Archives For

On September 15, Tesla won a big victory in Massachusetts. As we have previously chronicled at length on TOTM ( see, e.g., here, here, herehere, here and here), the car dealers are waging a state-by-state ground war to prevent Tesla from bypassing them and distributing directly to consumers. The dealers invoke 1950s-era franchise protection laws that are obsolete given the radical changes in automotive market in the intervening years and, in any event, have nothing to do with a company like Tesla that doesn’t use dealers at all. In Massachusetts State Automobile Dealers Ass’n, Inc. v. Tesla Motors MA, Inc., -2014 WL 4494167, the Supreme Judicial Court held that the dealers lacked standing to challenge Tesla’s direct distribution since the Massachusetts statute was intended to protect dealers from oppression by franchising manufacturers, not from competition by manufacturers who didn’t franchise at all. As we have previously detailed, there is no legitimate pro-consumer reason for prohibiting direct distribution.

What I found most telling about the Court’s decision was its quotation of a passage from the dealers’ brief. As readers may recall, the dealers have previously asserted that prohibiting direct distribution is necessary to break up the manufacturer’s “retail monopoly,” create price competition, and ensure that consumers get lower prices — arguments that are facially ludicrous as a matter of economics. But now listen to what the dealers have to say in Massachusetts:

Unless the defendants are enjoined, they will be allowed to compete unfairly with the dealers as their model of manufacturer owned dealerships with remote service centers will allow Tesla and Tesla MA financial savings which would not be available to Massachusetts dealers who must spend considerably to conform to Massachusetts law. This could cause inequitable pricing which also [could] cause consumer confusion and the inability to fairly consider the various automobiles offered.

Translation: Direct distribution leads to cost savings that are reflected in lower (“inequitable!”) prices to consumers.

Surely right, since a Justice Department study found that direct distribution could save over $2,200 per vehicle. But coming from the car dealers?  Who would have thunk it?

As Geoff posted yesterday, a group of 72 distinguished economists and law professors from across the political spectrum released a letter to Chris Christie pointing out the absurdities of New Jersey’s direct distribution ban. I’m heartened that both Governor Christie and his potential rival for the 2016 Republican nomination, Texas Governor Rick Perry, have made statements, here and here, in recent days suggesting that they would support legislation to allow direct distribution. Another potential 2016 Republican contender, has also joined the anti-protectionist fray. This should not be a partisan political issue. Hopefully, thinking people from both parties will realize that these laws help no one but the car dealers.

In the midst of these encouraging developments, I came across a March 5, 2014 letter from General Motors to Ohio Governor John Kasich complaining about proposed legislation that would carve out a special direct-dealing exemption for Tesla in Ohio. I’ve gotta say that I’m sympathetic to GM’s plight. It isn’t fair that Tesla would get a special exemption from regulations applicable to other car dealers. I’m not blaming Tesla, since I assume and hope that Tesla’s legislative strategy is to ask that these laws be repealed or that Tesla be exempted, not that the laws should continue to apply to other manufacturers. But the point of our letter is that no manufacturer should be subject to these restrictions. Tesla may have special reasons to prefer direct distribution, but the laws should be general—and generally permissive of direct distribution. The last thing we need is for a continuation of the dealers’ crony capitalism through a system of selective exemptions from protectionist statutes.

What was most telling about GM’s letter was its straightforward admission that allowing Tesla to engage in direct distribution would give Tesla a “distinct competitive advantage” and would create a “significant disparate impact” on competition in the auto industry. That’s just another way of saying that direct distribution is more efficient. If Tesla will gain a competitive advantage by bypassing dealers, shouldn’t we want all car companies to have that same advantage?

To be clear, there are circumstances were exempting just select companies from a regulatory scheme would give them a competitive advantage not based on superior efficiency in a social-welfare enhancing sense. For example, if the general pollution control regulations are optimally set, then exempting some firms will allow them to externalize costs and thereby obtain a competitive advantage, reducing net social welfare. But that would only be the case if the regulated activity is socially harmful, which direct distribution is not, as our open letter explained. The take-away from GM’s letter should be even more impetus for repealing the direct distribution bans across the board so that consumers can enjoy the benefit of competition among rival manufacturers who all have the right to choose the most efficient means of distribution for them.

