search results.

The same data can be examined to test the likelihood that a search engine will refer to content affiliated with a rival search engine.  Rather than exhibiting bias in favor of an engine’s own content, a “biased” search engine might conceivably be less likely to refer to content affiliated with its rivals.  The table below reports the likelihood (in odds ratios) that a search engine’s content appears in a rival engine’s results.

The first two columns of the table demonstrate that both Google and Yahoo content are referred to in the first search result less frequently in rivals’ search results than in their own.  Although Bing does not have enough data for robust analysis of results in the first position in E&L’s original analysis, the next three columns in Table 1 illustrate that all three engines’ (Google, Yahoo, and Bing) content appears less often on the first page of rivals’ search results than on their own search engine.  However, only Yahoo’s results differ significantly from 1.  As between Google and Bing, the results are notably similar.

E&L also make a limited attempt to consider the possibility that favorable placement of a search engine’s own content is a response to user preferences rather than anticompetitive motives.  Using click-through data, they find, unsurprisingly, that the first search result tends to receive the most clicks (72%, on average).  They then identify one search term for which they believe bias plays an important role in driving user traffic.  For the search query “email,” Google ranks its own Gmail first and Yahoo Mail second; however, E&L also find that Gmail receives only 29% of clicks while Yahoo Mail receives 54%.  E&L claim that this finding strongly indicates that Google is engaging in conduct that harms users and undermines their search experience.

However, from a competition analysis perspective, that inference is not sound.  Indeed, the fact that the second-listed Yahoo Mail link received the majority of clicks demonstrates precisely that Yahoo was not competitively foreclosed from access to users.  Taken collectively, E&L are not able to muster evidence of potential competitive foreclosure.

While it’s important to have an evidence-based discussion surrounding search engine results and their competitive implications, it’s also critical to recognize that bias alone is not evidence of competitive harm.  Indeed, any identified bias must be evaluated in the appropriate antitrust economic context of competition and consumers, rather than individual competitors and websites.  E&L’s analysis provides a useful starting point for describing how search engines differ in their referrals to their own content.  But, taken at face value, their results actually demonstrate little or no evidence of bias—let alone that the little bias they do find is causing any consumer harm.

As I’ll discuss in coming posts, evidence gathered since E&L conducted their study further suggests their claims that bias is prevalent, inherently harmful, and sufficient to warrant antitrust intervention are overstated and misguided.

Filed under: antitrust, business, economics, google, Internet search, law and economics, monopolization, technology Tagged: antitrust, Bing, google, search, search bias, Search Engines, search neutrality, Web search engine, Yahoo

[I am participating in an online “debate” at the American Constitution Society with Professor Ben Edelman.  The debate consists of an opening statement and concluding responses to be posted later in the week.  Professor Edelman’s opening statement is here.  I am cross-posting my opening statement here at TOTM.  This is my closing statement]

Professor Edelman’s opening post does little to support his case.  Instead, it reflects the same retrograde antitrust I criticized in my first post.

Edelman’s understanding of antitrust law and economics appears firmly rooted in the 1960s approach to antitrust in which enforcement agencies, courts, and economists vigorously attacked novel business arrangements without regard to their impact on consumers.  Judge Learned Hand’s infamous passage in the Alcoa decision comes to mind as an exemplar of antitrust’s bad old days when the antitrust laws demanded that successful firms forego opportunities to satisfy consumer demand.  Hand wrote:

we can think of no more effective exclusion than progressively to embrace each new opportunity as it opened, and to face every newcomer with new capacity already geared into a great organization, having the advantage of experience, trade connections and the elite of personnel.

Antitrust has come a long way since then.  By way of contrast, today’s antitrust analysis of alleged exclusionary conduct begins with (ironically enough) the U.S. v. Microsoft decision.  Microsoft emphasizes the difficulty of distinguishing effective competition from exclusionary conduct; but it also firmly places “consumer welfare” as the lodestar of the modern approach to antitrust:

Whether any particular act of a monopolist is exclusionary, rather than merely a form of vigorous competition, can be difficult to discern: the means of illicit exclusion, like the means of legitimate competition, are myriad.  The challenge for an antitrust court lies in stating a general rule for distinguishing between exclusionary acts, which reduce social welfare, and competitive acts, which increase it.  From a century of case law on monopolization under § 2, however, several principles do emerge.  First, to be condemned as exclusionary, a monopolist’s act must have an “anticompetitive effect.”  That is, it must harm the competitive process and thereby harm consumers.  In contrast, harm to one or more competitors will not suffice.

Nearly all antitrust commentators agree that the shift to consumer-welfare focused analysis has been a boon for consumers.  Unfortunately, Edelman’s analysis consists largely of complaints that would have satisfied courts and agencies in the 1960s but would not do so now that the focus has turned to consumer welfare rather than indirect complaints about market structure or the fortunes of individual rivals.

From the start, in laying out his basic case against Google, Edelman invokes antitrust concepts that are simply inapt for the facts and then goes on to apply them in a manner inconsistent with the modern consumer-welfare-oriented framework described above:

In antitrust parlance, this is tying: A user who wants only Google Search, but not Google’s other services, will be disappointed.  Instead, any user who wants Google Search is forced to receive Google’s other services too.  Google’s approach also forecloses competition: Other sites cannot compete on their merits for a substantial portion of the market – consumers who use Google to find information – because Google has kept those consumers for itself.

There are two significant errors here.  First, Edelman claims to be interested in protecting users who want only Google Search but not its other services will be disappointed.  I have no doubt such consumers exist.  Some proof that they exist is that a service has already been developed to serve them.  Professor Edelman, meet Googleminusgoogle.com.  Across the top the page reads: “Search with Google without getting results from Google sites such as Knol, Blogger and YouTube.”  In antitrust parlance, this is not tying after all.  The critical point, however, is that user preferences are being satisfied as one would expect to arise from competition.

