Guy Rub: Copyright or Contract?

Using software often means you have to sign a contract as a condition for using the software. This "end user license agreement," called a EULA, will lay out the terms under which the software can be used. For example, the EULA you sign to play a video game might say: "The player of this video game cannot cheat while playing the game."  What if you breach the EULA by playing the game using a commercially available cheating "bot"?  Is this copyright infringement? Or is this just a breach of contract?  This may seem obscure, but the question matters a lot.  For one thing, in this example, if the video game publisher has a copyright claim against the cheater, not just a contract claim, this could mean very large statutory damages versus no damages at all.

This is just one of many scenarios in which copyright owners use contracts to control the conditions of use, and whose breach may, or may not, give rise to copyright infringement. In his new article, Against Copyright Customization, Guy Rub addresses this thorny question—copyright or contract?—along with many closely related questions. For example: when is a software user a mere licensee versus an owner?  (Spoiler alert:  almost always!)  The article is forthcoming in Iowa Law Review and a draft can be downloaded on SSRN.

I interviewed Guy about the article. Here is a transcription.

CAH: Your paper is about how copyright owners use license agreements to get copyright infringement remedies for what would otherwise be simply a breach of contract.  Can you say more about when this issue comes up?  Is this just for digital works like software? Or can it happen with, say, physical books, as well?  

GUY RUB: We see this more with software and digital goods. But there is a "brick and mortar" version too. The classic case is a photographer, who authorizes a publisher to publish 5000 copies, and the publisher publishes 6000 copies.  Another version is a license that allows someone to install software on a single computer, and the end user installs the software on more than one computer. Or a license that says: "you can install this software as many times as you want, but you need to follow this complex algorithm to determine the precise payment for each installation..." and then the user messes up in their payments, not intentionally, but in breach of what the contract said. Is this copyright infringement or breach of contract?  If it's a breach of contract, the copyright owner just gets the gap in the payments, they get what was owed under the contract. But if it's copyright infringement, the owner gets way more damages than the gap.  These are all quite common cases.   

CAH: Ok, so can we say the situation arises any time that you have a copyrighted work that is used subject to a license that controls how the work can be used or copied?

GUY RUB: The components are that there is a copyrighted good, and there is some agreement that controls certain aspects of the transaction. I would not say that agreement is a "license," because I don't know if it's a license! 

CAH: Ah, let's return to that in a second.  For now, can you explain why it matters, at practical level, whether a claim for breach of the license is seen as copyright infringement or breach of contract?  

GUY RUB: It matters in two ways. One is straight up remedies. The remedies for copyright are more generous.  It is easier to get an injunction . You can get statutory damages, you can disgorge the benefit of the other side. You cannot do any of that in contract law.  The other reason why copyright is really valuable is the ability to go after after third parties, and especially relying on contributory infringement. That was MDY v. Blizzard in fact.  Blizzard, creator of World of Warcraft, was able to sue the maker of the cheating "bot," MDY Industries, not just the game players who were using the bot.  Blizzard was trying to get copyright remedies, but it was also relying on copyright contributory infringement to sue the seller of the bot, who could only be liable under a contributory theory.

CAH: What about tortious interference with contract?  

GUY RUB: The big issue there is preemption. Our favorite topic.  Tortious interference with contract is commonly considered preempted.

CAH: Oh, no. Ok. Let's not open that can of worms. Let's get into the weeds of the law. There are enough weeds here already. The weeds is a place you excel.

GUY RUB: That is the challenge. Writing all of these papers about the weeds, and making people still find it interesting!

CAH: You talk a lot in the paper about two leading cases from the 9th Circuit, both decided in 2010: Vernor v. Autodesk, and then the case you just mentioned, MDY v. Blizzard.  Each of these cases provides a crucial link in the chain that allows copyright owners to use contracts to get copyright remedies.  Starting with Vernor, that case, as I understand it, held that a software user will be treated as a licensee of the copy rather than an owner of a copy, so long as the copyright owner meets some pretty minimal requirements. The licensee status means that the so-called "essential step" defense does not apply, because the user is not an "owner."  Vernor is what set the stage for MDY, because if the end user is not an owner of their software, then merely using the software to play the game, outside the scope of the license, can be infringing, even without some other act that violates Section 106 like making non-essential copies or "derivative works"... 

