Should Copyright of Academic Works be Abolished?

Steven Shavell

doi:10.1093/jla/2.1.301

Should Copyright of Academic Works be Abolished?

Shavell, Steven 2010-01-01 00:00:00 SHOULD COPYRIGHT OF ACADEMIC WORKS BE ABOLISHED? Steven Shavell ABSTRACT T he conventional rationale for copyright of written works, that copyright is needed to foster their creation, is seemingly of limited applicability to the academic do- main. For in a world without copyright of academic writing, academics would still beneﬁ t from publishing in the major way that they do now, namely, from gaining scholarly esteem. Yet publishers would presumably have to impose fees on au- thors because publishers would no longer be able to proﬁ t from reader charges. If these author publication fees would actually be borne by academics, their incen- tives to publish would be reduced. But if the publication fees would usually be paid by universities or grantors, the motive of academics to publish would be unlikely to decrease (and could actually increase)—suggesting that ending academic copy- right would be socially desirable in view of the broad beneﬁ ts of a copyright-free world. If so, the demise of academic copyright should probably be achieved by a change in law, for the “open access” movement that effectively seeks this objective without modiﬁ cation of the law faces fundamental difﬁ culties. 1. INTRODUCTION Or iginal written works, and thus original written academic works, may 1 generally be copyrighted. However, the conventional rationale for copyright 1 Sam uel R. Rosenthal Professor of Law and Economics, Harvard Law School. I thank Yochai Benkler, Michael Fisher, Paul Goldstein, Stevan Harnad, Scott Hemphill, Louis Kaplow, Mark Lemley, Josh Lerner, Douglas Lichtman, Peter Menell, Proshanto Mukherji, Dotan Oliar, John Palfrey, Richard Posner, Stuart Shieber, Christopher Snyder, Peter Suber, and Jeannie Suk for comments, Jonathan Cooper, Steven Horowitz, James Kwok, Vincent Leow, and Stephanie Wiener for research assistance, and the John M. Olin Center for Law, Economics, and Business at Harvard Law School for research support. 2 See 17 U .S.C. § 102(a); Melville B. Nimmer & David Nimmer (2008, vol. 1 §§ 2.01–2.04). Moreover, academics themselves, rather than their universities, own the copyrights in their works. See Hays v. Sony Corp. of Am., 847 F.2d 412, 416 (7th Cir. 1988) (Posner, J.), noting that “virtually no one questioned that the academic author was entitled to copyright his writings” despite the work for hire doctrine, under which the academic’s employer would be entitled to copyright; Weinstein v. Univ. of Ill., 811 F.2d 1091, 1094 (7th Cir. 1987) (Easterbrook, J.), not- ing that “faculty own the copyrights in their academic work” regardless of the work for hire doctrine; William M. Landes & Richard A. Posner (2003, 272), “It is generally and we think correctly understood that academics, although employees of the university in which they teach and conduct their research, own the copyrights on their academic books and articles.” 302 ~ Shavell: Should Copyright Of Academic Works Be Abolished? of written works—that it stimulates their creation and publication by allow- ing authors to proﬁ t from their sale —is seeming ly of limited applicability to academic authors. The primary spur for academic writing ostensibly is not proﬁ t but rather scholarly esteem and professional advancement. Hence, one might be led to believe that copyright is not needed to encour- age academic writing. This view, though, is incomplete, because it does not reﬂ ect the point that if publishers were unable to beneﬁ t from copyright earnings, they would apparently have to impose fees on authors to cover their publishing costs. The possible effects of such publication fees must be evaluated along with other relevant factors to determine whether the beneﬁ ts of a copyright-free academic world make it socially preferable to our own. 2 I att empt to make this assessment here, beginning in Section 2 with an informal consideration of a stylized model of academic authorship and publication (the model is presented formally in an appendix), the major assumption of which is that academic authors desire readership of their works to gain esteem. I interpret the model in Section 3, taking into account a variety of issues that are not included in it. The analysis and discussion in these sections may be summarized as follows. 3 First, publication of academic works does indeed carry signiﬁ cant beneﬁ ts of recognition and of professional reward for academics that are independent of the existence of copyright; and copyright itself provides academics with little remuneration—academics generally do not earn any royalties from articles and only modest royalties from books. Hence, the afﬁ rmative motives of academics to publish would be little diminished in the absence of copyright. 4 Second, publishers would have to recover at least the ﬁ rst copy costs of selecting and preparing works for publication from a source other than 3 See U.S. Const. art. I, § 8, cl. 8, “The Congress shall have Power . . . To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries”; Eldred v. Ashcroft, 537 U.S. 186, 212 n.18 (2003), “copyright law celebrates the pr oﬁ t motive, recognizing that the incentive to proﬁ t from the exploitation of copyrights will redound to the public beneﬁ t by resulting in the prolifera- tion of knowledge . . . . The proﬁ t motive is the engine that ensures the progress of science.” (quoting Am. Geophysical Union v. Texaco, Inc., 802 F. Supp. 1, 27 (S.D.N.Y. 1992)); Mazer v. Stein, 347 U.S. 201, 219 (1954), “The economic philosophy behind the clause empowering Congress to grant patents and copyrights is the conviction that encouragement of individual effort by personal gain is the best way to advance welfare through the talents of authors and inventors in ‘Science and the useful Arts.’” See also P aul Goldstein (2005, vol. 1 §§ 1:41–:44); Peter S. Menell & Suzanne Scotchmer (2007, 1476–1482). Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 303 readers were there no copyright. Thus, as indicated, publishers would have to impose fees for publishing. If the fees incorporated ﬁ rst copy costs, they might lie in the range of one to several thousand dollars for an article and would be signiﬁ cantly higher for a book. Thir d, if academics would have to bear the publication fees themselves, 5 their incentives to publish would fall, possibly non-negligibly in view of the potential magnitude of the fees. Still, the social losses associated with such a decline in publications would be tempered by two considerations: discouraged publications would tend to be of lower than average quality; and written but unpublished works could still be posted on the Internet, so presumably would be available to readers in that form. Fourth, if academics would not have to bear publication fees—on the 6 assumption that universities or grantors would subsidize them—the incen- tive of academics to write and to publish articles would not fall and could rise. The reason that the elimination of copyright could, perhaps paradoxi- cally, augment incentives to publish articles is that readership of articles would grow in the absence of copyright, and thus the esteem that authors would derive from publication would tend to increase. The incentive of academics to write and to publish books, however, would not necessarily be promoted, because authors would forgo royalties. Yet since royalties are generally low, it is possible that incentives to write and publish books would also be enhanced. An enhancement of incentives to write due to the subsidy of publication fees is not necessarily socially desirable, though, because it could lead to the writing of works that are not socially valuable on net in view of the costs of publication—the subsidy could lead to a socially ex- cessive level of publication. This problem, however, could be addressed by conditioning subsidies on the quality of publication venues. Fifth, universities and grantors would have a motive to subsidize publi- 7 cation fees in the absence of copyright, for it would be in their interest to prevent a dilution in the incentives of faculty members and of grantees to write and publish works, at least works of reasonable quality. Furthermore, universities should not face great difﬁ culties in ﬁ nancing publication fees, for they would no longer have to purchase journal subscriptions or books. Finally, substantial social beneﬁ ts would ﬂ ow from a copyright-free 8 world. Unfettered access to academic works would mean that all works would become available on the Internet for any person to download for free, that print versions could be made at cost, and that the assembly of teaching materials would no longer be burdened by the task of securing 304 ~ Shavell: Should Copyright Of Academic Works Be Abolished? permissions from copyright holders. Further, the various legal and other costs of protecting copyright privileges would be avoided. 9 The conclusion that I draw from the foregoing is that if publication fees would be largely defrayed by universities and grantors, as I suggest would be to their advantage, then the elimination of copyright of academic works would be likely to be socially desirable: it would not compromise academic publication activity and would yield the social beneﬁ ts of a copyright-free regime. On the other hand, if publication fees would not be broadly subsi- dized, whether ending academic copyright would be socially advantageous is unclear. 10 I n Section 4, I examine the “open access” movement, which holds that academic publications should be costlessly obtainable from the Internet. Although open access initiatives have grown in recent years, they have also faced difﬁ culties. Namely, traditional journals are generally unwilling to grant immediate open access to published works, and most authors w ould not ﬁ nd it in their personal interest to publish in open access journals that sometimes charge publication fees rather than in the often more prestigious traditional journals that do not charge publication fees. Moreover, there is a fundamental divergence between individual and collective interests af- ﬂ icting the open access movement: if an individual academic publishes his or her own work as an open access article or book, the academic does not thereby obtain the broad social beneﬁ ts that would ﬂ ow from the general disappearance of academic copyright. This individual-versus-social incen- tive problem suggests that open access efforts to cabin copyright-associated charges for access will not succeed fully, or will do so only after a substan- tial delay. An outright change in law to terminate academic copyright is therefore likely the best policy if academic copyright is judged to be socially undesirable. 4 See, e.g., Berlin Declaration on Open Access to Knowledge in the Sciences and Humanities (2003); Bethesda Statement on Open Access Publishing (2003); Budapest Open Access Ini- tiative (2002), deﬁ ning open access as “the world-wide electronic distribution of the peer- reviewed journal literature and completely free and unrestricted access to it by all scientists, scholars, teachers, students, and other curious minds.” See generally John Willinsky (2005); Symposium (2006); Nicholas Bramble (2006); Nature, Web Focus (2004); Peter Suber (2007). These sources on open access initiatives emphasize journal literature, not academic books. 5 However, as I discuss in Section 4.1, traditional journals that hold copyright in their ar- ticles often do allow so-called “green” open access, that is, immediate free access to various pre-publication versions of articles and sometimes delayed free access to published versions. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 305 I n Section 5, I discuss how the elimination of academic copyright could be implemented and whether there would be serious legal or political dif- ﬁ culties in so doing. Befor e proceeding, let me note that I do not seek any striking novelty in 12 this article. That academics are motivated to write to gain scholarly respect, and therefore that copyright may not be required to encourage publication of academic works, has often been mentioned before. The contribution made here lies mainly in the development of this observation employ- ing a model in which explicit account is taken of the desire of academ- ics for esteem and of the publication fees that would likely emerge absent copyright; in the analysis suggesting that if universities or grantors would bear the publication fees, incentives to publish would not fall and might rise, and thus that abolishing copyright might well be socially desirable; and in the argument that the open access movement confronts serious problems in achieving its end, probably requiring a change in law to assure that outcome. 6 B ut as I note there, the author would possess the right to require proper attribution of his or her work and possibly certain ‘moral rights’, such as the right to prevent problematic abridgment of a work. 7 See, e.g. , Princeton Univ. Press v. Mich. Document Servs., 99 F.3d 1381, 1410 (6th Cir. 1996) (en banc) (Ryan, J., dissenting), “More than one hundred [academic] authors declared on the record that they write for professional and personal reasons such as making a contribution to a particular discipline, providing an opportunity for colleagues to evaluate and critique the authors’ ideas and theories, enhancing the authors’ professional reputations, and improving career opportunities. These declarants stated that the receipt of immediate monetary compen- sation such as a share of licensing fees is not their primary incentive to write.”; Yochai Benkler (2006, 47–48), describing information production in which authors produce a work for status or reputation beneﬁ ts; Carol Tenopir & Donald W. King (2000, 146–149), surveying scholar- ship on scholarly publication motivation and concluding that “[t]he two primary motivating factors of scholarly authors are (1) recognition for career advancement . . . , and (2) the desire to contribute to the body of knowledge in a ﬁ eld or to the archive of the scholarly knowledge in a ﬁ eld and to be recognized for their contribution by their peers”; Jessica Litman (2006, 782), commenting that authors are not motivated “by the incentives supplied by the copyright sys- tem”; Arnold Plant (1934, 169), observing that many authors, especially academics, write “to secure publicity” rather than to reap monetary reward, and thus that copyright is not needed to induce them to create new works; Uma Suthersanen (2003, 602), stating that one of the pri- mary purposes of writing journal articles is “achieving recognition or promotion”; Budapest Open Access Initiative (2002), referring to “the willingness of scientists and scholars to publish the fruits of their research in scholarly journals without payment.” 8 Ther e are, however, models of academic publication recognizing authors’ desire for readership, but assuming a regime of copyright (rather than comparing that regime to a regime without copyright). See, e.g., Mark J. McCabe & Christopher M. Snyder (2007). 306 ~ Shavell: Should Copyright Of Academic Works Be Abolished? 2. COP YRIGHT OF AC ADEMIC WORK S IN A S TYLIZED MODEL 13 I consider here the functioning of academic copyright and the possible so- cial desirability of its elimination in a stylized model of academic author- ship and publication. Because the purpose of the model is to elucidate the central effects of copyright on academic writing, I abstract from a variety of factors that will be of signiﬁ cance when I attempt to come to a realistic assessment of academic copyright in the next section. That section will be better appreciated by the reader in light of the analysis of the model, but he or she can still pass over this section without great loss of continuity. The crucial assumption of the model is that the utility of academic au- thors is increased by the readership of their published works because they gain esteem from readership. For simplicity, I suppose that an author’s util- ity from readership depends only on the number of readers (and thus, not also on the status of the readers or on the reputation of a publisher or a journal). I let a parameter t denot e the degree to which an author’s util- ity is raised by readership of a work. If t is small, the author cares slightly about readership, whereas if t is high, the author cares substantially about readership. I suppose t oo that an author’s utility depends on the conventional factors of income and expense and of writing effort. I assume that an author m ust devote positive effort and incur disutility to write a work and that its publication involves two types of cost for pub- lishers: an initial ﬁ xed, or “ﬁ rst copy” cost f, for reviewing, editing, proof- reading, and the like; and a per unit or marginal cost c of dissemination, such as the per article cost of printing a copy of a journal and of mailing it. This per unit c ost could be almost zero if the work is made available for downloading on an Internet site, an important case that I will discuss in the next section. 9 If the author obtained a beneﬁ t from writing rather than incurred disutility from doing so, the qualitative nature of the conclusions I would reach would not change. The reason is that the author who enjoys writing itself might still be discouraged from publishing in the absence of copyright by having to pay a fee in order to publish. 10 See Theodor e C. Bergstrom (2001, 187), “The costs of publishing a journal can be usefully partitioned into ﬁ rst copy costs and marginal subscriber costs .”; Donald W. King (2007, 94–96), discussing concepts of cost in journal publishing. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 307 The deﬁ nition of social welfare that I consider is a version of the standard utilitarian one, namely, the beneﬁ ts obtained from the publication of writ- ten works—both the beneﬁ ts obtained by readers and the utility obtained by authors because of the utility they derive from their readership – minus the costs of writing effort and of publication. 2.1. Copyright Let me now discuss the nature of the contract that an academic who has writ- 18 ten a work would rationally make with a publisher in a regime with copyright. (I will discuss later whether an academic would be motivated to write the work in the ﬁ rst place.) I will suppose that the contract involves two terms: a price at which the author’s work will be sold to readers in the marketplace; and a pay- ment term—either a royalty payment made by the publisher to the author, or possibly a publication fee paid by the author to the publisher. In some respects, the reader might ﬁ nd it convenient to suppose that the author’s work is a book rather than a journal article, for an author of a book would be able to bargain about the price at which it would sell, but the author of an article would not be thought to be able to bargain individually about the price at which a journal sells. N evertheless, the average desires of authors about journal prices should be reﬂ ected in journal prices, so that the model can be interpreted as applying to authors’ contracts with journals in an approximate sense. I n addition, I assume for simplicity that there are many competing pub- 19 lishers, implying that publishers will earn only enough to cover their costs. In other words, any publisher revenue from sales of a work in excess of pub- lisher costs will be paid to the author as a royalty, or any shortfall of sales from publisher costs will be paid by the author to the publisher as a fee. The author w ill choose a contract that maximizes his or her utility. That 20 is, among the possible contracts that allow publishers to cover their costs, the author will select the particular contract that makes him or her best off. The nature of this contract will depend on the degree t t o which the author cares about readership, as is illustrated in Figure 1 below. 11 F or a model in which journals rather than books are the venues for publication, see McCabe & Snyder (2007). It would be distracting from my central purpose—the comparison of a regime of copyright to a regime of no copyright—to have to take account of the complexities of such a model, rather than to address them informally, as I do, in the next sections. 12 This assumption does not affect the qualitative conclusions that I reach. In fact, many journals and publishers of books are of course proﬁ table, as I note in Section 5.2. To the degree that proﬁ ts are positive under copyright and thus that prices are higher than I assume in the model, the advantage of abolition of copyright would be enhanced. 308 ~ Shavell: Should Copyright Of Academic Works Be Abolished? Figure 1. Contract Prices Monopoly price p ($50) Price under copyright Price in the absence of copyright Per unit cost c ($2) t t t 1 2 3 Author’s degree of care t about readership T o explain, consider ﬁ rst an author for whom t is 0, in other words, the author is motivated to publish entirely by royalty payments and does not care about the level of readership per se. Such an author would choose the sales price that maximizes proﬁ t, the traditional monopoly price of eco- nomics. This price is labeled p in the graph, and for illustration, let us assume that it is $50. If t is positive, so that the author does care about read- ership, the author will generally want to increase readership from the level that would result if the monopoly price were charged. Thus, the author will want to set the price charged below the monopoly price even though this will mean some sacriﬁ ce of proﬁ ts and royalty payments. For example, sup- pose that by reducing the price from $50 to $49, readership would increase by 100 individuals and proﬁ ts and hence royalty payments would fall by $200. If the utility value to the author of having an additional 100 readers exceeds $200, the author would prefer the lower $49 price to the $50 mo- nopoly price despite the reduction in royalty payments. The best price for the academic author who cares about readership can be shown always to be less than the monopoly price, and is lo wer the greater is the degree t t o which 13 The monopoly pr ice is such that the marginal revenue equals marginal cost. See , e.g., Robert S. Pindyck & Daniel L. Rubinfeld (2009, 330). 14 Because the monopoly price maximizes proﬁ ts, a slight variation in the price would leave prof- its essentially unchanged. Hence, the effect of a small decrease in the price would be to increase readership with only a negligible lowering of proﬁ ts; and since the author values readership, the author would beneﬁ t thereby. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 309 the author cares about readership. This is why the graph of the price of the published work that the academic author would choose declines as t rises. The degree of care about readership t displayed in Figure 1 is the t at 22 which the price chosen by the author would yield revenues just equal to the publisher’s total costs, implying that the author would obtain no royalty payment from the publisher. Be yond t , the price favored by the author would not yield revenues sufﬁ cient to cover the publisher’s costs and would result in the author having to make a payment to the publisher. That is, if the degree to which an author cares about readership is at least t , the author will value readership enough that he or she will pay a publication fee to aug- ment readership by lowering the sales price substantially, even though the author could contract for a higher sales price and not have to pay to publish. At t authors w ould pay a publication fee of f in or der to lower the price to c . Indeed, if authors care enough about readership, if their t is at least t , they would give their work away for free—possibly by mounting it on the Internet and letting it be downloaded gratis—and have to pay a fee equal to the entire publisher costs. That corresponds to a price of zero on the graph. I n interpreting the graph, the reader might note that it is common 23 for an academic author of a book to negotiate with the publisher for the selling price to be lower than the publisher would like, because, as is con- sistent with the analysis, the author often wants larger readership rather than higher income from royalties. Also reﬂ ecting the graph and the point that authors value readership as well as not having to make pay- ments is that authors of journal articles generally do not collect royalties 15 Note that the graph shows that the price at t exceeds the unit cost c even though no proﬁ ts are made. The explanation is that the apparent proﬁ t of t – c on the units that are sold just offset the ﬁ rst copy costs f . 16 I do not consider negative prices—the possibility that authors would pay individuals to read their works. 17 See Am. Geophysical Union v. Texaco Inc., 802 F. Supp. 1, 27 (S.D.N.Y. 1992) (Leval, J.), “gener- ally [academic] authors have a far greater interest in the wide dissemination of their works than in royalties,” aff ’d , 60 F.3d 913 (2d Cir. 1994); Ralph R. Shaw (1951, 572), “The commer- cial publisher’s interest in publications is the sale of copies of the entire issue or volume for proﬁ t. If he does not sell copies at a proﬁ t, he will soon be a bankrupt ex-publisher. The author, however, may be interested in the widest possible dissemination of his writings, and if some- one were willing to reprint 10,000 copies of his article for free distribution, that would provide a great additional proﬁ t to the author in terms of professional credit.” I myself have bargained with a university press about the price of a book that I authored. I was urging a lower price for the book (and earlier release of an inexpensive paperback version) than the press was inclined to accept because of its interest in revenues and proﬁ ts. 310 ~ Shavell: Should Copyright Of Academic Works Be Abolished? Figure 2. Utility from Contracts Utility in absence of copyright, given subsidy Utility under copyright Utility in absence of copyright, no subsidy Utility AB C D 0 t t 0 2 Author’s degree of care t about readership and that journals retain copyright and charge prices to readers that allow them revenues sufﬁ cient not to have to impose fees for publishing on authors. 24 The utilit y that an author would obtain from publishing employing his or her preferred contract is displayed by the solid graph in Figure 2. (The other graphs will be discussed subsequently.) The author’s utility increases the higher is his or her t, even though the sales price charged falls, and even though, beyond t , authors pay a positive fee to have their works published. The argument explaining that the utility of authors must grow with t is essentially this. Compare an author with one level of t , say t , to an author with a higher level of t , say t . The t author could always make precisely H H the contract that the lower t author makes – name the same price and col- lect the same royalty (or pay the same fee). If the t author did this, the t H H author would be happier than the t author because, by hypothesis, the t L H author obtains more utility from having readers than the t author. In fact, the t author w ill obtain even greater utility than the amount just men- tioned, since the t author will prefer to choose a different contract with a lower sales price from that chosen by the t author. Hence, the t author L H must be better off, as claimed. 18 See the discussion in Section 3.1 below. 19 These levels of t are not shown in Figure 2. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 311 The solid graph in Figure 2 determines whether an author would want to invest the effort to create a work in the ﬁ rst place under copyright. In particular, an author would write a work if and only if the utility gained from his or her preferred contract would exceed the effort cost. In general, I assume that the effort costs vary for authors of each t , so that some authors of each type would choose to write and some would not. 2.2. No Copyright N ext let me consider the situation in a regime without copyright. In this 26 context, a contract between an author of a work and a publisher cannot specify a sales price exceeding the per unit cost c . The reason, of course, is competition: if the publisher named a higher price than c , it would sell nothing, as other publishers could copy the work and could sell at a lower price as their unit cost would be only c . Because, then, the publisher con- tracting with the author could only obtain a price of c for the work, the publisher would need to charge a publication fee to recoup the ﬁ rst copy cost f in order to cover its costs and to be willing to publish. Publication fees are borne by authors . Assume here that authors them- 27 selves would bear publication fees. Thus, if an author made a contract and the sales price were c , the author would have to pay a fee equal to the ﬁ rst copy cost f . Likewise, if the author made a contract and named a lower price than c, the author would have to pay more than f as a publication fee in order to cover the publisher’s costs. The sales pr ice that an author of a work would choose if he or she made 28 a contract can be seen from Figure 1. An author who would have chosen a price higher than c g iven copyright cannot name such a price in the ab- sence of copyright. The best feasible price for such an author to elect is c . Therefore, all authors with a t less than or equal to t would choose c and have to pay the ﬁ rst copy cost f in order to publish. Authors with higher t would have chosen a price below c and can still do so in the absence of copyright. Hence, the price chosen by authors in their preferred contracts is given by the dotted graph in Figure 1. 20 I assume that in the absence of copyright, the ﬁ rst publisher could not prevent copying by technological means, such as digital rights management. See Section 5.1. 21 This presumes that the per unit cost c is positive, that is, that publishing does not consist of mounting the work on the Internet. 312 ~ Shavell: Should Copyright Of Academic Works Be Abolished? Relatedly, we can see that the utility of authors from their contracts is given by the dotted graph in Figure 2. Note that authors with t belo w t obtain less utility because in the absence of copyright they are unable to make their otherwise-preferred contracts involving prices exceeding c . The utility loss suffered by such authors is greatest for authors with t of 0, for these authors would have named the highest price (the monopoly price) and would have earned the highest royalty, and the utility loss declines as t increases. For t at least t , authors suffer no utility loss, for the absence of copyright does not affect the contract that these authors would make, as the price would be c or belo w. 30 The height of the dotted graph tells us whether an author would choose to write in the absence of copyright; if the height of the graph exceeds the effort cost, the author would write and publish; otherwise the author would not write. 31 Publication f ees are subsidized . Now suppose that authors would not bear publication fees, because universities or grantors would do so (for rea- sons to be discussed in Section 3). In particular, assume that universities and grantors would pay a publication fee of f , so that authors could pub- lish for free. It can be shown that if an author wrote a work and made a contract, the author would choose the same contract price as he or she would in the absence of a subsidy of f . Th us the dotted graph in Figure 1 that shows a contract price of c for authors with t below t and then lower prices for higher t applies whether or not authors would bear publication fees. 32 H owever, the utility of authors is higher as a result of the subsidy, as is shown in the dashed graph in Figure 2. Speciﬁ cally, the utility of an author is higher by the amount f at eac h t than if the subsidy would not be paid in the absence of copyright; this is why the dashed graph is everywhere above the dotted graph by the distance f . One can also see geometrically that the dashed and solid graphs must intersect at some point, which I label t . Accordingly, to the left of t the inc entive to write is lower than under copyright and to the right of t the incentive to write is higher than under copyright. The sense behind this conclusion is as follows. For relatively low t , authors have to give up relatively high royalty revenues; and the greater 22 I assume that universities and grantors would not pay a higher fee, which would for instance allow a book costing c to be printed to be sold below cost. 23 This is established in the appendix. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 313 readership, which they value relatively little, combined with the subsidy they obtain in the absence of copyright does not offset the royalties they cede. For relatively high t, converse logic implies that authors are better off giving up royalty revenues in exchange for higher readership and the subsidy. A gain, the height of the graph of utility determines whether an author 33 would choose to write; if the height of the dashed graph exceeds the effort cost, the author would write and publish. 2.3. The Effect and the Social Desirability of Eliminating Copyright From what has been said, we can proceed to identify the effects and the 34 possible social desirability of elimination of copyright. I separately con- sider the case where publication fees would be paid by authors and the case where the fees would be subsidized. Publication f ees are borne by authors . In this case, we have seen from Figure 2 35 that eliminating copyright will reduce the incentive to write for authors in Region A, namely, those authors who would have selected a price above c under copyright. Some of these authors with reduced incentives in Region A will decide 36 not to write. Such authors will tend to have relatively low t, for they are the ones who would have earned relatively high royalties and thus who lose the most in the absence of copyright. When these authors are discouraged from writing, social welfare will fall. The measure of the social welfare loss suffered for a work that is not written is the value that would have been ob- tained by its author and its readers under copyright less the costs of writing and publishing it. The remaining authors in Region A will still elect to write, because 37 the decline in their utility from publishing will leave them with beneﬁ ts exceeding the cost of writing. These authors will likely be those with rela- tively high t , because they are the ones who would have earned the lowest royalties and thus give up the least in the absence of copyright. Social wel- fare will increase in regard to each of these authors. The reason is that these authors would have published at a price above c , whereas in the absence of copyright they will publish at a price of c , resulting in a beneﬁ cial rise in readership. To illustrate, suppose that the price given copyright would be $25 and that the price would fall to the per unit cost of $2 without copyright. Then individuals who place a value on reading the work that is in between $2 and $25 will purchase the work, whereas they would not 314 ~ Shavell: Should Copyright Of Academic Works Be Abolished? have at the price of $25, and their purchase will augment social welfare. For example, an individual who is willing to pay $20 for the work would purchase it at a price of $2, and this would raise social welfare by $18 (the $20 value placed on it by the person minus the $2 unit cost of providing it to the person). 38 I t is also clear from Figure 2 that authors in Region B, with t at least t , have incentives to write that are unaffected by elimination of copy- right. Hence, for them, the elimination of copyright has no social welfare consequence. 39 Whether elimination of copyright is socially desirable can now be expressed as follows. To the extent that academic authors are in Region A and would be discouraged from writing by the elimination of copy- right, that policy would be undesirable. But to the extent that authors are in region A and would not be discouraged from writing, the policy would be desirable. The numbers of authors in both groups and the sig- niﬁ cance of the associated social welfare changes must be evaluated to determine whether the elimination of copyright would on net be socially desirable. 40 Publication f ees are subsidized . In this case, we know from Figure 2 that in Region C, with t less than t , eliminating copyright reduces the incentive to write. 41 N ow in Region C, we can reason as we did in regard to Region A in the case just considered. That is, some of the authors in Region C, those with relatively low t, will be discouraged from writing by their loss of royalties and the elimination of copyright, causing social welfare to fall. The other authors in Region C will still write, and since the price of their works will fall to c , social welfare will rise. 42 I n Regions D and E, however, the incentive to write will increase relative to under copyright (because the dashed graph is above the solid graph), leading to a number of effects on social welfare. First, all authors in Region D who would have written given copyright will continue to write but will publish at a lower price c , and this will raise social welfare as has been ex- plained. Second, in both Regions D and E, there will be authors who will 24 A ctually, this statement is inexact. Social welfare would rise by more than $18 because, if the person obtained the work, the author would also gain utility from having an additional reader. For expositional ease, I abstract from this factor in discussing social welfare in the text, but I take it into account in the appendix. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 315 write only because copyright has been eliminated and publication fees subsidized. That these new writings will occur may either raise or lower so- cial welfare. In particular, the possibility that the new writings would lower social welfare arises because the subsidy of publication fees means that an author contemplating writing does not take into account the actual cost of publishing. W hether eliminating copyright in this case is socially desirable depends 43 on three groups of authors: those in Region C who would be discouraged from writing by the elimination of copyright and in respect to which social welfare would be lowered; those in Regions C and D who would write both in the presence and the absence of copyright, and in respect to which social welfare would rise (because the price would fall to c); and those in D and E who would write only in the absence of copyright, in respect to which social welfare could either rise or fall. 3. DISCUS SION OF THE MODEL— AND THE POS SIBLE CASE F OR ABOLITION OF AC ADEMIC COP YRIGHT In this section, I restate and interpret the analysis of the stylized model. 44 A cademic authors would still have a strong afﬁ rmative motivation to publish 45 in the absence of copyright—to gain scholarly esteem and to advance themselves professionally. A major assumption in the model was that academics have a positive incentive to publish regardless of the existence of copyright, es- pecially to gain scholarly esteem. That this reason for publishing is impor- tant is manifest to any person familiar with the academic environment. Academics plan and politic to present their works, they seek to publish their works in as highly ranked venues as possible, and they attempt in other ways as well to promote citations to their works; for many academ- ics, these efforts to achieve notice for one’s publications are ardent and consuming. I assume, therefore, that academics derive substantial value from publication of their works because of the respect and attention 25 For instance, suppose the following: the author’s disutility of writing would be 10, the author’s utility from readership would be 12, the publication fee (and ﬁ rst copy cost) would be 20, and the utility readers themselves would obtain from publication is 15. Then the author would write given that the fee is subsidized, for 12 exceeds 10. Yet writing is not socially desirable be- cause the social costs of writing are 10 + 20 = 30, exceeding the social beneﬁ ts of 12 + 15 = 27. 316 ~ Shavell: Should Copyright Of Academic Works Be Abolished? given to them by their readers, and that the y would beneﬁ t in this way in a world without copyright. 46 A nother source of utility to academics from publication that does not depend on the existence of academic copyright is that the promotion and salary of academics, and possibly their textbook-writing and consulting oppor tunities, are a function of their publication records. P ublication would bring these monetary beneﬁ ts to academics even if academic copyright were eliminated. 47 Consistent with the view that academics obtain signiﬁ cant value from publication that does not depend on the existence of copyright is the obser- vation that, in actuality, academics earn little royalty income from publica- tion. Publication of articles rarely results in receipt of royalties. Publication of academic books usually does result in some royalty payments, but this income tends to be minor. That academics in fact gener ally obtain only 26 See sour ces cited in note 7 above; Ian Rowlands & Dave Nicholas (2005, 17 & ﬁ g.7), ﬁ nd- ing that the most important factors to authors in choosing a journal for publication are its reputation and readership. 27 Academics would continue to beneﬁ t presuming that they would be cited as the authors of their works. In a world without copyright, I assume that there would still be a requirement to attribute copied material to its author. See Section 5.1 below. 28 See, e.g. , Weinstein v. Univ. of Ill., 811 F .2d 1091, 1094 (7th Cir. 1987) (Easterbrook, J.), noting that a university’s “demands—especially the demands of departments deciding whether to award tenure—will be ‘the motivating factor in the preparation of ’ many a scholarly work”; Landes & Posner (2003, 48), “Many authors derive substantial beneﬁ ts from publication that are over and beyond any royalties. This is true not only in terms of prestige, celebrity, and other forms of non- pecuniary income, but also in terms of pecuniary income in such forms as a higher salary for a professor who publishes than for one who does not, or greater consulting income . . . .” (footnote and emphasis omitted); Mark J. McCabe & Christopher M. Snyder (2004), “Prestige is the curren- cy of academia. Aside from the pure enjoyment of prestige, a scholar’s chances for promotion, ten- ure, a higher salary, etc., increase with his or her prestige. Prestige comes from doing high- quality research.” On the importance of publishing for salary, see, e.g., On ur Baser & Elda Pema (2003, 6), “Publications, we conclude, have a positive and diminishing effect on salaries.”; and William J. Moore, Robert J. Newman, & Geoffrey K. Turnbull (1998, 360-361), concluding that quantity and quality of publications have strong positive effects on academic economists’ salaries. 29 T extbooks could still be copyrighted if academic copyright were eliminated, as I discuss in Section 5.1 below. 30 See, e.g., Stephen R. Barnett et al. (1998, ¶ 1(e)), “The Author grants the above rights without claim of royalties or other compensation.”; Beth Luey (2002, 3), “Journals do not pay authors . . . .” 31 F or example, my understanding is that under a typical arrangement, an academic author would receive a royalty rate of 15 percent of revenue. If a book sold, say, 1,000 copies at a price of $60, the royalties received would be $9,000. If the book were 300 pages in length and the author spent an average of 3 hours a page, the implied hourly rate of pay received by the author would be $10. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 317 insubstantial or no royalty income from publishing yet they devote great effort to writing implies that the incentives to publish that are unrelated to copyright must be strong. Publication f ees, however, would probably be charged by publishers in the 48 absence of academic copyright, and the fees would be more than nominal. As emphasiz ed in the analysis of the model, publishers would be unable to proﬁ t from reader charges if academic copyright were eliminated. A pub- lisher of a book or journal would not be able to impose fees on readers for Internet access to its content, since all academic works would be made freely available by other parties. Likewise, a publisher would not be able to charge readers for print versions of works amounts exceeding printing and distri- bution costs, due to competition from other sellers of print versions. H ence, an initial publisher of a work would need to recover the ﬁ rst copy 49 costs it incurred—the expenses of arranging peer review, selecting, edit- ing, and formatting the work—from a source other than reader charges. Consequently, the publisher would have to impose fees on authors for publishing (assuming that the publisher’s operations would not be directly subsidized by universities or other donors). W hat would be the likely magnitude of the ﬁ rst copy costs and publish- 50 ing fees? First copy costs vary widely, from about $400 to $8,000 per article, with most falling in the $1,000 to $4,000 interval (King 2007). We can also gain some appreciation of the magnitude of ﬁ rst copy costs from the publication fees now charged by open access journals. As of April, 2008, fees charged to authors by major open access journals ranged from $1,250 32 I n other words, I am assuming that ﬁ rst publishers in today’s world would not be able to earn meaningful proﬁ ts merely from being ﬁ rst, given the ease of copying and the existence of the Internet. I am assuming also, as I noted, that initial publishers would not be able to purposely impede copying through technological means. See Section 5.1 belo w. In the past, proﬁ ts suf- ﬁ cient to allow publishers to succeed may have been possible in the absence of copyright. See Stephen Breyer (1970). 33 See also Wellcome Trust (2004, 11), discussing certain literature and commenting on meth- ods of ﬁ nancing scientiﬁ c publishing; Bernard Wysocki, Jr. (2005, A1), “Industry experts say typical per-article costs are between $3,000 and $4,000.” In terms of ﬁ rst copy costs per page, compare Bergstrom (2001, 187), noting that for economics journals, “ﬁ rst copy costs aver- age about $100 per page” and “marginal subscriber costs are about $.02 per subscriber per page,” with Joop Dirkmaat (2002, 229), criticizing Bergstrom’s methodology and estimating ﬁ rst copy costs to be at least $281 per page. For major international journals, these costs will be much higher than average owing to increased content in each issue to edit, more submissions to sift through, and greater expenditures on peer review for each submission. See Wellcome Trust (2004, 14); see, e.g., Wysocki (2005, A1), noting that the journal Science has a per ar ticle ﬁ rst-copy cost of approximately $10,000. 318 ~ Shavell: Should Copyright Of Academic Works Be Abolished? to $3,000 per article (King & Alvarado-Albertorio 2008, 264 table 7). For instance, Public Library of Science journals charge from $1,250 to $2,750 per article, and BioMed Central charges from nothing to $2,685 per ar- ticle, with $1,690 being standard ( Id. ). These open access journal fees may, however, understate their ﬁ rst copy costs because they sometimes receive income from grantors; and if the journals engage in less reviewing of sub- missions than traditional journals, their costs and fees may also understate the ﬁ rst copy costs of many journals of relevance to us. 51 Being longer than articles, books involve greater ﬁ rst copy costs, so that publishing fees for books would be greater. A conservative estimate is that if fees covered ﬁ rst copy costs for books, they would be at least $10,000. 52 It should also be observed that ﬁ rst copy costs are likely to fall over time because of publisher use of the Internet and of computer software. 34 See Lila G uterman (2004, A10, A11), noting that the Public Library of Science open access journals have beneﬁ ted from a $9 million grant from a private foundation; Colin Steele (2008), mentioning grants from the MacArthur Foundation and Mellon Foundation to university presses to print open access monographs; Peter Suber (2009a, § 5), reporting grants from the Wellcome Trust and the Mellon Foundation to open access journals. I myself have been in- volved in the launching of a new open access publication, the Journal of Legal Analysis , which is supported by grants from a private donor and by Harvard Law School. 35 Some indi viduals have expressed skepticism to me that ﬁ rst copy costs could be as large as I have described in this paragraph, suggesting that ﬁ rst copy costs should be negligible, in part because readers do not really demand copy-editing. If these individuals are correct in their as- sessment, that only reinforces the overall case I make in this article in favor of elimination of copyright; obviously, the lower are the costs of writing, the lower would be the fees for publica- tion and the less they would impede the motive to write and publish. 36 Dr yburgh Assoc. (2002) reports from a survey that the total ﬁ rst copy cost of an academic book is £7,391 (54 percent of which is for copyediting and typesetting), which at an exchange rate of $1.49, http://www.x-rates.com, is $11,013. “The upfront costs for publishing a mono- graph are . . . from about $20,000 on the low end to many multiples of that . . . .” Sanford G. Thatcher (2009, 62). Harvard University Press suggested in a conversation that the average ﬁ rst-copy cost per page is about $50, implying a ﬁ rst copy cost of $15,000 for a book of 300 pages. Telephone Interview with personnel, Harvard University Press, in Cambridge, Mass. (May 28, 2009). Also, it is reported that the copy editing costs of a page of an article average $85, suggesting that copy editing costs of a book of 300 pages would be over $20,000 (Tenopir & King 2000, 258 tbl. 51). 37 One mig ht think that ﬁ rst copy costs are also likely to fall because authors would undertake copy editing tasks themselves. But if authors were to do this, they would then add to the publication fee their effort costs of copy editing. Hence, it is not obvious that the effective publication fee would fall as a consequence of authors doing copy editing themselves. Indeed, if a typical author’s time is more valuable than that of a copy editor or a typical author is less efﬁ cient in performing copy editing, the effective fee would rise if authors do copy edit- ing (and for that reason, one would predict that copy editing would continue to be done by journals or by hired professionals). Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 319 Journals now usually accept online submissions of manuscripts and com- municate with authors and referees electronically, saving mailing costs. Also, journal administration is becoming automatic in many respects; journals are increasingly employing sophisticated software to facilitate the referee- ing and review process. Software is also, of course, employed in the editing and formatting of manuscripts; it is used to check spelling, grammar, and references, to prepare diagrams, and to format manuscripts. Nevertheless, copy editing costs are likely to remain a signiﬁ cant component of ﬁ rst copy costs, because of the inherent value of copy editing. A related mitigating point about costs is that various ﬁ xed costs of pub- 53 lishing would be avoided if academic copyright were eliminated. There would obviously be no need for legal protection of copyright; thus pub- lishers would not have to ensure that authors properly granted them copyright, to negotiate licensing arrangements, or to police for copyright infringement. Also, there would be no need for subscription management and for charging readers, except to the extent that print versions were sold. Because these ﬁ xed expenses would not be incurred in the absence of aca- demic copyright, the costs and publishing fees would be lower than might otherwise be thought. If academics w ould have to bear publication fees in the absence of copyright, 54 their incentive to write and to publish would fall. Let us provisionally assume that the publication fees just discussed would be borne directly by academics. Then we know as a logical matter that the incentive to publish 38 On the efﬁ ciencies of online administration of journals and of editorial software, see John B. Thompson (2005, 112–113), noting that the cumulative effect of the Internet and and com- puter software “has been a dramatic decline in the cost of typesetting”, and Priscilla Markwood (2006), describing a number of novel software techniques for publication and noting how technology advances in recent years, especially software for copyediting, have allowed many tasks to become automated. 39 J anet H. Fisher of MIT Press states that “Some discussions … have asserted that authors can handle everything—they don’t need copy editing, proofreading, or formatting of their com- puter ﬁ les. The argument is that authors are doing most of this themselves now anyway. I strongly disagree. Most peer-reviewed journals are copy edited, and most scholars who have been through the process are grateful for the careful eye of a copy editor to catch incomplete references, mistakes in mathematical calculations, mislabeled ﬁ gures, misnumbered footnotes, and even the occasional convoluted sentence. The quality of a scholar’s article would decline— in some cases signiﬁ cantly!—if copy editing was removed from the process.” Fisher (1995, 89). If this judgment is incorrect, copy editing turns out to be largely unnecessary, and ﬁ rst copy costs become much lower than predicted in the text, then the conclusion to which I come— that the elimination of academic copyright would be socially desirable—would again only be reinforced. 320 ~ Shavell: Should Copyright Of Academic Works Be Abolished? and thus to write would fall for all authors who today do not pay publica- tion fees and who grant copyright to their publishers. These authors, who are typical, would be made worse off by having to pay publication fees, so that their motive to publish and write would be lower in the absence of copyright. 55 Would the imposition of the publication fees under discussion have a signiﬁ cant effect on writing and publication? Would having to pay, say, $2,000 to publish an article or $10,000 to publish a book substantially re- duce writing and publication of articles and books? This is an empirical question about which we have, as far as I know, little hard information. My own intuition is that publication fees would exert an effect that would be measurable, and I ﬁ nd it plausible that the effect would be signiﬁ cant, especially for the many academics whose salaries are modest. However, this conclusion is premised on the assumption that the academics themselves would have to pay the fees, an assumption that I now relax. 56 If, however, academics would not have to bear publication fees—because univ ersities or grantors would pay them—their incentive to write and to publish articles would not fall and could even rise, and so too might their incentive to write and publish books . Now let us suppose that universities or grantors 40 See Sanford G. Thatcher (2008, 98), “It has long been a tradition in scholarly publishing for au- thors to transfer all rights to the publishers of their articles and even their books.”; Bo-Christer Björk & Turid Hedlund (2009), estimating that open access journals comprise 8.5 percent of all scholarly peer-reviewed journals worldwide; Bo-Christer Björk, Annikki Roosr & Mari Lauri (2008, 184), calculating that in 2006, 4.6 percent of articles worldwide were freely avail- able immediately, another 3.5 percent would be available after a period of delay, and another 11.3 percent were available in repositories. 41 Authors today are in principle able to pay publication fees to offset ﬁ rst copy costs and not to give copyright to publishers, but most choose to give copyright to publishers. Hence, it must be that they would be worse off if they did not transfer copyright to publishers and paid pub- lication fees. In terms of the model, these are authors with t less than t . They are the authors in region A of Figure 2. To the extent that there are authors who today pay full publication fees and do not transfer copyright, the elimination of copyright would make no difference. In terms of the model, these are authors with t at least t . 42 Ther e does exist survey evidence suggesting that most authors would be unwilling to pay any, or at any rate, more than $500 in fees to publish an article. See Robert Denicola (2006, 358 & n.38), collecting surveys. This evidence, however, is misleading: It does not imply that imposi- tion of publication fees would in fact have a pronounced negative effect on publication in a world without copyright, for unlike authors in that world, the survey respondents enjoy the option of publishing for free in traditional journals. See Sara Schroter & Leanne Tite (2006, 143), noting that 53 percent of surveyed authors “said it would be pointless to pay to publish in ‘this’ journal when other journals might publish it for free.” Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 321 would pay the publication fees. Then the inc entive of authors to write and to publish articles could in principle increase in the absence of copyright. The reason is that, as is generally true today, authors would pay nothing to publish articles, but authors should beneﬁ t from greater readership, and thus from greater esteem, because their articles would be available for free on the Internet and printed copies of articles and of journals would sell at cost due to competitive pressures. What the magnitude of the change in incentives would be is not clear—it depends on how much the authors would value the increase in readership —but its direction is. The case of books is different from that of articles, because authors 57 of books earn royalties today. H ence, an author of a book would lose royalties even though he or she would gain readership in a regime with- out copyright and subsidy of publication fees. The incentive to write and publish books would therefore increase if the beneﬁ t from greater read- ership outweighs the loss in royalty revenue, but would decrease in the converse situation. Still, an increase in publication of books would not be a surprise, in light of the low royalties earned from academic books and the probably substantial increase in readership that would come about in a world without copyright. Universities and grantors would have a motive to subsidize publication fees 58 in a world without copyright. U niversities would have a reason to pay the fees that would otherwise be charged to their faculty members for publishing in a world without copyright. Namely, universities would not want their 43 In particular, assume as in the model that they would pay fees equal to ﬁ rst copy costs, but would not pay more, such as if an author of a book wanted the printed version to be sold below the marginal cost of printing. 44 I n terms of the model, authors of articles in journals are roughly described by Figures 1 and 2 at the point t , where the price charged is above c and wher e authors pay nothing and receive nothing. 45 I n the model, an author would deﬁ nitely experience greater readership in the absence of copy- right because his or her work would be freely available on the Internet (or in printed form at a price equal to cost). However, in fact the author might not gain readership because not only would the author’s work become freely available, so would all other works that had been copy- righted, meaning greater competition for readership. I abstracted from these complications in the model. 46 If most of the readers about whom an author cares—say academics who are actively engaged in related research—would already see the article, then the increase in readership would matter little to the author. Yet this is not always the case. 47 I n terms of the model, authors of books are roughly described by Figures 1 and 2 at points below t , where the price charged is above c and where authors receive positive payments. 1 322 ~ Shavell: Should Copyright Of Academic Works Be Abolished? faculty to write and publish less. A university’s reputation and thus its abilit y to attract faculty, students, gifts, and grants, are obviously impor- tantly affected by the publication record of its faculty members. Hence, it would presumably be in the interest of universities to subsidize publica- tion fees to offset an undue dilution in publication incentives that would otherwise result from faculty having to bear publication fees. A proviso, however, is that universities might want to exercise some control over the quality of publications for which they pay publication fees (how so, to be discussed below). 59 Another proviso is that universities would not want a policy of subsidies to result in an excessive increase in publication fees by publishers. If universities simply paid publication fees, whatever their magnitude, an escalation of such fees could be imagined to occur. This problem could be ameliorated by granting each faculty member a fund for academic expenses including publication fees, for then faculty would have a motive to search for lower fees. U niversities could also police fees for unreasonableness. 60 The view that universities would wish to subsidize publication fees is consistent with the fact that universities today sometimes pay publication fees charged to faculty by open access journals and that, of course, univer- sities spend generally on support of faculty research. 61 It should also be noted that, on average, universities would not be ex- pected to face a serious liquidity problem in ﬁ nancing publication fees. In a world without copyright, universities would not have to spend the con- siderable amount they now do on subscriptions for journals and on new books. Today, the annual cost of a single journal subscription can exceed $20,000 (Denicola 2006, 352 & n. 4), and the annual library budgets of 48 A sa ving that a faculty member achieved in publication fees could be used by the faculty mem- ber on, say, travel to a conference or other academic expenses. Thus, even if, as would be desir- able, the fund is generous enough to cover all publication fees, a faculty member would still have a motive to search for low fees, other things being equal. The problem of the design of a faculty member fund for subsidy of publication fees is more complicated in a regime like our present one where copyright exists (mainly because the fund for fees cannot be fungible with other academic expenses, lest the faculty member simply avoid all fees by publishing in tradi- tional journals with no fees). On this matter, see Stuart M. Shieber (2009). 49 See Bramble (2006, 213); Leslie A. Harmel (2005, 568); Suber (2009a, § 5), reporting that the number of universities providing funds to pay publication fees for open access journals increased in 2008 from six to eleven; Universities UK and the Research Information Network (2009), containing recommendations for university payments of open-access publication fees. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 323 research universities are typically in the millions of dollars. Because pub- lishers would only be able to charge for printing and distribution costs in the absence of copyright, and because all publications would be freely avail- able on the Internet, universities would be expected to save substantially. On the whole, the amount that universities would save could exceed the amount they would spend on publication fees, for the subscription and new book prices now paid cover publisher costs and proﬁ ts, whereas the publication fees would cover only publisher costs. That is, university expen- ditures on publication fees could be less in a world without copyright than their expenditures today on subscriptions and book purchases, because universities would no longer be ﬁ nancing publisher proﬁ ts from academic works. H owever, those universities whose faculties are relatively produc- tive researchers would tend to pay more in publication fees than they would save on subscriptions and book purchases. Grantors too would be predicted to want to subsidize publication fees. 62 Since a grantor usually has as a speciﬁ c purpose the publication of research, its incentives to subsidize publication might be comparable to those of universities. Again, consistent with this prediction about grantor behav- ior is that several prominent research sponsors today fund author publi- cation costs; the Howard Hughes Medical Institute, the National Science Foundation, and the National Institutes of Health allow for this (Denicola 2006, 359). 50 At doctoral degree granting institutions in the United States, the mean expenditure in 2006 on library materials was $8.688 million, including $2.263 million on serials. See Mary Jane Petrowski (2007, 23). For example, in that year Brigham Young University spent $23.942 mil- lion on library materials ( id ., 25) ; Catholic University spent $5.122 million ( id., 27); Duke University spent $33.532 million ( id., 27); and Yale University spent $74.938 million ( id., 43). 51 T o amplify, suppose that universities are the only purchasers of academic works in a world with copyright and that in the absence of copyright, publication fees would equal only publisher costs. Then university expenditures on publication fees would fall in the absence of copyright by the amount of publisher proﬁ ts. However, to the extent that purchasers of academic works today are parties outside the university (and that such parties do not publish and thus would not pay publication fees), university expenditures on publication fees would fall by less than the amount of publisher proﬁ ts. Indeed, if the fraction of purchasers outside the university were sufﬁ ciently high, university expenditures on publication fees would rise in the absence of copyright. 52 F or example, Guterman (2004, A11) calculated that Duke University would pay slightly more in publication fees in an open access setting than it would save on its library journal budget. See also the discussion of this issue in Donald King and Carol Tenopir (2004). 53 See also BioMed Central (2009), noting thirty seven major funders of biomedical research that allow payment of article processing charges for open access publications. 324 ~ Shavell: Should Copyright Of Academic Works Be Abolished? Both uni versities and grantors could, as a partial substitute for paying authors publication fees, subsidize journals or book publishers directly, allowing them to reduce the publication fees they would need to charge authors. However, this policy would often be less attractive for universi- ties and grantors than payment of publication fees. The reason is that if a university or grantor subsidizes a publishing venue, then all authors who publish there are beneﬁ ted, not just the university’s faculty member or the grantee. Thus, from a university or grantor’s perspective, it is more effective to subsidize publication fees than journals or publishers. Hence, I would expect the primary way in which universities and grantors would support publishing is by the payment of publication fees. 64 The effect of elimination of academic copyright on the level of publication de- pends on the extent to which universities and grantors would subsidize publica- tion fees. Because of the motive of universities and grantors to subsidize these fees, the number of published works is unlikely to decline substantially and could increase . This conclusion follows from what has been said above. 65 T o the degree that publications would be discouraged by the elimination of academic copyright, the social losses would be limited because these publications would not ordinarily be of high quality. An article of high quality that has been accepted for publication is likely to be recognized as having this character by its author. The author will be able to make judgments about the merit of the article from such indicators as the ranking of the journal to which it was accepted, the reception of the article at presentations, and cita- tions to it in unpublished form. Furthermore, the author will be aware of his or her own general standing in the profession and record of success. If, then, authors generally know when their accepted articles are of high quality, they will value publication of such articles commensurately and thus be relatively unlikely to decide against publishing on account of hav- ing to bear a publication fee. Moreover, the probability of having to bear a publication fee should be comparatively low for high quality articles, for universities will be more likely to defray publication fees for articles published in upper tier journals and for faculty with better track records; additionally, high quality articles tend to be published by faculty at uni- versities with substantial resources. 66 Conversely, articles of poor quality that have been accepted for publi- cation are likely to be recognized as weak by their authors. These articles will often have been rejected for publication at a number of journals, will have been accepted only at a low-ranking journal, will not have made Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 325 much of an impression on colleagues at presentations or have been cited often, and will have been written by individuals without strong scholarly records. In all, it seems that articles that have been written that are of poor quality are the ones whose publication would tend to be discouraged by publication fees. I ha ve explained that if an article has been written and is of high rather 67 than poor quality, it would be relatively unlikely to be discouraged from be- ing published by publication fees. However, it should also be asked wheth- er the writing of high quality articles in the ﬁ rst place would be reduced by publication fees. When an author contemplates writing an article that would turn out to be of high quality, would the author decide against mak- ing the effort because of the possibility it would be poor and that the author would then be likely to have to pay publication fees? This is possible, but it must also be taken into account that when an author contemplates writ- ing an article, the author might have an idea of its quality, or at least might formulate one after undertaking preliminary work on the article. The importance of the claim that discouraged articles are unlikely to be 68 those of high quality seems substantial because of the great variability in quality of academic articles. There is evidence that article quality is con- centrated, that a small fraction of all published articles receives the great majority of citations. I f publication of this group of articles would only be discouraged in a minor way by publication fees, then the social loss associ- ated with the typical discouraged publication will be small. The foregoing points about articles apply to academic books as well, but 69 I conjecture with somewhat less force. The reasons are that an author’s abil- ity to forecast the quality of a book from the status of the publisher who accepted it or from citations to it in unpublished form seem less than for an article, and also that the prospect of having to pay publishing fees would weigh more heavily on an author contemplating writing a book than an article. T he social loss from a discouraged publication would also be limited because 70 an unpublished work could be posted on the Internet. Articles and books that are not published could and probably would be posted on an Internet site, such as that of an open access repository, a working paper series, or an 54 See e.g ., John P. A. Ioannidis (2006), documenting that for the sciences, including the social sciences, “the most inﬂ uential papers are extremely concentrated in a few journals” and not- ing that a core number of journals, constituting about 2 percent of all journals, publishes 95 percent of all cited articles. 326 ~ Shavell: Should Copyright Of Academic Works Be Abolished? author’s personal website. Internet posting would limit the social loss from a work’s not being published because it would be available for all to read. Furthermore, the ability of potential readers to locate the work would not be greatly compromised by its not being published because the author could post it on a site of appropriate subject matter relevance and because of the power of search engines. 71 Of course, the point that an unpublished work can be posted on the Internet is moot if the work would not have been written in the ﬁ rst place due to the prospect of having to pay a fee for publication. But this possibili- ty is itself made less likely by the opportunity to post works on the Internet, for that means that the author still would obtain utility from having written an article or book. 72 To the degree that publications would be encouraged by the elimination of academic copyright and subsidy of publication fees, either social gains or losses could be engendered. The latter problem might be offset by university and grantor efforts to condition a subsidy on quality. I observed above that in the absence of copyright but given subsidy of publication fees, some authors might decide to write who would not do so in a world with copyright. The generation of such new works would often be socially desirable but would also sometimes be socially undesirable. 73 The possibilit y of socially undesirable publication is an instance of the standard disadvantageous economic byproduct of a subsidy: because those receiving the subsidy do not take into full account the cost of what they purchase—here publication services—they might decide to make a pur- chase even though true costs exceed the beneﬁ ts. Consider, for example, a journal of very low quality that so few read as to make its operation socially undesirable in view of its publishing costs of $2,000 per article. Even so, this journal could be imagined to operate successfully if its publication fee of $2,000 would be borne by universities, for authors might well derive posi- tive beneﬁ ts from publishing in it—suppose they had written articles that were rejected in better journals and wanted to see their work in print. 74 S uch problems of socially excessive publication could be countered if universities (and grantors) imposed quality constraints on articles that they subsidize. They would have an incentive to do this as well, because 55 N ote that it is plausible that the journal would not operate in a world with copyright, for given the assumption that the journal has very few readers, it would have to charge authors publica- tion fees approximating $2,000 per article. It would be expected that few authors would want to pay this fee to publish in a journal known to be poor and having few readers. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 327 publications of questionable quality would not tend to enhance their repu- tations (but would cost as much as high quality publications). However, universities would not want to exercise control over the subsidy of publica- tion fees for articles in such a way that academic freedom, and expression of unpopular views, would be compromised. Suppose, though, that uni- versities base the granting of subsidies for an article on the identity of the journal that accepted the article—giving subsidies to any article accepted in a journal in a named list of journals of sufﬁ cient reputation. Then the content of the particular article itself could not directly inﬂ uence the sub- sidy decision and thus the threat to academic freedom would presumably be limited. Likewise, if universities base the granting of subsidies for a book on the identity of the publisher that accepted the book, not on the book itself, there should be no real threat to academic freedom in respect to the publication of books. Summary to this point: T he effect of elimination of copyright on authors’ incen- 75 tives to publish might not be negative overall—it might lead to more publications, due to subsidy of publication fees—and to the extent that it would discourage publications, the loss in social welfare would probably be limited. The preceding discussion suggests that the level of publications—especially of articles— would be unlikely to fall and might even increase in the absence of copy- right. If the number of published works were to fall, the decline would be unlikely to be substantial. The main reasons for these predictions are that universities and grantors would have an incentive to pay authors’ publica- tion fees and that readership of works and thus authors’ esteem would rise in the absence of copyright. Also, the social loss associated with discour- aged publications would probably be limited, for the quality of such publi- cations would tend to be low and unpublished written works could still be posted on the Internet. Further, although the effects on social welfare of a subsidy-induced increase in publications could be positive or negative, the latter possibility could be countered by the conditioning of subsidies on publication quality. Social beneﬁ ts from eliminating academic copyright—deriving from the free 76 availability of academic works. I f copyright of academic works were ended, a social beneﬁ t would be enjoyed associated with works that would still be published but that would otherwise have been copyrighted, and also with some of the works that would be published only because of the absence of copyright. In the absence of copyright, all these articles and books would presumably become instantly available on the Internet for individuals to 328 ~ Shavell: Should Copyright Of Academic Works Be Abolished? download freely. Also, print copies would often be produced and would sell for approximately production cost, due to competitive pressures. Hence, for instance, many new academic books would quickly become available as moderately priced hardbacks or as inexpensive paperbacks. Furthermore, teaching materials drawing on published works would become easy for academics to assemble because permissions for use would not have to be secured or royalties paid. 77 In considering these social beneﬁ ts, we need to make a realistic compari- son with the state of affairs under copyright. Today, universities subscribe to a large number of journals and make their content freely available to many in the university community through library and Internet access, and faculty members often have budgets that are designated for the purchase of journals and books. Moreover, copyright protection is signiﬁ cantly in- complete for articles. Journals are increasingly allowing free downloading of their articles from the Internet, at least after a window of time following publication. In addition to this voluntary provision of access to articles, there is much legally unpermitted exchange of articles because of the ease of copying and of emailing content over the Internet. Authors often main- tain personal websites from which their actual published articles can be downloaded despite this being a violation of publisher’s copyright privi- leges. Furthermore, as I will discuss in Section 4.1, pre-published versions of published articles can usually be located on the Internet and are often deposited in repositories. Hence, the difference that elimination of copy- right would make to the availability of articles should not be exaggerated; access to articles that are copyrighted is much greater than would be true were copyright protection insisted upon and rigorously enforced, and were it true that each individual had to pay for copyrighted works rather than a university paying for them. 78 N evertheless, the beneﬁ ts of abolition of copyright of articles are positive. Faculty and students do not have ready access to all published articles on the Internet and often face costly-in-time hurdles to locate what is in theory freely available. Articles in pre-published form are not equivalent to pub- lished versions. The assembly of teaching materials from articles and the 56 See, e.g., PubMed Central (2009), containing a list of journals in medicine and the life sci- ences, and showing that approximately half of the journals allow immediate access to their content and the remainder after two to thirty-six months, with twelve months being most common, and a similar listing at the Stanford University (2009) HighWire press website. 57 See Section 4.1. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 329 printing of them is often seriously constrained by copyright. F urther, many academics and students in institutions without substantial resources (includ- ing many small colleges and junior colleges in the United States and teaching institutions in other countries) cannot afford to pay for more than a narrow segment of journals. Additionally, there are numerous individuals who are not members of the university community but who wish to read academic works. When one takes these various observations into account and aggre- gates the beneﬁ ts of a copyright-free world over the relevant populations and the huge number of articles that are published, my supposition is that the sum would be substantial. Books seem to enjoy greater effective copyright protection than articles. 79 Academic books published under copyright today are not usually available 59 60 in electronic form, e ven though publishing practices are changing. And when books are offered in electronic versions, they must generally be pur- chased; unlike articles, books are not usually made freely available after a waiting period. Moreover, my personal impression is that there is a much lower ability of individuals to obtain electronic copies of books in pre- published version from the Internet than of articles. Hence, it appears that the difference that elimination of copyright would make for a typical book is greater than for a typical article, so that the proportional social beneﬁ t would be greater per book. W hether for articles or for books, it is evident that the existence of the 80 Internet greatly magniﬁ es the social beneﬁ ts that would be gained from elimination of copyright. Because of the Internet, all new academic works 58 See, e.g. , Princeton Univ. Press v. Mich. Document Servs., 99 F.3d 1381 (6th Cir. 1996) (en banc), afﬁ rming summary judgment against a commercial copy shop for copyright infringement be- cause it reproduced substantial segments of copyrighted materials in course packs; Basic Books, Inc. v. Kinko’s Graphics Corp., 758 F . Supp. 2d 1522 (S.D.N.Y. 1991), ﬁ nding Kinko’s liable to publishing houses for copyright infringement because it excerpted their copyrighted works, compiled them into course packets, and sold them to students; Ann Bartow (1998, 151), “[O]ver the past decade the scope of educational fair use has been dramatically compressed . . . . Fearful and litigation averse educational institutions . . . protect themselves by adopting copyright poli- cies that deﬁ ne a minimal scope of educational fair use, and make individual faculty members liable for any infringing photocopying activity outside of this petite orbit.” 59 A perusal of the websites of a number of university presses (Cornell University, Duke Univer- sity, Harvard University, University of Michigan, University of Texas) that I chose essentially at random shows that most sell few or no books in electronic form. The relatively limited list of books in electronic form offered by university presses may be found at the Association of American University Presses (2009) website. 60 On electronic publishing of academic books, see, e.g., Susan Gay (2007, 2). 330 ~ Shavell: Should Copyright Of Academic Works Be Abolished? would effectively become freely available to everyone. Befor e the age of the Internet, the elimination of copyright would only have meant that works would become available in the sense that journal publications and books would have fallen in price and that duplicated copies would have been mailed between individuals. 81 Finally , the elimination of copyright would mean that publishers would not invest resources in charging readers, making copyright arrangements with authors, granting licenses to others, or in protecting their copyrights through search for violations and through litigation. These savings would be distinct from the beneﬁ ts that we have been discussing due to free access to works and would be of some importance. 82 Conclusion. The judgment that I reach is that there is a plausible case for elimination of academic copyright, even though on a priori grounds one cannot come to the conclusion. I have offered arguments suggesting that the overall level of published academic articles and books might actually increase in the absence of academic copyright due to subsidy of publica- tion fees, but in any event, would not be likely to fall substantially. I have explained as well that any reduction that does occur would tend to be of relatively low quality works, and that most articles would be available on the Internet if not published, implying that the social loss per discouraged work would be relatively low. An increase in publications could be socially beneﬁ cial or detrimental, but the latter problem of excessive publication could be countered by limiting subsidization of publication fees to works 61 The a vailability of works copyrighted in the past is a different issue and one that I do not ad- dress, as the change in law I contemplate would abolish copyright only for future works, as I state in Section 5. In particular, therefore, I do not consider questions surrounding the social desirability of Google’s plan to digitize copyrighted and out-of-copyright books. See Google Book Settlement (2009). For conﬂ icting discussions of the impact the settlement will have on access to online books, see Robert Darnton (2009), criticizing the settlement for putting control of access to electronic versions of university books in the hands of a single for-proﬁ t company, and Paul Courant et al. (2009), presenting Paul Courant’s response to Robert Darnton that the settlement creates a valuable “universal bookstore,” and Darnton’s reply to Courant continuing to criticize the settlement for giving Google a de facto monopoly. 62 King (2007, 99, 104) discusses rights management and copyright protection activities as a component of indirect costs, which make up from 20 percent to 50 percent of total costs, and mentions an estimate that subscription management accounts for 7 percent of total costs. These costs would not be incurred in a copyright-free world. 63 As I explain in the model, whether eliminating academic copyright is socially desirable (in both the case where publication fees would and would not be subsidized), depends on the distribution of authors by type t , the cost of writing, c , and the values placed by readers on academic works. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 331 published in venues of reasonable quality. The social beneﬁ ts from ending copyright would be that the entire body of articles and books now under copyright, and thus the most valuable, highly cited works, would become available to all, for free downloading from the Internet, could be included in teaching materials without seeking permissions, and the like. Moreover, the costs of arranging and protecting copyrights would be avoided. In all, it does not seem unlikely that the beneﬁ ts would dominate the costs, and my prediction is that further analysis would validate this view. 4. THE FUND AMENTAL DIFFICULTIES FACED BY THE OPEN ACCES S MOVEMENT AND THE PR OB ABLE NEED F OR ABOLITION OF AC ADEMIC COP YRIGHT I n this section, I describe brieﬂ y the efforts of the open access movement 83 to make academic works freely available; I explain why these efforts are un- likely to succeed fully or without undue delay; and I suggest that a change in law may be needed to end academic copyright if the goal of unlimited access to academic publications is sought. 4.1. The Open Access Movement and Its Problems The most notable efforts to achieve unconstrained availability of academic 84 works fall under the heading of the open access movement. The announced aim of this movement is to assure that individuals can costlessly obtain published academic works, mainly journal literature. Open access initiatives have generally been limited in their objec- 85 tives, however. F requently, they seek to make freely available only pre- publication versions of articles, usually by having authors deposit these versions of their articles in repositories. What such initiatives achieve 64 See not e 4 above. As Peter Suber (2007) states, open access means the removal of “ price barriers (subscriptions, licensing fees, pay-per-view fees)” for readers and focuses on “peer-reviewed research articles.” 65 See Peter Suber (2009b) for a detailed description of open access initiatives. 66 P re-publication versions of articles range from “preprints,” versions that have not been ac- cepted for publication, to “postprints,” versions that have been approved for publication by peer review (but that may not have been copyedited); see Suber supra note 4. See, e.g., Na- ture Precedings (2009), allowing researchers to post preprint manuscripts; National Institutes of Public Health (2009) for its public access policy requiring grantees to “submit ﬁ nal peer- r eviewed journal manuscripts that arise from NIH funds to the digital archive PubMed Central upon acceptance for publication….” 332 ~ Shavell: Should Copyright Of Academic Works Be Abolished? over the status quo—in which discussion paper versions of articles have always been available—is not entirely clear, although it is positive, and the more so the closer is the accessible pre-publication version to the pub- lished article. In any case, a pre-publication version of an article is not a perfect substitute for the published article: readers will not necessarily know what is missing from a less than ﬁ nal version of an article, they will beneﬁ t from copyediting, they might want to cite ﬁ nal page numbers, and the like. Although some initiatives do aim to make actual published articles freely available, this is usually sought only after a delay, often six months or a year. 86 N ot only have open access initiatives been limited in their objectives, the initiatives have also encountered difﬁ culties in achieving their announced goals. Academics who have pledged support to an initiative or have been required to do so by their institutions have often been reluctant to follow through on their obligations. For example, a group of prominent scientists asked in 2001 for a boycott of scholarly journals that did not provide open access to publications within six months of publication (Public Library of Science 2001). Although over 30,000 scientists signed a promise to join the boycott, few actually did so and the boycott was viewed as a failure (Young 2002, A14). Recently, faculties at Harvard, MIT, and Stanford universities decided to require their members to publish their articles only if open access is allowed by publishers, but faculty members have been allowed to opt out of this “mandate,” presumably because of fear that they would resist it. R elatedly, the ability and willingness of universities that have 67 See, e.g., Howard Hughes Medical Institute (2007) for its policy requiring “its scientists to publish their original research articles in scientiﬁ c journals that allow the articles and sup- plementary materials to be made free accessible in a public repository within six months of publication”; the Scientiﬁ c Council of the European Research Council (2007) for its policy mandating that “all peer-reviewed publications from ERC-funded research projects be depos- ited on publication into an appropriate research repository where available, such as PubMed Central, ArXiv or an institutional repository, and subsequently made Open Access within 6 months of publication.” 68 H arvard University’s open access policies (Harvard University 2009), adopted in 2008 and 2009, apply at present to the Faculty of Arts and Sciences, Harvard Law School, and the John F. Kennedy School of Government. Under these policies each “faculty member grants to … Harvard … permission to make available his or her scholarly articles . . . to the public in an open-access repository.” But “[t]he Dean . . . will waive application of the policy to a particular article upon written request by a Faculty member ….” MIT also decided to employ an essen- tially identical open access policy (MIT News 2009). Stanford University’s School of Education adopted a similar open access policy (Stanford University, School of Education 2008). Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 333 adopted open access policies to sanction violations by their faculty has been limited. The open access movement has also sought to foster the development 87 of open-access academic journals, such as those in the Public Library of Science (Public Library of Science 2009). Open access journals generally do not copyright their works and often charge fees to authors to publish in or- der to cover their costs; thus they are examples of the kinds of journals that I described in Sections 2 and 3 that would exist in a world without copy- right. The number of open access journals has been rising, and they now constitute about 4 percent of all journals. However, open access journals are on the whole not among the most prestigious venues for publication (Thomas Hess et al. 2007). An important gauge of the effect of the open access movement is the 88 percentage of academic works published under conditions of open ac- cess. The fraction of journal publications that are immediately available on an open access basis is reported to be less than 5 percent (Björk, Roosr, & Lauri 2008), although almost all journal publications are available in at least some pre-published form. Finally, books are generally unavailable in open access format. 4.2. Explanation for the Problems The pr oblems and only partial success of the open access movement are not 89 difﬁ cult to explain. Obviously, traditional publishers would be unwilling to cede copyright privileges for nothing, for that would end their ability to charge for access. Traditional publishers would have to be paid to give up 69 For example, the Harvard, MIT, and Stanford open access policies, see preceding note, do not mention any sanctions for their violation. 70 The open ac cess website (Directory of Open Access Journals 2009) lists 4,050 open access jour- nals, and the Genamics Journal Seek (2009) website reports the total number of journals to be 94,589. However, Björk & Hedlund (2009) report that 8.5 percent of journals are open access. 71 See also B ramble (2006, 219), noting that “[c]ompared with more traditional modes of pub- lishing, Open Access journals and Open Access publishing/archiving in general suffer from a prestige gap”; Denicola (2006, 360), stating that “[w]hen authors who had never published in open-access journals were asked why they had not, sixty-nine percent said that their percep- tion that open-access journals have low prestige and low impact was an important or very important factor.” But see Public Library of Science (2005), noting, in addition, that young journals from the PLoS, particularly PLoS Biology, have made an impressive impact despite the prestige advantage of their more established competitors. 72 One survey reports that 97 percent of journals are “green,” with a majority allowing posting of postprints and the remainder posting of preprints (SHERPA RoMEO 2009a). 334 ~ Shavell: Should Copyright Of Academic Works Be Abolished? copyright, but as I discussed in Section 2.1, authors would not be expected to ﬁ nd it personally advantageous to do this; they will prefer for publishers to hold copyright in order to avoid having to pay fees for open access pub- lication. This is presumably why open access initiatives have been generally limited in nature and have not sought immediate open access to published articles themselves. 90 The r elative lack of success of open access journals is also not hard to understand. A primary difﬁ culty for an open access journal is, again, that authors will usually not want to pay publication fees to an open access publication; authors will tend to prefer to save these fees and not to beneﬁ t from the expanded readership that open access to their own works would bring. An additional problem for an open access publica- tion is that authors will be reluctant to publish with them due to their lack of prestige, as I mentioned above. Indeed, I suspect that many aca- demic authors would consider their lack of prestige a more serious dis- advantageous of open access publications than possibly having to pay publication fees. 91 I f open access publications are of lower quality than many tradition- al publications, the question arises why this would be so. The answer is presumably that open access publications were relatively recently begun and that high-quality publishing venues are difﬁ cult to establish. There is, though, no apparent intrinsic basis for believing that open access jour- nals would be of low quality. Open access journals can charge sufﬁ ciently high submission and publication fees to ﬁ nance a rigorous refereeing and editorial process; they can, and would generally be motivated to, exer- cise selectivity in the works that they accept for publication. H ence, the quality disadvantage of open access journals would be expected to wane over time. 73 Note the irony in this regard: academics’ desire for esteem is not only the factor that makes elimination of copyright socially desirable, it is also a factor that stands in the way of its elimination. 74 A number of commentators have suggested that open access journals would have an incentive to publish a greater volume of works in order to increase revenue—for their revenue would depend on publication fees—and that this would lead to a dilution of article quality. See, e.g., Karen Hunter (2004); and John Ewing (2004). This view does not make clear economic sense: if a journal degrades its quality, the fee that it will be able to charge each author will decline (for the esteem an author derives from publication depends on journal quality); hence, if a journal increases the number of acceptances and quality declines, its publication fee revenue and its proﬁ ts will not necessarily increase. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 335 A nother issue of relevance is the possibility that traditional journals would convert into open access journals. C onversion would, however, seem usually to go against the joint interests of authors and journals be- cause authors, as I have emphasized, would usually prefer to avoid publi- cation fees and to transfer copyright to journals so that they can impose charges on readers. Equivalently, if a traditional journal converted to an open access plan and charged publication fees, its submissions from authors would be likely to fall. A subset of authors, however, might prefer to pay publication fees in order to obtain open access for their articles, suggesting that a traditional journal might offer the option to an author of paying a publication fee and having that article freely available for downloading on the Internet. 4.3. The Fundamental Divergence between the Effects of Individual and Collective Efforts toward Open Access—and the Consequent Need for Legal Change One can e xpress the difﬁ culties faced by the open access movement in 93 terms of a basic divergence between the effects of individual and col- lective actions. If an individual academic insists on open access to his or her publication, this single action brings about no real change in the overall system of copyrighted works—that system continues to exist. And because the system of copyright continues to exist, the academic who obtains open access for his or her work does not as a consequence 75 This happened with the journal Nucleic Acids Research ; see Nucleic Acids Research 2009. See also Clair e Bird (2008), discussing and evaluating impact of Nucleic Acids Research ’s transition to open access, and noting that 68 other Oxford journals have since transitioned to some form of open access. A list of traditional journals that have converted to open ac- cess journals can be found at Open Access Directory (2009). On the other hand, some for- merly open access journals have recently switched to subscription models. See, e.g. , Journal of Visualized Experiments (2009), explaining that high production costs for video articles led the previously-open access Journal of Visualized Experiments to adopt a subscription model. 76 I am pr esuming that authors would have to pay the publication fees in saying this. Clearly, if universities would subsidize publication fees, journals that converted to an open access model would not lose submissions; they would gain submissions because their readership would increase. 