Archives For

In yesterday’s Super Tuesday primaries, Hillary Clinton won the two largest contests–California and New York–but the delegate count was close to even (perhaps Clinton even finished slightly behind) because Barack Obama won more states, albeit smaller ones.

The Clinton campaign argues that Clinton’s victories in larger, delegate-rich states suggest that she would be a more viable candidate than Obama in the general election. But does that conclusion really follow?

I don’t see why. States that are heavily Democratic, whether large or small, are very likely to vote for whoever is running on the Democratic ticket in November. The candidate who gives Democrats the best hopes of winning in the general election is the one who will do better among moderates, independents, and Republicans. So the fact that Clinton won in heavily Democratic states such as New York and California does not seem to be a very meaningful statistic for assessing whether she or Obama would do better in the general election. To the contrary, the candidates’ relative performance in more politically balanced or conservative states (regardless of size) would seem like a better indicator of general electability, especially in such states where independents and Republicans are permitted to participate in the Democratic primary. Judging by that measure, Obama is the candidate who seems more likely to do better in the general election.

To be sure, the campaign is far from over and much can still happen to influence voter opinion in the remaining primary contests and in the general election. My goal here is simply to debunk some spin by the Clinton campaign that seems to be based on an erroneous statistical inference.

A few days ago, I posted a comment about Starbucks’ recent disclosure that its average per store traffic has gone down slightly even though overall profits have gone up. I suggested a number of explanations for these phenomena consistent with a story that consumer taste for the Starbucks product has not diminished. One of these explanations was that Starbucks makes its product available to non-Starbucks retailers and that consumers may be turning increasingly to these other retailers. That is, consumers could be shifting away from Starbucks stores rather than away from the Starbucks product. In passing, I observed that Starbucks’ distribution of its product to non-Starbucks stores is proconsumer and belies the accusations of some opportunistic competitors who claim that Starbucks’ business model is anticompetitive.

Yesterday, Stephen Bainbridge helpfully pointed out that Starbucks owns its stores and does not franchise them (correcting my reference to Starbucks “franchisees”). At the same time, Stephen also notes that Starbucks does license holders of otherwise inaccessible real estate to sell Starbucks products, an arrangement that is economically indistinguishable from franchising. I appreciate this clarification from Stephen. But I also should say that the analysis and conclusions in my original post do not turn on the distinction between Starbucks owning or franchising its stores. The important point is that the Starbucks product is available at other places besides Starbucks stores, such as at Barnes & Noble. Whether these alternatives are created by franchise or license is immaterial. Stephen does not dispute my conclusion that the antitrust allegations against Starbucks are without merit. And though he presents the point as “a problem with [my] analysis,” Stephen’s correction is, as Josh explains, really factual rather than analytical.

Leaving aside the issue of why traffic is down at Starbucks stores, Stephen raises (or rather recycles from a superb earlier posting he wrote back in 2003) the important question of why some firms like Starbucks choose to own their own stores (i.e., to vertically integrate) while other firms such as Subway choose to franchise. Many great minds have weighed in on this question over the years, beginning (at least implicitly) with Ronald Coase’s great 1937 essay “The Nature of the Firm” (which uses variation in transaction costs to explain a firm’s choice between internal and external contracting) and continuing most recently with Josh’s and Larry Ribstein’s excellent postings in response to Stephen.

I wish to add here only one point to the analysis. Whatever its economic benefits, vertical integration has the added virtue that it can reduce a firm’s exposure to certain types of antitrust claims. While a franchisor and franchisee can be sued under Section 1 of the Sherman Act for anticompetitive vertical restraints (such as vertical price fixing or other forms of restricted distribution such as territorial restraints), a single firm or a parent and its wholly owned subsidiaries are immune from such suits under the Supreme Court’s Copperweld decision, which held that a firm cannot conspire with itself for purposes of satisfying the conspiracy element of Section 1. At the same time, even while reducing exposure under Section 1, vertical integration by a firm with an already high market share can increase the firm’s exposure to a monopolization (or attempted monopolization) claim under Section 2.

