Select Committee on Science and Technology Written Evidence


Supplementary evidence from the Public Library of Science (PLOS)

1.   We have received evidence that businesses read more papers than they publish and would therefore put less money into the system under an author-pays publishing model. What measures do you have in place to cope with this expected drop in revenue? (Q179)

  First, it is important not to measure the contribution businesses make to science and scientific publishing too narrowly. While companies will likely spend less to publish in open access journals than they currently spend on journal subscriptions, they will still make substantial financial contributions to science and scientific publishing through partnerships with universities and, more importantly, through taxes.

  PLoS and other open access publishers will not be directly affected by companies cancelling journal subscriptions. We have never relied on subscription revenues and thus no explicit measure needs to be put in place to deal with their loss. The open access financial model is simple and direct—there are costs associated with each journal and each paper, and these are recouped through publication fees. Companies who wish to publish their work in open access journals have to pay this fee, of course, but open access publishing does not require any additional financial contribution on their part.

  In their testimony Elsevier argue that the loss of subscription revenues from companies will place an increased financial burden on public institutions. We believe this is incorrect, and that the lost revenue from companies will be more than made up for by reduced costs that will result from the restoration of market forces to scientific publishing. Total revenue in STM publishing is estimated at between US$7 and US$10 billion—or at least US$4,000 per paper. The real costs of publishing a paper in a typical scientific journal are in the region of US$1-2,000, requiring a total expenditure of, at most, US$3 billion per year on scientific publishing. Even allowing for inefficiencies and profit, the total savings to the system will almost certainly exceed the roughly US$1 billion spent by companies on journal subscriptions.

  More importantly, the long-term benefits of open access to the UK economy and society far outstrip the short-term reduction of business contributions to scientific publishing. Unfettered access to the published scientific literature will stimulate and facilitate research and development—especially for small start-up companies for whom comprehensive access to the literature is prohibitively expensive—thereby increasing productivity. Furthermore, the availability of the scientific literature in a free and open form will spawn new industries in areas in which the UK is very strong (bioinformatics, text mining). The public will benefit from this increased productivity with, for example, more effective disease treatments developed more rapidly and less expensively. The UK treasury will benefit through increased tax revenues, allowing it to devote more funds to research, scientific publishing, and other important causes.

  Financial arguments aside, it is important to point out that companies should not have to pay to access the scientific literature. We have argued that members of the public—whose taxes support scientific research—should not have to pay again to access the results of research they sponsored. The same is true of companies. Furthermore, expecting companies to pay often exorbitant prices to access the scientific literature as a way to reduce the cost of publication of government-sponsored research adds a financial disincentive to private investment research and development—in effect, a rather perverse tax on the commercial investment in research that the UK government otherwise tries to encourage.

2.   What checks and balances have you built into the system to ensure that your open access journals do not accept more articles in order to generate an increased income? (Q189)

  The natural checks and balances in the academic "economy" will guarantee that the more selective journals will be the most widely and appreciatively read, and therefore the most prestigious and most attractive to authors. Scientists, who are very value conscious in their expenditure of research funds, will generally pay more to publish in a more prestigious venue or to make their article more appealing to readers. Indeed it is not uncommon for colour charges for an article in Nature to exceed £1,000, but authors are generally happy to pay them if a colour figure makes their article more attractive to readers. There is thus a strong incentive for PLoS—or any other publisher—to maintain high editorial standards. Authors send us their papers because publishing in our journals signifies that their paper has passed through a rigorous peer-review process and has been deemed by our editors to be of significant import and interest. If we were to relax these standards in the interest of generating more money, our audience (which includes all of our potential authors) would quickly notice the lowered standards and the value of publishing in and reading our journals would be diminished.

  Moreover, for PLoS, as for most scientific societies, publishing is a central part of a larger mission to promote scientific progress and education. To compromise this mission for the sake of higher revenue would be to betray our founding principles.

  Our entire editorial and business organisation is structured to ensure that all of our journals maintain high standards and a good reputation in the community. Every paper goes through rigorous peer review monitoring by highly trained and experienced professional editors working in partnership with active researchers who serve as voluntary academic editors. Any effort to compromise the integrity of this process for financial reasons would be counterproductive as well as contrary to our principles.

  This apparent incentive to lower editorial standards for economic gain is not unique to open access publishing. Many publishers justify subscription rates to potential subscribers in terms of the number of articles their journals contain, thereby creating an economic incentive for restricted-access journals to publish more papers. Despite this incentive, the scientific publishing system has retained its integrity for the same reasons it will retain its integrity under open access.

  Finally, the possibility that publishers would compete to provide a low-cost outlet for publication of larger volumes of articles would not necessarily be a problem—indeed, very valuable information (such as the results of clinical trials that produce negative or inconclusive results) is now often left unpublished because it is deemed not to have produced a scientific result important enough to interest the readers of a journal. Society would be well served by a cost-effective mechanism for recording these results in a public archive.

3.   It has been suggested that, were the UK to mandate an author-pays publishing model, it would pay for publications twice—once to publish its own work, and the second time to read the work published in countries which had not adopted the author-pays model. The Committee would value your thoughts on this matter. (Qq202, 205)

  This is incorrect. If this were to happen the UK would not be paying twice for the same thing, they would be paying for two different things.

