Archives For

Guest post by Dan Crane, responding to Steve’s post responding to Dan’s earlier post and Thom’s post on the appropriate liability rule for loyalty discounts.

Something that Thom and I both said in our earlier posts needs to be repeated at the outset:  I don’t know of anyone who disagrees with Steve and Josh that raising rivals’ costs (“RRC”) and economic analysis drawn from exclusive dealing law belong in an analysis of loyalty discounts.  There’s also no claim on the table that a loyalty discount that fails the “contestable share”/discount attribution test that Steve mentions should be treated anything like presumptively illegal.  The current debate is solely about whether there should be a price-cost screen in loyalty discount cases.  We aren’t even talking about what the measure of cost should be or how that screen should work (although, with Steve, I’m happy to assume marginal or average variable cost and the aforementioned contestable share/discount attribution approach for the sake of argument).  Josh and Steve are well justified in pointing out how aspects of RRC theory can apply in loyalty discount cases—but that doesn’t meet the objection that a screen should also apply.

It’s also important to recognize that the argument in favor of a price-cost screen for loyalty rebates does not need to entail a general argument in favor of a “profit sacrifice” theory for all monopolization offenses.  What we’re talking about here is unilaterally determined discounts to customers—something that is presumptively procompetitive, although potentially exclusionary under some circumstances.  Such discounts could be harmful if they resulted in customer foreclosure, but they would not result in customer foreclosure if the rival could profitably match the loyalty discount.  That is the point of the price-cost screen.  You might wonder why a rival would ever complain about a loyalty discount if they could profitably match it.  The reasons are many.  The rival might be losing sales because customers don’t like its product.  It might have failed for reasons completely apart from the accused firm’s loyalty discounts. It might be attempting to use antitrust law to thwart price competition, as a large body of literature suggests.  (See work by Will Baumol and Janusz Ordover, Preston McAfee and Nicholas Vakkur, and Edward Snyder and Tom Kauper, among others).

One thing I didn’t just mention—although it could often be true—is that the complaining rival isn’t an equally efficient competitor (“EEC”).  Steve is wrong to suggest that the price-cost test depends on adopting an EEC theory.  Although there is much merit to the EEC test (heck, even the Europeans have adopted it), one could formulate a version of the price-cost screen that simply requires the rival to show that the discount foreclosed a hypothetically equally efficient competitor or even this particular rival given its actual costs, as some have suggested.  The current argument is not over the formulation of the test, but whether we should dispense with a price-cost screen altogether in loyalty discount cases.

In any event, observe that the entire structure of modern predatory pricing law is premised on an EEC assumption.  If an incumbent firm with marginal costs of $50 and a current price of $100 faces entry by a new rival with marginal costs of $75 and drops its price to $74 in order to exclude the new rival, it enjoys categorical immunity under a long line of Supreme Court cases.  In another forum, Steve suggested that the difference in those cases is that the customer is getting the benefit of a lower price, so the law is hesitant to condemn the price as predatory.  But that exposes something problematic about Steve’s starting premise—he assumes that it’s uncertain whether loyalty discounts generally lower prices.  Prima facie, that seems wrong.  Customers routinely offer to trade loyalty for lower prices precisely because the prices are . . . lower.

Steve suggests that maybe loyalty discounts aren’t really discounts at all.  Maybe the seller, who was previously charging a price of $100, raises the price to $105 and then gives a discount back down to $100 in exchange for customer loyalty.  Steve notes that Thom and I didn’t consider this scenario.  That’s because Josh didn’t raise it in his speech.  It would have been very surprising if Josh had raised it in his speech, since Josh and I co-authored a paper several years ago debunking this same theory in the bundled discount context.  I discuss the “disloyalty penalty” theory at length in a forthcoming article in the Texas Law Review, really just extending the work that Josh and I started several years ago.

