Berkeley Law
Berkeley Law Scholarship Repository
Faculty Scholarship
12-1-1994
Of Property Rules, Coase, and Intellectual Property
Robert P. Merges
Berkeley Law
Follow this and additional works at: http://scholarship.law.berkeley.edu/facpubs
Part of the Law Commons
Recommended Citation
Robert P. Merges, Of Property Rules, Coase, and Intellectual Property, 94 Colum. L. Rev. 2655 (1994)
This Article is brought to you for free and open access by Berkeley Law Scholarship Repository. It has been accepted for inclusion in Faculty
Scholarship by an authorized administrator of Berkeley Law Scholarship Repository. For more information, please contact jcera@law.berkeley.edu.
OF PROPERTY RULES, COASE, AND INTELLECTUAL
PROPERTY
Robert P. Merges*
My colleagues in this Symposium have done a fine job of cataloguing
the many merits (and occasional lapses) of the paper by Jerry Reichman
on which I am asked to comment.' I cannot add much to this theme. I
would like instead to take the Reichman paper as an opportunity to comment on certain foundational issues in the law of intellectual property
rights (IPRs). Taking my cue from Reichman's characterization of his
approach as an "off-the-rack liability rule" regime, 2 in this Comment I
discuss the simple economics of the strong preference for injunctionsthe classic instance of a property rule-in IPR law. I conclude that proponents of IPR liability rules (including, apparently, Reichman) carry a
heavy burden.
According to the foundational literature on legal entitlements, 3 a
property rule is a legal entitlement that can only be infringed after bargaining with the entitlement holder. The holder thus sets the price for
infringing ex ante. Under a liability rule, by contrast, one may infringe
first, and a tribunal will determine the appropriate compensation in an ex
post proceeding. Ever since Calabresi and Melamed, transaction costs
have dominated the choice of the proper entitlement rule, with a liability
rule being the entitlement of choice when transaction costs are high.
Some of this thinking finds its way into Reichman's proposals. 4 To
the extent it does, I take issue. In my view, property rules can and do
work effectively in many situations involving IPRs. This is so because, in
the presence of high transaction costs, industry participants have an incentive to invest in institutions that lower the costs of IPR exchange.
Thus, at least in some cases, the costly bargaining occasioned by a strong
property rule leads to an administrative structure that serves much the
same function as a statutory liability rule. Consequently, the benefits of
lower transaction costs may be achieved without sacrificing the strong
property rule built into most IPRs.
* B.S. Carnegie-Mellon; J.D. Yale; LI.M. Columbia. I would like to thank Paul
Goldstein, Wendy Gordon, Steve Marks, Manuel Utset and the participants in this
Symposium for helpful comments, and Major League Baseball, whose strike eliminated my
best excuse for not working on this piece.
1. See J.H. Reichman, Legal Hybrids Between the Patent and Copyright Paradigm, 94
Colum. L. Rev. 1501 (1994).
2. Id. at 2533. "Off the rack" is a phrase that conjures up the literature on contract
default rules. Reichman does not pursue this theme, but it is applied to several related
areas in Robert P. Merges, Patent Law and Policy: Cases and Materials 898-900 (1992).
3. See Guido Calabresi & A. Douglas Melamed, Property Rules, Liability Rules, and
Inalienability: One View of the Cathedral, 85 Harv. L. Rev. 1089, 1092 (1972).
4. See Reichman, supra note 2, at 2447.
2655
2656
COLUMBIA LAW REVEW
[Vol. 94:2655
As explained in Part II, the pre-infringement bargaining called for by
a strong property rule is appropriate for the IPR field, where unique assets are exchanged. Part III briefly describes the exceptions to the strong
property rule baseline in the law of IPRs, the compulsory licensing provisions of the various IPR statutes. A general critique of those provisions is
also presented. Beforejumping into the discussion of property rules and
liability rules, however, it is appropriate to consider first whether entitlement rules really matter. This of course is the domain of the Coase Theorem, and of Part I, to which I now turn.
I. IPRs AND THE COASE
THEOREM: OR, WHY THE INDIFFERENCE
THEsis FAreS
A natural starting point when discussing the property/liability rule
distinction, or indeed any aspect of entitlements, is the Coase Theorem. 5
The Theorem says that in a world with zero transaction costs, initial rights
allocations are unimportant; they will be transferred to their highestvalue use through private bargains. 6 But where transaction costs are positive, this result cannot be assumed. Thus, ultimately Coase admonishes
us to pay attention to transaction costs in allocating property rights and
5. The theory was first stated, though not under that name, in RIH. Coase, The
Problem of Social Cost, 3 J.L. & Econ. 1 (1960). See generally Robert D. Cooter, Coase
Theorem, in 1 The New Palgrave: A Dictionary of Economics 457, 458 (John Eatwell et al.
eds., 1987) (presenting overview of Theorem).
6. Coase wrote the article primarily to refute the conventional wisdom associated with
A.C. Pigou regarding economic externalities. See Robert D. Cooter, The Coase Theorem,
in Allocation, Information and Markets: The New Palgrave 64, 69 (1987) (citing and
discussing A-C. Pigou, The Economics of Welfare (4th ed. 1932)). Externalities are effects
of one's behavior on others' activities. Pigou stated that externalities could be overcome
only through state intervention. For example, he argued that a tax was needed when a
business caused smoke that interfered with a neighboring business such as a laundry.
Without the tax, the smoke producer would not take into account the effects of the smoke
on the laundry; so the tax would be added to the smoke-producer to bring her private
incentives into line with the social good-that is, to "internalize" the negative externality of
smoke damage via a tax on smoke. One of Coase's insights was that the state could be
removed from the picture by giving the smoke producer (or the laundry owner) the right
to make smoke (or be smoke-free), and then permitting market exchanges. See R.H.
Coase, The Problem of Social Cost in The Firm, the Market, and the Law 95, 101, 105, 114
(1988). Furthermore, Coase concluded that it is irrelevant from the point of view of total
output who has the right, so long as transaction costs are low. See id. at 114 ("Such a
rearrangement will always take place if it will result in an increase in the value of
production."). Of course, one of Coase's main points was that economists such as Pigou
had ignored the costs of administering the government solution to the externality
problem. See id. at 154, 156. He showed this brilliantly by demonstrating that under the
same assumption of costless administration (i.e., zero transaction costs), there is no need
for government intervention.
1994] PROPERTY RULES, COASE, & 1NTELLECTUAL PROPERTY 2657
setting rules for their exchange.7 The purpose of this brief article is to do
just that in the context of IPRs. 8
Consider first a hypothetical to which the Coase Theorem applies.