The antitrust factor may well be one explanatory variable in Starbucks’ decision to own its stores rather than franchise them and Subway’s decision to franchise rather than own. Starbucks has a high and growing market share in the market for retail coffee, particularly if one defines the product market narrowly to include only “high end” coffee. It is plausible that a court could find it to have some degree of “market power” (though not “monopoly power”) in this market, which makes it vulnerable to Section 1 claims with respect to its arms’ length, vertical arrangements. Owning its own stores reduces that exposure, and since the stores have been owned from the start rather than acquired later there isn’t an apparent act of monopolization that would increase the firm’s exposure under Section 2. Subway, by contrast, seems much less dominant in its product space and indeed may not even have “market power” at all. If that is right, then Subway would benefit from vertical integration much less than Starbucks does–at least from an antitrust perspective.

Needless to say, other factors may well be at work besides the antitrust factor, as Josh, Larry, and Stephen ably suggest. My point here is only to add antitrust as another variable that may well explain firm franchising behavior in some range of cases.

Traffic at Starbucks shops open for 13 months or more is down one percent. Does this mean that the public is finally losing its appetite for Starbucks? Not necessarily.

While traffic is down, profits are up. Thus a more likely explanation for the new data is the firm’s price increase last summer rather than a change in consumer tastes. Another possibility is that even if demand *per store* is down, overall demand could still be constant or even up, given that Starbucks is always opening new stores whose sales to some extent dilute the revenues of existing stores. A third factor that may explain things, and this is the one that I want to focus on because it is of interest to antitrust lawyers, is that Starbucks stores do not enjoy exclusivity in the sale of Starbucks coffee.

You can go into a Barnes & Noble and other non-Starbucks stores and find a cafe selling “Starbucks coffee” even though the seller isn’t a Starbucks. The coffee itself and the surrounding atmosphere at such cafes (complete with their wireless hotspots to go with the lowfat lattes) are in many cases close to perfect substitutes for Starbucks stores. Sales at such places, if they are increasing at an ever faster rate, may well be diluting sales at Starbucks stores and hence may well account for the new data.

Starbucks’ willingness to sell its product widely rather than reserve it exclusively for its full-service franchisees suggests to me that the firm is competing aggressively in an “output enhancing,” pro-consumer way, rather than seeking to find ways to reduce output and raise price, as some opportunistic antitrust plaintiffs have erroneously alleged.

All therefore seems well for the Starbucks franchise–even if, as Jackie Mason has quipped, it is a bit much to ask customers both to clean up after themselves and also to leave a tip!


Keith Sharfman —  14 November 2007

Yet another major airline merger appears to be in the works: United and Delta. This calls for some antitrust analysis. A few months ago, Thom did a thorough job analyzing the antitrust aspects of AirTran’s proposed takeover of Midwest. The key point in Thom’s analysis was that assessment of an airline merger’s economic effects properly centers not on the merging parties’ overall market shares but rather on the extent to which the two firms compete head-to-head.

United and Delta are large carriers, the second and third largest in the industry. If one uses overall industry market shares to calculate HHI in the merger analysis, the transaction would seem presumptively unlawful. But if one looks at the actual routes on which the two airlines compete and the level of competition currently present on those routes from other carriers, the picture may look very different. If it is the case that the two firms now compete head-to-head only (or largely) on routes that are are served by a large number of carriers, then the firms’ high overall market shares may not matter very much.

That said, a note of caution. In a major airline asset acquisition some years ago, American/TWA, the firms argued that the transaction should be permitted on the ground that TWA (then in bankruptcy) was a “failing firm” and that therefore the transaction’s effect on HHI was not dispositive. The enforcement authorities (wrongly) bought into this argument and permitted the transaction, even though TWA’s airplanes would not have “left the industry” (the relevant standard under a failing firm theory) if they had been sold to the second highest bidder rather than to American. Commercial airplanes are a long term, durable capital good that can’t easily be converted into other uses. Sure TWA’s creditors wanted to maximize the value of TWA’s assets. But that’s not a reason to relax the requirements of antitrust law any more than it would be to permit a bankruptcy debtor to violate the Clean Water Act.

As with TWA, neither United’s nor Delta’s planes will disappear from the market if the deal is blocked, nothwithstanding the firms’ recent bankruptcies and the financial woes that chronically plague the industry. The United/Delta deal should be assessed solely on the basis of its competitive effects. The failing firm argument has no place here, and the parties should not assume that the enforcement authorities will treat them as generously as they treated American and TWA.

is here, over at eCCP, and differs somewhat from Thom’s.

The takeway excerpt is:

Credit Suisse has important implications for antitrust practice. The decision’s effect is to narrow the scope of antitrust law and to invite efforts by regulated industries to narrow it still further. The court’s “clearly incompatible” standard is new and (though it purports not to) seems to water down considerably the old “plain repugnancy” test of Gordon v. New York Stock Exchange, Inc. 422 U.S. 659, 682 (1975). Under the new incompatibility standard, there no longer has to be an actual conflict between antitrust and other federal law for antitrust implicitly not to apply. Even a mere regulatory overlap may now be sufficient to trigger antitrust immunity. (Recall that in Credit Suisse the Court assumed that both antitrust and the SEC disapproved of the tying and other practices in question, and yet the Court still considered the two bodies of law incompatible on account of the regulatory overlap.) ….

Going forward, the Court will need to tighten the rule in Credit Suisse if it wants antitrust to continue to operate as Congress intended it to in conjunction with the compartmentalized maze of federal regulatory law. No one thinks that securities firms should be exempt from the legal obligations that generally flow from non-securities law (antitrust aside). If we expect to hold securities and other regulated firms accountable for torts and breaches of contract, or for crimes and discrimination, then why not also hold them accountable for antitrust violations? If Congress says otherwise, that is one thing. But if Congress is silent on the question, a federal agency should not have have any more power than a state to confer antitrust immunity upon those that it regulates. Of states we require a clearly articulated policy that presents an actual conflict, not merely the possibility of future potential incompatibility. From federal agencies we should not expect any less.

Just yesterday, in its historic decision in Leegin, the Court strongly reaffirmed its confidence in the Rule of Reason’s workability by overturning Dr. Miles and extending the rule’s reach to vertical RPM. That workability should make us equally confident that antitrust can peacefully coexist with the reguatory state.

Josh, Thom, and I all predicted correctly that Dr. Miles would be overruled. We even predicted the vote count! I had hoped that Justice Breyer would join the majority, but instead he joined with Justices Ginsberg, Souter, and Stevens (as predicted) in dissent.

The opinion is here.

A few more thoughts to supplement Josh’s fine posting on the transcript of oral argument in Leegin.

I don’t understand Justice Breyer. He recognizes that there are at least some circumstances in which RPM helps consumers. Why isn’t that enough for Dr. Miles to be overruled?

Justice Breyer regards this as a “close case” (presumably for reasons of stare decisis rather than on the merits) and asks “what has changed?”

What has changed is our state of economic understanding. When Dr. Miles was decided in 1911, the proconsumer aspects of RPM were not yet recognized. And that was largely true even in 1966, a year that Justice Breyer has focused upon, given the publication that year of an academic book criticizing RPM, which seems (so far as Justice Breyer can tell) to have made all of the same arguments against RPM that are now being made today. But it was not and could not have been argued back in 1966 that it would be inconsistent to exempt some vertical restraints from the per se rule but not others. Only once the late 1970s came, when Continental overruled Schwinn, did it become clear how foolish Dr. Miles is. Now that we allow (subject to the rule of reason) non-price vertical restraints, it seems entirely crazy not to allow RPM as well. As then-Professor Posner argued back in the 1970s, RPM is likely in many cases to be a more efficient vertical restraint than the now permissible non-price ones. So banning RPM across the board on a per se basis does not seem to make any sense. And while many of the same arguments against RPM could have been made in 1966, there are some new ones now that didn’t exist then. Anyway, Dr. Miles was decided in 1911, not 1966. It’s not as though the Court considered overruling Dr. Miles in 1966 and decided not to. The Court is now squarely addressing whether Dr. Miles should be overruled for the first time. So any post-1911 developments in economic science are fair game; there’s no basis for excluding from debate what happens to have been known in 1966!

Justice Breyer is impressed by Professor Scherer’s examples of cases where RPM has been harmful to consumers. But how can these examples justify a per se rule? If Professor Scherer can persuasively demonstrate the benefits of banning RPM in the market, say, for blue jeans, then RPM should be banned in that context under the rule of reason. But this does not come close to showing that RPM is”always or almost always” harmful to consumers such that a per se rule is justified.

But enough about Justice Breyer. I also don’t understand Justice Stevens. What on earth does a horizontal conspiracy among New York distributors have to do with this case? As Justice Scalia said, what possible procompetitive benefit could be associated with a horizontal agreement on price? It’s very hard to think of one. But if Justice Stevens can think of one, then sure: apply the rule of reason in that context too!

Predictions are hard to make. But I don’t think Shubha Ghosh is right about Dr. Miles definitely have four votes in the bag. I say it’s one, not four. Justice Breyer says it’s a close case and may well come around. Justice Souter seems altogether uncertain. Justice Ginsberg if anything seems to be leaning the other way–given her interest about what arguments would be available to the plaintiff on remand assuming (as she appears to think likely) a loss in the Supreme Court. As I read the transcript, Dr. Miles can be confident only about Justice Stevens’ vote.

Today’s report that Sirius and XM plan to merge vindicates the antitrust analysis offered here last June.

Regulators should analyze the merger from a broad “audio market” perspective that includes terrestrial radio. Considering the extensive non-satellite content available to listeners, and considering as well the efficiencies associated with the Sirius/XM combination, it is reasonable to conclude that consumers will benefit from the deal and that regulators should therefore allow it (after careful review, to be sure).

The question I’m left with is: how long will it take for GM/Ford to follow?

Steamy Espresso

Keith Sharfman —  22 January 2007

A few weeks ago, I suggested that Belvi’s antitrust suit against Starbucks is weak and ought to be dismissed.

This report in today’s Seattle Times further strengthens the case for dismissal. Competition in the Seattle market for espresso is apparently more intense than Belvi’s complaint would have us believe!

Michael Abramowicz over at Concurring Opinions has an interesting post about the ongoing litigation between economists John Lott and Steven Levitt. Lott’s suit alleges that Levitt defamed him in his recent book Freakonomics by suggesting that Lott’s research on the relation between guns and crime could not be “replicated” by other scholars and in a subsequent email to an economist suggesting that Lott had paid $15,000 to the Journal of Law & Economics to publish in a special issue a series of articles supporting Lott’s views on guns. This week, a federal district court in Chicago granted Levitt’s motion to dismiss the claim concerning the statement in Freakonomics but denied his motion to dismiss the claim concerning the email.

Abramowicz suggests in his post that Lott’s “potential damages are almost certainly low” and that this case “though not technically frivolous” is “of a type that our legal system does not handle well” and “a vexatious use of the legal system, because the cost of bringing the claim seems much larger than any plausible reputational damage to Lott.”

Leaving the merits of the dispute aside, my question is this: if the cost of bringing the claim is really much larger than any damages that Lott may recover, then why is Lott pursuing the case? Isn’t Lott’s pursuit of the case strong evidence that he believes he could recover more than his costs?

Moreover, if indeed the claim is worth less than the cost of litigating it, why is Levitt vigorously defending the suit? Why have he and HarperCollins (his publisher) spent so much money disputing liability (e.g., by filing the motion to dismiss) rather than simply relaxing, knowing that damages won’t be very high? Isn’t it just as “vexatious” to dispute a vexatious claim as it is to assert one?

The answer, I think, is that defamation suits implicate subjective nonpecuniary interests that are difficult for courts to value. The formal legal remedies available in such cases are thus usually undercompensatory and pale in comparison to the reputational effects of winning or losing. While the financial stakes may be low, more is at stake than simply the money. The case is about reputation, not money. Hence the current legal quagmire. Even a generous financial settlement is therefore not likely to satisfy Mr. Lott, and by the same token an admission of having made a false statement is not something that Mr. Levitt would likely consider offering.

Here’s my suggestion for the most efficient way to end the dispute: in exchange for Lott’s agreement to dismiss the suit with prejudice, Levitt could agree to issue a statement not admitting to having defamed Lott but rather simply saying that he respects John Lott’s intellect and his work as an economist even though he remains skeptical about Lott’s work on guns and crime.

Hopefully a settlement along these lines is in the works, especially now that HarperCollins is out of the case and Levitt will have to start paying his lawyers out of his own pocket. But then again, settlement of the case would deprive bloggers of an interesting topic about which to comment!

Domain Name Hijacking

Keith Sharfman —  6 November 2006

Dan Solove over at Concurring Opinions reports on an insidious practice that unfortunately has become increasingly common: domain name hijacking.