The second error, as I noted in my first post, is to condemn vertical integration as inherently anticompetitive.  It is here that the retrograde character of Professor Edelman’s analysis (and other critics of Google, to be fair) shines brightest.  It reflects a true disconnect between the 1960s approach to antitrust which focused exclusively upon market structure and impact upon rival websites; impact upon consumers was nowhere to be found.  That Google not only produces search results but also owns some of the results that are searched is not a problem cognizable by modern antitrust.  Edelman himself—appropriately—describes Google and its competitors as “information services.”  Google is not merely a URL finder.  Consumers demand more than that and competition forces search engines to deliver.  It offers value to users (and thus it can offer users to advertisers) by helping them find information in increasingly useful ways.  Most users “want Google Search” to the exclusion of Google’s “other services” (and, if they do, all they need do is navigate over to http://googleminusgoogle.com/ (even in a Chrome browser) and they can have exactly that).  But the critical point is that Google’s “other services” are methods of presenting information to consumers, just like search.  As the web and its users have evolved, and as Google has innovated to keep up with the evolving demands of consumers, it has devised or employed other means than simply providing links to a set of URLs to provide the most relevant information to its users.  The 1960s approach to antitrust condemns this as anticompetitive foreclosure; the modern version recognizes it as innovation, a form of competition that benefits consumers.

Edelman (and other critics, including a number of Senators at last month’s hearing) hearken back to the good old days and suggest that any deviation from Google’s technology or business model of the past is an indication of anticompetitive conduct:

The Google of 2004 promised to help users “leave its website as quickly as possible” while showing, initially, zero ads.  But times have changed.  Google has modified its site design to encourage users to linger on other Google properties, even when competing services have more or better information.  And Google now shows as many fourteen ads on a page.

It is hard to take seriously an argument that turns on criticizing a company simply for looking different than it did seven years ago.  Does anybody remember what search results looked like 7 years ago?  A theory of antitrust liability that would condemn a firm for investing billions of dollars in research and product development, constantly evolving its product to meet consumer demand, taking advantage of new technology, and developing its business model to increase profitability should not be taken seriously.  This is particularly true where, as here, every firm in the industry has followed a similar course, adopting the same or similar innovations.  I encourage readers to try a few queries on http://www.bing-vs-google.com/– where you can get side by side comparisons – in order to test whether the evolution of search results and innovation to meet consumer preferences is really a Google-specific thing or an industry wide phenomenon consistent with competition.  Conventional antitrust analysis holds that when conduct is engaged in not only by allegedly dominant firms, but also by every other firm in an industry, that conduct is presumptively efficient, not anticompetitive.

The main thrust of my critique is that Edelman and other Google critics rely on an outdated antitrust framework in which consumers play little or no role.  Rather than a consumer-welfare based economic critique consistent with the modern approach, these critics (as Edelman does in his opening statement) turn to a collection of anecdotes and “gotcha” statements from company executives.  It is worth correcting a few of those items here, although when we’ve reached the point where identifying a firm’s alleged abuse is a function of defining what a “confirmed” fax is, we’ve probably reached the point of decreasing marginal returns.  Rest assured that a series of (largely inaccurate) anecdotes about Google’s treatment of particular websites or insignificant contract terms is wholly insufficient to meet the standard of proof required to make a case against the company under the Sherman Act or even the looser Federal Trade Commission Act.

• It appears to be completely inaccurate to say that “[a]n unsatisfied advertiser must complain to Google by ‘first class mail or air mail or overnight courier’ with a copy by ‘confirmed facsimile.’”  A quick search, even on Bing, leads one to this page, indicating that complaints may be submitted via web form.
• It is likewise inaccurate to claim that “advertisers are compelled to accept whatever terms Google chooses to impose.  For example, an advertiser seeking placement through Google’s premium Search Network partners (like AOL and The New York Times) must also accept placement through the entire Google Search Network which includes all manner . . . undesirable placements.”  In actuality, Google offers a “Site and Category Exclusion Tool” that seems to permit advertisers to tailor their placements to exclude exactly these “undesirable placements.”
• “Meanwhile, a user searching for restaurants, hotels, or other local merchants sees Google Places results with similar prominence, pushing other information services to locations users are unlikely to notice.”  I have strived in vain to enter a search for a restaurant, hotel, or the like into Google that yielded results that effectively hid “other information services” from my notice, but for some of my searches, Google Places did come up first or second (and for others it showed up further down the page).
• Edelman has noted elsewhere that, sometimes, for some of the searches he has tested, the most popular result on Google (as well, I should add, on other, non-“dominant” sites) is not the first, Google-owned result, but instead the second.  He cites this as evidence that Google is cooking the books, favoring its own properties when users actually prefer another option.  It actually doesn’t demonstrate that, but let’s accept the claim for the sake of argument.  Notice what his example also demonstrates: that users who prefer the second result to the first are perfectly capable of finding it and clicking on it.  If this is foreclosure, Google is exceptionally bad at it.

The crux of Edelman’s complaint seems to be that Google is competing in ways that respond to consumer preferences.  This is precisely what antitrust seeks to encourage, and we would not want a set of standards that chilled competition because of a competitor’s success.  Having been remarkably successful in serving consumers’ search demands in a quickly evolving market, it would be perverse for the antitrust laws to then turn upon Google without serious evidence that it had, in fact, actually harmed consumers.

Untethered from consumer welfare analysis, antitrust threatens to re-orient itself to the days when it was used primarily as a weapon against rivals and thus imposed a costly tax on consumers.  It is perhaps telling that Microsoft, Expedia, and a few other Google competitors are the primary movers behind the effort to convict the company.  But modern antitrust, shunning its inglorious past, requires actual evidence of anticompetitive effect before condemning conduct, particularly in fast-moving, innovative industries.  Neither Edelman nor any of Google’s other critics, offer any.

During the heady days of the Microsoft antitrust case, the big question was whether modern antitrust would be able to keep up with quickly evolving markets.  The treatment of the proferred case against Google is an important test of the proposition (endorsed by the Antitrust Modernization Commission and others) that today’s antitrust is capable of consistent and coherent application in innovative, high-tech markets.  An enormous amount is at stake.  Faced with the high stakes and ever-evolving novelty of high-tech markets, antitrust will only meet this expectation if it remains grounded and focused on the core principle of competitive effects and consumer harm.  Without it, antitrust will devolve back into the laughable and anti-consumer state of affairs of the 1960s—and we will all pay for it.

Filed under: antitrust, consumer protection, economics, error costs, exclusionary conduct, federal trade commission, google, monopolization, technology, tying

I don’t think so.

Let’s start from the beginning.  In my last post, I pointed out that simple economic theory generates some pretty clear predictions concerning the impact of a merger on rival stock prices.  If a merger is results in a more efficient competitor, and more intense post-merger competition, rivals are made worse off while consumers benefit.  On the other hand, if a merger is is likely to result in collusion or a unilateral price increase, the rivals firms are made better off while consumers suffer.

