The recent Consent Decree between the Department of Justice and Google sets a useful precedent for allowing, even requiring, oversight of Google, but it does not go far enough. It seems to embrace a flawed view of consumer welfare, and it therefore permits things in the name of consumer convenience that are harmful to competition. As importantly, it fails to understand the tremendous power of a search engine's owner to deny competitors fair access to consumers that both are seeking to serve.
The Decree therefore misses the greatest risks of the ITA acquisition and misses the points that the DoJ should have been seeking to regulate. This acquisition is not about a line extension of a software vendor into a new line of business, where the vendor and its current customers both want their profitable relationship to continue; it is about vertical encroachment by a search company into precisely those lines of business that searchers were trying to locate.
The acquisition will inevitably be extremely damaging to hotels, airlines, and travel service companies, of course. By extension, the principles embraced in the Consent Decree will allow Google to move into the sale of additional goods and services, and if these extensions are created organically rather than through acquisition, this consent decree may even be read as permitting the extensions to proceed without further review. These extensions would allow Google to create almost unlimited new revenue streams, from the direct sale of goods and services, by preferentially positioning its sites selling those goods and services above those of any and all competitors and weakening or destroying the competitive process.
Google would be able to use these additional revenue streams to cross-subsidize still more preemptive line extensions, and would be able to damage still more software vendors, including previously unassailable players like Microsoft and Amazon.
By damaging the competitive process the ITA acquisition inevitably harms consumers and should be blocked.
The ITA Consent Decree and Its Limitations
The recent ITA Consent Decree between the Department of Justice and Google contains something to please everyone; it is a compromise that seeks to achieve some sort of balance. Favorable comments can be found at Google's own site, of course. Google opponent FairSearch has a harsh comment on its home page, but still has a large number of supportive comments as well, principally because the DoJ has, for the first time, acknowledged the need to impose regulation on Google.
Of course, the Decree has something to disappoint everyone as well. Although it does establish an important precedent in putting Google on notice that its actions are going to be monitored, regrettably, we think that it falls far short of what is required. Indeed, the Decree appears to be severely flawed, and it appears to be based on a serious misunderstanding of consumer welfare in this case. As others have already noted, the Decree does not go far enough to protect sellers of travel services (hotels and airlines), and consequently it does not go far enough to protect consumers; any excessive fees charged to hotels and airlines will eventually be passed on to consumers. Ultimately, the Decree will need to be revised, and regulation of Google will need to be tightened. It is likely that the Decree itself should be rejected.
The Decree, as widely reported, allows Google to acquire ITA. Because Google is the largest search engine provider in the world, and ITA is one the most innovative and powerful providers of fare-calculations for online travel agencies, there was justifiably some concern expressed by the United States Department of Justice that the acquisition "would have substantially lessened competition among providers of comparative flight search websites in the United States, resulting in reduced choice and less innovation for consumers." The Decree lays out certain requirements for this acquisition to proceed, including the following:
And yet, as constructive as these terms sound, and as confident as the DoJ appears, the following risks were not adequately addressed:
We know from the Microsoft antitrust litigation that software interface designs are constantly changing as code is improved, and we know that competitors who are not given advance notice of changing interface design are always behind in incorporating the latest releases of software.
Our own experience is that it is very difficult to detect snooping when companies use a competitor's offering to perform an essential service. For example, one airline suspected that it was receiving insufficient revenue from a reservations system operated by a competitor, while the competitor claimed that it was giving the first airline more than its fair share of tickets. Only by going through actual tickets and examining the source of the reservations and performing complex data analyses on the huge data set were we able to figure out what was really happening.
The first airline was indeed being given more than its fair share of tickets, but it was mostly getting those tickets that were booked with deeply discounted fares, while the company that operated the reservations system was keeping the full fare customers for itself on days when it knew the flights were likely to sell out; that is, the first company was getting more than its fair share of tickets, but less than its fair share of revenues and profits. What exactly could the Google of Wi-Spy and Doodle do to competitors without detection if it chose, why exactly would it stop snooping now, and how exactly would competitors know if Google had accessed and abused their sensitive data?
This last point is certainly the most important, and will be addressed in more detail. The Consent Decree itself attempts to ensure that competitors can continue to use ITA software on equitable terms; it is more important to ask if competitors will even be able to exist, with or without access to ITA software, if Google is permitted to encroach through vertical integration into the competitors' lines of business. While the first two points are relevant in that they will need to be addressed to make the consent decree workable, the last point may actually require that the consent decree be rejected. Before examining the issue of fair access we first we need to consider some common, indeed pervasive, misconceptions that have come to dominate the Google antitrust debate.