Here’s how it works. The original owner of a popular website fails to renew its domain name prior to the expiration of the owner’s entitlement. An opportunistic “hijacker” then purchases the name and offers to sell it back to the original owner for a tidy sum. The original owner is then left with an unhappy choice: pay the hijacker off, or set up shop under a new domain name with the loss of traffic that such a switch inevitably entails.

The latest victim of such a hijacking scheme is Crescat Sententia, a popular blog that used to be located at but now has been forced to move to

Dan suggests that domain name hijacking of this sort may well be characterized as copyright infringement. But because the case for copyright protection isn’t clear cut, he wonders if there are other legal protections too.

Here’s my suggestion for another theory of liability: intentional interference with prospective economic advantage. The elements of that tort–(1) an economic relationship between the plaintiff and some third person containing the probability of future economic benefit to the plaintiff; (2) knowledge by the defendant of the existence of the relationship, (3) intentional acts on the part of the defendant designed to disrupt the relationship, (4) actual disruption of the relationship; and (5) damages to the plaintiff proximately caused by the acts of the defendant–all seem to be present here. The original website has an economic relationship with its existing readers or patrons; the hijacker knows about this relationship; the hijacker intentionally acts to disrupt the relationship by acquiring the domain name; the loss of the domain name actually disrupts the relationship by shutting down the old site without indicating where a new site, if any, is located; and the original owner is thereby damaged.

As matter of policy and economics, there isn’t any positive social value associated with domain name hijacking. Indeed, once transaction and switching costs are considered, the conduct actually entails social losses. One would therefore hope (or perhaps a la Posner even dare to predict) that the common law would forbid and deter such conduct. Applying the tort of intentional interference with prospective economic advantage would do just that. And so even if the copyright case against domain name hijacking isn’t airtight, the common law should come to the rescue.

Hijackers beware!

Espresso Exclusivity?

Keith Sharfman —  25 September 2006

Belvi Coffee and Tea Exchange cannot be serious. The firm is suing Starbucks for exclusive dealing in the Seattle and Bellevue, Washington real estate markets.

The suit alleges that Starbucks has leased real estate at above-market prices in exchange for commitments by the landlords to exclude other coffee shops from the building.

Let’s take Belvi’s allegations at face value and assume that Starbucks has a 73% share of the U.S. coffee shop industry, even though such a narrow product market definition seems implausible, given that Starbucks and other coffee shops compete for customers with many other types of shops, lounges, and restaurants. People buy coffee (and many of the other products that Starbucks sells) in all sorts of places besides coffee shops. Dunkin’ Donuts sells coffee and cake. So does McDonald’s. So does pretty much every diner and restaurant in the country.

But as I say, let’s take Belvi’s allegation of a 73% market share at face value. So what? The issue in this case is not whether Starbucks has monopoly or market power in the coffee shop market. That’s not in the least bit relevant. What matters is whether Starbucks has market power in real estate. And there’s not even an allegation of that here, much less any evidence.

Nothing stops Belvi from opening up as many shops as it wants to in any neighborhood where Starbucks is located, and if Starbucks charges too much people could always swing on over to Belvi. Surely it isn’t necessary for Belvi to be located in the very same building as Starbucks in order for Belvi to compete. (Has there ever been an antitrust case involving a retail industry in which the relevant geographic market is defined as a single building? I’m not aware of any.) Suppose that instead of leasing Starbucks owns the building. Would antitrust law require Starbucks to lease space to its competitor? That doesn’t seem very likely. A building isn’t an “essential facility” like a railroad track whose owner may be compelled to deal with a competitor. If outright ownership would entitle Starbucks to refuse to deal, why should an exclusive lease be treated any differently? It’s hard to see what would make an exclusive lease different from an outright sale.

Note the plausible procompetitive justification for the exclusivity that Starbucks obtains through these leases. Starbucks probably does lots of market research when deciding where to locate its stores. Why must Starbucks allow Belvi to free ride on that research?

If Starbucks had gotten an entire neighborhood to agree not to lease to other coffee shops, I could see Belvi’s point. But so long as Starbucks lacks the power and anyway has done nothing to prevent Belvi from locating next door, the case seems ludicrous and ought to be dismissed.