77 See, e.g. , Toby Green (2009, 75), reporting that of 263 publishers responding to a recent survey by the Association of Learned and Professional Society Publishers, 30 percent offer optional open access to authors. For instance, the publisher Springer offers this option to authors under its Open Choice program; authors can publish individual articles giving readers open access if they pay a fee of $3,000 (Springer 2008); see also, e.g., Elsevier (2009); Oxford Journals (2009); Taylor & Francis (2009); Wiley InterScience (2009). For a partial list of journals offering this option and their prices, see SHERPA RoMEO (2009b). 336 ~ Shavell: Should Copyright Of Academic Works Be Abolished? obtain the beneﬁ ts that would ﬂ ow from the general abolition of aca- demic copyright. That is, the academic does not enjoy free access to all academic works. Likewise, even if a university requires that there be open access to works published by its faculty members, the university will not then garner the beneﬁ ts of free access to all academic works—the univer- sity will still have to pay for journals and for books. Hence, it may well be that a collective action, namely, one that ends the legal right to copyright academic works, is needed to achieve open access, or at least to achieve that objective soon. 94 A related observation is that it may be rational for an individual aca- demic or for a individual university to refrain from committing to publish academic works only if open access is granted, yet to support a change in law to eliminate academic copyright. Indeed, I suspect that most academics and many universities are in this category. 5. HOW COULD THE ELIMINATION OF AC ADEMIC COP YRIGHT BE ACCOMPLISHED ? 95 I have suggested in the preceding section that if free access to published works is thought to be socially desirable, it would best be accomplished by a modiﬁ cation in copyright law rather than by the initiatives of the open access movement. Let me now brieﬂ y describe such an alteration in law and whether it would be difﬁ cult to effect. 5.1. A Law Eliminating Copyright in Academic Works 96 B y a law abolishing copyright in academic works, I mean a law denying the right of copyright for all newly authored academic works in the United States. I suppose also that the law would seek to bar use of technological methods intended to impede copying and that contractual arrangements to prevent copying would be made unenforceable. Hence, all new academic works could be copied and presumably would be made available free of charge in electronic form or sold at cost in printed form. Copied material would, however, have to include citations to its source and might be protected 78 The arguments suggesting that c opyright be abolished apply throughout the world, but I restrict attention here to the law in the United States. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 337 against certain forms of abridgment—otherwise the author could bring a legal action against the violator. The implementation of a law ending copyright of academic works would 97 require a deﬁ nition of academic works. Let me ﬁ rst consider articles (books will be addressed subsequently). An article could be deemed an academic work if it were to be published in an academic journal. To determine whether a journal is academic, four indicia could be employed: whether its authors are usually academics; whether its readers are mainly academics; the degree to which its content is academic in character (displays sophistication and knowledge of prior learning); and, most important, the magnitude of any royalties received by authors (low or no royalties would favor classiﬁ cation as academic). These four factors bear on the importance of esteem to authors and thus, 98 as has been explained at length above, on whether elimination of copyright would be unlikely to detrimentally diminish the volume of publications. If an article is going to be published in a journal containing works that are typically authored and read by academics and that exhibit the usual qualities of academic works, then the likelihood is that the purpose of the author of the article in question was to gain recognition from academics and also that it was not written in order to obtain royalties. Furthermore, and crucially, if the journal does not generally pay any, or more than mod- est, royalties, that would strongly reinforce the inference about the object of the author and the lack of need for copyright to have induced the writing of the article. 79 The author of a work would have the right to enjoin the distribution of unattributed material and to collect damages for harm done due to lack of attribution or other violations of rights. For a discussion of the basis for these rights, see Mark A. Lemley (1995). (Today, such legal rights are presumably largely unnecessary because copyright prevents others from copying work without permission.) But there would be no legal duty to cite the author for the author’s ideas. That duty would be enforced as it is now, by scholarly norms. Copyright law does not protect ideas. See , e.g. , Goldstein (2005, 1:52). 80 One possibility is that these criteria would be applied by an expert extra-judicial body, which would create and maintain a list of journals considered academic. Parties displeased with the classiﬁ cation of a journal as academic by the expert body could have a right of appeal to the courts. 81 This last fact or can be applied only in regard to a journal that was in existence before the pas- sage of the statute under consideration. For journals established after the passage of the statute, evidence about planned payments to authors could be examined. 338 ~ Shavell: Should Copyright Of Academic Works Be Abolished? For the most part, journals would be straightforward to classify accord- ing to the four criteria, and those journals that we now would naturally consider to be academic would clearly be treated as such under the criteria. It is true that in some instances, not all of the criteria would apply, but through direct consideration of the underlying issue of whether copyright is needed to induce publication, classiﬁ cation could still often be intelli- gently accomplished. Consider an existing journal that publishes articles on international affairs; that its readership is less than half academic; that its authors are usually not academics—the nonacademic authors are drawn from think tanks and the State Department; that its articles, while serious, can readily be understood by non-specialists; and that it pays nothing to its authors. Should this journal be considered academic? If we ask the crucial question whether copyright is needed to induce publication of the articles in the journal, our answer would be no, so that we ought to treat the journal as academic and not allow copyright. In particular, the journal’s nonaca- demic authors have positions suggesting that they would value the esteem of their readers: members of think tanks are usually highly educated indi- viduals, often have connections to academia, and in many respects appear to share the goals and ethos of academics; and some members of the State Department may have related objectives. That the readership of the journal is not predominantly academic is not necessarily dispositive, for this factor serves only as an indicator of whether the author wants esteem. As a general matter we would infer that if the audience of a journal is not academic, then gaining the esteem of that audience would not be highly valued by an author. But in the case of a journal in international affairs, the esteem that a writer would desire would plausibly be from a wider community than just academics interested in international affairs; presumably the author would be pleased to gain notice from readers in think tanks and members of the State Department. That the articles can be comprehended by nonspecialists is, in the context at issue, not surprising, for international affairs is not an abstruse or technical ﬁ eld. Finally, that authors do not receive payment for their articles demonstrates that they did not need royalties to publish and suggests that many would have been willing to pay publication fees in the absence of copyright. Furthermore, institutions such as think tanks and the State Department might wish to pay publication fees in order to encourage their employees to publish. 100 T urning now to books, an approach similar to that just described for articles could be attempted. That is, a book could be deemed an academic Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 339 work if it was to be published by a publisher whose books—or books pub- lished under an imprint or distribution plan—satisfy the four factors that I mentioned in regard to journals. Notably, then, most books published by university presses, and sold by them as academic rather than trade press books, would be classiﬁ ed as academic works. However, I suspect that the classiﬁ cation of books would be more difﬁ cult than the classiﬁ cation of articles because books generally do result in positive royalties. In any case, textbooks would not be considered academic, so could be 101 copyrighted. The premise in this regard is that most textbooks would prob- ably not be written in the absence of the prospect of proﬁ t from copyright. That is not to deny that some measure of esteem is derived by academics from authoring textbooks. It is rather that the strength of this factor does not seem very strong. Compilations of academic works, however, would generally not be ac- 102 corded copyright. Compilations are different from textbooks in that they require relatively little effort to create. Therefore, my suspicion is that disal- lowing copyright for compilations would have only limited effect on the volume of compilations generated. 5.2. Effecting the Elimination of Academic Copyright Ending academic c opyright seems straightforward to accomplish from a 103 legal perspective. Presently, copyright is authorized under Title 17 of the U.S. Code. To terminate academic copyright could be achieved under new legislation, deﬁ ning an exception for academic works. Other carve outs al- ready exist, and ther e is nothing conceptually difﬁ cult about crafting a 82 A number of different types of carve-outs are already codiﬁ ed. There are complete denials of copyright to works produced by certain actors. See, e.g., 17 U.S.C. § 105 (2006), denying copyright to works produced by the U.S. Government. There are denials of particular rights normally secured by copyright to certain types of works. See, e.g. , id. § 114, excluding the right of performance from copyrights in sound recordings; id. § 120, excluding from copyrights in certain architectural works the rights to prevent the making, distribution, or display of pictures of the work. And there are denials of particular rights normally secured by copyright when the right is violated for a particular purpose. See, e.g. , id. § 108, granting libraries and archives limited reproduction rights to preserve works or make them available for research at other libraries or archives, notwithstanding copyrights in those works; id. § 121, making reproduction and distribution of copyrighted works “in specialized formats exclusively for use by blind or other persons with disabilities” not an infringement. However, Paul Goldstein has suggested to me that elimination of copyright for academic works could lead to conﬂ ict with the obligations of the United States under the TRIPS Agreement. 340 ~ Shavell: Should Copyright Of Academic Works Be Abolished? legal exemption for future academic works. M oreover, doing so would not be constitutionally problematic because there is no constitutional right to copyright. Indeed, insofar as abolishing academic copyright could even spur the production of knowledge, it would be the constitutionally pre- ferred course according to the Copyright Clause’s preamble that copyright should be used “To promote the Progress of Science and useful Arts.” 104 Eliminating academic copyright also seems feasible from a political per- spective because of its likely endorsement by universities, academics, and students. One would expect universities to support abolition of copyright because of their general desire to promulgate knowledge, their wish for fac- ulty to be able to assemble teaching materials without their having to secure permissions from copyright holders, and their purely ﬁ nancial interest in not having to pay for subscriptions to journals and high prices for books. In the latter regard, recall that universities as a whole would probably save funds if academic copyright is eliminated, even assuming that they fully subsidize publication fees. This is because universities would no longer serve as the source of much of the large proﬁ ts that accrue to publishers of academic works. Academics themselves would be likely to favor abolition, especially if they felt that publication fees would often be subsidized by their universities and grantors (which as I explained above would be probable). Under this assumption, academics would probably be better off were copy- right eliminated: their works would remain costless for them to publish but would be more widely circulated; and they themselves would have free access to all academic works. Finally, students should exhibit a preference for elimination of copyright assuming that they would agree that it would not lead to a dilution in the quantity and value of academic works, for they would then be able to obtain all works for free on the Internet and at cost in print; furthermore, their instructors would face no copyright hurdles in assembling teaching materials, which would also beneﬁ t them. 83 Abolishing copyright for works already in existence may offend constitutional protections of property and thus could be subject to constitutional restraint. 84 U.S. Const. art. I, § 8, cl. 8. 85 On the university’s ﬁ nancial advantage (despite their subsidization of publication fees), see note 51 above. 86 The only obvious qualiﬁ cation is that an academic author who today would make substantial royalties from an academic book would forgo this income in the absence of copyright. How- ever, as I observed earlier, note 31 above, this is not the usual situation. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 341 Opposition w ould, of course, be expected to be voiced by for-proﬁ t pub- lishers of academic works, as their copyright-derived earnings and sales are substantial. For example, in 2008 Elsevier earned approximately $1 billion in operating proﬁ t on $3.145 billion in revenue. C opyright fuels much of these proﬁ ts by enabling commercial publishers to raise journal prices sub- stantially above cost, as evidenced by the much lower subscription prices for non-proﬁ t scholarly publications. Opposition might also be expressed by some learned societies, because they would have to cede the proﬁ ts that they now obtain from their own journals. Opposition c ould come as well from parties who would view the abolition of academic copyright as unde- sirable because it might lead to erosion of intellectual property rights in a wider domain. 6. CONCLUSION I have endeavored here to examine the effects of eliminating copyright of 106 academic works and the factors determining whether that change would be socially desirable. On the basis of a number of empirical judgments— notably, that universities and grantors would tend to subsidize publication fees—I suggested that ending academic copyright would be socially beneﬁ - cial. The reader may, of course, make different empirical assessments and 87 See Reed Elsevier (2009, 14, 124), reporting that in 2008 Elsevier obtained adjusted operating proﬁ ts of £568 million on revenue of £1,700 million, and suggesting an exchange rate of 1.85 U.S. dollars per pound sterling. For other academic publishers, see Wolters Kluwer (2009, 2, 3), reporting that in 2008 Wolters Kluwer earned operating proﬁ ts of €678 million on revenues of €3.374 billion, and John Wiley & Sons, Inc. (2008, 2), reporting that in 2008 John Wiley earned $222.99 million on revenues of $1.674 billion, and since its academic and scholarly revenues are 72 percent of the total, its estimated academic and scholarly earnings were $160.55 million. Because the annual reports do not allow one to differentiate between academic works and text- books and other works written for proﬁ t, the earnings ﬁ gures I have stated are overestimates of earnings on academic works but my conjecture is that they indicate suggest the order of magnitude of such earnings. 88 See, e.g. , Carl T. Bergstrom & Theodore C. Bergstrom (2004, 897 & tbl.1), determining that “the average price per page charged by commercial publishers is several times higher than that which is charged by professional societies and university presses.” 89 See John Willinsky (2004) on the dependence of learned societies on journal revenue. See also Guterman (2004, A12), observing that many learned societies use journal subscription rev- enues to ﬁ nance conferences and other activities and thus do not favor an open access and publication fee regime. Examples of learned societies that have stated opposition to or skepti- cism about open access publication are the American Physiological Society ( id. ), the American Mathematical Society (Ewing 2004), and the American Association for the Advancement of Science, which publishes Science (Wysocki 2005, A8). 342 ~ Shavell: Should Copyright Of Academic Works Be Abolished? come to a different conclusion. My principal goal was not to persuade the reader that my empirical judgments are correct but rather to identify and clarify the factors bearing on the social desirability of ending copyright of academic works. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 343 APPENDIX I pr esent below the analysis of the model informally discussed in Section 2. The assumptions are essentially as described there. In particular, academic authors derive utility from readership of their published works; publication is performed by publishers; authors beneﬁ t from wealth and they experi- ence disutility from writing effort. Deﬁ ne the following notation: x = number of readers of a published work; x ≥ 0; t = type of an author; t ≥ 0; tu ( x) = utilit y of an author of type t fr om x r eaders of a published work; u is positive, increasing in x , and concave in x ; y = income of an author; d = disutility of writing effort; d ≥ 0; F ( t , d ) = probability distribution function over t and d . A n author knows his t and d . The utility function of an author from a published work is tu ( x ) + y – d . (1) R eaders obtain utility from a published work but no utility from unpub- lished works. Let w ( x) = utilit y obtained by readers from a published work; w is positi ve, increasing in x with w' (0) > c , and concave in x . P ublishing a work involves a ﬁ rst copy cost and a marginal cost per unit. Deﬁ ne f = ﬁ rst copy cost of publishing a work; f > 0; c = marginal cost of distributing a work to a reader; c ≥ 0. H ence, the cost of publishing and distributing a work to a positive number x readers is f + cx . (2) 90 Thus I abstract in the model from the point that readers do in fact obtain utility from reading works that are unpublished but posted on the Internet. 91 Note that since w does not depend on t , the utility obtained from a published work is the same regardless of who writes the work. This simpliﬁ cation is useful in the model, but as the reader knows, I discussed the important implications of differences in quality of works in the text. 344 ~ Shavell: Should Copyright Of Academic Works Be Abolished? I f an author writes and publishes a work and it is read by x indi viduals, the contribution to social welfare will be tu ( x ) + w ( x ) – ( f + cx ) – d , (3) for the author will enjoy utility of tu ( x ), readers will obtain utility of w ( x ), publication costs will be f + cx , and writing effort will be d . A.1. Socially Ideal Outcomes Given that a work has been written and ﬁ rst copy costs for publication have been incurred, the socially optimal number of readers is the x maximizing tu ( x ) + w ( x ) – cx . (4) L et this optimal x be denot ed x ( t), which must be positive (because w' (0) > c ) and is determined by tu' ( x ) + w' ( x ) = c . (5) Note that x ( t ) exceeds the x at which w' ( x ) = c because the author derives utility from readership, and observe that x ( t) is incr easing in t. A work should be written, and then published, if and only if ∗ ∗ ∗ tu ( x ( t )) + w ( x ( t )) ≥ ( f + cx ( t )) + d . (6) A.2. Outcomes under Copyright Let me now ask what occurs under the regime of copyright as a function of an author’s type t and d. I assume that if an author wishes to publish, the author makes a contract with a publisher involving two terms: p = price at which the work will be sold by the publisher; p ≥ 0; r = payment by the publisher to the author. If r > 0, it will be described as a royalty payment, and if r < 0, it will be called a publication fee. The price p determines readership x ( p ) according to w '( x ) = p , (7) 92 Social welfare does not include the income of authors. The motivation is the standard one that payments between parties are transfers and do not themselves absorb resources. 93 Implicitly differentiate (5) with respect to t to obtain x ' ( t ) = – u' ( x )/[ tu"( x ) + w" ( x )] > 0. 94 I f (6) holds with equality, then whether the work is written does not affect social welfare, but for expositional convenience I adopt the convention that the work should be written and make similar assumptions below without further comment. 136 Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 345 because readers can be viewed as maximizing w ( x ) – px . We know that x '( p ) < 0 by implicit differentiation of (7). The utilit y of an author given a contract (but exclusive of writing effort d ) is tu ( x ( p )) + r . (8) I assume that publishers are numerous and perfectly competitive, so that they break even on their contracts, that is, px ( p ) = f + cx ( p ) + r . (9) Therefore, r = px ( p ) – [ f + cx ( p )]. The c ontract that an author chooses is determined by maximizing (8) subject to (9), that is, by maximizing tu ( x ( p )) + px ( p ) – [ f + cx ( p )] (10) with respect to p . For simplicity, I assume that (10) is concave in p . The ﬁ rst-order condition for the choice of p when it is positive is tu' ( x ( p )) x' ( p ) + x ( p ) + px' ( p ) = cx' ( p ) or tu' ( x ( p )) + x( p )/ x' ( p ) + p = c . (11) Denote the optimal choice of p given t as p ( t ), and observe that p (0) satisﬁ es x ( p )/ x' ( p) + p = c, that is, marginal revenue equals marginal cost, so that p(0) is the standar d monopoly price. For convenience, I assume that proﬁ ts would be positive at the monopoly price, p (0) x ( p (0)) – [ f + cx ( p (0))] > 0. (12) This assumption is r eﬂ ected in Figure 2, which shows the utility of the author at t = 0 as positive. Now it follows from implicit differentiation of (11) that p' ( t ) < 0; (13) that is, the optimal price of the work speciﬁ ed in the contract falls with t , as depicted in Figure 1. At a t sufﬁ ciently high, labeled t in Figure 1, p ( t ) must equal 0; using (11), this t is determined by t = { c – [ x (0)/ x' (0)]}/ u' ( x (0)), (14) 95 Differentiating with respect to t , we obtain w "( x ) x'( p ) = 1, so that x '( p ) = 1/ w "( x ) < 0. 96 The ﬁ rst-order condition (11) is of the form F ( p , t) = 0, and the second-order condition is F ( p , t ) < 0. Implicitly differentiating (11), we obtain F ( p , t ) p '( t ) + F ( p , t ) = 0. Hence p'( t ) = p p t –F ( p , t )/ F ( p , t ), and since F (p, t) = u' ( x ( p )) x' ( p ) < 0, we have p '( t ) < 0. t p t 145 346 ~ Shavell: Should Copyright Of Academic Works Be Abolished? which, note, is positive because x' (0) is negative. As in Figure 1, let t be such that p ( t ) = c. We know that 0 < t < t sinc e p ( t) is decr easing between 0 and 2 2 3 t . Also, as in Figure 1, let t be such that r ( t ) = 0. From (9), we know that 3 1 p ( t ) x ( p ( t )) – [ f + cx ( p ( t ))] = 0, implying that p ( t ) > c , and thus that t < t . 1 1 1 1 1 2 The utility of the author (exclusive of writing effort) who has chosen his preferred contract I will designate by U ( t) (wher e the subscript C denotes copyright); hence, U ( t ) = tu ( x ( p ( t ))) + p ( t ) x ( p ( t )) – [ f + cx ( p ( t ))]. (15) The utilit y U ( t) is incr easing in t, as is illustrated in Figure 2. To show this, suppose that t' > t and observe that if an author of type t' chooses p ( t ), then his utility would be t'u ( x ( p ( t))) + p ( t ) x ( p ( t)) – [ f + cx ( p ( t))], which exceeds U ( t) sinc e t' > t. Because U ( t') m ust exceed the utility obtained if p ( t ) is C C chosen, U ( t' ) > U ( t ). C C A n author will write if U ( t ) ≥ d, for then, if the author writes, the author’s beneﬁ t from publishing will be at least the writing cost. In summary, we have the following result. Proposition 1 : Under the regime of copyright, contract prices and the utilities of authors are as shown in Figures 1 and 2. Speciﬁ cally, (a) the contractual selling price of works p ( t ) declines with t between 0 and t (determined by (14)), where p (0) is the monopoly price and p ( t ) = 0; beyond t , p ( t ) = 0; 3 3 (b) there exist t and t , where 0 < t < t < t , such that r ( t ) = 0 and 1 2 1 2 3 1 p ( t ) = c ; hence, royalties are paid to authors for t < t and 2 1 publication fees are paid by authors for t > t ; (c) the utility U ( t) of authors (exclusive of writing cost) from the contracts they choose rises with t ; (d) authors choose to write and publish when U ( t ) ≥ d . A.3. Outcomes in the Absence of Copyright If there is no copyright and an author makes a contract with a publisher, the price cannot exceed c , because I assume that, immediately after publication, the work could be copied by other publishers and, as no ﬁ rst copy cost would be borne by them, competition would lead to the work being sold at its mar- ginal cost. I now consider the two regimes described in the text. In the ﬁ rst, authors bear publication fees, and I will sometimes refer to this simply as the 151 Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 347 no copyright regime. In the second, publication fees of f are paid by other en- tities, and I will sometimes refer to this as the no copyright/subsidy regime. Author s bear publication fees. The c ontract that an author who has writ- ten a work chooses is found by maximizing (10) with respect to p , where p is bounded by c . Denote by p ( t ) the price chosen by an author (where N stands for no copyright). We know that p ( t ) = c for t ≤ t because p ( t ) > c for such N 2 t and (10) is assumed to be concave in p . We also know that p ( t ) = p ( t ) for t > t because p ( t ) < c for such t . This is shown in Figure 1. Note that when p = c , r = – f . Hence, for t ≤ t authors who publish pay a publication fee of f and for t > t , authors who publish pay a publication fee of f + ( c – p ( t )) x ( p ( t )), so that fees exceed f and rise with t . L et the utility of an author (exclusive of writing cost) who publishes and has chosen his or her preferred contract be denoted U ( t ), so that U ( t ) = tu ( x ( c )) – f for t ≤ t N 2 = U (t) = tu(x(p(t))) + p(t)x(p(t)) – [f + cx(p(t))] for t > t . (16) C 2 Note that U ( t ) < U ( t ) for t < t , (17) N C 2 sinc e authors are forced to choose a price of c r ather than a higher price p ( t ). Observe also that U ( t ) is increasing in t , for it is increasing in t for t ≤ t , it is continuous at t , and we already know that it is increasing in t for 2 2 t > t . This is illustrated in Figure 2. Authors will choose to write, and then to publish, when U ( t )≥ d . We can summarize as follows. Proposition 2 : In the absence of copyright, contract prices and the utilities of au- thors are as shown in Figures 1 and 2. Speciﬁ cally, (a) the c ontractual selling price of works p ( t) equals c for t ≤ t N 2 and coincides with p ( t) for hig her t, so is declining until t , be- yond which it is 0; (b) authors pay a publication fee of f for t ≤ t and a fee of f + ( c – p ( t )) x ( p ( t )) for higher t ; (c) the utility U ( t) of authors (exclusive of writing cost) from the contracts they choose rises with t ; U ( t ) < U ( t ) for t < t ; U ( t ) = N C 2 N U ( t ) for t ≥ t ; and C 2 (d) authors choose to write and publish when U ( t ) ≥ d . 159 348 ~ Shavell: Should Copyright Of Academic Works Be Abolished? Authors do not bear publication fees . I now assume (for reasons given in Section 3) that universities or grantors pay publication fees of f (but not more). Hence, the contract that an author who has written a work is deter- mined by maximizing tu ( x ( p )) + px ( p ) – cx ( p ) over p ≤ c . This maximand differs from (10) by a constant, f, so that it must be maximized at the same p . In other words, p ( t ) described above also describes the choice of p when publication fees of f are not paid by authors. Note that in this case, when p = c , authors pay nothing, so that for t ≤ t authors who publish pay nothing and for t > t , authors who publish pay a publication fee of ( c – p ( t )) x ( p ( t )). Let the utility of an author (exclusive of writing cost) who publishes and has chosen his or her preferred contract be denoted U ( t ) (where NS NS stands for no copyright and subsidy of fee), so that U ( t ) = tu ( x ( c )) for t ≤ t NS 2 = U ( t ) + f = tu ( x ( p ( t ))) + p ( t ) x ( p ( t )) – cx ( p ( t )) for t > t . (18) C 2 In other words, U ( t ) = U ( t ) + f , (19) NS N as is shown in Figure 2. Authors will choose to write and to publish when U ( t ) ≥ d . Hence, we have NS Proposition 3 : In the no copyright/subsidy of publication fee regime, contract prices and the utilities of authors are as shown in Figures 1 and 2. Speciﬁ cally, (a) the contractual selling price of works p ( t ) equals c for t ≤ t N 2 and coincides with p ( t) for hig her t, so is declining until t , beyond which it is 0; (b) authors pay no publication fee for t ≤ t and a fee of ( c – p ( t )) x ( p ( t )) for higher t ; (c) the utility U ( t) of authors (exclusive of writing cost) from the NS contracts they choose rises with t ; also, U ( t ) = U ( t ) + f ; and NS N (d) authors choose to write and publish when U ( t ) ≥ d . NS A.4. Social Welfare in the Presence and in the Absence of Copyright Under copyright, if an academic writes and publishes, social welfare is ∫[ tu ( x ( p ( t ))) + w ( x ( p ( t ))) – ( f + cx ( p ( t ))) – d ] dF ( t , d ). (20) {( t , d )| U ( t ) ≥ d } C Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 349 N ote that the integrand, social welfare, must be positive whenever the author decides to write and publish. To see this, observe that U ( t ) ≥ d means that tu ( x ( p ( t))) + p ( t ) x ( p ( t)) – ( f + cx ( p ( t))) – d ≥ 0. But w ( x ( p ( t ))) > p ( t ) x ( p ( t )), so that social w elfare must be positive. The explanation is that the author captures less than the readers’ surplus, so that if the author decides to publish, it must be that that adds to social welfare. U nder the regime of no copyright and no subsidy of publication fees, social welfare is ∫ [tu ( x ( p ( t ))) + w ( x ( p ( t ))) – ( f + cx ( p ( t ))) – d ] dF ( t , d ). N N N (21) {( t , d )| U ( t ) ≥ d } H ere too social welfare must be positive when the author writes and pub- lishes, for U ( t ) ≥ d implies that U ( t ) ≥ d . N C Under the regime of copyright and payment of f as a publication fee, social welfare is ∫[ tu ( x ( p ( t ))) + w ( x ( p ( t ))) – ( f + cx ( p ( t ))) – d ] dF ( t , d ). N N N (22) {( t , d )| U ( t ) ≥ d } NS I n this case, however, social welfare is not necessarily positive when authors write and publish. For example, suppose that c = 0. Then the author will write and publish if tu ( x (0)) > d , whereas it is socially desirable to do only if tu ( x(0)) + w ( x(0)) > d +f. The former may hold even though the latter does not. The problem here is that the author does not incur the publication cost in deciding whether to write and publish. N ow I wish to explain why any of the three regimes might be best. To do so, it will be useful to make reference to Figures 1 and 2, which have been explained by what has been discussed above, except for t in Figur e 2. The point t is deﬁ ned to be that where U ( t ) and U ( t ) intersect, which we 0 NS C know must occur between 0 and t because U (0) < U (0) and U ( t ) > 2 NS C NS 2 U ( t ). T o show that t < t , observe that for any t in ( t , t ), we have U ( t ) ≤ C 2 0 1 1 2 C tu ( x ( t )) < tu ( x ( c )) = U ( t ). NS First, copyright might be the best of the regimes. Consider a t < t in Figure 2, that is, in region C. Here U ( t) e xceeds both U ( t) and U ( t ). C N NS 97 This follows because, for any x x x w(x)= ∫ w'(z)dz > ∫ w'(x)dz = px. 0 0 98 The point of intersection need not be unique, but taking that possibility into account would be tedious and would not affect the conclusions. Hence, I will assume above that it is unique. 179 350 ~ Shavell: Should Copyright Of Academic Works Be Abolished? Hence, there exists a d suc h that U ( t) > d > U ( t) and U ( t) > d > U ( t ). C N C NS For this ( t , d), social welfare is positive under copyright, for the author would write and publish (see the discussion following (20)), whereas the author would not be induced to write under either of the no copyright regimes. Since the distribution over t and d might be discrete—just the ( t , d ) that was discussed could be the only pair—copyright might be the best regime. Second, no copyright might be the best regime. Consider t < t in Figure a 2 2, that is, in region A. Choose d such that U ( t ) > d , so that the author will a N a a elect to write and publish. The author will also choose to write and publish under the other two regimes, as utility under them exceeds U ( t ) in region A. Social welfare at ( t , d ) under no copyright is higher than under copy- a a right because the price is lower; the price is c rather than p ( t ), and thus readership is higher. In particular, the difference in social welfare is [ t u ( x ( c )) + w ( x ( c)) – cx ( c)] – [ t u ( x ( p ( t ))) + w ( x ( p ( t )) – cx ( p ( t ))], which is posi- a a a a tive because social welfare (4) is concave in x and, from (5), is maximized at x > x ( c ). Social welfare at ( t , d ) under no copyright is the same as under no a a copyright/subsidy. Let t > t and c hoose d suc h that U ( t ) > d and social b 0 b NS b b welfare is negative as a result of the author writing and publishing (along the lines explained after (21)). The author would not write under copyright or no copyright, for as noted, under these regimes, an author chooses to write only if that would result in positive social welfare. If we have a discrete distri- bution, with some probability mass at ( t , d ) and the rest at ( t , d ) then no a a b b copyright is best: at ( t , d ), no copyright is tied with no copyright/subsidy, a a and no copyright is superior to copyright; at ( t , d ), no copyright is superior b b to no copyright/subsidy, and no copyright is tied with copyright. Thir d, no copyright/subsidy might be the best regime. That would be so if social welfare would be positive if the author would write under no copyright/subsidy but the author would not write under either of the other two regimes. This situation is possible. Let t > t and suppose that U ( t ) = d . Then at ( t , d ), we know that it must be strictly socially desirable for the author to write (see the discussion following (20)). Consider d + ε, where ε > 0. At d + ε, the author will not write under copyright, but it will still be strictly socially desirable to do so if ε is sufﬁ ciently small. Also, note that since we know that U ( t) > U ( t) (because t > t ), ε can NS C 0 also be chosen small enough that U ( t) > d + ε. In addition, note that NS U ( t ) ≥ U ( t). Hence, the author would publish under no copyright/ subsidy C N but not under copyright or no copyright, as claimed. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 351 Proposition 4 : (a) Social welfare under the three different regimes— copyright, no copyright, and no copyright/subsidy—are given in (20)–(22). (b) Any of the three regimes might be socially best. ■ 352 ~ Shavell: Should Copyright Of Academic Works Be Abolished? REFERENCES A ssociation of American University Presses. 2009. Electronic Publishing at University Presses. Available at http://www.aaupnet.org/resources/ electronic.html. Barnett, Stephen R., Robert A. Gorman, Marci A. Hamilton, & Lloyd Wein- reb. 1998. American Association of Law Schools, Model Author/Journal Agreement. Available at http://www.aals.org/deansmemos/98-24.html. Bar tow, Ann. 1998. Educational Fair Use in Copyright: Reclaiming the Right to Photocopy Freely. 60 U. Pitt. L. Rev . 149–230. Baser, Onur, & Elda Pema. 2003. The Return of Publications for Economics Faculty. 1(1) Econ. Bull. 1–13. Benkler, Yochai. 2006. The Wealth of Networks: How Social Production Trans- forms Markets and Freedom. New Haven: Yale University Press. Bergstrom, Carl T., & Theodore C. Bergstrom. 2004. The Costs and Beneﬁ ts of Library Site Licenses to Academic Journals. 101 Proc. Nat’l Acad. Sci. 897–902. Bergstrom, Theodore C. 2001. Free Labor for Costly Journals? 15(3) J. Econ. Persp. 183–198. Berlin Declar ation on Open Access to Knowledge in the Sciences and Humanities. 2003. Available at http://oa.mpg.de/openaccess-berlin/ berlindeclaration.html. Bethesda Statement on Open Access Publishing. 2003. Available at http:// www.earlham.edu/~peters/fos/bethesda.htm. B ioMed Central. 2009. Summary. Available at http://www.biomedcentral. com/info/authors/funderpolicies/. B ird, Claire. 2008. Oxford Journals’ Adventures in Open Access. 21 Learned Publishing 200–208. Björ k, Bo-Christer, & Turid Hedlund. 2009. Two Scenarios for How Scholarly Publishers Could Change Their Business Model to Open Access. 12(1) J. Electronic Publishing. Available at http://dx.doi. org/10.3998/3336451.0012.102. Björk, Bo-Christer, Annikki Roosr, & Mari Lauri. 2008. Global Annual Vol- ume of Peer Reviewed Scholarly Articles and the Share Available via Different Open Access Options, in Open Scholarship: Authority, Com- munity, and Sustainability in the Age of Web 2.0— Proceedings of the 12th International Conference on Electronic Publishing 178–186 (L eslie Chan & Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 353 Susanna Mornati eds.). Available at http://elpub.scix.net/data/works/ att/178_elpub2008.content.pdf. Bramble, Nicholas. 2006. Note, Preparing Academic Scholarship for an Open Access World. 20 Harv. J.L. & Tech. 209–233. B reyer, Stephen. 1970. The Uneasy Case for Copyright: A Study of Copy- right in Books, Photocopies, and Computer Programs. 84 Harv. L. Rev. 281–351. B udapest Open Access Initiative. 2002. Available at http://www.soros.org/ openaccess/read.shtml. Courant, Paul, Ann Kjellberg, J. D. McClatchy, Edward Mendelson, Margo Viscusi, Tappan Wilder, Steven Yenser, & Robert Darnton. 2009. Goog- le & Books: An Exchange. N.Y. Rev. Books, Mar. 26, 2009. Available at http://www.nybooks.com/articles/22496. Darnton, Robert. 2009. Google & the Future of Books. N.Y. Rev. Books, Feb. 12, 2009. Available at http://www.nybooks.com/articles/22281. Denicola, Robert C. 2006. Copyright and Open Access: Reconsidering Uni- versity Ownership of Faculty Research. 85 Neb. L. Rev. 351–382. Dir ectory of Open Access Journals. 2009. Available at http://www.doaj. org/. Dir kmaat, Joop. 2002. Comment, Pricing and Cost of Electronics Journals. 16(4) J. Econ. Persp . 227–230. Dryburgh Assoc., Ltd. 2002. The Costs of Learned Journal and Book Pub- lishing: A Benchmarking Study for ALPSP. The Association of Learned and Professional Society Publishers. Available at www.alsp.org.uk. Else vier. 2009. Elsevier Policy on Sponsored Articles. Available at http:// www.elsevier.com/wps/ﬁ nd/intro.cws_home/sponsoredarticles. E wing, John. 2004. The Orthodoxy of Open Access. Nature Web Focus: Access to the Literature, Sept. 13. Available at http://www.nature.com/ nature/focus/accessdebate. Fisher , Janet H. 1995. The True Costs of an Electronic Journal. 21(1) Serials Rev. 88–90. Ga y, Susan. 2007. ePublishing Trends in the Academic and Professional Book Market. 3(1) Volumes 2–4. Available at www.spi-bpo-com. Genamics Journal Seek. 2009. Available at http://journalseek.net. Goldstein, Paul. 2005. Goldstein on Copyright . 3d ed. New York: Aspen. Google Book Settlement. 2009. (Original settlement November 2008, amen- ded November 2009.) Available at http://www.googlebooksettlement. com/. 354 ~ Shavell: Should Copyright Of Academic Works Be Abolished? G reen, Toby. 2009. Book Review, Scholarly Publishing Practice 3: Academic Journals Publishers’ Policies and Practices in Online Publishing. 22(1) Learned Publishing 75–77. G uterman, Lila. 2004. The Promise and Peril of ‘Open Access.’ Chron. High- er Educ., Jan. 30, 2004, A10–A14. H armel, Leslie A. 2005. The Business and Legal Obstacles to the Open A ccess Publishing Movement for Science, Technical, and Medical Jour- nals. 17 Loy. Consumer L. Rev. 555–570. H arvard University. 2009. Harvard Open Access Policies, February 2008– June 2009. http://osc.hul.harvard.edu/OpenAccess/policytexts.php. H ess, Thomas, Rolf T. Wigand, Florian Mann, & Benedikt von Walter. 2007. Management Report 1/2007, Open Access & Science Publish- ing. Available at http://openaccess-study.com/Hess_Wigand_Mann_ Walter_2007_Open_Access_Management_Report.pdf. H oward Hughes Medical Institute. 2007. HHMI Announces New Policy for Publication of Research Article, June 26. Available at http://www.hhmi. org/news/20070626.html. Hunter, Karen. 2004. Open Access: Yes, No, Maybe. Nature , Web Focus: Access to the Literature: The Debate Continues. Available at http:// www.nature.com/nature/focus/accessdebate. I oannidis, John P. A. 2006. Concentration of the Most-Cited Papers in the Scientiﬁ c Literature: Analysis of Journal Ecosystems. 1(1) PLoS ONE e5. John Wiley & Sons, Inc. 2008. 2008 Annual Report. Available at http://www. wiley.com/legacy/annual_reports/ar_2008/. Journal of Visualized Experiments. 2009. Subscriptions FAQ. Available at http://www.jove.com/index/page.stp?name=UnderSubscription. King, Donald W. 2007. The Cost of Journal Publishing: A Literature Review and Commentary. 20(2) Learned Publishing 85–106. King, Donald W., & Frances M. Alvarado-Albertorio. 2008. Pricing and Other Means of Charging for Scholarly Journals: A Literature Review and Commentary. 21(4) Learned Publishing 248–272. K ing, Donald, & Carol Tenopir. 2004. An evidence-based assessment of the ‘author pays’ model. Nature , Web Focus: Access to the Literature. Avail- able at http://www.nature.com/nature/focus/accessdebate. Landes, William M., & Richard A. Posner. 2003. The Economic Structure of Intellectual Property Law. Cambridge, MA: Belknap Press. Lemley, Mark A. 1995. Rights of Attribution and Integrity in Online Com- munications. 1995 J. Online L . art. 2. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 355 Litman, Jessica. 2006. The Economics of Open Access Law Publishing. 10 Lewis & Clark L. Rev. 779–795. Luey, Beth. 2002. Handbook for Academic Authors . 4th ed. New York: Cam- bridge University Press. M arkwood, Priscilla. 2006. Paperless Workﬂ ows in Journal Production: A Management Perspective. 19(2) Learned Publishing 115–124. M cCabe, Mark J., & Christopher M. Snyder. 2004. The Best Business Model for Scholarly Journals: An Economist’s Perspective. Nature Web Focus: access to the literature, July 16, 2004. Available at http://www.nature. com/nature/focus/accessdebate. ———. 2007. Academic Journal Prices in a Digital Age: A Two-Sided Mar- ket Model. 7(1) Berkeley Electronic J. Econ. Analysis & Pol’y , art. 2. Avail- able at http://www.bepress.com/bejeap/vol7/iss1/art2. Menell, Peter S., & Suzanne Scotchmer. 2007. Intellectual Property Law , in Handbook of Law and Economics, ed. A. Mitchell Polinsky and Steven Shavell, 2:1473–1570. Amsterdam: North-Holland. MIT N ews. 2009. MIT Faculty Open Access to Their Scholarly Articles. Avail- able at http://web.mit.edu/newsofﬁ ce/2009/open-access-0320.html. Moore, William J., Robert J. Newman, & Geoffrey K. Turnbull. 1998. Do Academic Salaries Decline with Seniority? 16 J. Labor Econ. 352–366. National Institutes of Public Health. 2009. NIH Public Access Policy. Avail- able at http://publicaccess.nih.gov/. Nature , Web Focus. 2004. Access to the Literature: The Debate Continues. Available at http://www.nature.com/nature/focus/accessdebate/archive. html. N ature Precedings. 2009. Pre-publication Research and Preliminary Find- ings. Available at http://precedings.nature.com/. N immer, Melville B., & David Nimmer. 2008. Nimmer on Copyright. New York: M. Bender. Nucleic Acids Research. 2009. NAR’s Open Access Initiative. Avail- able at http://www.oxfordjournals.org/our_journals/nar/announce_ openaccess.html. Open Access Directory. 2009. Journals That Converted from TA to OA. Available at http://oad.simmons.edu/oadwiki/Journals_that_ converted_from_TA_to_OA. Oxford Journals. 2009. Oxford Open. A vailable at http://www .oxfordjournals. org/oxfordopen/. 356 ~ Shavell: Should Copyright Of Academic Works Be Abolished? Petrowski, Mary Jane. 2007. 2006 Academic Library Trends and Statistics. Chicago: Association of College and Research Libraries. Pind yck, Robert S., & Daniel L. Rubinfeld. 2009. Microeconomics. 7th ed. Upper Saddle River, NJ: Pearson/Prentice Hall. Plant, Arnold. 1934. The Economic Aspects of Copyright in Books. 1 Economica (n.s.) 167–195. Public Library of Science. 2001. Open Letter to Scientiﬁ c Publishers. Available at www.plos.org/about/letter.html. ——— . 2005. Press release, The First Impact for PLoS Biology (June 27). Available at http://news.google.com/news?hl=en&ned=us&q=%22PLo S+Biology%22+OR+%22Public+Library+of+Science+Biology%22&b tnG=Search+News. ——— . 2009. Why Submit Your Research to PLoS? Available at www.plos. org/journals/index.php. PubMed Central. 2009. PubMed Central (PMC) website. Available at http:// www.pubmedcentral.nih.gov/. R eed Elsevier. 2009. Annual Reports and Financial Statements 2008. Avail- able at http://www.reed-elsevier.com/annualreport08/. R owlands, Ian, & Dave Nicholas. 2005. New Journal Publishing Models: An International Survey of Senior Researchers. Available at http://www.ucl. ac.uk/ciber/ciber_2005_survey_ﬁ nal.pdf. Schroter, Sara, & Leanne Tite. 2006. Open Access Publishing and Author- Pays Business Models: A Survey of Authors’ Knowledge and Perceptions. 99 J. Royal Soc’y Med. 141–148. Scientiﬁ c Council of the European Research Council. 2007. ERC Scientiﬁ c Council Guidelines for Open Access. Available at http://erc.europa.eu/ pdf/ScC_Guidelines_Open_Access_revised_Dec07_ FINAL.pdf. Shaw, Ralph R. 1951. Copyright and the Right to Credit. 113 Science 571–573. SHERP A RoMEO. 2009a. Journal Policies—Summary Statistics So Far. Available at http://romeo.eprints.org/stats.php. ——— . 2009b. Publishers with Paid Options for Open Access. Available at http://www.sherpa.ac.uk/romeo/PaidOA.html. Shieber, Stuart M. 2009. Equity for Open-Access Journal Publishing. 7(8) PLoS Biology 1–3. Available at http://dx.doi.org/10.1371%2Fjournal. pbio.1000165. Springer. 2008. Springer Open Choice. Available at http://www.springer. com/authors/mypublication?SGWID=0-145303-6-572911-0. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 357 Stanfor d University. 2009. HighWire: Free Online Full-text Articles. Avail- able at http://highwire.stanford.edu/lists/freeart.dtl. Stanfor d University, School of Education. 2008. SUSE Open Access Motion. Available at http://ed.stanford.edu/suse/faculty/openaccess.html. St eele, Colin. 2008. Scholarly Monograph Publishing in the 21st Century: The Future More Than Ever Should Be an Open Book. 11(2) J. Electron- ic Publishing . Available at http://dx.doi.org/10.3998/3336451.0011.201. Suber, Peter. 2007. Open Access Overview. Available at http://www.earlham. edu/~peters/fos/overview.htm. ——— . 2009a. Open Access in 2008. 12(1) J. Electronic Publishing . Avail- able at http://dx.doi.org/10.3998/3336451.0012.104. ——— . 2009b. Timeline of the Open Access Movement. Available at http:// www.earlham.edu/~peters/fos/timeline.htm. S uthersanen, Uma. 2003. Copyright and Educational Policies: A Stakehold- er Analysis. 23 Oxford J. Legal Stud. 585–609. Symposium. 2006. Open Access Publishing and the Future of Legal Schol- arship. 10 Lewis & Clark L. Rev. 733–924. Taylor & Francis. 2009. iOpenAccess & NIH Policy. Available at http://www. tandf.co.uk/journals/iopenaccess.asp. Tenopir, Carol, & Donald W. King. 2000. Towards Electronic Journals: Re- alities for Scientists, Librarians, and Publishers. Washington, DC: Special Libraries Association. That cher, Sanford G. 2008. On the Author’s Addendum. 40 J. Scholarly Pub- lishing 97–115. ——— . 2009. From the University Presses—The Hidden Digital Revolu- tion in Scholarly Publishing: POD, SRDP, the “Long Tail,” and Open Ac- cess. 21(2) Against the Grain . Available at http://www.against-the-grain. com. Thompson, John B. 2005. Books in the Digital Age: The Transformation of Academic and Higher Education Publishing in Britain and the United States. Malden, MA: Polity. Universities UK and the Research Information Network. 2009. Paying for Open Access publication charges. Available at http://www.rin.ac.uk/ openaccess-payment-fees. W ellcome Trust. 2004. Costs and Business Models in Scientiﬁ c Research Publishing. Wiley InterScience. 2009. OnlineOpen. Available at http://www3.interscience. wiley.com/authorresources/onlineopen.html. 358 ~ Shavell: Should Copyright Of Academic Works Be Abolished? W illinsky, John. 2004. Scholarly Associations and the Economic Viability of Open Access Publishing. 4(2) J. Digital Info. Available at http://journals. tdl.org/jodi/article/viewArticle/104. ———. 2005. The Access Principle: The Case for Open Access to Research and Scholarship. Cambridge, MA: MIT Press. Wolters Kluwer. 2009. 2008 Annual Report. Available at http://www. wolterskluwer.com/2008reports/pdf/2008_Annual_Report_Wolters_ Kluwer.pdf. Wysocki, Bernard Jr. 2005. Peer Pressure: Scholarly Journals’ Premier Status Is Diluted by Web — More Research Is Free Online Amid Spurt of Start- Ups; Publishers’ Proﬁ ts at Risk — A Revolt on UC’s Campuses. Wall St. J., May 23, A1. Y oung, Jeffrey R. 2002. Journal Boycott over Online Access Is a Bust. Chron. Higher Educ., May 31, A34. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Legal Analysis Oxford University Press http://www.deepdyve.com/lp/oxford-university-press/should-copyright-of-academic-works-be-abolished-uTE02TT8mK