I pointed to this graph of Sprint and Clearwire stock prices increasing dramatically upon announcement of the merger to illustrate the point that it appears rivals are doing quite well:

The WSJ reports the increases at 5.9% and 11.5%, respectively.  In reaction to the WSJ and other stories highlighting this market reaction to the DOJ complaint, I asked what I think is an important set of questions:

How many of the statements in the DOJ complaint, press release and analysis are consistent with this market reaction?  If the post-merger market would be less competitive than the status quo, as the DOJ complaint hypothesizes, why would the market reward Sprint and Clearwire for an increased likelihood of facing greater competition in the future?

A few of our always excellent commenters argued that the analysis above was either incomplete or incorrect.  My claim was that the dramatic increase in stock market prices of Sprint and Clearwire were more consistent with a procompetitive merger than the theories in the DOJ complaint.

Commenters raised three important points and I appreciate their thoughtful responses.

First, the procompetitive theory does not explain the change in all stock market prices.  For example, readers pointed out that Verizon’s stock barely ticked downward, while smaller carriers MetroPCS and Leap both fell (.8% and 2.3%, respectively, according to the WSJ).  The procompetitive theory, the commenters argued, implies that Verizon and these other rivals should move upward.

Second, they argue that perhaps an exclusionary theory of the merger better explains these stock price reactions.  Indeed, the new 2010 Horizontal Merger Guidelines included (not without controversy) potential exclusionary effects (“Enhanced market power may also make it more likely that the merged entity can profitably and effectively engage in exclusionary conduct. “).  Rick Brunell of AAI writes:

Although the smaller carriers may gain in the short run due from a merger that raises prices, they also may lose in the long run due to its exclusionary effects, a theory that was front and center of Sprint’s opposition (and the smaller carriers’). Notably Verizon, which has no reason to fear exclusion and would have the most to lose if the merger were actually efficient, has not opposed the merger.”

Similarly, Matt Bodie writes:

Why wouldn’t the market’s reaction be a sign of this: (a) the AT&T/T-Mobile merger will give the new entity strong market power, (b) there are strong anticompetitive as well as efficiency gains from being bigger and having more market size, (c) the newly merged company would use that power to crush its weakest competitors, i.e. Sprint? After all, isn’t there a traditional story where monopolists cut prices to drive other competitors out, but then gradually raise price once their market power allows it, especially in industries with high barriers to entry?”

The basics of the exclusionary theory of the merger is that the anticompetitive harm is not coordination or unilateral price increases from the direct acquisition of market power, i.e. the elimination of competition from a close rival.  Rather, the exclusionary theory posits that the post-merger firm will have sufficient market power to exclude rivals from access to a critical input (e.g. backhaul) and, as Matt has it, “crush its weakest competitors.”  So to Matt, yes, there is that theory in antitrust.  But note that the post-merger share of the combined entity here would be nowhere close to traditional monopoly power standards required to make out a monopolization claim under Section 2 of the Sherman Act.  The new Guidelines do quasi-endorse the possibility of a Minority-Report like merger enforcement search for exclusion that doesn’t reach Section 2 standards post-merger, but might someday, but also needs to be stopped now.  But it is decidedly not standard in merger analysis. And this case is probably not a good test case for that theory; at least the DOJ thinks so.  But no, I don’t think the market reaction is reflecting concerns about exclusion.  More on that in a second.  But for now note that this is not simply a legal point.  While the law requires the demonstration of monopoly power for a Section 2 claim, the economic literature focusing upon exclusion also considers market power a necessary but not sufficient condition for competitive harm.  For the same reasons the exclusion claim would be rejected post-merger on legal grounds if we accept the market definition alleged by the DOJ, exclusion is unlikely as a matter of economics.

Put simply, the exclusionary theory’s proponents argue that it can explain the increase in Sprint’s stock price (reduced likelihood of future exclusion because of the DOJ challenge) and Verizon’s inconsequential reaction (it has “no reason to fear exclusion”).

Just so everybody is seeing the same thing — here is a chart with 5 days of trading including Verizon, Sprint, Clearwire, MetroPCS, Leap and the S&P 500.

Third, commenters argue that this simple analysis doesn’t account for other important factors.  NB writes:

Why did you choose Sprint particularly? Verizon, a larger and far more significant competitor, had its stock drop sharply in that same period you show Sprint “surging”. MetroPCS’s stock also dropped.

So what does it mean when a weak competitor’s stock jumps but two other competitors who are doing well have their stock drop? Other than that there are clearly more factors in play here?

Enough questions; time for answers.

Why Didn’t I Include the Exclusionary Theory of Harm?

I plead guilty.  Or at least guilty with an explanation.  I didn’t discuss the possibility of exclusion and whether it would better explain these market reactions than the theory that the merger is efficient or anticompetitive because it will facilitate coordination or unilateral price increases.  As it turns out, however, the reason is that the post was motivated by the following question:

How many of the statements in the DOJ complaint, press release and analysis are consistent with this market reaction?

Turns out, I’m in pretty good company in omitting this theory.  The DOJ didn’t allege it either.  As discussed above, the DOJ specifically alleged that the merger would result in coordinated effects in the national market and/or unilateral price increases.  Rick Brunell accurately points out that Sprint and AAI have both made these arguments.  Indeed, when I testified in the House on the merger, there were a lot of questions raised about exclusionary concerns.   But the bottom line is that they are not in the Complaint.  Apparently, those arguments did not persuade the Justice Department.  I have no intention on running from the interesting question posed by the commenters that the exclusion theory does a better job of explaining market price reactions.  That’s next.  But for now, let me say that I think there is a good reason the DOJ did not accept the Sprint / AAI invitation to adopt the exclusion theory.

Does Exclusion Do A Better Job of Explaining Verizon’s Non-Movement or Slight Fall? 

