Volume 1, No. 2, May 6, 2004

Comment on S. J. Liebowitz
Alternative Copyright Systems:
The Problems with a Compulsory License

Based on a Presentation at a Conference on
Promoting Markets in Creativity: Copyright in the Internet Age
Washington, DC | June 10, 2003

Sponsored by
The Progress & Freedom Foundation
and the
National Center for Technology and Law of the
George Mason University Law School (The Tech Center)

The criteria classically used to identify public goods, nonrivalrousness and nonappropriability, [1] reflect the respective concerns of those who would restrict and those who would expand intellectual property rights. Intellectual property scholars often emphasize the apparent nonrivalrousness of intellectual goods, observing that use by one person does not prevent or diminish use by another. [2] The appeal of arguments to limit the scope of pharmaceutical patents or shorten the copyright term stems from the familiar observation that once intellectual property exists, allocative efficiency can be satisfied only if it is priced at marginal cost. [3] Nonappropriability, meanwhile, is not the justification but the goal for those who would expand intellectual property rights. [4] Ensuring that inventors and authors receive as much of the value as possible of what they create optimizes the incentive for intellectual property creation. [5] Advocates of strong property protections for intellectual property worry that technological developments have increasingly facilitated the unauthorized appropriation of intellectual property, [6] particularly copyrighted works. Absent a strong regulatory and technological response, these commentators worry, the incentives to produce certain types of works might be undermined.

The criteria not only help frame existing debates, but also directly implicate a new one: Are copyrighted works public goods? This question has received little direct attention, [7] but the case seems at least as strong as for many other goods that are commonly thought to have public good characteristics. [8] Bridges, for example, are often considered to be public goods, even though many suffer from congestion and even though tolls can be used to reduce appropriability. Intellectual property, in contrast, generally meets the nonrivalrousness criterion. My using a book or software program does not decrease the utility that you might obtain from the same book or software program. Indeed, for some works, positive network effects mean that usage is not merely nonrivalrous, but also complementary. Intellectual property fits the nonappropriability criterion also, but not perfectly. Some creators of copyrighted works in particular are able, at least in practice, to prevent most free-riding. For an increasing number of works and an increasing number of consumers, however, the availability of digital copying technologies has dramatically increased copying. The more frequent copying becomes, the more closely copyrighted works resemble public goods.

Stan Liebowitz's paper, Alternative Copyright Systems: The Problems with a Compulsory License , [9] recognizes that the important question is not so much whether copyrighted works are public goods, but rather, assuming that they have some characteristics of public goods, whether the government can find means of efficiently furnishing them. Discussions of public goods often omit this critical consideration. Even if a bridge has public good characteristics, we might still prefer private construction and operation of bridges if the government is particularly incompetent at providing and running them, at least as long as there is sufficient slack in the nonappropriability criterion that private firms will have some incentive to enter the market. So too, indeed especially so, for copyright. Even if copyrighted works have public good characteristics, we might doubt that a Patronage Agency would hire the right employees to produce sufficiently appealing rap music or novels. Assuming that a government agency cannot efficiently amass the talent necessary to generate creative works, the question becomes whether government can create institutional mechanisms that will subsidize authors while still incentivizing production of copyrighted works of sufficiently high quality.

Alternative Copyright Systems offers skepticism and some much needed common sense in critiquing proposals that would require the government, either through general tax revenues or through taxes on ancillary products, to distribute money to copyright owners. [10] Liebowitz's analysis makes clear that the project is not a trivial one. A government compulsory license scheme might add new market inefficiencies while curing others, and moreover is potentially vulnerable to manipulation by private parties who seek to increase their payments. In the end, however, some form of experiment might be worthwhile as insurance should an expansion of copying erode the market to the point where even inefficient governmental provision would represent an improvement. In a world with private provision of bridges, the government would need to think about preparing to build bridges itself if a new technology allowed drivers to evade tolls. The market effect of copying so far has been relatively modest, but further technological improvements could have dramatic effects on the music industry and perhaps even the book-publishing industry.

Governmental subsidy, in addition, has at least the theoretical potential to represent an efficiency gain even if digital rights management systems were fully successful in protecting copyright owners' intellectual property rights. The deadweight loss attributable to monopoly pricing of copyrighted works is familiar to the literature on copyright, [11] and Liebowitz properly acknowledges that efficient levels of consumption can be achieved only if copyrighted goods are priced at zero cost. [12] If the government provides funding to copyright owners in exchange for the placement of works in the public domain, then deadweight loss is eliminated for those works. Similarly, the government can eliminate the deadweight loss attributable to prices above marginal cost for digital copying by compensating owners for a loss of the right to prevent file sharing. A government solution is superior to a private one as long as the inefficiencies attributable to government administration are less than the deadweight loss that the copyright regime imposes. Many governmental schemes would flunk even that test, and perhaps all would. [13] But this is not self-evident, and so it seems worthwhile at least to consider how a regime paying copyright owners might be administered.

