*
Intellectual property ordinarily passes from private ownership to the public domain, but it is possible to imagine regimes that would selectively convert some public domain property back to private ownership. Privatization might be useful in some cases to promote further scientific development of intellectual property, to provide incentives for unpatentable improvements, to encourage commercial experiments, to generate increased advertising or marketing, and to prevent overuse. Privatization has drawbacks too, including both the familiar drawbacks of private ownership and the rent-seeking that a system of privatization might entail. If selective privatization is appropriate, auctions may be the best mechanism for privatizing rights in a way that limits rent-seeking, and information markets might provide a useful means for determining when privatization is appropriate.
.............................................................................................4
c. Uncertainty of Commercial Success ...........................................................9
*
Associate Professor, George Washington University Law School. J.D., Yale Law School; B.A., Amherst College.
PRIVATIZING THE PUBLIC DOMAIN
4. Identification of Optimal Term and Optimal Scope .........................................30
...............................33
The path from the private domain to the public domain is almost always a one-way street.
1 Intellectual property rights may be held privately for limited periods of time, and then they fall into the public domain. Once in the public domain, an intellectual property right ordinarily cannot be reclaimed for private ownership. Typically, this is as it should be. At least in theory, intellectual property terms are designed to provide an adequate reward for the creation of the underlying intellectual property. Even if the designated term is suboptimally short for a particular category of intellectual property, once the intellectual property is created, retroactive increases in terms produce windfalls for private actors rather than social gains.
2 According to the conventional wisdom, society ex post is better off with intellectual property in the public domain rather than in private hands. Public ownership eliminates the transactions costs associated with identifying and negotiating with the owners of prior intellectual property rights, and goods and services utilizing the underlying intellectual property can be sold for a lower price, increasing access to the relevant goods.
This simple case for keeping the public domain public, however, is increasingly vulnerable. Ever since Edmund Kitch published his prospect theory of patent law,
3
intellectual property commentators have recognized that there might be ex post benefits to patent protection.
A patent owner, Kitch recognized, could “coordinate the search for technological and market enhancements of the patent’s value,” and so a patent could minimize the extent of duplicative investment in research and development activities and reduce transactions costs among
1
Two exceptions are discussed below: orphan drugs and abandoned trademarks. See infra Part I.A.1; Part I.C.
2
See, e.g.
, John F. Duffy, Intellectual Property Isolationism and the Average Cost Thesis , 83 T EX .
L.
R EV . 1077, 1094 (2005).
3
Edmund W. Kitch, The Nature and Function of the Patent System , 20 J.L.
& E
CON
. 265 (1977).
2
PRIVATIZING THE PUBLIC DOMAIN competing inventors.
4 Prospect theory remains controversial, 5 but even skeptics about prospect theory’s general applicability acknowledge that there may be some contexts in which patent protection does provide ex post social benefits.
6 Meanwhile, copyright scholars have recently recognized the possibility that there might be ex post benefits of copyrights as well.
7
Most prominently, William Landes and Richard Posner have suggested a regime of “indefinitely renewable copyrights” that conceivably would allow valuable copyrights forever to escape falling into the private domain.
8
Even the literature that suggests broad or long intellectual property rights does not endorse turning the road between the private and the public domains into a two-way street. This
Article will suggest that there may be circumstances in which social welfare would benefit from transferring intellectual property rights from the public domain back into the private domain. A rule barring such moves may nonetheless be justified. It may be difficult to establish a system that would determine with sufficient precision just which intellectual property rights that have fallen into the public domain ought to be privatized, and the costs of false positives might exceed the cost of false negatives inherent in a regime in which such moves are uniformly prohibited.
Moreover, the apparatus for determining which rights to privatize might itself generate various kinds of costs. Perhaps the existing bright-line approach—once public, always public—is optimal, but alternatives have received little consideration. This Article attempts to make the best possible case for the possibility of a legal regime that does allow occasional privatization of the public domain. It does not, however, offer a firm empirical claim about whether such a regime would represent a welfare improvement.
The Article will proceed as follows. Part I will seek to identify various contexts in which social welfare might rise if intellectual property in the public domain were transferred to private
4
Id.
at 276.
5
See, e.g.
, Roger P. Merges & Richard R. Nelson, On the Complex Economics of Patent Scope , 90 C OLUM .
L.
R EV . 839, 908
(1990).
6
Mark Lemley, Ex Ante Versus Ex Post Justifications for Intellectual Property , 71 U.
C
HI
.
L.
R
EV
. 129, 141 (2004) (“Prospect theory is needed when control over subsequent development is a necessary part of the incentive to produce the pioneering invention in the first place, as is arguably true with pharmaceuticals.”).
7
See, e.g.
, Michael Abramowicz, A Theory of the Derivative Right and Related Doctrines , M INN .
L.
R EV . (forthcoming 2005).
8
William M. Landes & Richard A. Posner, Indefinitely Renewable Copyright , 70 U.
C
HI
.
L.
R
EV
. 471 (2003).
3
PRIVATIZING THE PUBLIC DOMAIN ownership. The most important applications appear to be in patent law, but applications in copyright and trademark are also considered. Part II will consider possible problems with transferring intellectual property rights to private ownership. These problems fall into two categories: possible disadvantages of private relative to public ownership, and costs of the process of identifying rights to be privatized and transferring those rights to private ownership.
The Part concludes by identifying several goals that a system for selectively privatizing the public domain should meet. Finally, Part III considers possible strategies for respecting these goals, focusing especially on the possible use of auctions and information markets.
I.
P OTENTIAL A PPLICATIONS
Why might we ever want to take something out of the public domain and place it in private hands? Not to incentivize the creation of the intellectual property in the first place; by assumption it already exists. There are numerous reasons, however, that placing intellectual property rights in private hands might be beneficial, and this Part considers them, saving for the next part both obvious and nonobvious caveats. Within each category of intellectual property, the reasons fall in two general categories: first, the possibility that the intellectual property might need development that no one will have sufficient incentive to provide if the intellectual property remains in the public domain; and second, the possibility that privatizing intellectual property might relieve some form of a congestion or overuse problem.
A.
Patent
1.
Privatizing to Encourage Development a.
Further Scientific Assessment
The possibility that a patent in the public domain may need development is illustrated well by a prominent example of a rare regime that does sometimes return intellectual property from the public domain, the Orphan Drug Act.
9
The sponsors of the Act were concerned
9
Orphan Drug Act, Pub. L. No. 97-414, 96 Stat. 2049 (1982) (codified as amended at 21 U.S.C. §§ 360aa-360ee (1994), 26
U.S.C. § 45C (Supp. II 1994), 42 U.S.C. § 236 (1994)). For a comprehensive overview of the Act, see Gary A. Pulsinelli, The
Orphan Drug Act: What’s Right with It , 15 S
ANTA
C
LARA
C
OMPUTER
& H
IGH
T
ECH
.
L.J. 299 (1999).
4
PRIVATIZING THE PUBLIC DOMAIN specifically about rare diseases and conditions, now defined as affecting fewer than 200,000 people in the United States,
10
but the Act also applies to situations in which “there is no reasonable expectation that the cost of developing and making available in the United States a drug for such disease or condition will be recovered from sales in the United States of such drug.” 11 Under the Act, a sponsor may request that the drug be designated as a rare drug, and if so designated, then once someone receives approval from the Food and Drug Administration
(FDA) to sell a new drug, no one else can receive such approval for seven years.
12 Congress’s concern was that for these drugs, no one would have a sufficient incentive to go through the FDA approval process, including the trouble of performing necessary studies about efficacy, in the absence of some market protection.
Both patented and unpatented drugs are eligible for this protection, 13 though the protection will be redundant for drugs with at least seven years of patent term remaining. The
Act is thus especially important where a drug could not receive patent protection, presumably because the drug previously was patented and since had fallen into the public domain. When a new use is found for an old drug, the discoverer of the new use can obtain a process patent on the new use.
14
But such process patents may be difficult to enforce, because generic drug companies can still sell drugs in the public domain, as long as they do not advertise the drugs as being for the protected use.
15
In addition, sometimes a study may verify the effectiveness of a previously conceived use for a drug, and such verification may be a necessary step for drug approval. The patent system does not protect obvious inventions, 16 and it does not protect nonobvious insights and information about inventions that are in the prior art.
17 The Orphan Drug Act provides for
10
21 U.S.C. § 360bb(a)(2)(A) (2000).
11
Id.
§ 360bb(a)(2)(B).
12
Id.
§ 360cc(a).
13
Congress decided to make patented drugs eligible in 1985. See Orphan Drug Amendments of 1985, Pub. L. No. 99-91, 99 Stat.
387 (1985).
14
See, e.g.
, Joseph M. Reisman, Physicians and Surgeons as Inventors: Reconciling Medical Process Patents and Medical
Ethics , 10 H
IGH
T
ECH
.
L.J. 355, 389-91 (1995).
15
See Rebecca S. Eisenberg, The Problem of New Uses , 5 Y
ALE
J.
H
EALTH
P
OL
’
Y
L.
& E
THICS
717, 720 (2005) (“The discovery of a new use for an old drug might support a patent on a method of treatment, but such a patent offers little effective protection against generic competition once the drug itself is off-patent and may lawfully be sold for an older, unpatented use.”).
16
35 U.S.C. § 103(b) (2000).
17
See, e.g.
, In re Cruciferous Sprout Litigation, 301 F.3d 1343 (Fed. Cir. 2002) (finding invalid a patent for administering
5
PRIVATIZING THE PUBLIC DOMAIN intellectual property protection (albeit not in the form of a patent) in cases in which no one might have an incentive to perform otherwise unprotectable studies.
