Sue First, Negotiate Later

Just a brief post this week, as I have a perfect storm of non-work related happenings. So, I'll just say that I'm please to announce that my draft article Sue First, Negotiate Later will be published by the Arizona Law Review. The draft is on SSRN, and the longish abstract is below. I may blog about this in more detail in the future, but this is an introduction:

One of the more curious features of patent law is that patents can be challenged by anyone worried about being sued. This challenge right allows potential defendants to file a declaratory relief lawsuit in their local federal district court, seeking a judgment that a patent is invalid or noninfringed. To avoid this home-court advantage, patent owners may file a patent infringement lawsuit first and, by doing so, retain the case in the patent owner’s venue of choice. But there is an unfortunate side effect to such preemptive lawsuits: they escalate the dispute when the parties may want to instead settle for a license. Thus, policies that allow challenges are favored, but they are tempered by escalation caused by preemptive lawsuits. To the extent a particular challenge rule leads to more preemptive lawsuits, it might be disfavored.

This article tests one such important challenge rule. In MedImmune v. Genentech, the U.S. Supreme Court made it easier for a potential defendant to sue first. Whereas the prior rule required threat of immediate injury, the Supreme Court made clear that any case or controversy would allow a challenger to file a declaratory relief action. This ruling had a real practical effect, allowing recipients of letters that boiled down to, “Let’s discuss my patent,” to file a lawsuit when they could not before.

This was supposed to help alleged infringers, but not everyone was convinced. Many observers at the time predicted that the new rule would lead to more preemptive infringement lawsuits filed by patent holders. They would sue first and negotiate later rather than open themselves up to a challenge by sending a demand letter. Further, most who predicted this behavior—including parties to lawsuits themselves—thought that non-practicing entities would lead the charge. Indeed, as time passed, most reports were that this is what happened: that patent trolls uniquely were suing first and negotiating later. But to date, no study has empirically considered the effect of the MedImmune ruling to determine who filed preemptive lawsuits. This Article tests MedImmune’s unintended consequences. The answer matters: lawsuits are costly, and while “quickie” settlements may be relatively inexpensive, increased incentive to file challenges and preemptive infringement suits can lead to entrenchment instead of settlement.

Using a novel longitudinal dataset, this article considers whether MedImmune led to more preemptive infringement lawsuits by NPEs. It does so in three ways. First, it performs a differences-in-differences analysis to test whether case duration for the most active NPEs grew shorter after MedImmune. One would expect that preemptive suits would settle more quickly because they are proxies for quick settlement cases rather than signals of drawn out litigation. Second, it considers whether, other factors equal, the rate of short-lived case filings increased after MedImmune. That is, even if cases grew longer on average, the share of shorter cases should grow if there are more placeholders. Third, it considers whether plaintiffs themselves disclosed sending a demand letter prior to suing.

It turns out that the conventional wisdom is wrong. Not only did cases not grow shorter – cases with similar characteristics grew longer after MedImmune. Furthermore, NPEs were not the only ones who sued first and negotiated later. Instead, every type of plaintiff sent fewer demand letters, NPEs and product companies alike. If anything, the MedImmune experience shows that everyone likes to sue in their preferred venue. As a matter of policy, it means that efforts to dissuade filing lawsuits should be broadly targeted, because all may be susceptible

Jiarui Liu on the Dominance and Ambiguity of Transformative Use

The Stanford Technology Law Review just published an interesting new copyright article, An Empirical Study of Transformative Use in Copyright Law by Prof. Jiarui Liu. Here is the abstract:

This article presents an empirical study based on all reported transformative use decisions in U.S. copyright history through January 1, 2017. Since Judge Leval coined the doctrine of transformative use in 1990, it has been gradually approaching total dominance in fair use jurisprudence, involved in 90% of all fair use decisions in recent years. More importantly, of all the dispositive decisions that upheld transformative use, 94% eventually led to a finding of fair use. The controlling effect is nowhere more evident than in the context of the four-factor test: A finding of transformative use overrides findings of commercial purpose and bad faith under factor one, renders irrelevant the issue of whether the original work is unpublished or creative under factor two, stretches the extent of copying permitted under factor three towards 100% verbatim reproduction, and precludes the evidence on damage to the primary or derivative market under factor four even though there exists a well-functioning market for the use.

Although transformative use has harmonized fair use rhetoric, it falls short of streamlining fair use practice or increasing its predictability. Courts diverge widely on the meaning of transformative use. They have upheld the doctrine in favor of defendants upon a finding of physical transformation, purposive transformation, or neither. Transformative use is also prone to the problem of the slippery slope: courts start conservatively on uncontroversial cases and then extend the doctrine bit by bit to fact patterns increasingly remote from the original context.

This article, albeit being descriptive in nature, does have a normative connotation. Courts welcome transformative use not despite, but due to, its ambiguity, which is a flexible way to implement their intuitive judgments yet maintain the impression of stare decisis. However, the rhetorical harmony conceals the differences between a wide variety of policy concerns in dissimilar cases, invites casual references to precedents from factually unrelated contexts, and substitutes a mechanical exercise of physical or purposive transformation for an in-depth policy analysis that may provide clearer guidance for future cases.

This article builds on and extends prior empirical work in this area, such as Barton Beebe's study of fair use decisions from 1978 to 2005. And it provides a nice mix of interesting new empirical results and normative analysis that illustrates why fair use doctrine is (at least for me) quite challenging to teach. For example, Figure 1 illustrates how transformative use has cannibalized fair use doctrine since the 1994 Campbell v. Acuff-Rose decision endorsed its use:

Liu also examines data such as the win rate for transformative use over time, by circuit, and by subject matter. But I particularly like that Liu is not just counting cases, but also arguing that courts are using this doctrine as a substitute for in-depth policy analysis.

Does Administrative Patent Law Promote Innovation About Innovation?

I am at Texas Law today for a symposium on The Intersection of Administrative & IP Law, and my panel was asked to address the question: Does Administrative Patent Law Promote Innovation? I focused my remarks on a specific aspect of this: Does Administrative Patent Law Promote Innovation About Innovation? I think the short answer, at least right now, is "no."

There is a lot we don't know about the patent system. USPTO Regional Director Hope Shimabuku started her remarks today by saying that we know IP creates nearly 30 million jobs and adds $6.6 trillion to the U.S. economy each year, citing this USPTO report. But that's not what the report says. It looks at jobs and value from "IP-intensive industries," defined as ones with "IP-count to employment ratio is higher than the average for all industries considered." As the report acknowledges, it is unable to determine how much of these firms' performance is attributable to IP.

