Assessing the claims that the Google-AdMob merger will "leverage Google's dominance" and also kill kittens

Cite this Article
Geoffrey A. Manne, Assessing the claims that the Google-AdMob merger will "leverage Google's dominance" and also kill kittens, Truth on the Market (April 07, 2010),

News items continue to pile up suggesting that the FTC is likely to challenge Google’s acquisition of mobile application and website advertising provider, AdMob.  See this recent article from the Wall Street Journal.  News reports today contain this quote from an anonymous source:

“The staff (at the U.S. Federal Trade Commission) believes there is a significant competitive problem and they are prepared to make a recommendation to sue,” the person said, speaking on condition of anonymity.

Senator Herb Kohl also opined on the deal this week (having conducted his own thorough economic analysis, of course) and offered his assessment in a letter to the FTC:

Critics of this transaction worry that this deal will allow Google to merge with one of its biggest rival mobile advertising competitors, and leverage its dominance of PC-based search advertising market into the emerging mobile advertising market, particularly with respect to advertising embedded in smart phone applications.  The parties to this transaction argue, on the other hand, that there are ample competitors in the smart phone search and application-based advertising market, and also contend that the market for advertising on mobile devices is too nascent to determine that any transaction will lead to dominance by any one company.

* * *

Without reaching any conclusion as to whether the Google/AdMob transaction would create such dominance or would cause any substantial harm to competition, I believe it is essential that the FTC scrutinize this deal very closely to carefully examine this question.

* * *

The FTC should also pay close attention to the privacy interests implicated by this transaction, as the combined firm will gain access to a treasure trove of data on millions of consumers’ behavior, search and product preferences.    The FTC should assure itself that the deal, if approved, will have sufficient safeguards to protect consumers’ privacy

* * *

Thus the incipiency of the smart phone advertising market is not in itself a reason for the FTC to desist from taking any necessary action to enforce the antitrust laws or protect competition should it determine such action is necessary.

Because the acquisition has already received a second request, and is certainly being carefully assessed by the FTC (and Kohl would never suggest otherwise), I take Kohl’s urging that the FTC “scrutinize this deal very closely” to be the closest he wants to come to publicly telling the FTC to challenge the merger.  Kohl can perhaps be excused for imprecise and antitrust-irrelevant thinking, being a politician and all (even though one in charge of the Senate’s antitrust subcommittee).  But the basic content of Kohl’s letter is perfectly aligned with the claims of the deal’s critics to which he refers.  Unfortunately, there is just no support for the claims being made against the deal, and there is a substantial likelihood that intervention would be in error.

The “Leveraged Dominance” Claim

First of all, neither Kohl’s letter nor the “critics of this transaction” explain what it would mean for Google to “leverage its dominance of [the] PC-based search advertising market into the emerging mobile advertising market, particularly with respect to advertising embedded in smart phone applications,” but it sure sounds scary.  Leveraged dominance. Very bad.  Whatever it means.

If mobile application advertising competes with other forms of advertising offered by Google, then it represents a small fraction of a larger market and this transaction is competitively insignificant.  Moreover, acknowledging that mobile advertising competes with online search advertising does more to expand the size of the relevant market beyond the narrow boundaries it is usually claimed to occupy than it does to increase Google’s share of the combined market (although critics would doubtless argue that the relevant market is still “too concentrated”).  If it is a different market, on the other hand, then critics need to make clear how Google’s “dominance” in the “PC-based search advertising market” actually affects the prospects for competition in this one.  Merely using the words “leverage” and “dominance” to describe the transaction is hardly sufficient.  To the extent that this is just a breathless way of saying Google wants to build its business in a growing market that offers economies of scale and/or scope with its existing business, it’s identifying a feature and not a bug.  If instead it’s meant to refer to some sort of anticompetitive tying or “cross-subsidy” (see below), the claim is speculative and unsupported.