The Orphan Drug Act is a special-purpose statute, yet its limitation in ordinary cases to
“rare diseases” represents at best an imperfect proxy for situations in which market exclusivity may be warranted even within the pharmaceutical field. A drug in the public domain that might benefit millions of people conceivably could be ignored by pharmaceutical companies in the absence of a possibility of some form of intellectual property protection. This will especially be a danger where the probability that a study turns out to be successful is expected to be low. It may be socially worthwhile for private actors to embark on many low-probability experiments, but in the absence of patent protection, the private value of some risky experiments may be much less than their social value. Even in the absence of the Orphan Drug Act, the first company to obtain approval for a drug will have some lead-time advantage over other companies,
18
but depending on the cost of performing the studies and applying to the FDA for approval, that lead time advantage might not be enough. Of course, the reverse may sometimes be true too; some drugs might be designated as orphan drugs even though lead time would have given the sponsors of those drugs a sufficient market advantage to encourage the relevant studies.
Even if the Orphan Drug Act could be redesigned to identify exactly those drugs for which market exclusivity would be a social benefit, there may be advantages to similar arrangements outside the pharmaceutical context. The problem of inadequate incentives to experiment seems most obvious in the pharmaceutical context, because of the heavy need for post-invention experimentation with drugs. Experimentation, however, occurs in all scientific fields.
19
Ordinarily, any studies needed to assess an invention’s effectiveness are likely to occur within the original patent term, and in those cases providing additional exclusivity will lead to no benefits. But there may be cases in which a particular invention is abandoned, only to turn out to cruciferous sprouts based on a study that identified previously unknown cancer-fighting properties of the sprouts).
18
See generally William T. Robinson et al., First-Mover Advantages from Pioneering New Markets: A Survey of Empirical
Evidence , 9 R
EV
.
I
NDUS
.
O
RG
. 1 (1994) (providing an overview of the literature documenting the existence of first-mover advantages).
19
After all, patent law’s experimental use exception, part of the timing rules underlying patent law, apply to all inventive contexts, not just pharmaceuticals. See, e.g.
, City of Elizabeth v. American Nicholson Pavement Co., 97 U.S. 126 (1877)
(providing the classic statement of the doctrine, in a case involving experimentation on pavement).
6
PRIVATIZING THE PUBLIC DOMAIN be potentially useful later. In these cases, additional experimentation may be needed to assess whether the invention will work as anticipated.
It is not clear how often this will be the case outside the pharmaceutical context. There may be many drugs with numerous potential applications that will not be initially apparent, or may initially seem unlikely but seem more likely later once other related applications of the drugs have proven successful. With most inventions, however, by the time a patent is approved, the scientific properties and effects of the invention will generally be known. But this may not always be the case. Many patents are abandoned because they appear to be commercially useless, even though they may reflect genuine scientific advances, and this abandonment may occur before all scientific feasibility issues are resolved. For example, a new multicolored paving technology might be developed and fall into the public domain before the developer performs sufficient tests to determine how well that pavement would wear over time, but the pavement might later appear promising because of sudden consumer interest in colored pavements. Perhaps more of the technologies underlying patents that previously appeared worthless would eventually find uses if it were possible to return patent protection once an invention turned out to have potential uses, but the suitability of those uses demanded experimentation. The issue would be difficult to study empirically, because under current rules, it is not possible to restore patent protection. b.
Unpatentable Improvements
When complete scientific testing of a particular invention has not occurred before the invention falls into the public domain, in some cases the invention might need improvements. As long as an inventor can obtain an improvement patent on a public domain invention, then there is no reason to move the underlying technology out of the public domain. There may be cases, however, in which an invention needs improvements, but these improvements themselves might be unpatentable. Douglas Lichtman, for example, has noted that some inventions, though requiring a great deal of work, may be unpatentable.
20
He notes, for example, that semiconductor
20
Douglas Gary Lichtman, The Economics of Innovation: Protecting Unpatentable Goods , 81 M
INN
.
L.
R
EV
. 693, 712-14
7
PRIVATIZING THE PUBLIC DOMAIN chips are unpatentable, and that the Semiconductor Chip Protection Act provides only limited protection.
21
If improvements are unpatentable, there may be little incentive to improve an underlying public domain invention. One solution, of course, would be to establish some type of intellectual property protection for the improvements themselves, but if this is not feasible, it may be sensible to allow the underlying invention to be removed from the public domain.
There are at least two contexts in which providing patent or other intellectual property protection to the improvements may not be feasible. First, each of the underlying improvements may be obvious. For good reasons, the patent system excludes obvious inventions, but even obvious inventions require some work to develop. The standard for determining whether an invention is obvious is whether it would have been obvious to PHOSITA, the hypothetical person having ordinary skill in the art. Even PHOSITA will not work for free, and so it costs money to develop obvious inventions. A second mover may be able to copy the obvious improvements, by hiring someone with less skill than the PHOSITA who created the improvements in the first place. Even when an invention is obvious, the best mode for using the invention may be expensive to determine, 22 and a second mover in some cases may be able to reverse engineer the invention to discover the best mode.
23
Anticipating the second mover, the potential first mover may decide not to make the improvements, or perhaps even not to enter the market at all.
Second, the patent system may be too expensive for some relatively modest improvements. Patent prosecution costs approximately between $10,000 and $30,000 per patent, and for any given improvement, 24 that might not be worthwhile. A large number of improvements to a single invention might collectively be quite valuable. Patenting all of these
(1997).
21
Id.
(citing Semiconductor Chip Protection Act of 1984, 17 U.S.C. §§ 901-14 (1994).
22
The courts have noted that when an inventor discovers a new best mode after receiving a patent, there is no duty to disclose that new best mode. See, e.g.
, In re Gay, 309 F.2d 769, 772 (Ct. Cust. & Pat. App. 1962). If a new best mode is discovered after a patent has fallen into the public domain, and the information about the best mode is not sufficiently novel to entitle the discoverer to a new patent, second movers may be able to free ride on the actions of first movers.
23
For an analysis of reverse engineering, see Pamela Samuelson & Suzanne Scotchmer, The Law and Economics of Reverse
Engineering , 111 Y
ALE
L.J. 1575 (2002).
24
Mark Lemley, Rational Ignorance at the Patent Office , 95 N
W
.
U.
L.
R
EV
. 1495, 1498-99 (2001) (estimating the cost of a patent prosecution to be between $10,000 and $30,000 per patent).
8
PRIVATIZING THE PUBLIC DOMAIN improvements in combination in a single patent claim may not be an effective strategy, because then a copier could use most but not all of the improvements without infringing the initial patent.
When an invention is under private ownership, the patent owner may make improvements to the invention that would be patentable but aren’t worth patenting, but no one will have such incentives anymore once the invention falls into the public domain. In theory, the patent process itself might be scaled down for relatively modest inventions, but it is difficult to design bureaucratic processes that are sensitive to the amount of attention that each decision needs. In some cases, it might be more straightforward to place the underlying invention back into private hands, so that an inventor will have incentives to improve it without obtaining protection on the individual improvements. c.
Uncertainty of Commercial Success
Perhaps the most important type of experimentation on public domain inventions that the patent system will not protect is market experimentation. Business method patents, of course, may be available for business methods to the extent that the patent applications meet the ordinary standards of the patent system. But commercial uncertainty is not a sufficient basis for patentability. Indeed, to the contrary, an invention is more likely to be found to be nonobvious if the inventor can show that there was a long-felt unmet need for that invention.
25 A business method may well be anticipated in the prior art or obvious, but its prospects of commercial success may be uncertain. Someone considering whether to become the first to employ the business method must then weigh competing considerations. Being first in the market may provide a lead-time advantage, 26 but there are second-mover advantages too, because second movers can learn from the business’s mistakes.
27 A business model may have some chance of failure and some chance of success, but in the absence of patent protection, the first to employ the model cannot appropriate the full value of success, because success may lead to additional
25
For a discussion, see Robert P. Merges, Commercial Success and Patent Standards: Economic Perspectives on Innovation , 76
C
AL
.
L.
R
EV
. 803, 828 (1988).
26
See supra
27
See F. Scott Kieff, Property Rights and Property Rules for Commercializing Inventions , 85 M
INN
.
L.
R
EV
. 697, 708-09 (2001).
9
PRIVATIZING THE PUBLIC DOMAIN entry. So, in some cases a business model that in expected value terms would produce social welfare benefits from being tried will not be tried. As a result, moving the business method into private hands may be worthwhile.
An analysis of whether to move patents and unpatented inventions from the public domain to private ownership overlaps the question of whether new forms of intellectual property ought to be created. The concern about second-mover advantages preventing action by a first mover may suggest that there should be some sort of general protection against competitive pressures for first movers, a possibility that I plan to consider more in later work. Even if it is decided, however, that, all in all, creation of an additional intellectual property institution is undesirable, it may be possible in some circumstances to move a particular invention into private hands. This invention, of course, need not be a business method patent. It may be that exploitation of a previously patented invention, or of an invention that could never or simply was never patented, will not occur unless the invention is returned to private ownership. A mechanism for revitalizing patents or for relaxing some of the requirements of patent law in limited circumstances might be the most efficient means of accomplishing this. d.
Advertising and Marketing
Sometimes, a public domain invention will be sufficiently attractive that at least one business will use or sell the invention, even absent any intellectual property protection.
Nonetheless, there can be social benefits to placing such inventions into private hands, if marketing of the invention will be necessary to maximize the invention’s value. Generic drug companies, for example, will not generally spend much money advertising the drugs that they offer, because of a free-rider problem. Each company has an incentive to free-ride off the other companies’ investments. If the generic drug were returned to private ownership, then the owner would have incentives to engage in advertising and marketing. If, as much economic research suggests, 28 advertising and marketing are important vehicles through which consumers identify
28
See generally Timothy J. Muris, 2000 S UP .
C T .
E CON .
R EV . 265, 293-302 (2000) (supporting this view); Phillip Nelson,
Information and Consumer Behavior , 78 J.
P
OL
.
E
CON
. 311 (1970); George Stigler, The Economics of Information , 69 J.
P
OL
.
E
CON
. 213 (1961).
10
PRIVATIZING THE PUBLIC DOMAIN products and services from which they might benefit, privatizing public domain patent rights may usefully promote consumer information.
Not all economists’ stories about advertising are positive, of course, and it is possible that advertising might be socially excessive. Yet not all stories about socially excessive advertising counsel against privatization of the public domain. A number of economists, for example, have complained that sometimes competitors in production of a particular product may waste resources by generating competing advertisements that carry little information.