And the real answer is: we don't know. In an article I reviewed for Jotwell, economist Heidi Williams recently summarized: "we still have essentially no credible empirical evidence on the seemingly simple question of whether stronger patent rights—either longer patent terms or broader patent rights—encourage research investments." And even on smaller questions, the existing evidence base is weak.

As I explained in Patent Experimentalism, to make empirical progress we need some source of empirical variation. Economists often look for "natural experiments" with variation across time, across jurisdictions, or across similar technologies, and the closer that variation is to random, the easier it is to draw causal inferences. Of course, it's even better to have variation that is actually random, which is why I have joined other scholars in arguing for more use of randomized policy experiments.

The USPTO has a huge opportunity here to both improve the patent system and help address the key administrative law challenge of encouraging accurate and consistent decisions by a decentralized bureaucracy. There are many questions the agency could help answer using more randomization, as I discuss in Patent Experimentalism. During the panel today, I noted two potential areas: experimenting with the time spent examining a given patent (see this great forthcoming article by Michel Frakes and Melissa Wasserman) and with the possibility that examiner bias affects the gender gap in patenting (which fits within the agency's recent mandate from Congress). I noted ways that each could be designed as opt-in progress to encourage buy-in from applicants and from examiners.

But my main point was not that the USPTO should adopt one of these particular experiments—it was that the agency should study something in a way that allows us to draw rigorous inferences. Failing to do so seems like a tremendous missed opportunity.

Using Insurance to Deter Lawsuits

The conventional wisdom (my anecdotal experience, anyway) is that the availability of insurance fuels lawsuits. People that otherwise might not sue would use litigation to access insurance funds. I'm sure there's a literature on this. But most insurance covers both defense and indemnity - that is, litigation costs and settlements. But what if the insurance covered the defense and not any settlement costs? Would that serve as a disincentive to bring suit? It surely would change the litigation dynamic.

In The Effect of Patent Litigation Insurance: Theory and Evidence from NPEs, Bernhard Ganglmair (University of Mannheim - Economics), Christian Helmers (Santa Clara - Economics), Brian J. Love (Santa Clara - Law) explore this question with respect to NPE patent litigation insurance. The draft is on SSRN, and the abstract is here:

We analyze the extent to which private defensive litigation insurance deters patent assertion by non-practicing entities (NPEs). We do so by studying the effect that a patent-specific defensive insurance product, offered by a leading litigation insurer, had on the litigation behavior of insured patents’ owners, all of which are NPEs. We first model the impact of defensive litigation insurance on the behavior of patent enforcers and accused infringers. Assuming that a firm’s purchase of insurance is not observed by patent enforcers, we show that the mere availability of defense litigation insurance can have an effect on how often patent enforcers will assert their patents. Next, we empirically evaluate the insurance policy’s effect on the behavior of owners of insured patents by comparing their subsequent assertion of insured patents with their subsequent assertion of their other patents not included in the policy. We additionally compare the assertion of insured patents with patents held by other NPEs with portfolios that were entirely excluded from the insurance product. Our findings suggest that the introduction of this insurance policy had a large, negative effect on the likelihood that a patent included in the policy was subsequently asserted, and our results are robust across different control groups. Our findings also have importance for ongoing debates on the need to reform the U.S. and European patent systems, and suggest that market-based mechanisms can deter so-called “patent trolling.”

On reading the abstract, I was skeptical. After all, there are a bunch of reasons why more firms would defend against NPEs , why NPEs would be less likely to assert, and so forth. But the interesting dynamics of the patent litigation insurance market have me more convinced. Apparently, the insurance didn't cover any old lawsuit; instead, only specific patents were covered. So, the authors were able to look at the differences between firms asserting covered patents, firms that held both covered and non-covered patents, and firms that had no covered patents. Because each of these firms should be equally affected by background law changes, the differences should be limited to the role of insurance.

And that's what they find, unsurprisingly. Assertions of insured patents went down as compared to uninsured patents, and those cases were less likely to settle -- even with the same plaintiff. My one concern about this finding is that patents targeted for insurance may have been weaker in the first place (hence the willingness to insure), and thus there is self-selection. The paper presents some data on the different patents in order to quell this concern, but if there is a methodological challenge, it is here.

This is a longish paper for an empirical paper, in part because they develop a complex game theory model of the insurance purchasing, patent assertion, and patent defense. It is interesting and worth a read.

Foreign Meaning Matters: Brauneis and Moerland on Trademark's Doctrine of Foreign Equivalents

I was enjoying some siggi's® yogurt, and noticed, just below the trademark name siggi's®, an interesting piece of trivia: "skyr, that's Icelandic for thick yogurt!" You learn something new every day.

Robert Brauneis and Anke Moerland's recent article argues that it would not be good policy to allow the company that distributes siggi's ® yogurt to trademark the name SKYR for yogurt in the United States, even though most people in the United States do not currently know what the word "skyr" means. In short, they argue that when reviewing trademarks for purposes of distinctiveness, the U.S. Patent & Trademark Office (USPTO) and the courts should translate foreign terms that are generic or merely descriptive in their home country, because allowing such marks would cause unexpected harms for competition.

This is a fascinating paper that warrants serious thinking, and perhaps re-thinking, of how trademark law currently treats foreign terms.

What's the harm, we might ask? If a US company wants to sell thick yogurt under the trademark SKYR, and virtually no one in the US knows what SKYR means, surely this should be classified as fanciful (inherently distinctive) and receive strong protection.  At least this would be the answer provided by a typical "doctrine of foreign equivalents" analysis.

The Doctrine of Foreign Equivalents

Trademark law's so-called doctrine of foreign equivalents holds, at least in the Federal Circuit's interpretation, that the meaning of a mark in a non-English language should be considered in assessing distinctiveness, but only "where the ordinary American consumer," including English-only speakers and people proficient in the non-English language present in America, "would stop and translate the mark into English[.]"  See In re Spirits Intern., NV, 563 F. 3d 1347 (Fed. Cir. 2009). Obviously, there is usually going to be at least one native speaker in America who knows the meaning of the foreign term; but there has to be an "appreciable number" of native speaking consumers for their understanding to count. Id. (discussing the doctrine of foreign equivalents in the context of denying registration for geographically deceptively misdescriptive marks); see also Palm Bay Imports v. Veuve Clicquot Ponsardin, 396 F. 3d 1369 (Fed. Cir. 2005).