Professional Google critic, Scott Cleland, tries to put some meat on the bones:

  • An interesting and provocative post — by Bill Gurley of “Google Redefines Disruption: The ‘Less Than Free’ Business Model” — shows how Google-AdMob could be viewed in the larger antitrust context of Google attempting to extend its search advertising monopoly to wireless via predatory cross-subsidization.
  • Mr. Gurley raves about the unstoppability of Google’s “less than free” business model in wireless:
    • Google will give you ad splits on search if you use [the Google] version [of its android OS]!  That’s right; Google will pay you to use their mobile OS. I like to call this the “less than free” business model. This is a remarkable card to play. Because of its dominance in search, Google has ad rates that blow away the competition. To compete at an equally “less than free” price point, Symbian or windows mobile would need to subsidize.”
  • Mr. Gurley raises the issue, and the antitrust review of Google AdMob would provide a forum to investigate to see if Google is anti-competitively leveraging its market power in search advertising to extend its monopoly into mobile.
    • How Google intends to bundle its search, ad-serving, operating system, browser, etc. with AdMob will have big implications on how antitrust authorities would ultimately view this transaction. [Emphasis Cleland’s]

Predatory cross-subsidization”!! Why that’s scarier than leveraged dominance!  And it’s an interesting concept–especially because, unlike the normal predation claim, it purports not to require speculative, future recoupment to make the tactic worthwhile.  But, alas, there are still some serious problems.  First, even taking the argument on its face, as with all allegedly-predatory conduct, the immediate effect in the relevant market would be a reduction in prices–a subsidy.  In this case, the corresponding increase in prices would occur not in the merging parties’ market but elsewhere.  I imagine it’s hard to make out a merger case on the basis of the claim that the merger will lower prices in the merging parties’ market.  To do otherwise (to point to offsetting costs in another market) would create a serious problem for merger review, especially if one takes Philadelphia National Bank seriously: If efficiencies in other markets can’t save a merger, then surely cost increases in other markets can’t condemn a merger.  And if there’s a problem in the “PC-based search advertising market,” then that’s a different litigation, not this one.

Next, the predatory subsidy claim rests on the assertion that Google can charge monopoly rates in the PC-based search advertising market–and that these rates could go even higher to subsidize the predation.  First, there is no proof that Google does enjoy such a monopoly, nor the requisite flexibility to raise prices–no matter how many times Google’s critics state it as fact.  More important, there is also no demonstration that Google would have the ability to recoup its losses in the mobile-based market with price increases, even if it is a monopolist:  The claim requires questionable assumptions about demand elasticity and there is no reason to think either a rational monopolist or a competitive market participant would be able to offset the amount of revenue lost through predation in one market with higher prices in another in which it is presumably already maximizing profit.

To the extent that the claim is just, “well, monopolist Google already has scads of money, so it can afford to finance this ‘leveraging’ out of its current (monopoly) revenue,” the claim is absurd.  First off, then we’re just in the world of traditional predatory pricing.  Absent a strong demonstration of future recoupment, this is not likely a claim that would sit well with shareholders: Throwing away money, whether earned in a competitive market or a monopoly one, still hurts the bottom line.  Moreover, I can assure you we will never see the empirical analysis to support a speculative, future recoupment story (which is why such claims always rest on fluffy intent evidence rather than actual evidence).  Second, the ability to spend money is not unique to Google, whether you call it a monopolist or not.  Apple, beat out by Google in the competition for AdMob, also has scads of money.  So does Microsoft.  So does JPMorgan–and thus any company it is willing to lend to.  If the claim rests on Google’s unique ability to find the funds required to finance its costly empire-building, it is misplaced.  If it would be profitable for Google to incur the costs of both buying AdMob as well as engaging in predatory pricing in order to dominate the mobile advertising market, it would likely have been profitable for AdMob to borrow the money to do it itself (or else negotiate away Google’s expected gain in the sale price).  Otherwise we’re left with an argument that Google can do it more efficiently (in which case the claim cuts against challenging the merger), or else a claim that Google could be a more effective monopolist in mobile advertising, which is probably what Kohl is implying, but which is just hand-waving—and we still haven’t heard why this would be true.

Third, this predatory cross subsidy would be a tactic for obtaining market share in the secondary market–an alternative to, say, buying an existing competitor;  it’s not clear how buying AdMob facilitates this leveraging.  If you’re Google and you want to increase your presence in the mobile ad market, you might offer a subsidy to attract customers instead of buying AdMob, or you might buy AdMob and its customer base instead of fueling your expansion with customer subsidies.  At first cut, the tactics are substitutes, and the merger actually decreases the likelihood of predation.  (And don’t forget that buying AdMob will cost Google $750 million.  Predatory pricing on top of that would make the merger that much more costly).  Unless the merger itself somehow makes predation more effective (to the present value tune of $750 million), it’s hard to see the one as an explanation for the other.

Now, second cut, maybe you do both–maybe buying AdMob gives you some more market share, but you want even more, and so you cut prices on top of the merger in an effort to further increase scale.  But neither the ability nor the incentive to engage in anticompetitive predation in the mobile market is affected by the merger in this case.  We should challenge mergers only if they alter (for the worse) post-merger pricing incentives, and, if predatory cross-subsidization is the source of the concern here, this merger either doesn’t seem to alter the incentive to engage in predatory behavior or it actually reduces it.  It shouldn’t be enough to just scream the words “predatory cross-subsidization!” to justify a costly antitrust intervention.

As an additional point to support these claims, notice that the type of cross subsidization that Gurley is discussing (favorably, I might add) in the comments quoted by Cleland is one where one distinct group of customers (purchasers of PC-based online advertising) allegedly subsidizes a completely different group (cell phone manufacturers/carriers).  But to extend this claim to the AdMob merger case requires purchasers of PC-based online advertising to subsidize purchasers of mobile-based online advertising.  These might be somewhat distinct groups, but I bet there is enormous overlap.  It’s an odd subsidy (and one the relative elasticities are unlikely to support) that takes from Peter to pay . . . Peter.

This also points up the silliness of this kind of “predatory cross-subsidy” analysis.  The notion originates in (and, as far as I can tell, has remained until now completely relegated to) the world of regulated utilities, where a utility enjoying a government-mandated monopoly rate in market A convinces the government to raise the rate there to support the utility’s “predatory” expansion into unregulated, ancillary market B.  That dynamic is unique and it is not present here–in that case the price increase to finance the predation is mandated and enforced by the government, reducing the otherwise impossible problem of explaining where recoupment comes from.  Moreover, even in the regulated utility case, making out such a claim would rest on a demonstration that profit margins in the predatory market are lower than those in the monopolized market.  Leaving aside the enormous (and probably fatal) problem of arbitrarily allocating costs across markets to “demonstrate” these margins, it should always be the case that margins are smaller in competitive than in monopoly markets.  Demonstrating this differential would demonstrate nothing of antitrust relevance–and yet one can be certain that proponents of the claim would build their case on precisely such evidence.

Finally, notice that if Google and AdMob are in the same market (Cleland:  “Anyone that [sic] understands what Google does knows that Google does the same thing AdMob does, just less in the mobile space than AdMob does.”) for this analysis, critics need to revise their fundamental “market share” claims about Google–claims that never include any firms other than Google, Yahoo! and Microsoft in the denominator.

Error Costs

Kohl dismisses, of course, Google’s claims that the FTC should be reluctant to regulate such a new, dynamic and innovative industry.  But waiving off this concern is, while common these days, inappropriate and dangerous.  It is precisely in this sort of dynamic, innovative and not-yet-understood market where the risk and cost of deterring beneficial business models and strategies (to say nothing of technological progress) are highest.  To claim that the industry’s newness and dynamism are not a reason to forebear from intervention is ill-considered, unsupportable, and backward.  But don’t just take my word for it, or even Google’s:

The smartphone market is at an early stage of development.  That may appear to be a surprising statement – after all, mobile phones have been around since the 1970s, and have been common devices since the 1990s . . . .  But the principal functionality of and value proposition for the phone has changed.

Now . . . the value proposition has moved to the software stack.  As is clear from advertising by all of the major brands . . . people buy smartphones because they are fully functional computers that fit in the palm of your hand. . . . But what is most valuable is not the connection per se, but the new things that users can do with it – find nearby restaurants and movie theaters, send and receive email, and watch video, just to name a few.  The primary driver for adoption and sales in this market is the software on and available for the device.

The smartphone market is still in a nascent state; much innovation still lies ahead in this field.

This is Microsoft talking.  The specific discussion here relates to hardware/software innovation and patent rights, but the analysis is the same:  As the hardware technology, usage and market size, cost, and software for mobile phones changes, so do the strategy and profitability of the various business models that build up around them.  Whatever Google tries to do at this early stage of mobile phone market evolution, it will face challenges from competing business models not yet conceived of, changes in underlying hardware and software, and demographic/usage/consumer preference changes that will make any market power it might enjoy both fleeting and important in catalyzing the very competitive evolution that will undermine it.  Far from being irrelevant to the propriety of a merger challenge, the newness and dynamism of the mobile phone market is essential to this determination.

And on this point, as is well-known by now, Apple is believed to be announcing its own mobile advertising platform tomorrow.  This can only be viewed as an enormous competitive challenge to Google’s plans in this area.  But it may portend even more.  Here, for example, is one interesting discussion of what the future of competition in the mobile phone world might look like for Google and Apple:

Data Apple collects about users from its vaunted iPhone is so valuable that the company must build a special search engine just to keep Google from gleaning insight from that data, analysts say.  Piper Jaffray analyst Gene Munster said there is a 70 percent chance Apple will roll out a mobile search engine tailored for its iPhone within the next five years.