29 This, however, actually states a case for monopoly power, since the monopolist may not have to advertise as much. Another story is that monopolists may advertise excessively to protect their market position.
30
Yet this does not clearly counsel against revitalizing intellectual property rights either, because intellectual property protection might obviate the need for such protective advertising.
The worrisome story is that the additional advertising that a monopolist might perform will lure consumers into purchasing products that they do not need, for example by altering consumer tastes.
31
For example, a monopolist drug manufacturer might convince consumers that they need an ineffective drug. This may be a particularly strong danger when agency costs inhere in the purchasing decision, as for example when prescription drugs are covered by health insurance. The most pessimistic scenario is that advertising in fact may encourage consumption of products that are harmful to consumers or to third parties,
32
and a slightly less pessimistic scenario is that advertising may promote use by both those who would benefit from and those who would suffer from use of the relevant products.
33 The debate on advertising, of course, is too extensive to review fully here, but theoretical models suggest that it is possible that, from a social perspective, a monopolist may advertise too much or too little.
34
The case for removal of
29
See, e.g.
, Avinash Dixit & Victor Norman, Advertising and Welfare: Another Reply , 11 B ELL J.
E CON . 753 (1980).
30
See, e.g.
, Jeffry M. Netter, Excessive Advertising: An Empirical Analysis , 30 J.
I
NDUS
.
E
CON
.
361, 361 (1982) (citing sources).
31
For a model along these lines, see Nicholas Kaldor, The Economic Aspects of Advertising , 18 R
EV
.
E
CON
.
S
TUD
. 1 (1950).
32
See, e.g.
, Henry Saffer, Alcohol Advertising and Motor Vehicle Fatalities , 79 R
EV
.
E
CON
.
S
TAT
. 431 (1997) (estimating that a ban on alcohol advertising would save approximately 2000 to 3000 lives per year).
33
See, e.g.
, Bayer Agrees to Alter Aspirin Advertising , http://www.oag.state.ny.us/press/2000/jan/jan11a_00.html
(reporting on the settlement of New York Attorney General Eliot Spitzer’s lawsuit against Bayer aspirin for misleading advertising, based on the theory that the advertising failed to make clear that only certain adults would likely benefit from Bayer aspirin).
34
See, e.g.
, Y. Kotowitz & F. Mathewson, Informative Advertising and Welfare , 69 A
M
.
E
CON
.
R
EV
. 284 (1979) (offering a model that can produce both results, depending in part on whether consumers who have positive experiences with the product
11
PRIVATIZING THE PUBLIC DOMAIN intellectual property from the public domain can be based only on the latter scenario, but that is likely not inevitable.
While privatizing inventions in the public domain may be one way of encouraging socially useful advertising, it is not the only approach. The simplest alternative is to rely on trademark. For example, because Bayer has a sufficiently strong market position in the aspirin market, it has some incentive to market the allegedly beneficial health consequence of aspirin, or at least of its brand of aspirin.
35 Even if such advertising succeeds in conveying to consumers the existence of a drug that might benefit them, the consumers might recognize, either through their own research or from their doctors, that the trademarked version is just one of the available drugs. When an individual market participant cannot in effect advertise its own product without also promoting the products of others, it will underproduce advertising, at least from the perspective of itself and its industry.
An alternative approach is for different producers within an industry to create a joint advertising campaign. Individual producers, however, might refuse to join such a campaign, preferring to free ride on the efforts of others. A possible solution to this is for a trade association to lobby the government to advertise for it, using the revenues from a special purpose tax imposed on all members of the industry. In the United States, the Supreme Court has held that this counts as government speech and thus imposes no burden under the First Amendment.
36
There are dangers to such programs, however, particularly when the program may help some firms and hurt others, but the firms that benefit manage to persuade the government to impose the tax on all of the firms. Moreover, generic advertising in general may not be as effective as brand advertising,
37
and the governmental or quasigovernmental entity that administers the advertising campaign may not seek to create as aggressive a campaign as a monopolist would.
Governmental coordination of industry advertising may nonetheless present more modest risks help to recruit additional consumers to it).
35
Under a recent settlement, see supra
note 33, Bayer is not permitted to imply that the benefits of its aspirin product are unique.
A problem with this prohibition is that it reduces Bayer’s incentive to advertise the benefits of aspirin.
36
37
See Johanns v. Livestock Mktg. Ass’n, 125 S. Ct. 2055 (2005).
See, e.g.
, Harry M. Kaiser & Donald J. Liu, The Effectiveness of Generic Versus Brand Advertising: The Case of U.S. Dairy
Promotion , 20 R
EV
.
A
GRIC
.
E
CON
. 69 (1998).
12
PRIVATIZING THE PUBLIC DOMAIN associated with governmental interference than governmental selection of inventions to remove from the public domain.
38
2.
Privatizing to Reduce Overuse a.
Excessive Improvements of Inventions
Ever since Kitch announced his prospect theory, 39 it has been clear that one potential benefit of patents is that they might discourage excessive improvement to inventions. Once a particular invention exists, inventors will have an incentive to compete to improve the invention, and in the absence of patent protection, they might compete in excess of the social optimum.
Whatever the potential private rents from improvement of the invention, third parties will enter the competition to create improvements until those private rents are expected to be zero, corresponding to a normal economic profit.
40 A patentee, however, will have at least some ability to control the inventive process and to limit excessive invention, because a patentee can “block” the developers of an improvement from practicing the invention.
41 Because inventors cannot capture the full social value of their inventions,
42
it is possible that there will be too little improvement on an invention even in the absence of a patent. But there may be some instances in which a patent reduces wasteful competition to improve inventions.
The problem, of course, is that patents eventually fall into the public domain. A patentee who has efficiently limited excessive improvement of the patent during the patent term will no longer be able to do so once the patent term expires. To be sure, there is a fine line between improvement of a patented product and use of the patented invention in some other product, and the danger of excessiveness may be greater, if it exists at all, in the former case. Should an
38
See infra Part II.B.3.
39
See Kitch, supra
40
Mark Grady and Jay Alexander theorize that patent law strikes a balance between the inefficient rent dissipation associate with product improvements and the rent dissipation associated with patent races. See Mark F. Grady & Jay I. Alexander, Patent Law and Rent Dissipation , 78 V A .
L.
R EV . 305 (1992).
41
42
See, e.g.
, John F. Duffy, Rethinking the Prospect Theory of Patents , 71 U.
C
HI
.
L.
R
EV
. 439, 448 (2004).
For empirical analyses of economic spillovers from patenting, see Timothy F. Bresnahan, Measuring the Spillovers from
Technical Advance: Mainframe Computers in Financial Services , 76 A M .
E CON .
R EV . 742, 753 (1986); Edwin Mansfield et al.,
Social and Private Rates of Return from Industrial Innovations , 91 Q.J.
E
CON
. 221, 234 (1977); M
ANUEL
T
RAJTENBERG
,
E
CONOMIC
A
NALYSIS OF
P
RODUCT
I
NNOVATION
: T
HE
C
ASE OF
CT S
CANNERS
11-44 (1990).
13
PRIVATIZING THE PUBLIC DOMAIN invention be subject to excessive improvement upon falling into the public domain, however, restoring the patent to private ownership may be an antidote.
Of course, the same logic applies in a case in which a technology is born in the public domain, for example because the inventor does not apply for a patent, or cannot obtain a patent because the invention fails to meet one of the criteria for invention. John Duffy has argued that courts should be particularly willing to declare inventions obvious despite an absence of anticipation in the prior art when those inventions emerge as responses to exogenous shocks that suddenly make a particular technology more economically attractive.
43
Although this approach may generally be sensible, there may be cases in which patents have net benefits ex post, if a patentee can better control inventive efforts. Perhaps the patent system cannot efficiently identify such inventions ex ante and should allow them to fall into the public domain, but excessive invention might be easier to identify once it is already taking place. b.
Negative Externalities
Removal of a patent from the public domain may be justified when a product based on the patent is overused. One overuse scenario involves negative network externalities, the possibility that one person’s use of a product may make that product less valuable to others. Eric
Kades has recently argued that markets for antibiotics exhibit such negative network externalities, because use of antibiotics leads to bacterial resistance through evolution.
44
Kades thus suggests that antibiotics receive especially long patent terms, and also considers the possibility that antibiotics with expired patents might be repatented, for example through an auction mechanism.
45 Kades recognizes the possibility of alternative governmental interventions, such as the imposition of Pigovian taxes, 46 but an advantage of the patent approach is that the patent winner would have an incentive to trade off present and future use of the product, and might better be able than governmental officials to resist public pressure to lower prices in the
43
See, e.g.
, Duffy, supra
44
See Eric Kades, Preserving a Precious Resource: Rationalizing the Use of Antibiotics , 99 N
W
.
U.
L.
R
EV
. 611 (2005).
45
Id.
at 643-53.
46
Id.
at 635-43.
14
PRIVATIZING THE PUBLIC DOMAIN event of an epidemic.
47 Governmental officials might not have sufficient economic motivation to optimize the use of an antibiotic over its lifespan.
A similar argument could be used in any context in which a resource is exhaustible. It has long been understood that privatizing a fishery will reduce overfishing of that particular fishery,
48 but where privatization is impractical, a patent on a particular species of fish might provide a centralized alternative. Of course, this is a second-best solution. Private ownership of each of many fisheries should give each fishery owner appropriate incentives to conserve fish without creating monopoly pricing in the fish market. Similarly, OPEC is generally seen as harmful rather than beneficial to consumers, even though oil may be an exhaustible resource. Each owner of an oil field will have incentives to leave oil in the ground or store it rather than selling it immediately if the owner anticipates a shortage that will make oil extremely valuable, so conservation incentives already exist. It is only in the case of exhaustible intellectual property, which if in the public domain can be produced by anyone but to the detriment of all, that privatization of individual reserves of an exhaustible resource is not a policy option.
Nonetheless, returning inventions to private ownership is potentially useful in any context in which the inventions have negative externalities. Consider, for example, cigarettes. Secondhand smoke imposes costs on third parties, and if consumers irrationally discount the probability that they might become addicted, they may be too willing to smart smoking.
