Editor’s Note: This is the third post in an ACSblog debate on antitrust scrutiny of Google between Harvard Business School Professor Benjamin G. Edelman and George Mason University School of Law Professor Joshua D. Wright. This online debate follows a recent U.S. Senate hearing on whether Google’s business practices “serve consumers” or “threaten competition.”
By Benjamin G. Edelman, Assistant Professor, Harvard Business School
Professor Wright questions whether Google biases results towards its own services, and asks whether consumers are harmed even if Google does bias its results. I don’t find these questions so difficult, and while Professor Wright suggests we’d struggle to identify appropriate remedies, I see some straightforward solutions.
Let’s start with the question of whether Google biases its results towards its own services. On a whim, I ran a search for pop superstar Justin Bieber. Google’s top-most link promoted Google News (in oversized bold type). Down a few inches came a “Videos” section where three thumbnails and three video titles all linked to YouTube clips. (Less prominent links identified other services showing these same videos – links added only after critics flagged the problem of Google always directing this traffic to its own video site.) Lower, Google presented a block of Google Images results. In the analogous context of extra-prominent links to Google Finance, Google’s Marissa Mayer argued that the company should be permitted to put its own links first. “It seems only fair right, we do all the work for the search page and all these other things, so we do put it first.” Marissa doesn’t dispute that Google favors its own links – and she couldn’t, when Google’s links widely appear in prominent ways no other service enjoys.
And what of the consequences of Google’s bias? Professor Wright posits an “efficient bias” wherein Google usefully offers consumers its full suite of services. Certainly it’s handy to have a single Google password providing access to personalized search, finance, videos, and more. But this misses the serious harms of Google’s ever-broadening panoply of services.
Consider an advertiser, say a hotel, dissatisfied with high prices for Google’s dominant AdWords advertising service. If Google prominently features links to Expedia and Tripadvisor, the hotel can strike deals with those sites to promote its property – a plausible alternative to high prices for ads from Google. But consider Google’s recent changes to its search result format. Where Google used to link to Expedia, Google Hotel Finder now appears front-and-center – pushing Expedia links lower and less prominent. And where Google used to link to Tripadvisor, users now see Google Places – which requires hotels and booking services to pay Google to get direct booking links. (Adding insult to injury, Places also asks a hotel to bid against its competitors for ads on its own Google Place page. If the advertiser bids too low or refuses to participate, Google features competitors instead.) Sending less traffic to alternative advertising venues like Expedia and Tripadvisor, Google can raise prices with greater confidence, and advertisers have little means of escape. There’s nothing “efficient” about that; Google raises price above marginal cost, restricts supply, and takes its pound of flesh from advertisers who have little alternative.
Wright suggests we should focus on harm to consumers. In the long run, consumers certainly suffer when innovators can’t launch businesses or get financing for fear of Google blocking their opportunities. Who would launch a video sharing site, knowing that Google overwhelmingly sends video-related traffic to YouTube? And if savvy developers envisioned a new mapping site superior to Google Maps, perhaps with better printing or clearer instructions, that team would struggle to reach consumers since Google systematically features its own service whenever a search calls for a map. These foreclosures impede competition, slow innovation, and are a proper subject of antitrust inquiry.
Meanwhile, advertisers continue to suffer a particularly clear-cut harm – and since advertisers’ payments fuel Google’s $30+ billion annual revenue, antitrust authorities absolutely must consider their plight. As I argued in my opening piece, Google has been thug-like in its imposition of exceptionally harsh terms. Google offers no defense of its take-it-or-leave it terms; Google knows that even the largest advertisers have no viable alternative.
Professor Wright questions what remedy is appropriate for Google’s ever-expanding scope. I recently suggested several remedies for search bias, grounded in tried-and-true remedies antitrust authorities have applied in similar circumstances. For example, two decades ago, travel agents used reservation systems that were owned by airlines, and each airline’s reservation system favored its own flights – making it hard for travel agents or passengers to find the flight that actually best met their needs. Department of Justice litigation put a stop to this practice, disallowing reservation systems from sorting flights based on improper factors like carrier identity. The analogue here is that Google shouldn’t favor its own services just because they come from Google; putting Google Finance first because it’s most popular might be fine if it actually were most popular (it isn’t), but Google ought not put its services first just because they come from Google.
More recently, the European Commission required Microsoft to offer a “browser ballot box” to let users easily choose their preferred web browser, even a browser that competes with Microsoft’s own offering. Such a choice can also be provided within search results: When a user seeks information that matches a predefined vertical (like video, pictures, finance, or news), a drop-down box or other listing could let the user choose a preferred vendor. A user might choose Google for ordinary web search, but prefer Hulu’s video index, Yahoo’s stock quotes, Yelp’s local results, and Amazon’s product search. A bit of AJAX would let users switch their providers any time. Suddenly Google would be far less able to leverage its dominance in search to achieve dominance in other categories. That would be a major benefit to users, advertisers, and the entire online economy.
Two More Points to Consider
First, the "do they or do they not" bias results discussion is largely a distraction in modern antitrust analysis. The question is whether Google's search practices foreclose rivals sufficiently to raise barriers to entry and generate anticompetitive effects. Anecdotal evidence on these points is insufficient. But it is worth correcting the Mayer quote above; to save space, readers are referred to Danny Sullivan's correction here: http://searchengineland.com/survey-google-favors-itself-only-19-of-the-t....
Second, Professor Edelman gives me far too much credit when he writes "Professor Wright posits an “efficient bias” wherein Google usefully offers consumers its full suite of services." The idea that vertical integration or discrimination in favor of one's own products can be efficient is not my own. Credit may properly be attributed to Coase, Klein, Alchian, Hart, Holmstrom, Williamson, and even back to Cournot. These are old ideas. And distinguishing between foreclosure and efficient bias is at the heart of any modern attempt to diagnosis potentially exclusionary conduct under the antitrust laws.
It is in this light that the point that not only Google has evolved toward universal results and referral to its own content; but also Microsoft's Bing. Professor Edelman's own work demonstrates this; and subsequent analysis confirms it. But Google has market power one might object! In antitrust, the general conventional wisdom (for good economic reason) is that when firms with and without market power, i.e. when the industry, adopts a particular practice it is highly likely to be efficient. Such is the case here.
Remedies for Search Bias Good for Consumers? CRS Wasn't
You describe the remedies as "tried and true" --- but were they successful?
There has been ample study of the effects of the travel agent CRS remedy you appeal to. Sure, imposing a remedy is easy. But is it any good at improving consumer welfare? Alexander and Lee examine the CRS remedy and find that "[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." It gets worse. CRS regulations appear to have caused serious harm to the competitive process and made consumers worse off. Smith (1999) concludes that "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." Further, as with Google search, the CRS was imposed despite evidence that it had improved consumer welfare. 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). And in your discussion of the CRS model for regulation, you fail to mention that the DOT terminated the regulation in 2004 in light of its failure to improve competitive outcomes and a growing sense that they were making things worse, not better. Seems like an important fact to consider in the debate.
Post new comment