  By mandating open-access publication and providing funds to UK authors to publish in open-access journals, the UK government would be acting to ensure that the results of research it funds are freely accessible to scientists and the public in the UK and all other countries, furthering its goal of creating new knowledge and making it available for the benefit of its own citizens and the world.

  If the UK were the only country to make such a mandate, it would also have to pay to access the works of scientists from other countries, but this would be a separate expense for a different purpose. Furthermore, if the UK were to lead the way in mandating open access this would have a profound effect on an open access movement that is gaining momentum worldwide. Funding agencies across Europe have already provided strong endorsements of open access, as have private foundations and major universities across North America. We expect that most countries in Europe and North America (the source of most published papers) would rapidly follow the UK lead, and that publishers would see this event as the necessary catalyst for their transition to open access.

4.   In oral evidence you said that the costs of an author-pays publishing model would be lower than the costs of the traditional publishing model. Can you provide an analysis to illustrate this point? (Q205)

  Just to be clear, we were referring to the overall cost to the scientific community to publish its collective research output.

  As we and others detailed in our evidence, the costs of scientific journals have increased at a rate far greater than inflation. These increases have been accompanied by very healthy profits for most publishers. We argue strongly that these price increases are a consequence of an industry that is no longer responding to market forces, and that the restricted access business model is the prime cause of these inefficiencies. Open access will restore efficient market forces and, undoubtedly, reduce profits to a reasonable range at a considerable saving to the scientific community. This analysis is supported by stock analysts at firms like Bear Stearns and BNP Parabas who have consistently downgraded commercial publishers' stocks as open access has progressed, and by the publishers themselves whose strong opposition to open access reflects their belief that their current profit margins depend on maintenance of the subscription-based system.

  We have also carefully analyzed the costs of operation for a scientific publisher, as well as the costs of producing different types of scientific journals, and have concluded that our cost per paper will be roughly US$1,500 including fixed costs for the organisation and specific costs for each paper. Although we have not been in operation long enough to provide explicit data to support this cost, it is in line with estimates of other commercial and not-for-profit publishers of the cost of all of the steps leading to the online publication of a peer-reviewed, edited and formatted research article. Even if we assume we have underestimated our costs somewhat, it is difficult to imagine that they will come close to the lowest estimate (US$4,000) of the per article revenue across the industry. Although the elimination of excessive profits is one source of this difference, it is not the only source. For an open-access publisher, no money must be spent on soliciting and managing subscriptions—costs that can run up to 30% of budget for some publishers. More importantly, PLoS and BioMedCentral—the two largest open access publishers—make extensive use of computer technology to reduce our costs without compromising our standards. As open-access publishing becomes more widespread, market forces should act to promote the additional development of such technologies and to further drive down costs.

  Although we, and most others who have analyzed scientific publishing carefully, are convinced that open access will reduce costs, this is not the prime motivation for making this transition. Even if open access cost exactly the same as the current system, under open access we (the global scientific community) would be getting so much more for our money. Under the current system our US$7-10 billion per year covers the cost of peer-review, editing and production of all the papers we produce and buys a very limited kind of access to this material for a limited group of scientists and physicians, predominantly those affiliated with wealthy institutions in the developed world. Under open access the same (or less) money will cover the cost of peer-review editing and production of all the papers we produce and buy unlimited access to and use of this material for everyone in the world.

5.   Can you supply any data on the average cost per journal to the publisher of peer review?

  There are at least three components to cost of peer review: the cost of obtaining expert reviews from practicing scientists, the costs of managing the review process, and salaries for staff editors who are sufficiently expert to contribute to the evaluation of a paper's scientific merit.

  For the vast majority of journals the cost of peer-review itself is very small relative to the full cost of publication because it is done by academics in their own time and provided as a free service to the publisher. Some journals pay reviewers for their services, although this is rare in the sciences (eg the Lancet pays professional statisticians to review certain papers) and does not contribute significantly to the system-wide costs of peer review.

  The transition to electronic journal management systems (JMS), whereby manuscripts are submitted and exchanged electronically and reviews are submitted directly into this system, greatly reduces the overall cost of peer review. Some systems were developed by publishers, but increasingly they are licensed from vendors. The costs of these systems and their functionalities vary—PLoS, for example, pays its JMS vendor $20 per submitted article, a small marginal cost relative to the overall cost of publishing a paper.

  For a few top-tier journals like Nature or PLoS Biology, the main cost of peer review is the paid internal editorial staff whose professional judgment plays an important part in the peer review process as such journals attempt to insure that works they publish are of special significance and general interest. PLoS Biology has six expert professional reviewers and journals like Nature and Science have more than 25. These highly-trained editors can be expensive, and are the primary cost cited when journals like Science claim a cost of US$10,000 per article. It is worth noting, however, that most papers are published in journals with relatively little professional editorial oversight. For example, at a commercial publisher like Blackwell, one journal manager might oversee five-10 society publications. For our future PLoS community journals (with staffing and publication standards similar to most society journals), we estimate that peer review will cost no more than US$200 per article.