Last summer I blogged here at TOTM about the protectionist statutes designed to preempt direct distribution of Tesla cars that are proliferating around the country. This week, New Jersey’s Motor Vehicle Commission voted to add New Jersey to the list of states bowing to the politically powerful car dealers’ lobby.

Yesterday, I was on Bloomberg’s Market Makers show with Jim Appleton, the president of the New Jersey Coalition of Automotive Retailers. (The clip is here). Mr. Appleton advanced several “very interesting” arguments against direct distribution of cars, including that we already regulate everything else from securities sales to dogs and cats, so why not regulate car sales as well. The more we regulate, the more we should regulate. Good point. I’m stumped. But moving on, Mr. Appleton also argued that this particular regulation is necessary for actual reasons, and he gave two.

First, he argued that Tesla has a monopoly and that the direct distribution prohibition would create price competition. But, of course, Tesla does not have anything like a monopoly. A point that Mr. Appleton repeated three times over the course of our five minutes yesterday was that Tesla’s market share in New Jersey is 0.1%. Sorry, not a monopoly.

Mr. Appleton then insisted that the relevant “monopoly” is over the Tesla brand. This argument misunderstands basic economics. Every seller has a “monopoly” in its own brand to the same extent as Mr. Appleton has a “monopoly” in the tie he wore yesterday. No one but Tesla controls the Tesla brand, and no one but Mr. Appleton controls his tie. But, as economists have understood for a very long time, it would be absurd to equate monopoly power in an economic sense with the exclusive legal right to control something. Otherwise, every man, woman, child, dog, and cat is a monopolist over a whole bunch of things. The word monopoly can only make sense as capturing the absence of rivalry between sellers of different brands. A seller can have monopoly power in its brand, but only if there are not other brands that are reasonable substitutes. And, of course, there are many reasonable substitutes for Teslas.

Nor will forcing Tesla to sell through dealers create “price competition” for Teslas to the benefit of consumers. As I explained in my post last summer, Tesla maximizes its profits by minimizing its cost of distribution. If dealers can perform that function more efficiently than Tesla, Tesla has every incentive to distribute through dealers. The one thing Tesla cannot do is increase its profits by charging more for the retail distribution function than dealers would charge. Whatever the explanation for Tesla’s decision to distribute directly may be, it has nothing to do with charging consumers a monopoly price for the distribution of Teslas.

Mr. Appleton’s second argument was that the dealer protection laws are necessary for consumer safety. He then pointed to the news that GM might have prevented accidents taking 12 lives if it had recalled some of its vehicles earlier than it eventually did. But of course all of this occurred while GM was distributing through franchised dealers. To take Mr. Appleton’s logic, I should have been arguing that distribution through franchised dealers kills people.

Mr. Appleton then offered a concrete argument on car safety. He said that, to manufacturers, product recalls are a cost whereas, to dealers, they are an opportunity to earn income. But that argument is also facially absurd. Dealers don’t make the decision to issue safety recalls. Those decisions come from the manufacturer and the National Highway Traffic Safety Administration. Dealers benefit only incidentally.

The direct distribution laws have nothing to do with enhancing price competition or car safety. They are protectionism for dealers, pure and simple. At a time when Chris Christie is trying to regain credibility with New Jersey voters in general, and New Jersey motorists in particular, this development is a real shame.

Dan Crane is Sr. Professor of Law and Associate Dean for Faculty and Research at the University of Michigan Law School

I’m delighted that Josh and Maureen have launched a concerted effort to have the FTC articulate clear principles for Section 5 enforcement.  My own views on the proper scope of Section 5 are articulated in my book The Institutional Structure of Antitrust EnforcementI won’t attempt a comprehensive regurgitation here, but just offer three quick observations that may be relevant to the present debate.

First, the most important reason for the articulation of clear Section 5 principles is not to give greater guidance to the business community, although that’s important too.  The most important reason is to articulate principles of self-restraint that Article III courts can invoke in reviewing Commission decisions applying Section 5 in spaces that Sections 1 and 2 of the Sherman Act would not apply under current judicial doctrine.  History suggests that courts jealously guard their interpretations of the Sherman Act and are reluctant to allow the FTC to effectively override them based on assertions of Section 5 independence.  The courts rejected FTC efforts to wield an independent Section 5 in the late 70s and early 80s.  They will be inclined to do so again if the FTC merely asserts “Section 5 is a prophylactic statute; trust us to wield it to good ends.”  By articulating principles that delimit how far the FTC can go under Section 5, the FTC would provide courts assurances that meaningful judicial review can still occur.