There are many problems with this “disloyalty penalty” theory, including the empirical one that it doesn’t fit the pattern of almost any of the recent loyalty discount cases.  But there is also a problem of basic economics.  Unless it is engaging in limit pricing, the accused firm’s $100 price is a monopoly (or market power) profit-maximizing price.  By definition, any price increase will be unprofitable to the seller.  Obviously, the $105 price would be unprofitable.  But it’s also true that a price of $100 coupled with a new obligation to buy a certain percentage of requirements from the seller to achieve that price is unprofitable because it exceeds the profit-maximizing price.  The addition of a contractual term that restricts the buyer’s freedom is economically equivalent to a price increase if the buyer valued the prior freedom from the restriction (if the buyer didn’t value the prior freedom from the restriction there’s no effective price increase but also no anticompetitive effect, since the buyer wouldn’t have bought from the rival anyway).  Hence a price of $100 with loyalty term is effectively higher than a price of $100 without a loyalty term that restricts the buyer’s purchasing freedom.  By adding a loyalty term to obtain the $100 price, the seller exceeds its profit-maximizing monopoly price.

My claim is not that “penalty pricing” for disloyalty is impossible, but that the presumption should be that loyalty discounts are true discounts off the but-for price.  Loyalty discounts belong squarely in the “hospitability” tradition for unilaterally determined pricing structures—all those judicial decisions that talk about how important it is not to chill vigorous price competition.

Steve argues that loyalty discounts may “tie up customers” before competitors arrive on the scene.  I’m not sure what Steve means by “tie up customers.”  Suppose that a monopolist, knowing that rivals are about to enter the market, goes to all of its customers and offers  them a 5% discount if they will agree to purchase 95% of their requirements from the monopolist for the next three years.  At that point we have a partial exclusive dealing contract and the cost-price screen shouldn’t be required.  But, there, the exclusionary mechanism—the thing that keeps rivals from competing—is not the loyalty discount but rather the contractual commitment not to buy any more than 5% of requirements from rivals.  Customers would have to breach their contract in order to consider even the most advantageous offers from rivals.  The point that amici made in our Meritor v. Eaton brief was that when the claimed mechanism of exclusion is a price term and not a contractual restriction on purchasing from rivals, some version of the price-cost screen should apply.

The example I’ve just attributed to Steve (and sorry Steve if this is not what you have in mind) is not what we’re talking about in almost any of the current generation of loyalty discount cases.  In Meritor, for example, the Third Circuit acknowledged that the loyalty provisions at issue did not require customers to buy any of their requirements from Eaton.  It’s just that if the customers didn’t meet the loyalty thresholds, they would lose a possible rebate.  Meritor could compete for that business by offering its own counter-rebates so long as it wouldn’t have had to price unprofitably to do so.

Steve’s point about economies of scale is one that I covered in my post and is fully accounted for by the cost-price screen.  A rival who can profitably match a loyalty discount scheme is not foreclosed from operating at any particular scale.

The same is true of Steve’s point about loyalty discount schemes foreclosing a new seller’s ability to make incremental sales that don’t reduce the accused firm’s own sales.  Again, so long the rival can profitably match the discounts, there is no reason that output should be reduced.

Finally, Steve asserts that loyalty discounts obtained by intermediaries may not be passed onto ultimate consumers.  That’s equally true of conventional single-firm price reductions that are categorically immunized from antitrust liability under a long line of precedent.  One may not like the price-cost test in any context for that reason or others, but there’s nothing special about its application to loyalty discounts. The common denominator of all of these points is that loyalty discounts aren’t exclusionary unless they force rivals to price below cost in order to match the customer’s loss of the loyalty discounts if they fail to meet the loyalty threshold.

Steve thinks the price-cost screen exhibits “formalism”—that dreaded epithet in the post-realist world—but it’s actually just an expression of economic common sense.  Steve and Josh are excellent economists and it’s hard for me to imagine a case in which they would condemn a loyalty discount if there was undisputed evidence that the allegedly excluded rival could have completely neutralized the financial inducement of the loyalty discount by offering a counter-discount of its own without pricing below cost.  If they can offer an example of a circumstance where such a loyalty discount should be condemned, I would be very interested to hear it.  If they can’t, then they have implicitly adopted a version of the price-cost screen and, to repeat a point from my earlier post, all we’re haggling over is the price.

