Category Archives: Uncategorized

Innovation Policy Pluralism, or Innovation Policy Hybridism?

Jul 17, 2018 Kevin E. CollinsAdd a Comment

In previous work, Daniel J. Hemel and Lisa Larrimore Ouellette explored the range of tools available to regulators interested in promoting innovation. (See Ted Sichelman’s jot.) While legal scholars addressing innovation policy frequently focus solely on patent law—in fact, the term “intellectual property” is often employed as a synecdoche to refer to the broader scholarly field of innovation policy—Hemel and Ouellette argued that viewing patents, prizes, grants, and tax credits as imperfect substitutes allows the public goods problem that underlies innovation policy to be solved in a variety of different ways, each with its own advantages and disadvantages.

In their most recent work, Innovation Policy Pluralism, Hemel and Ouellette push their earlier argument one step further. They again increase the number of tools in the innovation-policy toolkit by developing a divide-and-recombine approach to intellectual property and its quasi-substitutes. They argue that any given tool for promoting innovation has two “separate and separable” components. First, it has an innovation incentive or a “payoff structure for the producers of knowledge goods.” Second, it has an allocation mechanism that “establish[es] the conditions under which consumers can use knowledge goods.” Hemel and Ouellette provide a thorough, clearly argued, and convincing analysis of the combinatorial possibilities that arise from this finer-grained analysis of the components of innovation-policy regimes.

More specifically, Hemel and Ouellette identify three conceptually distinct types of combinatorial possibilities. First, matching combines components of different types: it allows the innovation-incentive component from one conventional innovation-policy tool to be coupled to the allocation-mechanism component from another. For example, if the government were to buy patents from innovation producers upon issuance at the market price and make the innovation freely available to the public, the innovation incentive would resemble the innovation incentive created by patent law (because it rewards producers with a fraction of the social value that their innovation generates, as measured by consumers’ willingness to pay) but the allocation mechanism would resemble the open-access regime conventionally associated with prizes and grants. Second, mixing treats each of the components as continuous rather than binary variables and allows amalgams of distinct components of the same type. For example, focusing on the innovation-incentive component, an upfront grant that does not cover the full expense of research and development expenses may be combined with patent exclusivity that is weaker than conventional patent law provides. Third, layering recognizes that different types of innovation incentives and allocation mechanisms can be deployed at the national and international levels.

While they do produce and analyze the pros and cons of an array of specific proposals, Hemel and Ouellette view their primary contribution as a conceptual framework for talking about innovation policy in a more precise and nuanced manner. The framework helps us to both conceive new innovation-policy proposals and evaluate their normative merits. Just as importantly, it also helps us to understand the jumble of laws that currently promote innovation in the United States. To prove this latter point, Hemel and Ouellette use their framework in a final section to describe the innovation-policy regime that already exists in the pharmaceutical industry more succinctly and accurately than would be possible if taxes, grants, and patents were conceived of as siloed innovation-policy regimes.

To my mind, the article’s title is a bit misleading. (This is akin to praising with faint damnation, I know.) Hemel and Ouellette’s earlier work was about innovation-policy pluralism. There is no one-size-fits-all optimal solution to a public goods problem; different circumstances may call for different tools from the toolkit; there is value in recognizing the diversity of tools available for remedying an innovation-policy problem. Taking the argument to the next level, Innovation Policy Pluralism embraces what I think could more tellingly be labeled innovation-policy “hybridism.” Here, the very notion that there is a finite set of discrete tools in the toolkit becomes a contingent artifact of the conceptual framework conventionally used to think about innovation policy. By breaking these tools down into the components that they all share, the conceptual framework developed by Hemel and Ouellette reveals that we can match, mix and layer our way to hybrid tools that are more perfectly crafted to the specifications of any given innovation-policy problem.

IP Norms’ Dark Side

Jun 26, 2018 Dotan OliarAdd a Comment

The early legal literature on law and social norms tended to paint a rosy picture. Social norms were generally depicted as an optimal set of organically developed rules, informed by the experience of a close-knit community, and thus superior to formal law as a way of regulating behavior. Later scholars came to realize that nothing guarantees the optimality of social norms: they may perpetuate practices that no longer make sense, or they may advance the interests of certain groups but not social welfare. In such cases, formal law holds the promise of overriding suboptimal norms and moving society to a better place.

A few years back, I wrote, together with Christopher Sprigman, one of the first papers on social norms in intellectual property law in which we explored how standup comedians informally regulate the ownership and transfer of rights in jokes and comedic routines. While we thought that in that particular case the system of social norms largely brought good outcomes in terms of inducing creativity, that conclusion came at the end of reviewing their advantages relative to copyright law, but, importantly, also their relative disadvantages. For example, we noted aspects of the norms system that we thought were unattractive for creativity relative to formal copyright law: standups norms’ system does not recognize a term limit (so the norms’ system discourages dissemination and reuse of old materials), does not harbor a concept of fair use (so it discourages derivative and sequential creativity), and grants a scope of protection that exceeds that which is available under copyright’s idea/expression doctrine (so it discourages borrowing high level concepts and themes from others).

Since we wrote our paper, the law and social norms field of IP has grown considerably, and is, I believe, one of the most important developments in IP theory over the last decade, not to mention a fun and interesting one. Others have explored social norms that are copyright-like, patent-like, and trademark-like. More recently, the literature moved to make more general claims and observations based on a rich array of norms’ systems and case studies. One of the latest and important additions to this recent literature is the paper by Stephanie Bair and Laura Pedraza-Fariña which makes the claim about the dark side—i.e. counterproductive creativity-wise—of social norms more generally. In doing so, it builds on much social science and psychology literature.

Bair and Pedraza-Fariña emphasize that anti-innovation norms are harmful because they tend to hinder breakthrough innovation, namely innovation that recombines knowledge across boundaries of (often distant) creative communities and industries. They create barriers between bodies of knowledge, reinforcing the use of specialized, field-specific terminology rather than one that is shared among fields. They also tend to be over-enforced.

Anti-innovation norms come in three types: research priority norms, methodology norms, and evaluation norms. Research priority norms determine which problems are worth artistic or technological attention. Their dark side is that they implicitly suggest which problems are not worth members’ time and attention. This is particularly costly when norms discourage members from exploring intersectional problems.

Methodology norms dictate the proper ways, tools, and routines appropriate to solve the problems community members identified as important. Such norms can have a negative consequence to the extent that they teach against the use of alternative, non-conforming methodologies. Intersectional problems pose a specific challenge as they are often best addressed through the combination of methodologies.

Lastly, evaluation norms dictate how a creative community assesses the contribution of members. Good work may be that which is patented, or published in certain fora, or meets some other community standard. Their dark side lies in reinforcing priority and evaluation norms and in discouraging cross-disciplinary innovation.

Two psychological biases explain why anti-innovation norms emerge and persist. First, the status quo bias describes a preference for the current state of affairs over change. Second, conformity bias describes individuals’ tendency to forego exercising their own judgement in favor of following social norms.

Having illuminated social norms’ dark side, Bair and Pedraza-Fariña propose solutions. They believe that IP law, which affects the appropriability of the fruits of one’s labor, is limited in its ability to incentivize cross-boundary innovation. Instead, they emphasize the need to turn to other mechanisms that affect innovation, such as government grants and tax credits.

Conducting the first general treatment of the topic, this article is important not only for drawing attention to, and advancing the understanding of, social norms’ failures, its causes and its effects, and for suggesting possible solutions thereto, but also because the analysis opens up the door for many potential follow-on research paths, of which I would like to suggest a few. First, the article reinforces the understanding that law and social norms are just two ways to foster innovation, and that each comes with relative, rather than absolute, advantages and disadvantages. Law and social norms do not only affect how much is being created, but also the type of innovation that is taking place. For example, in standup comedy, early 20th century norms suggested that copying and reuse were legit and humor tended to be generic, but late 20th century norms suggested that originality was key, and the type of humor created tends to be personal, observational, and point of view driven. The law and social norms literature suggests that in some cases law and social norms may work in tandem and reinforce each other, but that in others they may crowd out one another. After recognizing that social norms have their disadvantages too, an additional step forward would be to develop a theory of the optimal mix of innovation law (IP and non-IP) and social norms.

Second, the authors focus on social norms’ dark side in hindering cross-boundary research. It is at least conceivable that social norms also have negative effects within disciplinary boundaries (as the last comedy example may suggest). Examining norms’ failure in other communities may suggest whether this is a more general point, and perhaps help to enrich the general theory of norms’ failure.

Third, and related, since the article comes at a stage where a considerable number of industry-specific case studies already exists, it likely presents the opportunity to further contextualize and concretize the discussion. For example, for all or some of the communities explored in the literature, scholars could identify particular anti-innovation norms, map the category they fall into, highlight the psychological biases that reinforce them, and propose particular IP and non-IP reforms that would be most effective in countering them. Indeed, it seems that there is much to be gained from paying attention to anti-innovation norms in IP.

Does Copyright Have a Framing Problem?

May 29, 2018 Michael W. CarrollAdd a Comment

Numerous provisions of the Copyright Act of 1976 (“1976 Act”) use the term “work” as a key referent for determining copyrightability, ownership, scope of rights, limitations on scope, and remedies. Yet, Congress did not provide a general-purpose definition of what counts as a “work,” even though it defined a plethora of arguably much less important terms. When the parties in litigation explicitly or implicitly disagree about the fundamental issue of what the plaintiff’s or defendant’s work is, what is a court to do?

This is a big and important question. While Kaminski and Rub do not provide a full response, they do frame the problem and illustrate how it plays out in many contexts. They demonstrate that courts have considerable flexibility in how to define the relevant work. Sometimes, courts use this flexibility to “zoom in” on particular facets of, for example, a design on a carpet that mixes public domain and original elements. Other times they “zoom out” to consider a work’s total concept and feel. Courts rarely defend their framing of the relevant work, and when they do, they do not use a shared set of criteria to justify their choices.

Kaminski and Rub have three goals for this article. First, they review numerous cases that illustrate the range of legal issues for which a court’s framing choice is consequential. They show the inconsistency of judicial framing choices and the rarity of justifications of these choices. Second, they argue that the doctrinal flexibility judges have exercised is more a feature than a bug in copyright law because it supplements a court’s ability to tailor the scope of rights or remedies in particular cases to better achieve copyright’s goals. Third, they argue that courts should more self-consciously exercise this framing flexibility and provide a justification for their choices. The authors close with a set of criteria courts might rely on to provide such a justification.

In their survey of the choices that are available to courts in framing the definition of the “work,” the authors start with the cases involving copying that is not wholesale, so-called “non-literal infringement.” The legal test is whether the two works are “substantially similar,” and how this is applied depends greatly on how the plaintiff’s and defendant’s works are framed. One of the more difficult issues in this analysis involves whether a character is a work independent of the larger narrative in which the character appears, and if it is a work, how is that work defined when the character, say, Sherlock Holmes, or the Batmobile, evolves? The article briefly discusses how the Ninth Circuit “zoomed out” from the many iterations of the Batmobile to focus on its general characteristics in the Batman series to define the work in DC Comics v. Towle.¹ As a reader, I wanted the authors to dive a little deeper into this issue to compare and contrast some of the other character cases.

The article then turns to questions of authorship, showing that courts have decided in motion picture cases to zoom out and treat the dominant author of the motion picture as the author, rejecting the theory that an actor creates a distinct work in a scene for which the actor is primarily responsible. The article then dives deeper into the useful article doctrine, arguing that the Supreme Court, in its enigmatic recent decision in Star Athletica LLC v. Varsity Brands, Inc.,² sent conflicting signals about how to frame the design features and the utilitarian aspects by zooming in on the chevrons on a cheerleading uniform in parts of the discussion and zooming out to look at the uniform as a whole in others.

Turning to the framing tests the courts have used, or could use, Kaminski and Rub argue that most courts use no test, showing the inconsistent framing in cases involving photographs copied from collections of photographs as an example. Courts sometimes ask whether something is a work or a constituent element by testing it for copyrightability, but the authors persuasively argue that this test does too little work. A market-based approach has some appeal, and some courts ask whether the putatively independent works can “live their own copyright life [sic].”³ A variant on this test is whether the putative work has independent economic value. The authors acknowledge the appeal and some utility of this approach, but they rightly caution that with increasing variety in bundling and unbundling of works in digital form, market signals are less likely to provide stable guidance.

The closing section of the article argues that judicial flexibility in framing the work is beneficial because it allows courts to tailor protection and remedies with respect to policy considerations such as: the incentives-access tradeoff, managing transaction costs, and channeling creativity. As a result, permitting inconsistent framing depending on the doctrinal context provides some benefit. Small units may be copyrightable, and therefore works, for the purposes of copyrightability, but if each copyrightable unit is eligible for a separate award of statutory damages, existing problems with proportionality in statutory damages would become greatly magnified, for example.

The authors set an ambitious task for themselves, and their ability to range across the copyright caselaw is impressive. Two caveats: As a reader, I would have liked to see greater engagement with Justin Hughes’s article on microworks,⁴ which the authors acknowledge that they are building on, particularly when considering the available framing tests. I also think that the framing choices they discuss in the subsections on fair use, statutory damages, and Section 201(c) are qualitatively related but distinct because the decisions in the cases discussed there focused more on choosing which work – the collective work or the contribution to the collective work – to focus on rather than whether to zoom in or out on aspects of the work, as in the substantial similarity or useful article cases. That said, Kaminski and Rub make a persuasive case that courts should more explicitly identify and justify their framing choices about the relevant work(s) when applying the wide range of doctrines tied to the “work.”

When Social Norms for Innovation Don’t Maximize Social Welfare

Apr 30, 2018 Christopher J. BuccafuscoAdd a Comment

For more than a decade, intellectual property scholars have demonstrated how informal social norms can efficiently substitute for formal legal protections to promote creativity and innovation. Now come Stephanie Bair and Laura Pedraza-Fariña to show that social norms can have a less happy effect. In some instances, they argue, informal social norms may impede innovation. Their forthcoming article, Anti-Innovation Norms, explores a selection of these and what might be done about them.

According to the standard economic story, IP rights exist to solve a public goods problem. In the absence of IP rights, rampant copying would reduce the ability of creators to recoup their investments, and society would face underinvestment in new works and inventions. IP rights solve this problem by enabling creators to charge supramarginal prices for those works and inventions. Although IP rights are costly, they produce net benefits by encouraging creativity that we wouldn’t otherwise have.

Yet a still-expanding list of studies has shown that in many industries, including fashion design, cooking, stand-up comedy, and tattooing, the lack of formal IP protection has not killed creativity. In these and other fields, robust social norms that promote creativity and attribution and that shun certain kinds of copying can replace formal legal protection more flexibly and at lower social cost. Norms can make free riding costly, and they support those engaged in creative work.

Professors Bair and Pedraza-Fariña do not gainsay the importance of social norms to limit the effects of free riding in many low-IP fields, but they suggest that scholars’ focus on free riding has caused them to miss other, less salutary, social norms that affect innovation. In particular, Professors Bair and Pedraza-Fariña demonstrate the existence of a group of norms that undermine an especially valuable source of innovation—disciplinary boundary crossing. While many of the most important innovations in science, technology, and the arts come from the merging of insights from previously unrelated fields, creators can face substantial disincentives to engage in boundary-crossing work.

Professors Bair and Pedraza-Fariña have, themselves, not been dissuaded from crossing disciplinary boundaries to substantiate their account. They turn first to sociological studies of science and the arts in order to identify three anti-innovation norms: research priority norms, methodology norms, and evaluation norms. These norms tell people within a given field (1) to work on projects that the field deems valuable, (2) to apply methods deemed acceptable to those projects, and (3) to judge the value of contributions to the field on terms that the field prefers. Drawing on examples from physics, medicine, and contemporary visual art, the authors demonstrate how anti-innovation norms create field-specific professional and reputational costs that discourage creators from pursuing potentially more valuable opportunities at or beyond the borders of their fields.

Continuing their own boundary crossing, Professors Bair and Pedraza-Fariña next turn to social psychology to explain possible sources of anti-innovation norms. They focus on two aspects of individual decision-making, the status quo bias and the conformity bias, as the likely culprits. For a variety of reasons, people tend to be strongly attached to the ways things are now, and it can be surprisingly difficult to get people to switch away from their present status. In addition, people generally like to conform their behavior to those in their social groups. Conformity aids group membership and solidarity and the benefits they confer. The authors speculate that these two tendencies can make it harder for individual creators to break away from accepted paths and procedures and to risk forging their own identities.

Having established the existence of anti-innovation norms, Professors Bair and Pedraza-Fariña return to their disciplinary home to consider whether law can do anything to solve the problems they’ve articulated. In large part, they are skeptical of IP law’s ability to particularly encourage boundary-crossing innovation, because copyright and patent tend to reward innovations without reference to whether they combine diverse insights or not. They argue that IP’s reliance on decentralized market signals, which is normally considered a feature, can also be a bug when those signals are filtered through a norm-inflected view of the technological or artistic environment. Although the authors suggest some tweaks to IP doctrine, their policy proposals focus instead on using grant funding and tax incentives to encourage and reward boundary-crossing creativity.

Professors Bair and Pedraza-Fariña have helpfully reminded IP scholars that social norms are not always a good thing for creativity. In addition to the norms against boundary crossing, other norms may discourage certain kinds of creators or certain kinds of creativity from emerging. As we develop a richer view of the sociology and psychology of innovation, we should pay attention to the costs of informal ordering as well as to its benefits.

Designing Design Patent Subject Matter

Mar 22, 2018 Mark McKennaAdd a Comment

Designs for articles of manufacture are eligible for design patent protection under section 171 of the Patent Act if they are “new, original and ornamental.” As is clear on the face of this provision, design patent subject matter is limited to designs for “articles of manufacture”—other types of design, including architectural design, would seem to be excluded. But what exactly is included? What is a “design for an article of manufacture”? In an important new article, The Article of Manufacture in 1877, Sarah Burstein provides critical historical context on that question.

Courts and the Patent Office have in recent years given “article of manufacture” a very broad interpretation. They have, for instance, accepted as design patent subject matter the design of a spray pattern for a water fountain, the design of the shank portion of a drill bit, and graphical user interface (GUI) designs (so long as they are depicted in a little rectangle to indicate—nominally—that they show up on a screen). As a subject matter threshold, then, the article of manufacture requirement has in recent years excluded very little from design patent protection.

But the Supreme Court’s decision in  Apple v. Samsung made the relevant article of manufacture central to determining damages, putting new pressure on design patent law’s understanding of that concept. The issue in Apple v. Samsung was the appropriate way to calculate damages in a case involving a “partial design”—that is, a case in which the design patent claimed only part of the article to which it was applied. Design patent has a special damages rule under which an infringer is “liable to the owner to the extent of his total profit, but not less than $250.” 35 U.S.C. § 289. Congress added that rule to the Patent Act after the Supreme Court’s decision in Dobson v. Hartford Carpet Co., 114 U.S. 439 (1885), in which the Court refused to allow the owner of a patented carpet design to recover the full profits from the sale of infringing rugs. According to the Court, profits attributable to the design had to be distinguished from those attributable to the “intrinsic merits of quality and structure” of the article. Thus, in order to recover the entire profits from the sale of articles bearing the patented design, those articles must have sold for higher prices than those without the design. Dobson, 114 U.S. at 444.

Since there is little dispute that Congress intended § 289 as a rejection of the result in Dobson, courts have long understood that section to prohibit apportionment of damages. In the view of the Federal Circuit, that meant that damages in a design patent case were calculated based on sales of the entire commercial product, even if the design patent at issue claimed only a small part of the design of that product (say, the rounded corners of a device, to pick a not-so-random example). But as the Supreme Court seemed to understand, that approach implicitly attributes zero value to the function of the device or any of its other features, a consequence that seems indefensible, particularly in the context of something so complex as a smartphone. Indeed, applying its “no apportionment” rule, the Federal Circuit upheld a jury award of Samsung’s entire profit on the sale of the infringing smartphones, even though those devices employed hundreds or thousands of other patented and unpatented components.

The Supreme Court purported to solve this problem by focusing on the meaning of “article of manufacture.” While the statute requires courts to award the defendant’s “total profits,” those profits must be determined in relation to the relevant article of manufacture to which the patented design was applied. Crucially, and contrary to the Federal Circuit’s assumption, the article of manufacture need not be the entire commercial product sold to consumers; it might be something less than that. So far, so good. Unfortunately, the Court gave no guidance as to when courts should treat the relevant article of manufacture as the commercial product or as some smaller unit, nor did it explain how one should identify the relevant article if it is less than the product as a whole.

It seems likely, however, that the claim will prove important here: Because a design must be for an article of manufacture, the article of manufacture will be defined in relation to the claimed aspects of the broader article shown in the drawings. So understood, “partial design” is a misnomer—patents that claim only a component of the article are not partial designs of an article; instead the article is, at least in some cases, defined by reference to the claim.

If it can be operationalized, the Court’s reconceptualization of the article of manufacture may prove an elegant solution to the remedies problem. But as Burstein shows in her article, if history is our guide to the proper definition of “article of manufacture,” there is more the Court could have done here. For while it’s true that not all articles of manufacture are complete commercial products, neither should all components be considered articles of manufacture.

Surveying early case law, Burstein concludes that “article of manufacture” was a term of art that did not simply refer to any article that was manufactured. (P. 42.) “Instead, it referred to a tangible item—other than a machine or a composition of matter—with a unitary structure made by humans and complete in itself for use or for sale.” (P. 43.) That definition didn’t exclude all components; indeed, several of the cases Burstein cites involve design patents on parts. But only those components that were manufactured separately and identifiable as discrete parts (even if ultimately combined into a bigger commercial product) counted.

That’s obviously not a self-applying rule. Among other things, its workability depends on courts’ ability to identify the claimed design as a relevant part of a product and not just as a feature of the product. That may sometimes be quite difficult. Still, conceiving of the article of manufacture in this way frames the inquiry very differently, and the implications go far beyond remedies. As Burstein highlights, the remedies problem is illustrative of a much bigger issue: Because the PTO and the courts have paid so little attention to what should count as a relevant article of manufacture, they have extended design patent protection to a number of designs that are subject to the Apple v. Samsung problem. Rather than solving it at the remedies stage, perhaps the better approach is to reconsider the scope of design patent subject matter.

Rethinking subject matter would go a long way toward addressing a number of other problems in design patent law. Accepting that one can patent small portions of the design of an article has made claiming more mysterious. That, in turn, has made it difficult for courts to compare claimed designs to the prior art to assess novelty and obviousness. It has exacerbated the difficulties courts face in giving a valid design its proper scope. Burstein acknowledges that there may be reasons to depart to some degree from the 1877 approach, particularly in light of the range of designs now impacted by design patent. But we can do so, she argues, without surrendering the idea that the article of manufacture has limiting content—the “article of manufacture” cannot just be whatever the claimant suggests it is. That’s a lesson that could profoundly affect all of our thinking about design patents for years to come.

What Is the Payoff from Public R&D Investments?

Feb 28, 2018 Lisa Larrimore OuelletteAdd a Comment

Intellectual property scholars have increasingly recognized that IP is only one of many legal tools for incentivizing innovation. In addition to facilitating transfers from consumers to innovators through a “shadow tax” on IP-protected goods, the U.S. government also directly rewards innovators with public funds through R&D tax incentives, innovation prizes, and—most significantly—well over $100 billion per year in direct federal and state R&D support. This direct public science funding, which primarily goes to grants and national laboratories, has long been viewed as important to U.S. economic growth. But federal R&D spending has been declining and is the subject of an increasingly partisan divide, making this a key moment to ask: What is the public’s return on this investment?

In an outstanding empirical analysis, corresponding author Danielle Li at MIT Sloan and her three coauthors—Pierre Azoulay at Sloan, Joshua Graff Zivin at UC San Diego Economics and Public Policy, and Bhaven Sampat at Columbia Public Health (collectively, “AGL&S”)—have tackled this question for grant funding by the U.S. National Institutes of Health (NIH). With a budget of over $30 billion per year for biomedical research, the NIH is the single largest research funder in the world. But assessing the causal impact of this investment is difficult, even when focusing only on its effect on private-sector patents. How can one measure information flows in fields full of serendipity and spillovers? Are grants merely correlated with private-sector advances? Does public funding “crowd out” private investment? AGL&S’s empirical design makes progress on each of these issues, and they conclude that each $10 million in NIH funding in fact generates 2.7 additional private-sector patents.

Prior work has studied outcomes in specific areas of NIH grant funding (which misses advances in different fields) or citations to university patents (which misses grant-funded academic results that are not patented). Instead, AGL&S look to publications resulting from these grants, and then to patents that cite these publications, which have been shown in a validation against survey results to be a better measure of the influence of public sector research than patent-to-patent citations. Out of the 153,076 NIH grants funded between 1980 and 2005, about 40% led to a publication cited by a patent; and out of the 232,276 private-sector life-science patents issued by the USPTO between 1980 and 2012, about 40% directly cited an NIH-funded publication.

A second empirical challenge is that public funding may target the most promising research areas, so increased grant funding might simply be correlated with increased patenting in that area rather than causing it. AGL&S take on this problem in two ways. First, they classify grants by disease (e.g., cancer), scientific approach (e.g., cell signaling), and time (e.g., 1990), allowing them to include fixed effects to account for these potential sources of endogeneity in funding. Second, they find effectively random variation in funding at the disease/science/time level based on grant applications whose review scores were just above or just below the NIH’s funding cutoffs, and they use the “windfall” funding received by a disease/science/time above the cutoffs as an instrument for total funding. Under both approaches, they find that additional funding increases the number of patents that cite NIH-funded publications.

The third challenge noted above is that grant-funded research might crowd out private-sector funding, such that the increase in private-sector follow-on patenting is accompanied by a decrease in patents that do not rely on NIH-funded results. To be clear, substitution of patents with grants is not necessarily problematic from an economic perspective: both are costly, and neither is optimal in all circumstances. But AGL&S show that grant funding spurs rather than substitutes for private-sector R&D. To study this effect, they use a separate dataset of patents that cite publications related to NIH-funded publications (using a keyword-based similarity measure). If NIH funding merely crowded out private research, the authors argue that one would not expect the award of an NIH grant to increase the number of patents in a research area, but instead, they find an increase here as well.

These results come with a number of caveats, the most important of which is that AGL&S focus on the outcome of only private-sector patenting, which ignores the many other benefits of public biomedical research. In the life sciences, more innovations are probably captured by the patent system than in other fields, but there are certainly medical innovations such as ICU hygiene checklists that have enormous social benefit that is not easily excludable with patents. This problem would be exacerbated in other areas of research, which is one reason that it would be difficult to replicate AGL&S’s analysis for grants from other funding agencies.

Of course, even within the life sciences, turning this result into a rate of return is difficult. AGL&S’s back-of-the-envelope calculations lead to tentative estimates on the order of a $20–$30 million increase in firm market value for each $10 million NIH grant. The true social rate of return might be significantly higher (e.g., if social value is significantly higher than firm market value, or if firms often build on uncited knowledge) or lower (e.g., if patents frequently cite to papers that are not real intellectual influences). But even if one focuses only on the results that can be determined with more precision—the positive effect on private-sector patenting and the lack of measurable crowding out—the results in this paper are important data points for the increasingly polarized discussion of direct government R&D funding.