Second, and in the same vein, the Commission needs to articulate principles not just about how far it can go under Section 5 but also about how far it cannot go.  It needs to say, in effect, “courts, here is how you will know if we crossed the line.”  These limitation principles need to be concrete enough that defendants have a reasonable opportunity to show through objective evidence that their conduct does not contravene the statute.  In other words, the Commission needs to explain how its view of Section 5 independence is not a plea for greater administrative discretion, which courts will be unlikely to afford, but for an expanded scope of antitrust coverage under principles that can be fairly contested in litigation.

Shifting to the substance of these limitation principles, my third point concerns one of the criteria proposed by Josh—that the challenged conduct have no cognizable efficiency benefits.  I agree with the thrust of Josh’s suggestion, but would suggest a small qualification.   There is virtually no anticompetitive conduct that doesn’t produce some efficiency.  Heck, even the proverbial blowing up of the competitor’s factory might product some efficiency (the rebuilt factory might be 3% more efficient than the old one and hence might spur greater competition in the long run).  Cartels often have some efficiency benefit—they reduce planning costs, smooth prices to customers, etc.  So if the criterion were that the challenged conduct had to be absolutely devoid of efficiency benefits, that might create a null set of independent Section 5 cases.  I would suggest, as a qualification, that the criterion be akin to that used to justify the per se rule—that the challenged conduct is so unlikely to have redeeming efficiencies that the law is justified in not inquiring into whether there are in fact efficiencies.  This is not to say that Section 5 cases would disallow inquiry into efficiencies, but rather that the Commission would need to show that any claimed efficiencies were so trivial or speculative compared to the clear competitive harms that the conduct was similar in kind to price fixing, market division, or bid rigging.  The paradigmatic Section 5 cases—invitations to collude and fraud—easily fit that bill.

In Continental T.V. v. GTE Sylvania (1977), Justice Powell observed that antitrust law should go easy on manufacturer restraints on dealer resale because manufacturers could always decide to integrate forward into distribution and bypass dealers altogether.  As anyone who has followed electric car manufacturer Tesla’s recent travails will know, Justice Powell’s observation is not true of the auto industry.  Dealer franchise laws in most states require car manufacturers to sell through independent dealers.  Tesla apparently would like to bypass the traditional dealership model and sell directly to customers, which is landing the company in legal hot waters in many states, including traditionally “free market” states like Texas, Virginia, and North Carolina.

Tesla is the offspring of the South African-American entrepreneur Elon Musk, who also brought us Pay-Pal and SpaceX.  The company’s luxury electric cars have caused a sensation in the auto industry, including a review by Consumer Reports calling Tesla’s Model S the best car it ever tested.  Extraordinarily for a startup, in the first quarter of 2013, Tesla Model S sales exceeded the top line offerings of the established German luxury brands, Mercedes, BMW, and Audi.  Indeed, more Teslas were sold than BMW 7 series and Audi A8s combined.

One would imagine that Tesla’s biggest problem would be economic and technological—creating the infrastructure for battery-pack swap and charging facilities necessary to persuade customers that powering their Teslas will be as seamless as pumping gas at a filling station.  (Telsa’s recently announced 90-second battery pack swap will go a long way in that direction).  Alas, Tesla’s major stumbling block seems to be more legal than technological.  Tesla wants to open its own showrooms and sell directly to customers.  The powerful car dealers’ lobby has been invoking decades-old dealer franchising laws to block Tesla’s progress, insisting that Tesla must sell through independent, franchised dealers like other car companies do.  Tesla has been lobbying for legislative reforms at the state level, thus far with mixed success.

The basic economics of the problem are straightforward.  As Ronald Coase taught us, whether a car manufacturer keeps the distribution function in house or buys distribution services on the market is a question of the agency and transactions costs of those respective forms of distribution.  There are many reasons why manufacturers might prefer to distribute through independent dealers.  This shifts the investment in distribution to someone other than the manufacturer, allowing the manufacturer to focus on its core competence in research and development and manufacturing.  It shifts managerial decisions to managers with local market knowledge.  It may create economies of scale or scope as dealers sell several different brands under a single roof.