Part I of this comment addresses some of the specific issues that Liebowitz identifies as counting against a government subsidization plan, focusing especially on the question of whether the government would be able to determine with sufficient accuracy how much money to give individual parties to compensate them for works placed in the public domain. I elaborate on some of the problems that Liebowitz identifies, but I also note that governmental simulation of market pricing for copyrighted works may not be as difficult as for many other goods. There are ready proxies for the popularity of copyrighted works, and while adopting any single proxy might invite manipulation by private parties, consumption of free works would provide considerable information about to enable relatively efficient governmental distribution of revenues to content producers.

Part II of this comment places the possibility of government subsidization of copyrighted works into a broader context by considering past proposals for patent buyouts. Those proposals seek to encourage the government to measure the value of patents and then pay patent holders for placing their patents into the public domain. In a past paper, I argued that these proposals are beset by technical challenges, as any system for calculating patent value will produce errors that will cause distortions in investment decisions. Errors, however, will be of less significance if they are made ex post, so that the directions of the errors cannot be anticipated, suggesting that ex post valuations of patents by flexible governmental decisionmakers may be more promising than ex ante valuations through rigid formulas. The challenge for the government in subsidizing intellectual property is technically simpler for copyrighted works than for patents, but the copyright context presents additional dangers inhering in governmental intervention that do not appear in the patent context. Most significantly, a copyright system that required subjective analyses of the quality of works might produce excessive governmental entanglement in creative processes.

Finally, Part III describes a scalable copyright regime that is responsive to the specific challenges of copyright. Unlike the "compulsory license" that Liebowitz considers, this system is opt-in, so copyright owners who wished to exploit their works in more traditional ways would still be able to do so. Copyright owners, however, also could choose to sell their works to private entities similar to collective rights organizations, but these entities would place the works in the public domain and ultimately would receive compensation based on governmental measures of the popularity of their portfolios of works. The key to such an approach is that the government would not need to assess the value of individual works, but the private organizations would still have an incentive to determine the extent to which individual works would be popular and thus increase the amount of money that they would receive from the government. An experiment with such an approach, launched by either the government or a private foundation, might help establish whether governmental payments to creators of copyrighted works could provide an adequate alternative incentive structure.

The most significant problem with copyright reward systems that Liebowitz identifies is informational. "Markets provide information," Liebowitz notes, "that is virtually impossible to determine in any other way." [14] The government is not likely to be able to develop a model that predicts with perfect accuracy how many consumers would have purchased a particular work at a given price. It may be more difficult still to calculate the full social value of works, including not only what consumers would have paid at the prices that would have been set by producers, but also how much consumer surplus those consumers would have received. Paying to copyright owners the full social value is a prerequisite if the goal of a subsidization scheme is to eliminate the nonappropriability problem altogether.

There are at least three significant aspects to this problem. The first, as Liebowitz notes, is technological. A reasonable starting point in determining compensation due would be a calculation of how many consumers obtained it. For digital works, however, it might be difficult to develop accurate counts of the number of downloads, as the original copyright owners would have strong incentives to download their own works repeatedly and thus increase the size of the subsidy that they will receive from the government. [15] A partial response is that alternative means of measurement might be employed, from consumer surveys to Nielsen-type ratings. But there might well be systematic differences between what users will download and either what they admit to downloading or what those who are selected as the allegedly random sample choose. Governmental selection of any single rating system would provide authors incentives to exploit that system's vulnerabilities.

The second problem is that of extrapolating from the number of consumers to an appropriate payment. Just because the same number of listeners download Song A and Song B does not necessarily mean that both are of equal value. Perhaps those who listen to Song A will tend to delete it after a few plays, while Song B will endure on its listeners' hard drives and affect their souls. Still thornier is the problem of translating listener enjoyment into dollar values. An approach that sought to track the market would place greater value on the auditory experiences of the rich than of the poor, perhaps finding a minute of classical music more valuable than a minute of rap. Yet just as administrative agencies are reluctant to value the lives of the rich at higher amounts than the lives of the poor in performing cost-benefit analysis, [16] so too might the government balk at market-mimicking pricing even if it were technologically feasible.

The third problem, to which Liebowitz alludes without fully exploring, is political. Liebowitz recognizes that a copyright subsidization system could reward authors alone or record companies and other producers of their content as well. If copying were freely allowed, however, there is a strong argument that the producers would not be necessary. Some commentators indeed have suggested that the primary benefit of Napster and post-Napster technologies is that they eliminate the need for content producers and other links in the distribution chain. [17] No federal program to subsidize copyrighted works, however, seems politically enactable in the absence of support from content producers. Authors may be too diffuse and poorly organized a group to exert enough lobbying power on their own. [18] The only feasible solutions thus might be the least achievable.

