The continuum of excludability and the limits of patents.

• Author: Kapczynski, Amy
• Position: Introduction through II. Illustrating the Three Influences on Excludability, p. 1900-1942

ESSAY CONTENTS INTRODUCTION I. UNDERSTANDING THE CONTINUUM OF EXCLUDABILITY A. Patents as a Solution to Information Goods' Appropriability Problem B. The Key Determinants of Excludability II. ILLUSTRATING THE THREE INFLUENCES ON EXCLUDABILITY A. Negative Information About Drugs B. Lifestyle Interventions C. Innovations in Healthcare Quality III. EXPLANATORY AND POLICY SIGNIFICANCE A. The Potential Distortions of Patents B. Specific Policy Prescriptions C. Broader Theoretical Implications CONCLUSION INTRODUCTION

Each year, nearly 30,000 people in the United States die from infections resulting from central-line catheters used for monitoring in intensive-care units (ICUs) in hospitals. (1) These deaths are in part a result of the growing problem of antibiotic-resistant infections in hospitals. (2) The intellectual property (IP) literature is replete with proposals to address that problem by incentivizing the creation of new antibiotics--proposals that sometimes have price tags in the billions of dollars. (3) But in 2006, a different kind of breakthrough was reported in the New England Journal of Medicine--a new technology that reduced the number of these infections by about two-thirds. (4) The technology was a humble checklist, featuring important and well-known hygienic practices such as hand washing and the use of antiseptic. (5) Clinical trials have shown that the intervention works in a range of settings, including in otherwise poor-quality ICUs. (6) While the mechanism is still somewhat unclear, it seems to work by giving nurses the authority to enforce the listed practices with doctors and by improving communication in hospitals in other ways. (7)

The checklist intervention is a classic information good: it is immaterial and was much more expensive to create (or to validate, which is often the more pertinent issue with medical interventions, as we will explain) than it is to copy. By any measure of social welfare, it is also a great intervention. It is cheap, has no known side effects, and prevents infections up front rather than simply treating them after the fact. If widely implemented in the United States, it could save more than 15,000 lives and $1 billion in treatment costs each year. (8) These figures are particularly impressive when compared to the billion--dollar figures often attached to proposals for new drug treatments.

Yet the checklist approach is unlikely to be well rewarded by even a very expansive patent system. Even if we assume that patent law permits a strong and enforceable patent on the intervention, it would be difficult for the creator to use that patent to appropriate any significant proportion of the social value created by the intervention. He would have to track behavior that is routine and to some extent cloaked by privacy norms related to doctor-patient relationships. Thus, the state of technology and of norms place limits on the freedom of our hypothetical patentee to enforce his patent. Of course, we should not overstate the case. An institutional factor would work in the inventor's favor: the existence of hospitals as an organized intermediary. This setup reduces the number of potential infringers and increases the likelihood of tangible evidence of infringing conduct by creating a "paper trail" of written or oral records of established hospital protocols. On balance, though, the checklist will be much less excludable than a drug.

Recognizing this shortcoming of the patent system draws attention to what we call "the continuum of excludability." As we elaborate below, excludability is not a binary quality, either "on" or "off' depending on the availability or absence of property rights. Rather, it is highly variable across information goods, and is affected not only by formal legal entitlements, but also by existing technologies for detecting or tracing such uses (and their costs); existing social norms regarding "acceptable" or "reasonable" enforcement efforts (in light of concerns about privacy, freedom of thought and speech, and so forth); and the existing institutions--or social roles, relations, and organizational forms--within which the predominant uses of the good will be made. Once we recognize that excludability is a continuous and not binary variable, an impressive array of information goods that are difficult to exclude even in the presence of patents comes into view. (9) Our central aim in this Essay is to develop, with examples in public health, our analytic understanding of the continuum of excludability, and to elucidate its substantial implications for innovation theory and policy.

The central justification for patent rights in the United States is economic in nature and is premised on the incentives that they provide to innovators. (10) But patents are only one strategy for incentivizing information production. Many others exist, and two in particular are much discussed in the economics literature: public funding (where a government agency either directly carries out research or sponsors others through grants), and financial inducement through prizes (where financial rewards are established, typically by the government, in exchange for specified information goods). (11) The most influential theoretical account of the advantages of patents over these other institutional approaches can be traced to the influential work of

Harold Demsetz. (12) In a 1969 article, Demsetz suggested that patents are plausibly superior to more directly government-led strategies for generating innovations because markets utilize dispersed private information more effectively than government actors can. (13) Because they link the magnitude and direction of innovation incentives to market prices, in other words, patents may be a better mechanism than reliance on government funding for ensuring that all truly valuable information goods-and only truly valuable information goods-are generated. Demsetz thought that this allocative advantage of patents could outweigh their acknowledged drawbacks, most prominently the fact that patents inefficiently curb the use of protected information. (14) Today, this is the most common justification for patents in the legal literature, (15) which we will call the "allocative" account.

External or foundational critiques could be made of the basic logic of the allocative case for patents. For example, one might challenge welfarism as a value, question the relationship between market value and social value, or be skeptical of the positive assumptions of the underlying "homo economicus" model of innovator motivations. We set aside such foundational objections here, and, for the sake of argument, accept the basic premises invoked by the conventional account. (16)

The main internal criticism of the allocative account in the existing literature points out that patents will systematically underreward research because they yield less than full appropriability (for example, because patents have a limited term and can be designed around, and because transaction costs interfere with market signaling). (17) These concerns, however, are not so much a criticism of the allocative case for patents as a worry over obstacles to its full realization. Here we develop an account of a different problem with the allocative case for patents. The link that Demsetz drew between exclusion rights and revenues for creators assumed that exclusion rights do not themselves introduce distortions into the equation between production and social value. But this is not the case.

We offer a series of detailed examples to show that some kinds of information goods will be much more difficult to exclude (and thus to commodify) than others. Importantly, there is no reason to expect that the ease of exclusion will be correlated with social value. Thus, patents themselves can have distortive effects, stemming from structural features of exclusion rights. Importantly, the problem of nonexcludability cannot be resolved by increasing patent scope or length, and it is asymmetrical with respect to different types of information goods. The continuum of nonexcludability thus means that property rights, even if perfected, and even assuming away conventional transaction costs, will create asymmetrical demand for different kinds of information goods. (18)

Other scholars have pointed out that patent systems fail to create goods whose value is difficult to appropriate in consumer markets. (19) But the fact that excludability operates in asymmetrical ways for different kinds of information goods has not been fully appreciated, nor have its implications for the standard allocative case for patents. Part I illuminates our analytic arguments in more detail, while Part II illustrates the continuum of excludability and its importance with three examples: hospital checklists, negative information about drugs, and information and interventions regarding "lifestyle" risk factors. All three cases identify highly nonexcludable information goods--and position them opposite pharmaceutical products--to show that even though the less excludable innovations may plausibly outperform the pharmaceutical alternatives (in terms of net social benefit provided), a patent system will tend to promote the excludable pharmaceutical approaches over the less excludable alternatives.

In Part III, we draw out the central implications that emerge from our analysis. We begin by developing the point that patent rights have the potential to predictably and systematically distort private investment decisions over innovations by overstating the value of highly excludable information goods and understating the value of highly nonexcludable ones. (20) As a result, patents will fail to incentivize many significant innovations (even if we were to increase the length or scope of their protection), and indeed may promote the production of less valuable inventions over the production of more valuable ones. The possible first-order...