So can we talk about Vernor and its implications?

GUY RUB: Vernor is a big big deal. It's an even bigger deal than MDY.  Vernor, to my mind, is what caused MDY. One of the main difficulties about some of the previous literature (though I am not punching at the previous literature very hard in my paper) is that it tried to fix MDY without fixing Vernor.  And I don't think you can fix MDY without fixing Vernor. Once you are at MDY, you are already lost. You are already gone. You are falling down. You can try to take a parachute. But you are falling. It's too late. So Vernor is actually the worst of it.  

But here's the thing. It's the same panel. It's the same judge writing the opinion.  They were decided a few months apart, but they were argued together on same day!

CAH: Oh wow.  Really? Vernor and MDY were argued the same day?

GUY RUB: Yes. The Ninth Circuit released MDY about three months after Vernor, but they were argued on the same day. Same judge. Same panel. I can only speculate, but ...  

CAH: Wow. 

GUY RUB: Okay. now Vernor. The problem with Vernor is that it answers this very difficult question of what is a "sale," and who is an "owner," in a way that suggests that, in essence, whatever the copyright owner says goes. If it says that this is not a sale of a copy, it's a license, and that the end user is just a licensee, not an owner, then that is what it is. It is slightly more than that. There is a three part test.  [Vernor holds that "a software user is a licensee rather than an owner of a copy where the copyright owner (1) specifies that the user is granted a license; (2) significantly restricts the user's ability to transfer the software; and (3) imposes notable use restrictions."] 

But this is really just "magic words." You tell the company to use the magic words, and nothing else matters.  So if you are a copyright owner, why wouldn't you just use the those magic words then, and make this a license?  It does happen sometimes, that a software company doesn't write what they are supposed to, and it is held to be a sale. Or maybe the software company can't produce the license or someone didn't agree to it. But it's a tiny line of cases. Vernor is actually a really pretty bright-line rule. That is the problem with it!

CAH: And Vernor matters because it says the user is not an owner, this means they use at the company's whim. Right?

GUY RUB:  Yes. Backing up, whether a user gets software under a sale or a license is really important, because Congress ties a lot of defenses to the status of ownership.  These are the "first sale" defense and, for software, the "essential step" defense. Congress bound these defenses to the status of ownership.  

CAH: First sale is essential to an owner's rights, because it says the owner of a copy of a work, that's lawfully acquired, can do what it wants with the copy, such as re-selling it or disposing of it, short of doing things like copying or other 106 acts.  Say more about "essential step"?

GUY RUB: The "essential step" defense is central to a software user's rights. It deals with the problem that to use software you need to make a copy of it.  Use alone is not copyright infringement under 106.  This is different from in patent law, where "using" is one of the patentee's exclusive rights.  In copyright, it is not. You should not be infringing by simply "using" a copyrighted good. But the problem is that, when we have computers, the nature of the beast with computers is they make copies, all the time, in the background.  So you need to make copies simply to use a copyrighted computer program.  

CAH: Right, to "use" a book, you don't need to make a copy of it. But to "use" software you do.

GUY RUY:  So that is why Congress intervened in 1980. Congress, following CONTU, passed the Section 117 "essential step" defense to deal with this, saying “it is not an infringement for the owner of a copy of a computer program" to make copies if the copy “is created as an essential step in the utilization of the computer program.” What this strives to do is say that the steps that go into using software that involve automatically making copies are not infringing.  

CAH: So the problem with Vernor is that it eliminates this shield. Any uses of the software that involve copying, even if "essential" to using the program, are infringing if performed outside that license.

One quick possibly stupid question: Does the rule of Vernor apply to physical goods too, not just to software?  In other words, did the court assume you can use the "magic words" for non-digital goods like books? Meaning sellers of books could say owners of books are now licensees of the books, not owners, and so the first sale defense wouldn't apply...