6.   How much, on average, does it cost to publish a journal article?

  The cost of publishing an article will also depend on many different factors and vary with the efficiency and scale of the publishing process (see Appendix 1—PLoS's recent white paper on Publishing Open Access Journals). Some of the factors that influence the cost of publishing include the cost of peer review, as discussed in Q.5, and the cost of processing both rejected and published articles, ie whether an electronic or human-managed system is in place and the journal's rejection rate (since rejected articles typically do not generate revenue). Once an article has been accepted for publication, additional variable costs come into play, including production processes such as graphic design and copyediting that journals may undertake and value-added components of the journal such as reviews or commentaries that add to the overall cost of the journal.

  We estimate that a typical article processed through our journal management system and production process costs between US$1,000-1,200, not including costs for peer review and print. Blackwell has estimated a figure of £1,250, which includes their current profit margin and cost of the print run. BioMedCentral estimates it processes articles for £550. The variations derive not from publishers' lack of honesty or transparency, but rather from rapid improvements in electronic publishing processes, variable value-added costs described above, and issues of scale—BioMedCentral, for example, currently publishes over 100 journals and so benefits from economies of scale more than PLoS currently does.

7.   Some disciplines are better funded than others. What measures do you have in place to ensure that there is no bias towards the richer, better funded disciplines under an author-pays model?

  The ideal system for scientific publishing is one in which all scientists can publish their works in an appropriate journal and access the works of their colleagues without impediments. The current system denies countless scientists access to journals they need for their research, especially scientists at less wealthy institutions or in poorly funded disciplines. Open access will correct this access inequity. But we must make sure that we do not create a system in which some scientists cannot publish their work in the appropriate journal.

  There are really two dangers here. First that individual authors will be shut out of particular open access journals because they can not pay the publication fee. And second that poorly funded disciplines will be poorly served by open access publishers responding to the lack of available funds in the discipline.

  The answer to the first problem—for PLoS at least—is simple. The ability to pay will never be a consideration in our decision to publish a paper. If a paper passes peer review, it will be published, regardless of author's ability to pay. We have factored into our business model and financial projections an expectation that 20% of our authors will be unable to pay our publication costs (this number is based on discussions of the frequency of requests for waivers of page charges from several journals, and is intentionally an overestimate). We expect waiver requests to come both from authors in developing countries as well as authors working in poorly funded disciplines in developed countries.

  Obviously the lost revenue from these articles must be made up elsewhere. One model is to increase charges to other authors by the amount required to make up the shortfall. Some have argued that this will result in the UK subsidising the publication of research from other countries and of well-funded disciplines subsidising publications from poorer disciplines. While this is true, we note that both of these subsidies already exist. Wealthy disciplines subsidise poorer disciplines by contributing disproportionately to library or institutional acquisition budgets. The UK subsidises the publication of authors in developing countries by subscribing to the journals in which they publish their work. Since the number of published articles from poorer countries is relatively small, subsidising their publication provides an inexpensive but high-leverage way for the UK to support development in poorer countries.

  There are also alternatives to having some authors subsidise the costs of those who can not afford to pay. Several foundations with an interest in promoting science in developing countries (eg The Open Society Institute) have allocated funds to subsidize publication charges for authors from these nations.

  There is some risk that open access publishers will respond to the relative lack of funds for publication in poorly funded disciplines by devoting insufficient resources to these fields. It will be difficult for an open access publisher to optimally serve fields that do not have sufficient resources to cover the costs of publishing their research output. (We note that the current system does not ideally serve many poorly funded disciplines.) PLoS—as an organization of scientists—is committed to serving the entire scientific community. But the best defence against a bias towards well-funded disciplines is for the funding agencies and research institutions who will primarily fund open access publishing to allocate sufficient funds to all disciplines to cover their publication costs. Publication budgets need not be allocated in direct proportion to research funding. In addition to, or instead of, including publication costs in research grants, a general fund could be created by each research organisation to provide publication funds to any of its scientists. Since publication costs will represent a modest fraction of research costs, this will not present a significant financial burden for institutions, and these costs will be offset by savings on subscriptions and other publication-related charges.

Closing statement

  What seemed an impossible ideal in 1836, when Antonio Panizzi, librarian of the British Museum, wrote: "I want a poor student to have the same means of indulging his learned curiosity, . . . of consulting the same authorities, . . . as the richest man in the kingdoms," is today within our reach. Today, we have the means to make humanity's treasury of knowledge freely available to scientists, teachers, students and the public around the world. New technology—most notably the Internet, with its capacity for distribution, storage, and retrieval of information in digital form—has brought us to the threshold of an era that scientists, philosophers and scholars have hoped for and dreamed about for millennia. At this extraordinary moment in the history of science and learning, the UK Parliament has a unique opportunity and, we hope, the vision and leadership to take us across this threshold.

  Every other argument we have made, although we believe each is compelling in its own right, is minor by comparison.

March 2004

previous page contents next page

House of Commons home page Parliament home page House of Lords home page search page enquiries index

© Parliamentary copyright 2004
Prepared 20 July 2004