I think proponents of the exclusion theory of the merger have a tough task here.  Notice that the prediction of the exclusionary theory is NOT that Verizon’s stock price will stay put or fall.  Instead, it is that it will increase post-merger.  While Brunell observes that Verizon need not fear post-merger exclusion itself, it would certainly be happy to free-ride on the allegedly imminent exclusionary efforts of the newly merged firm.  Post-Chicagoans often invoke the argument that “competition is a public good” when explaining why a downstream input provider has reason to go along with an upstream firm’s attempt to monopolize.   Bork argued that the downstream firm had no reason to engage in a contract with the upstream provider that would increase the likelihood that he would be facing an upstream monopolist (and thus worse terms of trade) tomorrow.  The classic Post-Chicago response is that each downstream firm doesn’t take into account the impact of his private decision to enter into such a contract with the would-be monopolist — that is, competition is a public good.  The flip side of this argument is that exclusion is a public good too!   To put it more concretely, if the post-merger combination of AT&T / T-Mobile were able to successfully exclude Sprint and smaller carriers such as MetroPCS and Leap, and thereby reduce competition, the clear implication of this theory is that Verizon would benefit.

The relevant economics here are not limited to the possibility that post-merger AT&T would successfully exclude Verizon.  Think about it: both Verizon and the post-merger firm would benefit from the exclusionary efforts and reduced competition.  However, Verizon would stand to gain even more!  After all, it isn’t paying the $39 billion purchase price for the acquisition (or any of the other costs of implementing an expensive exclusion campaign).  Thus, an announcement to block the would-be exclusionary merger — the one that would allow Verizon to outsource the exclusion of its rivals to AT&T on the cheap — wouldn’t happen.  Verizon stock should fall relative to the market in response to this lost opportunity.  The unilateral and coordinated effects theories in the DOJ complaint are at significant tension with the stock market reactions of firms like Sprint (and its affiliated venture, Clearwire).  The exclusion theory predicts a large decrease in stock price for Verizon with the announcement.   None of these comfortably fit the facts.  Verizon more or less tracks the S&P with a slight drop.  What about the smaller carriers?  Take a look at the chart.  MetroPCS barely moved relative to the market (in fact, may have increased relative to the market over the relevant time period); Leap is down a bit more than the market.   Here, with the smaller carriers there is not a lot of movement in any direction.  But, contra NB’s comment (“Verizon, a larger and far more significant competitor, had its stock drop sharply in that same period you show Sprint “surging”. MetroPCS’s stock also dropped.”), Verizon’s small fall relative to the market is nowhere near the magnitude of the positive effect on Sprint and Clearwire.

But what about competition?  Isn’t it true that if the merger was procompetitive a challenge announcement would likely mean less competition for Verizon and also predict an increase in stock price?  AAI’s comment tries to have this both ways.  If Verizon’s price stays still, its because it has nothing to fear from exclusion (contra the economics above); if it goes down, the DOJ announcement has decreased the likelihood of those coordinated effects Sprint and AAI argued were so likely (but then there is Sprint’s big jump); and if Verizon prices increase then it just means that we weren’t right in the first instance than they were safe from exclusion.  One is reminded of Tom Smith and his incredible bread machine.   But this leads to an interesting point.  Brunell and AAI (and perhaps other proponents of the DOJ challenge), as pointed out in the comments, appear to agree with me that stock market reactions are probative evidence of competitive effects.  Perhaps they believe that the exclusionary theory is a better explanation of the facts — I obviously don’t think so.  But we are where we are.  That theory is not alleged.  Now that we’ve observed the quite significant stock market reaction of Sprint to the challenge announcement.  Do we at least agree those facts are in tension with the coordinated effects theory made so prominent in the DOJ complaint???

Couldn’t There Be Other Important Factors Explaining Stock Price Movements Unrelated to the Competitive Implications of the DOJ’s Challenge?

To write the question is to answer it.  You bet there could be.  And indeed, I wrote in the first post that while the fairly dramatic stock price reactions of Sprint and Clearwire were probative, the post was not a full-blown event study that would account for those events, formulate a market model, and test for the abnormal returns surrounding the announcement controlling for other important events.  Further, not all competitors are created equal.  Under the efficiency story, the distribution of benefits will accrue proportionately to the rivals who were most likely to face increased competition post-merger (and now are more likely not to).  I certainly agree with Rick Brunell’s summary comment that the stock price evidence is somewhat “mixed.”  There are small and relatively ambiguous effects — once one includes the market performance — on the stock prices of Metro and Verizon.  Leap is more clearly down, even if by a small amount relative to the market.   There may well be a variety of factors unrelated to the announcement confounding effects here.  This is the reason we do real event studies in practice and why I do not believe the simple collection of evidence here warrants sweeping conclusions about the merits of the merger.

However, the DOJ complaint tells us that the important competitive players in the market — the “Big Four” — are AT&T, T-Mobile, Sprint, and Verizon.  Focusing upon the non-merging big 4, we see Sprint’s price going up dramatically and Verizon’s staying put.  The former is simply more consistent with procompetitive theories than the coordinated effects and unilateral effects theories alleged in the DOJ complaint.  One might expect an announcement to block a procompetitive merger to have a greater positive impact on Verizon stock.  But, as many have observed in the press, the impact of the merger upon Verizon is complicated by a number of factors, not the least of which is that the challenge announcement increases the likelihood that the DOJ is committed to challenging any future attempts to merger by Verizon.  Unless spectrum capacity is increased dramatically (see this excellent Adam Thierer post on this score) in the very near future it is difficult to see how the reduced ability to exercise that significant and valuable option would not also impact Verizon.  Thus, while not a slam dunk by any means, the procompetitive theory of the merger does a pretty decent job on the Big Four.   It certainly beats the coordination theory trumpeted in the Complaint.  As for the attempt of AAI and Sprint to salvage the DOJ complaint with the exclusionary theory — perhaps it is not too late to amend, but it isn’t there now and I’d warn the DOJ against including it.  With respect to the DOJ’s Big Four, the exclusionary theory is not only new and relatively controversial in the Guidelines, but also makes a strong prediction concerning a Verizon stock price increase that is inconsistent with the data.

There will certainly be more data as we move along.  And it should interesting to watch how things unfold both in the market and between the DOJ and FCC as well.  For now, however, color me unconvinced by the heavy reliance upon the structural, “Big 4 collusion” story leading the Complaint and the attempts to save it with exclusionary theories.