But there are also good reasons why a manufacturer might prefer to sell directly to consumers.  The manufacturer may be concerned that the dealers will focus more on short-term sales maximization rather than long-term investment in building the brand.  (This could be particularly concerning to a company like Tesla that is introducing a disruptive new technology that still needs to be proven in the market).  The manufacturer may worry that independent dealers will be insufficiently loyal and push other brands.  It may fret that local dealers will be unsophisticated about new technologies and that training and monitoring will be easier if retail distribution stays in house.

There is no a priori reason to favor one model or the other, and I have no idea whether Tesla is better off distributing through traditional dealer networks.  But I find it hard to fathom any good reason why the law should prohibit a car manufacturer from picking whatever distribution model it thinks best.  As a newcomer to these state dealer laws, I’ve been trying to keep an open mind that they might be supported by some legitimate policy concern and not pure protectionism.  Unfortunately, whenever a dealer-aligned speaker opens his mouth to defend these laws, the case that it’s just protectionism gets stronger.

One argument I’ve seen attributed to the auto dealers—and I sincerely hope that there’s some mistake and this is not actually an argument they’re making—is that creating “competition” in retail distribution of Tesla cars is necessary to prevent Tesla from price gouging customers. The idea that a vertically integrated manufacturer has a “monopoly” over the brand’s retail distribution that needs to be broken up by outsourcing the retail function to independent dealers is farcical.  If Tesla has market power, it will extract the full monopoly profit regardless of whether it sells to dealers or end users.  (It will be fully embedded in either the wholesale or resale price).  Since retail distribution is just a cost of doing business, Tesla will increase its monopoly profits by minimizing the cost of retail distribution since then it will sell more cars.  If anything, as economists have long recognized, outsourcing the retail distribution function to locally dominant automobile dealers could lead to double marginalization and increased prices.

A second argument is that having local dealers is necessary to ensure that customers are adequately served.  For example, Bob Glaser of the North Carolina Automobile Dealer’s Association has asserted that the restrictions are a form of “consumer protection,” since “a dealer who has invested a significant amount of capital in a community is more committed to taking care of that area’s customers.”  The obvious rejoinder is that Tesla has as much or more of an interest as the dealers in seeing that customers get the level of service they’re willing to pay for.   If Tesla gets a bad reputation for quality, it will fail.  I suppose that one might worry if  Tesla were a fly-by-night operation selling customers an expensive durable good at a high price and then fleeing with its profits and leaving customers without support.  But that’s obviously unlikely of a company that’s pouring billions of dollars into the creation of a new product and a recharging and battery swapping infrastructure.  Car manufacturers make considerably larger fixed capital investments than do dealers and I’m sure that the dealer failure and exit rate is considerably higher than that of manufacturers.

A related argument is that dealers play an important role in complying with local laws regarding titling and safety inspection.  But this argument doesn’t work either.  First, observe that at present most states only prohibit manufacturers from opening their own dealerships—they don’t prohibit online sales from outside the state.  (North Carolina recently passed a statute banning online sales as well).  There’s no reason why a manufacturer-owned dealership should be less capable of complying with local laws than an independent dealer.  Second, why should Internet sales involve evasion of state titling and safety inspection laws?  Internet sales can just as easily be subject to the same titling and inspection requirements as dealer-initiated sales.

Another argument I’ve heard is that prohibiting manufacturers from integrating forward into distribution is necessary to prevent them from competing unfairly with their own franchised dealers by undercutting them on price.  The logic of this argument is a little fuzzy. What would a manufacturer set up franchised dealers only to undercut them ruinously?  I suppose it might be some variant of the usual free-rider arguments—the manufacturer would set up independent dealers, free-ride on their local brand promotion, and then cut them out once the brand was established.  (Why the dealers can’t contractually bargain for protections from this isn’t clear).  But all of this is a lark for present purposes.  It clearly doesn’t apply to Tesla, which wants to avoid franchising altogether.  At most, if one were worried about “undercutting,” the rule should be a prohibition on manufacturer retail operations for manufacturers that also franchise, not for those that bypass franchising altogether.

