Archives For truth on the market

In Collins Inkjet Corp. v. Eastman Kodak Co. (2015) (subsequently settled, leading to a withdrawal of Kodak’s petition for certiorari), the Sixth Circuit elected to apply the Cascade Health Solutions v. PeaceHealth “bundled discount attribution price-cost” methodology in upholding a preliminary injunction against Kodak’s policy of discounting the price of refurbished Kodak printheads to customers who purchased ink from Kodak, rather than from Collins.  This case illustrates the incoherence and economic irrationality of current tying doctrine, and the need for Supreme Court guidance – hopefully sooner rather than later.

The key factual and legal findings in this case, set forth by the Sixth Circuit, were as follows:

Collins is Kodak’s competitor for selling ink for Versamark printers manufactured by Kodak. Users of Versamark printers must periodically replace a printer component called a printhead; Kodak is the only provider of replacement “refurbished printheads” for such printers. In July 2013, Kodak adopted a pricing policy that raised the cost of replacing Versamark printheads, but only for customers not purchasing Kodak ink. Collins filed suit, arguing that this amounts to a tying arrangement prohibited under § 1 of the Sherman Act, 15 U.S.C. § 1, because it is designed to monopolize the Versamark ink market. Collins sought a preliminary injunction barring Kodak from charging Collins’ customers a higher price for refurbished printheads. The district court issued the preliminary injunction, finding a strong likelihood that Kodak’s pricing policy was a non-explicit tie that coerced Versamark owners into buying Kodak ink and that Kodak possessed sufficient market power in the market for refurbished printheads to make the tie effective.

On appeal, Kodak challenges both the legal standard the district court applied to find whether customers were coerced into using Kodak ink and the district court’s preliminary factual findings. In evaluating the likelihood of success on the merits, the district court applied a standard that unduly favored Collins to determine whether customers were coerced into buying Kodak ink. The court examined whether the policy made it likely that all or almost all customers would switch to Kodak ink, but did not examine whether this would be the result of unreasonable conduct on Kodak’s part. A tying arrangement enforced entirely through differential pricing of the tying product contravenes the Sherman Act only if the pricing policy is economically equivalent to selling the tied product below cost. The record makes it difficult to determine conclusively Kodak’s ink production costs, but the available evidence suggests that Kodak was worse off when customers bought both products, meaning that it was in effect selling ink at a loss. Thus, Collins was likely to succeed on the merits even under the correct standard.  Furthermore, the district court was correct in its consideration of the other factors for a preliminary injunction. Accordingly, the preliminary injunction was not an abuse of discretion.

The Sixth Circuit’s Collins Inkjet opinion nicely illustrates the current unsatisfactory state of tying law from an economic perspective.  Unlike in various other areas of antitrust law, such as vertical restraints, exclusionary conduct, and enforcement, the Supreme Court has failed to apply a law and economics standard to tying.  It came close on two occasions, with four Justices supporting a rule of reason standard for tying in Jefferson Parish, and with a Supreme Court majority acknowledging that “[m]any tying arrangements . . . are fully consistent with a free, competitive market” in Independent Ink (which held that it should not be presumed that a patented tying product conveyed market power).  Nevertheless, despite the broad scholarly recognition that tying may generate major economic efficiencies (even when the tying product conveys substantial market power), tying still remains subject to a peculiar rule of limited per se illegality, which is triggered when:  (1) two separate products or services are involved; (2) the sale or agreement to sell one is conditioned on the purchase of the other; (3) the seller has sufficient economic power in the market for the tying product to enable it to restrain trade in the market for the tied product; and (4) a “not insubstantial amount” of interstate commerce in the tied product is affected.  Unfortunately, it is quite possible for plaintiffs to shoehorn much welfare-enhancing conduct into this multipart test, creating a welfare-inimical disincentive for efficiency-seeking businesses to engage in such conduct.  (The U.S. Court of Appeals for the D.C. Circuit refused to apply the per se rule to platform software in United States v. Microsoft, but other appellate courts have not been similarly inclined to flout Supreme Court precedent.)

Courts that are concerned with the efficient application of antitrust may nonetheless evade the confines of the per se rule in appropriate instances, by applying economic reasoning to the factual context presented and finding particular test conditions not met.  The Sixth Circuit’s Collins Inkjet opinion, unfortunately, failed to do so.  It is seriously problematic, in at least four respects.

First, the Sixth Circuit’s opinion agreed with the district court that “coercive” behavior created an “implicit tie,” despite the absence of formal contractual provisions that explicitly tied Kodak’s ink to sale of its refurbished printheads.

Second, it ignored potential vigorous and beneficial ex ante competition among competing producers of printers to acquire customers, which would have negated a finding of significant economic power in the printer market and thereby precluded per se condemnation.

Third, it incorrectly applied the PeaceHealth standard to the facts at hand due to faulty economic reasoning.  For a finding of anticompetitive (“exclusionary”) bundled discounting, PeaceHealth requires that, after all discounts are applied to the “competitive” product, “the resulting price of the competitive product or products is below the defendant’s incremental cost to produce them”.  In Collins Inkjet, all that was known was that Kodak “stood to make more money if customers bought ink from Collins and paid Kodak’s unmatched printhead refurbishment price than if they bought Kodak ink and paid the matched printhead refurbishment price.”  Absent additional information, however, this merely supported a finding that Kodak’s tied ink was priced below its average total cost, not below its (far lower) incremental cost.  (Applying PeaceHealth, the Collins Inkjet court attributed the printhead discount entirely to Kodak’s ink, the tied product.)  In short, absent this error in reasoning (ironically, the court justified its flawed cost analysis “as a matter of formal logic”), the Sixth Circuit could not have based a finding of anticompetitive conduct on the PeaceHealth precedent.