Guest post by Michigan Law’s Dan Crane. (See also Thom’s post taking issue with FTC Commissioner Josh Wright’s recent remarks on the appropriate liability rule for loyalty discounts).

A number of people on both sides of the ideological spectrum were surprised by FTC Commissioner Josh Wright’s recent speech advocating that the FTC reject the use of price-cost tests to assess the legality of loyalty discounts and instead pursue an exclusive dealing framework of analysis.  As the author of a brief (unsuccessfully) urging the Supreme Court to grant certiorari and reverse in Z.F. Meritor v. Eaton, I want respectfully to disagree with some of what Josh had to say.  But, first, two other observations.

First, I’m delighted that Josh is charting a course as Commissioner that defies some people’s expectations (even if they sometimes happen to be my own!).  Josh has long insisted on evidence-based analysis rather than simplistic theorizing or reductionist legal rules and his position on loyalty discounts is consistent with that theme.  Early in his term on the Commission, Josh is making it clear that he will exercise independent judgment, intellectual integrity, and a principled, non-ideological approach to decision-making.  That’s a nice rejoinder to those who believe that antitrust law reduces to simplistic right-left politics.  So kudos to Josh!

Second, Josh and I probably agree on 90% of what’s important about loyalty discounts.  We agree that loyalty discounts are usually competitively benign or procompetitive, but that they can sometimes be anticompetitive when they exclude rivals and create market power.  We also agree that exclusive dealing principles and analysis can be usefully deployed in loyalty discount cases (although I would only do so after a plaintiff satisfied a price-cost screen).  Finally, we also agree that unmodified predatory pricing rules—requiring the plaintiff to show that the defendant’s sales were below average variable cost—could potentially insulate some exclusionary loyalty discounts from antitrust scrutiny.

Where we differ is on the question of whether antitrust law should ever condemn a loyalty discount that the allegedly excluded rival could have met without pricing below cost.  To say that it should not is to say that there should be some sort of price-cost screen in place in loyalty discount cases.  Josh rejects the use of such screens.

One point of clarification:  Josh asserts that one of the central claims in favor of the price-cost test is its ease of administration.  Contrary to Josh’s suggestion, that is not an argument we made in our Meritor amicus brief.  As someone who has counseled clients and litigated these issues, I can attest that the discount attribution test (the variant of the price-cost test I support for loyalty discounts) is anything but easy to apply (which Josh himself recognizes with respect to the “contestable share” idea).  The virtue of the test is not its ease of administration, but that it requires plaintiffs to show that the discount scheme actually foreclosed them from competing.  Our point was about analytical discipline, not ease of administration.

This, I think, is the crucial difference between Josh and me.  Unless a rival would have to price below cost to match a loyalty discount, it is not foreclosed from competing for the business covered by the discount.  Josh wants to apply exclusive dealing analysis that looks at foreclosure without answering a question that, in my view, is necessary to discover whether there is any foreclosure at all—whether the rival could profitably match the discount.  A rival that has a profitable “predatory counterstrategy,” to quote Frank Easterbrook, isn’t foreclosed.

A thought experiment may be helpful.  Suppose that a firm with a 90% market share offered all of its customers a 0.0001% rebate if they purchased at least 80% of their requirements from the dominant firm.  No one could imagine that such a “loyalty discount” could exclude rivals, since even small rivals could easily make up the rebates foregone if customers forewent buying the 80% from the dominant firm.  We can make the rebate 0.001% with the same result.  And we can continue to pose successive iterations of the same question, increasing the discount incrementally, until we hit a point that someone could reasonably say “well now that could be exclusionary.”  Wherever we cross that Rubicon, we cross it because what was true at 0.0001%—that the small rival could laugh it off by shelling out a few dollars in a counter-discount—is no longer true.  To play this game is to conduct a competitive response sensitivity analysis of the very kind demanded by the attribution test. For present purposes, it’s unimportant where we draw the line; it’s the fact of the line-drawing that matters. To paraphrase Winston Churchill, we’ve already established what we are, now we’re just haggling over the price.