Filed under: antitrust, business, economics, exclusionary conduct, merger guidelines, mergers & acquisitions, monopolization, technology, telecommunications, wireless

In the last post, I discussed possible characterizations of Google’s conduct for purposes of antitrust analysis.  A firm grasp of the economic implications of the different conceptualizations of Google’s conduct is a necessary – but not sufficient – precondition for appreciating the inconsistencies underlying the proposed remedies for Google’s alleged competitive harms.  In this post, I want to turn to a different question: assuming arguendo a competitive problem associated with Google’s algorithmic rankings – an assumption I do not think is warranted, supported by the evidence, or even consistent with the relevant literature on vertical contractual relationships – how might antitrust enforcers conceive of an appropriate and consumer-welfare-conscious remedy?  Antitrust agencies, economists, and competition policy scholars have all appropriately stressed the importance of considering a potential remedy prior to, rather than following, an antitrust investigation; this is good advice not only because of the benefits of thinking rigorously and realistically about remedial design, but also because clear thinking about remedies upfront might illuminate something about the competitive nature of the conduct at issue.

Somewhat ironically, former DOJ Antitrust Division Assistant Attorney General Tom Barnett – now counsel for Expedia, one of the most prominent would-be antitrust plaintiffs against Google – warned (in his prior, rather than his present, role) that “[i]mplementing a remedy that is too broad runs the risk of distorting markets, impairing competition, and prohibiting perfectly legal and efficient conduct,” and that “forcing a firm to share the benefits of its investments and relieving its rivals of the incentive to develop comparable assets of their own, access remedies can reduce the competitive vitality of an industry.”  Barnett also noted that “[t]here seems to be consensus that we should prohibit unilateral conduct only where it is demonstrated through rigorous economic analysis to harm competition and thereby harm consumer welfare.”  Well said.  With these warnings well in-hand, we must turn to two inter-related concerns necessary to appreciating the potential consequences of a remedy for Google’s conduct: (1) the menu of potential remedies available for an antitrust suit against Google, and (2) the efficacy of these potential remedies from a consumer-welfare, rather than firm-welfare, perspective.

What are the potential remedies?

The burgeoning search neutrality crowd presents no lack of proposed remedies; indeed, if there is one segment in which Google’s critics have proven themselves prolific, it is in their constant ingenuity conceiving ways to bring governmental intervention to bear upon Google.  Professor Ben Edelman has usefully aggregated and discussed several of the alternatives, four of which bear mention:  (1) a la Frank Pasquale and Oren Bracha, the creation of a “Federal Search Commission,” (2) a la the regulations surrounding the Customer Reservation Systems (CRS) in the 1990s, a prohibition on rankings that order listings “us[ing] any factors directly or indirectly relating to” whether the search engine is affiliated with the link, (3) mandatory disclosure of all manual adjustments to algorithmic search, and (4) transfer of the “browser choice” menu of the EC Microsoft litigation to the Google search context, requiring Google to offer users a choice of five or so rivals whenever a user enters particular queries.

Geoff and I discuss several of these potential remedies in our paper, If Search Neutrality is the Answer, What’s the Question?  It suffices to say that we find significant consumer welfare threats from the creation of a new regulatory agency designed to impose “neutral” search results.  For now, I prefer to focus on the second of these remedies – analogized to CRS technology in the 1990s – here; Professor Edelman not only explains proposed CRS-inspired regulation, but does so in effusive terms:

A first insight comes from recognizing that regulators have already – successfully! – addressed the problem of bias in information services. One key area of intervention was customer reservation systems (CRS’s), the computer networks that let travel agents see flight availability and pricing for various major airlines. Three decades ago, when CRS’s were largely owned by the various airlines, some airlines favored their own flights. For example, when a travel agent searched for flights through Apollo, a CRS then owned by United Airlines, United flights would come up first – even if other carriers offered lower prices or nonstop service. The Department of Justice intervened, culminating in rules prohibiting any CRS owned by an airline from ordering listings “us[ing] any factors directly or indirectly relating to carrier identity” (14 CFR 255.4). Certainly one could argue that these rules were an undue intrusion: A travel agent was always free to find a different CRS, and further additional searches could have uncovered alternative flights. Yet most travel agents hesitated to switch CRS’s, and extra searches would be both time-consuming and error-prone. Prohibiting biased listings was the better approach.

The same principle applies in the context of web search. On this theory, Google ought not rank results by any metric that distinctively favors Google. I credit that web search considers myriad web sites – far more than the number of airlines, flights, or fares. And I credit that web search considers more attributes of each web page – not just airfare price, transit time, and number of stops. But these differences only grant a search engine more room to innovate. These differences don’t change the underlying reasoning, so compelling in the CRS context, that a system provider must not design its rules to systematically put itself first.

The analogy is a superficially attractive one, and we’re tempted to entertain it, so far as it goes.  Organizational questions inhere in both settings, and similarly so: both flights and search results must be ordinally ranked, and before CRS regulation, a host airline’s flights often appeared before those of rival airlines.  Indeed, we will take Edelman’s analogy at face value.  Problematically for Professor Edelman and others pushing the CRS-style remedy, a fuller exploration of CRS regulation reveals this market intervention – well, put simply, wasn’t so successful after all.  Not for consumers anyway.  It did, however, generate (economically) predictable consequences: reduced consumer welfare through reduced innovation. Let’s explore the consequences of Edelman’s analogy further below the fold.

History of CRS Antitrust Suits and Regulation

Early air travel primarily consisted of “interline” flights – flights on more than one carrier to reach a final destination.  CRSs arose to enable airlines to coordinate these trips for their customers across multiple airlines, which necessitated compiling information about rival airlines, their routes, fares, and other price- and quality-relevant information.  Major airlines predominantly owned CRSs at this time, which served both competitive and cooperative ends; this combination of economic forces naturally drew antitrust advocates’ attention.

CRS regulation proponents proffered numerous arguments as to the potentially anticompetitive nature and behavior of CRS-owning airlines.  For example, they claimed that CRS-owning airlines engaged in “dirty tricks,” such as using their CRSs to terminate passengers’ reservations on smaller, rival airlines and to rebook customers on their own flights, and refusing to allow smaller airlines to become CRS co-hosts, thereby preventing these smaller airlines from being listed in search results.  CRS-owning airlines faced further allegations of excluding rivals through contractual provisions, such as long-term commitments from travel agents.  Proponents of antitrust enforcement alleged that the nature of the CRS market created significant barriers to entry and provided CRS-owning airlines with significant cost advantages to selling their own flights.  These cost advantages purportedly derived from two main sources: (1) quality advantages that airline-owned CRSs enjoyed, as they could commit to providing comprehensive and accurate information about the owner airline’s flight schedule, and (2) joint ownership of CRSs, which facilitated coordination between airlines and CRSs, thereby decreasing the distribution and information costs.

These claims suffered from serious shortcomings including both a failure to demonstrate harm to competition rather than injury to specific rivals as well as insufficient appreciation for the value of dynamic efficiency and innovation to consumer welfare.  These latter concerns were especially pertinent in the CRS context, as CRSs arose at a time of incredible change – the deregulated airline industry, joined with novel computer technology, necessitated significant and constant innovation.  Courts accordingly generally denied antitrust remedies in these cases – rejecting claims that CRSs imposed unreasonable restraints on competition, denied access to an essential facility, or facilitated monopoly leverage.

Yet, particularly relevant for present purposes, one of the most popular anticompetitive stories was that CRSs practiced “display bias,” defined as ranking the owner airline’s flights above those of all other airlines.  Proponents claimed display bias was particularly harmful in the CRS setting, because only the travel agent, and not the customer, could see the search results, and travel agents might have incentives to book passengers on more expensive flights for which they receive more commission.  Fred Smith describes the investigations surrounding this claim:

These initial CRS services were used mostly by sophisticated travel agents, who could quickly scroll down to a customer’s preferred airline.  But this extra “effort” was considered discriminatory by some at the DOJ and the DOT, and hearings were held to investigate this threat to competition.  Great attention was paid to the “time” required to execute only a few keystrokes, to the “complexity” of re-designing first screens by computer-proficient travel agents, and to the “barriers” placed on such practices by the host CRS provider.

CRS Rules

While courts declined to intervene in the CRS market, the Department of Transportation (DOT) eagerly crafted rules to govern CRS operations.  The DOT’s two primary goals in enacting the 1984 CRS regulations were (1) to incentivize entry into the CRS market and (2) to prevent airline ownership of CRSs from decreasing competition in the downstream passenger air travel market.  One of the most notable rules introduced in the 1984 CRS regulations prohibited display bias.  The DOT changed both this rule and CRS rules as a whole significantly, and by 1997, the DOT required each CRS “(i) to offer at least one integrated display that uses the same criteria for both online and interline connections and (ii) to use elapsed time or non-stop itinerary as a significant factor in selecting the flight options from the database” (Alexander, 2004).  However, the DOT did not categorically forbid display bias; rather, it created several exceptions to this rule – and even allowed airlines to disseminate software that introduced bias into displays.  Additionally, the DOT expressly refused to enforce its anti-bias rules against travel agent displays.

Other CRS rules attempted to reinforce these two goals of additional market entry and preservation of downstream competition.  CRS rules specifically focused on mitigating travel agent switching costs between CRS vendors and reducing any quality advantage incumbent CRSs allegedly had.  Rules prohibited discriminatory booking fees and the tying of travel agent commissions to CRS use, limited contract lengths, prohibited minimum uses and rollover clauses, and required CRSs to give all participating carriers equal service upgrades.

Evidence of CRS Regulation “Success”?

The CRS regulatory experiment had years to run its course; despite the extent and commitment of its regulatory sweep, these rules failed to improve consumer outcomes in any meaningful way.  CRS regulations precipitated neither innovation nor entry, and likely incurred serious allocative efficiency and consumer welfare losses by attempting to prohibit display bias.

First, CRS regulations unambiguously failed in their goal of increasing ease of entry:

Only six CRS vendors offered their services to domestic airlines and travel agents in the mid-1980s. . . If the rules had actually facilitated entry, the number of CRS vendors should have grown or some new entrants should have been seen during the past twenty years.  The evidence, however, is to the contrary.  It remains that ‘[s]ince the [CAB] first adopted CRS rules, no firm has entered the CRS business.’  Meanwhile, there has been a series of mergers coupled with introduction of multinational CRS; the cumulative effect was to reduce the number of CRSs. . . Even if a regulation could successfully facilitate entry by a supplier of CRS services, the gain from such entry would at this point be relatively small, and possibly negative. (Alexander and Lee, 2004) (emphasis added).

As such, CRS regulations did not achieve one of their primary objectives – a fact which stands in stark contrast to Edelman’s declaration that CRS rules represent an unequivocal regulatory success.

Most relevant to the search engine bias analogy, the CRS regulations prohibiting bias did not positively affect consumer welfare.  To the contrary, by ignoring the reality that most travel agents took consumer interests into account in their initial choice of CRS operator (even if they do so to a lesser extent in each individual search they conduct for consumers), and that even if residual bias remained, consumers were “informed and repeat players who have their own preferences,” CRS regulations imposed unjustified costs.  As Alexander and Lee describe it

[T]he social value of prohibiting display . . . bias solely to improve the quality of information that consumers receive about travel options appears to be low and may be negative.  Travel agents have strong incentives to protect consumers from poor information, through how they customize their internal display screens, and in their choices of CRS vendors.

Moreover, and predictably, CRS regulations appear to have caused serious harm to the competitive process:

The major competitive advantage of the pre-regulation CRS was that it permitted the leading airlines to slightly disadvantage their leading competitors by placing them a bit farther down on the list of available flights.  United would place American slightly farther down the list, and American would return the favor for United flights.  The result, of course, was that the other airlines received slightly higher ranks than they would have otherwise.  When “bias” was eliminated, United moved up on the American system and vice versa, while all other airlines moved down somewhat.  The antitrust restriction on competitive use of the CRS, then, actually reduced competition.  Moreover, the rules ensured that the United/American market leadership would endure fewer challenges from creative newcomers, since any changes to the system would have to undergo DOT oversight, thus making “sneak attacks” impossible.  The resulting slowdown of CRS technology damaged the competitiveness of these systems.  Much of the innovative lead that these systems had enjoyed slowly eroded as the internet evolved.  Today, much of the air travel business has moved to the internet (as have the airlines themselves) (Smith, 1999).

These competitive losses occurred despite evidence suggesting that CRSs themselves enhanced competition and thus had the predictable positive impact for consumers.  For example, one study found that CRS usage increased travel agents’ productivity by an average of 41% and that in the early 1990s over 95% of travel agents used a CRS – indicating that travel agents were able to assist consumers far more effectively once CRSs became available (Ellig, 1991).  The rules governing contractual terms fared no better; indeed, these also likely reduced consumer welfare:

The prohibited contract practices–long-term contracting and exclusive dealing–that had been regarded as exclusionary might not have proved to be such a critical barrier to entry: entry did not occur, independently of those practices.  Evidence on the dealings between travel agents and CRS vendors, post-regulation, suggests that these practices may have enhanced overall allocative efficiency.  Travel agents appear to have agreed to some, if not all, restrictive contracts with CRS vendors as a means of providing those vendors with assurance that they would be repaid gradually, over time, for their up-front investments in the travel agent, such as investments in equipment or training (Alexander and Lee, 2004).

Accordingly, CRS regulations seem to have threatened innovation by decreasing the likelihood that CRS vendors would recover research and development expenditures without providing a commensurate consumer benefit.

Termination of Rules

The DOT terminated CRS regulations in 2004 in light of their failure to improve competitive outcomes in the CRS market and a growing sense that they were making things worse, not better – which Edelman fails to acknowledge and which certainly undermines his claim that regulators addressed this problem “successfully.”  From the time CRS regulations were first adopted in 1984 until 2004, the CRS market and the associated technology changed significantly, rapidly becoming more complex.  As the market increased in complexity, it became increasingly more difficult for the DOT to effectively regulate.  Two occurrences in particular precipitated de-regulation: (1) the major airlines divested themselves of CRS ownership (despite the absence of any CRS regulations requiring or encouraging divestiture!), and (2) the commercialization of the internet introduced novel forms of substitutes to the CRS system that the CRS regulations did not govern.  Online direct-to-traveler services, such as Travelocity, Expedia and Orbitz provide consumers with a method to choose their own flights, entirely absent travel agent assistance.  More importantly, Expedia and Orbitz each developed direct connection technologies that allow them to make reservations directly with an airline’s internal reservation system – bypassing CRS systems almost completely.  Moreover, Travelocity, Expedia, and Orbitz were never forced to comply with CRS regulations, which allowed them to adopt more consumer-friendly products and innovate in meaningful ways, obsoleting traditional CRSs.  It is unsurprising that Expedia has warned against overly broad regulations in the search engine bias debate – it has first-hand knowledge of how crucial the ability to innovate is.)

These developments, taken in harmony, mean that in order to cause any antitrust harm in the first instance, a hypothetical CRS monopolist must have been interacting with (1) airlines, (2) travel agents, and (3) consumers who all had an insufficient incentive to switch to another alternative in the face of a significant price increase.  Given this nearly insurmountable burden, and the failure of CRS regulations to improve consumer welfare in even the earlier and simpler state of the world, Alexander and Lee find that, by the time CRS regulations were terminated in 2004, they failed to pass a cost-benefit analysis.

Overall, CRS regulations incurred significant consumer welfare losses and rendered the entire CRS system nearly obsolete by stifling its ability to compete with dynamic and innovative online services.  As Ellig notes, “[t]he legal and economic debate over CRS. . . frequently overlooked the peculiar economics of innovation and entrepreneurship.”  Those who claim search engine bias exists (as distinct from valuable product differentiation between engines) and can be meaningfully regulated rely upon this same flawed analysis and expect the same flawed regulatory approach to “fix” whatever issues they perceive as ailing the search engine market.  Search engine regulation will make consumers worse off.  In the meantime, proponents of so-called search neutrality and heavy-handed regulation of organic search results battle over which of a menu of cumbersome and costly regulatory schemes should be adopted in the face of evidence that the approaches are more likely to harm consumers than help them, and even stronger evidence that there is no competitive problem with search in the first place.

Indeed, one benefit of thinking hard about remedies in the first instance is that it may illuminate something about the competitive nature of the conduct one seeks to regulate.  I defer to former AAG Barnett in explaining this point:

Put another way, a bad section 2 remedy risks hurting consumers and competition and thus is worse than no remedy at all. That is why it is important to consider remedies at the outset, before deciding whether a tiger needs catching. Doing so has a number of benefits.  …

Furthermore, contemplation of the remedy may reveal that there is no competitive harm in the first place.  Judge Posner has noted that “[t]he nature of the remedy sought in an antitrust case is often . . . an important clue to the soundness of the antitrust claim.”⁽⁴⁾ The classic non-section 2 example is Pueblo Bowl-O-Mat, where plaintiffs claimed that the antitrust laws prohibited a firm from buying and reinvigorating failing bowling alleys and prayed for an award of the “profits that would have been earned had the acquired centers closed.”⁽⁵⁾ The Supreme Court correctly noted that condemning conduct that increased competition “is inimical to the purposes of [the antitrust] laws”⁽⁶⁾–more competition is not a competitive harm to be remedied. In the section 2 context, one might wish that the Supreme Court had focused on the injunctive relief issued in Aspen Skiing–a compelled joint venture whose ability to enhance competition among ski resorts was not discussed⁽⁷⁾–in assessing whether discontinuing a similar joint venture harmed competition in the first place.⁽⁸⁾

A review of my paper with Geoff reveals several common themes among proposed remedies intimated by the above discussion of CRS regulations.  The proposed remedies consistently: (1) disadvantage Google, (2) advantage its rivals, and (3) have little if anything to do with consumers.  Neither economics nor antitrust history supports such a regulatory scheme; unfortunately, it is consumers that might again ultimately pay the inevitable tax for clumsy regulatory tinkering with product design and competition.

Filed under: antitrust, economics, federal trade commission, google, international center for law & economics, monopolization, technology Tagged: antitrust, Federal Trade Commission, google, search

This is part one of a two part series of posts in which I’ll address the problems associated with discerning an appropriate antitrust remedy to alleged search engine bias.  The first problem – and part – is, of course, how we should conceptualize Google’s allegedly anticompetitive conduct; in the next part, I will address how antitrust regulators should conceive of a potential remedy, assuming arguendo the existence of a problem at all.  Despite some commentators’ assumptions, I do not think the economics indicate any such problem exists.

The question of how to conceptualize Google’s business practices – even its business model! – remains the indispensible starting point for antitrust analysis, including potential remedies; doubly so in the wake of the FTC’s decision to formally investigate Google.  While the next part will focus more directly upon potential remedies that have been proposed by various Google critics, there is a fundamental link between how we conceptualize Google’s provision of search results for the purposes of antitrust analysis and the design of remedies.  Indeed, antitrust enforcers and scholars have taught that thinking hard about remedies upfront can and frequently should influence how we think about the competitive nature of the conduct at issue.  The question of how to conceptualize Google’s organic search results has sparked serious debate, as some have claimed that “Google’s behavior is harder to define” than traditional anticompetitive actions and represents “a new kind of competition.”  Some have also focused upon “search bias” itself as the relevant conduct for antitrust purposes.  Of course, as I’ve pointed out, these statements are not in line with modern antitrust economics and usually precede calls to deviate from traditional consumer-welfare-focused antitrust analysis.