Fourth, and more generally, the Sixth Circuit’s opinion, in its blinkered search for a “modern” (PeaceHealth) finely-calibrated test to apply in this instance, lost sight of the Supreme Court’s broad teaching in Reiter v. Sonotone Corp. that antitrust law was designed to be “a consumer welfare prescription.”  Kodak’s pricing policy that offered discounts to buyers of its printheads and ink yielded lower prices to consumers.  There was no showing that Collins Inkjet would likely be driven out of business, or, even if it were, that consumers would eventually be harmed.  Absent any showing of likely anticompetitive effects, vertical contractual provisions, including tying, should not be subject to antitrust challenge.  Indeed, as Professor (and former Federal Trade Commissioner) Joshua Wright and I have pointed out:

[T]he potential efficiencies associated with . . . tying . . . and the fact that [tying is] prevalent in markets without significant antitrust market power, lead most commentators to believe that [it is] . . . generally procompetitive and should be analyzed under some form of rule of reason analysis. . . .  [T]he adoption of a rule of reason for tying and presumptions of legality for [tying] . . . under certain circumstances may be long overdue.  

In sum, it is high time for the Supreme Court to take an appropriate case and clarify that tying arrangements (whether explicit or “coerced”) are subject to the rule of reason, with full recognition of tying’s efficiencies.  Such a holding would enable businesses to engage in a wider variety of efficient contracts, thereby promoting consumer welfare.

Finally, while it is at it, the Court should also consider taking a loyalty discount case, to reduce harmful uncertainty in this important area (caused by such economically irrational precedents as LePage’s, Inc. v. 3M) and establish a clear standard to guide the business community.  If it takes a loyalty discount case, the Court could beneficially draw upon Wright’s observation that “economic theory and evidence suggest[s] that instances of anticompetitive loyalty discounts will be relatively rare,” and his recommendation that “an exclusive dealing framework . . . be applied in such cases.”

My Office Door

Thom Lambert —  12 November 2015 — 5 Comments

University professors often post things on their office doors—photos, news clippings, conference posters, political cartoons.   I’ve never been much for that.  The objective, I assume, is to express something about yourself: who you are, what interests you, what values you hold.  I’ve never participated in this custom because I haven’t wanted to alienate students who might not share my views.  That’s not to suggest that I’m shy about those views.  I will—and regularly do—share them with students, even those who I know disagree with me.  But if posting my views on the door were to dissuade students from coming to me to discuss those views (and contrary ones), I would be losing the opportunity to have a meaningful dialogue.  Plus, my tastes veer toward minimalism, and doors covered with postings are ugly. Thus, no postings.

Until today.  My institution, the University of Missouri, is at a crossroads.  We can be a place where ideas—even unpopular ones—are  freely expressed, exchanged, and scrutinized.  Or we can be a place where everyone’s feelings are protected at all times.   It’s one or the other.

Tuesday morning, I opened an email and thought, “What a great prank. It looks so official!”  The email, which was from the MU Police, read as follows:

To continue to ensure that the University of Missouri campus remains safe, the MU Police Department (MUPD) is asking individuals who witness incidents of hateful and/or hurtful speech or actions to:

  • Call the police immediately at 573-882-7201. (If you are in an emergency situation, dial 911.)
  • Give the communications operator a summary of the incident, including location.
  • Provide a detailed description of the individual(s) involved.
  • Provide a license plate and vehicle descriptions (if appropriate).
  • If possible and if it can be done safely, take a photo of the individual(s) with your cell phone.

Delays, including posting information to social media, can often reduce the chances of identifying the responsible parties. While cases of hateful and hurtful speech are not crimes, if the individual(s) identified are students, MU’s Office of Student Conduct can take disciplinary action.

As it turns out, it was no joke.  Anyone on my campus who witnesses “hurtful speech” is directed to call campus police—individuals who carry guns, drive squad cars, and regularly arrest people. Now rest assured, “cases of hateful and hurtful speech are not crimes.”  They can give rise to, at most, “disciplinary action” by the MU Office of Student Conduct.  But still, isn’t it a bit unsettling—chilling, even—to think that if you say something “hurtful” at Mizzou (e.g., gay marriage is an abomination, affirmative action is unfair and hurts those it is ostensibly designed to help, Christians who oppose gay marriage are bigots, Islam is not a religion of peace, white men are privileged in a way that leads to undeserved rewards, culture matters in cultivating success, Republicans are dumb), the police may track you down and you may be required to defend yourself before the student conduct committee?  Perhaps the MU Police, or whoever crafted that email (let’s get real…it wasn’t the police), didn’t really mean that all hurtful speech is potentially problematic.  But if that’s the case, then why did they word the email as they did?  Pandering to an unreasonable element, maybe?

Contrast Mizzou’s approach to that taken by Purdue University.  The day after the Mizzou email, Purdue president Mitch Daniels reminded members of the Purdue community that their school actually stands for both tolerance AND free speech.  Here’s his letter:

Purdue Letter


The contrast between Mizzou and Purdue couldn’t be starker.  And it really, really matters.  I hope that posting these two documents on my door (along with this spot-on Wall Street Journal editorial) will not dissuade students from engaging in dialogue with me.  But I can’t be demure on this one.  So I now have—much to my aesthetic chagrin—a decorated office door.  Please come in and talk, even if you think I’m wrong.


Unless you live under a rock, you know that the president and chancellor of the University of Missouri, where I teach law, have resigned in response to protests over their failure to respond to several (three identified) racist incidents on campus. A group called Concerned Student 1950 staged a series of protests and demanded, among other things, that our president resign, that he publicly denounce his white privilege, that the university mandate a “comprehensive racial awareness and inclusion curriculum,” and that the percentage of black faculty and staff be increased to 10% by the 2017-18 school year. Last week, a student embarked on a hunger strike until Concerned Student 1950’s demands were met, and students in solidarity moved into a tent village on one of our quads.  Over the weekend, the black members of our football team threatened to boycott play until the president resigned, and our coach, facing four straight losses and little prospect of another victory this season, agreed to support and publicize the boycott.  Yesterday morning, Mizzou faculty supporting the Concerned Student group walked out of their classes and headed to the quad, where faculty and administrators joined protesting students in blocking media access to the tent village in the middle of our public quad. Around 10:00 AM, the president resigned, and protesting students danced on the quad.  Toward the end of the day, the chancellor announced that he will move from his position at the end of the year.

The Mizzou Faculty Council and the administration of the law school have expressed to Mizzou students that the faculty fully supports them.  We faculty members have been encouraged to express that support ourselves.  I want to do that now.