Loading next page...

References (72)

C. Steele (2008)
Scholarly monograph publishing in the 21st century: The future more than ever should be an open book
Journal of Electronic Publishing, 11
Sanford Thatcher (2008)
On the Author's Addendum
Journal of Scholarly Publishing, 40
K. Kelley (2008)
NIH public access policy
Brain, Behavior, and Immunity, 22
(2009)
Electronic Publishing at University Presses
R. Shaw (1951)
Copyright and the right to credit.
Science, 113 2942
(2007)
� OnlineOpen
W. Landes, R. Posner (2003)
The Economic Structure of Intellectual Property Law
L. Harmel (2005)
The Business and Legal Obstacles to the Open Access Publishing Movement for Science, Technical, and Medical Journals
Loyola Consumer Law Review, 17
U. Suthersanen (2003)
Copyright and Educational Policies: A Stakeholder Analysis
Oxford Journal of Legal Studies, 23
(2009)
NAR’s Open Access Initiative
(2007)
HHMI Announces New Policy for Publication of Research Article, June 26
Arnold Plant (1934)
The Economic Aspects of Copyright in Books
Economica, 1
(1998)
American Association of Law Schools, Model Author/Journal Agreement
(2009)
Subscriptions FAQ
J. Devine (2010)
Annual reports and financial statements
(2009)
Google & the Future of Books
(2009)
Journals That Converted from TA to OA
(2005)
Peer Pressure: Scholarly Journals’ Premier Status Is Diluted by Web — More Research Is Free Online Amid Spurt of StartUps; Publishers
(2004)
The Best Business Model for Scholarly Journals: An Economist’s Perspective
For a partial list of journals offering this option and their prices, see SHERPA RoMEO, Publishers with Paid Options for Open Access
S. Breyer (1970)
The Uneasy Case for Copyright: A Study of Copyright in Books, Photocopies, and Computer Programs
Harvard Law Review, 84
James Walther (2001)
Towards Electronic Journals: Realities for Scientists, Librarians and Publishers
(2004)
The Promise and Peril of ‘Open Access.
(2009)
Oxford Open
Peter Menell, Suzanne Scotchmer (2007)
Intellectual Property Law
, 2
(2009)
Elsevier Policy on Sponsored Articles
C. Bird (2008)
Oxford Journals' adventures in open access
Learned Publishing, 21
(2002)
Journal Boycott over Online Access Is a Bust
(2002)
The Costs of Learned Journal and Book Publishing: A Benchmarking Study for ALPSP
(2009)
Pre-publication Research and Preliminary Findings
(2009)
PubMed Central (PMC) website
(2009)
MIT Faculty Open Access to Their Scholarly Articles
T. Green (2009)
Scholarly Publishing Practice 3: Academic Journals Publishers' Policies and Practices in Online Publishing
Learn. Publ., 22
so that fees exceed f and rise with t. Let the utility of an author (exclusive of writing cost) who publishes and has chosen his or her
P. Courant (2009)
Google & books: An exchange
New York Review of Books, 56
(1995)
Rights of Attribution and Integrity in Online Communications
M. Nimmer, D. Nimmer (1963)
Nimmer on Copyright
(2009)
2008 Annual Report
Carl Bergstrom, T. Bergstrom (2004)
The costs and benefits of library site licenses to academic journals.
Proceedings of the National Academy of Sciences of the United States of America, 101 3
J. Thompson (2005)
Books In The Digital Age: The Transformation Of Academic And Higher Education Publishing In Britain And The United States
Ann Bartow (2004)
Educational Fair Use in Copyright: Reclaiming the Right to Photocopy Freely
University of Pittsburgh Law Review
Priscilla Markwood (2006)
Paperless workflows in journal production: a management perspective
Learned Publishing, 19
D. King, C. Tenopir (2004)
An Evidence-based Assessment of the "Author Pays" Model
(2008)
Springer Open Choice
(2009)
Harvard Open Access Policies, February 2008– June 2009
D. King (2007)
The cost of journal publishing: a literature review and commentary
Learned Publishing, 20
(2009)
Journal Policies—Summary Statistics So Far
Robert Denicola (2006)
Copyright and Open Access: Reconsidering University Ownership of Faculty Research
Nebraska law review, 85
Any of the three regimes might be socially best
Jessica Litman (2006)
The Economics of Open-Access Law Publishing
IO: Productivity
(2007)
ePublishing Trends in the Academic and Professional Book Market
B. Björk, A. Roos, Mariliis Lauri (2008)
Global annual volume of peer reviewed scholarly articles and the share available via different Open Access options
O. Baser, Elda Pema (2003)
The return of publications for economics faculty
Economics Bulletin, 1
Y. Benkler (2006)
The wealth of networks: how social production transforms markets and freedom
B. Luey (2009)
Handbook for Academic Authors: Frontmatter
(2004)
Open Access: Yes, No, Maybe
Rohit Jamdar, Kirti Jadhav, Megha Jadhav, Monika Jadhav, Mayur Ghotkar, Kaustubh Mane, G. Vambhurkar, S. Mohite (2005)
PHYTOCHEMICAL SCREENING AND IN-VITRO EVALUATION FOR ANTIMICROBIAL ACTIVITY OF BAUHINIA PURPUREA L. ETAHNOLIC EXTRACT
, 2020
Bladimir Borges (2008)
Berlin Declaration on Open Access to Knowledge in the Sciences and Humanities
, 4
(2009)
Paying for Open Access publication charges
Journal of Economic Perspectives—Volume 16, Number 4—Fall 2002—Pages 227–238 Comments Pricing and Cost of Electronics Journals
I. Rowlands, D. Nicholas (2005)
NEW JOURNAL PUBLISHING MODELS: AN INTERNATIONAL SURVEY OF SENIOR RESEARCHERS
D. King, Frances Alvarado-Albertorio (2008)
Pricing and other means of charging for scholarly journals: a literature review and commentary
Learned Publishing, 21
S. Shieber (2009)
Equity for Open-Access Journal Publishing
PLoS Biology, 7
(2006)
Open Access Publishing and the Future of Legal Scholarship
(2007)
Open Access Overview. Available at http://www.earlham
S. Schroter, L. Tite (2006)
Open access publishing and author-pays business models: a survey of authors' knowledge and perceptions.
Journal of the Royal Society of Medicine, 99 3
W. Moore, R. Newman, G. Turnbull (1998)
Do Academic Salaries Decline with Seniority?
Journal of Labor Economics, 16
B. Björk, Turid Hedlund (2009)
Two Scenarios for How Scholarly Publishers Could Change Their Business Model to Open Access
Journal of Electronic Publishing, 12
E. Zerhouni (2004)
NIH Public Access Policy
Science, 306
N. Bramble (2006)
Preparing Academic Scholarship for an Open Access World
Harvard Journal of Law & Technology, 20
(2001)
Free Labor for Costly Journals? 15(3) J
(2009)
HighWire: Free Online Full-text Articles