Therefore, so long as an ordinary American purchaser, including the 40,000 or so Icelanders residing in the United States, would not "ordinarily be expected to translate" skyr into "thick yogurt" there is no problem in granting a US yogurt company trademark rights to sell Skyr® yogurt.

Brauneis and Moerland's Critique

Not so fast, Professors Brauneis and Moerland argue.  Allowing trademarking of such terms can in fact have significant anticompetitive effects by keeping companies from marketing products across borders that incorporate the most efficient words to describe them.

As can be seen from the article's title, "Monopolizing Matratzen in Malaga: The Mistreatment of Distinctiveness of Foreign Terms in EU and US Trademark Law," this article is admirably international, addressing not just US trademark law, but EU law, and cites a plethora of court opinions from across nations. This post, in contrast, is directed primarily at the US situation, and comes from a concededly US-centric perspective.

The authors' central argument is that, in most instances, when someone tries to trademark a foreign term that is generic in the home language (or merely descriptive in the home language), the term should be translated from the foreign language into the domestic language in determining distinctiveness for purposes of protectability, regardless of whether consumers in the domestic country where protection is sought understand the term's meaning.  (Descriptive marks are important too, since under 15 U.S.C. § 1052(e)(1) and (f), they can only be protected with secondary meaning.)

In making their argument, the authors divide potentially trademarked generic/descriptive foreign terms into three categories, which move from least to most controversial.

Category 1: Terms Where Domestic Consumers Know What the Foreign Term Means

The first category of foreign terms they identify consists of foreign terms domestic consumers are likely to understand as the "common word in the foreign language" for a good or service, or as being merely "descriptive" of the good or service. (Brauneis and Moerland, at 3).

This is an uncontroversial category. Take the term BANK. Clearly "bank" is a generic term for "bank" in the English language. But it's also basically a generic term for bank in every other language, since the Spanish, French, German, and even the Japanese have very similar terms for bank. Thus, no one in any of these countries should be allowed to trademark the term "bank" (alone) for a bank. It is generic across the board. The result, and the caveat, is that companies like CITIBANK can use, and even trademark, the term "bank" within more distinctive composite marks, without worrying about likely confusion with a prior mark. Likewise, common French terms like sorbet (sorbet) or baguette (baguette) are so well known in the US, and presumably in other countries, that they would be generic regardless.

Courts in various countries have uniformly refused protection for generic or insufficiently distinctive terms when domestic consumers will quickly stop and translate. (3)

Category 2: "Proto-generic terms"

This is my favorite category, and I completely agree with the authors' conclusions here that what they call "proto-generic" terms should be translated from their foreign language into the domestic language.

The basic idea here is that sometimes "a foreign term may not yet be widely known in a particular market" because it has only recently been introduced into that market. For example, take SKYR.  The Greek yogurt craze got going in only around 2005 thanks to Chobani, and the Icelandic yogurt follow-on is even newer.  So (on concededly minimal evidence) I'd consider SKYR a “proto-generic” foreign term that will, in the next couple years, be recognizable to an appreciable number of US consumers if it's not already.*  Indeed, the authors point to a German case to this effect: SKYR should be translated from Icelandic because German consumers don't yet know its meaning but soon will given the movement of the yogurt market. (6).

* Incidentally, the company that distributes siggi's® operates under the trade name, The Icelandic Milk and Skyr Corporation, a private company based in New York.

The upshot is that, for this category, courts are wrong when they decline to translate from the foreign term, because there will be anticompetitive effects in the not-too-distant future if only one company can market goods under a soon-to-be-generic term. This rule is especially appropriate given that trademark law tends to be forward-looking, assessing not just how the competitive landscape looks today, but how it will look in the future. Examples of this include where courts assess "the likelihood that the prior owner will bridge the gap" in the confusion analysis, and also the Lanham Act's willingness to find that a term is abandoned once it becomes generic, even if it was not born that way.

The authors identify a few opinions taking the erroneous view, but they note that many courts already get this category right. They cite several cases where courts held "proto-generic" foreign terms, such as OTOKOYAMA for Japanese dry sake, should be translated even though US consumers may not quite yet know the meaning. (5-6).  This means a lot of these proto-generic foreign terms (or proto-descriptive terms lacking secondary meaning), will not be eligible.

The Federal Circuit is a more interesting case. (The Federal Circuit's view is important since it has jurisdiction over appeals from rejections by examiners and the Trademark Trial and Appeal Board.) To the extent the Federal Circuit's holding in In re Spirits, mentioned above, suggests that translating foreign terms is required in the distinctiveness analysis only if ordinary American consumers would currently stop and translate the mark into English, this would be wrong. That said, there is room for debate on precisely what the Federal Circuit held in In re Spirits.

The issue in In Re Spirits was whether the term MOSKOVSKAYA, which means "of or from Moscow," was primarily "geographically deceptively misdescriptive" when used to sell vodka not made or sold in Moscow, under 15 U.S.C. § 1052(e)(3). The Federal Circuit affirmed the finding that an "appreciable number" of "ordinary American purchasers," including hundreds of thousands of Russian speakers in the US, would currently understand MOSKOVSKAYA to mean "of or from Moscow." In re Spirits, 563 F.3d. at 1351-1352. The court did not directly hold on the issue of whether future conceptions of a term should be considered. But the court did go on to find MOSKOVSKAYA for vodka that is not made or sold in Moscow was nonetheless a valid trademark, because the deception must also be a "material" factor in consumers' purchasing decisions, and this requires that the "appreciable number" of native speakers also be a "substantial portion of the intended audience." Id. at 1353 (citing In re California Innovations, Inc., 329 F.3d 1334, 1341 (Fed. Cir. 2003).  In this case, the 706,000 Russian speakers only made up one-quarter of one percent of the whole US population in the market for vodka,  which was not "substantial" enough. Id.

In Professors Brauneis and Moerland's view, this holding effectively dispatches with the category of "proto-generic" foreign terms with respect to trademark distinctiveness analysis. If 706,000 people speak Russian in the US today and understand that MOSKOVSKAYA means "of or from Moscow," but this was not seen as sufficient, at least in the context of a § 1052(e)(3) rejection, this effectively says that the doctrine of foreign equivalents does not require a court to assess whether a foreign term whose meaning is known to only a few people now will nonetheless be known to many people in the near future. The authors note that at least "[o]ne scholar has argued that the reasoning of [In re Spirits] requires the distinctiveness of foreign words to be assessed solely on the basis of current domestic consumer understanding." (21) (citing Serge Krimnus, The Doctrine of Foreign Equivalents at Death’s Door, 12 N.C.J.L. & Tech. 159, 161 (2010)).

Another way to see this opinion, however, is that the Federal Circuit was only concerned about correcting how § 1052(e)(3) geographically deceptively misdescriptive rejections are performed, not about distinctiveness rules generally. So, for example, if someone applies to register SKYR (which is not a geographic term at all) for yogurt, this could still be rejected even under In re Spirits as "merely descriptive" without secondary meaning or "generic" for thick yogurt, despite that fact that only 40,000 people currently know the term, and 40,000 is a very small portion of the US yogurt market. The reason is that "materiality" is not at issue for ordinary, non-geographic "merely descriptive" or "primarily geographically descriptive" assessments under § 1052(e)(1) and § 1052(e)(2). So if the question is whether MOSKOVSKAYA is merely descriptive for a restaurant that sells Russian food, this might still be translated into Russian for purposes of assessing whether the term is "merely descriptive" or "primarily geographically descriptive" of the food offerings.

In any case, needless to say, if the Federal Circuit gets the opportunity to speak on this precise point of law in the future, the authors would presumably urge the Federal Circuit to hold that the doctrine of foreign equivalents can be used in distinctiveness assessments, even if ordinary American consumers don't currently know the term's meaning, so long as it is "reasonable to assume that that might be the case in the future." (7) (quoting a Court of Justice of the European Union (CJEU) opinion taking this stance).

Category 3: No One in the Domestic Country Knows What the Term Means, And They Probably Never Will 

The final category is the most controversial. This category consists of foreign generic or descriptive terms that consumers do not understand now, and that they will not likely understand in future.

I am initially skeptical. What is the problem if a German company wants to sell violins in the US using a trademark consisting of the German term for violin, GEIGE, if no one in the US knows what it means? Surely US consumers will just see GEIGE as a highly distinctive mark (fanciful) that identifies a source, not a product, making it protectable under the Lanham Act.  The Federal Circuit would apparently not translate such a term.  See, e.g., Palm Bay Imports, 396 F. 3d at 1377 ("When it is unlikely that an American buyer will translate the foreign mark and will take it as it is, then the doctrine of foreign equivalents will not be applied.").

The authors argue, however, that there is, in fact, a big problem here.  Their explanation is quite artful, and worth quoting in full.

[T]he Spanish-language term for “mattress” is “colchón,” which, unlike the English word, has no relation or similarity to the German term “Matratze.” If Spanish consumers are unlikely ever to make the jump from “colchón” to “Matratze,” how could protecting “Matratzen” as a trademark in Spain have any anticompetitive effects? 

The answer, in one sentence, is that producers that have developed composite-mark branding incorporating the foreign term [in this case, German producers] will find it difficult or impossible to use that branding in the domestic market [here, in Spain, where trademark rights are sought], and that is a significant limitation on competition that will also affect domestic consumers.

(Brauneis and Moerland, 8). The example they give of this problem is where a German company selling mattresses under the composite name MATRATZEN CONCORD wishes to sell into the Spanish market. The German company will be prevented from doing so due to the already-trademarked MATRATZE brand mattress. The Spanish trademark office was fine with it, because the term was highly distinctive to Spanish speaking consumers; but now German companies cannot sell into the US.

Here is an example of the problem the authors have identified that should resonate with yogurt lovers.  Imagine the USPTO permits a US company to obtain a trademark for the term SKYR for thick yogurt. Assume further that "ordinary US consumers" do not currently know what the term skyr means, and almost certainly never will, given the stark difference from the US term for "yogurt."

But then imagine that an Icelandic company wants to sell yogurt under a composite mark that incorporates the Icelandic term for yogurt. The one that I think of is FYNDIO SKYR ("fyndio appears to mean "funny" in Icelandic). Surely, the Icelandic company should be able to call its product, loosely, Funny Yogurt® and be able to sell its Funny Yogurt® in the US too, without changing its name. After all, composite marks incorporating generic terms are exceptionally common. (8-9) ("Consider, for example, CITIBANK, MASTERCARD, NESCAFÉ...and of course MATRATZEN CONCORD.").

But now that SKYR has been trademarked and branded in connection with a US yogurt company, American consumers will see FYNDIO SKYR not as meaning "funny yogurt," but as indicating a source. They will likely see an affiliation with the previously trademarked SKYR, or even believe the source of the two yogurts is the same.

Indeed, as the authors compellingly observe, the fact that US consumers do not know the generic meaning for the foreign term potentially creates even more of a problem than if they did know it, because now the term is highly distinctive! Indeed, "from [America consumers'] point of view, it is a fanciful, newly-coined word[,]" and this "will only increase the likelihood of confusion between two brand names that incorporate the word, and will thus increase the likelihood that the new entrant faces ... anticompetitive burdens." (10).

Note that the problem is not limited only to "composite marks" in the classic sense, like CITIBANK or SURGICENTER. Rather, the authors define "composite-mark branding" quite broadly to include "branding that incorporates two or more words," whether or not separated by a space, "in which at least one of the words is generic or descriptive in the language that the brand initially targets..." (8). This appears to sweep in marks that combine a generic term with a strong brand name, such as TOYOTA TORAKKU. (Torakku means "truck" in Japanese). True, the problem is likely to be more severe for composite marks that combine two generic terms than for marks that include an actual trademark term, since the presence of a strong brand name like Toyota will typically alleviate consumer confusion. But still, if a foreign generic term like TORAKKU becomes protected as a trademark in the United States, there could be anticompetitive problems in this scenario too, if Toyota tried to market a truck called something like TOYOTA TORAKKU.

Objections to Consider

I thought of several questions and objections to the authors' arguments, especially as respects their third category: generic foreign terms that no one in the US is ever going to stop and translate. Each was answered by the authors over email. I'll share each objection and response below.

My first question is, isn't the descriptive fair use defense going to resolve the issue? As effectuated under Lanham Act § 1115(b)(4), use of a term that is another firm’s trademark won’t be infringement if, in relevant part, "the use of the [term] charged to be an infringement is a use, otherwise than as a mark ...of a [term] which is descriptive of and used fairly and in good faith only to describe the goods or services of [defendant.]" Why won't this defense work in the CONCORD MATRATZEN case, where the German term for mattress is being used solely to describe the good sold, or in the FYNDIO SKYR case, where skyr is used only to describe the yogurt?

I am somewhat unsatisfied with the authors' explanation in their article of why descriptive fair use is insufficient to deal with the problem. The authors argue a mere defense is insufficient, because there is a "clear difference between having an acknowledged freedom to use a term ... and having to seek such a freedom through a descriptive fair use [defense,]" which entails "the legal costs of being involved in proceedings in a foreign and possibly unknown legal system." (16). But I feel like that can largely be said any time we give protection to merely descriptive terms that develop secondary meaning, such as FISH FRI. Descriptive fair use is a balancing act. On the one hand, a term that comes to indicate a source to US consumers, like MATRATZEN, will get protection, and on the other hand, a German defendant wishing to use a composite mark like CONCORD MATRATZEN in good faith merely to describe their mattress-related product gets a defense. Why can't this balancing be done in all these cases?

Professor Brauneis provided a pretty compelling response to this point via email, which I'll quote with permission, because it is exceptionally clear and has great images.

Under US law, at least, I think descriptive fair use would be very difficult to invoke as a defense in cases in which the defendant was using the word as part of a composite mark.  In those cases, the defendant will typically be presenting the descriptive or generic word with the same prominence, and in the same font, and so on, as the distinctive portion of the composite mark.  For example, the CAFÉ in NESCAFÉ is presented no differently than the NES: 

 

and in the MATRATZEN CONCORD logo, the word MATRATZEN is in the same rounded-off font as CONCORD, and is obviously coordinated with it, scaled so that the two words are exactly the same length:  

 

I think those uses would fare terribly in a descriptive fair use test.  As the Second Circuit has formulated it in, e.g., Kelly-Brown v. Winfrey, 717 F.3d 395 (2013), to mount a successful descriptive fair use defense, a defendant must prove that its use was made "(1) other than as a mark, (2) in a descriptive sense, and (3) in good faith." Id. at 308. As for what "use as a mark" means, the court asks "whether [the defendant was"] using the term 'as a symbol to attract public attention,'" or in an alternative formulation, whether the defendant was "attempting to build an association with consumers between the [term and the defendant]."  Id. 

Suppose that other companies had managed to register CAFÉ and MATRATZEN as trademarks.  If I were counseling Nescafé or Matratzen Concord, I would warn them that if they used the logos I've pasted in above, they would be running a very high risk of failing the descriptive fair use test, because a court would likely find that they are using the words "café" and "Matratzen" to attract public attention and to build consumer association between those words and the defendant as a source.  First, the words are integrated with the distinctive part of the defendant's mark. Second, they are in a foreign language, and if the defendant's purpose was description, there are perfectly good domestic language words (in English, coffee and mattress) for the defendant to use.  Thus, I'm not as optimistic as you are that descriptive fair use would save the day.

My second objection is actually more of a question. I was uncertain after reading the article whether the authors are also concerned about domestic companies, or just about foreign companies. For example, what if Chobani, a New York-based company founded by a Greek individual, wants to sell Chobani Skyr®? Chobani means "shepherd" in Greek and is a conceptually as well as commercially strong mark. But this composite mark would present a problem for Chobani if SKYR has already been trademarked as a "fanciful" term by another yogurt company. Should Chobani have less of a right to use a generic Icelandic term than an Icelandic company?

Here is Professor Brauneis' response, which I think is a crucial clarification to the scope of their argument.

If the word is "proto-generic," such as (I would argue) SKYR is, then I think domestic companies have to have the same right.  An employee from a US or a German dairy company might have travelled to Iceland, sampled the local product, and decided to introduce it in the US or Germany, and it shouldn't matter that the company involved is not based in Iceland.

This means the authors think Chobani, the New York-based company, should be able to sell Chobani Skyr® in the US, just as much as an Icelandic company should be able to sell Findio Skyr® in the US. Their argument that we should avoid letting companies get trademarks on the generic foreign term is just as much protective of US companies as foreign ones. Note, however, that the descriptive fair use approach would arguably come out differently here, and be less protective of the US company than of the Icelandic one. If we let another company like siggis® trademark SKYR, Chobani would probably have a harder time arguing its use was merely to describe the product than an Icelandic company for whom SKYR actually means yogurt.

This makes sense to me. For proto-generic terms, there is a very real concern that a term will become generic or merely descriptive in the near future even if it isn't now. Given my points earlier about trademark law as forward-looking, I think considering future impacts on competition is appropriate.

Professor Brauneis concedes, however, that

[i]t is a more difficult case if the foreign language term falls into our third category.  Suppose that no one in the US speaks German, and one US firm has for some time been using the word GEIGE as a brand name for its violins. Since English has a perfectly fine word for violins, it is hypothetically unlikely that "geige" will ever gain currency as a term in the US. Now another US firm enters the violin market and wants to call itself GRANITE GEIGEN. Or perhaps a company from China, say, MengXiang, has never been in the German market, but now enters the US market and wants to sell its violins under the band name MENGXIANG GEIGEN. Should there be a different result than if an established German company, like GÖTTINGER GEIGENLADEN, wants to enter the US market? 

For now, I just have to say that I'm not sure about this one. Our rationale, of avoiding burdens on established foreign brands, does not apply when a domestic or a third-country company wants to market goods under a foreign term that is not and likely will not be understood by domestic consumers.  And the existing domestic user of GEIGE for violins has an appealing story about how its existing goodwill is going to be destroyed.  So I will let my co-author solve this :-), or at least see what she has to say.

Professor Moerland got the last word, and suggested that, in principle, other entities, whether from the home market (US) or from Germany or from China, should be able to put products on the market that use a foreign descriptive or generic word, even though it is not from their own country and there is a perfectly acceptable, valid domestic alternative term.  This discussion opens the door to points about international relations and the rights of foreign speakers to use their home language, which I'll leave the authors to tackle another day.

***

In any case, this was a fantastic article. It was exceptionally clear and engaging, introducing an intellectual mystery and running through the intricacies of an area of law that, in the authors' hands, becomes fascinating. I am especially grateful to the authors for providing such clear and candid responses to my questions, which advanced the discussion and improved the post.

IP and the Right to Repair

I ran across an interesting article last week that I thought I would share. It's called Intellectual Property Law and the Right to Repair, by Leah Chan Grinvald (Suffolk Law) and Ofer Tur-Sinai (Ono Academic College). A draft is on SSRN and the abstract is here:

In recent years, there has been a growing push in different U.S. states towards legislation that would provide consumers with a “right to repair” their products. Currently 18 states have pending legislation that would require product manufacturers to make available replacement parts and repair manuals. This grassroots movement has been triggered by a combination of related factors. One such factor is the ubiquity of microchips and software in an increasing number of consumer products, from smartphones to cars, which makes the repair of such products more complicated and dependent upon the availability of information supplied by the manufacturers. Another factor is the unscrupulous practices of large, multinational corporations designed to force consumers to repair their products only through their own offered services, and ultimately, to manipulate consumers into buying newer products instead of repairing them. These factors have rallied repair shops, e-recyclers, and other do-it-yourselfers to push forward, demanding a right to repair.

Unfortunately, though, this legislation has stalled in many of the states. Manufacturers have been lobbying the legislatures to stop the enactment of the right to repair laws based on different concerns, including how these laws may impinge on their intellectual property rights. Indeed, a right to repair may not be easily reconcilable with the United States’ far-reaching intellectual property rights regime. For example, requiring manufacturers to release repair manuals could implicate a whole host of intellectual property laws, including trade secret. Similarly, employing measures undercutting a manufacturer's control of the market for replacement parts might conflict with patent exclusivity. Nonetheless, this Article’s thesis holds that intellectual property laws should not be used to inhibit the right to repair from being fully implemented.

In support of this claim, this Article develops a theoretical framework that enables justifying the right to repair in a manner that is consistent with intellectual property protection. In short, the analysis demonstrates that a right to repair can be justified by the very same rationales that have been used traditionally to justify intellectual property rights. Based on this theoretical foundation, this Article then explores, for the first time, the various intellectual property rules and doctrines that may be implicated in the context of the current repair movement. As part of this overview, this Article identifies those areas where intellectual property rights could prevent repair laws from being fully realized, even if some of the states pass the legislation, and recommends certain reforms that are necessary to accommodate the need for a right to repair and enable it to take hold.

I thought this was an interesting and provocative paper, even if I am skeptical of the central thesis. I should note that the first half of the paper or so makes the normative case, and the authors do a good job of laying out the case.

Many of the topics are those you see in the news, like how laws that forbid breaking DRM stop others from repairing their stuff (which now all has a computer) or how patent law can make it difficult to make patented repair parts.

The treatment of trade secrets, in particular, was a useful addition to the literature. As I wrote on the economics of trade secret many years ago, my view is that trade secrecy doesn't serve as an independent driver of innovation because people will keep their information secret anyway. Thus, any innovation effects are secondary, in the sense that savings made from not having to protect secrets so carefully can be channeled to R&D. But there was always a big caveat: this assumes that firms can "keep their information secret anyway," and that there's no forced disclosure rule.

So, when this article's hypothesized right to repair extended to disclosure of manuals, schematics, and other information necessary to repair, it caught my eye. On the one hand, as someone who has been frustrated by lack of manuals and reverse engineered repair of certain things, I love it. On the other hand, I wonder how requiring disclosure of such information would change the incentive to dynamics. With respect to schematics, companies would probably continue to create them, but perhaps they might make a second, less detailed schematic. Or, maybe nothing would happen because that information is required anyway. But with respect to manuals, I wonder whether companies would lose the incentive to keep detailed records of customer service incidents if they could not profit from it. Keeping such records is costly, and if repairs are charged to customers, it might be better to reinvent the wheel every time than to pay to maintain an information system that others will use. I doubt it, though, as there is still value in having others repair your goods, and if people can repair their own, then the market becomes even more competitive.

While the paper discusses the effect on the incentive to innovate with respect to other forms of IP, it does not do so for trade secrets.

With respect to other IP, the paper seems to take two primary positions on the effect of immunizing IP infringement for repair. The first is that the right to repair can also promote the progress, and thus it should be considered as part of the entire system. While I agree with the premise from a utilitarian point of view, I was not terribly convinced that the right to repair would somehow create incentives for more development that would outweigh initial design IP rights. It might, of course, but there's not a lot of nuanced argument (or evidence) in either direction.

The second position is that loosening IP rights will not weaken "core" incentives to develop the product in the first place, because manufacturers will still want to make the best/most innovative products possible. I think this argument is incomplete in two ways. Primarily, it assumes that manufacturers are monolithic. But the reality is that multiple companies design parts, and their incentive to do so (and frankly their ability to stay in business) may well depend on the ability to protect designs/copyright/etc. At the very least, it will affect pricing. For example, if a company charged for manuals, it may be because it had to pay a third party for each copy distributed. Knowing that such fees are not going to be paid, the original manual author will charge more up front, increasing the price of the product (indeed, the paper seems to assume very little effect on original prices to make up for lost repair revenue). Secondarily, downstream repairs may drive innovation in component parts. For example, how repairs are done might cause manufacturers to not improve parts for easy repair. The paper doesn't seem to grapple with this nuance.

This was an interesting paper, and worth a read. It's a long article - the authors worked hard to cover a large number of bases, and it certainly made me think harder about the right to repair.

Using Antitrust to Fix Broken Markets

The prescription drug market is a real mess, in my view. At the very least, it is complicated, and prices for drugs seem to be higher in the U.S. than elsewhere (though teasing that out is hard, since very few pay sticker price and insurance companies negotiate deals). But I don't know what the solution is, and I'm not convinced anyone else does, either. The recent article that triggers my thoughts on this is: A Non-Coercive Approach to Product Hopping, 33 Antitrust 102 (2018), by Michael Carrier (Rutgers) and Steve Shadowen (Hilliard & Shadowen LLP). Their paper is on SSRN, and the abstract is here:

The antitrust analysis of product hopping is nuanced. The conduct, which consists of a drug company’s reformulation of its product and encouragement of doctors to switch prescriptions to the reformulated product, sits at the intersection of antitrust law, patent law, the Hatch-Waxman Act, and state substitution laws, and involves uniquely complicated markets with different buyers (insurance companies, patients) and decision-makers (physicians).

In Doryx, Namenda, and Coercion, Jack E. Pace III and Kevin C. Adam applaud some courts’ use of a product-hopping analysis that finds liability only where there is an element of coercion. In this response, we explain that the unique characteristics of pharmaceutical markets render such a coercion-based approach misguided. We also show that excessively deferential analyses would give brand-name drug firms free rein to evade generic-promoting regulatory regimes. Finally, we offer a conservative framework for analyzing product hopping rooted in the economics and realities of the pharmaceutical industry.

This very brief response essay does a good job of highlighting some of the difficult nuances associated with product hopping (something I'll describe more in a minute) in the prescription drug market. I don't necessarily agree with it - I actually disagree with the final proposal. I share this here because it prompted me to think more closely about an issue I had mostly ignored, and I respect Mike Carrier and think that he does about as good a job at presenting this particular viewpoint as anyone.

Still, I'm troubled by the use of the blunt weapon of antitrust against product hopping, even as I have misgivings about other areas of this market. Product hopping occurs when a name brand (read, patented) drug is pulled from the market in favor of a "new and improved" product (that is also patented). The concern is that the removal of the old brand from the market just as generics are arriving will fail to trigger mandatory generic substitution rules, because doctors can't prescribe the old name brand. Instead, the fear is that doctors will only fill prescriptions for the new, improved (though the parties debate the improved point) drug, for which there is no generic to substitute. Further, they won't ever prescribe the generic once the name brand is off the market.

Here's what gives me a bit of trouble about this. While I am a big fan of automatic generic substitution laws, I am skeptical that they should be used as innovation policy, to the point where the inability of a generic to take advantage of it creates an antitrust injury and also forces a company to make a product that it may or may not want to make for profit or other reasons. In other words, a system that requires a company to make a product so that other companies can sell a substitute product that sellers are required by law to sell is a broken system indeed.

The question is, where is the system broken? Some will, no doubt say that it is the patent system and exclusivity. Surely that plays a role. But I'd like to point to three other points of market failure that drug makers point to. None of these are really new, but I'm taking the blogger's prerogative to talk about them.

1. I think it is a huge assumption that doctors will only prescribe the new drug and won't prescribe the generic once it hits the market. The argument in the article above is that because doctors don't have to pay for it, they have no incentive to cost minimize. Which the incentive is certainly diminished in theory, this is not my experience with (many) doctors at all. I have had many doctors prescribe me (or my wife) "older" generic versions of drugs because they were cheaper. One funny thing is that often times those older versions were awful - like a drug my wife had to take as an awful tasting liquid because the pill version cost 3 or 4 times as much and wasn't covered by the (generally good) insurer, or a blood pressure medication I had to take because it was the standards--until I could show that I had a rare side effect.

2. I think insurance companies can regulate much of this through formularies. If you make the new drug non-formulary (or even brand cost), people will migrate to the cheaper generic substitute, even if there is no prescription available. Even if doctors don't pay, insurers sure do, and they have every incentive to make sure that generic substitutes are used if the new drug is not really an improvement.

3. I bristle a bit at the notion that generic companies must rely on substitution laws to get doctors to prescribe. I realize that's how it is done now, and while substitution laws are a good thing, there is nothing stopping generics from telling doctors why the new-fangled drug is no better than the one that was just removed from the market for the same money. We make every other industry do this. I suspect that litigation is cheaper than advertising, and while I am happy to give the industry a leg up, I am wary of giving it a pass on basic business requirements, and I am wary to taking something that's a regulatory windfall to generics (even if it is good for consumers) and making it an affirmative innovation policy. .

4. Taking this last point further, from an innovation standpoint, is there any reason why generics can't innovate their own product hopping drug? Knowing that a deadline is coming, why can't they innovate (or license) their own extended relief version in addition? Indeed, I take an extended release version of a generic drug. It costs a fortune (before insurance), and the generic is benefiting from having created/licensed it. I suppose a more salient example is Mylan's Epi-Pen, which uses a protected delivery system for a generic drug. While most people are not thrilled with Mylan's pricing strategy, it illustrates the basic point I'm trying to make: generic manufacturers are not helpless victims of the system stacked against them, they are active, rent-seeking participants willing to exploit systematic flaws in their favor.


Finally, I will say that all of my arguments are empirically testable, as are the arguments on the other side. If folks can point to studies that have demonstrated actual physician, consumer, and insurer behavior in product hopping cases, I will be happy to post here and assess!

AOC on Pharma & Public Funding

Congresswoman Alexandria Ocasio-Cortez has already gotten Americans to start teaching each other about marginal taxation, and now she has started a dialog about the role of public funding in public sector research:

This week I conducted my first-ever line of questioning at an Oversight committee hearing, which focused on the skyrocketing costs of pharmaceuticals.

Here’s what happened ⬇️ https://t.co/M6XuL61zXN

— Alexandria Ocasio-Cortez (@AOC) February 1, 2019

Big Pharma companies often say that drugs “need” to be expensive to fund research. What they don’t tell you is that they use *publicly-funded research* for expensive, privatized drugs.

This week on Oversight, @RoKhanna and I exposed how this scheme works:pic.twitter.com/ELbllU3Fxy

— Alexandria Ocasio-Cortez (@AOC) February 1, 2019

In these short videos (which email subscribers to this blog need to click through to see), Ocasio-Cortez and Ro Khanna are seen asking questions during a Jan. 29 House Oversight and Reform Committee hearing, "Examining the Actions of Drug Companies in Raising Prescription Drug Prices." So far, @AOC's three tweets about this issue have generated over 7,000 comments, 58,000 retweets, and 190,000 likes.

Privatization of publicly funded research through patents is one of my main areas of research, so I love to see it in the spotlight. There are enough concerns with the current system that the government should be paying attention. But as I explain below, condensing Ocasio-Cortez and Khanna's questions into a headline like "The Public, Not Pharma, Funds Drug Research" is misleading. Highlighting the role of public R&D funding is important, but I hope this attention will spur more people to learn about how that public funding interacts with private funding, and why improving the drug development ecosystem involves a lot of difficult and uncertain policy questions. This post attempts to explain some key points that I hope will be part of this conversation.

1. Pharmaceutical development currently depends on both public and private funding, with public funding playing the largest role in basic sciences discoveries that indirectly lead to new drugs.

According to the National Science Foundation, out of the $495 billion spent on U.S. R&D in 2015, the federal government funded $121 billion—about one-quarter. Figuring out the portion of this R&D that might be relevant to the pharmaceutical industry is difficult, but most of it would be funded by the National Institutes of Health (NIH), which has an annual budget around $39 billion. The U.S. pharmaceutical industry performed over $58 billion in R&D in 2015, with 99.8% of this funding coming from private industry.

How significant are these different funding sources for new drug approvals? It depends on how you count. As Suzanne Scotchmer has explained, new discoveries build on old ones, and the pharmaceutical industry is no exception. If you consider all of the basic science and research tools that make new medical discoveries possible, every innovation can surely be traced to public R&D. (For an accessible read on this topic, see here.) If you are interested in the more direct costs of developing a specific new drug, however, most of the price is still being paid by the private sector.

To illustrate the distinction between direct and indirect influence: Bhaven Sampat and Frank Lichtenberg found that of drugs approved by the FDA from 1998 to 2005, 9% of all drugs and 17% of "priority-review" drugs (those reflecting "significant improvements") have at least one public-sector patent (i.e., a patent based on publicly funded R&D), though they will typically have private-sector patents as well. But a much higher portion reflected more indirect influence of public R&D: half of all drugs and two-thirds of priority-review drugs had a private-sector patent that cited a public-sector patent or publication. An even more expansive measure of indirect influence was used in the 2018 PNAS study by researchers at Bentley University that Khanna is referring to when he asks about "the study that shows that between 2010 and 2016, of every drug, all 210 drugs that were approved by the FDA, were funded by the NIH or public money." As the study notes, ">90% of this funding represents basic research related to the biological targets for drug action." This basic science is crucial, and expensive. But it does not mean that the private sector does not fund drug R&D: later-stage research also carries an enormous pricetag.

Assessing the impact of public R&D is difficult, but some cutting-edge economic research is making progress on these questions. I recently reviewed one such study, which concludes that each $10 million in NIH funding in fact generates 2.7 additional private-sector patents, as opposed to being merely correlated with private-sector advances or crowding out private investment. This study helps justify the large federal expenditures on biomedical R&D, but it does not address the relative importance of public and private sector R&D investments.

2. Late-stage pharmaceutical development requires expensive clinical trials that are primarily financed by private firms, which depend on having sufficient patent protection to recoup development costs.

Ocasio-Cortez and Khanna's comments sound similar to the "paying twice" critique of patents on publicly funded inventions: Why should U.S. taxpayers have to pay supracompetitive prices on patented products when they have already paid for the initial research? This critique seems most compelling for products that are actually covered by public-sector patents—e.g., a patent on a federally funded university invention that is exclusively licensed to a private firm.

But the problem is that many inventions are very far from commercialized products. A university might patent a promising drug candidate based on studies in petri dishes or animals, but the FDA won't approve the drug until it has been through expensive clinical trials that demonstrate safety and efficacy in humans. Currently, most clinical trials are conducted by for-profit firms, and as Ben Roin has explained, pharmaceutical companies screen any drugs with insufficient patent protection out of their development pipelines. Daniel Hemel and I have questioned why there is not more government investment in clinical trials (see p. 570 here), but given the current institutional structures for drug development, if publicly funded drug candidates couldn't be patented and exclusively licensed to private companies, many of them would never make it to market.

This commercialization argument is the primary justification for allowing federally funded inventions to be patented and exclusively licensed, as is permitted under the Bayh–Dole Act of 1980. To be clear, however, this theory cannot justify the Act's present scope. As I have explained in work with Ian Ayres, many federally funded inventions do not require exclusivity to be brought to market. Universities and other recipients of federal research funding should either limit exclusivity to just what is needed for commercialization or do more to justify greater patent rights—and perhaps pressure from Congress can help with this.

3. Some pharmaceuticals generate profits that far exceed their risk-adjusted cost of development, but this doesn't mean that pharmaceutical profits are uniformly too high—in some cases, expected profits seem too low.

Many pharmaceutical prices seem difficult to justify from a public welfare perspective. A recent study of 99 cancer drugs approved by the FDA from 1989 to 2017 found a return of $14.50 per $1 of risk-adjusted R&D spending. Researchers at Yale estimated that Gilead's hepatitis C drugs brought in forty times their development cost in just the first 27 months. Specific instances of pharmaceutical profiteering have led to headlines and public outcry.

But what we don't see are the drugs that are never developed because firms don't expect a sufficient return. These nonexistent drugs can't be counted, but a clever empirical study by Eric Budish, Ben Roin, and Heidi Williams demonstrated R&D investment distortion away from cancer drugs with shorter effective patent protection. Other empirical work has demonstrated the link between expected profits and R&D in other ways: Medicare Part D significantly increased pharmaceutical R&D for drugs targeting older patients; policies focused on increasing returns for developing vaccines and orphan drugs were effective; and introduction of a new drug in a given country seems to be delayed by price controls but accelerated by strong patent rights.

In short, incentives matter. There are lots of ways in which market incentives aren't well aligned with social welfare, and many good arguments that drug development incentives should be less dependent on patent rights. Think, for example, of unpatentable interventions, or products with negative externalities, or innovations for populations with low ability to pay. But policymakers considering pharmaceutical price regulation should recognize that current incentives are not uniformly too high. As summarized by Darius Lakdawalla, "whether innovation is too high or too low is a first-order—perhaps the first-order—policy question in the economics of the pharmaceutical industry."

4. Requiring private firms that make use of public research to provide the public with a return on investment is more complicated than it sounds.

Ocasio-Cortez seems particularly concerned that "the public is acting as an early investor, putting tons of money in the development of drugs that then become privatized, and then they receive no return on the investment that they have made." But it's not obvious that this kind of return on investment makes sense.

First, it would be logistically and legally difficult to require some kind of return based on all the indirect influence discussed above, such as basic science findings that don't result in a patent. For publicly funded research that does result in a patent, the patenting institution (e.g., a university) does receive a return in the form of equity or royalties. But (1) this is a relatively small return, with many university technology transfer offices operating in the red (see p. 291-94 here), and (2) Congress has directed patenting institutions to reinvest any net income in science research and education, so in this sense, the public is receiving a return—it is just not going back through the federal budget.

Congress could require a more significant return on investment, such as requiring firms that profit from publicly supported research to repay those grants—and perhaps more—out of their profits. I have discussed some arguments in favor of such a system, but I have also explained (in work with Daniel Hemel) that "if government-set rewards are targeted toward knowledge goods that the market underestimates, then reducing the market-generated reward for these knowledge goods may be counterproductive" (p. 579 here).

5. There are a lot of open empirical questions about privatization of publicly funded research and its connection to pharmaceutical prices.

There are open questions about pharmaceutical economics, and about patents in general, and about how best to spend public R&D money (though thinking longer term is probably a good step!). And there are lots of questions about how these areas intersect under the Bayh–Dole Act, and whether that system could be improved.

I am convening a small group of leading economists and legal scholars next month to try to reach consensus on what is currently known, what the important open questions are, and what policy recommendations we might have. But given the empirical uncertainty, it is particularly important that any policy changes be implemented in a way that aids robust evaluation. And I welcome dialog with scholars or policymakers interested in tackling these problems.