Imagine a patent infringer who can produce a patented product more
cheaply than the inventor who owns the patent. Whether a transaction is
required to produce the efficient outcome depends on whether the patentee has a legal right, but the efficient outcome will be the same. If the
inventor holds the right, the higher-valuing infringer can be expected to
negotiate a license. If the infringer has the right, she can market her
product without negotiating a license. The infringer's superior product
will reach the market in either case, although the parties' wealth differs
depending on who has the right. 9
So the Coase Theorem applies to at least some inventor/infringer
interactions. At first blush, then, we might be tempted to conclude that
allocation of initial entitlements between rival inventors-hence, IPRs
themselves-is irrelevant to an efficient outcome: the indifference thesis
holds. But before we make this leap, we must catalogue some important
differences between intellectual property and the sorts of physical externalities with which Coase was concerned. In the first place, unlike his
famous smoke or cattle examples,' 0 it is not always easy to detect externalities in the invention context. The problem of detecting externalities in
the IPR field implicates an additional difference between IPRs and other
Coasian interactions. As Arrow has pointed out in his "paradox of information," without a property right, the licensor is in a pickle: if in trying
to strike a deal she discloses her idea (e.g., the technology she invented),
she has nothing left to sell, but if she does not disclose anything the buyer
has no idea what is for sale." Patents (and to a lesser extent trade
7. See Thrl.inn Eggertsson, Economic Behavior and Institutions 105 (1990) ("Coase's
main contribution . . . was to arouse our awareness of the implications of positive
transaction costs.").
8. Although I did discuss this issue in my patent law casebook, see Merges, supra note
2, at 776-77, 1 was not the first to do so. See Stanley M. Besen et al., Copyright Liability for
Cable Television: Compulsory Licensing and the Coase Theorem, 21 J.L. & Econ. 67
(1978).
See also John W. Schlicher, Patent Law: Legal and Economic Principles
§§ 2.04-2.05 (1992) (applying Coase Theorem to laws protecting information production
and bargaining problems in agreements to produce information).
9. Obviously, I am using "infringer" loosely in this sentence to preserve the symmetry
of the examples. One who has the right to use something cannot be an infringer. Perhaps
"second comer" or "follow-on creator" would be better. A straightforward numerical
example can be found in Merges, supra note 2, at 766-77.
10. For details of the smoke example, see supra note 6.
11. See Kenneth J. Arrow, Economic Welfare and the Allocation of Resources for
Invention, in The Rate and Direction of Inventive Activity 609, 615 (National Bureau of
Economic Research ed., 1962). Many examples of this well-known phenomenon can be
found; for one drawn from history, see Tom D. Crouch, The Bishops Boys: A Life of
Wilbur and Orville Wright 346-48 (1989) (describing Wright brothers' dilemma when,
prior to filing for a patent on their airplane design, they were approached by the
government to reveal their invention; they proposed a large up-front payment or "bond" to
solve the problem).
2658
COLUMBIA LAW REVIEW
[Vol. 94:2655
secrets) protect the licensor's property so she can confidently discuss it in
the context of striking a Coasian bargain. A farmer adjacent to a cattle
ranch will normally have no trouble determining when cattle have trampled her crops, for purposes of assessing the need for (and price of) a
compensatory exchange. In the IPR context, there is no smoky soot or
wandering cattle to serve as an unambiguous marker, although a direct
copy of an apparent feature may appear on the market in some cases.
Creators very often work far away from each other, and at different times.
Prior creative works or inventions may be incorporated into products
unintentionally.
Detection aside, in many cases there will be disagreements about
whether and to what degree the prior work added to the value of the
subsequent one. Prior creations may be minor inputs that are subsequently incorporated into a larger product, or they may take on a form
similar to a common "ancestor" work. In the case of patented inventions,
an infringer may have no way of knowing that her own independent invention is an infringement, or that, at the time she makes her investment
decisions, a patent even exists. 12 In such a setting, there will be many
more arguments over the existence of externalities than in the examples
used by Coase. 13 To the extent that IPRs deal with abstract rights over
creative works, there will often be profound debate about whether there
are any externalities at all that must be "internalized" via a property right.
Even assuming they agree on the existence of an externality, the two
creators are unlikely to agree on its severity. A second creator who has
drawn on an earlier work, for example, is much more likely to believe (or
at least argue) that the unique elements in her work or product are responsible for whatever success it has had. Anyone who has negotiated an
IPR license (or the sale of any unique asset) can only yearn for the sort of
mutually agreed-upon schedule of harms and benefits Coase draws up in
12. The same is true in trademark law. Like one who independently invents an
already patented device, one who begins using a similar trademark is an infringer
regardless of the fact that the trademark owner's business was completely unknown to the
infringer. This situation might be described as a form of asymmetric information, but the
point in the text is meant to go further. Even where the two parties, the creators of the
prior and subsequent works, agree that the latter had access to information about the
former's work, there may be disagreements over whether and to what extent knowledge of
this work assisted in the creation of the subsequent work, or adds value to it.
13. Cf. Douglass C. North, Structure and Change in Economic History 16 (1981). As
North points out:
Throughout history there has almost always been an immense gap between the
private and the social returns to invention and innovation. The problem is one of
specifying property rights over ideas and their application to economic activity; it
has been more difficult to devise property rights over technological development
than over products or resource inputs. The difficulty of measuring the
dimensions of intellectual property and innovations and of enforcing any such
property rights has been a basic reason for the divergence between private and
social benefits.
1994] PROPERTY RULES, COASE, & INTELLECTUAL PROPERTY 2659
his hypotheticals. The problems with producing such a schedule in the
IPR field result from the abstract quality of the benefits conferred by
prior works and the cumulative, interdependent nature of works covered
by IPRs. Valuation, then, is at least as great a problem as detection.
One corollary of the valuation problem is the heightened possibility
of strategic behavior. Coasian bargaining assumes that the parties will
equitably divide-rather than fight over-the cooperative surplus to be
gained via the bargain. This equitable division is unrealistic in many
14
Indeed, I have argued
cases, as noted by Robert Cooter and others.
elsewhere that overlapping IPRs on distinct aspects of a single product
can create very knotty valuation problems.' 5 The famous case of blocking
patents, a most interesting institution from the property rights perspective and a textbook case of bilateral monopoly in action, often creates
problems of this sort. (When A owns a patent on a broad class of products and B owns a patent on a specific product in that class, A and B are
14. See Robert D. Cooter, The Cost of Coase, 11 J. Legal Stud. 1, 23 (1982).
15. See Robert P. Merges, Intellectual Property Rights and Bargaining Breakdown:
The Case of Improvement Inventions and Blocking Patents, 61 Tenn. L. Rev. (forthcoming
1994). I recommend occasionally excusing the subservient patent holder (B in the
example in the text) from infringement liability when the holdup problem is severe. The
holdup may be severe, for example, when the invention covered by an improvement
patent, which cannot be implemented without permission of the holder of an original,
dominant patent, contributes a very large proportion of the total value of the original-plusimprovement combination. Historically, bargaining breakdowns have often occurred in
this situation, the social costs of which justify a sort of "patent fair use" or "efficient
infringement" principle. Fortunately, while such situations are serious given that they
often involve significant new technologies, they are relatively rare. In the great run of
cases, the ingenious institution of blocking patents balances the rights of original creators
and subsequent improvers rather nicely. Interestingly, no such institution exists in the law
of copyrights under which a follow-on creator would be prohibited from appropriating and
adding to the copyrighted material of an original creator. It has been argued that to deny
the derivative artist a copyright in his additional creative efforts, to the extent they are
pervaded by the original work, diminishes his incentives to create new works based upon
copyrighted material:
The rule [denying copyrightability for unauthorized derivative works] is, however,
hard to justify when applied to derivative works such as the motion picture in
Sheldon v. MGM in which the underlying work represents only a small part of the
value of the derivative work but, because it underlies the whole, will defeat
copyright protection for the entire derivative work. Just as an injunction against
the motion picture gave plaintiff there a greater return than was needed to
induce his investment in the underlying work, so depriving the motion picture
owner of all protection against others will give it far less return than is needed to
justify investment in the derivative work.