Some people have quite fairly complained that Tesla shouldn’t be given special exemptions when other car companies are bound by the dealer restrictions.  For sure, but that cuts in favor of amending or repealing these laws altogether, not enforcing them against Tesla.  If anything, if it’s true that Tesla would obtain a competitive advantage by bypassing traditional dealer networks, consumers should want this advantage available to all car companies.  To put it other way, this argument is basically an admission that the dealer laws are raising car prices.

The last argument I’ve heard—and it’s a real doozy—is that independent dealers are civic bastions of local communities and therefore deserve to be specially protected.  Never mind the fact that many auto dealerships are owned and operated by large regional chains rather than local Boy Scouts troopmasters.  Why on God’s green earth should we single out automobiles for economic protectionism in order to subsidize local civic participation?  Why stop with automobiles?  Why not household appliances, twinkies, and lingerie?   And who is to say that Tesla will be any less civic minded than franchised auto dealers?  Further, if the model of direct distribution is so superior to franchised distribution that eliminating legal restrictions would put the dealers out of business, there must be something systematically inefficient about franchised distribution.  In that case, both consumers and local communities would be better served if state legislatures just levied a tax on auto sales and distributed them pro rata to local civic organizations.

Since the arguments for dealer laws are so weak, I’m left with the firm impression that this is just special interest rent-seeking of the worst kind.  It’s a real shame that Tesla—seemingly one of the most innovative, successful, and environmentally correct American industrial firms of the last decade—is going to have to spend tens of millions of dollars and may eventually have to cut shady political deals for the right to sell its own products.  I’m ordinarily a fan of federalism and states’ rights, but if the current debacle continues, it may be necessary for Congress to step in with preemptive federal legislation.

Dan’s final post responding to Steve’s latest postOther posts in the series: DanSteveDanSteveDan, and Thom.

It seems that it’s time to wind down and that a further tit-for-tat might not be productive, so I’ll close with a final comment on the first point that Steve makes—one that may undergird much of our disagreement.  Steve asserts that “the $71 payment would fail a test of comparing the rival’s price and cost, but that it is not the test.  The test compares the monopolist’s price and cost.”  That would only be true if we were applying an unmodified predatory pricing rule to loyalty rebates—a position that I’ve never advocated in these posts are elsewhere.  If we applied the attribution test that Steve and I have been assuming, the question would be whether the rival could profitably remain in the market given the price it would have to charge to neutralize the effect of the monopolist’s rebate operating at the contestable share and scale.  And, since Steve and I have now agreed that using the rival’s rather than the monopolist’s costs is admissible if we don’t insist on an EEC component, then my statement that the $71 could fail the price-cost screen is accurate.  But if the effective price the rival would have to meet were $69 and not $71, there wouldn’t be any foreclosure—which is why the screen makes sense.

Thanks, Steve, for this impromptu exchange.  I hope that both our fair points and grievous errors have been educational to ourselves and to others.

Dan’s next installment, responding to Steve’s latest post responding to Dan’s latest post on the appropriate liability rule for loyalty discounts. Other posts in the series: SteveDan, and Thom.

I’m happy to keep going back in forth with Steve until we wear out our welcome at TOTM, or simply wear out. [Keep 'em coming! - ed.]

(1) There’s neither input foreclose nor output foreclosure if a rival can neutralize a loyalty discount without pricing unprofitably.  The rival who can profitably compete despite the loyalty discount structure will remain in the market; there is no input exclusion.  As I understand him, Steve posits that an RRC effect would arise if the guilty firm tried to use loyalty rebates to raise the effective price the rival had to pay for “distribution services” with the effect of raising general market prices.  But that theory doesn’t work.  First, why would a firm that can profitably match a loyalty discount have to increase its price in order to “pay for” the loyalty discount?  Second, the theory requires the victimized rival to simultaneously increase its “payments” to distributors in the form of discounts even while it is increasing its prices to those same distributors in equal amount in order to offset the payments.   A rival who feels “pressured” to make a $100 loyalty payment to distributors and consequently raises his price to the distributors by $100 to offset has done . . . nothing.