* * *

Building its own iPhone-centric search engine would help Apple shield Google from its App Store data, Munster said in a March 30 research note.

“We believe Apple could utilize data unavailable to Google, data generated by the company’s App Store, to create a mobile centric search engine, which would be a unique offering to Google’s search engine,” Munster wrote.

Apple lacks the experience and engineering wherewithal to build a large, scalable search engine. There are alternatives to Google such as Microsoft Bing, which was rumored to replace Google on the iPhone. With Google the default search service on Apple’s newly released iPad, it seems unlikely that Apple will in fact replace Google with Bing or anything else on the iPhone.

But Munster said Apple could buy a search startup with a Web index, such as Cuil, using its index as the seed for its own search engine. Mobile search startup Taptu would also be a good possibility for Apple because it focuses its index on touch-enabled Websites.

* * *

IDC analyst Hadley Reynolds added that local search is the initial killer app for mobile, noting that the Quattro ad platform could make for an attractive environment for advertisers, particularly when paired with Taptu’s touch-screen approach to search.

Meanwhile, iPhone apps from Google and Bing are still delivering long lists of links, which are hard to deal with in small screen real estate. Moreover, these apps do not have access to the data that Apple has piling up in its iTunes Store and on its devices, Reynolds told eWEEK.

“The next-generation services will be much more touch-enabled than what Google and Bing are offering, and they will take users to sites that offer the most attractive info consumption models to their audiences,” Reynolds added.

“Apple is in an inside position to tap into the current pent-up demand for better mobile search, and add a new competitive differentiation from other search providers and device makers.”

Assumptions about the trajectory of competition in this area that don’t take account of these dynamics and the immense uncertainty surrounding them are both integral to making the case against the Google-AdMob merger and also, unfortunately, unsupportable.  The risk and cost of error here are extremely high, particularly when the opposition resorts to hand-waving, troubling assumptions and political pressure to push its positions.


A final, quick word about privacy.  Kohl’s letter states that

The FTC should also pay close attention to the privacy interests implicated by this transaction, as the combined firm will gain access to a treasure trove of data on millions of consumers’ behavior, search and product preferences.    The FTC should assure itself that the deal, if approved, will have sufficient safeguards to protect consumers’ privacy.

No one has put forth an antitrust-relevant theory to support these sorts of claims.  I will have more to say on this issue if it continues to garner the undeserved attention it has enjoyed thus far.  But here is a short response:

The data in question currently exists.  The claim here is that the same amount of data in the hands of one firm instead of two presents a problem, and that any such combination must be accompanied by “safeguards to protect consumers’ privacy.”  There is no indication why privacy is more in danger when the two databases are combined.  There is no definition of “privacy,” for that matter.  Is the fear that my data is more likely to be unintentionally released into the public domain?  I don’t see why this is any more likely if Google controls both databases.  Is the fear that my data is more likely to be used in Google’s decision-making when combined than when separate?  First, I see no reason why this would be so, and second, this offers huge potential benefits, if true.  How does it help me to “safeguard” my privacy by making the products I use otherwise less valuable to me?  Privacy’s optimum is certainly not the maximum, and the optimum differs for every person.  How is this to be incorporated into an antitrust analysis?   Related to this, the implication of this kind of approach is that any efficiency that might be realized from a single firm having access to a larger or more robust database of information is not cognizable, but is, in fact, a bug and not a feature.  This would threaten to condemn some efficiency-enhancing mergers by disregarding a potentially-important source of efficiency by labeling it a “privacy degradation” instead of an efficiency.  Finally (for now), where concentration of data entails the pooling of lots of people’s data, why is this of any concern to me or any other individual?  Is my privacy any more at risk if Google has access to another 10 million people’s data?  If anything the opposite is true.

Until proponents of incorporating privacy analysis into antitrust review–especially merger review–put forward anything resembling an antitrust-relevant theory of how mergers (or other conduct) could harm privacy instead of just parroting what amounts to an unsupported conduct-structure-performance assertion, the FTC should not “pay close attention to the privacy interests implicated by this transaction.”

If this merger ends up challenged it will almost certainly be because the FTC can’t or won’t resist the political and public pressure to “do something” about Google, but the agency would have a tough time proving its case.  These silly leverage and predation theories won’t go anywhere, and a case would have to be built on a crabbed market definition and a straightforward market share/HHI analysis–problematic as that is.  Let’s hope the news reports are wrong.