GUY RUY: I don't know. There is nothing in the case that says it only applies to software. It was a software case, although it was not online software, it had to do with discs, so it was an "old software" case.  The software industry was pushing this argument for decades before, so clearly they had software on their minds. But there is nothing in the opinion that says the holding does not apply outside of the software world. Not a word there. 

I think this application of Vernor to software is extremely offensive, because the "essential step defense" is a software mechanism. Unlike the first sale doctrine, which applies to everything, the essential step defense is a mechanism that Congress created to deal with software and to balance the rights of software users and the rights of the software industry. It was specifically designed to prevent this precise situation, and now the Ninth Circuit says it does not apply in this precise situation? To say, if you only use the magic words, then the essential step defense goes away? That is just crazy. Now everything is up to the contract. The statutory defenses are completely subject to the contact. If the contract says it is not a sale, it is not a sale. 

And what is weird is that they call it a "license," because you do not usually "license" physical copyright goods like books. So it is bizarre. The court bends over backward to create this legal structure just to allow the software industry to avoid the clear intent of Congress.  How can you say essential step de facto does not apply to software? It is all about software.

CAH:  So Vernor is step one in the chain of reasoning.  Let's talk about MDY.  Same day, same panel, same judge, like you said. MDY builds on Vernor, because it answers the question of, okay, if a works is deemed only licensed and not owned, when does the licensor/copyright owner get a copyright infringement remedy versus a contract remedy? How does the MDY court answer this question?  And what is your criticism of this approach? 

GUY RUB:  The facts of MDY are that Blizzard [maker of the World of Warcraft (WoW) video game] wanted to go after the maker of a bot that players were using to cheat in WoW. To go after the bot maker, MDY, Blizzard needs copyright for the contributory infringement claim, like we discussed. Now the users of the bot are infringing just by playing the game with the bot in breach of the end user agreement they signed, because playing means making copies, and because the users are not an owner, they don’t get the essential step defense. That is Vernor. So that would make MDY a contributory infringer, because it's selling the bot to the players knowing they are cheating.

The Ninth Circuit does not like this.  The court thinks this gives copyright owners too much control. And I say: "Yes, you you gave them that control in Vernor! I know it's too much. You did it!" But the Ninth Circuit tries to fix it in MDY with the "nexus" test.  

CAH: And what is the nexus test, in theory?

GUY RUB: The MDY nexus test says that users who breach a license while using the software are liable for copyright infringement, but only if there is "a nexus between the condition and the licensor's exclusive rights of copyright."  The Ninth Circuit decided there was no copyright infringement here because there was no "nexus" between the breached license provision (the user promises “not to create or use bots"), and any of the copyright owners exclusive rights under 106. The court says using a bot while playing "does not alter or copy WoW software..." 

CAH: Sorry, backing up.  I thought the rule is just that the court has to decide whether the breached provision is a "condition" or a "covenant" of the license. If the breached provision was a mere "covenant," the plaintiff is entitled to sue only for a breach of contract. If the provision was a "condition" for the license, the licensor will be able to sue for copyright infringement.  I thought that actually made sense. I wrote a really bad paper on this in law school, back in 2010 when it just came out, and I argued the condition-covenant thing did make sense, because a covenant is just a promise, whereas a condition is integral to whether obligations are triggered under the contract. So if someone breaches a condition of the license by using the game while cheating, they have no license to use the game at all. They are liable for copyright infringement when they make those unlicensed automatic, essential copies.  That part makes some sense to me, or it used to...  

So how does "nexus" even fit in here? Isn't the real question just: is it a condition or is it a covenant?

GUY RUB: The covenant versus condition distinction is actually well established in contract law. That is not necessarily problematic. But to figure out whether something is a condition versus a covenant the court in MDY says there has to be a "nexus" between the breached provision and a statutory copyright exclusive right, like copying or making a derivative work.  

This is where the wheels come off.  "Nexus" is not part of contract law. The Ninth Circuit made it up out of thin air. What the court thinks it is doing is saying, okay, so the users are licensees, and they breached the license, but this can't mean Blizzard gets a copyright claim. So the court is saying, this is only a "condition" of the license if there is a "nexus" to copyright exclusive rights.  Once you pass that "nexus" test, then you can get a copyright claim. However, if there is no nexus, the provision is just a covenant, and you only get a contract claim.

CAH: Ah, okay. So, your interpretation of the court's opinion is that the "nexus" requirement is what the breached provision must meet to be considered a "condition" at all... It's confusingly worded. The Ninth Circuit just says: "[F]or a licensee's violation of a contract to constitute copyright infringement, there must be a nexus between the condition and the licensor's exclusive rights of copyright."  But you're saying the "nexus" finding is central. There has to be a nexus to a copyright exclusive right.  

Does that make sense?

GUY RUB: Well, no. Because the users did, in fact, create temporary copies upon using the game. Copying is an exclusive right.  Using means making potentially-infringing copies because of Vernor.  So even though they were not altering the software, or making "non-essential" copies, they were making copies that were outside the license.  So it's almost like the court is inventing a test specifically to avoid the result it had generated in Vernor.   

CAH: So how does the court explain this then, why there is not a nexus?

GUY RUB: The court seems to suggest there is no nexus because the bot itself does not create more copies. I do not fully understand what the court was trying to do with that. The bot does not make more copies. That is true. But the users are making copies while using the bot... 

CAH: The users are making thousands of copies every second. And there is no essential step defense. That is gone. So I am still not getting it. The provision has to be connected to an act that is covered under copyright, such as copying. The provision says no cheating. The users are cheating while making copies, which are infringing copies because the court said so in Vernor. 

GUY RUB: I think the court is thinking, does the word "bot" sound like copyright to you? No, so no nexus. But if the contract had said, "you cannot make copies while using a bot," then would there be a nexus? Or if the contract said, "as soon as you use a bot, your license to make copies expires?"  

CAH: Oh, I get it, so the rule is "nexus" means something that sounds "copyrighty."  At a high level I think I get what nexus means. If a license says, "you can make 6000 copies of our book," and defendant makes 10,000 copies, or doesn't pay for the copies, that seems like a nexus, that seems like copyright infringement. But if a license says, "you can make copies of our book, but you can't be sitting on the toilet while you're doing it..." that doesn't sound like copyright infringement.

GUY RUB: Well it turns out we do not know what sounds copyrighty.  The court used the word "nexus" in an informal way, and I do not know what the word means. I never understood it.  When I started this project, I believed in good faith it meant something. I thought that if I just read the caselaw I would figure out what it means. Now I see, no one knows what it means. Courts do not know what it means.  Apparently in the last ten years since the decision came out, no one understood it.  The judges say: "I think there is a nexus," or "I think there is no nexus"... Why?  They don't say.

CAH: OK, so the nexus test is gobbledygook. That's what my professor said about my paper in law school. And I was like: but it's gobbledygook!  So if it's gobbledygook then how do courts really decide whether something is copyright infringement or breach of contract? What are they doing if the nexus rule does not actually work?

GUY RUB: It is chaos. I looked at a decade of caselaw after MDY. 

[In the paper, Guy found 44 opinions that used MDY to try and decide when a breach of a license agreement triggers copyright liability. He finds that they were inconsistent in outcome, and that none defines "nexus" or explains what it means. They may quote MDY, but then they basically ignore it, or they even come out the opposite from what MDY said.  For example, in Ticketmaster v. Prestige Entm't, 315 F. Supp. 3d 1147 (C.D. Cal. 2018), the defendant used licensed code and breached promise not to use bots. The defendant argued this was a mere covenant and not a condition, like in MDY, so should just be a breach of contract. The court held it was copyright infringement, because the bot developers engaged in copying, an exclusive right. The court distinguished MDY, saying there were policy concerns there that were not present in this case. "That’s not true," Guy writes. The policy concerns were "exactly the same..."]