Josh is surely right that loyalty discounts can raise rivals’ costs.  That could happen in one of two ways.  First, if a small firm were prevented from reaching efficient scale, or second if a firm were forced to ramp up to an inefficiently large scale in order to meet a competitor’s loyalty discounts.  But neither of those scenarios holds if the rival is able to compete against the loyalty discounts without pricing below cost.  The small firm will not be prevented from reaching minimum efficient scale if it can increase its share by profitably competing against the loyalty discount.  And the second firm will not be rushed into increasing its scale if it can compete profitably at a smaller scale.  In either case, the RRC mechanism is forcing the firms to price below their costs.

At the end of the day, I suspect that Josh—using whatever analytical tools he associates with exclusive dealing analysis—would be highly unlikely to condemn any loyalty discount in a case where the rival could profitably match the discounts.  That gives me assurances as to Josh, but not as to all other players in the legal system, many of whom are eager to jettison the discipline of price-cost screens so that they can get onto the “real meat” of the case—like inflammatory internal e-mails employing metaphors of coercion that Judge Posner has aptly labeled “compelling evidence of predatory intent to the naïve.”  So I remain highly confident that we’re in good hands with Josh, but worry about what others may do with his words.

I just finished watching the FTC webcast announcing the Intel settlement and did a quick read over the agreement itself. Some quick high-level reactions:

  1. The tone of the press conference was triumphant, of course. Leibowitz claimed that the FTC got 22 out of 26 of the remedies proposed in the complaint and that Intel, which had previously criticized the proposed remedies as unprecedented, was suddenly making the remedies “precedented.” Further study required here, but it’s far too glib to count victory based on 22 out of 26. Many of the proposed remedies contained suggestive, open-ended language which, if interpreted reasonably expansively, would have gone far beyond this settlement.
  2. To my ear, there was a big change in emphasis from the theory of the complaint. The complaint was predominantly about Intel’s exclusivity and rebating practices with customer with some deception theories thrown in to make it sound like a proper FTC case. The settlement is much more about intellectual property restrictions that prevent AMD and Via from outsourcing manufacturing when they become capacity constrained.
  3. Section 5 of the FTC Act: Leibowitz made a special point of reiterating his view that Section 5 is “a penumbra around the Sherman Act.” I happen to agree with that view, but it’s an open question whether this settlement really advances this view. It’s notable that the FTC has brought several Section 5 cases in the last few years and hasn’t chosen to litigate any of them all the way. Not saying it’s a bad decision, just pointing out that the status of Section 5 remains open after this settlement.
  4. Predatory design: This is an aspect of the settlement that I really can’t stomach. It makes me nervous to think that the FTC is going to have an open-ended right to decide that Intel’s design changes are predatory because they do not provide “any actual benefit” to the product. Benefits, like beauty, are often in the eye of the beholder.

More to follow.

On the campaign trail, Barack Obama made an issue of the ostensibly lax state of antitrust enforcement during the Bush administration. Christine Varney’s first public act as head of the Antitrust Division was to withdraw the Bush Antitrust Division’s unilateral
monopolization report and announce that trustbusting against dominant firms was back on the agenda.  Expectations were high for an antitrust revival.  The antitrust world waited for the first shoe to drop in Washington.

So far, the shoe hasn’t dropped.  For all of the sound and fury, things have been relatively tame at the Antitrust Division.  At least from the outsider perspective, things look like business as usual: anti-cartel enforcement, consent decrees on controversial mergers.

Things have been a little different at the FTC, with the filing of the Intel case.  But it’s hard to put that one on the Obama administration, since the complaint was authorized by three commissioners appointed in previous administrations.  And, in any case, the FTC is supposed to be a politically independent commission (thanks to Humphrey’s Executor).

Last November, the St. Petersburg Times interviewed me for their service, which follows up on politicians’ campaign promises.  They wanted to know whether Obama had lived up to his campaign promise to reinvigorate antitrust enforcement.  At that time, I was cautious, telling them: “We haven’t seen the filing of any big cases, but that could be because it takes time to develop those cases, and they’re still working on them.”

