Sacrificing Consumer Welfare in the Search Bias Debate

Josh Wright —  22 April 2011

Eric Clemons and Nehal Madhani have a posted a series of short blog posts on the Huffington Post focusing on Google, antitrust, and more specifically, vertical integration and search (Part I, Part II, and Part III).   The articles contain much of the standard hand-wringing about vertical integration and its impact on consumer welfare.  This is an issue I’ve written about here and here.  On its own, their analysis adds little new insight to understanding these issues from an antitrust perspective.  The series of posts are noteworthy, however, for two reasons.   The first is that Clemons and Madhani offer the very awkward economic argument, but one that I’ve observed with increasing frequency when it comes to search, that consumers cannot be trusted to make decisions that improve their own welfare.  For example, Clemons and Madhani concede that many of the practices they and others have criticized with respect to search provide consumer benefits.  However, Clemons and Madhani write:

It should be obvious that consumers are not always the best judges of their own welfare, and it should be obvious that consumers have extreme difficulty judging whether actions will improve their welfare if the results follow from complex interactions, occur after a significant delay, or both. The argument that government should help consumers through regulation sounds paternalistic, but if consumers were always the best judges of their long-term welfare, we would not have problems with smoking or obesity.

We know that consumers often make bad decisions when an experience is immediately pleasurable and when harm is deferred or the relationship between cause and effect are complex and not immediately visible. Free software is pleasurable. This free software is funded through excessive charges imposed on companies that need to pay to be found through search; consumers cannot readily observe the harm that comes from these excessive charges because the complex mechanisms by which these charges are passed along to consumers are not directly observable.

One need not be an extreme adherent to the notion that consumers generally behave rationally and in their self-interest to reject the contention that consumer preferences ought to be given zero weight in the analysis.  That’s not entire fair.  Clemons and Madhani do not give zero weight to consumer preferences; they appear to assign consumers’ valuation of these services negative weight.  That is, the authors argue not only that consumers are simply not capable of knowing whether they derive greater value from X or Y, but also that the fact that consumers choose X, ceteris paribus,  increases the likelihood that X is engaged in anticompetitive behavior.  Quite odd.  And certainly unrelated to any economic conception of consumer welfare.

So what is left of consumer welfare analysis without the consumers?  Its not a trick question.  How does one know that these consumers and their faulty preferences are actually harmed in practice?   Harm to rivals is sufficient, Clemons and Madhani appear to answer:

Google’s revision to its search engine, code-named Panda, substantially reduced the visibility of low quality sites, which is definitely a good thing. But the Panda release also seems to have slammed Ciao.co.uk, a Microsoft-owned company, and a potential competitor as a pricing comparison site, which had been leading an EU competition case against Google.

The real argument that Clemons and Madhani make is that consumers’ initial happiness with increased convenience from search bias helping them find what they are looking for will result in harm to rivals, such as online travel sites that want to be found, that these firms will pay higher fees that are invisible to consumers though they are passed on to them in the form of higher prices, and consumers will “remain content and oblivious to harm.”  It remains unclear as to how consumers, even ones that do not always behave in their best interest, will remain oblivious to higher prices.  There are lots of business models in which consumers do not know the internal cost structure of operating a business, or what percentage of what fees are ultimately passed on.  But the assumption that consumers are altogether insensitive to prices and price competition is an extreme and unrealistic one.

In any event, the explicit paternalism embedded in the approach here are obvious, and in my opinion misguided.  But just entertaining the argument on the merits for a moment, Clemons and Madhani concede that search bias and vertical integration provide at least some real short-term consumer benefits that must be weighed against the long-term competitive risks.  They write: “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.”

This is not altogether unrelated to the right economic approach, i.e. the weighing of short-term benefits against long-term risks.  However, the authors make a key mistake in assessing the risks of harm to competition in the future, i.e. conflating harm to rivals with harm to the competitive process.  As I’ve written:

An antitrust inquiry would distinguish harm to competitors from harm to competition; it would focus its inquiry on whether bias impaired the competitive process by foreclosing rivals from access to consumers and not merely whether various listings would be improved but for Google’s bias.  The answer to that question is clearly yes.  The relevant question, however, is whether that bias is efficient.   Evidence that other search engines with much smaller market shares, and certainly without any market power, exhibit similar bias would suggest to most economists that the practice certainly has some efficiency justifications.  Edelman ignores that possibility and by doing so, ignores decades of economic theory and empirical evidence.  This is a serious error, as the overwhelming lesson of that literature is that restrictions on vertical contracting and integration are a serious threat to consumer welfare.

One might believe, given that Clemons and Madhani at least pay lip service to the notion that search bias improves consumer welfare, that they might advocate some sort of fact-intensive rule of reason analysis that attempted to ferret out anticompetitive examples of search bias or vertical integration from those that are pro-competitive.  Such an approach would at least be consistent with the economic literature on vertical integration — which tells a much different tale about the competitive effects of the practice.  The economic literature on vertical restraints and vertical integration provides no support for ex ante regulation arising out of the concern that a vertically integrating firm will harm competition through favoring its own content and discriminating against rivals.  Economic theory suggests that such arrangements may be anticompetitive in some instances, but also provides a plethora of pro-competitive explanations.  Lafontaine & Slade explain the state of the evidence in their recent survey paper in the Journal of Economic Literature:

We are therefore somewhat surprised at what the weight of the evidence is telling us. It says that, under most circumstances, profit-maximizing vertical-integration decisions are efficient, not just from the firms’ but also from the consumers’ points of view. Although there are isolated studies that contradict this claim, the vast majority support it. Moreover, even in industries that are highly concentrated so that horizontal considerations assume substantial importance, the net effect of vertical integration appears to be positive in many instances. We therefore conclude that, faced with a vertical arrangement, the burden of evidence should be placed on competition authorities to demonstrate that that arrangement is harmful before the practice is attacked. Furthermore, we have found clear evidence that restrictions on vertical integration that are imposed, often by local authorities, on owners of retail networks are usually detrimental to consumers. Given the weight of the evidence, it behooves government agencies to reconsider the validity of such restrictions.

John Maynard Keynes is quoted as saying “When the facts change, I change my mind. What do you do, sir?”  Given the economic evidence on vertical integration, and the degree to which it conflicts with Clemons and Madhani’s priors on its competitive effects, one might expect them to support a rule that at least made an attempt to retain the consumer benefits of vertical integration.  The second noteworthy point about their post, however, is just how much their policy position diverges from the lessons gleaned from the economic literature described above.  The authors would ban vertical integration altogether!  In their own words:

For this reason, we believe that search engines should be forbidden to engage in vertical integration into comparison of goods and services, or into direct sale of goods and services. We feel that this should be blocked irrespective of the mechanism used to create the vertical integration; search engine providers should not be allowed to develop this vertical integration internally nor acquire it through acquisition, regardless of assurances offered to regulators of fair pricing or Chinese walls.

Understanding the competitive economics of vertical integration and vertical contractual arrangements can be quite complex because there are generally anticompetitive theories of the conduct as well as efficiency explanations.  One must be very careful with the facts in these cases to avoid conflating harm to rivals arising from competition on the merits, with harm to competition arising out of exclusionary conduct.  Misapplication of even this nuanced approach can generate significant consumer harm in the form of error costs and reduced incentive to innovate.  Reasonable minds can differ on where to draw the lines on these issues, where to impose safe harbors, and how to allocate burdens.  An outright ban on vertical integration has none of these virtues and rejects mainstream antitrust economics and available evidence.  The unifying theme holding together the two noteworthy points in these posts is that both imply sacrificing consumer welfare in the interest of firms.