Publisher: Oxford University Press
Copyright: © Published by Oxford University Press.
Subject: Articles
ISSN: 2161-7201
eISSN: 1946-5319
DOI: 10.1093/jla/2.1.301
Publisher site: See Article on Publisher Site

Abstract

SHOULD COPYRIGHT OF ACADEMIC WORKS BE ABOLISHED? Steven Shavell ABSTRACT T he conventional rationale for copyright of written works, that copyright is needed to foster their creation, is seemingly of limited applicability to the academic do- main. For in a world without copyright of academic writing, academics would still beneﬁ t from publishing in the major way that they do now, namely, from gaining scholarly esteem. Yet publishers would presumably have to impose fees on au- thors because publishers would no longer be able to proﬁ t from reader charges. If these author publication fees would actually be borne by academics, their incen- tives to publish would be reduced. But if the publication fees would usually be paid by universities or grantors, the motive of academics to publish would be unlikely to decrease (and could actually increase)—suggesting that ending academic copy- right would be socially desirable in view of the broad beneﬁ ts of a copyright-free world. If so, the demise of academic copyright should probably be achieved by a change in law, for the “open access” movement that effectively seeks this objective without modiﬁ cation of the law faces fundamental difﬁ culties. 1. INTRODUCTION Or iginal written works, and thus original written academic works, may 1 generally be copyrighted. However, the conventional rationale for copyright 1 Sam uel R. Rosenthal Professor of Law and Economics, Harvard Law School. I thank Yochai Benkler, Michael Fisher, Paul Goldstein, Stevan Harnad, Scott Hemphill, Louis Kaplow, Mark Lemley, Josh Lerner, Douglas Lichtman, Peter Menell, Proshanto Mukherji, Dotan Oliar, John Palfrey, Richard Posner, Stuart Shieber, Christopher Snyder, Peter Suber, and Jeannie Suk for comments, Jonathan Cooper, Steven Horowitz, James Kwok, Vincent Leow, and Stephanie Wiener for research assistance, and the John M. Olin Center for Law, Economics, and Business at Harvard Law School for research support. 2 See 17 U .S.C. § 102(a); Melville B. Nimmer & David Nimmer (2008, vol. 1 §§ 2.01–2.04). Moreover, academics themselves, rather than their universities, own the copyrights in their works. See Hays v. Sony Corp. of Am., 847 F.2d 412, 416 (7th Cir. 1988) (Posner, J.), noting that “virtually no one questioned that the academic author was entitled to copyright his writings” despite the work for hire doctrine, under which the academic’s employer would be entitled to copyright; Weinstein v. Univ. of Ill., 811 F.2d 1091, 1094 (7th Cir. 1987) (Easterbrook, J.), not- ing that “faculty own the copyrights in their academic work” regardless of the work for hire doctrine; William M. Landes & Richard A. Posner (2003, 272), “It is generally and we think correctly understood that academics, although employees of the university in which they teach and conduct their research, own the copyrights on their academic books and articles.” 302 ~ Shavell: Should Copyright Of Academic Works Be Abolished? of written works—that it stimulates their creation and publication by allow- ing authors to proﬁ t from their sale —is seeming ly of limited applicability to academic authors. The primary spur for academic writing ostensibly is not proﬁ t but rather scholarly esteem and professional advancement. Hence, one might be led to believe that copyright is not needed to encour- age academic writing. This view, though, is incomplete, because it does not reﬂ ect the point that if publishers were unable to beneﬁ t from copyright earnings, they would apparently have to impose fees on authors to cover their publishing costs. The possible effects of such publication fees must be evaluated along with other relevant factors to determine whether the beneﬁ ts of a copyright-free academic world make it socially preferable to our own. 2 I att empt to make this assessment here, beginning in Section 2 with an informal consideration of a stylized model of academic authorship and publication (the model is presented formally in an appendix), the major assumption of which is that academic authors desire readership of their works to gain esteem. I interpret the model in Section 3, taking into account a variety of issues that are not included in it. The analysis and discussion in these sections may be summarized as follows. 3 First, publication of academic works does indeed carry signiﬁ cant beneﬁ ts of recognition and of professional reward for academics that are independent of the existence of copyright; and copyright itself provides academics with little remuneration—academics generally do not earn any royalties from articles and only modest royalties from books. Hence, the afﬁ rmative motives of academics to publish would be little diminished in the absence of copyright. 4 Second, publishers would have to recover at least the ﬁ rst copy costs of selecting and preparing works for publication from a source other than 3 See U.S. Const. art. I, § 8, cl. 8, “The Congress shall have Power . . . To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries”; Eldred v. Ashcroft, 537 U.S. 186, 212 n.18 (2003), “copyright law celebrates the pr oﬁ t motive, recognizing that the incentive to proﬁ t from the exploitation of copyrights will redound to the public beneﬁ t by resulting in the prolifera- tion of knowledge . . . . The proﬁ t motive is the engine that ensures the progress of science.” (quoting Am. Geophysical Union v. Texaco, Inc., 802 F. Supp. 1, 27 (S.D.N.Y. 1992)); Mazer v. Stein, 347 U.S. 201, 219 (1954), “The economic philosophy behind the clause empowering Congress to grant patents and copyrights is the conviction that encouragement of individual effort by personal gain is the best way to advance welfare through the talents of authors and inventors in ‘Science and the useful Arts.’” See also P aul Goldstein (2005, vol. 1 §§ 1:41–:44); Peter S. Menell & Suzanne Scotchmer (2007, 1476–1482). Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 303 readers were there no copyright. Thus, as indicated, publishers would have to impose fees for publishing. If the fees incorporated ﬁ rst copy costs, they might lie in the range of one to several thousand dollars for an article and would be signiﬁ cantly higher for a book. Thir d, if academics would have to bear the publication fees themselves, 5 their incentives to publish would fall, possibly non-negligibly in view of the potential magnitude of the fees. Still, the social losses associated with such a decline in publications would be tempered by two considerations: discouraged publications would tend to be of lower than average quality; and written but unpublished works could still be posted on the Internet, so presumably would be available to readers in that form. Fourth, if academics would not have to bear publication fees—on the 6 assumption that universities or grantors would subsidize them—the incen- tive of academics to write and to publish articles would not fall and could rise. The reason that the elimination of copyright could, perhaps paradoxi- cally, augment incentives to publish articles is that readership of articles would grow in the absence of copyright, and thus the esteem that authors would derive from publication would tend to increase. The incentive of academics to write and to publish books, however, would not necessarily be promoted, because authors would forgo royalties. Yet since royalties are generally low, it is possible that incentives to write and publish books would also be enhanced. An enhancement of incentives to write due to the subsidy of publication fees is not necessarily socially desirable, though, because it could lead to the writing of works that are not socially valuable on net in view of the costs of publication—the subsidy could lead to a socially ex- cessive level of publication. This problem, however, could be addressed by conditioning subsidies on the quality of publication venues. Fifth, universities and grantors would have a motive to subsidize publi- 7 cation fees in the absence of copyright, for it would be in their interest to prevent a dilution in the incentives of faculty members and of grantees to write and publish works, at least works of reasonable quality. Furthermore, universities should not face great difﬁ culties in ﬁ nancing publication fees, for they would no longer have to purchase journal subscriptions or books. Finally, substantial social beneﬁ ts would ﬂ ow from a copyright-free 8 world. Unfettered access to academic works would mean that all works would become available on the Internet for any person to download for free, that print versions could be made at cost, and that the assembly of teaching materials would no longer be burdened by the task of securing 304 ~ Shavell: Should Copyright Of Academic Works Be Abolished? permissions from copyright holders. Further, the various legal and other costs of protecting copyright privileges would be avoided. 9 The conclusion that I draw from the foregoing is that if publication fees would be largely defrayed by universities and grantors, as I suggest would be to their advantage, then the elimination of copyright of academic works would be likely to be socially desirable: it would not compromise academic publication activity and would yield the social beneﬁ ts of a copyright-free regime. On the other hand, if publication fees would not be broadly subsi- dized, whether ending academic copyright would be socially advantageous is unclear. 10 I n Section 4, I examine the “open access” movement, which holds that academic publications should be costlessly obtainable from the Internet. Although open access initiatives have grown in recent years, they have also faced difﬁ culties. Namely, traditional journals are generally unwilling to grant immediate open access to published works, and most authors w ould not ﬁ nd it in their personal interest to publish in open access journals that sometimes charge publication fees rather than in the often more prestigious traditional journals that do not charge publication fees. Moreover, there is a fundamental divergence between individual and collective interests af- ﬂ icting the open access movement: if an individual academic publishes his or her own work as an open access article or book, the academic does not thereby obtain the broad social beneﬁ ts that would ﬂ ow from the general disappearance of academic copyright. This individual-versus-social incen- tive problem suggests that open access efforts to cabin copyright-associated charges for access will not succeed fully, or will do so only after a substan- tial delay. An outright change in law to terminate academic copyright is therefore likely the best policy if academic copyright is judged to be socially undesirable. 4 See, e.g., Berlin Declaration on Open Access to Knowledge in the Sciences and Humanities (2003); Bethesda Statement on Open Access Publishing (2003); Budapest Open Access Ini- tiative (2002), deﬁ ning open access as “the world-wide electronic distribution of the peer- reviewed journal literature and completely free and unrestricted access to it by all scientists, scholars, teachers, students, and other curious minds.” See generally John Willinsky (2005); Symposium (2006); Nicholas Bramble (2006); Nature, Web Focus (2004); Peter Suber (2007). These sources on open access initiatives emphasize journal literature, not academic books. 5 However, as I discuss in Section 4.1, traditional journals that hold copyright in their ar- ticles often do allow so-called “green” open access, that is, immediate free access to various pre-publication versions of articles and sometimes delayed free access to published versions. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 305 I n Section 5, I discuss how the elimination of academic copyright could be implemented and whether there would be serious legal or political dif- ﬁ culties in so doing. Befor e proceeding, let me note that I do not seek any striking novelty in 12 this article. That academics are motivated to write to gain scholarly respect, and therefore that copyright may not be required to encourage publication of academic works, has often been mentioned before. The contribution made here lies mainly in the development of this observation employ- ing a model in which explicit account is taken of the desire of academ- ics for esteem and of the publication fees that would likely emerge absent copyright; in the analysis suggesting that if universities or grantors would bear the publication fees, incentives to publish would not fall and might rise, and thus that abolishing copyright might well be socially desirable; and in the argument that the open access movement confronts serious problems in achieving its end, probably requiring a change in law to assure that outcome. 6 B ut as I note there, the author would possess the right to require proper attribution of his or her work and possibly certain ‘moral rights’, such as the right to prevent problematic abridgment of a work. 7 See, e.g. , Princeton Univ. Press v. Mich. Document Servs., 99 F.3d 1381, 1410 (6th Cir. 1996) (en banc) (Ryan, J., dissenting), “More than one hundred [academic] authors declared on the record that they write for professional and personal reasons such as making a contribution to a particular discipline, providing an opportunity for colleagues to evaluate and critique the authors’ ideas and theories, enhancing the authors’ professional reputations, and improving career opportunities. These declarants stated that the receipt of immediate monetary compen- sation such as a share of licensing fees is not their primary incentive to write.”; Yochai Benkler (2006, 47–48), describing information production in which authors produce a work for status or reputation beneﬁ ts; Carol Tenopir & Donald W. King (2000, 146–149), surveying scholar- ship on scholarly publication motivation and concluding that “[t]he two primary motivating factors of scholarly authors are (1) recognition for career advancement . . . , and (2) the desire to contribute to the body of knowledge in a ﬁ eld or to the archive of the scholarly knowledge in a ﬁ eld and to be recognized for their contribution by their peers”; Jessica Litman (2006, 782), commenting that authors are not motivated “by the incentives supplied by the copyright sys- tem”; Arnold Plant (1934, 169), observing that many authors, especially academics, write “to secure publicity” rather than to reap monetary reward, and thus that copyright is not needed to induce them to create new works; Uma Suthersanen (2003, 602), stating that one of the pri- mary purposes of writing journal articles is “achieving recognition or promotion”; Budapest Open Access Initiative (2002), referring to “the willingness of scientists and scholars to publish the fruits of their research in scholarly journals without payment.” 8 Ther e are, however, models of academic publication recognizing authors’ desire for readership, but assuming a regime of copyright (rather than comparing that regime to a regime without copyright). See, e.g., Mark J. McCabe & Christopher M. Snyder (2007). 306 ~ Shavell: Should Copyright Of Academic Works Be Abolished? 2. COP YRIGHT OF AC ADEMIC WORK S IN A S TYLIZED MODEL 13 I consider here the functioning of academic copyright and the possible so- cial desirability of its elimination in a stylized model of academic author- ship and publication. Because the purpose of the model is to elucidate the central effects of copyright on academic writing, I abstract from a variety of factors that will be of signiﬁ cance when I attempt to come to a realistic assessment of academic copyright in the next section. That section will be better appreciated by the reader in light of the analysis of the model, but he or she can still pass over this section without great loss of continuity. The crucial assumption of the model is that the utility of academic au- thors is increased by the readership of their published works because they gain esteem from readership. For simplicity, I suppose that an author’s util- ity from readership depends only on the number of readers (and thus, not also on the status of the readers or on the reputation of a publisher or a journal). I let a parameter t denot e the degree to which an author’s util- ity is raised by readership of a work. If t is small, the author cares slightly about readership, whereas if t is high, the author cares substantially about readership. I suppose t oo that an author’s utility depends on the conventional factors of income and expense and of writing effort. I assume that an author m ust devote positive effort and incur disutility to write a work and that its publication involves two types of cost for pub- lishers: an initial ﬁ xed, or “ﬁ rst copy” cost f, for reviewing, editing, proof- reading, and the like; and a per unit or marginal cost c of dissemination, such as the per article cost of printing a copy of a journal and of mailing it. This per unit c ost could be almost zero if the work is made available for downloading on an Internet site, an important case that I will discuss in the next section. 9 If the author obtained a beneﬁ t from writing rather than incurred disutility from doing so, the qualitative nature of the conclusions I would reach would not change. The reason is that the author who enjoys writing itself might still be discouraged from publishing in the absence of copyright by having to pay a fee in order to publish. 10 See Theodor e C. Bergstrom (2001, 187), “The costs of publishing a journal can be usefully partitioned into ﬁ rst copy costs and marginal subscriber costs .”; Donald W. King (2007, 94–96), discussing concepts of cost in journal publishing. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 307 The deﬁ nition of social welfare that I consider is a version of the standard utilitarian one, namely, the beneﬁ ts obtained from the publication of writ- ten works—both the beneﬁ ts obtained by readers and the utility obtained by authors because of the utility they derive from their readership – minus the costs of writing effort and of publication. 2.1. Copyright Let me now discuss the nature of the contract that an academic who has writ- 18 ten a work would rationally make with a publisher in a regime with copyright. (I will discuss later whether an academic would be motivated to write the work in the ﬁ rst place.) I will suppose that the contract involves two terms: a price at which the author’s work will be sold to readers in the marketplace; and a pay- ment term—either a royalty payment made by the publisher to the author, or possibly a publication fee paid by the author to the publisher. In some respects, the reader might ﬁ nd it convenient to suppose that the author’s work is a book rather than a journal article, for an author of a book would be able to bargain about the price at which it would sell, but the author of an article would not be thought to be able to bargain individually about the price at which a journal sells. N evertheless, the average desires of authors about journal prices should be reﬂ ected in journal prices, so that the model can be interpreted as applying to authors’ contracts with journals in an approximate sense. I n addition, I assume for simplicity that there are many competing pub- 19 lishers, implying that publishers will earn only enough to cover their costs. In other words, any publisher revenue from sales of a work in excess of pub- lisher costs will be paid to the author as a royalty, or any shortfall of sales from publisher costs will be paid by the author to the publisher as a fee. The author w ill choose a contract that maximizes his or her utility. That 20 is, among the possible contracts that allow publishers to cover their costs, the author will select the particular contract that makes him or her best off. The nature of this contract will depend on the degree t t o which the author cares about readership, as is illustrated in Figure 1 below. 11 F or a model in which journals rather than books are the venues for publication, see McCabe & Snyder (2007). It would be distracting from my central purpose—the comparison of a regime of copyright to a regime of no copyright—to have to take account of the complexities of such a model, rather than to address them informally, as I do, in the next sections. 12 This assumption does not affect the qualitative conclusions that I reach. In fact, many journals and publishers of books are of course proﬁ table, as I note in Section 5.2. To the degree that proﬁ ts are positive under copyright and thus that prices are higher than I assume in the model, the advantage of abolition of copyright would be enhanced. 308 ~ Shavell: Should Copyright Of Academic Works Be Abolished? Figure 1. Contract Prices Monopoly price p ($50) Price under copyright Price in the absence of copyright Per unit cost c ($2) t t t 1 2 3 Author’s degree of care t about readership T o explain, consider ﬁ rst an author for whom t is 0, in other words, the author is motivated to publish entirely by royalty payments and does not care about the level of readership per se. Such an author would choose the sales price that maximizes proﬁ t, the traditional monopoly price of eco- nomics. This price is labeled p in the graph, and for illustration, let us assume that it is $50. If t is positive, so that the author does care about read- ership, the author will generally want to increase readership from the level that would result if the monopoly price were charged. Thus, the author will want to set the price charged below the monopoly price even though this will mean some sacriﬁ ce of proﬁ ts and royalty payments. For example, sup- pose that by reducing the price from $50 to $49, readership would increase by 100 individuals and proﬁ ts and hence royalty payments would fall by $200. If the utility value to the author of having an additional 100 readers exceeds $200, the author would prefer the lower $49 price to the $50 mo- nopoly price despite the reduction in royalty payments. The best price for the academic author who cares about readership can be shown always to be less than the monopoly price, and is lo wer the greater is the degree t t o which 13 The monopoly pr ice is such that the marginal revenue equals marginal cost. See , e.g., Robert S. Pindyck & Daniel L. Rubinfeld (2009, 330). 14 Because the monopoly price maximizes proﬁ ts, a slight variation in the price would leave prof- its essentially unchanged. Hence, the effect of a small decrease in the price would be to increase readership with only a negligible lowering of proﬁ ts; and since the author values readership, the author would beneﬁ t thereby. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 309 the author cares about readership. This is why the graph of the price of the published work that the academic author would choose declines as t rises. The degree of care about readership t displayed in Figure 1 is the t at 22 which the price chosen by the author would yield revenues just equal to the publisher’s total costs, implying that the author would obtain no royalty payment from the publisher. Be yond t , the price favored by the author would not yield revenues sufﬁ cient to cover the publisher’s costs and would result in the author having to make a payment to the publisher. That is, if the degree to which an author cares about readership is at least t , the author will value readership enough that he or she will pay a publication fee to aug- ment readership by lowering the sales price substantially, even though the author could contract for a higher sales price and not have to pay to publish. At t authors w ould pay a publication fee of f in or der to lower the price to c . Indeed, if authors care enough about readership, if their t is at least t , they would give their work away for free—possibly by mounting it on the Internet and letting it be downloaded gratis—and have to pay a fee equal to the entire publisher costs. That corresponds to a price of zero on the graph. I n interpreting the graph, the reader might note that it is common 23 for an academic author of a book to negotiate with the publisher for the selling price to be lower than the publisher would like, because, as is con- sistent with the analysis, the author often wants larger readership rather than higher income from royalties. Also reﬂ ecting the graph and the point that authors value readership as well as not having to make pay- ments is that authors of journal articles generally do not collect royalties 15 Note that the graph shows that the price at t exceeds the unit cost c even though no proﬁ ts are made. The explanation is that the apparent proﬁ t of t – c on the units that are sold just offset the ﬁ rst copy costs f . 16 I do not consider negative prices—the possibility that authors would pay individuals to read their works. 17 See Am. Geophysical Union v. Texaco Inc., 802 F. Supp. 1, 27 (S.D.N.Y. 1992) (Leval, J.), “gener- ally [academic] authors have a far greater interest in the wide dissemination of their works than in royalties,” aff ’d , 60 F.3d 913 (2d Cir. 1994); Ralph R. Shaw (1951, 572), “The commer- cial publisher’s interest in publications is the sale of copies of the entire issue or volume for proﬁ t. If he does not sell copies at a proﬁ t, he will soon be a bankrupt ex-publisher. The author, however, may be interested in the widest possible dissemination of his writings, and if some- one were willing to reprint 10,000 copies of his article for free distribution, that would provide a great additional proﬁ t to the author in terms of professional credit.” I myself have bargained with a university press about the price of a book that I authored. I was urging a lower price for the book (and earlier release of an inexpensive paperback version) than the press was inclined to accept because of its interest in revenues and proﬁ ts. 310 ~ Shavell: Should Copyright Of Academic Works Be Abolished? Figure 2. Utility from Contracts Utility in absence of copyright, given subsidy Utility under copyright Utility in absence of copyright, no subsidy Utility AB C D 0 t t 0 2 Author’s degree of care t about readership and that journals retain copyright and charge prices to readers that allow them revenues sufﬁ cient not to have to impose fees for publishing on authors. 24 The utilit y that an author would obtain from publishing employing his or her preferred contract is displayed by the solid graph in Figure 2. (The other graphs will be discussed subsequently.) The author’s utility increases the higher is his or her t, even though the sales price charged falls, and even though, beyond t , authors pay a positive fee to have their works published. The argument explaining that the utility of authors must grow with t is essentially this. Compare an author with one level of t , say t , to an author with a higher level of t , say t . The t author could always make precisely H H the contract that the lower t author makes – name the same price and col- lect the same royalty (or pay the same fee). If the t author did this, the t H H author would be happier than the t author because, by hypothesis, the t L H author obtains more utility from having readers than the t author. In fact, the t author w ill obtain even greater utility than the amount just men- tioned, since the t author will prefer to choose a different contract with a lower sales price from that chosen by the t author. Hence, the t author L H must be better off, as claimed. 18 See the discussion in Section 3.1 below. 19 These levels of t are not shown in Figure 2. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 311 The solid graph in Figure 2 determines whether an author would want to invest the effort to create a work in the ﬁ rst place under copyright. In particular, an author would write a work if and only if the utility gained from his or her preferred contract would exceed the effort cost. In general, I assume that the effort costs vary for authors of each t , so that some authors of each type would choose to write and some would not. 2.2. No Copyright N ext let me consider the situation in a regime without copyright. In this 26 context, a contract between an author of a work and a publisher cannot specify a sales price exceeding the per unit cost c . The reason, of course, is competition: if the publisher named a higher price than c , it would sell nothing, as other publishers could copy the work and could sell at a lower price as their unit cost would be only c . Because, then, the publisher con- tracting with the author could only obtain a price of c for the work, the publisher would need to charge a publication fee to recoup the ﬁ rst copy cost f in order to cover its costs and to be willing to publish. Publication fees are borne by authors . Assume here that authors them- 27 selves would bear publication fees. Thus, if an author made a contract and the sales price were c , the author would have to pay a fee equal to the ﬁ rst copy cost f . Likewise, if the author made a contract and named a lower price than c, the author would have to pay more than f as a publication fee in order to cover the publisher’s costs. The sales pr ice that an author of a work would choose if he or she made 28 a contract can be seen from Figure 1. An author who would have chosen a price higher than c g iven copyright cannot name such a price in the ab- sence of copyright. The best feasible price for such an author to elect is c . Therefore, all authors with a t less than or equal to t would choose c and have to pay the ﬁ rst copy cost f in order to publish. Authors with higher t would have chosen a price below c and can still do so in the absence of copyright. Hence, the price chosen by authors in their preferred contracts is given by the dotted graph in Figure 1. 20 I assume that in the absence of copyright, the ﬁ rst publisher could not prevent copying by technological means, such as digital rights management. See Section 5.1. 21 This presumes that the per unit cost c is positive, that is, that publishing does not consist of mounting the work on the Internet. 312 ~ Shavell: Should Copyright Of Academic Works Be Abolished? Relatedly, we can see that the utility of authors from their contracts is given by the dotted graph in Figure 2. Note that authors with t belo w t obtain less utility because in the absence of copyright they are unable to make their otherwise-preferred contracts involving prices exceeding c . The utility loss suffered by such authors is greatest for authors with t of 0, for these authors would have named the highest price (the monopoly price) and would have earned the highest royalty, and the utility loss declines as t increases. For t at least t , authors suffer no utility loss, for the absence of copyright does not affect the contract that these authors would make, as the price would be c or belo w. 30 The height of the dotted graph tells us whether an author would choose to write in the absence of copyright; if the height of the graph exceeds the effort cost, the author would write and publish; otherwise the author would not write. 31 Publication f ees are subsidized . Now suppose that authors would not bear publication fees, because universities or grantors would do so (for rea- sons to be discussed in Section 3). In particular, assume that universities and grantors would pay a publication fee of f , so that authors could pub- lish for free. It can be shown that if an author wrote a work and made a contract, the author would choose the same contract price as he or she would in the absence of a subsidy of f . Th us the dotted graph in Figure 1 that shows a contract price of c for authors with t below t and then lower prices for higher t applies whether or not authors would bear publication fees. 32 H owever, the utility of authors is higher as a result of the subsidy, as is shown in the dashed graph in Figure 2. Speciﬁ cally, the utility of an author is higher by the amount f at eac h t than if the subsidy would not be paid in the absence of copyright; this is why the dashed graph is everywhere above the dotted graph by the distance f . One can also see geometrically that the dashed and solid graphs must intersect at some point, which I label t . Accordingly, to the left of t the inc entive to write is lower than under copyright and to the right of t the incentive to write is higher than under copyright. The sense behind this conclusion is as follows. For relatively low t , authors have to give up relatively high royalty revenues; and the greater 22 I assume that universities and grantors would not pay a higher fee, which would for instance allow a book costing c to be printed to be sold below cost. 23 This is established in the appendix. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 313 readership, which they value relatively little, combined with the subsidy they obtain in the absence of copyright does not offset the royalties they cede. For relatively high t, converse logic implies that authors are better off giving up royalty revenues in exchange for higher readership and the subsidy. A gain, the height of the graph of utility determines whether an author 33 would choose to write; if the height of the dashed graph exceeds the effort cost, the author would write and publish. 2.3. The Effect and the Social Desirability of Eliminating Copyright From what has been said, we can proceed to identify the effects and the 34 possible social desirability of elimination of copyright. I separately con- sider the case where publication fees would be paid by authors and the case where the fees would be subsidized. Publication f ees are borne by authors . In this case, we have seen from Figure 2 35 that eliminating copyright will reduce the incentive to write for authors in Region A, namely, those authors who would have selected a price above c under copyright. Some of these authors with reduced incentives in Region A will decide 36 not to write. Such authors will tend to have relatively low t, for they are the ones who would have earned relatively high royalties and thus who lose the most in the absence of copyright. When these authors are discouraged from writing, social welfare will fall. The measure of the social welfare loss suffered for a work that is not written is the value that would have been ob- tained by its author and its readers under copyright less the costs of writing and publishing it. The remaining authors in Region A will still elect to write, because 37 the decline in their utility from publishing will leave them with beneﬁ ts exceeding the cost of writing. These authors will likely be those with rela- tively high t , because they are the ones who would have earned the lowest royalties and thus give up the least in the absence of copyright. Social wel- fare will increase in regard to each of these authors. The reason is that these authors would have published at a price above c , whereas in the absence of copyright they will publish at a price of c , resulting in a beneﬁ cial rise in readership. To illustrate, suppose that the price given copyright would be $25 and that the price would fall to the per unit cost of $2 without copyright. Then individuals who place a value on reading the work that is in between $2 and $25 will purchase the work, whereas they would not 314 ~ Shavell: Should Copyright Of Academic Works Be Abolished? have at the price of $25, and their purchase will augment social welfare. For example, an individual who is willing to pay $20 for the work would purchase it at a price of $2, and this would raise social welfare by $18 (the $20 value placed on it by the person minus the $2 unit cost of providing it to the person). 38 I t is also clear from Figure 2 that authors in Region B, with t at least t , have incentives to write that are unaffected by elimination of copy- right. Hence, for them, the elimination of copyright has no social welfare consequence. 39 Whether elimination of copyright is socially desirable can now be expressed as follows. To the extent that academic authors are in Region A and would be discouraged from writing by the elimination of copy- right, that policy would be undesirable. But to the extent that authors are in region A and would not be discouraged from writing, the policy would be desirable. The numbers of authors in both groups and the sig- niﬁ cance of the associated social welfare changes must be evaluated to determine whether the elimination of copyright would on net be socially desirable. 40 Publication f ees are subsidized . In this case, we know from Figure 2 that in Region C, with t less than t , eliminating copyright reduces the incentive to write. 41 N ow in Region C, we can reason as we did in regard to Region A in the case just considered. That is, some of the authors in Region C, those with relatively low t, will be discouraged from writing by their loss of royalties and the elimination of copyright, causing social welfare to fall. The other authors in Region C will still write, and since the price of their works will fall to c , social welfare will rise. 42 I n Regions D and E, however, the incentive to write will increase relative to under copyright (because the dashed graph is above the solid graph), leading to a number of effects on social welfare. First, all authors in Region D who would have written given copyright will continue to write but will publish at a lower price c , and this will raise social welfare as has been ex- plained. Second, in both Regions D and E, there will be authors who will 24 A ctually, this statement is inexact. Social welfare would rise by more than $18 because, if the person obtained the work, the author would also gain utility from having an additional reader. For expositional ease, I abstract from this factor in discussing social welfare in the text, but I take it into account in the appendix. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 315 write only because copyright has been eliminated and publication fees subsidized. That these new writings will occur may either raise or lower so- cial welfare. In particular, the possibility that the new writings would lower social welfare arises because the subsidy of publication fees means that an author contemplating writing does not take into account the actual cost of publishing. W hether eliminating copyright in this case is socially desirable depends 43 on three groups of authors: those in Region C who would be discouraged from writing by the elimination of copyright and in respect to which social welfare would be lowered; those in Regions C and D who would write both in the presence and the absence of copyright, and in respect to which social welfare would rise (because the price would fall to c); and those in D and E who would write only in the absence of copyright, in respect to which social welfare could either rise or fall. 3. DISCUS SION OF THE MODEL— AND THE POS SIBLE CASE F OR ABOLITION OF AC ADEMIC COP YRIGHT In this section, I restate and interpret the analysis of the stylized model. 44 A cademic authors would still have a strong afﬁ rmative motivation to publish 45 in the absence of copyright—to gain scholarly esteem and to advance themselves professionally. A major assumption in the model was that academics have a positive incentive to publish regardless of the existence of copyright, es- pecially to gain scholarly esteem. That this reason for publishing is impor- tant is manifest to any person familiar with the academic environment. Academics plan and politic to present their works, they seek to publish their works in as highly ranked venues as possible, and they attempt in other ways as well to promote citations to their works; for many academ- ics, these efforts to achieve notice for one’s publications are ardent and consuming. I assume, therefore, that academics derive substantial value from publication of their works because of the respect and attention 25 For instance, suppose the following: the author’s disutility of writing would be 10, the author’s utility from readership would be 12, the publication fee (and ﬁ rst copy cost) would be 20, and the utility readers themselves would obtain from publication is 15. Then the author would write given that the fee is subsidized, for 12 exceeds 10. Yet writing is not socially desirable be- cause the social costs of writing are 10 + 20 = 30, exceeding the social beneﬁ ts of 12 + 15 = 27. 316 ~ Shavell: Should Copyright Of Academic Works Be Abolished? given to them by their readers, and that the y would beneﬁ t in this way in a world without copyright. 46 A nother source of utility to academics from publication that does not depend on the existence of academic copyright is that the promotion and salary of academics, and possibly their textbook-writing and consulting oppor tunities, are a function of their publication records. P ublication would bring these monetary beneﬁ ts to academics even if academic copyright were eliminated. 47 Consistent with the view that academics obtain signiﬁ cant value from publication that does not depend on the existence of copyright is the obser- vation that, in actuality, academics earn little royalty income from publica- tion. Publication of articles rarely results in receipt of royalties. Publication of academic books usually does result in some royalty payments, but this income tends to be minor. That academics in fact gener ally obtain only 26 See sour ces cited in note 7 above; Ian Rowlands & Dave Nicholas (2005, 17 & ﬁ g.7), ﬁ nd- ing that the most important factors to authors in choosing a journal for publication are its reputation and readership. 27 Academics would continue to beneﬁ t presuming that they would be cited as the authors of their works. In a world without copyright, I assume that there would still be a requirement to attribute copied material to its author. See Section 5.1 below. 28 See, e.g. , Weinstein v. Univ. of Ill., 811 F .2d 1091, 1094 (7th Cir. 1987) (Easterbrook, J.), noting that a university’s “demands—especially the demands of departments deciding whether to award tenure—will be ‘the motivating factor in the preparation of ’ many a scholarly work”; Landes & Posner (2003, 48), “Many authors derive substantial beneﬁ ts from publication that are over and beyond any royalties. This is true not only in terms of prestige, celebrity, and other forms of non- pecuniary income, but also in terms of pecuniary income in such forms as a higher salary for a professor who publishes than for one who does not, or greater consulting income . . . .” (footnote and emphasis omitted); Mark J. McCabe & Christopher M. Snyder (2004), “Prestige is the curren- cy of academia. Aside from the pure enjoyment of prestige, a scholar’s chances for promotion, ten- ure, a higher salary, etc., increase with his or her prestige. Prestige comes from doing high- quality research.” On the importance of publishing for salary, see, e.g., On ur Baser & Elda Pema (2003, 6), “Publications, we conclude, have a positive and diminishing effect on salaries.”; and William J. Moore, Robert J. Newman, & Geoffrey K. Turnbull (1998, 360-361), concluding that quantity and quality of publications have strong positive effects on academic economists’ salaries. 29 T extbooks could still be copyrighted if academic copyright were eliminated, as I discuss in Section 5.1 below. 30 See, e.g., Stephen R. Barnett et al. (1998, ¶ 1(e)), “The Author grants the above rights without claim of royalties or other compensation.”; Beth Luey (2002, 3), “Journals do not pay authors . . . .” 31 F or example, my understanding is that under a typical arrangement, an academic author would receive a royalty rate of 15 percent of revenue. If a book sold, say, 1,000 copies at a price of $60, the royalties received would be $9,000. If the book were 300 pages in length and the author spent an average of 3 hours a page, the implied hourly rate of pay received by the author would be $10. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 317 insubstantial or no royalty income from publishing yet they devote great effort to writing implies that the incentives to publish that are unrelated to copyright must be strong. Publication f ees, however, would probably be charged by publishers in the 48 absence of academic copyright, and the fees would be more than nominal. As emphasiz ed in the analysis of the model, publishers would be unable to proﬁ t from reader charges if academic copyright were eliminated. A pub- lisher of a book or journal would not be able to impose fees on readers for Internet access to its content, since all academic works would be made freely available by other parties. Likewise, a publisher would not be able to charge readers for print versions of works amounts exceeding printing and distri- bution costs, due to competition from other sellers of print versions. H ence, an initial publisher of a work would need to recover the ﬁ rst copy 49 costs it incurred—the expenses of arranging peer review, selecting, edit- ing, and formatting the work—from a source other than reader charges. Consequently, the publisher would have to impose fees on authors for publishing (assuming that the publisher’s operations would not be directly subsidized by universities or other donors). W hat would be the likely magnitude of the ﬁ rst copy costs and publish- 50 ing fees? First copy costs vary widely, from about $400 to $8,000 per article, with most falling in the $1,000 to $4,000 interval (King 2007). We can also gain some appreciation of the magnitude of ﬁ rst copy costs from the publication fees now charged by open access journals. As of April, 2008, fees charged to authors by major open access journals ranged from $1,250 32 I n other words, I am assuming that ﬁ rst publishers in today’s world would not be able to earn meaningful proﬁ ts merely from being ﬁ rst, given the ease of copying and the existence of the Internet. I am assuming also, as I noted, that initial publishers would not be able to purposely impede copying through technological means. See Section 5.1 belo w. In the past, proﬁ ts suf- ﬁ cient to allow publishers to succeed may have been possible in the absence of copyright. See Stephen Breyer (1970). 33 See also Wellcome Trust (2004, 11), discussing certain literature and commenting on meth- ods of ﬁ nancing scientiﬁ c publishing; Bernard Wysocki, Jr. (2005, A1), “Industry experts say typical per-article costs are between $3,000 and $4,000.” In terms of ﬁ rst copy costs per page, compare Bergstrom (2001, 187), noting that for economics journals, “ﬁ rst copy costs aver- age about $100 per page” and “marginal subscriber costs are about $.02 per subscriber per page,” with Joop Dirkmaat (2002, 229), criticizing Bergstrom’s methodology and estimating ﬁ rst copy costs to be at least $281 per page. For major international journals, these costs will be much higher than average owing to increased content in each issue to edit, more submissions to sift through, and greater expenditures on peer review for each submission. See Wellcome Trust (2004, 14); see, e.g., Wysocki (2005, A1), noting that the journal Science has a per ar ticle ﬁ rst-copy cost of approximately $10,000. 318 ~ Shavell: Should Copyright Of Academic Works Be Abolished? to $3,000 per article (King & Alvarado-Albertorio 2008, 264 table 7). For instance, Public Library of Science journals charge from $1,250 to $2,750 per article, and BioMed Central charges from nothing to $2,685 per ar- ticle, with $1,690 being standard ( Id. ). These open access journal fees may, however, understate their ﬁ rst copy costs because they sometimes receive income from grantors; and if the journals engage in less reviewing of sub- missions than traditional journals, their costs and fees may also understate the ﬁ rst copy costs of many journals of relevance to us. 51 Being longer than articles, books involve greater ﬁ rst copy costs, so that publishing fees for books would be greater. A conservative estimate is that if fees covered ﬁ rst copy costs for books, they would be at least $10,000. 52 It should also be observed that ﬁ rst copy costs are likely to fall over time because of publisher use of the Internet and of computer software. 34 See Lila G uterman (2004, A10, A11), noting that the Public Library of Science open access journals have beneﬁ ted from a $9 million grant from a private foundation; Colin Steele (2008), mentioning grants from the MacArthur Foundation and Mellon Foundation to university presses to print open access monographs; Peter Suber (2009a, § 5), reporting grants from the Wellcome Trust and the Mellon Foundation to open access journals. I myself have been in- volved in the launching of a new open access publication, the Journal of Legal Analysis , which is supported by grants from a private donor and by Harvard Law School. 35 Some indi viduals have expressed skepticism to me that ﬁ rst copy costs could be as large as I have described in this paragraph, suggesting that ﬁ rst copy costs should be negligible, in part because readers do not really demand copy-editing. If these individuals are correct in their as- sessment, that only reinforces the overall case I make in this article in favor of elimination of copyright; obviously, the lower are the costs of writing, the lower would be the fees for publica- tion and the less they would impede the motive to write and publish. 36 Dr yburgh Assoc. (2002) reports from a survey that the total ﬁ rst copy cost of an academic book is £7,391 (54 percent of which is for copyediting and typesetting), which at an exchange rate of $1.49, http://www.x-rates.com, is $11,013. “The upfront costs for publishing a mono- graph are . . . from about $20,000 on the low end to many multiples of that . . . .” Sanford G. Thatcher (2009, 62). Harvard University Press suggested in a conversation that the average ﬁ rst-copy cost per page is about $50, implying a ﬁ rst copy cost of $15,000 for a book of 300 pages. Telephone Interview with personnel, Harvard University Press, in Cambridge, Mass. (May 28, 2009). Also, it is reported that the copy editing costs of a page of an article average $85, suggesting that copy editing costs of a book of 300 pages would be over $20,000 (Tenopir & King 2000, 258 tbl. 51). 37 One mig ht think that ﬁ rst copy costs are also likely to fall because authors would undertake copy editing tasks themselves. But if authors were to do this, they would then add to the publication fee their effort costs of copy editing. Hence, it is not obvious that the effective publication fee would fall as a consequence of authors doing copy editing themselves. Indeed, if a typical author’s time is more valuable than that of a copy editor or a typical author is less efﬁ cient in performing copy editing, the effective fee would rise if authors do copy edit- ing (and for that reason, one would predict that copy editing would continue to be done by journals or by hired professionals). Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 319 Journals now usually accept online submissions of manuscripts and com- municate with authors and referees electronically, saving mailing costs. Also, journal administration is becoming automatic in many respects; journals are increasingly employing sophisticated software to facilitate the referee- ing and review process. Software is also, of course, employed in the editing and formatting of manuscripts; it is used to check spelling, grammar, and references, to prepare diagrams, and to format manuscripts. Nevertheless, copy editing costs are likely to remain a signiﬁ cant component of ﬁ rst copy costs, because of the inherent value of copy editing. A related mitigating point about costs is that various ﬁ xed costs of pub- 53 lishing would be avoided if academic copyright were eliminated. There would obviously be no need for legal protection of copyright; thus pub- lishers would not have to ensure that authors properly granted them copyright, to negotiate licensing arrangements, or to police for copyright infringement. Also, there would be no need for subscription management and for charging readers, except to the extent that print versions were sold. Because these ﬁ xed expenses would not be incurred in the absence of aca- demic copyright, the costs and publishing fees would be lower than might otherwise be thought. If academics w ould have to bear publication fees in the absence of copyright, 54 their incentive to write and to publish would fall. Let us provisionally assume that the publication fees just discussed would be borne directly by academics. Then we know as a logical matter that the incentive to publish 38 On the efﬁ ciencies of online administration of journals and of editorial software, see John B. Thompson (2005, 112–113), noting that the cumulative effect of the Internet and and com- puter software “has been a dramatic decline in the cost of typesetting”, and Priscilla Markwood (2006), describing a number of novel software techniques for publication and noting how technology advances in recent years, especially software for copyediting, have allowed many tasks to become automated. 39 J anet H. Fisher of MIT Press states that “Some discussions … have asserted that authors can handle everything—they don’t need copy editing, proofreading, or formatting of their com- puter ﬁ les. The argument is that authors are doing most of this themselves now anyway. I strongly disagree. Most peer-reviewed journals are copy edited, and most scholars who have been through the process are grateful for the careful eye of a copy editor to catch incomplete references, mistakes in mathematical calculations, mislabeled ﬁ gures, misnumbered footnotes, and even the occasional convoluted sentence. The quality of a scholar’s article would decline— in some cases signiﬁ cantly!—if copy editing was removed from the process.” Fisher (1995, 89). If this judgment is incorrect, copy editing turns out to be largely unnecessary, and ﬁ rst copy costs become much lower than predicted in the text, then the conclusion to which I come— that the elimination of academic copyright would be socially desirable—would again only be reinforced. 320 ~ Shavell: Should Copyright Of Academic Works Be Abolished? and thus to write would fall for all authors who today do not pay publica- tion fees and who grant copyright to their publishers. These authors, who are typical, would be made worse off by having to pay publication fees, so that their motive to publish and write would be lower in the absence of copyright. 55 Would the imposition of the publication fees under discussion have a signiﬁ cant effect on writing and publication? Would having to pay, say, $2,000 to publish an article or $10,000 to publish a book substantially re- duce writing and publication of articles and books? This is an empirical question about which we have, as far as I know, little hard information. My own intuition is that publication fees would exert an effect that would be measurable, and I ﬁ nd it plausible that the effect would be signiﬁ cant, especially for the many academics whose salaries are modest. However, this conclusion is premised on the assumption that the academics themselves would have to pay the fees, an assumption that I now relax. 56 If, however, academics would not have to bear publication fees—because univ ersities or grantors would pay them—their incentive to write and to publish articles would not fall and could even rise, and so too might their incentive to write and publish books . Now let us suppose that universities or grantors 40 See Sanford G. Thatcher (2008, 98), “It has long been a tradition in scholarly publishing for au- thors to transfer all rights to the publishers of their articles and even their books.”; Bo-Christer Björk & Turid Hedlund (2009), estimating that open access journals comprise 8.5 percent of all scholarly peer-reviewed journals worldwide; Bo-Christer Björk, Annikki Roosr & Mari Lauri (2008, 184), calculating that in 2006, 4.6 percent of articles worldwide were freely avail- able immediately, another 3.5 percent would be available after a period of delay, and another 11.3 percent were available in repositories. 41 Authors today are in principle able to pay publication fees to offset ﬁ rst copy costs and not to give copyright to publishers, but most choose to give copyright to publishers. Hence, it must be that they would be worse off if they did not transfer copyright to publishers and paid pub- lication fees. In terms of the model, these are authors with t less than t . They are the authors in region A of Figure 2. To the extent that there are authors who today pay full publication fees and do not transfer copyright, the elimination of copyright would make no difference. In terms of the model, these are authors with t at least t . 42 Ther e does exist survey evidence suggesting that most authors would be unwilling to pay any, or at any rate, more than $500 in fees to publish an article. See Robert Denicola (2006, 358 & n.38), collecting surveys. This evidence, however, is misleading: It does not imply that imposi- tion of publication fees would in fact have a pronounced negative effect on publication in a world without copyright, for unlike authors in that world, the survey respondents enjoy the option of publishing for free in traditional journals. See Sara Schroter & Leanne Tite (2006, 143), noting that 53 percent of surveyed authors “said it would be pointless to pay to publish in ‘this’ journal when other journals might publish it for free.” Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 321 would pay the publication fees. Then the inc entive of authors to write and to publish articles could in principle increase in the absence of copyright. The reason is that, as is generally true today, authors would pay nothing to publish articles, but authors should beneﬁ t from greater readership, and thus from greater esteem, because their articles would be available for free on the Internet and printed copies of articles and of journals would sell at cost due to competitive pressures. What the magnitude of the change in incentives would be is not clear—it depends on how much the authors would value the increase in readership —but its direction is. The case of books is different from that of articles, because authors 57 of books earn royalties today. H ence, an author of a book would lose royalties even though he or she would gain readership in a regime with- out copyright and subsidy of publication fees. The incentive to write and publish books would therefore increase if the beneﬁ t from greater read- ership outweighs the loss in royalty revenue, but would decrease in the converse situation. Still, an increase in publication of books would not be a surprise, in light of the low royalties earned from academic books and the probably substantial increase in readership that would come about in a world without copyright. Universities and grantors would have a motive to subsidize publication fees 58 in a world without copyright. U niversities would have a reason to pay the fees that would otherwise be charged to their faculty members for publishing in a world without copyright. Namely, universities would not want their 43 In particular, assume as in the model that they would pay fees equal to ﬁ rst copy costs, but would not pay more, such as if an author of a book wanted the printed version to be sold below the marginal cost of printing. 44 I n terms of the model, authors of articles in journals are roughly described by Figures 1 and 2 at the point t , where the price charged is above c and wher e authors pay nothing and receive nothing. 45 I n the model, an author would deﬁ nitely experience greater readership in the absence of copy- right because his or her work would be freely available on the Internet (or in printed form at a price equal to cost). However, in fact the author might not gain readership because not only would the author’s work become freely available, so would all other works that had been copy- righted, meaning greater competition for readership. I abstracted from these complications in the model. 46 If most of the readers about whom an author cares—say academics who are actively engaged in related research—would already see the article, then the increase in readership would matter little to the author. Yet this is not always the case. 47 I n terms of the model, authors of books are roughly described by Figures 1 and 2 at points below t , where the price charged is above c and where authors receive positive payments. 1 322 ~ Shavell: Should Copyright Of Academic Works Be Abolished? faculty to write and publish less. A university’s reputation and thus its abilit y to attract faculty, students, gifts, and grants, are obviously impor- tantly affected by the publication record of its faculty members. Hence, it would presumably be in the interest of universities to subsidize publica- tion fees to offset an undue dilution in publication incentives that would otherwise result from faculty having to bear publication fees. A proviso, however, is that universities might want to exercise some control over the quality of publications for which they pay publication fees (how so, to be discussed below). 59 Another proviso is that universities would not want a policy of subsidies to result in an excessive increase in publication fees by publishers. If universities simply paid publication fees, whatever their magnitude, an escalation of such fees could be imagined to occur. This problem could be ameliorated by granting each faculty member a fund for academic expenses including publication fees, for then faculty would have a motive to search for lower fees. U niversities could also police fees for unreasonableness. 60 The view that universities would wish to subsidize publication fees is consistent with the fact that universities today sometimes pay publication fees charged to faculty by open access journals and that, of course, univer- sities spend generally on support of faculty research. 61 It should also be noted that, on average, universities would not be ex- pected to face a serious liquidity problem in ﬁ nancing publication fees. In a world without copyright, universities would not have to spend the con- siderable amount they now do on subscriptions for journals and on new books. Today, the annual cost of a single journal subscription can exceed $20,000 (Denicola 2006, 352 & n. 4), and the annual library budgets of 48 A sa ving that a faculty member achieved in publication fees could be used by the faculty mem- ber on, say, travel to a conference or other academic expenses. Thus, even if, as would be desir- able, the fund is generous enough to cover all publication fees, a faculty member would still have a motive to search for low fees, other things being equal. The problem of the design of a faculty member fund for subsidy of publication fees is more complicated in a regime like our present one where copyright exists (mainly because the fund for fees cannot be fungible with other academic expenses, lest the faculty member simply avoid all fees by publishing in tradi- tional journals with no fees). On this matter, see Stuart M. Shieber (2009). 49 See Bramble (2006, 213); Leslie A. Harmel (2005, 568); Suber (2009a, § 5), reporting that the number of universities providing funds to pay publication fees for open access journals increased in 2008 from six to eleven; Universities UK and the Research Information Network (2009), containing recommendations for university payments of open-access publication fees. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 323 research universities are typically in the millions of dollars. Because pub- lishers would only be able to charge for printing and distribution costs in the absence of copyright, and because all publications would be freely avail- able on the Internet, universities would be expected to save substantially. On the whole, the amount that universities would save could exceed the amount they would spend on publication fees, for the subscription and new book prices now paid cover publisher costs and proﬁ ts, whereas the publication fees would cover only publisher costs. That is, university expen- ditures on publication fees could be less in a world without copyright than their expenditures today on subscriptions and book purchases, because universities would no longer be ﬁ nancing publisher proﬁ ts from academic works. H owever, those universities whose faculties are relatively produc- tive researchers would tend to pay more in publication fees than they would save on subscriptions and book purchases. Grantors too would be predicted to want to subsidize publication fees. 62 Since a grantor usually has as a speciﬁ c purpose the publication of research, its incentives to subsidize publication might be comparable to those of universities. Again, consistent with this prediction about grantor behav- ior is that several prominent research sponsors today fund author publi- cation costs; the Howard Hughes Medical Institute, the National Science Foundation, and the National Institutes of Health allow for this (Denicola 2006, 359). 50 At doctoral degree granting institutions in the United States, the mean expenditure in 2006 on library materials was $8.688 million, including $2.263 million on serials. See Mary Jane Petrowski (2007, 23). For example, in that year Brigham Young University spent $23.942 mil- lion on library materials ( id ., 25) ; Catholic University spent $5.122 million ( id., 27); Duke University spent $33.532 million ( id., 27); and Yale University spent $74.938 million ( id., 43). 51 T o amplify, suppose that universities are the only purchasers of academic works in a world with copyright and that in the absence of copyright, publication fees would equal only publisher costs. Then university expenditures on publication fees would fall in the absence of copyright by the amount of publisher proﬁ ts. However, to the extent that purchasers of academic works today are parties outside the university (and that such parties do not publish and thus would not pay publication fees), university expenditures on publication fees would fall by less than the amount of publisher proﬁ ts. Indeed, if the fraction of purchasers outside the university were sufﬁ ciently high, university expenditures on publication fees would rise in the absence of copyright. 52 F or example, Guterman (2004, A11) calculated that Duke University would pay slightly more in publication fees in an open access setting than it would save on its library journal budget. See also the discussion of this issue in Donald King and Carol Tenopir (2004). 53 See also BioMed Central (2009), noting thirty seven major funders of biomedical research that allow payment of article processing charges for open access publications. 324 ~ Shavell: Should Copyright Of Academic Works Be Abolished? Both uni versities and grantors could, as a partial substitute for paying authors publication fees, subsidize journals or book publishers directly, allowing them to reduce the publication fees they would need to charge authors. However, this policy would often be less attractive for universi- ties and grantors than payment of publication fees. The reason is that if a university or grantor subsidizes a publishing venue, then all authors who publish there are beneﬁ ted, not just the university’s faculty member or the grantee. Thus, from a university or grantor’s perspective, it is more effective to subsidize publication fees than journals or publishers. Hence, I would expect the primary way in which universities and grantors would support publishing is by the payment of publication fees. 64 The effect of elimination of academic copyright on the level of publication de- pends on the extent to which universities and grantors would subsidize publica- tion fees. Because of the motive of universities and grantors to subsidize these fees, the number of published works is unlikely to decline substantially and could increase . This conclusion follows from what has been said above. 65 T o the degree that publications would be discouraged by the elimination of academic copyright, the social losses would be limited because these publications would not ordinarily be of high quality. An article of high quality that has been accepted for publication is likely to be recognized as having this character by its author. The author will be able to make judgments about the merit of the article from such indicators as the ranking of the journal to which it was accepted, the reception of the article at presentations, and cita- tions to it in unpublished form. Furthermore, the author will be aware of his or her own general standing in the profession and record of success. If, then, authors generally know when their accepted articles are of high quality, they will value publication of such articles commensurately and thus be relatively unlikely to decide against publishing on account of hav- ing to bear a publication fee. Moreover, the probability of having to bear a publication fee should be comparatively low for high quality articles, for universities will be more likely to defray publication fees for articles published in upper tier journals and for faculty with better track records; additionally, high quality articles tend to be published by faculty at uni- versities with substantial resources. 66 Conversely, articles of poor quality that have been accepted for publi- cation are likely to be recognized as weak by their authors. These articles will often have been rejected for publication at a number of journals, will have been accepted only at a low-ranking journal, will not have made Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 325 much of an impression on colleagues at presentations or have been cited often, and will have been written by individuals without strong scholarly records. In all, it seems that articles that have been written that are of poor quality are the ones whose publication would tend to be discouraged by publication fees. I ha ve explained that if an article has been written and is of high rather 67 than poor quality, it would be relatively unlikely to be discouraged from be- ing published by publication fees. However, it should also be asked wheth- er the writing of high quality articles in the ﬁ rst place would be reduced by publication fees. When an author contemplates writing an article that would turn out to be of high quality, would the author decide against mak- ing the effort because of the possibility it would be poor and that the author would then be likely to have to pay publication fees? This is possible, but it must also be taken into account that when an author contemplates writ- ing an article, the author might have an idea of its quality, or at least might formulate one after undertaking preliminary work on the article. The importance of the claim that discouraged articles are unlikely to be 68 those of high quality seems substantial because of the great variability in quality of academic articles. There is evidence that article quality is con- centrated, that a small fraction of all published articles receives the great majority of citations. I f publication of this group of articles would only be discouraged in a minor way by publication fees, then the social loss associ- ated with the typical discouraged publication will be small. The foregoing points about articles apply to academic books as well, but 69 I conjecture with somewhat less force. The reasons are that an author’s abil- ity to forecast the quality of a book from the status of the publisher who accepted it or from citations to it in unpublished form seem less than for an article, and also that the prospect of having to pay publishing fees would weigh more heavily on an author contemplating writing a book than an article. T he social loss from a discouraged publication would also be limited because 70 an unpublished work could be posted on the Internet. Articles and books that are not published could and probably would be posted on an Internet site, such as that of an open access repository, a working paper series, or an 54 See e.g ., John P. A. Ioannidis (2006), documenting that for the sciences, including the social sciences, “the most inﬂ uential papers are extremely concentrated in a few journals” and not- ing that a core number of journals, constituting about 2 percent of all journals, publishes 95 percent of all cited articles. 326 ~ Shavell: Should Copyright Of Academic Works Be Abolished? author’s personal website. Internet posting would limit the social loss from a work’s not being published because it would be available for all to read. Furthermore, the ability of potential readers to locate the work would not be greatly compromised by its not being published because the author could post it on a site of appropriate subject matter relevance and because of the power of search engines. 71 Of course, the point that an unpublished work can be posted on the Internet is moot if the work would not have been written in the ﬁ rst place due to the prospect of having to pay a fee for publication. But this possibili- ty is itself made less likely by the opportunity to post works on the Internet, for that means that the author still would obtain utility from having written an article or book. 72 To the degree that publications would be encouraged by the elimination of academic copyright and subsidy of publication fees, either social gains or losses could be engendered. The latter problem might be offset by university and grantor efforts to condition a subsidy on quality. I observed above that in the absence of copyright but given subsidy of publication fees, some authors might decide to write who would not do so in a world with copyright. The generation of such new works would often be socially desirable but would also sometimes be socially undesirable. 73 The possibilit y of socially undesirable publication is an instance of the standard disadvantageous economic byproduct of a subsidy: because those receiving the subsidy do not take into full account the cost of what they purchase—here publication services—they might decide to make a pur- chase even though true costs exceed the beneﬁ ts. Consider, for example, a journal of very low quality that so few read as to make its operation socially undesirable in view of its publishing costs of $2,000 per article. Even so, this journal could be imagined to operate successfully if its publication fee of $2,000 would be borne by universities, for authors might well derive posi- tive beneﬁ ts from publishing in it—suppose they had written articles that were rejected in better journals and wanted to see their work in print. 74 S uch problems of socially excessive publication could be countered if universities (and grantors) imposed quality constraints on articles that they subsidize. They would have an incentive to do this as well, because 55 N ote that it is plausible that the journal would not operate in a world with copyright, for given the assumption that the journal has very few readers, it would have to charge authors publica- tion fees approximating $2,000 per article. It would be expected that few authors would want to pay this fee to publish in a journal known to be poor and having few readers. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 327 publications of questionable quality would not tend to enhance their repu- tations (but would cost as much as high quality publications). However, universities would not want to exercise control over the subsidy of publica- tion fees for articles in such a way that academic freedom, and expression of unpopular views, would be compromised. Suppose, though, that uni- versities base the granting of subsidies for an article on the identity of the journal that accepted the article—giving subsidies to any article accepted in a journal in a named list of journals of sufﬁ cient reputation. Then the content of the particular article itself could not directly inﬂ uence the sub- sidy decision and thus the threat to academic freedom would presumably be limited. Likewise, if universities base the granting of subsidies for a book on the identity of the publisher that accepted the book, not on the book itself, there should be no real threat to academic freedom in respect to the publication of books. Summary to this point: T he effect of elimination of copyright on authors’ incen- 75 tives to publish might not be negative overall—it might lead to more publications, due to subsidy of publication fees—and to the extent that it would discourage publications, the loss in social welfare would probably be limited. The preceding discussion suggests that the level of publications—especially of articles— would be unlikely to fall and might even increase in the absence of copy- right. If the number of published works were to fall, the decline would be unlikely to be substantial. The main reasons for these predictions are that universities and grantors would have an incentive to pay authors’ publica- tion fees and that readership of works and thus authors’ esteem would rise in the absence of copyright. Also, the social loss associated with discour- aged publications would probably be limited, for the quality of such publi- cations would tend to be low and unpublished written works could still be posted on the Internet. Further, although the effects on social welfare of a subsidy-induced increase in publications could be positive or negative, the latter possibility could be countered by the conditioning of subsidies on publication quality. Social beneﬁ ts from eliminating academic copyright—deriving from the free 76 availability of academic works. I f copyright of academic works were ended, a social beneﬁ t would be enjoyed associated with works that would still be published but that would otherwise have been copyrighted, and also with some of the works that would be published only because of the absence of copyright. In the absence of copyright, all these articles and books would presumably become instantly available on the Internet for individuals to 328 ~ Shavell: Should Copyright Of Academic Works Be Abolished? download freely. Also, print copies would often be produced and would sell for approximately production cost, due to competitive pressures. Hence, for instance, many new academic books would quickly become available as moderately priced hardbacks or as inexpensive paperbacks. Furthermore, teaching materials drawing on published works would become easy for academics to assemble because permissions for use would not have to be secured or royalties paid. 77 In considering these social beneﬁ ts, we need to make a realistic compari- son with the state of affairs under copyright. Today, universities subscribe to a large number of journals and make their content freely available to many in the university community through library and Internet access, and faculty members often have budgets that are designated for the purchase of journals and books. Moreover, copyright protection is signiﬁ cantly in- complete for articles. Journals are increasingly allowing free downloading of their articles from the Internet, at least after a window of time following publication. In addition to this voluntary provision of access to articles, there is much legally unpermitted exchange of articles because of the ease of copying and of emailing content over the Internet. Authors often main- tain personal websites from which their actual published articles can be downloaded despite this being a violation of publisher’s copyright privi- leges. Furthermore, as I will discuss in Section 4.1, pre-published versions of published articles can usually be located on the Internet and are often deposited in repositories. Hence, the difference that elimination of copy- right would make to the availability of articles should not be exaggerated; access to articles that are copyrighted is much greater than would be true were copyright protection insisted upon and rigorously enforced, and were it true that each individual had to pay for copyrighted works rather than a university paying for them. 78 N evertheless, the beneﬁ ts of abolition of copyright of articles are positive. Faculty and students do not have ready access to all published articles on the Internet and often face costly-in-time hurdles to locate what is in theory freely available. Articles in pre-published form are not equivalent to pub- lished versions. The assembly of teaching materials from articles and the 56 See, e.g., PubMed Central (2009), containing a list of journals in medicine and the life sci- ences, and showing that approximately half of the journals allow immediate access to their content and the remainder after two to thirty-six months, with twelve months being most common, and a similar listing at the Stanford University (2009) HighWire press website. 57 See Section 4.1. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 329 printing of them is often seriously constrained by copyright. F urther, many academics and students in institutions without substantial resources (includ- ing many small colleges and junior colleges in the United States and teaching institutions in other countries) cannot afford to pay for more than a narrow segment of journals. Additionally, there are numerous individuals who are not members of the university community but who wish to read academic works. When one takes these various observations into account and aggre- gates the beneﬁ ts of a copyright-free world over the relevant populations and the huge number of articles that are published, my supposition is that the sum would be substantial. Books seem to enjoy greater effective copyright protection than articles. 79 Academic books published under copyright today are not usually available 59 60 in electronic form, e ven though publishing practices are changing. And when books are offered in electronic versions, they must generally be pur- chased; unlike articles, books are not usually made freely available after a waiting period. Moreover, my personal impression is that there is a much lower ability of individuals to obtain electronic copies of books in pre- published version from the Internet than of articles. Hence, it appears that the difference that elimination of copyright would make for a typical book is greater than for a typical article, so that the proportional social beneﬁ t would be greater per book. W hether for articles or for books, it is evident that the existence of the 80 Internet greatly magniﬁ es the social beneﬁ ts that would be gained from elimination of copyright. Because of the Internet, all new academic works 58 See, e.g. , Princeton Univ. Press v. Mich. Document Servs., 99 F.3d 1381 (6th Cir. 1996) (en banc), afﬁ rming summary judgment against a commercial copy shop for copyright infringement be- cause it reproduced substantial segments of copyrighted materials in course packs; Basic Books, Inc. v. Kinko’s Graphics Corp., 758 F . Supp. 2d 1522 (S.D.N.Y. 1991), ﬁ nding Kinko’s liable to publishing houses for copyright infringement because it excerpted their copyrighted works, compiled them into course packets, and sold them to students; Ann Bartow (1998, 151), “[O]ver the past decade the scope of educational fair use has been dramatically compressed . . . . Fearful and litigation averse educational institutions . . . protect themselves by adopting copyright poli- cies that deﬁ ne a minimal scope of educational fair use, and make individual faculty members liable for any infringing photocopying activity outside of this petite orbit.” 59 A perusal of the websites of a number of university presses (Cornell University, Duke Univer- sity, Harvard University, University of Michigan, University of Texas) that I chose essentially at random shows that most sell few or no books in electronic form. The relatively limited list of books in electronic form offered by university presses may be found at the Association of American University Presses (2009) website. 60 On electronic publishing of academic books, see, e.g., Susan Gay (2007, 2). 330 ~ Shavell: Should Copyright Of Academic Works Be Abolished? would effectively become freely available to everyone. Befor e the age of the Internet, the elimination of copyright would only have meant that works would become available in the sense that journal publications and books would have fallen in price and that duplicated copies would have been mailed between individuals. 81 Finally , the elimination of copyright would mean that publishers would not invest resources in charging readers, making copyright arrangements with authors, granting licenses to others, or in protecting their copyrights through search for violations and through litigation. These savings would be distinct from the beneﬁ ts that we have been discussing due to free access to works and would be of some importance. 82 Conclusion. The judgment that I reach is that there is a plausible case for elimination of academic copyright, even though on a priori grounds one cannot come to the conclusion. I have offered arguments suggesting that the overall level of published academic articles and books might actually increase in the absence of academic copyright due to subsidy of publica- tion fees, but in any event, would not be likely to fall substantially. I have explained as well that any reduction that does occur would tend to be of relatively low quality works, and that most articles would be available on the Internet if not published, implying that the social loss per discouraged work would be relatively low. An increase in publications could be socially beneﬁ cial or detrimental, but the latter problem of excessive publication could be countered by limiting subsidization of publication fees to works 61 The a vailability of works copyrighted in the past is a different issue and one that I do not ad- dress, as the change in law I contemplate would abolish copyright only for future works, as I state in Section 5. In particular, therefore, I do not consider questions surrounding the social desirability of Google’s plan to digitize copyrighted and out-of-copyright books. See Google Book Settlement (2009). For conﬂ icting discussions of the impact the settlement will have on access to online books, see Robert Darnton (2009), criticizing the settlement for putting control of access to electronic versions of university books in the hands of a single for-proﬁ t company, and Paul Courant et al. (2009), presenting Paul Courant’s response to Robert Darnton that the settlement creates a valuable “universal bookstore,” and Darnton’s reply to Courant continuing to criticize the settlement for giving Google a de facto monopoly. 62 King (2007, 99, 104) discusses rights management and copyright protection activities as a component of indirect costs, which make up from 20 percent to 50 percent of total costs, and mentions an estimate that subscription management accounts for 7 percent of total costs. These costs would not be incurred in a copyright-free world. 63 As I explain in the model, whether eliminating academic copyright is socially desirable (in both the case where publication fees would and would not be subsidized), depends on the distribution of authors by type t , the cost of writing, c , and the values placed by readers on academic works. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 331 published in venues of reasonable quality. The social beneﬁ ts from ending copyright would be that the entire body of articles and books now under copyright, and thus the most valuable, highly cited works, would become available to all, for free downloading from the Internet, could be included in teaching materials without seeking permissions, and the like. Moreover, the costs of arranging and protecting copyrights would be avoided. In all, it does not seem unlikely that the beneﬁ ts would dominate the costs, and my prediction is that further analysis would validate this view. 4. THE FUND AMENTAL DIFFICULTIES FACED BY THE OPEN ACCES S MOVEMENT AND THE PR OB ABLE NEED F OR ABOLITION OF AC ADEMIC COP YRIGHT I n this section, I describe brieﬂ y the efforts of the open access movement 83 to make academic works freely available; I explain why these efforts are un- likely to succeed fully or without undue delay; and I suggest that a change in law may be needed to end academic copyright if the goal of unlimited access to academic publications is sought. 4.1. The Open Access Movement and Its Problems The most notable efforts to achieve unconstrained availability of academic 84 works fall under the heading of the open access movement. The announced aim of this movement is to assure that individuals can costlessly obtain published academic works, mainly journal literature. Open access initiatives have generally been limited in their objec- 85 tives, however. F requently, they seek to make freely available only pre- publication versions of articles, usually by having authors deposit these versions of their articles in repositories. What such initiatives achieve 64 See not e 4 above. As Peter Suber (2007) states, open access means the removal of “ price barriers (subscriptions, licensing fees, pay-per-view fees)” for readers and focuses on “peer-reviewed research articles.” 65 See Peter Suber (2009b) for a detailed description of open access initiatives. 66 P re-publication versions of articles range from “preprints,” versions that have not been ac- cepted for publication, to “postprints,” versions that have been approved for publication by peer review (but that may not have been copyedited); see Suber supra note 4. See, e.g., Na- ture Precedings (2009), allowing researchers to post preprint manuscripts; National Institutes of Public Health (2009) for its public access policy requiring grantees to “submit ﬁ nal peer- r eviewed journal manuscripts that arise from NIH funds to the digital archive PubMed Central upon acceptance for publication….” 332 ~ Shavell: Should Copyright Of Academic Works Be Abolished? over the status quo—in which discussion paper versions of articles have always been available—is not entirely clear, although it is positive, and the more so the closer is the accessible pre-publication version to the pub- lished article. In any case, a pre-publication version of an article is not a perfect substitute for the published article: readers will not necessarily know what is missing from a less than ﬁ nal version of an article, they will beneﬁ t from copyediting, they might want to cite ﬁ nal page numbers, and the like. Although some initiatives do aim to make actual published articles freely available, this is usually sought only after a delay, often six months or a year. 86 N ot only have open access initiatives been limited in their objectives, the initiatives have also encountered difﬁ culties in achieving their announced goals. Academics who have pledged support to an initiative or have been required to do so by their institutions have often been reluctant to follow through on their obligations. For example, a group of prominent scientists asked in 2001 for a boycott of scholarly journals that did not provide open access to publications within six months of publication (Public Library of Science 2001). Although over 30,000 scientists signed a promise to join the boycott, few actually did so and the boycott was viewed as a failure (Young 2002, A14). Recently, faculties at Harvard, MIT, and Stanford universities decided to require their members to publish their articles only if open access is allowed by publishers, but faculty members have been allowed to opt out of this “mandate,” presumably because of fear that they would resist it. R elatedly, the ability and willingness of universities that have 67 See, e.g., Howard Hughes Medical Institute (2007) for its policy requiring “its scientists to publish their original research articles in scientiﬁ c journals that allow the articles and sup- plementary materials to be made free accessible in a public repository within six months of publication”; the Scientiﬁ c Council of the European Research Council (2007) for its policy mandating that “all peer-reviewed publications from ERC-funded research projects be depos- ited on publication into an appropriate research repository where available, such as PubMed Central, ArXiv or an institutional repository, and subsequently made Open Access within 6 months of publication.” 68 H arvard University’s open access policies (Harvard University 2009), adopted in 2008 and 2009, apply at present to the Faculty of Arts and Sciences, Harvard Law School, and the John F. Kennedy School of Government. Under these policies each “faculty member grants to … Harvard … permission to make available his or her scholarly articles . . . to the public in an open-access repository.” But “[t]he Dean . . . will waive application of the policy to a particular article upon written request by a Faculty member ….” MIT also decided to employ an essen- tially identical open access policy (MIT News 2009). Stanford University’s School of Education adopted a similar open access policy (Stanford University, School of Education 2008). Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 333 adopted open access policies to sanction violations by their faculty has been limited. The open access movement has also sought to foster the development 87 of open-access academic journals, such as those in the Public Library of Science (Public Library of Science 2009). Open access journals generally do not copyright their works and often charge fees to authors to publish in or- der to cover their costs; thus they are examples of the kinds of journals that I described in Sections 2 and 3 that would exist in a world without copy- right. The number of open access journals has been rising, and they now constitute about 4 percent of all journals. However, open access journals are on the whole not among the most prestigious venues for publication (Thomas Hess et al. 2007). An important gauge of the effect of the open access movement is the 88 percentage of academic works published under conditions of open ac- cess. The fraction of journal publications that are immediately available on an open access basis is reported to be less than 5 percent (Björk, Roosr, & Lauri 2008), although almost all journal publications are available in at least some pre-published form. Finally, books are generally unavailable in open access format. 4.2. Explanation for the Problems The pr oblems and only partial success of the open access movement are not 89 difﬁ cult to explain. Obviously, traditional publishers would be unwilling to cede copyright privileges for nothing, for that would end their ability to charge for access. Traditional publishers would have to be paid to give up 69 For example, the Harvard, MIT, and Stanford open access policies, see preceding note, do not mention any sanctions for their violation. 70 The open ac cess website (Directory of Open Access Journals 2009) lists 4,050 open access jour- nals, and the Genamics Journal Seek (2009) website reports the total number of journals to be 94,589. However, Björk & Hedlund (2009) report that 8.5 percent of journals are open access. 71 See also B ramble (2006, 219), noting that “[c]ompared with more traditional modes of pub- lishing, Open Access journals and Open Access publishing/archiving in general suffer from a prestige gap”; Denicola (2006, 360), stating that “[w]hen authors who had never published in open-access journals were asked why they had not, sixty-nine percent said that their percep- tion that open-access journals have low prestige and low impact was an important or very important factor.” But see Public Library of Science (2005), noting, in addition, that young journals from the PLoS, particularly PLoS Biology, have made an impressive impact despite the prestige advantage of their more established competitors. 72 One survey reports that 97 percent of journals are “green,” with a majority allowing posting of postprints and the remainder posting of preprints (SHERPA RoMEO 2009a). 334 ~ Shavell: Should Copyright Of Academic Works Be Abolished? copyright, but as I discussed in Section 2.1, authors would not be expected to ﬁ nd it personally advantageous to do this; they will prefer for publishers to hold copyright in order to avoid having to pay fees for open access pub- lication. This is presumably why open access initiatives have been generally limited in nature and have not sought immediate open access to published articles themselves. 90 The r elative lack of success of open access journals is also not hard to understand. A primary difﬁ culty for an open access journal is, again, that authors will usually not want to pay publication fees to an open access publication; authors will tend to prefer to save these fees and not to beneﬁ t from the expanded readership that open access to their own works would bring. An additional problem for an open access publica- tion is that authors will be reluctant to publish with them due to their lack of prestige, as I mentioned above. Indeed, I suspect that many aca- demic authors would consider their lack of prestige a more serious dis- advantageous of open access publications than possibly having to pay publication fees. 91 I f open access publications are of lower quality than many tradition- al publications, the question arises why this would be so. The answer is presumably that open access publications were relatively recently begun and that high-quality publishing venues are difﬁ cult to establish. There is, though, no apparent intrinsic basis for believing that open access jour- nals would be of low quality. Open access journals can charge sufﬁ ciently high submission and publication fees to ﬁ nance a rigorous refereeing and editorial process; they can, and would generally be motivated to, exer- cise selectivity in the works that they accept for publication. H ence, the quality disadvantage of open access journals would be expected to wane over time. 73 Note the irony in this regard: academics’ desire for esteem is not only the factor that makes elimination of copyright socially desirable, it is also a factor that stands in the way of its elimination. 74 A number of commentators have suggested that open access journals would have an incentive to publish a greater volume of works in order to increase revenue—for their revenue would depend on publication fees—and that this would lead to a dilution of article quality. See, e.g., Karen Hunter (2004); and John Ewing (2004). This view does not make clear economic sense: if a journal degrades its quality, the fee that it will be able to charge each author will decline (for the esteem an author derives from publication depends on journal quality); hence, if a journal increases the number of acceptances and quality declines, its publication fee revenue and its proﬁ ts will not necessarily increase. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 335 A nother issue of relevance is the possibility that traditional journals would convert into open access journals. C onversion would, however, seem usually to go against the joint interests of authors and journals be- cause authors, as I have emphasized, would usually prefer to avoid publi- cation fees and to transfer copyright to journals so that they can impose charges on readers. Equivalently, if a traditional journal converted to an open access plan and charged publication fees, its submissions from authors would be likely to fall. A subset of authors, however, might prefer to pay publication fees in order to obtain open access for their articles, suggesting that a traditional journal might offer the option to an author of paying a publication fee and having that article freely available for downloading on the Internet. 4.3. The Fundamental Divergence between the Effects of Individual and Collective Efforts toward Open Access—and the Consequent Need for Legal Change One can e xpress the difﬁ culties faced by the open access movement in 93 terms of a basic divergence between the effects of individual and col- lective actions. If an individual academic insists on open access to his or her publication, this single action brings about no real change in the overall system of copyrighted works—that system continues to exist. And because the system of copyright continues to exist, the academic who obtains open access for his or her work does not as a consequence 75 This happened with the journal Nucleic Acids Research ; see Nucleic Acids Research 2009. See also Clair e Bird (2008), discussing and evaluating impact of Nucleic Acids Research ’s transition to open access, and noting that 68 other Oxford journals have since transitioned to some form of open access. A list of traditional journals that have converted to open ac- cess journals can be found at Open Access Directory (2009). On the other hand, some for- merly open access journals have recently switched to subscription models. See, e.g. , Journal of Visualized Experiments (2009), explaining that high production costs for video articles led the previously-open access Journal of Visualized Experiments to adopt a subscription model. 76 I am pr esuming that authors would have to pay the publication fees in saying this. Clearly, if universities would subsidize publication fees, journals that converted to an open access model would not lose submissions; they would gain submissions because their readership would increase. 77 See, e.g. , Toby Green (2009, 75), reporting that of 263 publishers responding to a recent survey by the Association of Learned and Professional Society Publishers, 30 percent offer optional open access to authors. For instance, the publisher Springer offers this option to authors under its Open Choice program; authors can publish individual articles giving readers open access if they pay a fee of $3,000 (Springer 2008); see also, e.g., Elsevier (2009); Oxford Journals (2009); Taylor & Francis (2009); Wiley InterScience (2009). For a partial list of journals offering this option and their prices, see SHERPA RoMEO (2009b). 336 ~ Shavell: Should Copyright Of Academic Works Be Abolished? obtain the beneﬁ ts that would ﬂ ow from the general abolition of aca- demic copyright. That is, the academic does not enjoy free access to all academic works. Likewise, even if a university requires that there be open access to works published by its faculty members, the university will not then garner the beneﬁ ts of free access to all academic works—the univer- sity will still have to pay for journals and for books. Hence, it may well be that a collective action, namely, one that ends the legal right to copyright academic works, is needed to achieve open access, or at least to achieve that objective soon. 94 A related observation is that it may be rational for an individual aca- demic or for a individual university to refrain from committing to publish academic works only if open access is granted, yet to support a change in law to eliminate academic copyright. Indeed, I suspect that most academics and many universities are in this category. 5. HOW COULD THE ELIMINATION OF AC ADEMIC COP YRIGHT BE ACCOMPLISHED ? 95 I have suggested in the preceding section that if free access to published works is thought to be socially desirable, it would best be accomplished by a modiﬁ cation in copyright law rather than by the initiatives of the open access movement. Let me now brieﬂ y describe such an alteration in law and whether it would be difﬁ cult to effect. 5.1. A Law Eliminating Copyright in Academic Works 96 B y a law abolishing copyright in academic works, I mean a law denying the right of copyright for all newly authored academic works in the United States. I suppose also that the law would seek to bar use of technological methods intended to impede copying and that contractual arrangements to prevent copying would be made unenforceable. Hence, all new academic works could be copied and presumably would be made available free of charge in electronic form or sold at cost in printed form. Copied material would, however, have to include citations to its source and might be protected 78 The arguments suggesting that c opyright be abolished apply throughout the world, but I restrict attention here to the law in the United States. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 337 against certain forms of abridgment—otherwise the author could bring a legal action against the violator. The implementation of a law ending copyright of academic works would 97 require a deﬁ nition of academic works. Let me ﬁ rst consider articles (books will be addressed subsequently). An article could be deemed an academic work if it were to be published in an academic journal. To determine whether a journal is academic, four indicia could be employed: whether its authors are usually academics; whether its readers are mainly academics; the degree to which its content is academic in character (displays sophistication and knowledge of prior learning); and, most important, the magnitude of any royalties received by authors (low or no royalties would favor classiﬁ cation as academic). These four factors bear on the importance of esteem to authors and thus, 98 as has been explained at length above, on whether elimination of copyright would be unlikely to detrimentally diminish the volume of publications. If an article is going to be published in a journal containing works that are typically authored and read by academics and that exhibit the usual qualities of academic works, then the likelihood is that the purpose of the author of the article in question was to gain recognition from academics and also that it was not written in order to obtain royalties. Furthermore, and crucially, if the journal does not generally pay any, or more than mod- est, royalties, that would strongly reinforce the inference about the object of the author and the lack of need for copyright to have induced the writing of the article. 79 The author of a work would have the right to enjoin the distribution of unattributed material and to collect damages for harm done due to lack of attribution or other violations of rights. For a discussion of the basis for these rights, see Mark A. Lemley (1995). (Today, such legal rights are presumably largely unnecessary because copyright prevents others from copying work without permission.) But there would be no legal duty to cite the author for the author’s ideas. That duty would be enforced as it is now, by scholarly norms. Copyright law does not protect ideas. See , e.g. , Goldstein (2005, 1:52). 80 One possibility is that these criteria would be applied by an expert extra-judicial body, which would create and maintain a list of journals considered academic. Parties displeased with the classiﬁ cation of a journal as academic by the expert body could have a right of appeal to the courts. 81 This last fact or can be applied only in regard to a journal that was in existence before the pas- sage of the statute under consideration. For journals established after the passage of the statute, evidence about planned payments to authors could be examined. 338 ~ Shavell: Should Copyright Of Academic Works Be Abolished? For the most part, journals would be straightforward to classify accord- ing to the four criteria, and those journals that we now would naturally consider to be academic would clearly be treated as such under the criteria. It is true that in some instances, not all of the criteria would apply, but through direct consideration of the underlying issue of whether copyright is needed to induce publication, classiﬁ cation could still often be intelli- gently accomplished. Consider an existing journal that publishes articles on international affairs; that its readership is less than half academic; that its authors are usually not academics—the nonacademic authors are drawn from think tanks and the State Department; that its articles, while serious, can readily be understood by non-specialists; and that it pays nothing to its authors. Should this journal be considered academic? If we ask the crucial question whether copyright is needed to induce publication of the articles in the journal, our answer would be no, so that we ought to treat the journal as academic and not allow copyright. In particular, the journal’s nonaca- demic authors have positions suggesting that they would value the esteem of their readers: members of think tanks are usually highly educated indi- viduals, often have connections to academia, and in many respects appear to share the goals and ethos of academics; and some members of the State Department may have related objectives. That the readership of the journal is not predominantly academic is not necessarily dispositive, for this factor serves only as an indicator of whether the author wants esteem. As a general matter we would infer that if the audience of a journal is not academic, then gaining the esteem of that audience would not be highly valued by an author. But in the case of a journal in international affairs, the esteem that a writer would desire would plausibly be from a wider community than just academics interested in international affairs; presumably the author would be pleased to gain notice from readers in think tanks and members of the State Department. That the articles can be comprehended by nonspecialists is, in the context at issue, not surprising, for international affairs is not an abstruse or technical ﬁ eld. Finally, that authors do not receive payment for their articles demonstrates that they did not need royalties to publish and suggests that many would have been willing to pay publication fees in the absence of copyright. Furthermore, institutions such as think tanks and the State Department might wish to pay publication fees in order to encourage their employees to publish. 100 T urning now to books, an approach similar to that just described for articles could be attempted. That is, a book could be deemed an academic Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 339 work if it was to be published by a publisher whose books—or books pub- lished under an imprint or distribution plan—satisfy the four factors that I mentioned in regard to journals. Notably, then, most books published by university presses, and sold by them as academic rather than trade press books, would be classiﬁ ed as academic works. However, I suspect that the classiﬁ cation of books would be more difﬁ cult than the classiﬁ cation of articles because books generally do result in positive royalties. In any case, textbooks would not be considered academic, so could be 101 copyrighted. The premise in this regard is that most textbooks would prob- ably not be written in the absence of the prospect of proﬁ t from copyright. That is not to deny that some measure of esteem is derived by academics from authoring textbooks. It is rather that the strength of this factor does not seem very strong. Compilations of academic works, however, would generally not be ac- 102 corded copyright. Compilations are different from textbooks in that they require relatively little effort to create. Therefore, my suspicion is that disal- lowing copyright for compilations would have only limited effect on the volume of compilations generated. 5.2. Effecting the Elimination of Academic Copyright Ending academic c opyright seems straightforward to accomplish from a 103 legal perspective. Presently, copyright is authorized under Title 17 of the U.S. Code. To terminate academic copyright could be achieved under new legislation, deﬁ ning an exception for academic works. Other carve outs al- ready exist, and ther e is nothing conceptually difﬁ cult about crafting a 82 A number of different types of carve-outs are already codiﬁ ed. There are complete denials of copyright to works produced by certain actors. See, e.g., 17 U.S.C. § 105 (2006), denying copyright to works produced by the U.S. Government. There are denials of particular rights normally secured by copyright to certain types of works. See, e.g. , id. § 114, excluding the right of performance from copyrights in sound recordings; id. § 120, excluding from copyrights in certain architectural works the rights to prevent the making, distribution, or display of pictures of the work. And there are denials of particular rights normally secured by copyright when the right is violated for a particular purpose. See, e.g. , id. § 108, granting libraries and archives limited reproduction rights to preserve works or make them available for research at other libraries or archives, notwithstanding copyrights in those works; id. § 121, making reproduction and distribution of copyrighted works “in specialized formats exclusively for use by blind or other persons with disabilities” not an infringement. However, Paul Goldstein has suggested to me that elimination of copyright for academic works could lead to conﬂ ict with the obligations of the United States under the TRIPS Agreement. 340 ~ Shavell: Should Copyright Of Academic Works Be Abolished? legal exemption for future academic works. M oreover, doing so would not be constitutionally problematic because there is no constitutional right to copyright. Indeed, insofar as abolishing academic copyright could even spur the production of knowledge, it would be the constitutionally pre- ferred course according to the Copyright Clause’s preamble that copyright should be used “To promote the Progress of Science and useful Arts.” 104 Eliminating academic copyright also seems feasible from a political per- spective because of its likely endorsement by universities, academics, and students. One would expect universities to support abolition of copyright because of their general desire to promulgate knowledge, their wish for fac- ulty to be able to assemble teaching materials without their having to secure permissions from copyright holders, and their purely ﬁ nancial interest in not having to pay for subscriptions to journals and high prices for books. In the latter regard, recall that universities as a whole would probably save funds if academic copyright is eliminated, even assuming that they fully subsidize publication fees. This is because universities would no longer serve as the source of much of the large proﬁ ts that accrue to publishers of academic works. Academics themselves would be likely to favor abolition, especially if they felt that publication fees would often be subsidized by their universities and grantors (which as I explained above would be probable). Under this assumption, academics would probably be better off were copy- right eliminated: their works would remain costless for them to publish but would be more widely circulated; and they themselves would have free access to all academic works. Finally, students should exhibit a preference for elimination of copyright assuming that they would agree that it would not lead to a dilution in the quantity and value of academic works, for they would then be able to obtain all works for free on the Internet and at cost in print; furthermore, their instructors would face no copyright hurdles in assembling teaching materials, which would also beneﬁ t them. 83 Abolishing copyright for works already in existence may offend constitutional protections of property and thus could be subject to constitutional restraint. 