I want to express my support for the students for a couple of reasons.  First, I really love Mizzou.  It’s a special place full of wonderful students.  I visited at the University of Minnesota a  few years back, and I couldn’t wait to get back to Mizzou.  Our students reflect the amazing diversity of our state: inner city kids from St. Louis and Kansas City, kids from the suburbs, kids with southern accents from the bootheel, kids with near-Minnesota accents from the northern part of the state, rich kids from fancy prep schools, poor kids who went to public school in the inner city or farm towns.  Unlike so many public law schools, Mizzou has kept its operating costs and its tuition at reasonable levels, so an education here is really open to just about all qualified students from the state.  (Minnesota’s in-state law school tuition is $41,222; Missouri’s is $19,545.)  We mix everyone together and end up with a wonderful student body.  I simply adore my students.

Second, I want to support students who have been the subject of racist remarks because I, too, have experienced the pain of being mocked, criticized, ridiculed, etc. for who I am.  I was a not-very-athletic gay kid who attended the very traditional and somewhat jockish Fairview Christian Academy.  I followed that up with Wheaton College, Billy Graham’s alma mater.  For most of my formative years, I was continually reminded that I was deficient, flawed, damned.  Express slurs were few and far between (though they occurred), but I was not accepted for who I am.  I know the pain of exclusion, and I want both to provide an empathetic ear to my students who feel excluded and to sound a prophetic voice against those who discriminate.

But I could not really support my Mizzou students in this difficult time if I did not point out a few things.

First–The top administrators of a school of 35,000 people cannot prevent all instances of racism.  Ignorant, mean people are sometimes going to yell slurs from their pick-up trucks when they drive through campus.  Drunken frat boys are occasionally going to say ugly things.  When you ambush the homecoming parade, to which parents have brought their small children for a rah-rah college experience, some people are not going to be nice to you.  Those ambushed may be taken aback and may not say all the right things.  People who draw things with poop are especially hard to control. Be prepared: The people who replace the deposed president and chancellor at Mizzou are unlikely to prevent every racist incident on our campus.

Second–The U.S. Constitution forbids state institutions from employing racial quotas.  Having been involved in hiring at Mizzou for a number of years, I can assure that we bend over backward to fill open positions with qualified minority applicants. It is highly unlikely that Concerned Student 1950’s demand that the percentage of black faculty and staff at Mizzou be raised to 10% by 2017-18 can be implemented in a manner consistent with constitutional obligations.  You should know that.

Third–Free speech means more than the freedom to express views with which you agree.  I honestly think most Mizzou students understand this point, but I’m afraid that the administrator and communications professor in this video don’t grasp it.  Lest you be misled by their ill-advised bullying, you should know that the First Amendment is for everyone.

Fourth–Unreasonable demands have consequences.  We will survive this, but Mizzou has been badly weakened.  I can’t imagine that the press accounts from the last week will help with minority student and faculty recruitment next year.  That’s a shame, because based on my encounters with a great many minority students and professors at Mizzou over the past twelve years, I believe most have had good experiences.  Perhaps they haven’t been honest with me.  Or perhaps the situation has changed in the last couple of years.  If so, I’m terribly sorry to hear that. But, following the events of the last week, I can’t imagine that next year will be better.

Fifth–Regardless of your take on the events of the last week, I hope you will not let bitterness reign in your hearts.  Unlike many of my gay friends from conservative religious backgrounds, I chose years ago not to write off those people who were once unkind to me.  I’m glad I made that choice.  I hope any Mizzou student who is currently feeling marginalized for any reason will keep calm, carry on, give others the benefit of the doubt, and be open to reconciliation.

So, Mizzou students, I support you.  But I will not coddle you.  You’re adults and should be treated as such.


shepherd-joannaTruth on the Market is delighted to welcome our newest blogger, Joanna Shepherd. Joanna is a Professor of Law at Emory School of Law and holds an adjunct position in the Emory Economics Department (where she also earned her PhD). At the law school she teaches Torts, Law and Economics, Analytical Methods for Lawyers, and Legal and Economic Issues in Health Policy. She also frequently teaches economics courses to law professors and federal and state judges.

Joanna is also a senior scholar at the International Center for Law and Economics.

Joanna’s research focuses on various law & econ topics. Her recent research has examined issues related to the healthcare and pharmaceutical industries, tort reform, litigation practice, and judicial behavior. Her works has appeared in the Michigan Law Review, Vanderbilt Law Review, Southern California Law Review, New York University Law ReviewDuke Law JournalUCLA Law Review, The Journal of Legal Studies, The Journal of Law & Economics, The American Law & Economics Review, Health Matrix, and The American Journal of Law & Medicine, among others. Joanna is also an author of the textbook, Economic Analysis for Lawyers, with Henry Butler and Christopher Drahozal. Her research has been discussed in numerous newspapers, including the Wall Street Journal and the New York Times, and has been cited by several courts including the Supreme Court.

You can find links to Joanna’s scholarship on her SSRN page.

Welcome Joanna!

Tomorrow (August 24, 2015) marks once and future TOTM’er Josh Wright’s last day as an FTC Commissioner. Starting tomorrow and continuing throughout the week, Truth on the Market will be hosting a symposium —  a collection of commentaries and contributions — honoring Josh’s tenure at the FTC. We’ve invited contributions from a range of luminaries, including academics, practitioners, former FTC officials, and the like. Watch this space for the contributions, and feel free to add your own thoughts in the comments to the posts. Links to the posts will be collected here.

Monday’s posts will commence with contributions from

  • Richard Epstein,
  • Jon Jacobson,
  • Tom Hazlett, and
  • Keith Hylton

— with many more to come!


Uber is currently facing a set of plaintiffs who are seeking class certification in the Northern District of California (O’Connor, et. al v. Uber, #CV 13-3826-EMC) on two distinct grounds. First, the plaintiffs allege that Uber systematically deprived them of tips from riders by virtue of how the service is presented to end-users and how compensation is given to the riders in violation of the California Unfair Competition Law, Cal. Bus. & Prof. Code § 17200 et seq. Second, the plaintiffs claim that Uber misclassified its drivers – all 160,000 of them in California over the last five years – by failing to give them the legal definition of “employee” and, following from this, deprived said “employees” of reimbursement for things like mileage, gas, and other wear-and-tear on their vehicles (not to mention the shadow of entitlements like benefits and worker’s comp).

Essentially, claim one is based on the notion that Uber informs passengers that gratuity is included in the total cost of the car service and that there is no need to tip the driver. However, according to the plaintiffs, Uber either failed to collect this gratuity, or by failing to differentiate between the gratuity and the fee for the ride, and then collecting its own 20% cut of the total fee, the company improperly retained some of the gratuity for itself. In truth, it’s not completely clear from the complaint exactly how the plaintiffs are calculating allegedly withheld tips. Uber does a good job in its motion to defeat certification of pointing out, on the one hand, that there is no such thing as a “standard tip,” and, on the other hand, that the assessment of the tip issue would require so much individualized examination — from figuring out whether drivers were told that they could be tipped or not, to figuring out if drivers actually were consistently tipped — that the common issues proper to class examination would be overwhelmed.

The real meat of this case, however, and the issue with the most effect on both Uber’s bottom line as well as on the future of sharing platforms generally, is whether the drivers should be classified as employees or not.

Uber’s motion to defeat certification is, logically enough, based on attacking the commonality and typicality requirements of Rule 23. The main thrust of Uber’s motion is that not only would the four named plaintiffs be inappropriate to represent the 160,000 member class of allegedly harmed drivers, but also no such plaintiffs could represent such a class as the relationship between Uber and its drivers is so diverse that no common questions or issues would control the proceeding. In support of its position, Uber introduced the sworn declarations of over 400 Uber drivers from California, each detailing a unique situation that would either make them not in line with the harms alleged by the named plaintiffs, or squarely opposed to them.

Further, there were seventeen different contracts involved in the relationship between Uber and the 160,000 drivers swept up into the suit, which would make identifying common questions exceedingly difficult. Even terms that are common across agreements, Uber claims, would have enough distinction between them to make class certification impossible. For instance, Uber cited numerous examples from its different agreements where tipping was permitted, and others where it was not mentioned at all. Similarly, Uber cited examples where the right to terminate rested solely with Uber, and others where the right to terminate was by mutual consent between Uber and the driver.  Further, Uber claims that the employment test from Borello (the case that governs employee classification in California) requires a fact-based examination of each driver’s particular circumstances owing to the wide variation in contract terms — further making class certification inappropriate.

Uber’s arguments are all sound, and I sincerely hope that it defeats the class certification. But the case itself represents an ongoing and persistent problem for Uber and sharing economy platforms across the United States (and the world, really). The core of that problem is simply this: are you an employee or a contractor? A heading from Uber’s motion stands out to me as emblematic of this problem:

The Named Plaintiffs Are Not Typical Of the Putative Class Because There Is No Typical Uber Driver

There is no typical Uber driver because Uber is just a platform, the definitions of our antiquated legal system notwithstanding. The real value proposition of sharing platforms is that they enable normal folks — that is, people outside of a typically defined industry — to take part in an industry that was previously dominated by firms (and replete with considerable barriers to entry). As the Northern District of California observes in Cotter v. Lyft, trying to fit a sharing economy worker of today into yesterday’s notion of “employees” and “contractors” is akin to “be[ing] handed a square peg and asked to choose between two round holes.” In the same passage, that court observed that “[t]he test the California courts have developed over the 20th Century for classifying workers isn’t very helpful in addressing this 21st Century problem.”


The claims of the plaintiffs in the Uber class action notwithstanding, there is nothing inherently “employee”-like about an Uber driver, and there are plenty of opportunities for sharing economy workers to not be quite so “contractor”-like either.  What we really need is some creative thinking, and an application of legal principles (as opposed to tired categories) to the new reality of the 21st century in order to come up with a third way (and maybe a fourth and fifth way, as well…) of regulating labor relationships. If we must have classes, consider it the entrepreneurial class.

Uber’s business model is a great example of how an employee definition doesn’t quite make sense. The party that contracts with Uber might not even be an individual, but a corporation that, even without Uber’s platform, would be providing private ride services. Particularly with UberBlack, private companies use Uber’s lead generation platform merely to supplement their own marketing efforts. Obviously converting these companies and their own employees into “employees” of Uber is ludicrous.

However, even for the more common example that many people will first think of — the guy down the street with a car and some time on his hands — sticking him into the employee category may or may not make sense. First, as an employee he will be handed a whole raft of potential benefits that have corresponding obligations for Uber. Those obligations — like disability, health benefits, time off, etc — will come at a cost, which will typically mean less money earned for that sometimes-driver as those costs are passed on in the form of either increased prices (and a reduction in ridership) or reduced wages. For many people, this will decrease their marginal earnings to the point where it won’t make sense for them to drive anymore.

Second, for many people it may lead to an outright conflict that either prevents them from being a driver, or else locks them into a single platform, thus harming competition in the marketplace. A driver who is Uber’s “employee” may be in violation of her duties of loyalty to Uber if she takes rides from the Lyft platform (and multi-homing is extremely common in this space). Similarly, employers – in particular state and municipal governments – frequently have strict rules on outside employment, and a determination that driving for Uber makes you an “employee” of the company may effectively preclude drivers by virtue of their actual employer’s policies.

Further, I believe it’s notable that many employment tests in the United States are extremely multi-factor; the Borello case from CA outlines thirteen distinct considerations, for instance.  The utter complexity of fitting a worker into an “employee” classification suggests that even this old, familiar notion of what it is to be an “employee” is not quite as clear as we often presume, but is more of a “catch-all” category. The sharing-economy platforms from companies like Uber and Lyft will only exacerbate this problem — and serve to make its problematic consequences more pointed.

But even the definition of “contractor” is inapplicable to these drivers. In the case at hand, Uber was accused of treating drivers as employees because it provided suggestions about how to earn higher ratings from riders, and because it offered “on-boarding” programs that give new drivers an orientation. This general training is not a need unique to Uber, however. Consider Instacart’s recent announcement that it would re-classify some of its employees in Boston as part-time workers. In large part, it seems clearly to be the case that the company decided to make this move for purely strategic, legal reasons. In actuality, it wanted simply to be able to guarantee that there would be some minimum level of quality for the people who provided services through its network. This might involve orientation meetings, intermittent trainings, and some minor direction on how a shopper should perform his or her work (for instance, pick produce last so that it remains fresh). There is no obvious reason why providing this sort of guidance should force a company to destroy all of the unique and socially beneficial qualities of its offerings by being forced into classifying on-demand workers as “employees.”

The sharing economy promises to remove the transaction costs that have for quite a long time chained employees to firms. On their own, individuals simply cannot obtain enough information that would enable them to realize a fully self-defined work environment. It’s an accident of history (and technology) — of scarce resources and scarcer information — that the model of work has revolved around selling one’s services to an employer. But technology is now rendering this model inefficient compared to the alternatives — and our legal system should not get in its way. Canadian courts have begun experimenting with a third classification of worker — the “dependent” worker, a classification that may or may not work here — and so too should our courts and legislatures start thinking about a new classification. It makes no sense to drag down cutting-edge 21st century work and life models with depression-era notions of what it means to earn a living.

In its June 30 decision in United States v. Apple Inc., a three-judge Second Circuit panel departed from sound antitrust reasoning in holding that Apple’s e-book distribution agreement with various publishers was illegal per se. Judge Dennis Jacobs’ thoughtful dissent, which substantially informs the following discussion of this case, is worth a close read.

In 2009, Apple sought to enter the retail market for e-books, as it prepared to launch its first iPad tablet. Apple, however, confronted an e-book monopolist, Amazon (possessor of a 90 percent e-book market share), that was effectively excluding new entrants by offering bestsellers at a loss through its popular Kindle device ($9.99, a price below what Amazon was paying publishers for the e-book book rights). In order to effectively enter the market without incurring a loss itself (by meeting Amazon’s price) or impairing its brand (by charging more than Amazon), Apple approached publishers that dealt with Amazon and offered itself as a competing e-book buyer, subject to the publishers agreeing to a new distribution model that would lower barriers to entry into retail e-book sales. The new publishing model was implemented by three sets of contract terms Apple asked the publishers to accept – agency pricing, tiered price caps, and a most-favored-nation (MFN) clause. (I refer the reader to the full panel majority opinion for a detailed discussion of these clauses.) None of those terms, standing alone, is illegal. Although the publishers were unhappy about Amazon’s below-cost pricing for e-books, no one publisher alone could counter Amazon. Five of the six largest U.S. publishers (Hachette, HarperCollins, Macmillan, Penguin, and Simon & Schuster) agreed to Apple’s terms and jointly convinced Amazon to adopt agency pricing. Apple also encouraged other publishers to implement agency pricing in their contracts with other retailers. The barrier to entry thus removed, Apple entered the retail market as a formidable competitor. Amazon’s retail e-book market share fell, and today stands at 60 percent.

The U.S. Department of Justice (DOJ) and 31 states sued Apple and the five publishers for conspiring in unreasonable restraint of trade under Sherman Act § 1. The publishers settled (signing consent decrees which prohibited them for a period from restricting e-book retailers’ ability to set prices), but Apple proceeded to a bench trial. A federal district court held that Apple’s conduct as a vertical enabler of a horizontal price conspiracy among the publishers was a per se violation of § 1, and that (in any event) Apple’s conduct would also violate § 1 under the antitrust rule of reason.   A majority of the Second Circuit panel affirmed on the ground of per se liability, without having to reach the rule of reason question.

Judge Jacobs’ dissent argued that Apple’s conduct was not per se illegal and also passed muster under the rule of reason. He pointed to three major errors in the majority’s opinion. First, the holding that the vertical enabler of a horizontal price fixing is in per se violation of the antitrust laws conflicts with the Supreme Court’s teaching (in overturning the per se prohibition on resale price maintenance) that a vertical agreement designed to facilitate a horizontal cartel “would need to be held unlawful under the rule of reason.” Leegin Creative Leather Prods, Inc. v. PSKS, Inc. 551 U.S. 877, 893 (2007) (emphasis added).   Second, the district court failed to recognize that Apple’s role as a vertical player differentiated it from the publishers – it should have considered Apple as a competitor on the distinct horizontal plane of retailers, where Apple competed with Amazon (and with smaller player such as Barnes & Noble). Third, assessed under the rule of reason, Apple’s conduct was “overwhelmingly” procompetitive; Apple was a major potential competitor in a market dominated by a 90 percent monopoly, and was “justifiably unwilling” to enter a market on terms that would assure a loss on sales or exact a toll on its reputation.

Judge Jacobs’ analysis is on point. The Supreme Court’s wise reluctance to condemn any purely vertical contractual restraint under the per se rule reflects a sound understanding that vertical restraints have almost always been found to be procompetitive or competitively neutral. Indeed, vertical agreements that are designed to facilitate entry into an important market dominated by one firm, such as the ones at issue in the Apple case, are especially bad candidates for summary condemnation. Thus, the majority’s decision to apply the per se rule to Apple’s contracts appears particularly out of touch with both scholarship and marketplace realities.

More generally, as Professor Herbert Hovenkamp (the author of the leading antitrust treatise) and other scholars have emphasized, well-grounded antitrust analysis involves a certain amount of preliminary evaluation of a restraint seen in its relevant factual context, before a “per se” or “rule of reason” label is applied. (In the case of truly “naked” secret hard core cartels, which DOJ prosecutes under criminal law, the per se label may be applied immediately.) The Apple panel majority panel botched this analytic step, in failing to even consider that Apple’s restraints could enhance retail competition with Amazon.

The panel majority also appeared overly fixated on the fact that some near-term e-book retail prices rose above Amazon’s previous below cost levels in the wake of Apple’s contracts, without noting the longer term positive implications for the competitive process of new e-book entry. Below-cost prices are not a feature of durable efficient competition, and in this case may well have been a temporary measure aimed at discouraging entry. In any event, what counts in measuring consumer welfare is not short term price, but whether expanded output is being promoted by a business arrangement – a key factor that the majority notably failed to address. (It appears highly probable that the fall in Amazon’s e-book retail market share, and the invigoration of e-book competition, have generated output and welfare levels higher than those that would have prevailed had Amazon maintained its monopoly. This is bolstered by Apple’s showing, which the majority does not deny, that in the two years following the “conspiracy” among Apple and the publishers, prices across the e-book market as a whole fell slightly and total output increased.)

Finally, Judge Jacobs’ dissent provides strong arguments in favor of upholding Apple’s conduct under the rule of reason. As the dissent stresses, removal of barriers to entry that shield a monopolist, as in this case, is in line with the procompetitive goals of antitrust law. Another procompetitive effect is the encouragement of innovation (manifested by the enablement of e-book reading with the cutting-edge functions of the iPad), a hallmark and benefit of competition. Another benefit was that the elimination of below-cost pricing helped raise authors’ royalties. Furthermore, in the words of the dissent, any welfare reductions due to Apple’s vertical restrictions are “no more than a slight offset to the competitive benefits that now pervade the relevant market.” (Admittedly that comment is a speculative observation, but in my view very likely a well-founded one.) Finally, as the dissent points out, the district court’s findings demonstrate that Apple could not have entered and competed effectively using other strategies, such as wholesale contracts involving below-cost pricing (like Amazon’s) or higher prices. Summing things up, the dissent explains that “Apple took steps to compete with a monopolist and open the market to more entrants, generating only minor competitive restraints in the process. Its conduct was eminently reasonable; no one has suggested a viable alternative.” In closing, even if one believes a more fulsome application of the rule of reason is called for before reaching the dissent’s conclusion, the dissent does a good job in highlighting the key considerations at play here – considerations that the majority utterly failed to address.

In sum, the Second Circuit panel majority wore jurisprudential blinders in its Apple decision. Like the mesmerized audience at a magic show, it focused in blinkered fashion on a magician’s sleight of hand (the one-dimensional characterization of certain uniform contractual terms), while not paying attention to what was really going on (the impressive welfare-enhancing invigoration of competition in e-book retailing). In other words, the majority decision showed a naïve preference for quick and superficial characterizations of conduct at the expense of a nuanced assessment of the broader competitive context. Perhaps the Second Circuit en banc will have the opportunity to correct the panel’s erroneous understanding of per se and rule of reason analysis. Even better, the Supreme Court may wish to step in to ensure that its thoughtful development of antitrust doctrine in recent years – focused on actual effects and economic efficiency, not on superficial condemnatory labels that ignore marketplace benefits – not be undermined.

imageI am of two minds when it comes to the announcement today that the NYC taxi commission will permit companies like Uber and Lyft to update, when the companies wish, the mobile apps that serve as the front end for the ridesharing platforms.

My first instinct is to breathe a sigh of relief that even the NYC taxi commission eventually rejected the patently ridiculous notion that an international technology platform should have its update schedule in anyway dictated by the parochial interests of a local transportation fiefdom.

My second instinct is to grit my teeth in frustration that, in the face of the overwhelming transformation going on in the world today because of technology platforms offered by the likes of Uber and Lyft, anyone would even think to ask the question “should I ask the NYC taxi commission whether or not I can update the app on my users’ smartphones?”

That said, it’s important to take the world as you find it, not as you wish it to be, and so I want to highlight some items from the decision that deserve approbation.

Meera Josh, the NYC Taxi Commission chairperson and CEO, had this to say of the proposed rule:

We re-stylized the rules so they’re tech agnostic because our point is not to go after one particular technology – things change quicker than we do – it’s to provide baseline consumer protection and driver safety requirements[.]

I love that the commission gets this. The real power in the technology that drives the sharing economy is that it can change quickly in response to consumer demand. Further, regulators can offer value to these markets only when they understand that the nature of work and services are changing, and that their core justification as consumer protection agencies necessarily requires them to adjust when and how they intervene.

Although there is always more work to be done to make room for these entrepreneurial platforms (for instance, the NYC rules appear to require that all on-demand drivers – including the soccer mom down the street driving for Lyft – be licensed through the commission), this is generally forward-thinking. I hope that more municipalities across the country take notice, and that the relevant regulators follow suit in repositioning themselves as partners with these innovative companies.

If you haven’t been following the ongoing developments emerging from the demise of Grooveshark, the story has only gotten more interesting. As the RIAA and major record labels have struggled to shut down infringing content on Grooveshark’s site (and now its copycats), groups like EFF would have us believe that the entire Internet was at stake — even in the face of a fairly marginal victory by the recording industry. In the most recent episode, the issuance of a TRO against CloudFlare — a CDN service provider for the copycat versions of Grooveshark — has sparked much controversy. Ironically for CloudFlare, however, its efforts to evade compliance with the TRO may well have opened it up to far more significant infringement liability.

In response to Grooveshark’s shutdown in April, copycat sites began springing up. Initially, the record labels played a game of whac-a-mole as the copycats hopped from server to server within the United States. Ultimately the copycats settled on, using a host and registrar outside of the country, as well as anonymized services that made direct action against the actual parties next to impossible. Instead of continuing the futile chase, the plaintiffs decided to address the problem more strategically.

High volume web sites like Grooveshark frequently depend upon third party providers to optimize their media streaming and related needs. In this case, the copycats relied upon the services of CloudFlare to provide DNS hosting and a content delivery network (“CDN”). Failing to thwart Grooveshark through direct action alone, the plaintiffs sought and were granted a TRO against certain third-parties, eventually served on CloudFlare, hoping to staunch the flow of infringing content by temporarily enjoining the ancillary activities that enabled the pirates to continue operations.

CloudFlare refused to comply with the TRO, claiming the TRO didn’t apply to it (for reasons discussed below). The court disagreed, however, and found that CloudFlare was, in fact, bound by the TRO.

Unsurprisingly the copyright scolds came out strongly against the TRO and its application to CloudFlare, claiming that

Copyright holders should not be allowed to blanket infrastructure companies with blocking requests, co-opting them into becoming private trademark and copyright police.

Devlin Hartline wrote an excellent analysis of the court’s decision that the TRO was properly applied to CloudFlare, concluding that it was neither improper nor problematic. In sum, as Hartline discusses, the court found that CloudFlare was indeed engaged in “active concert and participation” and was, therefore, properly subject to a TRO under FRCP 65 that would prevent it from further enabling the copycats to run their service.

Hartline’s analysis is spot-on, but we think it important to clarify and amplify his analysis in a way that, we believe, actually provides insight into a much larger problem for CloudFlare.

As Hartline states,

This TRO wasn’t about the “world at large,” and it wasn’t about turning the companies that provide internet infrastructure into the “trademark and copyright police.” It was about CloudFlare knowingly helping the enjoined defendants to continue violating the plaintiffs’ intellectual property rights.

Importantly, the issuance of the TRO turned in part on whether the plaintiffs were likely to succeed on the merits — which is to say that the copycats could in fact be liable for copyright infringement. Further, the initial TRO became a preliminary injunction before the final TRO hearing because the copycats failed to show up to defend themselves. Thus, CloudFlare was potentially exposing itself to a claim of contributory infringement, possibly from the time it was notified of the infringing activity by the RIAA. This is so because a claim of contributory liability would require that CloudFlare “knowingly” contributed to the infringement. Here there was actual knowledge upon issuance of the TRO (if not before).

However, had CloudFlare gone along with the proceedings and complied with the court’s order in good faith, § 512 of the Digital Millennium Copyright Act (DMCA) would have provided a safe harbor. Nevertheless, following from CloudFlare’s actual behavior, the company does now have a lot more to fear than a mere TRO.

Although we don’t have the full technical details of how CloudFlare’s service operates, we can make some fair assumptions. Most importantly, in order to optimize the content it serves, a CDN would necessarily have to store that content at some point as part of an optimizing cache scheme. Under the terms of the DMCA, an online service provider (OSP) that engages in caching of online content will be immune from liability, subject to certain conditions. The most important condition relevant here is that, in order to qualify for the safe harbor, the OSP must “expeditiously [] remove, or disable access to, the material that is claimed to be infringing upon notification of claimed infringement[.]”

Here, not only had CloudFlare been informed by the plaintiffs that it was storing infringing content, but a district court had gone so far as to grant a TRO against CloudFlare’s serving of said content. It certainly seems plausible to view CloudFlare as acting outside the scope of the DMCA safe harbor once it refused to disable access to the infringing content after the plaintiffs contacted it, but certainly once the TRO was deemed to apply to it.

To underscore this point, CloudFlare’s arguments during the TRO proceedings essentially admitted to knowledge that infringing material was flowing through its CDN. CloudFlare focused its defense on the fact that it was not an active participant in the infringing activity, but was merely a passive network through which the copycats’ content was flowing. Moreover, CloudFlare argued that

Even if [it]—and every company in the world that provides similar services—took proactive steps to identify and block the Defendants, the website would remain up and running at its current domain name.

But while this argument may make some logical sense from the perspective of a party resisting an injunction, it amounts to a very big admission in terms of a possible infringement case — particularly given CloudFlare’s obstinance in refusing to help the plaintiffs shut down the infringing sites.

As noted above, CloudFlare had an affirmative duty to to at least suspend access to infringing material once it was aware of the infringement (and, of course, even more so once it received the TRO). Instead, CloudFlare relied upon its “impossibility” argument against complying with the TRO based on the claim that enjoining CloudFlare would be futile in thwarting the infringement of others. CloudFlare does appear to have since complied with the TRO (which is now a preliminary injunction), but the compliance does not change a very crucial fact: knowledge of the infringement on CloudFlare’s part existed before the preliminary injunction took effect, while CloudFlare resisted the initial TRO as well as RIAA’s efforts to secure compliance.

Phrased another way, CloudFlare became an infringer by virtue of having cached copyrighted content and been given notice of that content. However, in its view, merely removing CloudFlare’s storage of that copyrighted content would have done nothing to prevent other networks from also storing the copyrighted content, and therefore it should not be enjoined from its infringing behavior. This essentially amounts to an admission of knowledge of infringing content being stored in its network.

It would be hard to believe that CloudFlare’s counsel failed to advise it to consider the contributory infringement issues that could arise from its conduct prior to and during the TRO proceedings. Thus CloudFlare’s position is somewhat perplexing, particularly once the case became a TRO proceeding. CloudFlare could perhaps have made technical arguments against the TRO in an attempt to demonstrate to its customers that it didn’t automatically shut down services at the behest of the RIAA. It could have done this in good faith, and without the full-throated “impossibility” argument that could very plausibly draw them into infringement litigation. But whatever CloudFlare thought it was gaining in taking a “moral” stance on behalf of OSPs everywhere with its “impossibility” argument, it may well have ended up costing itself much more.

Nearly all economists from across the political spectrum agree: free trade is good. Yet free trade agreements are not always the same thing as free trade. Whether we’re talking about the Trans-Pacific Partnership or the European Union’s Digital Single Market (DSM) initiative, the question is always whether the agreement in question is reducing barriers to trade, or actually enacting barriers to trade into law.

It’s becoming more and more clear that there should be real concerns about the direction the EU is heading with its DSM. As the EU moves forward with the 16 different action proposals that make up this ambitious strategy, we should all pay special attention to the actual rules that come out of it, such as the recent Data Protection Regulation. Are EU regulators simply trying to hogtie innovators in the the wild, wild, west, as some have suggested? Let’s break it down. Here are The Good, The Bad, and the Ugly.

The Good

The Data Protection Regulation, as proposed by the Ministers of Justice Council and to be taken up in trilogue negotiations with the Parliament and Council this month, will set up a single set of rules for companies to follow throughout the EU. Rather than having to deal with the disparate rules of 28 different countries, companies will have to follow only the EU-wide Data Protection Regulation. It’s hard to determine whether the EU is right about its lofty estimate of this benefit (€2.3 billion a year), but no doubt it’s positive. This is what free trade is about: making commerce “regular” by reducing barriers to trade between states and nations.

Additionally, the Data Protection Regulation would create a “one-stop shop” for consumers and businesses alike. Regardless of where companies are located or process personal information, consumers would be able to go to their own national authority, in their own language, to help them. Similarly, companies would need to deal with only one supervisory authority.

Further, there will be benefits to smaller businesses. For instance, the Data Protection Regulation will exempt businesses smaller than a certain threshold from the obligation to appoint a data protection officer if data processing is not a part of their core business activity. On top of that, businesses will not have to notify every supervisory authority about each instance of collection and processing, and will have the ability to charge consumers fees for certain requests to access data. These changes will allow businesses, especially smaller ones, to save considerable money and human capital. Finally, smaller entities won’t have to carry out an impact assessment before engaging in processing unless there is a specific risk. These rules are designed to increase flexibility on the margin.

If this were all the rules were about, then they would be a boon to the major American tech companies that have expressed concern about the DSM. These companies would be able to deal with EU citizens under one set of rules and consumers would be able to take advantage of the many benefits of free flowing information in the digital economy.

The Bad

Unfortunately, the substance of the Data Protection Regulation isn’t limited simply to preempting 28 bad privacy rules with an economically sensible standard for Internet companies that rely on data collection and targeted advertising for their business model. Instead, the Data Protection Regulation would set up new rules that will impose significant costs on the Internet ecosphere.

For instance, giving citizens a “right to be forgotten” sounds good, but it will considerably impact companies built on providing information to the world. There are real costs to administering such a rule, and these costs will not ultimately be borne by search engines, social networks, and advertisers, but by consumers who ultimately will have to find either a different way to pay for the popular online services they want or go without them. For instance, Google has had to hire a large “team of lawyers, engineers and paralegals who have so far evaluated over half a million URLs that were requested to be delisted from search results by European citizens.”

Privacy rights need to be balanced with not only economic efficiency, but also with the right to free expression that most European countries hold (though not necessarily with a robust First Amendment like that in the United States). Stories about the right to be forgotten conflicting with the ability of journalists to report on issues of public concern make clear that there is a potential problem there. The Data Protection Regulation does attempt to balance the right to be forgotten with the right to report, but it’s not likely that a similar rule would survive First Amendment scrutiny in the United States. American companies accustomed to such protections will need to be wary operating under the EU’s standard.

Similarly, mandating rules on data minimization and data portability may sound like good design ideas in light of data security and privacy concerns, but there are real costs to consumers and innovation in forcing companies to adopt particular business models.

Mandated data minimization limits the ability of companies to innovate and lessens the opportunity for consumers to benefit from unexpected uses of information. Overly strict requirements on data minimization could slow down the incredible growth of the economy from the Big Data revolution, which has provided a plethora of benefits to consumers from new uses of information, often in ways unfathomable even a short time ago. As an article in Harvard Magazine recently noted,

The story [of data analytics] follows a similar pattern in every field… The leaders are qualitative experts in their field. Then a statistical researcher who doesn’t know the details of the field comes in and, using modern data analysis, adds tremendous insight and value.

And mandated data portability is an overbroad per se remedy for possible exclusionary conduct that could also benefit consumers greatly. The rule will apply to businesses regardless of market power, meaning that it will also impair small companies with no ability to actually hurt consumers by restricting their ability to take data elsewhere. Aside from this, multi-homing is ubiquitous in the Internet economy, anyway. This appears to be another remedy in search of a problem.

The bad news is that these rules will likely deter innovation and reduce consumer welfare for EU citizens.

The Ugly

Finally, the Data Protection Regulation suffers from an ugly defect: it may actually be ratifying a form of protectionism into the rules. Both the intent and likely effect of the rules appears to be to “level the playing field” by knocking down American Internet companies.

For instance, the EU has long allowed flexibility for US companies operating in Europe under the US-EU Safe Harbor. But EU officials are aiming at reducing this flexibility. As the Wall Street Journal has reported:

For months, European government officials and regulators have clashed with the likes of Google, and Facebook over everything from taxes to privacy…. “American companies come from outside and act as if it was a lawless environment to which they are coming,” [Commissioner Reding] told the Journal. “There are conflicts not only about competition rules but also simply about obeying the rules.” In many past tussles with European officialdom, American executives have countered that they bring innovation, and follow all local laws and regulations… A recent EU report found that European citizens’ personal data, sent to the U.S. under Safe Harbor, may be processed by U.S. authorities in a way incompatible with the grounds on which they were originally collected in the EU. Europeans allege this harms European tech companies, which must play by stricter rules about what they can do with citizens’ data for advertising, targeting products and searches. Ms. Reding said Safe Harbor offered a “unilateral advantage” to American companies.

Thus, while “when in Rome…” is generally good advice, the Data Protection Regulation appears to be aimed primarily at removing the “advantages” of American Internet companies—at which rent-seekers and regulators throughout the continent have taken aim. As mentioned above, supporters often name American companies outright in the reasons for why the DSM’s Data Protection Regulation are needed. But opponents have noted that new regulation aimed at American companies is not needed in order to police abuses:

Speaking at an event in London, [EU Antitrust Chief] Ms. Vestager said it would be “tricky” to design EU regulation targeting the various large Internet firms like Facebook, Inc. and eBay Inc. because it was hard to establish what they had in common besides “facilitating something”… New EU regulation aimed at reining in large Internet companies would take years to create and would then address historic rather than future problems, Ms. Vestager said. “We need to think about what it is we want to achieve that can’t be achieved by enforcing competition law,” Ms. Vestager said.

Moreover, of the 15 largest Internet companies, 11 are American and 4 are Chinese. None is European. So any rules applying to the Internet ecosphere are inevitably going to disproportionately affect these important, US companies most of all. But if Europe wants to compete more effectively, it should foster a regulatory regime friendly to Internet business, rather than extend inefficient privacy rules to American companies under the guise of free trade.


Near the end of the The Good, the Bad, and the Ugly, Blondie and Tuco have this exchange that seems apropos to the situation we’re in:

Bloeastwoodndie: [watching the soldiers fighting on the bridge] I have a feeling it’s really gonna be a good, long battle.
Tuco: Blondie, the money’s on the other side of the river.
Blondie: Oh? Where?
Tuco: Amigo, I said on the other side, and that’s enough. But while the Confederates are there we can’t get across.
Blondie: What would happen if somebody were to blow up that bridge?

The EU’s DSM proposals are going to be a good, long battle. But key players in the EU recognize that the tech money — along with the services and ongoing innovation that benefit EU citizens — is really on the other side of the river. If they blow up the bridge of trade between the EU and the US, though, we will all be worse off — but Europeans most of all.