GUY RUB: One particularly inconsistent line of cases is about whether a failure to pay royalties is just a breach of contract or also copyright infringement. The plaintiff says, you can make copies, but only for an agreed-upon payment amount. The defendant makes copies but fails to pay the amount.  A little more than half of the cases say this is only a breach of contract. But the others say it is copyright infringement.

CAH: Wait, to me those failure-to-pay-royalty cases sound like copyright infringement. The defendant has a license to make copies for a certain price, and then it makes copies without meeting the price. The price is a condition, right? And there's a "nexus," if you want, because the license provision says "you can make copies, but only at this price," and this is clearly connected to the exclusive right to copy. 

GUY RUB: Well, in the majority of those cases, I think it was seven or eight, the courts held the duty to pay was a mere covenant, so no copyright infringement. A minority of the courts said what the court in MDY suggested is the right answer, that a violation of the duty to pay is copyright infringement. [The MDY case said in footnote 4: "A licensee arguably may commit copyright infringement by continuing to use the licensed work while failing to make required payments, even though a failure to make payments otherwise lacks a nexus to the licensor's exclusive statutory rights. We view payment as sui generis, however, because of the distinct nexus between payment and all commercial copyright licenses, not just those concerning software."]

CAH:  Yes, exactly, in the paper you discuss all these cases after MDY that go the opposite way from MDY.  I especially liked the case about the gown. Because I was thinking about how your paper applies to licenses that require attribution. Those seem like qualitatively different situations from the video game cheating scenario. To me, it seems like classic copyright infringement, to publish an article without attribution, when you signed a license that requires attribution. So the gown case agrees with that!  

[In the paper, Guy discusses the case, Etemadi v. Metro. Fashion Week, No. CV 17-1549 PA (GJSX), 2017 WL 5592901, (C.D. Cal. June 21, 2017). The copyright owner says: "you can display the gown, so long as you attribute me."  And the defendant makes a public display of the gown without attribution in breach of that oral license. This is a like MDY in a way, even though it is not software. There is a license agreement and an action that triggers a 106 right, the public display of the gown, and the licensee did the display outside the scope of the license. Yet the court says this is copyright infringement. This is the opposite of what the court held in MDY.]  

GUY RUB: I originally was going to call the paper "A Story of Legal Epicycles." Epicycles is a concept from astronomy. The Aristotelean astronomers thought Earth was at the center of the universe. But then people started to look at the sky, and they realized the stars and planets were not where they were supposed to be. So they invented new motions called "epicycles" to describe the planetary orbits in a way that made them fit.  But then more observations happened, and people said, "well, it doesn't even look like that, the stars are still not where they are supposed to be..." So the astronomers added more epicycles, and more and more... They were trying to fix the latest issue without fixing the underlying problem. The underlying problem is that you put the Earth in the wrong place. If you put the Sun in the right place, it just fixes itself.

So, yes, you can add another rule and another rule to solve Vernor. You can make up MDY's rule. And you can add more rules to deal with books, another for video games. We can have this complex system which would be totally unpredictable. Instead of fixing the underlying problem they invent epicycles. So I appreciate what the Ninth Circuit was trying to do. They created the mess in Vernor, and now, in MDY, they are trying to fix it. But the solution is worse. There is just layer upon layer of fixing it. No one can predict what might happen when a dispute is litigated.

CAH: So what is your solution?  I assume Plan A is to overrule Vernor? But do you have a Plan B?

GUY RUB: If Vernor were infallible, and we could not change it, then MDY would be maybe a good attempt to resolve it, maybe a Plan B. But Plan A is to overrule Vernor.

CAH: But there are two problems, right, not one? The Vernor problem is: Is a user of software an owner who gets an "essential step" defense, or a licensee who doesn't. Without Vernor, things would be easier in a sense. The user of software would just be an owner, they would get the "essential step" defense.  If they signed an end user agreement this would generally just be a breach of contract only, unless the user goes on to do some other act in Section 106 like making non-essential-step copies or derivative works or public displays. So true, without Vernor, the defendant easily wins in MDY, since there's no exclusive right violated, now that making the automatic copies is shielded, assuming using the bot is not making a derivative work.