49
If so, privatizing a patent to tobacco might increase social welfare. The argument above concerning advertising suggests that this is not inherently so; a monopolist tobacco producer might advertise more than competitive producers, and this might increase the number of smokers. It is possible, however, that a tobacco patent might increase tobacco prices and eventually decrease use.
An alternative, of course, is to impose a sin tax on tobacco. The patent approach might, however, have lower transactions costs. It would eliminate the need for tax collection (except
47
Id.
at 656-57.
48
See, e.g.
, H. Scott Gordon, The Economic Theory of a Common-Property Resource: The Fishery , 62 J.
P
OL
.
E
CON
. 124, 130-
131 (1954).
49
See, e.g.
, Jon D. Hanson & Kyle D. Logue, The Costs of Cigarettes: The Economic Case for Ex Post Incentive-Based
Regulation , 107 Y
ALE
L.J. 1163, 1181-223 (1998) (noting informational difficulties that might lead consumers to make suboptimal decisions about smoking).
15
PRIVATIZING THE PUBLIC DOMAIN perhaps in the form of an auction used to distribute the patent initially 50 ), and the costs associated with licensing tobacco companies to use the patent are likely to be considerably lower. The patent approach creates a potentially efficient private enforcement mechanism, as the patentee would have incentives to identify and sue third parties seeking to sell black market cigarettes.
The patent approach, however, does not necessarily provide optimal pricing incentives, and so a significant benefit of the patent solution to the antibiotic problem would be lost. On the other hand, if the patent approach does not lead to a sufficiently large drop in the use of tobacco, the government could also impose a tax on tobacco use, though the creation of a monopolist might make it easier for the industry to resist efficient taxation.
Patents on harmful inventions (or inventions that cause both harms and benefits) may be particularly useful if the government wishes to impose liability on companies for the harmful consequences of those inventions. For example, Jon Hanson and Kyle Logue have suggested strict liability for cigarette manufacturers, on the ground that this would improve their incentives in designing safer cigarettes and in pricing their products.
51
A potential problem with their approach is the need to identify the manufacturer of the particular cigarettes that harmed individual smokers.
52
Awarding a patent on tobacco, coupled with the requirement that the patentee be strictly liable for the consequences of cigarette use, would ease this administrative problem. The patentee would charge high licensing fees to pay eventual judgments, and the litigation process would be more consolidated than it otherwise would be.
B.
Copyright
Some of the same considerations that may make a case for restoring patents to the public domain may be applied in the case of expired copyrights. Copyrights in general, however, seem less likely to be good candidates for rescue from the public domain, for two reasons. First, the
50
See infra Part III.A.
51
Hanson & Logue, supra
52
Id.
at 1287-91 (discussing challenges in assessing causation). Hanson and Logue, in a separate paper with an additional coauthor, describe a “smokers compensation” system designed to address these problems. See Jon D. Hanson, Kyle D. Logue &
Michael S. Zamore, Smokers’ Compensation: Toward a Blueprint for Federal Regulation of Cigarette Manufacturers , 22 S.
I
LL
.
U.
L.J.
519 (1998).
16
PRIVATIZING THE PUBLIC DOMAIN standard for copyright protection is low, 53 and no formalities are required to obtain its benefit.
54 It is true, of course, that copyright has some controversial exclusions, most notably the exclusion of facts, 55 but this again is more properly a question of whether copyright scope should be extended than of whether public domain works should be privatized. Second, the copyright term is very long. By the time works fall into the public domain under the existing term, they are likely to have relatively little value. Nonetheless, the theoretical possibility that restoration might be efficient in some cases remains, and this possibility might be particularly relevant if copyright terms were shortened in the future, as some have proposed for certain classes of goods, such as computer software.
56
1.
Marketing and Advertising
The possibility that uncopyrighted works might receive too little marketing is parallel to the analogous argument for patents.
57
William Landes and Richard Posner offer a straightforward example of this possibility:
Imagine a novel published many years ago in which copyright has expired. The novelist is rediscovered and there is a surge in demand for his novels. Since no publisher could establish a property right in them, the incentives of publishers to publish and promote them might well be inadequate from a social standpoint.
58
They note that the problem will be particularly acute with public domain authors whose works have only a low probability of success.
59
Although Landes and Posner use these points primarily in favor of an argument for potentially allowing renewable copyright terms, they recognize that “[a] complete solution would require that the ‘resurrectors’ of old works on which
53
See, e.g.
, Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53 (1884) (finding a photograph to be sufficiently original to receive copyright protection).
54
See 17 U.S.C. § 401 (2000) (clarifying that copyright notice is optional, but will prevent a defendant from mitigating damages through a defense of innocent infringement).
55
See, e.g.
, Feist Publications, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340 (1991) (finding a telephone directory to be unprotectable).
56
See, e.g.
, Mark Aaron Paley, A Model Software Petite Patent Act , 12 S ANTA C LARA C OMPUTER & H IGH T ECH .
L.J. 301, 317
(1996).
57
See supra Part I.B.1.
58
Landes & Posner, supra
59
Id. at 489.
17
PRIVATIZING THE PUBLIC DOMAIN copyright had expired without renewal, like finders in the law of real property, be allowed to obtain copyright in those works.”
60
The same empirical questions about the efficiency of a monopolist’s advertising expenditures are relevant in the copyright context as well. The point, however, may be yet more obvious. With competition among publishers, any particular title will likely sell for close to marginal cost, and margins will likely not be large enough to lead any publishers to engage in advertising. In the patent context, dominant trademarks (such as Bayer for aspirin) provided some incentive to advertise, but trademarks are less likely to play such a role for copyright, because each revived copyrighted work and creator has an independent (but public domain) trademark. There may be circumstances in which a publisher will create a series of revived works—Penguin Classics, for example—and then invest somewhat in marketing those works as a group, but the incentive to seek to revive any particular forgotten work or author is low.
Restoring copyright will increase those incentives, though if excessive advertising is a danger, it is possible that the increase in incentives will not be socially optimal.
The copyright context, however, highlights an empirical caveat about revitalizing abandoned intellectual property that seems less prominent in the patent context. Owners of intellectual property do not have monopolies, but rather compete in monopolistically competitive markets. Most if not all patents have substitutes of one kind of another, and copyright markets are overflowing with potential substitutes. The product differentiation literature makes clear that entry into a monopolistically competitive market of another product will not necessarily improve welfare.
61 If the anticipated revenues for a new product will come largely at the expense of existing products, the product may be produced even if the additional benefits to consumers are less than the cost of its production, since much of the cost will be borne by existing producers. In competitive markets, we dismiss such “demand diversion” or “business stealing” as mere pecuniary externalities without welfare consequences, but the product differentiation models of
60
Id.
at 493-94. Landes and Posner note the possibility that there might be excessive incentives to “bank” public domain works and suggest several antidotes, though interestingly they do not suggest the antidote of an auction. See infra Part III.A (discussing auctions).
61
See, e.g.
, Steven C. Salop , Monopolistic Competition with Outside Goods , 10 B
ELL
J.
E
CON
. 141 (1979).
18
PRIVATIZING THE PUBLIC DOMAIN spatial competition show that this is too facile an assumption in the context of monopolistic competition.
62
Perhaps it would be better if potential entrants diverted their energies into competitive fields in which demand diversion was less salient.
Because the models also allow for the possibility that the standard view that entry into any particular differentiated product market will raise consumer welfare, 63 it is difficult to draw firm conclusions about the benefits of entry. Economists should be particularly concerned about the danger of excess entry in intellectual property markets, because of the low marginal cost of reproducing intellectual property,
64
but even in copyright, the heterogeneity and complexity of differentiated product markets may it difficult to assess copyright policies, including the optimality of revitalizing existing works. The cost of revitalizing and marketing a public domain work can be seen as akin to the cost of entering a differentiated product market, and this may be a socially wasted expenditure if, to take the extreme possibility, the expenditures simply shift consumers from some works to others from which they derive no additional utility. On the other hand, even if there is excessive entry into copyright markets, revitalization of an old work may be cheaper than creation of a new work, so if revitalization of one work indirectly discourages creation of a new work, welfare might still rise.
There are yet more complications, such as the question of whether consumption of copyrighted works may produce other externalities. Perhaps classics better promote virtue than newer works, and more virtuous citizens will end up contributing more to society as a whole.
Many arguments along these lines or pointing in the reverse direction may easily be constructed, but they are difficult to evaluate. There also remain uncertainties about which public domain works are the best candidates for copyright restoration. Landes and Posner seem to assume that such a regime would focus on forgotten works, yet it may be that well-remembered works have far greater potential, but potential that has not been fully captured. We remember William
62
For a discussion, see Michael Abramowicz, An Industrial Organization Approach to Copyright Law , 46 W
M
.
& M
ARY
L.
R
EV
.
33 (2004).
63
See, e.g.
, Avinash K. Dixit & Joseph E. Stiglitz, Monopolistic Competition and Optimum Product Diversity , 67 A
M
.
E
CON
.
R EV . 297 (1977).
64
See Abramowicz, supra
19
PRIVATIZING THE PUBLIC DOMAIN
Shakespeare, but perhaps his plays would be read and performed even more if a copyright holder to the plays had incentives to generate excitement about the Bard currently reserved for the likes of Harry Potter.
2.
Congestion Externalities
Landes and Posner also worry about the possibility that overuse of copyrighted works might lead to “congestion externalities.” They argue, for example, that if Mickey Mouse were to fall into the public domain, we might have so much Mickey Mouse that the public’s demand for
Mickey might fall rather than rise. “Not only would the public rapidly tire of Mickey Mouse,” they argue, “but his image would be blurred, as some authors portrayed him as a Casanova, others as catmeat, others as an animal-rights advocate, still others as the henpecked husband of
Minnie.” 65 The argument is parallel to the argument about antibiotics in the patent context. If
Mickey is an exhaustible resource, then placing Mickey in private hands will provide the copyright owner appropriate incentives to use the resource relatively slowly. This may provide an argument, at least in some contexts, both for very long copyright terms and for restoring some copyrights to private ownership.
As with all of the contexts assessed in this Article, the case in which restoration seems justified is likely to be the exception rather than the rule. The case may be strongest for copyrighted characters and derivative works employing these characters, because some degree of scarcity may help maximize consumer buzz when a new work emerges. Copyright’s strong derivative right, the exclusive right to create adaptations of a copyrighted work, may give a copyright owner sufficient time to make the sequel that everyone has waited for, rather than rushing to compete with second-rate adaptations.
66 The argument seems much weaker when the question is whether there might be excessive distribution of any particular copyrighted work.
67
The greater danger seems likely to be that there may be too many works of a particular type.
Although an entire genre of copyrighted works may be beset by excessive entry, stemming such
65
Landes & Posner, supra
66
See Abramowicz, supra
note 7 (providing an extended version of this argument).
67
See Abramowicz, supra
20
PRIVATIZING THE PUBLIC DOMAIN entry may be more feasible when a single copyright, such as that over Mickey Mouse, defines the relevant field.
3.
Uncopyrightable Compilations
Selective transfer of rights to private ownership potentially might be useful not merely in cases in which copyright has expired, but also in cases in which copyright never could be obtained. In the Feist case, the Supreme Court held that there is no copyright in facts, even where a great deal of effort was expended in assembling a compilation of facts, if the compilation lacked originality.
68
Even if this rule is justified as a general matter of copyright policy, it might make sense to provide for deviations from it in specific cases in which copiers have such significant advantages that there appears to be underdevelopment of databases.
Underdevelopment might mean that although there is a demonstrated need for a particular compilation, no one has assembled the compilation, or it might mean that no one appears to be updating or improving a compilation because of the risk of competition.
The arguments for providing protection in these situations might appear to be congruent to the arguments for extending copyright generally to compilations of facts. The rules of copyrightability, however, generally depend on the characteristics of the work for which copyright is sought, rather than on conditions in particular copyright markets. A regime of selective intellectual property restoration, however, presumably would need to be sensitive to market conditions. Thus, if such a project has the potential to improve economic efficiency, it also may be sensible to consider extending it to selective intellectual property grants where intellectual property ordinarily might not be available. Of course, such a regime might mean that some database compilations might receive protection at their creation, or potentially even before their creation, so selective copyright grants will lead to the broadening of the copyright umbrella.
68
See supra
21
PRIVATIZING THE PUBLIC DOMAIN
C.
Trademarks
Trademark is the area of intellectual property in which the case for restoration of lost rights is weakest. The reason is that trademark rights have at least the potential to be perpetual, as long as a trademark owner continues to use the trademark in commerce. There is thus relatively little reason to worry that valuable trademarks will fall into the public domain.
Trademark nonetheless provides an important symmetry to the analysis here, in part because trademark, in contrast to patent and copyright, does provide a vehicle through which intellectual property that falls into the public domain can be restored to private ownership: the doctrine of abandonment. A trademark is deemed abandoned “[w]hen its use has been discontinued with intent not to resume such use.” 69
In contrast to patent law, where an abandoned patent remains in the public domain, anyone may be able to resuscitate an abandoned trademark and then receive trademark protection for it.
70
The theory, presumably, is that abandoned trademarks that still appear to have some positive association with the public may nonetheless remain useless unless someone can take dominion over it. A trademark that has some positive goodwill associated with it, or that does not but happens to provide a good fit for a new product or service, will nonetheless not be adopted if the adopter expects others to be able to harness whatever goodwill the new product or service provides. But it is possible to imagine leaving a trademark in the public domain, allowing anyone to harness its existing goodwill and create (or destroy) additional goodwill, sharing the mark in common.
71
The theory that it is justifiable to return the trademark to private ownership recognizes that common ownership likely will not optimize its value, because private ownership maximizes development of the trademark. Thus, at least superficially, the trademark abandonment issue is parallel to the patent and copyright contexts.
69
15 U.S.C. § 1127 (2000).
70
Perhaps the most famous case of an abandoned trademark is Major League Baseball Properties, Inc., v. Sed Non Olet
Denarius, Ltd.
, 817 F. Supp. 1103 (S.D.N.Y. 1993), in which a bar called “The Brooklyn Dodger” successfully persuaded the court that when the Dodgers moved to Brooklyn, they abandoned the “Brooklyn Dodger” mark. For an overview of trademark abandonment doctrine, see Annotation, Abandonment of Trademark or Tradename , 3 A.L.R.2d 1226 (2005). The general rule is that “[i]f some person uses the abandoned mark or name in a particular area to the exclusion of others he gains title to it as though he were the original owner of it; but if two or more persons adopt it at about the same time neither gains title to it.” Id.
§ 2.
71
For an analysis suggesting that trademarks, though privately owned, have some features of public goods, see David Barnes,
Trademarks as Congestible Public Goods (2005) (unpublished manuscript, on file with author).
22
PRIVATIZING THE PUBLIC DOMAIN
What accounts for this difference in treatment of abandoned intellectual property? The benefits of keeping private control over trademarks are relatively high. Trademarks need development, either through explicit advertising or through production of a good or service in a way that produces goodwill. Although patents and copyrights also may benefit from marketing activities, a trademark will have virtually no value absent some effort to build a positive association for consumers with the trademark. At the same time, the benefits of having trademarks in the public domain are relatively low. Intellectual property underlying lapsed patents and copyrights may continue to have some value even if no marketing activity directly supports them. Some companies may well have sufficient incentives to sell formerly patented and copyrighted goods even in the absence of intellectual property protection. But consumers do not purchase trademarks; they purchase the goods or services the trademarks represent. There will be few contexts in which a trademark will be used commercially at all if the trademark sits in the public domain.
There is at least one context in which a trademark may fall into the public domain and yet not be recoverable, and that is when the trademark has become generic.
72 When this is the case, because the public has a strong association between the former trademark and the class of goods, it would be difficult for competitors to describe their products to the public without using the trademark. Thus, the trademark does have value in the public domain. If the law’s treatment of generic marks is to be justified, this value must be considerable, because there are obvious inefficiencies associated with the doctrine. A mark holder may be less willing to invest in the mark if there is a risk that the mark may become generic, and a trademark holder may engage in advertising expenditures simply to prevent a mark from becoming generic.
73
Such expenditures might well be inefficient, if the trademark holder would not have engaged in them in the absence of a risk of losing the trademark.
72
See, e.g.
, The Murphy Door Bed Co., Inc. v. Interior Sleep Sys., Inc., 874 F.2d 95, 101 (2d Cir. 1989) (concluding that a trademark had become generic).
73
David H. Melilli, A Good Mark Is Hard to Find, Even Harder to Protect Against Genericide , 12 J. C ONTEMP .
L EGAL I SSUES
185, 187-88 (2001) (discussing Xerox’s advertising campaign to inform the public that “Xerox” is a trademark rather than a generic term).
23
PRIVATIZING THE PUBLIC DOMAIN
Conceivably, generic trademarks might be recovered from the public domain.
74 A case for allowing such recovery might be that the generic term itself could benefit from advertising or marketing. Usually, however, this would be the case because the public needs to be informed about the benefits of the underlying technology, not because the trademark itself needs development, since by assumption the trademark has already become a generic term for the product. Thus, if there is a case to be made that advertising and marketing should be increased for an unpatented product, it might make more sense to privatize a patent covering the relevant technology than to privatize the trademark. Privatizing the trademark might be ineffective in any event, because eventually the public will likely adapt to an alternative generic name for the product.
II.
P ROBLEMS
Part I has identified some advantages of moving intellectual property out of the public domain, and it has hinted simultaneously at some problems with such moves. This part will identify those problems more clearly, explaining why at least in the ordinary case, privatization of the public domain is not justified. The most obvious reason is that usually society will gain more from leaving the intellectual property in the public domain, because there are a variety of considerations that will outweigh those sketched out above. These reasons are sketched in Part
II.A. Even where property ideally should be privately owned, there may be good reasons to leave it in the public domain, and Part II.B summarizes these.
A.
Problems Associated with Private Ownership
1.
Deadweight Loss
The most obvious problem of private ownership of intellectual property is the problem of deadweight loss.
75 A patentee does not by virtue of the patent sell a larger number of inventions at the competitive price, but instead sells fewer at some higher monopoly price. Those consumers
74
Under current law, a generic mark can be recovered only if it ceases to be generic. See Singer Mfg. Co. v. Redlick, 109 F.
SSupp. 623 (S.D. Cal. 1952) (finding that “Singer” was no longer generic).
75
See, e.g.
, J
EAN
T
IROLE
, T
HE
T
HEORY OF
I
NDUSTRIAL
O
RGANIZATION
67-68 (1988) (identifying the deadweight loss created by monopoly pricing).
24
PRIVATIZING THE PUBLIC DOMAIN who value the invention at more than its marginal cost but less than this monopoly price thus will not be able to obtain the invention. And so too in copyright, indeed perhaps especially so in copyright, because the marginal cost of reproducing copyrighted works is generally low. Some commentators have noted that intellectual property owners may not quite be charging a monopoly price, because they face competition from substitutes, 76 but there is no denying that intellectual property protection will lead to some increase in price and decrease in quantity. This may be a worthwhile sacrifice given the dynamic benefits of the intellectual property system in providing incentives to create new works. Once works are already in the public domain, however, the reward justification makes little sense. If deadweight loss is to be tolerated, it ought to be because there is some countervailing social benefit to private ownership.
2.
Transactions Costs
Another advantage of the public domain is that it may reduce transactions costs. When intellectual property is protected by a patent or a copyright, someone who wishes to use the protected property must negotiate with the rights holder. As Landes and Posner note, sometimes it may be difficult to even find the rights holder, and these “tracing costs” increase as the length of a protection term is extended.
77 These costs are not a primary concern with restored intellectual property rights, since a company that goes to the trouble of acquiring an interest in property formerly in the public domain will likely have a continuing incentive to make its identity known.
78
The greater concern is with the interactions that take place after the property owner is found. In the copyright context, costs might not be all that high, because any particular new work will likely not require more than a handful of licenses.
79 In the patent context, however, any given new invention might combine large numbers of past inventions, and if all of those past
76
See Edmund W. Kitch, Elementary and Persistent Errors in the Economic Analysis of Intellectual Property , 53 V AND .
L.
R
EV
. 1727 (2000).
77
See Landes & Posner, supra
note 8, at 477 (“Enormous tracing costs would be incurred by any would-be publisher of a new
translation of the Iliad if the heirs of Homer could enforce copyright in the work . . . .”).
78
Landes and Posner note that “modest institutional reforms,” such as requiring reregistration of copyright owners ever few years, would largely solve problems associated with tracing costs. Id.
79
Id.
at 479 (noting that the problem might be more severe “for composite works, such as anthologies”).
25
PRIVATIZING THE PUBLIC DOMAIN inventions were privately owned, transactions costs could be prohibitive. An associated difficulty is the danger of an anticommons or a “patent thicket,” 80
as the owner of each of a number of patent inputs may have the power to prevent the use of a new invention and therefore may insist on a large portion of the surplus from that invention. There is considerable debate in the literature about how great a concern the anticommons is, 81 and whether collective licensing organizations may serve as an effective antidote.
82
Nonetheless, in both the patent and copyright contexts, placing a relatively small number of inventions back into private ownership for limited terms would not be likely to increase costs dramatically.
3.
Duplication from Inventing Around
The ability of the owner of an intellectual property right to employ a holdout strategy is constrained by the possibility that the party seeking to use the intellectual property will be able to do without it. In the copyright context, a movie producer seeking to incorporate a particular song might either create or license another song instead. In the patent context, the would-be licensee may incorporate an alternative technology already developed by another third party or created because of the refusal to license. The Federal Circuit has sometimes suggested that such
“inventing around” is a benefit of the patent system, 83 and it might well be, if the inventions provide social surplus that the inventor cannot capture. But inventing around may be socially inefficient wasteful duplication.
84
If creating a patent on the wheel leads some to reinvent the wheel rather than licensing it, the result may well be inefficient.
80
See, e.g.
, Michael A. Heller & Rebecca S. Eisenberg, Can Patents Deter Innovation? The Anti-Commons in Biomedical
Research , 280 S
CIENCE
698 (1998).
81
See, e.g.
, Richard A. Epstein & Bruce N. Kuhlik, Is There a Biomedical Anticommons?
, R EGULATION , Summer 2004, at 54
(offering skepticism); John P. Walsh et al., Working Through the Patent Problem , 299 S
CIENCE
1021 (2003) (reporting an empirical study indicating that patent rights were generally not preventing medical innovation).
82
See, e.g.
, Carl Shapiro, Navigating the Patent Thicket: Cross Licenses, Patent Pools and Standard Setting , in 1 I
NNOVATION
P
OL
’
Y
& E
CON
.
119 (2001).
83
See, e.g.
, State Indus., Inc. v. A.O. Smith Corp., 751 F.2d 1226, 1236 (Fed. Cir. 1985).
84
See, e.g.
, Lewis Kaplow, The Patent-Antitrust Intersection: A Reappraisal , 97 H
ARV
.
L.
R
EV
. 1813, 1869 (1984)
26
PRIVATIZING THE PUBLIC DOMAIN
4.
Motivation to Exploit
A final potential danger associated with privatizing the public domain is monopolist complacency. Some critics of Kitch’s prospect theory have worried that monopolists may be less inclined to inclined than those who face continued competition.
85
As a result, an invention might not be exploited as thoroughly under private ownership as under competitive ownership. For example, a patent owner might not improve the patent in as many ways as would occur if the invention were in the public domain, or might not seek out as many potentially profitable applications for the patented invention. Of course, we have seen that there may sometimes be a danger of excessive exploitation of intellectual property, but reduced exploitation could well reduce social welfare. A related concern is that a monopolist will not be as willing to take risks, approving safe projects but rejecting risky ones that might have more promise on average.
86
B.
Problems of Transferring to Private Ownership
The inherent problems associated with private relative to public ownership of intellectual property suggest that in the ordinary case, privatization will not be justified. This still, however, may leave some cases in which the benefits of private ownership are sufficiently great to outweigh these costs. Even where this is the case, however, a regime allowing transfer to private ownership may not be justified. The most obvious reason is the possibility of false positives.
Perhaps a legal regime leading to privatization will exact greater costs from inadvisable privatization than it will provide beneficial privatization. Even placing aside the inevitability of some errors, however, the privatization process itself might impose some costs. This section considers some of those costs.
1.
Rent-Seeking Behavior
A significant concern is rent-seeking behavior. If there is some positive rent to be earned from privatization of a public domain resource, these rents will likely be dissipated in the effort to obtain that resource. If the privatization decision were made by a private agency, for example,
85
See, e.g.
, Merges & Nelson, supra note x, at 887 (providing an example of how competition may reduce complacency).
86
See Tim Wu, Intellectual Property, Innovation, and Decision Architectures , V
A
.
L.
R
EV
. (forthcoming 2005).
27
PRIVATIZING THE PUBLIC DOMAIN private entities would spend money trying to persuade the agency to privatize as long as the marginal benefits of such expenditures exceeded the marginal costs. Some such money might be well spent. Just as society presumably benefits from litigants investing in legal services to refine arguments and educate judges, so too might one type of rent-seeking behavior entail preparation of detailed economic studies and analyses. But direct political contributions and other relatively inefficient forms of rent-seeking are also inevitable.
The literature on rent dissipation does not suggest, of course, that rent dissipation will always be complete. If one entity is in a better position than others to privatize a particular invention, then the race among the private parties is unlikely to fully dissipate the private benefits associated with privatization. And even if it did fully dissipate those benefits, there might be public benefits in excess of what is privately appropriable, a particularly strong concern in the patent context. Nonetheless, a sensible general strategy for discouraging inefficient rentseeking is to provide a regime that eliminates windfall gains from privatization. That is, a private party should have to pay the government for much of the expected private benefit. Exactly how much is uncertain, however, particular because there may be some social benefit to encouraging private parties to bring particular properties that would benefit from privatization to the government’s attention.
2.
Investment Predictability
A further risk of privatizing the public domain is that the prospect of such privatization might discourage some socially useful investments. For example, an inventor may be hesitant to incorporate a public domain invention into a new invention if there is a sufficiently high probability that the public domain invention will later be privatized. If the inventor were to incorporate the public domain invention and then privatization occurred, the inventor might have to pay patent damages, or worse might face an injunction, and because of the difficulty of changing products that have already been created, might end up paying far more to the new private patent owner than initially would have been necessary. Similarly, a publisher might not
28
PRIVATIZING THE PUBLIC DOMAIN be willing to undertake the modestly profitable project of publishing a new edition of a public domain work if the publisher expects the work to be privatized.
A related concern is that the owner of a trademark might not be willing to invest in the trademark if there is some chance that the underlying intellectual property will fall into private hands. Suppose, for example, that I am considering launching a new brand of aspirin, a project that will require me to invest in assets, including advertising but also the establishment of manufacturing plants, that cannot easily be liquidated. If aspirin suddenly becomes patented, then I may still seek a license from the new patentee, but the licensing fee will likely cost me a great deal of the expected surplus. And so, when privatization is possible, I will be less likely to invest in the new brand in the first place, perhaps lowering social welfare even if privatization never occurs.
There are several related kinds of antidotes to this problem, besides what is undoubtedly the best approach in many cases, not allowing privatization where there are substantial existing uses. One antidote is that there might be some form of grandfathering for those who have created intellectual property relying on the public domain before the privatization. Such grandfathering might be limited in duration, but long enough to allow at least some recovery of investment. Of course, this is at best only a partial solution. A second is that privatization might be limited only to cases in which few if any preexisting users will be affected. For example, the government might require agreement among a high percentage of current users, perhaps based on market share, before proceeding with a privatization plan. Those users might, for example, agree that whoever receives the patent will pay compensation to the others. And a third antidote would be to institute some form of compensation scheme directly. The compensation might be paid by the government, analogous to the Takings Clause. Or it might be paid directly by the patent holder.
It might be sensible to combine the second and third antidotes, allowing privatization only with substantial agreement of existing users and just compensation for the few hold-outs who refuse to go along. Naturally, the familiar problems of determining just compensation will limit the effectiveness of any such approach.
29
PRIVATIZING THE PUBLIC DOMAIN
A compensation plan ought not be too generous, because a too generous compensation plan will lead to entry in anticipation of intellectual property being removed from the public domain. If an invention is to be privatized, someone not previously in the market might hastily enter the market, in the hope of receiving greater compensation than the marketplace would allow if privatization did not take place. Of course, if compensation is based on expected future profits in the absence of privatization and this can be measured adequately, then there will be no perverse incentives. But expected future profits, especially of recent market entrants, may be difficult to measure, and so any compensation plan may rely on crude proxy variables, such as market share. In sum, if the compensation plan is too stingy, there will be inadequate entry, but if it is too generous, there will be excessive entry, and it may be difficult to get compensation just right.
3.
Identification of Optimal Right Holder
An additional danger of privatization is that even if private ownership is theoretically desirable, the privatization process might end up placing ownership into the wrong hands. This concern combines with the concern about rent-seeking, which is destructive not only because of the resources that are wasted in lobbying activities, but also because it might lead to poor decisions, both about whether to privatize and about who should own the right. Allowing sales of the right will greatly increase the chance that the intellectual property right eventually is owned by whoever can use the right most efficiently. But there are transactions costs to sales. Moreover, the ability to sell a right may encourage those who cannot use the right efficiently themselves to lobby for it to be assigned to them, increasing rent-seeking. Thus, ideally a privatized intellectual property right should be assigned through a mechanism that identifies the optimal intellectual property right owner effectively and in a manner that is relatively invulnerable to rent-seeking.
4.
Identification of Optimal Term and Optimal Scope
The designer of an intellectual property privatization system also must face some questions associated with the design of any intellectual property system, specifically the length
30
PRIVATIZING THE PUBLIC DOMAIN of the term and the scope of the right. There are arguments for both longer and shorter terms than would be customary for the relevant type of intellectual property. On one hand, the intellectual property may already have received a full term, and so only a shorter term may be needed. On the other hand, the same characteristics that make privatization optimal even placing aside the reward function of intellectual property protection may tend to counsel long intellectual property terms. For example, if privatization is desirable because the underlying intellectual property needs advertising and marketing, that may continue to be true well into the future, and a long term may be necessary to provide adequate returns to advertising and marketing activities.
There may also be arguments for providing broader or narrower scope to a privatized intellectual property right. For example, if a goal of privatization is to reduce wasteful duplicative efforts, then it may be sensible to offer relatively broad intellectual property rights as part of the privatization, because narrow rights would discourage designing around activities. On the other hand, providing broad intellectual property rights may aggravate concerns about interfering with investment expectations. For example, if the owner of a privatized patent receives rights not only to the patent itself but also to close economic substitutes of that patent, then developers of such substitutes will, in the absence of compensation, suffer economically.
Indeed, it may be sensible to offer narrower intellectual property scope to protect incidental users. For example, privatized intellectual property protection might not extend to uses of the intellectual property that begin before the privatization and account for a small proportion of the value of the privatized resource. Relatedly, intellectual property remedies might be narrower; perhaps the new owner of a formerly public domain right should be denied injunctive relief.
A significant problem in setting optimal term and scope is that they may vary from one intellectual property right to another. This problem may be even more severe in the case of privatized intellectual property than for intellectual property more broadly, since the motives for privatization may be relatively heterogeneous. It thus should not be surprising if legislative privatization schemes reflect relatively arbitrary decisions, and indeed that is true at least for the
Orphan Drug Act, which provides a seven-year patent term. Presumably, if the term were longer, pharmaceutical companies would invest in more studies and produce more drugs targeting rare
31
PRIVATIZING THE PUBLIC DOMAIN diseases. Just as it is possible that a longer patent term will result in inventive activity occurring earlier and thus maybe even in an invention being placed in the public domain sooner,
87
so too might a longer protection term lead to orphan drugs being available generically sooner. A longer term, however, produces familiar costs, such as deadweight loss,
88
and these costs may be unnecessary if the government awards orphan drug rights in cases in which companies would have had sufficient incentives to produce the orphan drugs just as quickly in any event.
89
There does not appear to be any obvious justification for the large gap between the familiar 20-year patent term and the seven-year Orphan Drug Act term,
90
though the drug context highlights a subtle case for shorter intellectual property terms beyond that presented above when an invention already exists in the public domain. When the cost and benefits of an invention are known in advance, raising patent terms beyond a certain point will not produce social losses, because longer terms will simply lead to inventors joining patent races even earlier.
91
There may be situations, however, in which an exogenous shock makes development of an invention seem more attractive than it did previously, and in such a case the patent term may be longer than necessary to justify development of the invention. Exogenous shocks seem particularly likely in the case of inventions that previously seemed worthless, but suddenly seem to have potential, for example as a result of developments in another field. And so, a seven-year term may be enough of an inducement to development, even if a previous patent owner had inadequate incentives before the exogenous shock. But there will be cases also in which seven years is not enough to encourage socially useful behavior.
87
See Duffy, supra
88
See supra Part II.A.1.
89
In general, this seems unlikely to be the case. If the Orphan Drug Act results in the completion of studies for a drug in 2005, then presumably money would have been invested and the studies completed earlier if a longer term were available.
90
One theory might be that it is less cumbersome to develop orphan drugs because the substances already exist, and progress may not be necessary on their synthesis. But process patents receive the usual 20-year term, so there is at least an inconsistency.
91
See John F. Duffy, A Minimum Optimal Patent Term (unpublished manuscript, 2004), available at http://www.ssrn.com/abstract=354282 (visited Mar. 21, 2004).
32
PRIVATIZING THE PUBLIC DOMAIN
5.
Constitutionality
A final challenge in the design of an intellectual property privatization system is that the system must pass constitutional muster. In the United States, the danger is that a court might find that Congress does not have the power to provide retroactive intellectual property protection to inventions that have already fallen into the public domain. In Eldred v. Ashcroft , 92 the Supreme
Court upheld the Copyright Term Extension Act,
93
which extended copyright terms for both current and past copyrights. The Court noted that the Act did not grant perpetual copyright and found that Congress’s decision was rational.
94
This suggests that rational privatization efforts will also be upheld, though one might seek to distinguish Eldred on the ground that the CTEA did not attempt to revive copyrights that had already fallen into the public domain.
95
Even if an intellectual property scheme were held beyond Congress’s power under the Intellectual Property
Clause, they might be upheld under the Commerce Clause.
96
A full assessment of such issues, of course, is beyond the scope of this Article.
III.
S TRATEGIES FOR S ELECTIVELY P RIVATIZING THE P UBLIC D OMAIN
The ultimate question is whether it is possible to enact a system for selectively privatizing the public domain that would produce more benefits than costs. This empirical question is, of course, impossibly complex, especially in the absence of experimentation. Perhaps the most that can be said is that some regimes for selective privatization are likely to be better than others. In particular, an effective regime should avoid undue government discretion in deciding when privatization is appropriate and who should receive the privatized right, because government discretion will necessarily increase rent-seeking expenditures. That does not mean that a
92
537 U.S. 186 (2003).
93
Pub. L. No. 105-298 (105th Cong. 1998).
94
537 U.S. at 201-08.
95
Similar issues have arisen with copyright restoration under the relatively narrow terms of § 104A of the Copyright Act, which restores copyrights previously denied because of lack of formalities. For a discussion of this provision, see Adam P. Segal,
Zombie Copyrights: Copyright Restoration Under the New § 104A of the Copyright Act , 13 S
ANTA
C
LARA
C
OMPUTER
& H
IGH
T ECH .
L.J. 71 (1997).
96
See generally Thomas B. Nachbar, Intellectual Property and Constitutional Norms , 104 C OLUM .
L.
R EV . 272 (2004) (arguing that Congress may regulate intellectual property under the commerce power, because there is no generally applicable norm that would suggest that the limits on the intellectual property power should apply to other powers).
33
PRIVATIZING THE PUBLIC DOMAIN discretionary regime is inherently inefficient; perhaps it would be possible to design an independent agency that would make efficient decisions and not be vulnerable to interest group manipulation. But the case for selective privatization is much stronger if it is possible to design mechanisms that will lead to at least roughly optimal decisions with a minimum of governmental interference. This section will highlight two market strategies that might be useful in privatization efforts: auctions and information markets. Auction mechanisms can help assign a property right efficiently, and although information markets also might be useful for such a task, they may be especially useful for determining whether privatization should occur in the first place.
A.
Auctions
Probably the most straightforward means of distributing an intellectual property right that is to be privatized is through auction. Auction mechanisms are generally efficient means for moving assets to their highest-valuing users. That does not mean that auctions will necessarily produce maximal allocative efficiency, however. Even assuming that the private party who will benefit the most from a privatized intellectual property right is the party who will most benefit society by holding that right, we cannot be sure that auctions will identify the correct party. The problem is that auctions rely entirely on each party’s own private assessment of value. Familiar pathologies of behavioral economics, such as optimism and overconfidence biases,
97
thus threaten to undermine efficiency. Perhaps of greatest concern is the winner’s curse, 98 though not because this will necessarily mean that the highest bidder will be overpaying for the intellectual property right, a boon to society. Rather, the winner’s curse reflects that each party estimates the common value component of the auctioned property right with some error. If the value of the intellectual property right is separated into this common value component and a private value component that varies from one user to another, then the auction winner may be the party that
97
See generally Jennifer Arlen, The Future of Behavioral Economic Analysis of Law , 51 V AND .
L.
R EV . 1765, 1773-75 (1998)
(discussing overoptimism biases).
98
See R
ICHARD
H.
T
HALER
, The Winner’s Curse , in T
HE
W
INNER
’
S
C
URSE
: P
ARADOXES AND
A
NOMALIES OF
E
CONOMIC
L
IFE
50,
51-52 (1992).
34
PRIVATIZING THE PUBLIC DOMAIN offers the highest estimation of the common value component, rather than the private party whose (potentially erroneous) estimation of its own private valuation is highest. Nonetheless, auction mechanisms in general provide stronger incentives for allocative efficiency than would a system in which government officials simply assign a property right to a private party.
An additional virtue of auctions is that they provide a means of dissipating rents and thus avoiding windfalls to private parties. In a second-price auction, for example, each party will bid an amount equal to its expected valuation, because the winner only pays the amount of the second-highest bid.
99
The auction winner’s expected benefit will thus be equal only to the difference between its valuation and the second-highest valuation. The incentives to lobby for an auction are thus relatively limited, at least if the common value component of the intellectual property right accounts for a large portion of its value. Auctions may still induce a fair amount of rent-seeking when it is clear before the auction which party will be the highest-valuing user. In this case, participation in the auction may be quite limited, because there is little incentive to participate if one knows that one will not win. In theory, auction design can give incentives to participate (for example, by providing some compensation to the second-highest bidder), though it may be difficult for governments to design auctions properly.
Although the primary purpose of an auction is to distribute a monopoly intellectual property right, in theory an auction also might be used as a means of determining whether to distribute a monopoly right at all. For example, the government might simultaneously hold two auctions, one auction offering a monopoly intellectual property right, and one auction offering three rights to participate as oligopolists. The ratio of the highest bid in the monopoly auction to the sum of the three highest bids in the oligopoly auction provides some information about the private value to having a monopoly right. For example, when a patent needs a great deal of advertising and coordination among three oligopolists seems difficult to coordinate, this ratio seems likely to be high. Thus, the government might agree to consummate the monopoly auction only when the ratio exceeds a certain level. This approach is not without its problems; perhaps
99
See generally William Vickrey, Counterspeculation, Auctions, and Competitive Sealed Tenders , 16 J.
F
IN
.
8 (1961)
(introducing second-price auctions).
35
PRIVATIZING THE PUBLIC DOMAIN the ratio will reflect other attributes of the particular market, such as the ability of a monopolist to price discriminate. Nonetheless, it highlights that auctions may be designed to generate proxies for different market conditions.
I recently proposed that a similar design be used for patent reauctions, that is for auctions for patent term extensions.
100 Under this proposal, with a few years remaining in the patent term, a patentee would be able to request a patent term extension, perhaps of five years duration. The patentee, however, would be allowed to win the patent only by outbidding the top bidder by, say,
25%. To discourage patentees from calling for auctions that they do not expect to win, a patentee who does not outbid the top bidder by that margin would have to pay a fine. The purpose of this auction design is to encourage patentees to call for patent term extensions only when the benefits of continuous patent ownership are relatively high. A third-party winner of the auction would be allowed after the original patent term expired to sell the patent back to the original patentee, so third parties should be expected to bid the greater of their own valuation of the patent and the original patentee’s valuation, less negotiation and transfer costs. Thus, when the original patentee can outbid a third party by a substantial margin, that indicates not merely that the original patentee values the patent more than third parties, but that the original patentee particularly values being assured of ownership continuity. Presumably, this will be because the original patentee wishes to invest further in the development of the patent, for example by advertising, in ways that would produce returns both during the patent term and afterward.
This proposal does not have apply directly to the problem of privatizing intellectual property that is already public domain, because when such privatization occurs, there is no initial patent owner. It would be possible, however, to design a system with similar objectives. For example, any party might be able to call for two auctions, one for a new five-year term and one for an additional five-year term beyond that, with the two auctions to be held sequentially. The winner of the first five-year term would be able to win the second five-year term only by exceeding the next highest bid by a substantial percentage. As before, to make the mechanism
100
See Michael Abramowicz, Patent Auctions (2005) (unpublished manuscript, on file with author).
36
PRIVATIZING THE PUBLIC DOMAIN work, the auctions must be consummated even where the winner of the first term does not bid enough to win the second term. The trick is providing incentives to call for the auction only when this seems likely to be the case. With patent extensions, giving the decision to the patentee, with a penalty to be paid if the patentee then could not exceed the bid for the second term, was sufficient. But when there is no single party with an appropriate incentive, it may be necessary to offer a bonus for calling for an auction in which a single party wins both terms with a penalty for calling for one that does not yield that result. Alternatively, some other mechanism, such as an information market, might be used to predict the result of an auction.
B.
Information Markets
An information market is a mechanism for generating consensus predictions from a group.
101 The most common forms of information markets are literally security markets, with traders buying and selling propositions. For example, in the Iowa Electronic Markets, traders may buy and sell shares corresponding to different candidates in a particular election, and the share of whichever candidate eventually wins will be sold for $1.
102
At any given time, the price at which shares are exchanged provides an indication of the probability that a particular candidate will win.
More recent proposals for information markets do not involve markets and trading of securities at all. Broadly conceived, an information market is a mechanism to induce individuals to offer sequential predictions, with individuals profiting when they can improve on previous predictions and losing money when they move the consensus prediction in the wrong direction.
A relatively simple approach, Robin Hanson’s market scoring rule, requires the information market sponsor to announce a formula by which the final predictor will be rewarded.
103 Anyone may offer a prediction (with the new prediction becoming the new consensus prediction) by agreeing to pay the previous predictor according to the same formula.
101
See Michael Abramowicz, Information Markets, Administrative Decisionmaking, and Predictive Cost-Benefit Analysis , 71 U.
C
HI
.
L.
R
EV
.
933 (2004).
102
See http://www.biz.uiowa.edu/iem/ (last visited Aug. 25, 2005).
103
See Robin D. Hanson, Combinatorial Information Market Design , 5 I
NFO
.
S
YS
.
F
RONTIERS
107 (2003).
37
PRIVATIZING THE PUBLIC DOMAIN
Preliminary studies indicate that information markets are relatively accurate, at least compared to other means of eliciting expert opinion,
104
and laboratory studies indicate that information markets are relatively difficult to manipulate.
105 The theoretical insight is that someone who tries to manipulate the market, in the hope of affecting some public policy result, will lose money on the market manipulation and invite counter-speculators to push the market price back to its appropriate value. Manipulation is likely to be especially hard if traders can identify one another and can assess whether other market participants have analyses and data to support their trades.
106
Repeat players who have earned reputations for making accurate predictions are less likely to invite counter-speculation than new participants who do not seek to explain their trading patterns in a convincing way. Although further experimentation is necessary, information markets (even if they are subject to some degree of manipulation) may be a useful general purpose predictive tool.
Information markets thus may be useful as a means of predicting the consequences of privatizing intellectual property. Conditional information markets can be used to make predictions based on some contingency, so information markets could be used to make predictions conditional on a particular intellectual property right being privatized and conditional on its not being privatized.
107 The challenge is in determining what exactly should be predicted.
Where an intellectual property right corresponds to a discrete consumer product (such as a pharmaceutical drug), a market might be used to predict the total consumer surplus from use of that product. This requires some ex post proxy for consumer surplus. For example, an ex post survey might be used to determine what portions of consumers used a product and the most the consumers would have paid for it. This might be a noisy proxy, but that is of relatively little
104
See, e.g.
, Joyce Berg, et al, Results from a Dozen Years of Election Futures Markets Research (Nov. 2000), online at http://www.biz.uiowa.edu/iem/archive/BFNR_2000.pdf (visited Aug. 25, 2005).
105
See Robin Hanson et al., Information Aggregation and Manipulation in an Experimental Market (July 12, 2005), available at http://hanson.gmu.edu/biastest.pdf.
106
A mechanism that can encourage this is a deliberative information market, one in which one is compensated based on whether one moves the market price over a relatively short period of time.
See Michael Abramowicz, Deliberative Information Markets for Small Groups , in I
NFORMATION
M
ARKETS
: A N
EW
W
AY OF
M
AKING
D
ECISIONS IN THE
P
UBLIC AND
P
RIVATE
S
ECTORS
(Robert
Hahn & Paul Tetlock eds., AEI-Brookings Press forthcoming 2005).
107
For a discussion of conditional information markets, see Abramowicz, supra
38
PRIVATIZING THE PUBLIC DOMAIN concern, as long as the ex ante predictors cannot anticipate the direction and extent of any error in the ex post measurement. An alternative noisy proxy would simply be a subjective evaluation of the net social benefits produced by the privatization.
108 Under this approach, an intellectual property right would be privatized only when the market predicts net benefits rather than net costs.
These applications of information markets are admittedly particularly speculative, and more experience must be gained with information markets before they can be integrated with confidence into intellectual property institutions. A somewhat more straightforward approach might combine an information market with the double auction mechanism described in the previous section. That is, once again, two auctions might be held sequentially, one for a five-year term and the second for a five-year term immediately following. This would occur, however, only if an information market predicted with sufficient confidence that holding two auctions would result in the high bidder in the first auction offering the highest bid by a substantial margin in the second auction. Because the information market determines whether the auctions will occur, there is no need in this formulation to require that the winner of the first auction bid substantially more than the next highest bidder in the second auction to win the second term. But the auctions will only be held when this is expected with a high probability to be the case. Of course, because information markets generally require subsidy, we still need some mechanism for determining when to launch these information markets, but this is a second-order problem.
109
At least four regimes thus bear comparison. The first is a regime in which governmental officials make decisions on a case-by-case basis about whether to privatize intellectual property, and then auction rights selected for privatization. The use of an auction would substantially dissipate rents and thus reduce the room for rent-seeking, but there remain substantial questions about whether governmental officials would make sufficiently good decisions to make the regime have greater costs than benefits. The second, which for most technologies is the current
108
For a defense of using information markets to aggregate subjective assessments, see id.
at 997-1010.
109
Conceivably, we might use relatively low subsidy information markets to determine whether it would be advisable to have higher subsidy information markets, which in turn would determine whether auctions should be held.
39
PRIVATIZING THE PUBLIC DOMAIN regime, does not allow selective privatization of the public domain. Given the current state of legal technology, this status quo for now should be preserved.
If information market technology improves and becomes more reliable, however, the third and fourth regimes might become feasible. The third regime would use information markets in combination with the double auction mechanism described above, and the fourth regime would use information markets to anticipate some ex post proxy for consumer surplus or social welfare conditional on privatization occurring and not occurring. This last regime imposes the most demands on the information market mechanism, but has the strongest potential to take into account a wide range of considerations, such as whether additional advertising, even if privately optimal, would be socially optimal for a particular intellectual property right.
IV.
C ONCLUSION
My purpose in discussing selective privatization of intellectual property, and especially in suggesting that auctions and information markets might play around, is not to make an immediate policy proposal. Rather, it is to suggest that improvements in information market technology may someday dramatically improve the feasibility of selective privatization of the public domain. There is a broader point: Law and economics has developed to a point where at least the basic tradeoffs in intellectual property policy are well understood, though we know that there are likely to be some major theoretical surprises. This leaves two challenges. The first, widely recognized challenge, is improved empirical analysis that can help optimize the term and scope of intellectual property rights. The second, previously ignored challenge is the use of capital market mechanisms, likely including auctions and information markets, to bring such empirical analysis to bear not just on the overall terms of intellectual property law, but on specific intellectual property rights.
Ideally, legal technology should allow incorporation of theoretical and empirical knowledge about specific markets and intellectual creations to permit customization of the terms of intellectual property protection. In this world, we would not have a number of discrete intellectual property regimes—patent, copyright, and trademark to name the most prominent—
40
PRIVATIZING THE PUBLIC DOMAIN but a broad concept of intellectual property with different instantiations in different markets. The literature on intellectual property has largely ignored the possibility that legal technology itself might improve, and the vast majority of intellectual property scholarship focuses (appropriately enough, given our limited past experience) on optimizing intellectual property regimes within existing institutions, such as the patent office and the trial courts. The challenge is to use market mechanisms to create new, better institutions. Many scholars may believe this to be an impossible challenge, and indeed there are large obstacles. Even if new institutions seem promising, it is difficult to find ways of allowing modest experimentation on them, for there is no marketplace for law itself. My hope, however, is that scholars will strive to imagine radically different institutional arrangements. Perhaps in two or three generations of such scholarship, some consensus may be possible initially on limited experimentation with intellectual property regimes and incentive structures quite different from those that exist today.
41