Paul Goldstein, Derivative Rights and Derivative Works in Copyright, 30J. Copyright Soc'y
209 (1983) (citing Sheldon v. MGM, 309 U.S. 390 (1940)). See also Wendy J. Gordon,
Toward a Jurisprudence of Benefits: The Norms of Copyright and the Problem of Private
Censorship, 57 U. Chi. L. Rev. 1009 (1990). Perhaps the best explanation for the lack of a
doctrine of"blocking copyrights" is copyright law's policy favoring the reputational interest
of authors: by requiring ex ante licensing of anyone who wishes to incorporate a
copyrighted work into another work, the law ensures that the owner of the copyrighted
work will completely control all manifestations of it.
2660
COLUMBIA LAW REVIEW
[Vol. 94:2655
said to hold blocking patents with respect to the specific product of B's
patent: neither can commercialize the product covered by B's patent
without first obtaining a license from the other.) I recommend that the
law sometimes excuse holders of subservient blocking patents (B in the
example above) from liability, as a way of influencing right-holders to
reach agreement. Whatever the policy solution, however, the problem of
blocking patents teaches a valuable lesson: the Coasian picture of two
economic agents agreeing on the joint costs and benefits of each level of
production is out of place in at least some, and perhaps most, cases involving IPRs. 1 6 Ultimately, valuation is significantly trickier than that described by Coase and others in hypotheticals of the farmer/rancher
variety.
So far, I have described problems of detection and valuation. Another potential problem with the Coasian construct as it applies to intellectual property is its avowed lack of concern for the distribution of resources. Coase sets out to show that total output in an economy is
unaffected by the initial allocation of rights. The farmer and the cattle
rancher, for example, will bargain to produce the same mix of products
regardless of who holds the right. 17 At the same time, the one holding
the right will have more money after the bargain is struck. 18
16. If more proof is needed, consult a decision setting damages for patent
infringement-a form of ex post valuation that is obviously difficult and inexact. See
Merges, supra note 2, at ch. 9.
Valuation problems also extend beyond patent law, they are pervasive in copyright as
well. Consider the complex cases that try to apportion the relative contributions of
infringing and non-infringing portions of an aggregate copyrighted work. In the typical
case, the holder of a copyright on a work seeks to enjoin commercialization of a larger
work that incorporates it. The classic example is Sheldon v. Metro-Goldwyn Pictures Corp.,
309 U.S. 390 (1940), discussed supra note 15. Plaintiff's play was copied by MGM in a
motion picture that also included substantial original contributions by MGM-such as a
famous cast and significant advertising expenditures. The Supreme Court affirmed the
lower court's order that plaintiff be awarded only one-fifth of defendant's profits. See id.
at 409. The Court's heroic attempt to block plaintiff's holdup attempt was laudable.
Nevertheless, the one-fifth royalty appears completely arbitrary. See id. at 407-09
(estimates by expert witnesses from film industry of percentage of value contributed by
play ranged from 0% to 12%; court settled on 20% to insure no harm to copyright holder).
This arbitrariness, together with the felt necessity of litigation up to the Supreme Court,
provides indirect evidence of the severity of valuation problems.
17. See Cooter, supra note 6, at 66 ("As with ordinary goods, the gains from trading
legal entitlements are not exhausted until each entitlement is held by the party who values
it the most."); Merges, supra note 2, at 766-67 (applying this "strong form" of the Coase
theorem to intellectual property, specifically patents).
18. Coase claims in a later commentary that, in fact, any differences in the
distribution of resources will be eliminated as the cost of related resources rises (or falls) to
reflect the new income distribution. See Coase, supra note 6, at 157, 170-74. The
limitations of this point are best captured in a direct quote, taken from a passage
discussing the rancher/farmer example described earlier in the text[I]f the rule of liability is known, the amount that will have been paid to acquire
the land will reflect this, less being paid for the ranching land and more for the
farming land when compensation has to be paid [to the farmers] than when it
1994] PROPERTY RULES, COASE, & INTELLECTUAL PROPERTY 2661
Although the overall goal of intellectual property law is often described in allocational efficiency terms (i.e., to increase economic output
by overcoming market failures associated with the public goods quality of
creative works), there is often an undercurrent of concern with the distribution of resources. Therefore, the law would seem to favor allocation of
rights to creators rather than infringers, even if the parties will bargain to
the efficient outcome regardless of the allocation. As a result of this builtin distributional bias in intellectual property, in some cases an inventor
reaps royalties even though total output is the same as it would be if the
infringer were not liable.
The problem of transaction costs is the most blatant one in applying
the Coase theorem to IPRs. Despite a few brave attempts to assume away
the obvious,19 those who have considered the application of the Coase
20
theorem to IPRs have noted the pervasive presence of transaction costs.
Thus, analysis conventionally begins with a discussion of where rights
should reside in the first instance, given high (and intractable) transaction costs. In addition, these arguments are sometimes taken a measure
further: they are used to justify the imposition of compulsory licenses,
that is, liability rules. Thus, a common rationale for the several statutory
compulsory licenses in copyright law is that they are needed in order for
certain types of exchange to take place. Transaction costs preclude the
formation of a market for certain types of rights;2 1 in the absence of statudoes not have to be paid. The wealth of the land-owners would thus remain the
same, changes in the amount paid for the land offsetting the changes in the flow
of payments brought about by a difference in the legal [rule] ....
Id. at 172. Even if such an adjustment were possible in some markets, it is very hard to
imagine the prices of inputs to creators or infringers (e.g., labor, art supplies, and lab
equipment) changing in response to a change in the legal rule on protection of creative
works, given the diversity and diffuseness of these inputs and the fact that so many have
alternate uses outside the creation of protectable works. Thus, the locus of rights in the
creator/infringer context will almost surely have distributional consequences.
19. SeeJack Hirshleifer, The Private and Social Value of Information and the Reward
to Inventive Activity, 61 Am. Econ. Rev. 561, 571-72 (1971) (arguing that, since inventors
have "inside information" about their inventions, they can reap gains by investing in assets
that their inventions will make more valuable and selling short assets that their inventions
will make less valuable; because of this inside information, inventors might be
overcompensated for their inventions); Ben T. Yu, Potential Competition and Contracting
in Innovation, 24 J.L. & Econ. 215 (1981) (presenting model in which activities of
prospective innovators are coordinated by central authority to reduce rent dissipation from
competitive inventing).
20. See, e.g., Schlicher, supra note 8, §§ 2.04-2.05 (1992).
21. See, e.g., Ralph Oman, The Compulsory License Redux: Will It Survive in a
Changing Marketplace?, 5 Cardozo Arts & Ent. L. Rev. 37 (1986) (describing the interplay
between transaction costs and the need for compulsory licensing); Lorna Veraldi, Note,
Cable Television's Compulsory License: An Idea Whose Time Has Passed?, 25 N.Y.L. Sch.
L. Rev. 925, 949-50 (1980) (concluding, on the basis of transaction costs, that there is no
evidence that the cable compulsory license, if abolished, would be replaced with market
transactions).
2662
COLUMBIA LAW REVIEW
[Vol. 94:2655
torily mandated transactions, none would take place. Some of
Reichman's comments reflect a similar judgment. 22
I have argued elsewhere that this reasoning is flawed in some cases.23
Indeed, I suggest that the presence of high transaction costs does not halt
exchanges but encourages both producers and users to invest in institutions that lower the cost of certain types of exchanges. Even though
there are many dispersed buyers (and sellers) of IPRs, and even though
the transaction costs of IPR exchanges are otherwise high, 24 the strong
property rule baseline often works quite well. The frequency of contracting in many markets for IPRs-an underdeveloped theme in most of
the entitlements literature 2 5 -gives rise to a myriad of institutions
(broadly defined) designed to streamline the exchange of property
rights. Institutions such as ASCAP 26 and patent pools 27 arise when firms
modify the strong property rule baseline of intellectual property law by
contractinginto liability rules. More importantly for property rights theory
in general, the property rule facilitates this adjustment. By contrast, statutory liability rules work against the flexible, voluntary institutions that are
formed to overcome the costs faced by transactors.
While there is some evidence that these institutions are beginning to
form in the software field (particularly multimedia products such as CD-
22. See Reichman, supra note 2, at 2447.
23. See Robert P. Merges, Contracting into Liability Rules: Institutions Supporting
Transactions in Intellectual Property Rights 22-24 (1994) (unpublished manuscript on file
with the Columbia Law Review). I am not alone: in comments on an earlier draft of this
paper, Paul Goldstein noted that transaction costs provide a dubious rationale for many of
the compulsory license provisions of the copyright statute. See also WendyJ. Gordon, Fair
Use as Market Failure: A Structural and Economic Analysis of the Betamax Case and Its
Predecessors, 82 Colum. L. Rev. 1600, 1613 (1982) (arguing that other factors can explain
some compulsory licenses, e.g., antitrust concerns).
24. Calabresi and Melamed assert that these conditions are likely to lead to a liability
rule:
Often the cost of establishing the value of an initial entitlement by negotiation is
so great that even though a transfer of the entitlement would benefit all
concerned, such a transfer will not occur. If a collective determination of the
value were available instead, the beneficial transfer would quickly come about.
Calabresi & Melamed, supra note 3, at 1106.
25. For example, to illustrate the desirability of liability rules, Calabresi and Melamed
describe a one-shot "eminent domain" example where adjoining landowners sell their
individual parcels to a single buyer. No repeat-play examples are given, which is not
surprising given the authors' emphasis on delineating a typology of initial entitlements.
See id. at 1106-07.
26. ASCAP, the American Society of Composers, Authors, and Publishers, is the
private copyright organization that collects composers' performance rights for licensing to
radio stations and nightclubs.
27. Patent pools are industry-wide agreements, often accompanied by administrative
structures, to centralize all firms' patents for automatic out-licensing or to cross-license
each others' patents.
1994] PROPERTY RULES, COASE, & INTELLECTUAL PROPERTY 2663
ROM) 28 and biotechnology, 29 the best evidence that they will do so is the
history of other industries. Since its infancy, radio broadcasting has involved an enormous transactional load to operate effectively. Likewise,
many technology-intensive industries in the early twentieth century developed only through the use of a myriad of technologies whose patents
were owned by separate firms. Yet, in each case, institutions, such as ASCAP and patent pools, emerged over time to facilitate many different combinations of separate IPRs. Until we know for sure that software and biotechnology are exceptions, we should, for the most part, give these
industries a chance to evolve similar institutional arrangements. 30
I cannot argue these points in detail here. I mention them because
the study of IPR institutions calls into question the static nature of the
usual Coase Theorem construct. The initial assignment of rights receives
all the attention under this analysis. Subsequent market and institutional
28. As one commentator recently noted:
[A group of] photographers [is] trying to flesh out a new set of copyright
arrangements for photographs .... [T]his sort of licensing has been going on in
the music recording industry for years, managed by a company [sic] known as
ASCAP ....
"Most [CD-ROM publishers] I talk to are very interested in having the
industry get together to make some practical [licensing] standards," [Jane] Kinne [,
of Comstock, a New York stock photography agency] says . . . . "They're not
interested in investing a lot of money in a product that is going to be shot down a
couple of years later in a lawsuit."
Paul Karon, Electronic Publishing Faces Legal Traps over Copyrights, InfoWorld, Mar. 9,
1992, at S-70.
29. So-called "materials transfer agreements," or "MTAs," contracts under which
biotechnology laboratories exchange cell lines and other research tools and findings, are
now very common. See Charles E. Lipsey et al., Protecting Trade Secrets in Biotechnology,
in Protecting Trade Secrets 1986, at Exhibit K (PLI Patent, Copyright, Trademarks, &
Literary Property Course Handbook Series No. 224, 1986). A uniform, standardized MTA
has been proposed by the Association of University Technology Managers (AUTM),
suggesting that standard contractual terms-a form of transaction-cost reducing industry
coordination-are beginning to emerge. See Interview with Sandy Shotwell, AUTM
member and participant in project to draft the Uniform Biotechnology Material Transfer
Agreement (UBMTA), in Washington, D.C. (Feb. 1994).
30. Waiting for these industries to develop institutional arrangements does not mean
that we should allow property rights to proliferate unchecked in these industries. Software
patents, for example, may well be a bad idea, as would be the granting of patents on "raw"
gene sequences such as those the National Institute of Health applied for in 1992 (and
subsequently dropped). See Rebecca S. Eisenberg, Technology Transfer and the Genome
Project: Problems with Patenting Research Tools, 5 Risk: Health, Safety, & Environment
163-75 (1994); Robert P. Merges, Patenting Human DNA and Legal Issues Related to the
Experimental Use Exception (Nov. 1, 1993) (unpublished report submitted to the Office
of Technology Assessment of the U.S. Congress, on file with the Columbia Law Review). I
mean only to suggest that, when the legal system makes a reasoned decision to grant
certain IPRs, courts should enforce these rights through injunctions (i.e., a property rule)
and thereby encourage private transactions.
COLUMBIA LAW REVEW[
2664
[Vol. 94:2655
adjustments are either ignored or subsumed in a sweeping price adjust3
ment scenario. '
II.
INCORPORATING TRANSACTION Cosrs: PROPERTY vs. LIABILITY RULES
Coase's point was that transaction costs determine where rights
should reside. The property/liability rule dichotomy put forth by Guido
Calabresi and A. Douglas Melamed systemized thinking about rights allocations in the presence of transaction costs. 3 2 They argued that the ex-
tent and nature of transaction costs in a particular case dictate whether
one of the parties to the Coasian bargain ought to have an absolute property right or simply the right to collect damages caused by the other
party's encroachment (i.e., a "liability rule"). This doctrine holds that the
following factors point toward adoption of a property rule: few parties,
difficult valuation problems, and otherwise low transaction costs. Other
factors indicate that a liability rule might better effectuate the bargain:
many parties (especially when one party has the power to "holdup" the
entire enterprise), the likelihood of strategic bargaining, and otherwise
33
high transaction costs.
Thus, in the Calabresi and Melamed framework, a property rule allows the right-holder to set her own asking price through ex ante negotiations when someone begins to interfere with the holder's activities.3 4
Under a liability rule, however, a court sets the price in a proceeding that
typically takes place after the right has been infringed. The choice of
which rule to apply depends on the degree and nature of the transaction
5
costs.