But there is also the second MDY problem, right? You still have to answer the question: Is the breached provision a condition or a covenant? I agree that Vernor changes the game. But Vernor is not the only thing that lets copyright owners "customize" what is allowed and what is not, right? Copyright owners can still have users, publishers, whoever, sign contracts that say: "You can make copies, but you need to satisfy X condition" or "You can make copies, but you cannot do Y." Then the user makes copies, and they don't satisfy X condition, or they do Y. Overruling Vernor does not help with that situation.  We are dealing with owners of a copy, sure, even if it's software. But those owners want to make more copies or do other acts that otherwise would infringe, like a public performance. So you still have to answer the "condition" versus "covenant" problem, when there is an infringing act outside the scope of the license, right? We still need to know: is this breach of contract or is it copyright infringement?

GUY RUB: Yes. And in my mind this is just a question of contract law.  Take this example: an agreement says that "you have authority to translate my book into language X" and you translate it to language Y. Translating the book entails making a derivative work. Now we need to decide if this is infringement. It should be. You took an infringing act outside the scope of the license. You signed that agreement. You agreed to it. You do not have permission to translate. If you do it anyway, that is copyright infringement.

But notice that these cases are not about a software user. This is a publisher, a sophisticated party to a contract.  They agreed to it. Assuming there is an infringing act somewhere, not covered by fair use, then I am okay with that being copyright infringement. But only because there is an underlying infringing act somewhere. This is not the video game situation where using means copying, and it's automatic. Sophisticated parties have a choice!

CAH: Ok, so the video game user is a very different situation, because all they are doing is playing using the bot. They are not making intentional copies. That is why we have essential step.   But surely there are situations where it is not so easy. The CC license example is interesting to me. You can make copies but with attribution only. You make copies without attribution...

GUY RUB: I am fine with that being copyright infringement although it depends how exactly the license is drafted. Those copyleft contracts typically say this is a condition for the license. And in that case that is infringement, so long as it is not a fair use, and of course you need to be making additional copies. It can't just be "use."

CAH: Ok, hmm, what about "you can make copies but you also have to say really good things about us."  Then you make copies, and you do not say good things about the company.  Copyright infringement?

GUY RUB: That is where you start to push it. I think in this case the question is how much fair use can play a role for one thing.

CAH: I don't see how fair use saves you. You do not have a fair use reason to make copies. It's plain old copying. You just are not saying good things about the company like you were supposed to.

GUY RUB: Fair enough. Well, there also need to be a valid contract. At some point the contract defenses will kick in. Public policy, maybe reflecting First Amendment values. So if the contract says, "if you give me your first born child you can make copies of this book," that is a contract law problem. The contract defenses will prevent that.  Otherwise, when the contract is enforceable, I am sort of fine with it, so long as we are talking about a publisher not a user. Again, users do not typically make copies.

CAH: Okay. So, to summarize, a video game company says in a EULA, "you cannot make copies while playing the game if you say bad things about the game, and this is a condition of your license." And a player plays, and they say terrible things about the game. That is not copyright infringement.  It's a condition, and there's a nexus, probably, whatever that is. But in your universe where Vernor is overruled, we do not even need to ask whether this is a condition or a covenant. This is just a user, they're not making copies, except as required to play, and that's allowed under essential step. 

GUY RUB: Right. Although in this example even under contract law there would probably be an issue. For copyright, there should not even be a question.

CAH: Do you have any parting words?

GUY RUB: Vernor gives too much power to owners. MDY is an epicycle that tried to fix that problem. It does not work. Many copyright professors like MDY because it narrows Vernor, but my view is that the court reached the right outcome for completely nonsensical reasons. Is there a nexus or not a nexus? No one knows. I don't know. The plaintiff should have lost because the court should have overruled Vernor.  The software users were owners entitled to the essential step defense. There was no copyright infringement.

CAH:  Well on that note, Guy, thank you so much for taking the time to do this interview. It was extremely illuminating. I still do not feel like I understand what "nexus" means any better than I did in 2010. But I feel vindicated in my confusion, because the courts don't understand, and you have now explained, in a way I do understand, what is going on. 