(2) Steve posits “that economic analysis is the same” for loyalty discounts and contractual commitments not to buy from rivals.  Really?  I suppose at a high level of generality one would say that in both cases the question is whether the price or contract forecloses competitors, but it can’t possibly be that an exclusive dealing contract that prohibits purchases from a rival has the same effect as a 0.0001% rebate if the customer purchases exclusively from the seller.  Once an exclusive dealing contract is in place, the rival can’t compete without inducing breach of contract.  With a pure loyalty discount, the rival can always compete so long as it can profitably neutralize the loyalty discount with its own pricing terms.  You see this in the real world—customers operate in environments where multiple sellers are offering loyalty discounts.  The customers freely switch between suppliers (sometimes showing loyalty and obtaining the discounts; sometimes deciding to forgo the discounts and prefer variety).  This is not true of a market locked up with exclusive dealing agreements.

(3) Steve presents four examples of how an auction for contractual exclusivity results in a single firm obtaining contractual exclusivity with anticompetitive effects due to asymmetry of incentives, all of which begs the question of how loyalty discounts without contractual exclusivity achieve the same effect.  Moreover, all Steve shows is that if you applied unmodified predatory pricing analysis to the exclusivity auction, the contracts wouldn’t be illegal since the winning bidder could show that its revenues exceeded its costs.  He doesn’t show how, in the absence of contractual exclusivity, the winning bidder would satisfy the attribution test we’ve been assuming.  For example, in Scenario (a), if there isn’t contractual exclusivity but merely a discount structure that would require the entrant to pay rebates of $71 even while earning revenues of $70, then the rebate system would fail the price-cost screen and we could proceed to all of the foreclosure/RRC analysis that Steve has so magnificently pioneered and illustrated.  The same is true in each of Steve’s examples.  None of his examples shows a circumstance where the new entrant could obtain retailer distribution services without inducing breach of contract and without pricing below its cost and yet there would be an anticompetitive effect from a loyalty discount scheme.

(4) In his first post, Steve said that “the price-cost test is premised on the erroneous idea that only equally efficient competitors are worth protecting.”  My last post disputed that claim.  Steve now says that “the price-cost test is better framed as a measure of ‘profit-sacrifice,’ and EEC is simply a misleading way to express the test.”  So I think we’re now in agreement that one could accept a version of the price-cost screen even while believing that antitrust law should not necessarily be bounded by an EEC component.  As to profit sacrifice, more in a moment.

(5) I don’t want to get pulled into defending an EEC component in this discussion, since I think we’re now in agreement that it’s not necessary to support some version of the price-cost screen.  In any event, when I said that the test has “merit,” I wasn’t saying that it should be applied categorically to all types of exclusion claims.  For example, (and I’m not committing to this), institutional context might matter to whether an EEC test should apply.  One might reasonably believe that it’s dangerous to allow less efficient rivals to seek treble damages for their exclusion from the market for all kinds of incentives and decisional reasons but that the government should be free to mount exclusion theories that do not depend on a showing that EECs were excluded.  Again, I’m not arguing that this should be the case, only explaining the non-committal nature of my “merit” comment.

(6) (I’ll collapse my response to Steve’s points 6-8 into one).  The discount penalty theory is premised on the assumption that the monopolist increases its list price above the profit-maximizing monopoly level and then offers a concession back down to the monopoly level, combined with an onerous term restricting the customer’s freedom of choice among suppliers.  I take it that Steve accepts the point from my last post that a price of x plus loyalty restriction is effectively higher than a price of x without a loyalty restriction.  Steve says that a monopolist might choose to exceed the profit-maximizing price and thereby forego profits as a predatory investment.  In that case, the loyalty discount program must be of short duration and the monopolist has to be highly confident of recoupment, just as in conventional predation cases.  Observe that the profit sacrifice is not merely the charging of the $105 “penalty” price but also the charging of the $100 plus loyalty restriction price.  Thus, the monopolist is not saved from profit sacrifice, as Steve posits, merely by threatening a price of $105 that it never has to charge.  Also, observe that the strategy is much more risky and expensive to the monopolist than to the customer.  By definition, the monopoly price the customer is paying is one where the customer will be willing to substitute to other products at just a slightly higher price, whereas the monopolist stands to lose highly profitable sales by exceeding its monopoly profit-maximizing price.  Finally, let me underline as I did in my last post that I’m not claiming that “penalty pricing” is economically impossible.  Rather, I’m making the point that “penalty pricing” is an expensive and risky strategy for monopolists, that we should therefore not assume that it will be routinely deployed, and that the starting presumption should be that most loyalty discounts are true reductions from the but-for price.

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.