Section 2 Symposium: David Evans on "Tying as Antitrust's Greatest Intellectual Embarrassment"

David Evans —  7 May 2009

evansDavid Evans is Head, Global Competition Policy Practice, LECG; Executive Director, Jevons Institute for Competition Law and Economics, and Visiting Professor, University College London; and Lecturer, University of Chicago.

I’d like to propose a contest for the greatest intellectual embarrassment of antitrust. Let me name the first contestant—tying, which some of you know has been one of my favorite for years. Here’s why. First, there is no persuasive theoretical or empirical evidence that tying is a business practice that is likely to harm consumers.  (This is not the blog to deal with Professor Elhauge’s provocative paper except to say that it does not alter this view.)  There is work that says it could be, under stringent conditions, and one can point to cases where maybe the practice has been used in a harmful way.  Yet the courts have put tying in the same antitrust category as price fixing when done by a firm with some market power.   Second, the courts, lacking any analytical framework for detecting bad behavior, have developed a mechanical test for tying that doesn’t have any connection whatsoever to any of the plausible theories of when and why tying might be bad.  The test leads to false positives almost by design.  Third, tying has led to one of the most ridiculous antitrust remedies of all time—namely the  European Commission’s insistence that Microsoft expend effort creating and offering a product–a version of Windows that didn’t include Microsoft’s media player technology—that no one wants. Now, I understand that others will have their own candidates. But to beat mine your challenge is you must show a complete lack of theoretical or empirical support; a really bad legal test; and a remedy that better demonstrates the bankruptcy of the law.   The challenge is on.

5 responses to Section 2 Symposium: David Evans on "Tying as Antitrust's Greatest Intellectual Embarrassment"


    I’ll admit to not having read Einer’s provocative paper, which I’ll try to remedy. But I argued at the Spring Meeting a couple of years ago that the only rational way to look at tying is through the Section 2 lens — and then, not as some separate species of “anticompetitive” conduct, but rather as one would look at any other conduct at Section 2. Tying shares with RPM that it’s really unilateral conduct shoehorned into a Section 1 analytical framework, and, as a result, it generates an unfortunate mix of increased transactions costs for businesses trying to comply with incoherent legal rules, and litigation results that likely bear little relationship to the actual competitive merits of the practice.

    Josh Wright 8 May 2009 at 6:02 pm

    I agree with David that the greatest intellectual embarrassment of antitrust LAW is Jefferson Parish.

    From an earlier thread on TOTM on “the worst antitrust decision that remains good law,” I wrote:

    Robert Bork has noted: “the connoisseur of bad antitrust opinions must take into account Fortner I, Utah Pie, Sealy, Schwinn, Procter & Gamble, Von’s Grocery, and many others.” There are also the exclusive dealing cases finding liability with tiny foreclosure rates. Not to mention Jefferson Parish, which I’ve argued (along with others) is ripe for reversal… .

    My best guess is that the Jefferson Parish decision is probably the worst remaining culprit on a consumer welfare basis as it interferes with all sorts of efficiency enhancing marketing and distribution efforts. The practice is prevalent in the modern economy, and there is an economic consensus (see also here) on the view that tying is generally pro-competitive and that a per se approach (even a “modified” one) is inappropriate.

    However, if we broaden “antitrust law” to cover possible Congressional action, a per se rule against minimum RPM as contemplated by the pending legislation would give Jefferson Parish a run for its money, and in my view, immediately leap to the most prominent position in antitrust’s hall of shame.

    Michael Salinger 8 May 2009 at 1:04 pm

    Apparently, Chairman Leibowitz and Commissioners Harbour and Rosch do not agree. They would have the world believe that the DOJ Report was a right-wing screed to gut Section 2 and that they were merely providing a more balanced view. I would cite their section on tying along with their section on predatory pricing as evidence that they are the radicals in this debate.

    Howard Marvel makes a similar point about exclusive dealing as the point David and I have made about tying: both practices are common in competitive markets. The observation is important, but it is not sufficient to establish that we need to be more concerned with false positives than false negatives. So what distinguishes tying from exclusivity?

    I would argue that there are four reasons why different standards should apply to them. The most important is administrability. Almost every product could theoretically be broken into more than one distinct product. Without articulating a limiting principle, something that no one has done with tying, it is hard to know what is allowed and what is not. Second, and related, there is common and then there is common. Exclusivity is common, but it is not as common as tying, which implicates virtually everyting sold. Third, the efficiencies of tying are more subtle than have generally been acknowledged. I am confident that the most common example used to illustrate the efficiency of tying is the sale of shoes in pairs. I have no doubt that selling a pair of shoes saves packaging costs, but that only explains why the bundle exists. Tying is not just offering the bundle. It is doing so and not offering one of the components. One needs to explain the efficiency of not offering the individual shoes to the small number of people who want just one. This comes down to measuring a cost of a more complex product offering, a phenomenon that is real but often hard to measure. The agency issues that lie at the heart of the efficiency from exclusivity might also be hard to measure, but I suspect they are better understood than the efficiencies of tying. The fourth is a point raised by David’s post. The theory of anticompetitive harm from exclusivity is pretty straightforward. It entails paying a trading partner not to deal with a rival. Economists have had to strain to articulate a theory of anticompetitive harm from tying. The models exist, but all I conclude from them is that anticompetitive tying is not a theoretical impossibility.

    antitrust guy 8 May 2009 at 9:03 am

    I think you need to say that it’s the “per se rule against tying”, not tying in itself. Give it rule of reason treatment and let Prof. Elhauge bring the next tying case.

    Thom Lambert 7 May 2009 at 1:07 pm

    As for current embarrassments, it may be impossible to top the per se (or quasi per se) rule against tying. If I’m allowed to mention past embarrassments, I’d point to the per se rule against vertical maximum resale price maintenance (Albrecht v. Herald).

    One other current embarrassment, at least as implemented, is Brown Shoe’s instruction that markets may be defined on the basis of “practical indicia.” This doesn’t look so silly on its face, but it has led to some embarrassing conclusions, such as the FTC’s conclusion that “premium natural and organic supermarkets” are a separate market from conventional grocery stores that sell the same products. To reach that conclusion, the FTC, enabled by Brown Shoe’s practical indicia comments, went on a fishing expedition through Whole Foods’ and Wild Oats’ business documents, looking for statements highlighting the chains’ uniqueness. Surprise, surprise — it found some!