I look forward to the next one! 

Lucy Xiaolu Wang on the Medicines Patent Pool

Patent pools are agreements by multiple patent owners to license related patents for a fixed price. The net welfare effect of patent pools is theoretically ambiguous: they can reduce numerous transaction costs, but they also can impose anti-competitive costs (due to collusive price-fixing) and costs to future innovation (due to terms requiring pool members to license future technologies back to the pool). In prior posts, I've described work by Ryan Lampe and Petra Moser suggesting that the first U.S. patent pool—on sewing machine technologies—deterred innovation, and work by Rob Merges and Mike Mattioli suggesting that the savings from two high tech pools are enormous, and that those concerned with pools thus have a high burden to show that the costs outweigh these benefits. More recently, Mattioli has reviewed the complex empirical literature on patent pools.

Economics Ph.D. student Lucy Xiaolu Wang has a very interesting new paper to add to this literature, which I believe is the first empirical study of a biomedical patent pool: Global Drug Diffusion and Innovation with a Patent Pool: The Case of HIV Drug Cocktails. Wang examines the Medicines Patent Pool (MPP), a UN-backed nonprofit that bundles patents for HIV drugs and other medicines and licenses these patents for generic sales in developing countries, with rates that are typically no more than 5% of revenues. For many diseases, including HIV/AIDS, the standard treatment requires daily consumption of multiple compounds owned by different firms with numerous patents. Such situations can benefit from a patent pool for the diffusion of drugs and the creation of single-pill once-daily drug cocktails. She uses a difference-in-differences method to study the effect of the MPP on both static and dynamic welfare and finds enormous social benefits.

On static welfare, she concludes that the MPP increases generic drug purchases in developing countries. She uses "the arguably exogenous variation in the timing of when a drug is included in the pool"—which "is not determined by demand side factors such as HIV prevalence and death rates"—to conclude that adding a drug to the MPP for a given country "increases generic drug share by about seven percentage points in that country." She reports that the results are stronger in countries where drugs are patented (with patent thickets) and are robust to alternative specifications or definitions of counterfactual groups.

On dynamic welfare, Wang concludes that the MPP increases follow-on innovation. "Once a compound enters the pool, new clinical trials increase for drugs that include the compound and more firms participate in these trials," resulting in more new drug product approvals, particularly generic versions of single-pill drug cocktails. And this increase in R&D comes from both pool insiders and outsiders. She finds that outsiders primarily increase innovation for new and better uses of existing compounds, and insiders reallocate resources for pre-market trials and new compound development.

Under these estimations, the net social benefit is substantial. Wang uses a simple structural model and estimates that the MPP for licensing HIV drug patents increased consumer surplus by $700–1400 million and producer surplus by up to $181 million over the first seven years of its establishment, greatly exceeding the pool's $33 million total operating cost over the same period. Of course, estimating counterfactuals from natural experiments is always fraught with challenges. But as an initial effort to understand the net benefits and costs of the MPP, this seems like an important contribution that is worth the attention of legal scholars working in the patent pool area.

It's Hard Out There for a Commons

I just finished reading a fascinating draft article about the Eco-Patent Commons, a commons where about 13 companies put in a little fewer than 100 patents that could be used by any third party. A commons differs from cross-licensing or other pools in a couple of important ways. First, the owner must still maintain the patent (OK, that's common to licensing, but different from the public domain). Second, anyone, not just members of the commons, can use the patents (which is common to the public domain, but different from licensing).