3
IPR cases fit the criteria set up by Calabresi and Melamed for application of a property rule: (1) there are only two parties to the transaction;
(2) the costs of a transaction between the parties are otherwise low; and,
most importantly, (3) a court called on to set the terms of the exchange
would have a difficult time doing so quickly and cheaply, given the specialized nature of the assets and the varied and complex business environments in which they are deployed. Because each asset covered by an IPR
is in some sense unique-a characteristic guaranteed by various requirements for protectability in intellectual property statutes-it is difficult for
a court in an infringement case to properly value the right-holder's loss.
Hence, the parties should be left to make their own deal. For example,
recall the problems the court faced in Sheldon v. Metro-Goldwyn Pictures
Corp.3 6 valuing the contribution made by the copyrighted play to the film
that incorporated some of its plot elements. Courts are simply not wellsituated to make difficult valuations of this sort; surely the parties could
31. See Coase, supra note 6, at 170-74.
32.
33.
34.
35.
36.
See
See
See
See
309
Calabresi & Melamed, supra note 3, at 1106-07.
id. at 1106-08.
id. at 1105.
id. at 1106-10.
U.S. 390 (1940); for further discussion, see supra notes 15 and 16.
1994] PROPERTY RULES, COASE, & INTELLECTUAL PROPERTY 2665
be expected to do at least as well given their intimate knowledge of the
play, the film, and relevant industry norms. Property rules insure that
this knowledge is brought to bear in cases in which valuation is difficult.
A liability rule should prevail, by contrast, when it is relatively easy to
determine the value of an exchange by reference to an objective market
price.3 7 Contracts cases provide the paradigmatic example. Often, a
court can accurately calculate the appropriate compensation for the injured party. This ability to calculate damages accurately has the added
benefit-often touted as the primary rationale for money damages in
contract cases-of encouraging breach where the breaching party can
both fully compensate the injured party and enter into a substitute transaction with someone else who values the breaching party's performance
more highly. This is the concept of the efficient breach: a compensatory
remedy predicated on the ease of valuing the costs of breach and the
desirability (i.e., Pareto optimality) of directing the breaching party's per-
38
formance where it is most valued.
A persistent minority, however, advocates a property rule (i.e., specific performance remedy) as a baseline remedy even in contracts cases.
(It is interesting that many of the arguments made from this position
mirror conventional wisdom in the IPR field.3 9) In this context and
others, recent scholars have become ever more enamored of the logic of
property rules. 40 Even the foundational work of A. Mitchell Polinsky,
37. See Cabresi & Melamed, supra note 3, at 1106-08.
38. Less commonly, a court will order specific performance. This equitable remedy
requires the breaching party to fully perform the contract, rather than merely to
compensate the injured party for the costs of the breach. A number of celebrated law
review authors have argued that the frequency of the two remedies should be reversed,
with specific performance being designated as the default remedy. According to scholar
Thomas Ulen,
[t]his method of protecting entitlements-a method that the authors [Calabresi
and Melamed] call one of property rules-is the most efficient means when the
level of transaction costs between the parties in conflict is low. Only in those
circumstances is it possible for voluntary exchange to determine which of the
competing uses is of higher value.
Thomas S. Ulen, The Efficiency of Specific Performance: Toward a Unified Theory of
Contract Remedies, 83 Mich. L. Rev. 341, 367 (1984). See also Anthony T. Kronman,
Specific Performance, 45 U. Chi. L. Rev. 351 (1978); Alan Schwartz, The Case for Specific
Performance, 89 Yale LJ. 271 (1979). But so far, courts have, for the most part, adhered to
the traditional default of money damages for breach. See Edward Yorio, In Defense of
Money Damages for Breach of Contract, 82 Colum. L. Rev. 1365 (1982). The interesting
point is that many of the arguments made by the proponents of specific performance as
the proper norm in contract law have long been made-and accepted-in intellectual
property law.
39. See Schwartz, supra note 38; Ulen, supra note 38.
40. The interesting contribution of Haddock, McChesney, and Spiegel, for example,
sets out a general theory supporting the dominance of property rules based on an
Edgeworth Box model. See David D. Haddock et al., An Ordinary Economic Rationale for
Extraordinary Legal Sanctions, 78 Cal. L. Rev. 1 (1990). On the Edgeworth Box, a simple
model of the bargaining range between a seller and buyer, and the related notion of the
"contract curve," see Donald N. McCloskey, The Applied Theory of Price 88-90 (1982). In
2666
COLUMBIA LAW REVIEW
[Vol. 94:2655
who argued forcefully that in many situations there is little difference between property rules and liability rules, 4 1 supports the strong property
rule baseline for IPRs. Polinsky argued that a property rule is superior to
a liability rule when "the court lacks information about both damages and
benefits." 42 Without accurate information, the damages may be set below
the Haddock et al. model, a liability rule returns the plaintiff to the indifference curve she
was on at the outset of the transaction. This rule deprives that party of the opportunity to
capture some of the surplus of a mutually advantageous bargain. (This bargain is usually
predicted to lie on the intersection of two indifference curves reflecting higher utility for
each party.) In this, Haddock et al. resembles several earlier papers. Polinsky, in a
discussion of strategic bargaining under both property and liability rules, frequently
mentions the availability of "gains from trade." See, e.g., A. Mitchell Polinsky, Resolving
Nuisance Disputes: The Simple Economics of Injunctive and Damage Remedies, 32 Stan.
L. Rev. 1075, 1092 (1980). Another paper points out that when two neighboring activities
involve benefits as well as costs to the two parties, and a liability rule merely places one
party in the position she would be in absent any interaction with the other party, each
party might prefer to bear liability. This would then allow that party to reap all the
proximity gains while paying out a measure of damages that, by assumption, yields net
profits. See Susan Rose-Ackerman, I'd Rather Be Liable Than You: A Note on Property
Rules and Liability Rules, 6 Int'l Rev. L. & Econ. 255, 255 (1986). David Schap argues that
when court-set damages systematically fail to reflect the plaintiff's full economic costs, for
example, when the plaintiff suffers "psychic costs" that are difficult for a court to value,
property rules should be preferred. See David Schap, The Nonequivalence of Property
Rules and Liability Rules, 6 Int'l Rev. L. & Econ. 125, 129 (1986). Note the applicability of
Schap's thesis to IPRs which, I have argued, are rife with valuation problems. The failure
of liability rules to allocate fairly the bargaining surplus justifies the award of
"extraordinary" (i.e., supracompensatory) damages in certain cases, and points toward a
property rule in others. While this is a different rationale for a property rule than the one
presented here, it places similar emphasis on the importance of bargaining.
At the same time, the assumption that a liability rule merely returns the plaintiff to the
same utility level she occupied before the transaction does not ring true in at least one
situation with which I am familiar: damages in patent cases. The benchmark remedy of
the patentee's "reasonable royalty" is calculated on the basis of a "hypothetical
negotiation" between patentee and infringer, conducted prior to the commencement of
infringement. See Merges, supra note 2, at 786-88. This measure of damages, though
clearly a liability rule (since it applies to past infringements), most assuredly does not seek
to return the patentee to her initial indifference curve; it clearly envisions giving the
patentee a share of the value added by the infringer. It, therefore, attempts to mimic a
property rule, thus eliminating the criticism of liability rules levelled by Haddock et al.
41. See Polinsky, supra note 40, at 1111. For example, Polinsky showed that strategic
bargaining, which many commentators had associated exclusively with property rules, is
also present in many cases under a liability rule. This follows from the observation that
there will often be a difference between the court-set damages and the plaintiff's actual
injury, which creates the same opportunities for (and costs of) bargaining as a property
rule. Interestingly, one common practice in the music industry appears to confirm this
observation. This is the practice of bargaining around the statutory compulsory licensing
regime for "mechanical" or "cover" royalties via the Harry Fox Agency, a New York
"clearinghouse" for performance rights. See Merges, supra note 23, at 31 (noting that,
unlike many of the compulsory licensing provisions described below, the statute permits
parties to contract around the statutory rate). See generally Ian Ayres & Eric Talley,
Solomonic Bargaining- Dividing a Legal Entitlement to Facilitate Coasean Trade, 104 Yale
LJ. (forthcoming 1994) (describing bargaining in the shadow of a liability rule).
42. Polinsky, supra note 40, at 1112.
1994] PROPERTY RULES, COASE, & INTELLECTUAL PROPERTY 2667
the actual level of harm, encouraging the "injurer" (or infringer) to engage in an excessive level of activity-in our case, increased infringement.
Thus, the typical IPR transaction meets Polinsky's conditions for the dominance of a property rule, despite his general point that there is often less
difference between these forms of entitlement than others had
43
presumed.
III.
EXCEPTIONS TO THE STRONG PROPERTY
RuLE
FOR
IPRs
All familiar with the IPR field recognize the strong presumption in
favor of injunctions. As one court put it, in ruling that an injunction is
the normal remedy for patent infringement: "Without the right to obtain
an injunction, the right to exclude granted to the patentee would have
only a fraction of the value it was intended to have, and would no longer
be as great an incentive to engage in the toils of scientific and technological research." 44 Copyright shares the same baseline. As Paul Goldstein
says, "[c] ourts exercise their statutory authority to grant.temporary injunctive relief more readily in copyright actions than in other intellectual
property cases. Final injunctive relief, though similarly discretionary, is
ordinarily available." 45 Moreover, as Goldstein points out, copyright
shares similar rationales for its strong property rule:
[A] Ithough coercive relief [which includes injunctions] will sometimes overcompensate the copyright owner, monetary relief
alone will often undercompensate it. Damages are hard to
prove, the infringer's profits may be less than the copyright
owner could have earned in the same market, and statutory
damages... may systematically undercompensate 4the
copyright
6
owner for its business risk in producing the work.
Despite widespread acceptance of a property rule, departures from it
are not unknown in the IPR context. It is worth reviewing these briefly,
and thereby implicitly asking whether Reichman has justified his own
proposed departure in his paper.
43. Alan Schwartz pursued a similar tack in a recent article, where he argues in favor
of enforcement of contractual provisions stipulating a specific performance remedy. See
Alan Schwartz, The Myth That Promisees Prefer Supracompensatory Remedies: An
Analysis of Contracting for Damage Measures, 100 Yale LJ. 369, 371 (1990). Schwartz
presumes that the presence of such a term in a contract is a sure sign that the parties have
had difficulty valuing the potential loss from nonperformance of the contract; regardless
of the merits of this as a general proposition, the statutorily-defined status of IPRs as
unique assures that it is realistic in the cases of interest to me.
44. Smith Int'l, Inc. v. Hughes Tool Co., 718 F.2d 1573, 1577 (Fed. Cir.), cert. denied,
464 U.S. 996 (1983). See generally BradfordJ. Duft, Patent Preliminary Injunctions and
the United States Court of Appeals for the Federal Circuit, 65J. Pat. Off. Soc'y 131 (1983)
(arguing that judicial discretion to grant preliminary injunctions is neither prejudice nor
whim). The Court of Appeals for the Federal Circuit has softened the "per se" injunction
rule of Smith to a limited extent in more recent cases. See Merges, supra note 2, at 749-77.
45. 2 Paul Goldstein, Copyright: Principles, Law and Practice § 11.0, at 247-48
(1989) (footnotes omitted).
46. Id. at 249 (footnotes omitted).
2668
COLUMBIA LAW REVIEW
[Vol. 94:2655
There are two general exceptions to the strong property rule for
IPRs. Compulsory licenses coerce IPR holders to transfer certain rights
to certain classes of licensees. In addition, various doctrines, best explained as responses to market failure of one
kind or another, excuse
47
infringement of otherwise protected works.
Defenders of the exceptions say high transaction costs provide ajustification. I shall not touch on the second exception, market failure,
though there is much of interest to say about it. Regarding compulsory
licenses, I have, as noted earlier, taken exception with those who would
rush to implement these licenses before institutions have a chance to
emerge to reduce IPR transaction costs.
Compulsory license provisions set automatic terms for IPR transactions. In general, they are available for patents only as a remedy for violation of the antitrust laws. 48 In copyright law, however, although they are
still the exception, they are found more frequently. Currently the copyright statute provides for compulsory licensing of four classes of works:
(1) musical recordings; 49 (2) songs played on jukeboxes; 50 (3) certain
47. In the eyes of some commentators, a middle ground of ongoing, compensated
infringement should be permitted: a liability rule crafted byjudges. See Pierre N. Leval,
Toward a Fair Use Standard, 103 Harv. L. Rev. 1105, 1131, 1134 (1990) (prescribing no
injunction remedy in cases where private unpublished papers of authors or other public
figures are infringed in course of writing biography). In a case later reversed by the
Second Circuit, Salinger v. Random House, Inc., Judge Leval wrote (in dictum): "With all
respect for the rights of a copyright owner, a publication should not be enjoined where the
degree of infringement is trifling and inconsequential." Salinger v. Random House, Inc.,
650 F. Supp. 413, 426 (S.D.N.Y. 1986), rev'd, 811 F.2d 90 (2d Cir. 1987). See also Abend v.
MCA, Inc., 863 F.2d 1465, 1478, 1479 (9th Cir. 1988), aff'd sub nom. on other grounds,
Stewart v. Abend, 495 U.S. 207 (1990) (copyrighted story, basis of film "Rear Window,"
directed by Alfred Hitchcock and starring Jimmy Stewart and Grace Kelly, infringed by
showing of film after copyright grant lapsed; but copyright holder entitled only to
continuing royalty on film profits, rather than injunction). See generally Ralph S. Brown,
Civil Remedies for Intellectual Property Invasions: Themes and Variations, Law &
Contemp. Probs., Spring 1992, at 45, 51-52 (identifying two classes of cases where
preliminary injunctions are sometimes denied: (1) unpublished papers; and (2)
copyrighted works incorporated into larger works where the license for the copyrighted
work is terminated under the grant-renewal provision of the Copyright Act).
Overtones of a market failure theme may be heard where infringing parodies qualify
for the fair use defense. See Fisher v. Dees, 794 F.2d 432, 437 (9th Cir. 1986) ("Parodists
will seldom get permission from those whose works are parodied. Self-esteem is seldom
strong enough to permit the granting of permission even in exchange for a reasonable fee
.... The parody defense to copyright infringement exists precisely to make possible a use
that generally cannot be bought."); Harriette K. Dorsen, Satiric Appropriation and the Law
of Libel, Trademark, and Copyright: Remedies Without Wrongs, 65 B.U. L. Rev. 923
(1985) (collecting cases). See generally Gordon, supra note 23; Robert P. Merges, Are You
Making Fun of Me? Notes on Market Failure and the Parody Defense in Copyright, 21 Am.
Intell. Prop. L. Ass'n QJ. 305 (1993).
48. See Merges, supra note 2, at 906-08.
49. See 17 U.S.C. § 115 (1988). See infra notes 59-64 and accompanying text.
50. See 17 U.S.C. § 116 (Supp. V 1993). See also Scott M. Martin, The Berne
Convention and the U.S. Compulsory License for Jukeboxes: Why the Song Could Not
Remain the Same, 37 J. Copyright Soc'y U.S.A. 262, 315-26 (1989) (describing recent
1994] PROPERTY RULES, COASE, & INTELLECTUAL PROPERTY 2669
cable television transmissions; 5 ' and (4) certain uses of copyrighted works
by public television. 52 Although compulsory licensing is relatively rare,
these provisions cover a substantial portion of the copyrighted works in
circulation.
Paul Goldstein's important recent treatise on copyright law53 cites
the conventional justification of these compulsory licensing provisions on
the basis of transaction costs. The Act, he states, contains provisions
which "set a legislatively predetermined or administratively prescribed
rate for specified uses, thus removing negotiation costs as an element of
copyright transactions." 54 But Goldstein has also noted, perceptively, that
compulsory licensing provisions may prevent the creation of technologies
and organizational innovations that would efficiently administer the
rights-clearance process.5 5 To the extent this is correct, one must be hesitant to endorse compulsory licensing.
Indeed, the history of collective rights organizations such as
ASCAP 5 6 supports the main theoretical point raised earlier: that a property rule for IPRs can be transformed into a voluntary liability rule, in the
negotiated agreement between the collective rights societies and the jukebox trade group
for a voluntary licensing system in light of changes to United States copyright law
necessitated by United States adherence to the main international copyright union, the
Berne Convention).
51. See 17 U.S.C. § 111(d) (1988 & Supp. V 1993) (cable retransmissions). See the
related provision on satellite retransmissions by "Superstations" at 17 U.S.C. § 119 (1988 &
Supp. V 1993). Under section 111, a cable system that rebroadcasts a non-network signal
from a television station must pay compulsory royalties. Prior to 1993, these royalties were
set by an administrative group known as the Copyright Royalty Tribunal. Under recent
legislation, the functions of the Tribunal will be performed by special arbitration panels
convened by the Register of Copyrights. See 17 U.S.C. § 801 (1988 & Supp. V 1993).
Interestingly, the new panels, if successful, might be made to approximate the bargaining
groups that arrive at industry-wide liability rules as studied in this paper. If they avoid
capture, then the legislatively-mandated royalty-setting apparatus may end up being a
codification of the type of process that would have arisen in the absence of legislation.
52. See 17 U.S.C. § 118 (1988 & Supp. V 1993).
53. See 1 Goldstein, supra note 45.
54. Id. at 19. Paul Goldstein recently noted that
Congress has always tempered the extension of copyright with recognition of the
problem of transaction costs-the problem that some individual uses of
copyrighted works will be so dispersed that any eventual license royalties will not
repay the expense of enforcement and negotiation.
Paul Golstein, Copyright, Law & Contemp. Probs., Spring 1992, at 79, 84 [hereinafter
Goldstein, Copyright]. Goldstein has emphasized that he believes the transaction cost
rationale to be dubious in at least some cases. See supra note 23.
55. See Goldstein, Copyright, supra note 54.
56. ASCAP licenses the music performance rights of composers and licenses them out
to radio stations, television networks, nightclubs, bars, and the like, almost always on the
basis of a "blanket license" covering ASCAP's complete repertoire. See Buffalo
Broadcasting Co. v. ASCAP, 546 F. Supp. 274, 277 (S.D.N.Y. 1982), rev'd, 744 F.2d 917 (2d
Cir. 1984), cert. denied, 469 U.S. 1211 (1985). Note that an international study of IPR
collectives by a group of economists found striking similarities in their basic organization
and administration. See Stanley Besen et al., An Economic Analysis of Copyright
Collectives, 78 Va. L. Rev. 385-90 (1992).
2670
COLUMBIA LAW RbVEWl
[Vol. 94:2655
form of an effective institution to carry out IPR transactions. ASCAP acts
as a central depository for members' rights to control public performances of their works. It issues "blanket licenses" covering the relevant
copyrights of all members of the society to radio and television stations
and other entertainment outlets. It then monitors the songs played and
divides up the total receipts among all members on the basis of a complex pro rata formula. Broadcast Music Incorporated (BMI) is a rival organization, founded expressly to compete with ASCAP, that operates similarly.5 7 These are only two of the seventy-two music-related collective
rights organizations that operate in some 182 countries.5 8 Each, to a varying degree, regularizes transactions among holders of strong "property
rule" entitlements. In many cases these organizations also establish compensation schemes that operate according to agreed-upon formulas, such
as the broadcast-monitoring based system run by ASCAP. These preagreed compensation formulae are akin in many ways to statutory compulsory licensing formulae, which is why I call the institutions that administer them instances of a voluntary, or contractual, liability rule.
One may accept all that has been said so far, yet still object to the
notion that strong property rules are to be preferred in the IPR context.
The objection might be based on the observation that at most what has
been demonstrated in the preceding paragraphs is that property rules
can form the basis of operative IPR-exchange mechanisms. What has not
been shown, it might be argued, is that statutory liability rules are worse.
While one instance cannot serve as a complete response to this objection,
I hope that a brief review of the compulsory license for "cover" versions of
recorded compositions will illustrate the deficiencies of compulsory
licenses.
The history of the mechanical compulsory license begins with the
Supreme Court's 1908 ruling in White-Smith Music Publishing Co. v. Apollo
Co.5 9 that player piano rolls were not "copies" of copyrighted music.
Congress responded by recognizing recording and mechanical reproduction rights as part of the bundle of rights secured by copyright law.60 The
1909 Act, however, reflected congressional concern with the market
power conferred on the sole dominant music roll firm by virtue of the
new right: under the Act, any manufacturer of recordings or mechanical
reproductions could use a musical composition as long as that composition had already been licensed for mechanical reproduction and as long
57. On the formation of BMI, see Lee C. White, Musical Copyrights v. The Anti-Trust
Laws, 30 Neb. L. Rev. 50, 54-55 (1950) (describing formation of BMI by radio group, the
National Association of Broadcasters, in 1941 in response to steep price hike by ASCAP).
58. See David Sinacore-Guinn, Collective Administration of Copyrights and
Neighboring Rights 5 (1993).
59. 209 U.S. 1, 16-18 (1908).
60. See Copyright Act of 1909, ch. 320, 35 Stat. 1075 (1909), now superseded by 17
U.S.C. § 115 (1982) (providing that any copyright owner who licensed a work for
reproduction is subsequently subject to compulsory licensing of that work).
1994] PROPERTY RULES, COASE, & INTELLECTUAL PROPERTY 2671
as the manufacturer paid a royalty to the copyright owner. 61 At the time
of the 1909 Act one piano roll firm, the Aeolian Company, dominated
the industry. 62 Congress enacted the mechanical license in response to
Aeolian's perceived market power.
