• 03 August 2009

Intellectual Property for Market Experimentation

Michael Abramowicz & John F. Duffy

Posted in , , , , , ,

Why did it take decades from the time inventors first developed wheeled suitcases before they were put on the market?  Why haven’t the courts concluded that trademarks like “Band-Aid” and “Rollerblade” are now generic?  Why did many analysts doubt the business wisdom of launching Netflix even after customer subscriptions exceeded expectations?  Why do we ban competitors from selling baseball hats featuring the logos of Major League Baseball teams even when they are not pretending to sell authorized merchandise?  Why does trade secret protection extend to information like customer lists, when dissemination of such information might promote competition?  Why do we need a special legal regime to encourage the testing of “orphan drugs” even if those drugs are in the public domain?

The answers to these questions all lie in the scope and limits of intellectual property law’s relationship to market experimentation.  To those familiar with intellectual property law, the phrase “intellectual property for market experimentation” may sound like an oxymoron.  Intellectual property law has long been understood principally as a means of encouraging the production of different forms of technological information and creative works.  Patent law and trade secret law protect inventions, while copyright law protects books, art, and music.  Trademark law, rather than protecting innovations, is justified by minimizing consumer search costs and confusion about products.  However, once these various intellectual property laws have protected the relevant information or symbols, the forces of competition are assumed to operate without further interference.  Intellectual property, on the traditional account, is not needed to promote market experimentation.  The invisible hand itself does that.

The Connection Between Technology and Market Experimentation

But there is a parallel between market and technological experimentation that suggests that intellectual property law could encourage market experimentation in much the same way that it encourages technological experimentation.

Consider two similarly situated companies.  Company A is contemplating testing thousands of materials to see if it can find one that could make standard light bulb filaments cheaper.  No one knows for sure which, if any, will work.  The testing process is expected to cost $100,000,000 and holds about a 50% chance of yielding a successful new technology.  If testing is successful and Company A obtains exclusive rights to the new material for a period of years, the firm would gain over $200,000,000.

Company B, meanwhile, is considering promoting a more efficient type of light bulb that was discovered long ago but was never marketed effectively.  A marketing campaign costing $100,000,000 is expected to hold about a 50% chance of convincing consumers to accept the bulb, which produces a different tint of light from conventional bulbs and which, though cheaper in the long run, requires higher up-front costs.  If the campaign is successful and the firm obtains exclusive rights to the newly commercialized bulbs for a period of years, the firm would reap over $200,000,000.

The social benefits and costs of providing exclusive property rights are about equal in these two cases.  It makes no economic difference whether the source of risk arises from scientific or market uncertainty.  The applicable legal rules, however, present a paradox of differential treatment:  Company A likely can secure exclusive rights, while Company B likely cannot.  This distinction may help explain why some technologies exist for years before being effectively commercialized.

Indeed, the hypothetical Company B may help explain the historical experience with compact fluorescent light bulbs.  The basic technology associated with these bulbs is now more than a half-century old and thus in the public domain.  By the 1980s, several companies were engaged in limited manufacturing of this type of bulb.  Yet, despite the availability of the technology, the bulbs remained for decades only a small fraction of the market.  The new bulbs faced numerous marketing and consumer-information hurdles, many of which remain to some extent unsolved to this day.  Perhaps existing market barriers could be overcome with extensive marketing campaigns, product demonstrations, and warranties, but with hundreds of competing manufacturers, no one firm would reap any significant rewards from such efforts if they made the new product successful.

This absence of protection suggests that the existing level of market experimentation may be inefficiently low.  Empirical assessment of this is difficult; one cannot easily measure what does not exist.  Instead, the empirical evidence must come from businesses that have been created.

The Netflix example advances the case because it is an example of a business that easily might not have existed but for the possibility of sufficient legal and nonlegal first-mover advantages.  Also essential was the persistence of its multimillionaire founder, who could personally afford to capitalize the initial business and who had a sufficient track record of success that he could attract venture capital (albeit not without difficulty) as the business was developing.  We cannot identify with certainty business ideas that would have been successful if only they had been implemented, but we can show that even businesses like Netflix that proved to be phenomenally successful may at one time have appeared to be marginal projects or likely losers.  Indeed, even after Netflix took off against the odds, there were many analysts in 2002 and beyond who doubted that it would be able to survive competition from Blockbuster and Wal-Mart.  Netflix’s own disclosures of market data during its 2002 initial public offering have been blamed for attracting Blockbuster and Amazon.com as competitors, further showing that competitors do try to free ride on the market experiments of other firms.

We can also see inadequate protection for market experimentation at work in explaining why some seemingly obvious products and product features take so long to emerge.  A possible example of this phenomenon is luggage with wheels, a feature that came into common use only in the late twentieth century.  The idea that adding wheels to luggage might be useful is old in the art.  A 1914 patent, for example, protected a device that secured wheels to a suitcase.  Even such a patent would provide little market exclusivity, given the myriad of other ways that one might attach wheels to a suitcase.

The chief impediment to wheeled suitcases, as in the case of the light bulb example discussed above, appears to have been not technological uncertainty but market uncertainty:  here, uncertainty about consumer demand.  If a widespread market experiment with wheeled luggage proved successful, established luggage companies would surely copy the innovation.  This was, of course, precisely what happened once a suitcase with wheels was finally marketed successfully.  While we cannot eliminate the possibility that other factors, such as technological complications, may have contributed to the delayed widespread introduction of luggage wheels, the inadequacy of incentives to engage in market experiments seems likely to have played some role.

Intellectual Property Protection for Market Innovation

Whatever the empirical evidence, there are strong theoretical arguments that incentives for market innovation are insufficient.  Industrial organization scholars have long recognized that in many contexts, free entry will not lead to socially optimal entry.  Similarly, the literature on business management has recognized that, while first movers enjoy significant advantages, second movers have significant advantages too.  A market experiment can produce information of a type generally ignored by the literature—information about whether consumer demand and other market conditions will permit commercial success.  Straightforward economic models can show that a system of free competition with no exclusive rights for market experiments may provide inadequate incentives to induce socially valuable experiments such as the introduction of a new (but technologically uninnovative) product or service or even the marketing of an old product or service to a new geographic market.  Intellectual property protection for market experimentation could raise social welfare, even if the exclusive rights impose substantial losses.

Modern intellectual property theory posits that exclusivity imposes a static cost, but that the dynamic benefit of encouraging information production and dissemination may make this cost worth bearing.  This same logic applies to market experimentation.  The effective launch of a new product or service may require substantial investments.  Commercial success or failure produces information about market demand and supply upon which competitors can often free ride.  To be sure, early experimenters will gain some first-mover advantages, as they also do with technological innovations, but late-entering competitors obtain two important second-mover advantages against early market experimenters.

First, they do not have to bear the cost of investing in market development.  Second, they can copy the first experimenter’s market successes and avoid repeating its failures.  Once such market information is created, consumers would benefit from competition, but without a sufficient guarantee of exclusivity there may be no one with enough incentive to undertake the risky initial investment in developing and testing the market.  Just as patents encourage risky but ultimately beneficial technological experimentation, some form of intellectual property protection could result in a socially beneficial increase in market experimentation and entrepreneurial activity.

Economists, ranging from leading twentieth-century champions of free markets to modern analysts of matters as seemingly diverse as high-tech entrepreneurship and economic growth in developing economies, have long recognized this parallel between market experimentation and technological experimentation but, along with legal theorists, have failed to draw a connection to intellectual property law.  Leading economist Frank Knight, for example, viewed trade secret law and the patent system as important means of addressing the problem that, “owing to the low cost of indefinitely multiplying an idea, it is usually difficult to capitalize an increase in productive power.”1 Knight recognized that existing intellectual property law would make it difficult for an innovator to “secur[e] any permanent gain” from an “improvement of business organization and methods” because such an improvement was “usually neither patentable nor capable of being kept secret.”2

Yet, after thoughtfully identifying this problem, Knight largely dodged solving it.  He asserted that “there is no evidence of any unwillingness to make expenditures in this form of improvement,” even though that “fact” was puzzling to him.  Perhaps Knight simply assumed that the large amount of market innovation precluded the possibility that there might be inefficiently little of it.

Professors Scott Shane and S. Venkataraman, two leading scholars in the study of entrepreneurship, also recognize the parallel between market and technological experimentation without recognizing how intellectual property law might help.  Shane and Venkataraman view entrepreneurship as “the crucial engine” driving change in a capitalist society.  They also recognize that entrepreneurship is tied to the production and exploitation of new information, and they expressly note that the opportunities for entrepreneurship “need not be restricted to . . . technological developments.”3 They acknowledge the appropriability problem of information generated by entrepreneurship, noting that the new “information diffuses to other members of society who can imitate the innovator and appropriate some of the innovator’s entrepreneurial profit.”4

Yet they accept the legal structure addressing the appropriability problem as a given; Shane and Venkataraman view entrepreneurial activity as depending upon factors such as “[t]he provision of monopoly rights, as occurs with patent protection or an exclusive contract,” and “the slowness of information diffusion or the lags in the timeliness with which others recognize information.”5

Objections to Intellectual Property Rights for Market Innovations

There are at least two general objections to the argument for intellectual property rights for market experimentation that may help explain why it has received almost no attention in the past.  The first was voiced long ago by Friedrich Hayek as a general objection to any form of a government-sanctioned exclusive right designed to promote innovation:  The free market already provides abundant incentives to experiment and innovate.6 This Hayekian position draws no distinction between technological and market experimentation; thus, it would resolve the paradox of differential treatment present in the law by abolishing intellectual property rights for innovation generally. The Hayekian position has remained a polar position, and, while free competition remains an important alternative to intellectual property in some circumstances, all developed nations now recognize intellectual property rights as a means of fostering technological experimentation.

The second objection maintains that even if free competition is suboptimal in encouraging market experimentation, the remedy of expanding intellectual property protection would be worse than the original disease.  This argument is more formidable than the Hayekian objection.  Even a limited grant of exclusive rights may impose significant costs on society.  Because society generally relies on exclusive rights to encourage technological experimentation, however, it would be highly surprising, given the fundamental similarities between technological and market experimentation, if the optimal policy choice for encouraging market experimentation were always to rely upon whatever natural first-mover advantages exist in a particular market and never to deploy some form of exclusive rights.

Current Protections for Market Innovations

It is thus not surprising that exclusive rights have been used in some circumstances to encourage market experimentation.  British “patents of importation,” which provided exclusive rights for the importer of a technology already existing outside of Britain, were permissible for hundreds of years.  Likewise, exclusive franchise agreements are a widely employed means by which franchisors encourage new franchisees to risk developing a business in new geographic locations.

Although there is no general public law analogue to franchise agreements, and although the United States’s legal system was the first to reject patents of importation, the goal of promoting market experimentation is not as alien to modern American intellectual property doctrine as it may first seem.  Perhaps as much by accident as by design, our existing systems of intellectual property already include several doctrines that are difficult to explain unless the relevant intellectual property rights are recognized as partially advancing the goal of encouraging market experimentation.

For example, some observers have argued that productive American industry is being overrun by “patent trolls,” companies that produce no actual products but merely obtain and enforce patents.  The conventional theory of the patent system maintains that the basic quid pro quo for obtaining exclusive patent rights is the disclosure of the technology set forth in the patent document itself.  Under that theory, the concern about patent trolls seems inexplicable:  Someone who makes a sufficient disclosure and obtains a valid patent cannot be gaming the system.  But if the patent system is recognized as having mixed goals—both spurring the disclosure of technological information and fostering actual investment in real-world market experiments—then the concern over patent trolls makes sense.  The law should be more generous to firms that have both made technological disclosure in patent documents and risked assets in launching new businesses based on the technology.

A.     Market Innovation and Trademark Law

Trademark law can also be understood as advancing the goal of market experimentation.  The generally accepted purpose of trademark theory is to economize on consumer search costs. But trademarks are also central to allowing an entrant into a new market to maintain its market share in the face of competition.  If, for example, any competitor were permitted to use the label “Netflix” to describe services similar to Netflix’s, then Netflix likely would lose its market share much more rapidly, since the Netflix product would seem less distinctive and attractive.  The goal of encouraging market experiments like Netflix therefore provides an additional justification for trademark protection.

In their classic analysis of the economics of trademark law, William Landes and Richard Posner note that prices for brand-named goods have “seemed to some economists and more lawyers an example of the power of brand advertising to bamboozle the public and thereby promote monopoly.”7 Landes and Posner point out that the presence of differentiated prices between brand and generic products may not imply deadweight costs if the basis for the difference is that consumers are paying extra for guarantees of high-quality manufacturing or to avoid the expense of determining whether alternatives are in fact of equal quality.  In the pharmaceuticals market, though, the absence of empirical evidence showing that generic drugs have dramatically inferior quality as compared to brand drugs makes Landes and Posner’s empirical claim difficult to verify.

The market experimentation theory provides a more satisfactory response to the argument that there are costs associated with trademark protection.  Trademarks can serve a useful function even if many consumers, acting solely in their own private interests, are irrationally loyal to a brand.  If there is some static inefficiency to consumers’ preferences for brand names, early market entrants will expect a greater market share and, therefore, engage in more market experimentation.

Market experimentation theory thus may help explain our courts’ general reluctance to commit “genericide” by concluding that a trademark has become generic.  Everyday language suggests that many consumers use “Band-Aid” and “Rollerblades” generically, and yet they persist as trademarks.  Some scholars have argued that this aspect of the law cannot be reconciled with trademark’s goal of reducing consumer search costs.  The market experimentation theory suggests that even so, the first-mover advantages provided by such trademarks encourage market entry for future potential products.  Similarly, protection for branding, such as with Major League Baseball caps, does not seem explained by the theory that consumers associate the New York Mets logo with high quality millinery.  But it may make sense to protect these hats to encourage creation of other businesses that might furnish branding opportunities later.

B.     Market Innovation and Trade Secret Law

Market experimentation theory also helps make sense of trade secret law.  A puzzle of trade secret law is why it extends to nontechnological information, such as customer lists and sales figures.  This aspect of trade secret law is difficult to explain through its two traditional justifications:  that it provides incentives to produce technological information or that it reduces social costs associated with “self-help” remedies.  Indeed, conventional theories of competition might suggest that there should be an affirmative obligation to disclose such information, perhaps justifying a Freedom of Information Act for the private sector.

Explaining these aspects of trade secret law as an appropriate social subsidy to encourage market experimentation makes for a more solid foundation.  The law protects whatever business data can be hidden, thus discouraging subsequent entry, increasing a first entrant’s expected share of rents, and creating stronger incentives for the market experiments that produce the data.  Trade secret law may be overinclusive—it protects copycat businesses too—but, in general, innovators are the businesses that have the most information worth protecting.

In short, existing trademark and trade secret law can be more fully explained by consideration of market experimentation.

C.     Market Innovation and Patent Law

Patent law, on the other hand, does not provide well-tailored incentives for market experimentation.  The problem is that the granting of patents is not dependent upon the extent to which an innovation will promote market experimentation.  This holds true even with—perhaps especially with—the modern advent of so-called business method patents.  Such patents might be granted for innovations that are based on new technologies but that would have been created and marketed even without patent protection.  On the other hand, such patents could be refused for being technologically obvious, non-novel, or outside the scope of patentable subject matter, even though some grant of economic exclusivity might be needed to encourage the risky testing of such ideas in the marketplace.

A more reasonable system would be open to granting exclusive rights based on market, rather than technological innovation, but only in those cases where the market success is truly doubtful—in other words, where the market innovation is nonobvious.  This general problem has been recognized only in narrow areas, such as in protection for “orphan drugs.”8 Part of the concern with such drugs is that no firm may have the incentive to invest in the Food and Drug Administration’s approval process if they won’t receive exclusive rights, and so the Orphan Drug Act provides a limited term of exclusivity even for drugs that otherwise have fallen into the public domain.9 Legal scholars have not previously recognized that the logic of the Orphan Drug Act could be more broadly applicable to the patent system.10


Why don’t intellectual property theory and property rights theory more generally contemplate property protection for market experimentation?  It is not because information about market success is a type of information that inherently needs no protection.  Rather, it may be because our property rights institutions, as currently designed, are poorly suited to afford such protection.  Institutional limitations have become theoretical limitations that inhibit a clear conceptualization of the entire field of intellectual property.  With the theoretical limitations exposed, perhaps intellectual property scholars can begin to consider whether it might be possible to imagine new institutions or smaller institutional changes that might improve protection for market experimentation.dingbat


Copyright © 2009 New York University Law Review.

Michael Abramowicz is Professor of Law at George Washington University Law School.

John F. Duffy is Oswald Symister Colclough Research Professor of Law at George Washington University Law School.

The authors thank Martin Adelman, Shamnad Basheer, Robert Brauneis, Richard Pierce, David Post, Michael Risch, Roger Schechter, and Alex Tabborak, as well as participants in workshops at the George Mason University School of Law, the Boston University School of Law, the University of Pennsylvania Law School, and the University of Michigan Law School for helpful comments and suggestions. We also thank Megan Keane for excellent research assistance.

This Editorial is based on the following full-length Article:   Michael Abromowicz & John F. Duffy, Intellectual Property For Market Experimentation, 83 N.Y.U. L. REV. 337 (2008).

  1. FRANK H. KNIGHT, RISK, UNCERTAINTY, AND PROFIT 372 (Sentry Press 1964) (1921).
  2. Id. at 372-73.
  3. Scott Shane & S. Venkataraman, The Promise of Entrepreneurship as a Field of Research, 25 ACAD. MGMT. REV. 217, 221 (2000).
  4. Id.
  5. Id.
  6. F.A. Hayek, The Use of Knowledge in Society, 35 AM. ECON. REV. 519, 524 (1945).
  7. William M. Landes & Richard A. Posner, Trademark Law: An Economic Perspective, 30 J.L. & ECON. 265, 274 (1987).
  8. The more familiar definition of “orphan drug” is a drug for a disease that “affects less than 200,000 persons in the United States,” but the statute also includes cases in “which there is no reasonable expectation that the cost of developing . . . a drug for such a disease or condition will be recovered from sales.” 21 U.S.C. § 360bb(a)(2) (2006).
  9. See id. § 360cc(a).
  10. A recent article, however, recognizes that the logic underlying the Orphan Drug Act could be applied more broadly in the pharmaceutical context. Benjamin N. Roin, Unpatentable Drugs and the Standards of Patentability, 87 TEX. L. REV. (forthcoming 2009), available at http://ssrn.com/abstract-1127742.

Post a Comment (all fields are required)

You must be logged in to post a comment.