I’m now ready to abandon my earlier caution and declare that there is some apparent dissonance between Obama’s campaign rhetoric and the reality on the ground at DOJ.  Mind you, I’m not criticizing for now: just observing.

Assuming I’m right about the dissonance, what’s the story?  My leading contender — one that is the subject of a forthcoming essay in the Antitrust Law Journal — is that antitrust enforcement is almost always put on the back burner during major economic crises.  Although I have little inside information about decision-making inside the beltway, I have some reason to believe that the White House has put the brakes on the Antitrust Division because of the economic circumstances.  (If any readers of this blog have contrary knowledge, I would be happy to stand corrected).

It also may be that I’m jumping the gun and the DOJ has a slew of big cases waiting in the wings.  They are certainly conducting some big investigations that could lead to Microsoftesque cases.  One of the downsides of blogging about something not happening is that it could happen the next day.  (One of the nice things about a blog is that your prior statement can then be updated the next day).

But this last possibility raises a different and quite subtle issue: how do we measure the real effectiveness or vigor of antitrust enforcement, when its chief function is deterrence?  In a world of perfect deterrence, there would be no antitrust cases filed.  This is a paradox of law enforcement.  The Bush Administration collected record fines in cartel cases, but that might be nothing more than evidence of an explosion of cartel behavior because of a perception that enforcement would be lax.  Perhaps the current mildness out of the Antitrust Division is merely the consequence of the President and AAG having threatened large companies with severe sanctions for misbehaving with the consequence that the ostensible offenders retreated from their worst practices to wait out the current administration.  If so, the dissonance between rhetoric and practice is merely the sound of deterrent success.

I don’t pretend to have firm answers to the questions I posed.  These sorts of questions are best analyzed with the benefit of hindsight after the close of the relevant events rather than midstream.  So let’s see how the story develops.

Whoa There, Big Fellows!

Dan Crane —  21 April 2010

The DoJ/FTC revised merger guidelines, released as a draft for public comment yesterday, have me scratching my head. I need to spend more time with them before I come to any strong views, but the obvious issue-spotter is the elimination of market definition as a necessary step in the analysis.

So we all know that market definition is sometimes or often a formalistic and technical step that ends up bearing little relation to the analysis that follows. I’m happy to be persuaded that we should abandon it as a necessary prerequisite in all kinds of antitrust cases.

Problem is, the official dogma from the courts is still that defining a relevant market is a “necessary prerequisite” to finding liability under the Clayton Act. See Judge Walker in Oracle, for example, quoting the Supreme Court in duPont. So if I were grading the draft Guidelines as an exam, I would put a big X through the statement on the FTC website that “market definition is not an end itself or a necessary starting point of merger analysis, but instead a tool that is useful to the extent it illuminates the merger?s likely competitive effects.”

Of course, only a couple of contestable merger cases get tried in the courts a year and the second request threat gives the agencies effective power to stop mergers that they might not be able to block in court, so maybe we shouldn’t worry too much if the Guidelines don’t track the law. But do we want the agencies to lose the habit of thinking in relevant market terms? As long as the law requires it, I should think that the agencies should make themselves go through the exercise. Otherwise, they are going to develop intellectual laziness on market definition over time. Or, they will develop cases for internal assessment based on other theories and then have to scramble to invent relevant market theories of the case if it heads to litigation.

Maybe the agencies think that a guidelines revision that downplays market definition will lead the courts to downplay it as well. That strikes me as possible, but a gamble. More likely, the revised Guidelines will result in more Oracle-PeopleSofts, where the agencies lose the case on market definition. But, with spring in the air here in Michigan, I’m feeling optimistic and susceptible to persuasion.

Someone persuade me I’m wrong.

Let Sleeping Dogs…

Dan Crane —  27 October 2009

I feel no great urgency to revise the Guidelines.  True enough, they’re more of an analytical thought experiment than an accurate description of how merger review takes place in the agencies, but who’s really fooled?  Perhaps some business people think that the Guidelines are a computer program waiting for the introduction of the relevant data to spit out the answer, but most sophisticated executives contemplating a merger will understand that the Guidelines are just a beginning point for conversation.

