I did not intend for this to become a series (Part I), but I underestimated the supply of analysis simultaneously invoking “search bias” as an antitrust concept while waving it about untethered from antitrust’s institutional commitment to protecting consumer welfare. Harvard Business School Professor Ben Edelman offers the latest iteration in this genre. We’ve criticized his claims regarding search bias and antitrust on precisely these grounds.
For those who have not been following the Google antitrust saga, Google’s critics allege Google’s algorithmic search results “favor” its own services and products over those of rivals in some indefinite, often unspecified, improper manner. In particular, Professor Edelman and others — including Google’s business rivals — have argued that Google’s “bias” discriminates most harshly against vertical search engine rivals, i.e. rivals offering search specialized search services. In framing the theory that “search bias” can be a form of anticompetitive exclusion, Edelman writes:
Search bias is a mechanism whereby Google can leverage its dominance in search, in order to achieve dominance in other sectors. So for example, if Google wants to be dominant in restaurant reviews, Google can adjust search results, so whenever you search for restaurants, you get a Google reviews page, instead of a Chowhound or Yelp page. That’s good for Google, but it might not be in users’ best interests, particularly if the other services have better information, since they’ve specialized in exactly this area and have been doing it for years.
I’ve wondered what model of antitrust-relevant conduct Professor Edelman, an economist, has in mind. It is certainly well known in both the theoretical and empirical antitrust economics literature that “bias” is neither necessary nor sufficient for a theory of consumer harm; further, it is fairly obvious as a matter of economics that vertical integration can be, and typically is, both efficient and pro-consumer. Still further, the bulk of economic theory and evidence on these contracts suggest that they are generally efficient and a normal part of the competitive process generating consumer benefits. Vertically integrated firms may “bias” their own content in ways that increase output; the relevant point is that self-promoting incentives in a vertical relationship can be either efficient or anticompetitive depending on the circumstances of the situation. The empirical literature suggests that such relationships are mostly pro-competitive and that restrictions upon firms’ ability to enter them generally reduce consumer welfare. Edelman is an economist, with a Ph.D. from Harvard no less, and so I find it a bit odd that he has framed the “bias” debate outside of this framework, without regard to consumer welfare, and without reference to any of this literature or perhaps even an awareness of it. Edelman’s approach appears to be a declaration that a search engine’s placement of its own content, algorithmically or otherwise, constitutes an antitrust harm because it may harm rivals — regardless of the consequences for consumers. Antitrust observers might parallel this view to the antiquated “harm to competitors is harm to competition” approach of antitrust dating back to the 1960s and prior. These parallels would be accurate. Edelman’s view is flatly inconsistent with conventional theories of anticompetitive exclusion presently enforced in modern competition agencies or antitrust courts.
But does Edelman present anything more than just a pre-New Learning-era bias against vertical integration? I’m beginning to have my doubts. In an interview in Politico (login required), Professor Edelman offers two quotes that illuminate the search-bias antitrust theory — unfavorably. Professor Edelman begins with what he describes as a “simple” solution to the search bias problem:
I don’t think it’s out of the question given the complexity of what Google has built and its persistence in entering adjacent, ancillary markets. A much simpler approach, if you like things that are simple, would be to disallow Google from entering these adjacent markets. OK, you want to be dominant in search? Stay out of the vertical business, stay out of content.
The problems here should be obvious. Yes, a per se prohibition on vertical integration by Google into other economic activities would be quite simple; simple and thoroughly destructive. The mildly more interesting inquiry is what Edelman proposes Google ought provide. May, under Edelman’s view of a proper regulatory regime, Google answer address search queries by providing a map? May Google answer product queries with shopping results? Is the answer to those questions “yes” if and only if Google serves up some one else’s shopping results or map? What if consumers prefer Google’s shopping result or map because it is more responsive to the query. Note once again that Edelman’s answers do not turn on consumer welfare. His answers are a function of the anticipated impact of Google’s choices to engage in those activities upon rival vertical search engines. Consumer welfare is not the center of Edelman’s analysis; indeed, it is unclear what role consumer welfare plays in Edelman’s analysis at all. Edelman simply applies his prior presumption that Google’s conduct, even if it produces real gains for consumers, is or should be actionable as an antitrust claim upon a demonstration that Google’s own services are ranked highly on its own search engine — even if Google-affiliated content is ranked highly by other search engines! (See Danny Sullivan making that point nicely in this post). Edelman’s proscription ignores the efficiencies of vertical integration and the benefits to consumers entirely. It may be possible to articulate a coherent anticompetitive theory involving so-called search bias that could then be tested against the real world evidence. Edelman has not.
Professor Edelman’s other quotation from the profile of the “academic wunderkind” that drew my attention was the following answer in response to the question “which search engine do you use?” After explaining that he probably uses Google and Bing in proportion to their market shares, Professor Edelman is quoted as saying:
If your house is on fire and you forgot the number for the fire department, I’d encourage you to use Google. When it counts, if Google is one percent better for one percent of searches and both options are free, you’d be crazy not to use it. But if everyone makes that decision, we head towards a monopoly and all the problems experience reveals when a company controls too much.
By my lights, there is no clearer example of the sacrifice of consumer welfare in Edelman’s approach to analyzing whether and how search engines and their results should be regulated. Note the core of Professor Edelman’s position: if Google offers a superior product favored by all consumers, and if Google gains substantial market share because of this success as determined by consumers, we are collectively headed for serious problems redressable by regulation. In these circumstances, given the (1) lack of consumer lock-in for search engine use, (2) the overwhelming evidence that vertical integration is generally pro-competitive, and (3) the fact that consumers are generally enjoying the use of free services — one might think that any consumer-minded regulatory approach would carefully attempt to identify and distinguish potentially anticompetitive conduct so as to minimize the burden to consumers from inevitable false positives. With credit to antitrust and its hard-earned economic discipline, this is the approach suggested by modern antitrust doctrine. U.S. antitrust law requires a demonstration that consumers will be harmed by a challenged practice — not merely rivals. It is odd and troubling when an economist abandons the consumer welfare approach; it is yet more peculiar that an economist not only abandons the consumer welfare lodestar but also argues for (or at least presents an unequivocal willingness to accept) an ex ante prohibition on vertical integration altogether in this space.
I’ve no doubt that there are more sophisticated theories of which creative antitrust economists can conceive that come closer to satisfying the requirements of modern antitrust economics by focusing upon consumer welfare. Certainly, the economists who identify those theories will have their shot at convincing the FTC. Indeed, Section 5 might even open the door to theories ever-so slightly more creative and more open-ended that those that would be taken seriously in a Sherman Act inquiry. However, antitrust economists can and should remain intensely focused upon the impact of the conduct at issue — in this case, prominent algorithmic placement of Google’s own affiliated content its rankings — on consumer welfare. Because Professor Edelman’s views harken to the infamous days of antitrust that cast a pall over any business practice unpleasant for rivals — even if the practice delivered what consumers wanted. Edelman’s theory is an offer to jeopardize consumers and protect rivals, and to brush the dust off antiquated antitrust theories and standards and apply them to today’s innovative online markets. Modern antitrust has come a long way in its thinking over the past 50 years — too far to accept these competitor-centric theories of harm.