I see two useful conceptual constructs in evaluating “search bias” within the antitrust framework.  Recall that “search bias” typically translates to allegations that Google favors its own affiliated content over that of rivals.  For example, a search query on Google for “map of Arlington, VA” might turn up a map of Arlington from Google Maps in the top link.  These allegations usually concede that we would expect Bing Maps if we ran the same search on Bing.  The complaints from vertical search engines and travel services like Expedia particularly center around the notion that Google’s “entry” into various spaces  –  such as travel services – supported by prominent search rankings disadvantages rivals and may lead to their exit.

Observant readers will note my use of scare quotes around “entry.”  This is not coincidental.  It is not obvious to me that Google necessarily enters a new sector (much less a well-defined antitrust product market) when it directs a user to content in a new format– such as a map, video, or place page.  Google’s primary function is search; users rely on search engines to reduce search and information costs.  I think it is at least as likely that Google’s attempts to provide this content by any chosen metric is simply an attempt to do their cardinal job better: answering user queries with relevant information at a minimum of cost.  Holding that threshold issue aside for a moment, in my mind, there are two ways to classify that conduct in the antitrust framework.

First, one might conceive of search bias allegations as “vertical integration” or vertical contractual activity.  I’ve explored this conception at significant length both in blog posts (see, e.g. here and here) as well as a longer article with Geoff.  The classic antitrust concern in this setting is that a monopolist might foreclose rivals from an input the rivals need to compete effectively.  For example, Google owns YouTube; Google could prominently place YouTube results when users enter queries seeking video content.  (Ignore for the moment that YouTube will necessarily rank highly on other search engines because it is the leading site for video content).  Within this vertical integration framework, there is a standard analysis for understanding when competitive concerns might arise, the conditions that must be satisfied for those concerns to warrant scrutiny, a deeply embedded understanding that harm to rivals must be distinguished from demonstrable harm to competition, and an equally deeply held understanding that these vertical arrangements and relationships are often, even typically, pro-competitive (e.g., in the YouTube example vertical integration likely leads to reduced latency and faster provision of video content).

Second, one might conceptualize organic search results as the product of Google’s algorithm and thus falling into the category of conduct analyzed as “product design” for antitrust purposes.  This algorithm faces competition from other search algorithms and vertical search engines to deliver relevant results to consumers.  It is the design of the algorithm that ranks Google-affiliated content, according to the complaints, preferentially and to the disadvantage of rivals. I explore both beneath the fold.

The two conceptions are not mutually exclusive.  The antitrust implications of the two different conceptions of Google’s organic search are significant.  Courts and agencies generally give wide latitude to product design decisions, through with some prominent exceptions (Microsoft, FTC v. Intel).  Courts are skeptical to intervene on the basis of complaints about product design by rivals because they concerned that such intervention will chill innovation.  Concern for false positives play a central part in the analysis, as do concerns that any remedy will involve judicial oversight of product innovation.  Plaintiffs can and do, from time to time, win these cases, but the product-design conception carries with it a heavy deference for design decisions.

The “vertical” (in the antitrust sense) conception of Google’s search results requires us to think about the economics of algorithmic search ranking, placement choices, and the economics of vertical relationships between a content provider and a search engine.  There are many economic reasons for vertical contractual relationships between such content or product providers and retailers.  Coca-Cola pays retailers for promotional shelf space, manufacturers compensate retailers by granting them exclusive territories, and product manufacturers and distributors often enter into exclusive relationships in which the distributor does not simply feature or promote the manufacturer’s product, but does so to the exclusion of all of the manufacturer’s rivals.

The anticompetitive narrative of Google’s conduct focuses heavily on that prominent placement within Google’s rankings, e.g. the first link or one towards the top of the page, results in a substantial amount of traffic.  This is no doubt true; it is not a sufficient condition for proving competitive harm.  It is equally true that eye-level and other premium level shelf space in the supermarket generates more sales than other placements within the store.  There is good economic reason for manufacturers to pay retailers for premium shelf space (see Klein and Wright, 2007); and evidence that these arrangements are good for consumers (Wright, 2008).  Retailers’ shelf space decisions, and decisions to promote one product over another, are often influenced by contractual incentives; and it is a good thing for consumers.   Now consider the case when the retailer shelf space decision is influenced not by contractual incentive and compensation, but by ownership.  This is really just a special case – as ownership aligns the incentives (like the contract would) of the manufacturer and retailer.  For example, a supermarket might promote its own private label brand in eye-level shelf space.  Alternatively, in a category management relationship, a retailer might delegate a specific manufacturer as “category captain” and allow it significant influence over product selection and shelf space placement decisions.  Note that in the case of exclusive relationships, the presumption that such arrangements are pro-competitive applies to shelf placement that would entirely exclude a rival from the shelf, not just demote it.

In economics, the theoretical and empirical verdict is in about these sorts of vertical contractual relationships: while they can be anticompetitive under some circumstances, the appropriate presumption is that they are generally pro-competitive and a part of the normal competitive process until proven otherwise.  How we conceptualize placement of search results, including those affiliated with the search engine (e.g. Google Maps on Google or Bing Maps on Bing), should influence how we think about the appropriate burden of production facing would-be antitrust plaintiffs, including the Federal Trade Commission.

Indeed, these two models offer important trade-offs for antitrust analysis.  To wit, in my view, the vertical integration model provides a still difficult, but relatively easier case for potential rivals to make under existing case law, but it also integrates efficiencies directly into the analysis.  For example, vertical integration and exclusive dealing cases accept as a starting point the notion that such arrangements are often efficient.  On the other hand, while potential plaintiffs have a tougher initial burden in a product design case, the focus often turns to how the design impacts interoperability and whether the defendant can defend its technical design choices.   Having explored the potential conceptual constructs for characterizing Google’s conduct for the purpose of antitrust analysis, my next post will link those concepts to a discussion of potential remedies, exploring the proposed remedies for Google’s conduct, a relevant historical parallel to today’s “search bias” debate raised by some as a model of regulatory success, and a discussion of the economic non sequiturs surrounding the case against Google as juxtaposed against these proposed remedies.

Filed under: antitrust, economics, federal trade commission, google, monopolization, technology