Could the beginning point be clearer or conform more closely to agency practice?  Sure, but that doesn’t mean that revision of the Guidelines is justified.  With hindsight, the First Amendment could be a little better worded, but no one wants to tinker with it now–who knows what would result?  I’m sufficiently satisfied with current merger practice in the agencies that I don’t care that much what’s in the Guidelines and I am worried about the unpredictable results that could obtain if we started tinkering.  Let sleeping dogs . . .

But if we are going to revise, then my pet issue is the asymmetrical treatment of the probabilities on anticompetitive effects and offsetting efficiencies–a point on which Joe Simons and I are planning a fuller analysis.  The Guidelines seem to suggest that if the probability of anticompetitive effects of magnitude 100 is 50% and the probability of offsetting efficiencies of magnitude 100 is 50%, then the merger should be challenged, since a greater quantum of proof is required for efficiencies than for anticompetitive effects.  This makes no sense to me–everything else being equal, efficiencies that would be passed on to consumers and market power increases should be given equal weight and not assigned separate probability standards.

craneDaniel Crane is a Professor of Law at Cardozo Law School (soon to be at University of Michigan Law School).

Bundled discounts have been one of the hottest monopolization topics of the last decade. Much of the trouble began with the Third Circuit’s en banc decision in LePage’s v. 3M, which reversed an earlier 2-1 panel decision which in turn had overturned a plaintiff’s jury verdict largely based on 3M’s bundled discounts. After the Solicitor General’s amicus curiae brief asked the Supreme Court to deny cert on the grounds that there wasn’t sufficient scholarship on bundled discounts, there was a flurry of legal and economic scholarship, the overwhelming majority of which was highly critical of LePage’s.

Over the past five years or so, it seemed that a consensus was emerging that some sort of discount reallocation or attribution test should be used as a screen in bundled discounting cases. There are various formulations of the test, but in general it requires the plaintiff to show that defendant priced the competitive product below cost after the discounts on the non-competitive product are reallocated to the competitive market. Versions of that test have been adopted by a variety of commentators, agencies, and courts, including the DOJ in its Section 2 report, the Antitrust Modernization Commission, the Areeda-Hovenkamp treatise, and the Ninth Circuit’s PeaceHealth decision. I have been—and continue to be—a staunch defender of some formulation of that test.

Just when I thought we were close to reaching a strong majority position on bundled discounts, along comes a significant new article by Einer Elhauge (to be published this coming December in the Harvard Law Review) challenging the entire basis of the theory. Einer argues that bundled discounts manifest anticompetitive “power effects” if the unbundled price for the linking product exceeds the but-for price level (i.e., the price the defendant would charge in the absence of the bundle) and that such bundles should be treated as tie-ins.

Einer’s article is sure to attract lots of attention and give courts and perhaps the agencies pause in adopting the until-now consensus position on bundled discounts. Although I profoundly disagree with much of Einer’s analysis, it is a provocative and important article. Josh Wright and I are planning a full response at a later date. For the moment, let me just preview one responsive angle. Continue Reading…

Daniel Crane is a Professor of Law at Cardozo Law School (soon to be at University of Michigan Law School).

I must confess that my basic reaction to the Section 2 report was disappointment.  It’s not that I find much fault with the report itself–a few quibbles yes, but generally I find it quite satisfactory–but that after all of the time and effort put into the joint hearings by the FTC, the FTC wasn’t able to join the report.  Moreover, the shrill dissenting statement by three commissioners will probably prevent the report from playing influencing judicial decisions or legislation.

That’s a real pity.  Many staff members at both agencies and many outsiders (myself included) put a lot of work into the hearings and the report.  Contrary to the suggestion in the dissenting statement, I saw no evidence that the hearings were stacked against more interventionist perspectives on Section 2.  That certainly was not the case at the bundled discount hearings at which I testified.  To the contrary, my general impression was that the hearings were constructive and made substantial progress toward agreement on some basic principles.  Certainly, they lacked the rancor that characterized the release of the report.

Continue Reading…