In sum, any attempt by the federal government to estimate the social value associated with copyrighted works will be at best imperfect and might result in excess payments to content producers relative to authors. Market valuation, however, is imperfect too, especially when performers are unable to appropriate the full value of their work as a result of copying. If college students are particularly likely to copy music, for example, then, at least in the long run, the result will be less of the type of music that these students tend to prefer. Serves them right, perhaps. But the punishment for piracy sins will be borne by a class of people that likely includes many who pay for their music. Moreover, the existence of a prisoners' dilemma equilibrium in which everyone or almost everyone copies does not furnish a justification for that equilibrium. To the extent that government intervention can avoid the result of complete defection, it should do so.

Even if copying were uniform across different categories of music (or of another class of copyrighted works), the market is unlikely to produce the optimal mix of works. As I have argued in detail elsewhere, [19] there may be inefficiently low differentiation among copyrighted works. New entrants into markets for copyrighted works will seek, to the extent that rationality underlies the entry process, [20] to create works that will maximize their profits, even if some other choice would maximize social welfare. For example, a new boy band might be able to lure enough customers from other boy bands to ensure a greater customer base than a new experimental music form, but the profits are at least partly "business stealing," [21] in industrial organization jargon, rather than market expansion through product differentiation. Ordinarily, active governmental intervention to correct for such effects seems likely to do more harm than good. But once government is in the business of making payments for the production of copyrighted works, it might seek to give higher rewards to relatively original works that produce relatively little business stealing. Rewarding originality rather than mere market share has the potential to be efficiency-increasing.

The imperfections of a program providing for subsidization while allowing consumer copying must thus be balanced against the imperfections of the market, including the deadweight loss associated with copyright, any distortions from asymmetric appropriation of copyrighted works by consumers, and any other anomalies that might arise under conditions of imperfect competition. The ultimate assessment is necessarily empirical, and Liebowitz's prediction that governmental allocation would impose net costs might well be justified. But it might not. Assessing the value of copyrighted works, after all, is not nearly as difficult a task as estimating the value of most commodities traded in markets, especially for a narrow class of copyrighted works such as sound recordings. At least the government should be capable of estimating the relative importance of different songs. Download counts provide just one of many means of assessing the popularity of different recordings, and while the government should be wary of relying exclusively on any single measurement that might be manipulated by authors or publishers, agencies might be able to develop reasonably accurate assessments by considering a variety of different proxies and measurement techniques.

Just as the imperfections of traditional copyright markets must be considered, however, Liebowitz is right to emphasize the drawbacks of taxation. Taxes produce distortions in economic activity. Liebowitz properly notes that a program funded from general tax revenues would produce the "general inefficiencies involved with the tax system," [22] and these inefficiencies can be substantial. [23] In addition, any governmental system invites the problem of rent-seeking activity, [24] as individual copyright owners seek to maximize the payments that the government makes to them, and the direct costs of such activity must be factored in as well. These considerations weaken the case for government subsidization, but a comparison of these costs with the others remains an uncertain empirical matter.

Excise taxes on ancillary products may be an alternative to funding a copyright subsidy program directly from general revenues, but a problematic one. The appeal of the approach arises from its targeted nature. If a perfect complement to a class of copyrighted works can be identified, then the consumers of those works can pay for the right to have access to those works through the targeted tax. The Audio Home Recording Act, for example, provided that makers of digital recording equipment would provide compensation to artists whose songs consumers would then be allowed to copy, in exchange for which the equipment makers would not be liable for such copying. [25] With the advent of CD burner peripherals, however, copying can be achieved through general-purpose computers, and many people use computers without copying music. A tax on recordable compact disks may reduce the problem of overinclusion, but only somewhat, [26] and it magnifies the problem of underinclusion, as many copiers of music will choose other media, such as hard disks.

There is yet a more serious problem with the excise tax approach: It tends to undermine the benefit of government subsidization, the reduction of deadweight loss. If computers become more expensive as a result of a transfer payment, then while computer users may have access to unlimited music, those who cannot afford computers will receive no benefit. Of course, the problem of an inability to afford copying technology applies to any proposal to place copyrighted works in the public domain, but the tax may increase the number who will choose not to make the initial capital investment even as the cost of the underlying computer technology declines. Similarly, if an excise tax is placed on recording media, then a marginal cost of consumption will apply to each downloaded song. The excise tax approach thus not only produces distortions (indeed, distortions that are likely to be greater than those imposed by a tax on income [27]), but it also adds an access charge for use of works in the public domain, making it not so efficiency-enhancing after all. This approach thus seems inferior to funding through general taxation, though the question once again is empirical.