The hope for the commons was that it would aid in diffusion of green patents, but it was not to be. The draft by Jorge Contreras (Utah Law), Bronwyn Hall (Berkeley Econ), and Christian Helmers (Santa Clara Econ) is called Green Technology Diffusion: A Post-Mortem Analysis of the Eco-Patent Commons. A draft is on SSRN. Here is the abstract:

We revisit the effect of the “Eco-Patent Commons” (EcoPC) on the diffusion of patented environmentally friendly technologies following its discontinuation in 2016, using both participant survey and data analytic evidence. Established in January 2008 by several large multinational companies, the not-for-profit initiative provided royalty-free access to 248 patents covering 94 “green” inventions. Hall and Helmers (2013) suggested that the patents pledged to the commons had the potential to encourage the diffusion of valuable environmentally friendly technologies. Our updated results now show that the commons did not increase the diffusion of pledged inventions, and that the EcoPC suffered from several structural and organizational issues. Our findings have implications for the effectiveness of patent commons in enabling the diffusion of patented technologies more broadly.

The findings were pretty bleak. In short, the patents were cited less than a set of matching patents, and many of them were allowed to lapse (which implies lack of value). Their survey-type data also showed a lack of importance/diffusion.

What I really love about this paper, though, is that there's an interpretation for everybody in it. For the "we need strong rights" group, this failure is evidence of the tragedy of the commons. If nobody has the right to fully profit on the inventions, then nobody will do so, and the commons will go fallow.

But for the "we don't need strong rights" group, this failure is evidence that the supposedly important patents were weak, and that it was better to essentially make these public domain than to have after the fact lawsuits.

For the "patents are useless" group, this failure shows that nobody reads patents anyway, and so they fail in their essential purpose: providing information as a quid pro quo for exclusivity.

And for the middle ground folks, you have the conclusions in the study. Maybe some commons can work, but you have to be careful about how you set them up, and this one had procedural and substantive failings that doomed the patents to go unused.

I don't know the answer, but I think cases studies like this are helpful for better understanding how patents do and do not disseminate information, as well as learning how to better structure patent pools.

Equitable Servitudes and Post-Sale Restrictions

I have continued to find the issue of post-sale restrictions vexing. On the one hand, I think that there are sound economic reasons for them. On the other hand, I really don't like them, especially when they limit what should otherwise be reasonable and free activities.

The Supreme Court's recent cases in this area have made it more difficult to enforce such restrictions, but they have done so in a way that leaves open the possibility that some restrictions might apply while also not giving much guidance about when.

A recent article takes this on. Tim Scott (King & Spalding) has published The Availability of Post-Sale Contractual Restrictions in the Wake of Impression Products, Inc. v. Lexmark, 581 U.S. 1523 (2017) in Les Nouvelle. It is available on SSRN as well, and the abstract is below. I write about this article in part because the abstract is uninspiring when compared to the high quality analysis in the article. Don't be fooled.

In Impression Products, Inc. v. Lexmark International Inc., 581 U.S. ___, 137 S. Ct. 1523 (2017), the United States Supreme Court reaffirmed the patent exhaustion rule; i.e., patent rights are exhausted upon the first sale of the patented item such that the patentee has no rights to impose any post-sale conditions or limitations on the use of the product, at least under the patent laws. Id. at 1529. In doing so, the Court left open the question of whether such conditions or limitations could be imposed as a matter of contract law. Thus, the “restrictions in Lexmark’s contracts with its customers may have been clear and enforceable under contract law, but they do not entitle Lexmark to retain patent rights in an item that it has elected to sell.” Id. at 1531 (emphasis added). In summarizing its decision in Quanta Computer, Inc. v. LG Electronics, Inc., 533 U.S. 617 (2008) ruling that the sale of computer components exhausted the plaintiffs patent rights in those components, the Court noted that it reached that conclusion “without so much as mentioning the lawfulness of the contract.” Lexmark, 137 S. Ct. at 1533. And it later summarized its holding by stating that “whatever rights Lexmark retained are a matter of the contracts with its purchasers, not the patent law.” Id. (emphasis added).

This is my kind of article: it's short, it gets right to business, and it is thorough despite its brevity. The article introduces the cases, discusses contract v. property rights, discusses equitable servitudes, surveys the literature on equitable servitudes on personal property (including pros and cons), proposes alternatives to equitable servitudes (and points to critiques), and then discusses how the alternatives might have applied to Lexmark's activities in Impression Products.

Not bad for eight journal pages. This is recommended reading for anyone who wants a quick background on the state of post-sale restrictions.