Misconceptions About Google and Antitrust
The discussion of Google and antitrust has shifted in recent weeks from relevant market (is Google 65% of paid search, 32% of online advertising, or 3% of all advertising?) and anticompetitive behavior (is Google engaging in "preemptive line extensions," unfairly squeezing out competition by pricing its non-search offerings below cost, or unfairly listing its own offerings above those of competitors?) to consumer happiness and consumer convenience. Mainstream publications have adopted the consumer happiness argument.
There are three problems with this analysis:
The relationship between consumer happiness and consumer welfare, and the relationship between current and future consumer welfare, are explored in a previous post.
What Form Could Harm Take?
We believe that shifting the discussion away from Google's share of the relevant market and possible antitrust violations to consumer happiness as an indication of consumer welfare is dangerously misleading. Surely a search engine can keep consumers happy, with convenience, among other things. And surely, by bypassing competitors and redirecting consumers to its more convenient web offerings it could grow its share. Then, it could use its market power to demand discounts, which it would share with consumers, much as hotels.com did. Ultimately, competing websites would lose importance, and indeed they might fail, further increasing Google's market power. Then, Google could demand excessive commissions, perhaps as high as the 30% commission once charged by hotels.com. Consumers would not complain; they would be receiving apparent discounts, discounts below hotels and airlines published prices. Indeed, consumers would be happy with the added convenience and apparently lower price.
Notice the progression:
It really would not possible for the hotels' and airlines' own direct distribution sites to compete, or for travel integration sites like Kayak or Orbitz to compete, unless consumers were willing to examine and compare the offerings of each URL separately rather than using search at all. Not only is there no limit to what Google could charge travel sites to "not be not found," which is the problem existed before Google's move into search; now Google would have a reason to hide some or all of these sites no matter what they are willing to pay.
The nature of harm to the competitive process, and the nature of consumer harm that would inevitably result, are explored in a previous post.
The Bottom Line -- Harm and Remedies
Consumers appear to be well-served at present, and consumers may even believe they are well-served at present, but ultimately and inevitably they will not be well-served in the future if the Consent Decree is adopted as written. As competition in travel distribution is destroyed, and it will be, travel providers will pay more and more for distribution. As Google's power in travel distribution increases, and it will, Google will be able to provide consumers with larger and larger apparent discounts, further increasing Google's power and further increasing consumers' perceived satisfaction with Google. And yet travel providers will inevitably pass through some of these higher distribution costs, making consumers' apparent discounts entirely illusory. Consumers will be happy with the apparent discounts, but they will indeed be worse off.
As a general matter, vertical integration by search engines creates grave risks for the competitive process, because such vertical integration increases the risk of search engines placing its product or service listings above those of others. We learned from the experience of Sabre, Apollo, and the airlines that depended upon them that control of access to customers can be a overwhelmingly powerful tool to use against competitors and an irresistible temptation to use, and we know how similar consumers' use of search today can be to travel agents use of search a quarter of a century ago. Indeed, these allegations were raised in an EU complaint against Google last year by the founders of the British price comparison site, Foundem, and the issues were also raised in an investigation by the Texas Attorney General. The complaint alleged that Google provided its own products with prime placement on its own search engine to the detriment of competing services. Similar abuses, and similar complaints, appear inevitable if Google's move into travel services is allowed to proceed.
We believe that search is an essential facility for reaching consumers, and that without access to search it is not possible to operate any business that requires that consumers be able to find it to be able to fairly compare and fairly assess its offerings. We believe that there is no compelling reason to permit vertical integration by the owner of such an essential facility into any comparison-based business or into any direct retailing-based business.
The benefits to consumers are illusory; the harm to a consumer from having to compare several URLs, fairly presented, and click upon one is immeasurably small, while the risks of damage to the competitive process, as outlined above, are real, large, and measurable. The risks are simply too great, even if all search engines could be counted on to act in a manner free from any alleged evil; the history of Google merely underscores the fact that search power corrupts, and that absolute search power corrupts absolutely, regardless of initial purity of intentions.
We do understand that our position on this is still in the minority. Most players in the travel industry are focusing on the fact that they will still have access to ITA's software, and not on the fact that, like Ciao.co.uk, they may not continue to have access to their own customers. We may not get much support from the DoJ; their own former Chief Economist Bruce Owen apparently believes that "vertical integration is welfare enhancing, even when market power is present." But we are not entirely alone. States' Attorneys General, including Rhode Island and Texas, and some members of Congress, are starting to look at what Google can now do.
Google's acquisition of ITA should be blocked. Subsequently, all attempts at vertical integration by a search engine provider into comparison-based businesses or into any direct retailing-based businesses should likewise be blocked.
Eric Clemons is the founder and Project Director for the Wharton School's Sponsored Research Project on Information: Strategy and Economics, within the Program For Global Strategy and Knowledge Intensive Organizations, which supports his research. Microsoft has recently agreed to participate in the Project. Nehal Madhani is an attorney in New York. Previously, he founded an online classifieds service for college students and advised small businesses and nonprofits on legal and strategic considerations.