Copyright theory often parallels patent theory, and indeed a considerable patent law literature considers the possibility that the government might pay patent holders the value of their inventions, in exchange for those inventions being placed in the public domain. [28] In a recent article, I criticized a number of such proposals, largely by identifying flaws in various attempts to describe algorithms by which the government might calculate the value of patents. [29] So many considerations are relevant to assessing the value of a patent that any attempt to do so will lead to shortchanging some and overvaluing others. Such misestimation threatens to yield adverse selection in a hypothetical opt-in system, with the most overvalued patents those that patent holders agree to place in the public domain. [30] Meanwhile, in a hypothetical mandatory system serving as an alternative to the patent system, research decisions would be based on imperfect governmental formulas rather than on market tests, meaning that governmental mistakes would directly produce inefficiencies in resource allocation.

I suggested, however, that the attempts might be more ambitious than necessary. The authors of the proposals assumed that governmental discretion was necessarily problematic, given the danger that governmental officials might undercompensate or overcompensate patent holders. [31] But governmental errors matter only if patent holders or prospective patent holders know the direction of these errors and change their decisionmaking as a result. If governmental decisions are made years after patents are placed into the public domain, patent holders will make their decisions based on an assessment of the reward that an average governmental official would give, so only systematic errors lead to distortions. Ex post evaluation also may be more accurate, because it can consider years of actual experience of public use of the patent placed into the public domain.

Such an approach may impose some risk on patent holders, but not risk of a magnitude greater than what already inheres in the research and development process. [32] Although I was ultimately agnostic about whether any particular administrative arrangement would have greater benefits than costs, the possibility of a patent prize system relying at least partly on governmental discretion in ex post decisions appeared promising. At least, it seemed to me, an experiment with an opt-in system might be used to assess the system on a modest scale in some area for which the deadweight costs of patents seem particularly salient, such as AIDS drugs. An experiment might help determine the risk cost imposed by the system, as well as whether problems like rent-seeking might be overcome. [33]

Given my optimism about the possibility that retrospective evaluation might make a patent prize system viable, it might seem that I would be at least as optimistic about copyright prizes. The task of valuing patents, after all, is far more complicated than that of valuing copyrights. Although copyrighted works may substitute for one another to some extent, the degree of substitution is far more idiosyncratic for patents, so patent valuation depends in part on factors like whether other companies will be successful in attempts to invent around patents and develop competing products. [34] Moreover, the relationship between patents, whose values must be assessed, and products, whose usage is far more easily observed, is complicated, at least outside the pharmaceutical context. A single patent might be licensed in any number of products, making more difficult both prospective valuation of a patent and retrospective assessment of what a patent would have been worth had it not been placed in the public domain. It would not be straightforward to assess the extent to which a patent placed in the public domain was even used in products, let alone the significance of the patent's contribution to those products. In the copyright context, by contrast, there exists more-or-less a one-to-one relationship between copyrights and works used by consumers.

An additional reason that a copyright prize system might seem a stronger case than a patent prize system is that patents more than copyrights may play significant roles besides encouraging inventive activity. A patent, for example, may help to consolidate and coordinate further research activity developing the initial invention, [35] and a patent may help provide incentives to commercialize an invention that otherwise might not be developed if second-mover advantages outweigh first-mover advantages. [36] These problems are not necessarily fatal to proposals for patent prize systems, but such systems at least must take them into account, for example by giving the original inventor some incentive to market the invention even after it is placed in the public domain. These considerations are smaller in the copyright context. Copyrights may help to coordinate further developments as a result of the expansive derivative right, [37] but the need for such coordination may be less than in the patent context, where almost any invention can stand to be improved. Moreover, commercialization may be less critical to successful operation of a copyright regime, because in the absence of commercialization activity, word of mouth and music sharing could help popular songs find audiences.

While a copyright prize system should thus be technically more feasible than a patent prize system, there may be a greater need for a patent prize system, at least assuming that the problem of unauthorized copying of copyrighted works does not become more severe than it is today. It may be, for example, that price discrimination is more likely with some copyrights than would usually be the case with patents, as lower-valuing consumers of copyrighted works can attend a movie theater at a relatively inconvenient time, or wait until a particular copyrighted work falls in price. [38] Because price discrimination by itself helps to reduce deadweight loss, [39] a copyright prize system would have less deadweight loss to eliminate. The empirics here are complex, however. Price discrimination conceivably could decrease allocative efficiency. [40] Moreover, patentees employ some degree of price discrimination as well, [41] and the extent to which such price discrimination helps reduce deadweight loss is controversial. [42]

Perhaps a more significant concern is that there may be less need to encourage additional inventive activity in the copyright area than in the patent area. Studies have suggested that the social value of patents is a large multiple of the value that patent holders are able to obtain, [43] indicating that the amount of inventive activity in that area is inefficiently low. Despite the inability of copyright holders to appropriate the full value of their works, there seems to be no shortage of copyrighted works. Would we be better off if there were twice as many copyrighted works of equal quality, once we take into account that the resources to create those works must be diverted from other activities? [44] The business-stealing effect means that it is possible for the number of copyrighted works to be socially excessive. [45] Even if copyright does not encourage an excessive number of works, the marginal work encouraged by a copyright prize program might be of relatively little value.

Finally, and perhaps most importantly, governmental involvement in a copyright prize system raises intrinsic difficulties. Even placing aside the First Amendment limitations on governmental subsidy of speech activity, [46] copyrighted works serve important democratic functions, and the involvement of government in selecting which copyrighted works to subsidize, and thus which genres or points of view to promote, is troubling. Many governments, including some with proud democratic traditions, do subsidize the dissemination of copyrighted works through state-run television stations. Arguably, such subsidization is more troubling than subsidization of music or books, because news programming has a direct influence on public opinion. In this country, however, governmental involvement in subsidizing even fine art has been controversial precisely because of the content of some of the works being subsidized. [47]

Subsidizing copyrighted works through individual examination of the works' quality might be impractical in any event, given the large number of works potentially eligible for buyouts. Combined with the controversy that would inhere in evaluation of particular works, this might seem to suggest that the government would need to employ some predetermined procedure, perhaps based on something as crude as statistics on how many copies of a song have been downloaded. But we have then come full circle. I have argued that a prize system works best when evaluations are made ex post and governmental decisionmakers enjoy considerable flexibility in making decisions. The problems such flexibility creates, however, seem to make systems relying on discretion unmanageable.

There is perhaps an escape from the dilemma, a way to retain the benefits of flexibility without empowering the government to promote the speech that it prefers. The key is to retain governmental flexibility in choosing among different methodologies for assessing works, but to require the government to evaluate portfolios of copyrighted works rather than individual works. The system would work as follows: The government would set aside some amount of money for copyright buyouts (perhaps only a few million dollars in an initial experiment), specifying a particular genre, such as sound recordings. Private entities, similar to collective licensing organizations, [48] would then compete to convince authors to place their works in the public domain (or just to permit file sharing), for example by paying the authors cash or by promising publicity. For each such agreement, the work would be added to the private entity's portfolio. At some later time, the government would rate each private entity's portfolio's contribution to the public and distribute the money set aside, plus any interest accrued, in proportion to such ratings.

If the process for rating a portfolio merely consisted of aggregating governmental decisionmakers' subjective preferences for particular works, this process would solve the problem of governmental entanglement only incompletely. At least the production of works would not depend on the preferences of any particular governmental decisionmakers, for if the delay period is sufficiently long, copyright owners will not be able to anticipate who the decisionmakers will be. Nonetheless, copyright owners might anticipate that the decisionmakers' preferences would reflect systematic biases relative to consumer preferences. There is little need, however, for subjective evaluation. The government agency could develop a variety of proxies for a random sample of works in the portfolio, based on downloads, surveys, and the like, or perhaps even more rigorous econometric studies that would seek to identify the extent to which products' markets were merely the result of business-stealing. Although the agency also might consider whether the private entity sought to manipulate any of these proxies, [49] it would rate the economic contribution of the portfolio as a whole.

The private entities, of course, would have incentives to achieve the best portfolios for the least amount of money by seeking out those works that will be expected to do best in whatever set of survey instruments that the government eventually decides to deploy. The approach has three significant additional advantages. First, by granting the government decisionmakers flexibility, the approach ensures that the deficiencies of any single formula established in advance will not lead to distortions. [50] Second, because the process would still be based on proxies, there would be no need for idiosyncratic evaluation of the underlying works. This avoids the specter of excessive governmental entanglement, and the random selection of works by the retrospective government evaluators greatly eases the administrative burden of such a system. Third, because the amount of money is fixed, any systematic tendency by governmental officials to overvalue or undervalue works will have no effect on the number of works placed in the public domain or on the government's expenditures.

Would the benefit of such a system exceed the cost? The total cost is easy to determine: the amount that the government has set aside (adjusted for any tax distortions created by taxation), plus any cost of paying the retrospective governmental decisionmakers. The money set aside indirectly pays for both the works placed in the public domain and for the efforts of private entities in assessing those works. The greater the costs associated with such evaluation, the fewer works private entities will be willing to buy to place in the public domain. The risk associated with uncertainty about the government's eventual evaluations similarly will depress the number of works added to the public domain. That risk is significant, but it may be small, because private entities can diversify their portfolios of works, and because those entities' shareholders can diversify their financial portfolios to minimize their total exposure.

The principal benefit is the reduction in deadweight loss as the placement of the copyrighted works in the public domain increases the public's access to those works. Even a relatively modest program conceivably could produce large benefits, because the value that consumers receive from copyrighted works presumably declines over the number of copyrighted works that they consume, just as utility is an increasing but concave function of wealth. The program might produce a sufficiently robust public domain to satisfy the needs of many consumers who otherwise would choose to purchase no or only a very small number of copyrighted works even though they would still obtain some benefit from additional works. The program could benefit connoisseurs of the relevant genre of copyrighted works as well, not just by saving money for which the government will now have to find a substitute, but more significantly by increasing the incentives to produce such works.

Hayek's skepticism about government's ability to substitute for the information generated by markets is well-founded. The government's need to substitute for markets is greater in some areas than in others, however. Intellectual property imposes a deadweight loss that must count as an inherent market inefficiency to be balanced against any inefficiencies of governmental administration. The appropriability problem in copyright, meanwhile, leads to the possibility that incentives to produce works may be inadequate. Given this combination, the possibility of a governmental program that complements the existing copyright system by encouraging the placement of certain works in the public domain deserves study. Alternative Copyright Systems provides a critical first step in such a study, noting problems with solutions that have been suggested so far. In this work, I have attempted to provide a second step by suggesting the possibility of an opt-in system and by arguing that ex post rewards made to private entities that accumulate portfolios of public domain works might be superior to the existing proposals. Whether the benefits would exceed the costs, as well as whether still further improvements might be offered, remains for further research.

* Associate Professor of Law, George Mason University, Visiting Associate Professor of Law, George Washington University

1. The terminology used for these criteria varies. See, e.g. , Earl R. Brubaker, Free Ride, Free Revelation, or Golden Rule? , 18 J.L. & Econ. 147, 148 (1975) (using the word "nonexcludability" instead of "nonappropriability"). Some definitions of "public good" focus exclusively on the nonrivalry criterion. See, e.g. , Robert B. Ekelund, Jr. & Robert D. Tollison, Economics 516 (3d ed. 1991).
2. See, e.g. , William W. Fisher III, Reconstructing the Fair Use Doctrine , 101 Harv. L. Rev. 1659, 1700 (1988) (noting that intellectual property "can be used and enjoyed by unlimited persons without being 'used up'").
3. Michael J. Meurer, Copyright Law and Price Discrimination , 23 Cardozo L. Rev. 55, 97-98 (2001) ("Allocative efficiency requires that the price of a copyrighted work be equal to the marginal cost of producing and distributing that work. Intuitively, efficiency calls for access to a work by all consumers with a valuation above the marginal cost of supplying a work. This outcome is achieved when the price equals marginal cost.").
4. It can serve as a justification as well, however. See, e.g. , Wendy J. Gordon, Assertive Modesty: An Economics of Intangibles , 94 Colum. L. Rev. 2579, 2587-88 (1994) ("[P]ublic goods-things that can be shared by many without physical diminution, and for which it is difficult to exclude nonpayors-can give rise to a pattern in which consumers will get less of the good than they would otherwise be willing to pay for. Such market failures can be costly enough to justify the law in imposing restraints on copying, such as the law of copyright, patent, and unfair competition.").
5. Intellectual property law, along with other branches of law such as contract, is what makes appropriation possible and thus ensures that there is some incentive for creative endeavours. Shi-Ling Hsu explains:
   
   This is not to say that a resource that has been costly to develop or cultivate, but can thereafter be nonrival in consumption, should be an open access resource. If the party developing or cultivating a valuable resource that is nonrival in consumption cannot recoup the costs of development or cultivation, there will never exist any incentive to develop or cultivate the resource. This issue can be most clearly seen in the area of intellectual property, where the knowledge created by patentable ideas and the value of copyrighted works are nonrival in consumption, but nevertheless the subject of protection from appropriation.
   
   Shi-Ling Hsu, A Two-Dimensional Framework for Analyzing Property Rights Regimes , 36 U.C. Davis L. Rev. 813, 849 n.149 (2003).
6. See, e.g. , Trotter Hardy, Property (and Copyright) in Cyberspace , 1996 U. Chi. Legal Forum. 217, 258-59.
7. But see David J. Brennan, Fair Price and Public Goods: A Theory of Value Applied to Retransmission , 22 Int'l Rev. L. & Econ. 347, 349-51 (2002) ( discussing the public goods nature of copyrighted works in the context of a discussion of retransmission rights) .
8. As James Buchanan has noted, virtually no goods are pure public goods. See James M. Buchanan, The Demand and Supply of Public Goods 49-50 (1968).
9. Stan J. Liebowitz, Alternative Copyright Systems: The Problems with a Compulsory License , in this volume p. ___.
10. See, e.g. , Neil W. Netanel, Impose a Noncommercial Use Levy to Allow Free P2P File-Swapping and Remixing (U. Tex. Law Sch. Pub. Law Research Paper No. 44, 2002), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=352560 (last visited June 5, 2003); William W. Fisher, An Alternative Compensation System , http://cyber.law.harvard.edu/people/tfisher/PTKChapter6.pdf (last visited July 15, 2003).
11. See, e.g. , Wendy J. Gordon, Authors, Publishers, and Public Goods: Trading Gold for Dross , 36 Loy. L.A. L. Rev. 159, 161 n.5 (2002) (discussing deadweight loss); Alireza Jay Naghavi & Günther G. Schulze, Bootlegging in the Music Industry: A Note , 12 Eur. J.L. & Econ. 57, 62-63 (2001) (" The static effect [of copyright] is that the protected artistic product or the intellectual property receives only suboptimal dissemination, i.e. it is underconsumed. Given that a product has been produced, it is optimal from a welfare point of view to sell it at its marginal cost, which is typically very low.").
12. Liebowitz, supra note 9 (manuscript at 8) ("Therefore, an economic requirement for efficient consumption of a nonrivalrous good is that any consumers who would like to consume it (e.g., a particular song) be allowed to consume the good.").
13. Liebowitz does not conclude definitively that a governmental compulsory license scheme would be infeasible. See Liebowitz, supra note 9 (manuscript at 2) ("[I]t is unclear that a compulsory license would have the net positive impacts that have been attributed to it.").
14. Id. (manuscript at 15).
15. Id. (manuscript at 12).
16. For a discussion of this issue, see David A. Hoffman & Michael P. O'Shea, Can Law and Economics Be Both Practical and Principled? , 53 Ala. L. Rev. 335, 359 (2002). Cf. Stephen F. Williams, Squaring the Vicious Circle , 53 Admin. L. Rev. 257, 266 (noting that cost-benefit analysis values the preferences of the wealthy more than those of the poor in issues other than life valuation).
17. See, e.g. , Raymond Shih Ray Ku, The Creative Destruction of Copyright: Napster and the New Economics of Digital Technology , 69 U. Chi. L. Rev. 263 (2002).
18. Perhaps the most notable legal victory of authors in recent years was judicial. See N.Y. Times Co. v. Tasini, 533 U.S. 483 (2001).
19. See Michael Abramowicz, Copyright Redundancy (unpublished manuscript, on file with author).
20. For an argument that bounded irrationality often leads to overentry into markets, see Avishalom Tor, The Fable of Entry: Bounded Rationality, Market Discipline, and Legal Policy , 101 Mich. L. Rev. 482 (2002). Irrationality might well be greater in markets for copyrighted works than in most other markets, given the force of optimism biases in this area.
21. For a discussion of business stealing, sometimes called demand diversion, see N. Gregory Mankiw & Michael D. Winston, Free Entry and Social Inefficiency , 17 Rand J. Econ. 48 (1986).
22. Liebowitz, supra note 9 (manuscript at 10).
23. One estimate suggests the cost of a marginal dollar increase in the income tax is between 17 and 56 cents. See Barton H. Thompson, Jr., The Endangered Species Act : A Case Study in Takings & Incentives , 49 Stan. L. Rev. 305, 355 (1997). With such wide confidence intervals, conclusive assessment of virtually any governmental program is difficult.
24. See generally Gordon Tullock, The Welfare Costs of Tariffs, Monopolies and Theft , 5 W. Econ. J. 224 (1967) (providing the seminal account of rent-seeking).
25. Pub. L. No. 102-563, 106 Stat. 4244 (codified at 17 U.S.C. §§ 1001-1010). See generally David M. Hornik, Recent Development Combating Software Privacy: The Softlifting Problem , 7 Harv. J.L. & Tech. 377, 405-09 (1994) (discussing the Act).
26. Liebowitz points out, "Not all blank CDs . are used to copy MP3 files." Liebowitz, supra note 9 (manuscript at 10).
27. See Louis Kaplow & Steven Shavell, Why the Legal System Is Less Efficient Than the Income Tax in Redistributing Income , 23 J. Legal Stud. 667, 680 (1994) (arguing that excise taxes are generally more ineffient than income taxes).
28. Steve P. Calandrillo, An Economic Analysis of Intellectual Property Rights: Justifications and Problems of Exclusive Rights, Incentives, and the Alternative of a Government-Run Reward System , 9 Fordham Intell. Prop. Media & Ent. L.J. 301 (1998); Robert C. Guell & Marvin Fischbaum, Toward Allocative Efficiency in the Prescription Drug Industry , 73 Milbank Q. , June 1995, at 213; Michael Kremer, Patent Buyouts: A Mechanism for Encouraging Innovation , 113 Q.J. Econ. 1137 (1998); Michael Polanvyi, Patent Reform , 11 Rev. Econ. Stud. 61 (1944); Steven Shavell & Tanguy van Ypersele, Rewards Versus Intellectual Property Rights , 44 J.L. & Econ. 525 (2001).
29. See Michael Abramowicz, Perfecting Patent Prizes , 56 Vand. L. Rev. 115 (2003). For other criticisms of these proposals, see F. Scott Kieff, Property Rights and Property Rules for Commercializing Inventions , 85 Minn. L. Rev. 697 (2001); and Douglas Gary Lichtman, Pricing Prozac: Why the Government Should Subsidize the Purchase of Patented Pharmaceuticals , 11 Harv. J.L. & Tech. 123 (1997).
30. See Abramowicz, supra note 29 , at 193-96.
31. The industrial organization literature has focused on the danger of undercompensation. See, e.g. , Jean Tirole , The Theory of Industrial Organization 401 (1988) ("Because the investor's investment is sunk at [the research and development] stage, the inventor is subject to the hold-up problem . . . . The administrative and judicial bodies in charge of prizes generally estimate the values of inventions very conservatively.").
32. See Abramowicz, supra note 29 , at 211-18.
33. Id. at 209-11 (discussing this issue).
34. Id. at 190-93.
35. See Edmund W. Kitch, The Nature and Function of the Patent System , 20 J.L. & Econ. 265 (1977).
36. Kieff, supra note 29 , at 708.
37. See 17 U.S.C. § 106(2). For a comment on the breadth of the derivative right, see Jed Rubenfeld, The Freedom of Imagination: Copyright's Constitutionality , 112 Yale L.J. 1, 50 (2002).
38. This is a form of second-degree price discrimination, which applies when a seller separates buyers whose valuations are hidden, typically by altering some aspect of the quality of the product sold. See Meurer, supra note 3 , at 71-75.
39. See, e.g. , Lichtman, supra note 29 , at 133 n.25 ("The more a producer is able to price discriminate-the more he is able to identify low-valuing consumers and charge them a correspondingly low price-the fewer the number of consumers needlessly excluded from the market.").
40. For example, an increase in movie price for relatively convenient showtimes and a decrease for inconvenient ones will reduce the consumer surplus of high-valuing users and will produce less consumer surplus for low-valuing users than paying marginal cost at a convenient showtime. See Meurer, supra note 3 , at 90-102 (explaining that price discrimination by copyright owners can increase or reduce allocative efficiency).
41. See, e.g. , Arti K. Rai, The Information Revolution Reaches Pharmaceuticals: Balancing Innovation Incentives, Cost, and Access in the Post-Genomics Era , 2001 U. Ill. L. Rev. 173, 201 (arguing that health insurance promotes price discrimination for pharmaceuticals).
42. Pharmaceutical companies have themselves pledged to offer some drugs at reduced prices for the poor, but critics have charged that the companies have not succeeded in providing the products. See Stephanie Storm & Matt Fleischer-Black, Drug Maker's Vow to Donate Cancer Medicine Falls Short , N.Y. Times , June 5, 2003, at A1.
43. See, e.g. , M. Ishaq Nadiri, Innovations and Technological Spillovers (Nat'l Bureau of Econ. Research, Working Paper No. 4423, 1993); Timothy F. Bresnahan, Measuring the Spillovers from Technical Advance: Mainframe Computers in Financial Services , 76 AM. Econ. Rev. 742, 753 (1986).
44. For an article that takes into account the resource reallocation effects of copyright, see Glynn S. Lunney, Jr., Reexamining Copyright's Incentives-Access Paradigm , 49 Vand. L. Rev. 483 (1996).
45. See, e.g. , Steven C. Salop, Monopolistic Competition with Outside Goods , 10 Bell J. Econ. 141 (1979) (offering a model of a differentiated product market that results in excessive entry).
46. See generally Robert C. Post, Subsidized Speech , 106 Yale L.J. 151 (1996).
47. See, e.g. , National Endowment for the Arts v. Finley, 524 U.S. 569 (1998) (considering the constitutionality of governmental funding of the arts).
48. See generally Robert P. Merges, Contracting into Liability Rules: Intellectual Property and Collective Rights Organizations , 84 Cal. L. Rev. 1293 (1996).
49. This should be a far easier task than checking whether individual authors sought to manipulate the proxies for their works, although it is not necessarily invulnerable to fraud. Individual authors might still seek to manipulate the proxies, even though they would have no direct financial incentive to do so, but this will matter only to the extent that there are differential levels of manipulation across genres.
50. Liebowitz illustrates the possibility of such distortions by providing a history of mediator-based rates. Liebowitz, supra note 9 (manuscript at 16).