There were soon hundreds of record companies, mooting the origi-
nal rationale for the license. 63 Yet the record companies had come to
rely on the license as a cheap source of material. One might have expected a titanic battle over the preservation of the license, pitting the
entrenched record companies against composers and consumers. Yet no
such battle ever took place. Composers, apparently accepting the license
as an immovable lesson in interest group influence, simply live with it to
observers that it continues
this day-despite the cries of knowledgeable
64
to distort the market for compositions.
It seems this industry (and others like it) 65 is locked into a suboptimal liability rule, due to the difficulty of rooting out a compulsory li61. See id. § 1(e).
62. As described by a knowledgeable source:
In anticipation of the [legislation], Aeolian had entered into exclusive contracts
with many of the major music publishers of that time for the right to make
mechanical reproductions of the works in the repertoires of the publishers. [As a
result,] few other manufacturers of piano rolls could acquire the licenses they
required to compete in the piano roll business. To thwart this potential
monopoly, Congress provided in the 1909 copyright law [for the compulsory
license].
Al Kohn & Bob Kohn, The Art of Music Licensing 310-11 (1992).
63. See Ralph S. Brown & Robert C. Denicola, Cases on Copyright 497 (5th ed. 1990).
See also David E. Kronemyer andJ. Gregory Sidak, The Structure and Performance of the
U.S. Record Industry, in Entertainment, Publishing and the Arts Handbook 263, 266 (John
D. Viera & Robert Thorne eds., 1986) ("A surge in demand for prerecorded music in the
late 1950s and middle 1960s spawned numerous labels . .
").
64. See Register of Copyrights, 87th Cong., 1st Sess., Copyright Law Revision: Report
on the General Revision of the U.S. Copyright Law 35 (House Comm. Print 1961):
Removal of the compulsory license would be likely to result in a royalty rate, fixed
by free negotiation, of more than the present statutory ceiling of 2 cents. The
record companies would, of course, lose the advantage of the lower rate ....
[W]e would conclude that the 2-cent ceiling denies authors and publishers the
compensation due them for the use of their works.
65. It is certainly not unique in this respect. The jukebox compulsory license, which
had also become outdated, proved very hard to eliminate as well. Like record companies,
jukebox proprietors were well organized politically. See Brown & Denicola, supra note 63,
at 501:
Attempts to repeal the jukebox license] were frequent but unsuccessful. The
operators who supply machines and the taverns and other establishments who by
custom get half the take constitute a broad political base, in contrast to the
concentration of the popular music industry in a few urban centers. They had
the further advantage of the status quo.
Scott Martin, supra note 50, at 315-27, however, describes the recent negotiated
agreement between the collective rights societies and the jukebox trade group for a
voluntary licensing system in light of changes to United States copyright law necessitated by
United States adherence to the main international copyright union, the Berne
Convention.
2672
COLUMBIA LAW REVIEW
[Vol. 94:2655
cense. This persistence should not be a surprise; studies in the public
choice tradition predict just such a status quo bias. Although it is of
course not impossible for music publishers and composers to eliminate
the license from the statute, the recording companies seem to have at
least enough political muscle to block these efforts.6 6
The difficulty of dislodging compulsory licenses even in the face of
radically changed circumstances is one of the primary reasons to favor
voluntary institutions of the sort I describe earlier in this Comment.
For present purposes, I want to discuss two implications of my observation that compulsory licenses for IPRs are sometimes embraced rashly.
First, I find it curious that scholars who have studied legal issues arising
out of innovations and other creative works have failed to realize that
transactional difficulties are no less capable of calling forth new responses than technological or creative impasses. Perhaps it is an example
of Oliver Williamson's point that
organizational innovations are under67
studied and underappreciated.
Second, I have come to see that the rush to implement liability rules
in the face of transaction costs reveals a general weakness in property
rights theory. Most writing on property rights theory treats only a small
number of enforcement techniques and technologies. Furthermore, the
literature assumes that these techniques are either exogenously determined or subject only to minor optimization. 68 One of a small set of
66. At least one experienced witness to the political economy of the IPRs recognizes
this. In a recent speech, Rep. William Hughes, then-Chair of the House Subcommittee on
Intellectual Property, observed that
I have explored ways to repeal the compulsory licenses but have not met with
great success yet. Hope springs eternal, though. There are various reasons for
this present lack of success, depending upon which license we are talking about.
There is, though, one common thread: some companies, after having adapted to
life under a compulsory license, are more comfortable with the devil they know
than with the devil they don't. I understand this. But it is interesting that
business people who complain about government interference seem reluctant to
live in a free copyright market.
Representative William Hughes, before the Copyright Society of U.S.A., in 46 Pat.
Trademark and CopyrightJ. (BNA) 526.
67. See Oliver E. Williamson, Markets and Hierarchies: Analysis and Antitrust
Implications 192-93 (1975).
68. See, e.g., Yoram Barzel, Economic Analysis of Property Rights 73-74 (1989)
(citing two limited techniques for reducing policing costs: economies of scale in
measuring product attributes, and elimination of duplicate measurements); 1 Harold
Demsetz, Ownership, Control, and the Firm: The Organization of Economic Activity 31,
41-43 (1988) (describing as "methods which are likely to arise in the market and which
will lower the required police cost," vertical integration and bundling of hard-to-police
commodities with others that are not); Gary D. Libecap, Contracting for Property Rights
16-17 (1989) (discussing incentives to more precisely define property rights after external
shocks such as changes in factor prices or production costs, enforcement technology, or
political parameters); Robert C. Ellickson, Property in Land, 102 Yale LJ. 1315, 1320 n.14
(1993) ("In this Article, a land regime is treated as a dependent variable that is affected by
technologies, scale efficiencies, risks, ideologies, and other variables regarded as
independent."). Libecap comes closest to my interpretation of institution formation as a
1994] PROPERTY RULES, COASE, & IN=TLLECTUAL PROPERTY 2673
enforcement techniques is assumed to be a more or less natural response
to increasing asset valuation coupled with enforcement difficulties. 69 My
point is twofold: institutions are enforcement technologies too, and they
are often generated intentionally to reduce transaction costs and thus increase the value of assets. To the extent institution-building investment is
pervasive, we need to incorporate the possibility of a wide range of actively generated endogenous enforcement "technologies"-especially institutions-into our analysis of property rights.
CONCLUSION
In this brief paper, I have trained my attention on the application of
property rights theory to intellectual property. I hope I have clarified
some foundational issues submerged in proposals such as Reichman's. I
also hope I have hinted at the potential gains in property rights theory
that will flow from a study of the distinct characteristics of intellectual
property.
response to high transaction costs; he emphasizes incentives to change property rights
(including enforcement techniques) in the political arena. Yet his focus on the legislature
as the primary source of enforcement techniques differs substantially from my emphasis on
non-legislative institutions.
69. Thus the canonical instance of barbed wire which, once developed, changed the
dynamics of western land use and hence the value and nature of property rights in western
land. For example, the fine recent article by Robert Ellickson includes this discussion:
The efficiency thesis predicts that innovations in technologies for marking,
defending, and proving boundaries lead to more parcelization because they
reduce the transaction costs of private property regimes. According to this view;
for example, Glidden's invention of barbed wire in 1874 should have stimulated
more subdivision of rangeland in the American West. And this indeed appears to
have occurred.
Ellickson, supra note 68, at 1330 (footnotes omitted).