84 U.S. Const. art. I, § 8, cl. 8. 85 On the university’s ﬁ nancial advantage (despite their subsidization of publication fees), see note 51 above. 86 The only obvious qualiﬁ cation is that an academic author who today would make substantial royalties from an academic book would forgo this income in the absence of copyright. How- ever, as I observed earlier, note 31 above, this is not the usual situation. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 341 Opposition w ould, of course, be expected to be voiced by for-proﬁ t pub- lishers of academic works, as their copyright-derived earnings and sales are substantial. For example, in 2008 Elsevier earned approximately $1 billion in operating proﬁ t on $3.145 billion in revenue. C opyright fuels much of these proﬁ ts by enabling commercial publishers to raise journal prices sub- stantially above cost, as evidenced by the much lower subscription prices for non-proﬁ t scholarly publications. Opposition might also be expressed by some learned societies, because they would have to cede the proﬁ ts that they now obtain from their own journals. Opposition c ould come as well from parties who would view the abolition of academic copyright as unde- sirable because it might lead to erosion of intellectual property rights in a wider domain. 6. CONCLUSION I have endeavored here to examine the effects of eliminating copyright of 106 academic works and the factors determining whether that change would be socially desirable. On the basis of a number of empirical judgments— notably, that universities and grantors would tend to subsidize publication fees—I suggested that ending academic copyright would be socially beneﬁ - cial. The reader may, of course, make different empirical assessments and 87 See Reed Elsevier (2009, 14, 124), reporting that in 2008 Elsevier obtained adjusted operating proﬁ ts of £568 million on revenue of £1,700 million, and suggesting an exchange rate of 1.85 U.S. dollars per pound sterling. For other academic publishers, see Wolters Kluwer (2009, 2, 3), reporting that in 2008 Wolters Kluwer earned operating proﬁ ts of €678 million on revenues of €3.374 billion, and John Wiley & Sons, Inc. (2008, 2), reporting that in 2008 John Wiley earned $222.99 million on revenues of $1.674 billion, and since its academic and scholarly revenues are 72 percent of the total, its estimated academic and scholarly earnings were $160.55 million. Because the annual reports do not allow one to differentiate between academic works and text- books and other works written for proﬁ t, the earnings ﬁ gures I have stated are overestimates of earnings on academic works but my conjecture is that they indicate suggest the order of magnitude of such earnings. 88 See, e.g. , Carl T. Bergstrom & Theodore C. Bergstrom (2004, 897 & tbl.1), determining that “the average price per page charged by commercial publishers is several times higher than that which is charged by professional societies and university presses.” 89 See John Willinsky (2004) on the dependence of learned societies on journal revenue. See also Guterman (2004, A12), observing that many learned societies use journal subscription rev- enues to ﬁ nance conferences and other activities and thus do not favor an open access and publication fee regime. Examples of learned societies that have stated opposition to or skepti- cism about open access publication are the American Physiological Society ( id. ), the American Mathematical Society (Ewing 2004), and the American Association for the Advancement of Science, which publishes Science (Wysocki 2005, A8). 342 ~ Shavell: Should Copyright Of Academic Works Be Abolished? come to a different conclusion. My principal goal was not to persuade the reader that my empirical judgments are correct but rather to identify and clarify the factors bearing on the social desirability of ending copyright of academic works. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 343 APPENDIX I pr esent below the analysis of the model informally discussed in Section 2. The assumptions are essentially as described there. In particular, academic authors derive utility from readership of their published works; publication is performed by publishers; authors beneﬁ t from wealth and they experi- ence disutility from writing effort. Deﬁ ne the following notation: x = number of readers of a published work; x ≥ 0; t = type of an author; t ≥ 0; tu ( x) = utilit y of an author of type t fr om x r eaders of a published work; u is positive, increasing in x , and concave in x ; y = income of an author; d = disutility of writing effort; d ≥ 0; F ( t , d ) = probability distribution function over t and d . A n author knows his t and d . The utility function of an author from a published work is tu ( x ) + y – d . (1) R eaders obtain utility from a published work but no utility from unpub- lished works. Let w ( x) = utilit y obtained by readers from a published work; w is positi ve, increasing in x with w' (0) > c , and concave in x . P ublishing a work involves a ﬁ rst copy cost and a marginal cost per unit. Deﬁ ne f = ﬁ rst copy cost of publishing a work; f > 0; c = marginal cost of distributing a work to a reader; c ≥ 0. H ence, the cost of publishing and distributing a work to a positive number x readers is f + cx . (2) 90 Thus I abstract in the model from the point that readers do in fact obtain utility from reading works that are unpublished but posted on the Internet. 91 Note that since w does not depend on t , the utility obtained from a published work is the same regardless of who writes the work. This simpliﬁ cation is useful in the model, but as the reader knows, I discussed the important implications of differences in quality of works in the text. 344 ~ Shavell: Should Copyright Of Academic Works Be Abolished? I f an author writes and publishes a work and it is read by x indi viduals, the contribution to social welfare will be tu ( x ) + w ( x ) – ( f + cx ) – d , (3) for the author will enjoy utility of tu ( x ), readers will obtain utility of w ( x ), publication costs will be f + cx , and writing effort will be d . A.1. Socially Ideal Outcomes Given that a work has been written and ﬁ rst copy costs for publication have been incurred, the socially optimal number of readers is the x maximizing tu ( x ) + w ( x ) – cx . (4) L et this optimal x be denot ed x ( t), which must be positive (because w' (0) > c ) and is determined by tu' ( x ) + w' ( x ) = c . (5) Note that x ( t ) exceeds the x at which w' ( x ) = c because the author derives utility from readership, and observe that x ( t) is incr easing in t. A work should be written, and then published, if and only if ∗ ∗ ∗ tu ( x ( t )) + w ( x ( t )) ≥ ( f + cx ( t )) + d . (6) A.2. Outcomes under Copyright Let me now ask what occurs under the regime of copyright as a function of an author’s type t and d. I assume that if an author wishes to publish, the author makes a contract with a publisher involving two terms: p = price at which the work will be sold by the publisher; p ≥ 0; r = payment by the publisher to the author. If r > 0, it will be described as a royalty payment, and if r < 0, it will be called a publication fee. The price p determines readership x ( p ) according to w '( x ) = p , (7) 92 Social welfare does not include the income of authors. The motivation is the standard one that payments between parties are transfers and do not themselves absorb resources. 93 Implicitly differentiate (5) with respect to t to obtain x ' ( t ) = – u' ( x )/[ tu"( x ) + w" ( x )] > 0. 94 I f (6) holds with equality, then whether the work is written does not affect social welfare, but for expositional convenience I adopt the convention that the work should be written and make similar assumptions below without further comment. 136 Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 345 because readers can be viewed as maximizing w ( x ) – px . We know that x '( p ) < 0 by implicit differentiation of (7). The utilit y of an author given a contract (but exclusive of writing effort d ) is tu ( x ( p )) + r . (8) I assume that publishers are numerous and perfectly competitive, so that they break even on their contracts, that is, px ( p ) = f + cx ( p ) + r . (9) Therefore, r = px ( p ) – [ f + cx ( p )]. The c ontract that an author chooses is determined by maximizing (8) subject to (9), that is, by maximizing tu ( x ( p )) + px ( p ) – [ f + cx ( p )] (10) with respect to p . For simplicity, I assume that (10) is concave in p . The ﬁ rst-order condition for the choice of p when it is positive is tu' ( x ( p )) x' ( p ) + x ( p ) + px' ( p ) = cx' ( p ) or tu' ( x ( p )) + x( p )/ x' ( p ) + p = c . (11) Denote the optimal choice of p given t as p ( t ), and observe that p (0) satisﬁ es x ( p )/ x' ( p) + p = c, that is, marginal revenue equals marginal cost, so that p(0) is the standar d monopoly price. For convenience, I assume that proﬁ ts would be positive at the monopoly price, p (0) x ( p (0)) – [ f + cx ( p (0))] > 0. (12) This assumption is r eﬂ ected in Figure 2, which shows the utility of the author at t = 0 as positive. Now it follows from implicit differentiation of (11) that p' ( t ) < 0; (13) that is, the optimal price of the work speciﬁ ed in the contract falls with t , as depicted in Figure 1. At a t sufﬁ ciently high, labeled t in Figure 1, p ( t ) must equal 0; using (11), this t is determined by t = { c – [ x (0)/ x' (0)]}/ u' ( x (0)), (14) 95 Differentiating with respect to t , we obtain w "( x ) x'( p ) = 1, so that x '( p ) = 1/ w "( x ) < 0. 96 The ﬁ rst-order condition (11) is of the form F ( p , t) = 0, and the second-order condition is F ( p , t ) < 0. Implicitly differentiating (11), we obtain F ( p , t ) p '( t ) + F ( p , t ) = 0. Hence p'( t ) = p p t –F ( p , t )/ F ( p , t ), and since F (p, t) = u' ( x ( p )) x' ( p ) < 0, we have p '( t ) < 0. t p t 145 346 ~ Shavell: Should Copyright Of Academic Works Be Abolished? which, note, is positive because x' (0) is negative. As in Figure 1, let t be such that p ( t ) = c. We know that 0 < t < t sinc e p ( t) is decr easing between 0 and 2 2 3 t . Also, as in Figure 1, let t be such that r ( t ) = 0. From (9), we know that 3 1 p ( t ) x ( p ( t )) – [ f + cx ( p ( t ))] = 0, implying that p ( t ) > c , and thus that t < t . 1 1 1 1 1 2 The utility of the author (exclusive of writing effort) who has chosen his preferred contract I will designate by U ( t) (wher e the subscript C denotes copyright); hence, U ( t ) = tu ( x ( p ( t ))) + p ( t ) x ( p ( t )) – [ f + cx ( p ( t ))]. (15) The utilit y U ( t) is incr easing in t, as is illustrated in Figure 2. To show this, suppose that t' > t and observe that if an author of type t' chooses p ( t ), then his utility would be t'u ( x ( p ( t))) + p ( t ) x ( p ( t)) – [ f + cx ( p ( t))], which exceeds U ( t) sinc e t' > t. Because U ( t') m ust exceed the utility obtained if p ( t ) is C C chosen, U ( t' ) > U ( t ). C C A n author will write if U ( t ) ≥ d, for then, if the author writes, the author’s beneﬁ t from publishing will be at least the writing cost. In summary, we have the following result. Proposition 1 : Under the regime of copyright, contract prices and the utilities of authors are as shown in Figures 1 and 2. Speciﬁ cally, (a) the contractual selling price of works p ( t ) declines with t between 0 and t (determined by (14)), where p (0) is the monopoly price and p ( t ) = 0; beyond t , p ( t ) = 0; 3 3 (b) there exist t and t , where 0 < t < t < t , such that r ( t ) = 0 and 1 2 1 2 3 1 p ( t ) = c ; hence, royalties are paid to authors for t < t and 2 1 publication fees are paid by authors for t > t ; (c) the utility U ( t) of authors (exclusive of writing cost) from the contracts they choose rises with t ; (d) authors choose to write and publish when U ( t ) ≥ d . A.3. Outcomes in the Absence of Copyright If there is no copyright and an author makes a contract with a publisher, the price cannot exceed c , because I assume that, immediately after publication, the work could be copied by other publishers and, as no ﬁ rst copy cost would be borne by them, competition would lead to the work being sold at its mar- ginal cost. I now consider the two regimes described in the text. In the ﬁ rst, authors bear publication fees, and I will sometimes refer to this simply as the 151 Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 347 no copyright regime. In the second, publication fees of f are paid by other en- tities, and I will sometimes refer to this as the no copyright/subsidy regime. Author s bear publication fees. The c ontract that an author who has writ- ten a work chooses is found by maximizing (10) with respect to p , where p is bounded by c . Denote by p ( t ) the price chosen by an author (where N stands for no copyright). We know that p ( t ) = c for t ≤ t because p ( t ) > c for such N 2 t and (10) is assumed to be concave in p . We also know that p ( t ) = p ( t ) for t > t because p ( t ) < c for such t . This is shown in Figure 1. Note that when p = c , r = – f . Hence, for t ≤ t authors who publish pay a publication fee of f and for t > t , authors who publish pay a publication fee of f + ( c – p ( t )) x ( p ( t )), so that fees exceed f and rise with t . L et the utility of an author (exclusive of writing cost) who publishes and has chosen his or her preferred contract be denoted U ( t ), so that U ( t ) = tu ( x ( c )) – f for t ≤ t N 2 = U (t) = tu(x(p(t))) + p(t)x(p(t)) – [f + cx(p(t))] for t > t . (16) C 2 Note that U ( t ) < U ( t ) for t < t , (17) N C 2 sinc e authors are forced to choose a price of c r ather than a higher price p ( t ). Observe also that U ( t ) is increasing in t , for it is increasing in t for t ≤ t , it is continuous at t , and we already know that it is increasing in t for 2 2 t > t . This is illustrated in Figure 2. Authors will choose to write, and then to publish, when U ( t )≥ d . We can summarize as follows. Proposition 2 : In the absence of copyright, contract prices and the utilities of au- thors are as shown in Figures 1 and 2. Speciﬁ cally, (a) the c ontractual selling price of works p ( t) equals c for t ≤ t N 2 and coincides with p ( t) for hig her t, so is declining until t , be- yond which it is 0; (b) authors pay a publication fee of f for t ≤ t and a fee of f + ( c – p ( t )) x ( p ( t )) for higher t ; (c) the utility U ( t) of authors (exclusive of writing cost) from the contracts they choose rises with t ; U ( t ) < U ( t ) for t < t ; U ( t ) = N C 2 N U ( t ) for t ≥ t ; and C 2 (d) authors choose to write and publish when U ( t ) ≥ d . 159 348 ~ Shavell: Should Copyright Of Academic Works Be Abolished? Authors do not bear publication fees . I now assume (for reasons given in Section 3) that universities or grantors pay publication fees of f (but not more). Hence, the contract that an author who has written a work is deter- mined by maximizing tu ( x ( p )) + px ( p ) – cx ( p ) over p ≤ c . This maximand differs from (10) by a constant, f, so that it must be maximized at the same p . In other words, p ( t ) described above also describes the choice of p when publication fees of f are not paid by authors. Note that in this case, when p = c , authors pay nothing, so that for t ≤ t authors who publish pay nothing and for t > t , authors who publish pay a publication fee of ( c – p ( t )) x ( p ( t )). Let the utility of an author (exclusive of writing cost) who publishes and has chosen his or her preferred contract be denoted U ( t ) (where NS NS stands for no copyright and subsidy of fee), so that U ( t ) = tu ( x ( c )) for t ≤ t NS 2 = U ( t ) + f = tu ( x ( p ( t ))) + p ( t ) x ( p ( t )) – cx ( p ( t )) for t > t . (18) C 2 In other words, U ( t ) = U ( t ) + f , (19) NS N as is shown in Figure 2. Authors will choose to write and to publish when U ( t ) ≥ d . Hence, we have NS Proposition 3 : In the no copyright/subsidy of publication fee regime, contract prices and the utilities of authors are as shown in Figures 1 and 2. Speciﬁ cally, (a) the contractual selling price of works p ( t ) equals c for t ≤ t N 2 and coincides with p ( t) for hig her t, so is declining until t , beyond which it is 0; (b) authors pay no publication fee for t ≤ t and a fee of ( c – p ( t )) x ( p ( t )) for higher t ; (c) the utility U ( t) of authors (exclusive of writing cost) from the NS contracts they choose rises with t ; also, U ( t ) = U ( t ) + f ; and NS N (d) authors choose to write and publish when U ( t ) ≥ d . NS A.4. Social Welfare in the Presence and in the Absence of Copyright Under copyright, if an academic writes and publishes, social welfare is ∫[ tu ( x ( p ( t ))) + w ( x ( p ( t ))) – ( f + cx ( p ( t ))) – d ] dF ( t , d ). (20) {( t , d )| U ( t ) ≥ d } C Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 349 N ote that the integrand, social welfare, must be positive whenever the author decides to write and publish. To see this, observe that U ( t ) ≥ d means that tu ( x ( p ( t))) + p ( t ) x ( p ( t)) – ( f + cx ( p ( t))) – d ≥ 0. But w ( x ( p ( t ))) > p ( t ) x ( p ( t )), so that social w elfare must be positive. The explanation is that the author captures less than the readers’ surplus, so that if the author decides to publish, it must be that that adds to social welfare. U nder the regime of no copyright and no subsidy of publication fees, social welfare is ∫ [tu ( x ( p ( t ))) + w ( x ( p ( t ))) – ( f + cx ( p ( t ))) – d ] dF ( t , d ). N N N (21) {( t , d )| U ( t ) ≥ d } H ere too social welfare must be positive when the author writes and pub- lishes, for U ( t ) ≥ d implies that U ( t ) ≥ d . N C Under the regime of copyright and payment of f as a publication fee, social welfare is ∫[ tu ( x ( p ( t ))) + w ( x ( p ( t ))) – ( f + cx ( p ( t ))) – d ] dF ( t , d ). N N N (22) {( t , d )| U ( t ) ≥ d } NS I n this case, however, social welfare is not necessarily positive when authors write and publish. For example, suppose that c = 0. Then the author will write and publish if tu ( x (0)) > d , whereas it is socially desirable to do only if tu ( x(0)) + w ( x(0)) > d +f. The former may hold even though the latter does not. The problem here is that the author does not incur the publication cost in deciding whether to write and publish. N ow I wish to explain why any of the three regimes might be best. To do so, it will be useful to make reference to Figures 1 and 2, which have been explained by what has been discussed above, except for t in Figur e 2. The point t is deﬁ ned to be that where U ( t ) and U ( t ) intersect, which we 0 NS C know must occur between 0 and t because U (0) < U (0) and U ( t ) > 2 NS C NS 2 U ( t ). T o show that t < t , observe that for any t in ( t , t ), we have U ( t ) ≤ C 2 0 1 1 2 C tu ( x ( t )) < tu ( x ( c )) = U ( t ). NS First, copyright might be the best of the regimes. Consider a t < t in Figure 2, that is, in region C. Here U ( t) e xceeds both U ( t) and U ( t ). C N NS 97 This follows because, for any x x x w(x)= ∫ w'(z)dz > ∫ w'(x)dz = px. 0 0 98 The point of intersection need not be unique, but taking that possibility into account would be tedious and would not affect the conclusions. Hence, I will assume above that it is unique. 179 350 ~ Shavell: Should Copyright Of Academic Works Be Abolished? Hence, there exists a d suc h that U ( t) > d > U ( t) and U ( t) > d > U ( t ). C N C NS For this ( t , d), social welfare is positive under copyright, for the author would write and publish (see the discussion following (20)), whereas the author would not be induced to write under either of the no copyright regimes. Since the distribution over t and d might be discrete—just the ( t , d ) that was discussed could be the only pair—copyright might be the best regime. Second, no copyright might be the best regime. Consider t < t in Figure a 2 2, that is, in region A. Choose d such that U ( t ) > d , so that the author will a N a a elect to write and publish. The author will also choose to write and publish under the other two regimes, as utility under them exceeds U ( t ) in region A. Social welfare at ( t , d ) under no copyright is higher than under copy- a a right because the price is lower; the price is c rather than p ( t ), and thus readership is higher. In particular, the difference in social welfare is [ t u ( x ( c )) + w ( x ( c)) – cx ( c)] – [ t u ( x ( p ( t ))) + w ( x ( p ( t )) – cx ( p ( t ))], which is posi- a a a a tive because social welfare (4) is concave in x and, from (5), is maximized at x > x ( c ). Social welfare at ( t , d ) under no copyright is the same as under no a a copyright/subsidy. Let t > t and c hoose d suc h that U ( t ) > d and social b 0 b NS b b welfare is negative as a result of the author writing and publishing (along the lines explained after (21)). The author would not write under copyright or no copyright, for as noted, under these regimes, an author chooses to write only if that would result in positive social welfare. If we have a discrete distri- bution, with some probability mass at ( t , d ) and the rest at ( t , d ) then no a a b b copyright is best: at ( t , d ), no copyright is tied with no copyright/subsidy, a a and no copyright is superior to copyright; at ( t , d ), no copyright is superior b b to no copyright/subsidy, and no copyright is tied with copyright. Thir d, no copyright/subsidy might be the best regime. That would be so if social welfare would be positive if the author would write under no copyright/subsidy but the author would not write under either of the other two regimes. This situation is possible. Let t > t and suppose that U ( t ) = d . Then at ( t , d ), we know that it must be strictly socially desirable for the author to write (see the discussion following (20)). Consider d + ε, where ε > 0. At d + ε, the author will not write under copyright, but it will still be strictly socially desirable to do so if ε is sufﬁ ciently small. Also, note that since we know that U ( t) > U ( t) (because t > t ), ε can NS C 0 also be chosen small enough that U ( t) > d + ε. In addition, note that NS U ( t ) ≥ U ( t). Hence, the author would publish under no copyright/ subsidy C N but not under copyright or no copyright, as claimed. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 351 Proposition 4 : (a) Social welfare under the three different regimes— copyright, no copyright, and no copyright/subsidy—are given in (20)–(22). (b) Any of the three regimes might be socially best. ■ 352 ~ Shavell: Should Copyright Of Academic Works Be Abolished? REFERENCES A ssociation of American University Presses. 2009. Electronic Publishing at University Presses. Available at http://www.aaupnet.org/resources/ electronic.html. Barnett, Stephen R., Robert A. Gorman, Marci A. Hamilton, & Lloyd Wein- reb. 1998. American Association of Law Schools, Model Author/Journal Agreement. Available at http://www.aals.org/deansmemos/98-24.html. Bar tow, Ann. 1998. Educational Fair Use in Copyright: Reclaiming the Right to Photocopy Freely. 60 U. Pitt. L. Rev . 149–230. Baser, Onur, & Elda Pema. 2003. The Return of Publications for Economics Faculty. 1(1) Econ. Bull. 1–13. Benkler, Yochai. 2006. The Wealth of Networks: How Social Production Trans- forms Markets and Freedom. New Haven: Yale University Press. Bergstrom, Carl T., & Theodore C. Bergstrom. 2004. The Costs and Beneﬁ ts of Library Site Licenses to Academic Journals. 101 Proc. Nat’l Acad. Sci. 897–902. Bergstrom, Theodore C. 2001. Free Labor for Costly Journals? 15(3) J. Econ. Persp. 183–198. Berlin Declar ation on Open Access to Knowledge in the Sciences and Humanities. 2003. Available at http://oa.mpg.de/openaccess-berlin/ berlindeclaration.html. Bethesda Statement on Open Access Publishing. 2003. Available at http:// www.earlham.edu/~peters/fos/bethesda.htm. B ioMed Central. 2009. Summary. Available at http://www.biomedcentral. com/info/authors/funderpolicies/. B ird, Claire. 2008. Oxford Journals’ Adventures in Open Access. 21 Learned Publishing 200–208. Björ k, Bo-Christer, & Turid Hedlund. 2009. Two Scenarios for How Scholarly Publishers Could Change Their Business Model to Open Access. 12(1) J. Electronic Publishing. Available at http://dx.doi. org/10.3998/3336451.0012.102. Björk, Bo-Christer, Annikki Roosr, & Mari Lauri. 2008. Global Annual Vol- ume of Peer Reviewed Scholarly Articles and the Share Available via Different Open Access Options, in Open Scholarship: Authority, Com- munity, and Sustainability in the Age of Web 2.0— Proceedings of the 12th International Conference on Electronic Publishing 178–186 (L eslie Chan & Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 353 Susanna Mornati eds.). Available at http://elpub.scix.net/data/works/ att/178_elpub2008.content.pdf. Bramble, Nicholas. 2006. Note, Preparing Academic Scholarship for an Open Access World. 20 Harv. J.L. & Tech. 209–233. B reyer, Stephen. 1970. The Uneasy Case for Copyright: A Study of Copy- right in Books, Photocopies, and Computer Programs. 84 Harv. L. Rev. 281–351. B udapest Open Access Initiative. 2002. Available at http://www.soros.org/ openaccess/read.shtml. Courant, Paul, Ann Kjellberg, J. D. McClatchy, Edward Mendelson, Margo Viscusi, Tappan Wilder, Steven Yenser, & Robert Darnton. 2009. Goog- le & Books: An Exchange. N.Y. Rev. Books, Mar. 26, 2009. Available at http://www.nybooks.com/articles/22496. Darnton, Robert. 2009. Google & the Future of Books. N.Y. Rev. Books, Feb. 12, 2009. Available at http://www.nybooks.com/articles/22281. Denicola, Robert C. 2006. Copyright and Open Access: Reconsidering Uni- versity Ownership of Faculty Research. 85 Neb. L. Rev. 351–382. Dir ectory of Open Access Journals. 2009. Available at http://www.doaj. org/. Dir kmaat, Joop. 2002. Comment, Pricing and Cost of Electronics Journals. 16(4) J. Econ. Persp . 227–230. Dryburgh Assoc., Ltd. 2002. The Costs of Learned Journal and Book Pub- lishing: A Benchmarking Study for ALPSP. The Association of Learned and Professional Society Publishers. Available at www.alsp.org.uk. Else vier. 2009. Elsevier Policy on Sponsored Articles. Available at http:// www.elsevier.com/wps/ﬁ nd/intro.cws_home/sponsoredarticles. E wing, John. 2004. The Orthodoxy of Open Access. Nature Web Focus: Access to the Literature, Sept. 13. Available at http://www.nature.com/ nature/focus/accessdebate. Fisher , Janet H. 1995. The True Costs of an Electronic Journal. 21(1) Serials Rev. 88–90. Ga y, Susan. 2007. ePublishing Trends in the Academic and Professional Book Market. 3(1) Volumes 2–4. Available at www.spi-bpo-com. Genamics Journal Seek. 2009. Available at http://journalseek.net. Goldstein, Paul. 2005. Goldstein on Copyright . 3d ed. New York: Aspen. Google Book Settlement. 2009. (Original settlement November 2008, amen- ded November 2009.) Available at http://www.googlebooksettlement. com/. 354 ~ Shavell: Should Copyright Of Academic Works Be Abolished? G reen, Toby. 2009. Book Review, Scholarly Publishing Practice 3: Academic Journals Publishers’ Policies and Practices in Online Publishing. 22(1) Learned Publishing 75–77. G uterman, Lila. 2004. The Promise and Peril of ‘Open Access.’ Chron. High- er Educ., Jan. 30, 2004, A10–A14. H armel, Leslie A. 2005. The Business and Legal Obstacles to the Open A ccess Publishing Movement for Science, Technical, and Medical Jour- nals. 17 Loy. Consumer L. Rev. 555–570. H arvard University. 2009. Harvard Open Access Policies, February 2008– June 2009. http://osc.hul.harvard.edu/OpenAccess/policytexts.php. H ess, Thomas, Rolf T. Wigand, Florian Mann, & Benedikt von Walter. 2007. Management Report 1/2007, Open Access & Science Publish- ing. Available at http://openaccess-study.com/Hess_Wigand_Mann_ Walter_2007_Open_Access_Management_Report.pdf. H oward Hughes Medical Institute. 2007. HHMI Announces New Policy for Publication of Research Article, June 26. Available at http://www.hhmi. org/news/20070626.html. Hunter, Karen. 2004. Open Access: Yes, No, Maybe. Nature , Web Focus: Access to the Literature: The Debate Continues. Available at http:// www.nature.com/nature/focus/accessdebate. I oannidis, John P. A. 2006. Concentration of the Most-Cited Papers in the Scientiﬁ c Literature: Analysis of Journal Ecosystems. 1(1) PLoS ONE e5. John Wiley & Sons, Inc. 2008. 2008 Annual Report. Available at http://www. wiley.com/legacy/annual_reports/ar_2008/. Journal of Visualized Experiments. 2009. Subscriptions FAQ. Available at http://www.jove.com/index/page.stp?name=UnderSubscription. King, Donald W. 2007. The Cost of Journal Publishing: A Literature Review and Commentary. 20(2) Learned Publishing 85–106. King, Donald W., & Frances M. Alvarado-Albertorio. 2008. Pricing and Other Means of Charging for Scholarly Journals: A Literature Review and Commentary. 21(4) Learned Publishing 248–272. K ing, Donald, & Carol Tenopir. 2004. An evidence-based assessment of the ‘author pays’ model. Nature , Web Focus: Access to the Literature. Avail- able at http://www.nature.com/nature/focus/accessdebate. Landes, William M., & Richard A. Posner. 2003. The Economic Structure of Intellectual Property Law. Cambridge, MA: Belknap Press. Lemley, Mark A. 1995. Rights of Attribution and Integrity in Online Com- munications. 1995 J. Online L . art. 2. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 355 Litman, Jessica. 2006. The Economics of Open Access Law Publishing. 10 Lewis & Clark L. Rev. 779–795. Luey, Beth. 2002. Handbook for Academic Authors . 4th ed. New York: Cam- bridge University Press. M arkwood, Priscilla. 2006. Paperless Workﬂ ows in Journal Production: A Management Perspective. 19(2) Learned Publishing 115–124. M cCabe, Mark J., & Christopher M. Snyder. 2004. The Best Business Model for Scholarly Journals: An Economist’s Perspective. Nature Web Focus: access to the literature, July 16, 2004. Available at http://www.nature. com/nature/focus/accessdebate. ———. 2007. Academic Journal Prices in a Digital Age: A Two-Sided Mar- ket Model. 7(1) Berkeley Electronic J. Econ. Analysis & Pol’y , art. 2. Avail- able at http://www.bepress.com/bejeap/vol7/iss1/art2. Menell, Peter S., & Suzanne Scotchmer. 2007. Intellectual Property Law , in Handbook of Law and Economics, ed. A. Mitchell Polinsky and Steven Shavell, 2:1473–1570. Amsterdam: North-Holland. MIT N ews. 2009. MIT Faculty Open Access to Their Scholarly Articles. Avail- able at http://web.mit.edu/newsofﬁ ce/2009/open-access-0320.html. Moore, William J., Robert J. Newman, & Geoffrey K. Turnbull. 1998. Do Academic Salaries Decline with Seniority? 16 J. Labor Econ. 352–366. National Institutes of Public Health. 2009. NIH Public Access Policy. Avail- able at http://publicaccess.nih.gov/. Nature , Web Focus. 2004. Access to the Literature: The Debate Continues. Available at http://www.nature.com/nature/focus/accessdebate/archive. html. N ature Precedings. 2009. Pre-publication Research and Preliminary Find- ings. Available at http://precedings.nature.com/. N immer, Melville B., & David Nimmer. 2008. Nimmer on Copyright. New York: M. Bender. Nucleic Acids Research. 2009. NAR’s Open Access Initiative. Avail- able at http://www.oxfordjournals.org/our_journals/nar/announce_ openaccess.html. Open Access Directory. 2009. Journals That Converted from TA to OA. Available at http://oad.simmons.edu/oadwiki/Journals_that_ converted_from_TA_to_OA. Oxford Journals. 2009. Oxford Open. A vailable at http://www .oxfordjournals. org/oxfordopen/. 356 ~ Shavell: Should Copyright Of Academic Works Be Abolished? Petrowski, Mary Jane. 2007. 2006 Academic Library Trends and Statistics. Chicago: Association of College and Research Libraries. Pind yck, Robert S., & Daniel L. Rubinfeld. 2009. Microeconomics. 7th ed. Upper Saddle River, NJ: Pearson/Prentice Hall. Plant, Arnold. 1934. The Economic Aspects of Copyright in Books. 1 Economica (n.s.) 167–195. Public Library of Science. 2001. Open Letter to Scientiﬁ c Publishers. Available at www.plos.org/about/letter.html. ——— . 2005. Press release, The First Impact for PLoS Biology (June 27). Available at http://news.google.com/news?hl=en&ned=us&q=%22PLo S+Biology%22+OR+%22Public+Library+of+Science+Biology%22&b tnG=Search+News. ——— . 2009. Why Submit Your Research to PLoS? Available at www.plos. org/journals/index.php. PubMed Central. 2009. PubMed Central (PMC) website. Available at http:// www.pubmedcentral.nih.gov/. R eed Elsevier. 2009. Annual Reports and Financial Statements 2008. Avail- able at http://www.reed-elsevier.com/annualreport08/. R owlands, Ian, & Dave Nicholas. 2005. New Journal Publishing Models: An International Survey of Senior Researchers. Available at http://www.ucl. ac.uk/ciber/ciber_2005_survey_ﬁ nal.pdf. Schroter, Sara, & Leanne Tite. 2006. Open Access Publishing and Author- Pays Business Models: A Survey of Authors’ Knowledge and Perceptions. 99 J. Royal Soc’y Med. 141–148. Scientiﬁ c Council of the European Research Council. 2007. ERC Scientiﬁ c Council Guidelines for Open Access. Available at http://erc.europa.eu/ pdf/ScC_Guidelines_Open_Access_revised_Dec07_ FINAL.pdf. Shaw, Ralph R. 1951. Copyright and the Right to Credit. 113 Science 571–573. SHERP A RoMEO. 2009a. Journal Policies—Summary Statistics So Far. Available at http://romeo.eprints.org/stats.php. ——— . 2009b. Publishers with Paid Options for Open Access. Available at http://www.sherpa.ac.uk/romeo/PaidOA.html. Shieber, Stuart M. 2009. Equity for Open-Access Journal Publishing. 7(8) PLoS Biology 1–3. Available at http://dx.doi.org/10.1371%2Fjournal. pbio.1000165. Springer. 2008. Springer Open Choice. Available at http://www.springer. com/authors/mypublication?SGWID=0-145303-6-572911-0. Spring 2010: Volume 2, Number 1 ~ Journal of Legal Analysis ~ 357 Stanfor d University. 2009. HighWire: Free Online Full-text Articles. Avail- able at http://highwire.stanford.edu/lists/freeart.dtl. Stanfor d University, School of Education. 2008. SUSE Open Access Motion. Available at http://ed.stanford.edu/suse/faculty/openaccess.html. St eele, Colin. 2008. Scholarly Monograph Publishing in the 21st Century: The Future More Than Ever Should Be an Open Book. 11(2) J. Electron- ic Publishing . Available at http://dx.doi.org/10.3998/3336451.0011.201. Suber, Peter. 2007. Open Access Overview. Available at http://www.earlham. edu/~peters/fos/overview.htm. ——— . 2009a. Open Access in 2008. 12(1) J. Electronic Publishing . Avail- able at http://dx.doi.org/10.3998/3336451.0012.104. ——— . 2009b. Timeline of the Open Access Movement. Available at http:// www.earlham.edu/~peters/fos/timeline.htm. S uthersanen, Uma. 2003. Copyright and Educational Policies: A Stakehold- er Analysis. 23 Oxford J. Legal Stud. 585–609. Symposium. 2006. Open Access Publishing and the Future of Legal Schol- arship. 10 Lewis & Clark L. Rev. 733–924. Taylor & Francis. 2009. iOpenAccess & NIH Policy. Available at http://www. tandf.co.uk/journals/iopenaccess.asp. Tenopir, Carol, & Donald W. King. 2000. Towards Electronic Journals: Re- alities for Scientists, Librarians, and Publishers. Washington, DC: Special Libraries Association. That cher, Sanford G. 2008. On the Author’s Addendum. 40 J. Scholarly Pub- lishing 97–115. ——— . 2009. From the University Presses—The Hidden Digital Revolu- tion in Scholarly Publishing: POD, SRDP, the “Long Tail,” and Open Ac- cess. 21(2) Against the Grain . Available at http://www.against-the-grain. com. Thompson, John B. 2005. Books in the Digital Age: The Transformation of Academic and Higher Education Publishing in Britain and the United States. Malden, MA: Polity. Universities UK and the Research Information Network. 2009. Paying for Open Access publication charges. Available at http://www.rin.ac.uk/ openaccess-payment-fees. W ellcome Trust. 2004. Costs and Business Models in Scientiﬁ c Research Publishing. Wiley InterScience. 2009. OnlineOpen. Available at http://www3.interscience. wiley.com/authorresources/onlineopen.html. 358 ~ Shavell: Should Copyright Of Academic Works Be Abolished? W illinsky, John. 2004. Scholarly Associations and the Economic Viability of Open Access Publishing. 4(2) J. Digital Info. Available at http://journals. tdl.org/jodi/article/viewArticle/104. ———. 2005. The Access Principle: The Case for Open Access to Research and Scholarship. Cambridge, MA: MIT Press. Wolters Kluwer. 2009. 2008 Annual Report. Available at http://www. wolterskluwer.com/2008reports/pdf/2008_Annual_Report_Wolters_ Kluwer.pdf. Wysocki, Bernard Jr. 2005. Peer Pressure: Scholarly Journals’ Premier Status Is Diluted by Web — More Research Is Free Online Amid Spurt of Start- Ups; Publishers’ Proﬁ ts at Risk — A Revolt on UC’s Campuses. Wall St. J., May 23, A1. Y oung, Jeffrey R. 2002. Journal Boycott over Online Access Is a Bust. Chron. Higher Educ., May 31, A34.

Journal

Journal of Legal Analysis – Oxford University Press

Published: Jan 1, 2010

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Should Copyright of Academic Works be Abolished?

Should Copyright of Academic Works be Abolished?

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Should Copyright of Academic Works be Abolished?

Should Copyright of Academic Works be Abolished?

References (72)

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies