House of Commons - Science and Technology

Select Committee on Science and Technology Tenth Report

7 Should the author pay?

142. Growing frustration with the STM journals market has led to the emergence of the author-pays publishing model. Under this system, the costs of publishing are met by the author, rather than the reader or subscriber. Articles would undergo the same peer review process as is employed in the current system. Once published, they would be available online, free of charge, to anyone. It is a common misconception that the author would have to meet the publication costs personally.[262] In practice, however, either the research funder or the research institution would pay the fee. Surveys that suggest that up to 50% of authors would not be prepared to pay a publication fee thus miss the point.

143. Memoranda submitted to this inquiry have tended to focus on the arguments for and against the author-pays publishing model rather than the wider issues surrounding scientific publications. During oral evidence we found it difficult, at times, to persuade witnesses to answer questions on any other subject. To a certain extent we found that the eagerness of parties either to promote or condemn the system of author payments hampered a more sophisticated discussion of the issues involved. For the Government either to endorse or dismiss the new publishing model would be too simplistic. Without any Government action, some authors are already choosing to publish in journals that use author payments to recover costs. Author-pays publishing is a phenomenon that has already arrived: it is for the Government and others to decide how best to respond.

Sustainability and costs: the evidence so far

144. Current author-pays publishers argue that the existing publishing model is based on a print environment that has been superseded by digitisation. The Public Library of Science (PLoS) argued that, "when each copy of a journal represented a significant cost for printing and distribution, it made sense for recipients to pay for each copy delivered. With the Internet now the most effective and widely used medium for communicating the results of scientific research, charging for use is now economically irrational and limiting access to subscribers is needlessly restrictive".[263] They argue that the author-pays model is more sustainable than the subscriber-pays model in the long term: "by treating the costs of publication as costs of research and including funds in research grants, monies available for publication will scale with publication expenses".[264] Both PLoS and the Wellcome Trust also maintain that the author-pays publishing model costs less than the current model overall.[265]

145. There is only limited evidence on which to base an assessment of whether or not the author-pays publishing model is sustainable or cheaper in the long term. This is largely because author-pays publishing is a recent phenomenon. In total, author-pays publishing currently represents approximately 5% of the total journals market.[266] BioMed Central was established in 2001. PLoS has been publishing its journal, PLoS Biology, since October 2003. Several other publishers have been conducting selective author-pays experiments. IoPP, for example, created one of the first journals to adopt this model, the New Journal of Physics, in 1998. More recently, Oxford University Press (OUP) announced that "there remains a dearth of factual information […] to support the arguments about the potential advantages and disadvantages of Open Access publishing. During the course of 2004 we therefore intend to run a series of experiments to test a range of business models".[267] It created a new author-pays journal, Evidence-based Complementary and Alternative Medicine, that was first published in June 2004. In addition, from July 2004, its Journal of Experimental Botany has levied a publication charge.

146. Very few independent studies into the sustainability of the new publishing model have been conducted to date. Experiments tend to be carried out at the level of individual publishers, and the results are not always shared. In April 2004, the Wellcome Trust published a report produced by SQW, independent economic consultants, on Costs and business models in scientific research publishing. The report concluded that author-pays publishing was not only more sustainable than subscriber-pays but also that it cost up to 30% less overall.[268] This is debatable: although the analysis of the costs cited by the Wellcome Trust appears to be sound, some costs are missing. For example, author-pays publishers would need sales forces to negotiate with research funders to pay fees on behalf of their researchers; they would also need a system for tracking and collecting all those fees. We do not have any evidence to enable us to assess these costs relative to the costs of selling subscriptions, collecting subscription fees and negotiating licences that are currently incurred by subscriber-pays publishers. Many publishers and other organisations have been quick to discredit the calculations on which the report was based. Sally Morris, Chief Executive of ALPSP, stated in the British Medical Journal that "it is suggested that publishing costs could be reduced by up to 30% by a move to Open Access. This is nonsense; most of the saving would be due to a move to online-only. Indeed, reduction of publishing revenues by 30% would put many very valuable journals out of business".[269] The stance of both the Wellcome Trust's report and the critiques of it may in part be motivated by the agendas of those concerned. Without further evidence it is impossible to accept as authoritative any of the cost analyses of author-pays publishing that have been carried out to date.

147. Growing frustration with the STM journals market has led to the emergence of the author-pays publishing model. Under this system, the costs of publishing are met by the author, rather than the reader or subscriber. Articles would undergo the same peer review process as is employed in the current system. Once published, they would be available online, free of charge, to anyone. It is a common misconception that the author would have to meet the publication costs personally.[270] In practice, however, either the research funder or the research institution would pay the fee. Surveys that suggest that up to 50% of authors would not be prepared to pay a publication fee thus miss the point.

148. Subscriber-pays publishers quoted us higher costs per article. Wiley told us that $1,500 (£825) would be the lowest putative cost per article, claiming that, for prestigious journals the cost would be much higher: "it is likely that a wholesale shift to a supply-side model as opposed to the current demand-side model would lead to higher article fees being required for the best journals since rejection rates are higher and quality processing would still command a premium".[271] Reed Elsevier told us that "even the highest article fees charged by Open Access publishers today ($1,500) cover only about 40% - 60% of the estimated total costs to publish an article of the quality that researchers are used to today".[272] Nature Publishing Group gave us the highest estimate, stating that the cost of publishing in Nature would be $10,000-$30,000 (£5,500—£16,500) per article published because of its 90% rejection rate.[273] Professor Fry, who opposes author-pays publishing, told us that publication costs would add 8% per year to grant costs, based on a cost of £3,500 per article.[274] Such high figures were contested by author-pays publishers. PloS, for example, told us "although many publishers have claimed that they would need to charge authors in excess of $4,000 (£2,200) to support journals by author payment, this discussion has been hampered by a lack of information sharing about the real nature of these costs".[275]

149. We suspect that the costs per article of author-pays publishing supplied to us by commercial publishers are exaggerated. Many opponents of author-pays publishing claim that the existence of subsidies demonstrates that the author-pays model is not financially viable on a long-term basis. However, PLoS told us that subsidies were only necessary as an initial investment to give the new publishing model its impetus: "we do not expect to be in profit for another 4-5 years but it is unlikely that other publishers would have to bear similar costs once open access is accepted as a credible publishing model".[276] In addition, not all journal costs would need to be met by author payments. BioMed Central generates other sources of revenue from: subscription access to commissioned articles; sales of paper copies of journals to libraries; advertising and sponsorship; and a range of subscription-based services, including literature reviews and evaluation, databases, and other software research aids.[277] All these factors would reduce the charge levied on authors.

150. Given the variety and provenance of the figures that we have received, it is clear that little is yet known of the costs involved in author-pays publishing. An understanding of these costs would be necessary before the UK embarked on any large-scale implementation of this business model because of their implications for research funders. We agree with the Director General of the Research Councils that "I do not think we have sufficient analysis of this and understanding of it in a sector that is still significantly less than 5 per cent and is being subsidised".[278] The evidence produced so far suggests that the author-pays model could be viable. We recommend that Government mobilise the different interest groups to support a comprehensive independent study into the costs associated with author-pays publishing. The study could be used to inform Government policy and strategy.

Benefits of the author-pays model

151. One of the consequences of the author-pays publishing model is that it would change the role and function of libraries. Author-pays publishing moves the costs of publication rather than removing them from the system entirely. Any transition to an author-pays model would entail the transfer of some of the library's funds to the research funders to enable them to meet publication costs. Although libraries would no longer face the mounting strain of increasing journal price increases, they would have a reduced overall budget.

152. The removal of the library from the funding equation would have the added benefit of bringing academics into the transaction. As is outlined in paragraphs 11 and 106—107, academics do not pay for the journals they read under the current system. The cost of a journal is not a factor in their decision about where to publish or what to read, separating demand from price in the STM journals market. Under an author-pays model, the cost of publication would become a consideration for the academic because it would derive from their research grants. This would make it difficult for journals to compete if they raised their charges beyond what were perceived to be reasonable levels. An article posted on a Nature discussion forum notes that, under this model, "an author, deciding where to publish, is likely to consider different journals of similar quality as close substitutes".[279] As a report by the Wellcome Trust explains: "for economic efficiency, it is better for individuals to incur the true cost of their activities. They will then modify their behaviour in some way so that the costs they incur from those actions is in some measure balanced by the benefit they receive".[280] This argument has been used by some to back up claims that the pressure to keep author charges down would reduce quality (rejecting papers costs money). This is examined in paragraphs 169—174.

153. SPARC Europe told us that author-pays publishing was intrinsically sustainable: "[it] provides a financial model that scales with increases in research funding".[281] This compares favourably with the current situation in which research funding increases at a greater rate than the funds made available for libraries to purchase research output in the form of scientific publications.

154. It is important here to sound a note of caution. Many of the arguments posed in favour of author-pays publishing have a strong ideological streak. As an example of the ideological language used by the Open Access movement, PloS posited that "publications describing publicly funded research belong in the public domain, where they can do the greatest good for science and humanity".[282] Whilst we would endorse this principle wholeheartedly, there are practical considerations that must also be understood. These considerations are explored in the remainder of this chapter.

Extending access?

155. Author-pays publishing is shaped by the perception that the current publishing model limits and inhibits access to scientific publishing. By making journal articles free at the point of use, author-pays publishers hope to solve this problem. Nonetheless, there have been claims that, because the new publishing model relies on the internet for its delivery, it has the effect of restricting access to those who are able to access and use a computer. In oral evidence, Sir Crispin Davis told us that "open access would today have the result of reducing accessibility to scientific research because it is only available on the internet […] "this would exclude some 20-25 per cent of scientists; globally it would exclude 50 per cent of scientists".[283] We are not convinced by this argument for a number of reasons. Firstly, very few researchers would now be able to work without the aid of a computer, suggesting that the vast majority of journal readers would have ready access to online journals. Secondly, in the developed world, even those who do not own a computer can gain access to one, in particular through public libraries. Thirdly, for the reasons outlined in paragraphs 45—48, online journals provide the developing world with a cheaper and more efficient means of obtaining scientific information, even when the number of computers available is limited.

156. Author-pays publishing would bring the greatest potential increase in access for groups of users that do not habitually subscribe to journals or belong to subscribing institutions. Virginia Barbour, formerly Molecular Medicine Editor at The Lancet told us that patients in particular would benefit: "everyone should be able to read freely the primary research (often funded by public agencies) upon which decisions concerning their health care are made".[284] The departmental evidence also identified groups of users that would experience increased access, writing that "free access to research papers via the internet would be particularly useful for key groups such as women taking career breaks. Constant developments within many subjects mean that people on career breaks may often lose touch with the major developments in their subject".[285] Encouraging a public that is more scientifically literate and assisting women in their pursuit of successful careers in scientific research have been two of the Committee's longstanding concerns. We support, in principle, any measure that seeks to further these aims.

157. It is still too early to tell whether or not author-pays publishing has had the effect of increasing access to scientific publications. BioMed Central reports a higher usage rate for its articles than that for articles published in subscriber-pays journals.[286] Since January 2004 OUP has published its Nucleic Acids Research database on an author-pays basis. It reported that "during the six months following initial publication, the full-text of Database issue articles were, on average, downloaded 52% more frequently than the average number of full-text downloads of other [subscriber-pays] articles published in NAR". OUP cautions that "it is difficult to be sure whether this difference is solely due to the absence of access restrictions or whether the Database issue would have generated higher than average access use had access been restricted to subscribers".[287] Nonetheless, OUP's recent announcement that it would extend the author charge to all sections of Nucleic Acids Research suggests that OUP is confident that the experiment has been successful.[288] There is also conflicting evidence. IoPP, for example reports that "the electronic readership of NJP [New Journal of Physics] is not significantly higher than our regular subscription journals".[289]

158. A recent study by Thomson ISI, a private sector information provider, examined the citation rates of author-pays journals compared to their subscriber-pays competitors. It found that they "are cited at a level that indicates they compete favourably with similar journals in their field" but also that "the wide distribution of these OA journals has not yet been shown to have any appreciable effect on their appearance in lists of cited references in other journals".[290] This is not surprising for two reasons. Firstly, citation rates are indicators, albeit imperfect, of a journal's usefulness to the research community, not primarily its accessibility. Given that author-pays publishing is likely to increase access for users, such as patients, who are unlikely to cite the material that they read in other journals, it is not surprising that they have not shown a marked increase in citation rates. Secondly, author-pays journals are relatively new. It takes time for new journals to achieve recognition and, consequently, citations. That author-pays journals are able to compete with subscriber-pays journals in terms of citations is an encouraging indication of high levels of quality.

159. Although early indications are positive, it is too early to assess the impact that author-pays publishing has had on access to scientific publications.

Access for researchers in the developing world

160. On a recent visit to Malawi we heard from the Forestry Research Institute of Malawi, part of the University of Malawi, and others that researchers currently have only very limited print journal resources. The internet has the potential to vastly increase the number and range of journal articles available to them but they are unable to afford subscription charges. A shift towards free online access, via the author-pays publishing model, would greatly increase access to journal articles for researchers in developing countries.

161. There are concerns that, by transferring the costs of the system from readers to authors in the developing world, the author-pays model would reduce the visibility of research generated there because of the inability of many authors to pay the publication fee. Blackwell Publishing told us that "the author charge is a barrier to publication which will favour richer countries and organisations and make it difficult to publish a journal with authors from, say Eastern Europe and the Developing World".[291] Currently publishers operate a number of admirable schemes to facilitate access to scientific publications in poorer countries. We see no reason why such schemes could not be adapted to meet the needs of an author-pays publishing model, with publishers paying for authors from the developing world to publish, rather than to read, articles. Whereas researchers in developing countries need access to as large a proportion of global research output as possible, the proportion of published articles originating in such countries is currently very small. In addition, many research projects in the developing world are either funded by or carried out in collaboration with partners in developed countries who can afford to pay publication charges. Developing countries would therefore be disadvantaged less by paying publication charges for a very small number of authors than they would be by buying subscriptions to all the journals that they need. By the same token, publishers could assist developing countries most efficiently by paying their publication fees and granting them free online access to journals.

162. The author-pays publishing model would be extremely advantageous to researchers in developing countries, enabling them to keep abreast of research conducted elsewhere. Financially, author charges would be less burdensome to researchers in the developing world than current subscription rates. If the author-pays model were to prevail, publishers, Government agencies and other donors would need to adapt existing schemes, such as HINARI, AGORA and INASP-PERI, to meet the demands of the altered cost recovery model.

Practicalities

Ability to pay

163. Publishing costs money. The author-pays publishing model transfers costs from the end user to the producer. There is some concern that, just as currently there are people who cannot afford to pay to read scientific journals, there are also those who would not be able to afford to publish in them. This is not a factor of the author's personal financial situation. As Vitek Tracz noted in oral evidence, "people imagine that the situation is that we suddenly ask authors to take some money from their petty cash, or away from their children and give it to some publisher who is going to publish them. That is not at all the situation. […] Just as, most commonly, scientists do not have to subscribe, so most commonly scientists do not have to pay personally".[292] If it were made clear to authors that publication costs would be met by their research funder, it is likely that their reluctance to pay would be significantly reduced.

164. RCUK observed that there was already a need for scientists to pay some publishing fees.[293] In order to give researchers a free choice of the journal in which they publish, funding to pay those fees will need to be formally identified. We believe that such funding should originate with the research funder, not the institution. By making publication costs a proportion of total research costs, this would facilitate effective planning and budgeting. Were research institutions to pay for publication costs, they would not be able to scale the model in the same way, leading to budgetary uncertainty. In oral evidence, the DGRC told us that "we are considering the possibility of making funds available to those authors that wish to go the open-access route, so they are in no sense penalised against other routes".[294] We were glad to hear that the Government has recognised the need for funds to be made available for authors who choose to publish in an author-pays journal. Given that author-pays publishing represents a growing proportion of the market, the Government will need to move swiftly from considering the issue to taking action. Currently, only 5% of all publishing takes place using the author-pays model. The funding needed to facilitate this activity at present is, therefore, limited. Were the practice to become more widespread, the amount of funding needed would increase. Such increases would, however, be balanced by savings in the library budget.

165. We recommend that the Research Councils each establish a fund to which their funded researchers can apply should they wish to publish their articles using the author-pays model. The Research Councils will need to be funded by OST to take account of this increase in costs. We hope that industry, charity and other Government funders will consider similar measures. The issue of free-riders in the publishing system will be discussed in paragraphs 175—177.

166. Publication costs paid by public funding bodies would not cover the costs of all potential authors. The City University Centre for Information Behaviour and the Evaluation of Research (CIBER) told us that not all funding sources are acknowledged, making it difficult for the funding bodies to pay for publication. According to their statistics, 35% of published articles currently have no financial acknowledgements, "either explicit or implicit from their corporate addresses".[295] The author payment may also prove difficult for charitably-funded researchers, because the charity may not have the extra funds available to cover publication costs. Differential levels of funding between disciplines further complicate the situation. ALPSP stated that "the financial feasibility of Open Access may well vary between disciplines; in some areas, particularly in the sciences, research is expensive, research grants are accordingly large, and a few thousand pounds to publish a couple of articles may make little difference to the total sum funded. In other areas, however, such as the humanities, research is relatively inexpensive and grants are therefore small or even non-existent".[296] Professor Hitchin of Oxford University noted a problem with mathematics in particular.[297] Research Councils for disciplines that require only limited funding should be funded to enable them to pay for publication costs where necessary.

167. All witnesses agreed that the ability of the author to pay should not be a factor in the publisher's decision about which articles to publish. At the moment, author-pays publishers waive author charges for those that cannot afford them. PloS told us that "if an author cannot pay, then the fee must be waived, if the peer review process judges that the article is worthy of publication".[298] PLoS editors are currently "blind" to the author's financial situation when making decisions about acceptances and rejections. In its author-pays experiment with the Nucleic Acids Research database, OUP found that, out of a total of 142 articles, 90% of authors paid the publication charge. Of the 14 authors that were granted a waiver, four were from developing countries and the remainder were from developed countries, but lacked sufficient funding to pay the charge.[299] If these statistics are indicative of general trends, under an author-pays model, the 90% of authors that pay would be subsidising the waiver for the 10% of authors that did not. This would inevitably, but slightly, push up the amount charged to paying authors.

168. The variation in the ability of authors to pay to publish is an important factor in any consideration of the author-pays model. It would not, however, have a significant adverse impact on the likely success of such a model were it to be implemented on a national or international scale. Were it to adopt the author-pays model, the UK Government eventually would make savings in the budgets currently allocated to libraries (see paragraphs 188 to 189 for an explanation of the costs to the UK of the implementation of the author-pays system if other countries did not follow suit). These savings would need to be redirected to research funders to enable them to meet the costs of publication. As PloS told us in answers to supplementary questions, "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".[300] By a similar token, Government should encourage charitable and private sector funders to meet the publication costs of their funded researchers where applicable.

Peer review in an author-pays system

169. Many memoranda expressed concern that the author-pays publishing model would compromise the integrity of peer review. This was the position adopted by all of the commercial publishers that we spoke to on 1 March. It can be summed up by a statement made by the Publishers Association that "once financial or any other type of patronage is introduced, independence and objectivity [are] compromised".[301] This argument was dismissed by Harold Varmus, of PloS, in oral evidence as "rubbish […]We have reviewers who make the determinations about what we are going to accept, who have no direct interest in the fate of our journal, but the most important thing is that we, as publishers of open access journals, want our journals to be high quality. It is the only way we are going to succeed".[302] Peer review is carried out by academics who are unpaid and whose reputation is influenced by their participation in the peer review process. In both the subscriber-pays and the author-pays publishing models, the process of reviewing articles for publication is, and needs to remain, independent of the mechanisms used by publishers to recover costs. We would envisage this remaining the case even if our recommendation in paragraph 70 were to be implemented by publishers.

170. The success of the journal is a key factor in understanding the impact of author-pays publishing on peer review. As is outlined in paragraphs 7—9 above, in order to further their career and disseminate their research to the most distinguished peer audience possible, academics seek to publish in journals that have high impact factors and are well regarded in their field. The academics that we spoke to agreed that, generally speaking, a journal would struggle to survive if the standards of the articles that it accepted for publication were low. Professor Crabbe told us that "science is a relatively close-knit area and within the field if something is wrong then people talk and if the journal does not instantly produce some sort of retraction or correction then people just will not go to publish in that journal".[303] This statement outlines two main principles: firstly that the scientific community is itself able to assess the merits of published articles; and secondly, that currently if a journal is perceived to be substandard it will be less successful. CURL told us that "the survival and growth of the BMC [BioMed Central] journals will depend, as with all journals, on the quality of submissions accepted and the number of citations to articles in their journals. This discipline should ensure that peer review remains stringent: acceptance of articles based on ability to pay will very quickly prove self-defeating".[304] Both PloS and BioMed Central told us that they wanted to publish successful journals. Whilst this may be true for the higher end of the market, in theory, if author-pays publishing were to become the dominant model, there is a risk that some parts of the market would be able to produce journals quickly, at high volume and with reduced quality control and still succeed in terms of profit, if not reputation. Such journals would cater for those academics for whom reputation and impact were less important factors than publication itself. This situation may take some time to develop because it is unlikely to occur whilst the author-pays journals are still striving to prove their worth in a mixed market.

171. We heard that payment for publication gave publishers a financial incentive to publish more articles because of the need to cover fixed costs, irrespective of the quantity or quality of submissions received. Sir Crispin Davis, for example, told us that "if you are receiving potential payment for every article submitted there is an inherent conflict of interest that could threaten the quality of the peer review system".[305] Blackwell Publishing told us that, by contrast, "the subscription based model favours rejection, which is especially important, for example, in clinical medicine, where the risks associated with publishing substandard material are higher".[306] It should be noted that subscriber-pays publishers also have an incentive to publish ever greater numbers of research articles, because, as is shown in paragraph 52 of this Report, increases in the volume of articles are used to justify price increases. For the reasons outlined above, we believe that the publishing process needs to have inbuilt checks and balances that would mitigate against the acceptance of an increasing volume of substandard articles.

172. In order to succeed, most author-pays publishers, like everyone else, will have to publish articles of a high quality. It is not, therefore, within the interest of journals at the higher end of the market to lessen the rigour of peer review. Nonetheless, there is a risk that lower quality journals might seek to reduce their quality threshold in order to generate profit. Were the author-pays publishing model to prevail it would be vital to ensure that peer review was not compromised in order to retain confidence in the integrity of the publishing process.

173. The stringency of peer review, a potential strength of author-pays publishing, has also created problems for author-pays publishers when determining their payment models. John Cox Associates, a publishing consultancy, noted that in the author-pays model, because of the need to peer review every article, successful authors would be paying for the processing of unsuccessful articles: "if the author whose paper is accepted for publication bears the whole cost of publication, this means that he/she bears the cost of processing papers that are rejected; it seems inequitable that a successful author should bear the costs involved in dealing with others whose work is judged not to be worthy of publication. This will result in the most prestigious journals being the most expensive".[307] The Biochemical Society also noted that "The Biochemical Journal, along with many other journals has a rejection rate of 60%. It seems invidious that the authors of accepted papers should subsidise authors of rejected papers".[308] A 90% rejection rate is the reason given by Nature Publishing Group for its very high estimate of the publication charge for accepted articles.

174. The payment model is easily adapted to subvent this inequality. Under a revised system, authors would pay a small fee upon submission of their article in order to cover the costs of peer review. If their article were accepted, they would then pay a second fee to cover the costs of publication. Neither PloS nor BioMed Central employ this system. Vitek Tracz explained that, in the case of BioMed Central "the reason we do not do it is because we are still a young industry and we worry that if we start charging for submission it will be harder for us to persuade authors to do it".[309] This is probably true in a mixed market in which authors have the choice of publishing in a journal which charges no fee. The reluctance of authors to pay a submission fee would largely be overcome, however, if funds were made available for this purpose by the research funder. It is also worth noting that a submission fee might act as a disincentive to authors submitting articles speculatively, with little confidence that they will be published. This would reduce the need to reject so many articles and, in turn, reduce costs. The introduction of a submission fee would be an important step towards ensuring the quality of scientific publications and we strongly recommend that author-pays publishers introduce this system.

Free-riders

175. A substantial proportion of the total number of journal readings derives from the industrial and corporate sectors. Reed Elsevier, for example, estimates that 25% of its revenue and 20% of annual global STM journal spending derives from these sectors.[310] The Pharma Documentation Ring described the usefulness of journals to pharmaceutical companies: "new drugs, new cures against diseases can only be successfully and efficiently developed when the present scientific knowledge is easily available and reliable. […] Scientific publications are a key resource for pharmaceutical research".[311] Companies pay substantially more for access to publications than public sector organisations, which are usually given discounted rates. Yet, for various reasons, the proportion of published articles that originates from the industrial and commercial sectors is much lower than the proportion of articles that they read. As a consequence of this discrepancy, the Biochemical Society told us that, "in the open-access world it would appear that the only real winners are going to be corporate pharmaceutical companies who would no longer have to pay to access information".[312] There is concern that author-pays publishing would allow companies to become free-riders in the publishing process.

176. The author-pays publishers that we took evidence from were unable to suggest a satisfactory solution to this problem. PloS told us that the commercial sector "pay their taxes, the taxes go to government, the government agencies pay for publication and want the industries to see the results of research because one of the reasons we do medical research is to support industrial efforts in making new products which help to improve the health of the nation".[313] This is true: giving industry access to research findings can have significant benefits for the economy and would assist the transfer of knowledge between academia and industry. It might also have the effect of attracting increased industry funding into research. Nonetheless, there are practical implications of allowing the commercial and industrial sectors to read the output from scientific research free of charge. PloS argued that "we have never relied on subscription revenues and thus no explicit measure needs to be put in place to deal with their loss".[314] It is indisputable, however, that a significant proportion of publishing costs are currently paid for by industry. Even though they are disputed, publishing costs do not fall below a minimum level whatever the publishing model. If industrial subscriptions are taken out of the system, the publishing process will be substantially less well off. Even taking into account a reduction in publisher profit margins, this would mean that the UK was paying more overall to publish articles than it would pay to read them.

177. The commercial and industrial sectors currently contribute significant funds to the publishing process through payments for journal subscriptions. Much of this money would be lost to the system if an author-pays model were to prevail. This is one of the key issues that needs to be addressed before the wholescale transition to an author-pays model can be supported. Government, publishers and industry need to work together to identify a solution to this problem in order to avoid a disproportionate increase in the amount of money that Government invests directly or indirectly in the publishing process.

Impact on learned and professional societies

178. Learned societies use income generated by their publishing activities to subsidise other society activities, such as conferences, maintaining professional standards and supporting education in schools. Many learned societies expressed concern that the author-pays model would jeopardise their publishing income and, ultimately, their other activities. The British Pharmacological Society told us that "In 2002—3 we spent over £850,000 on promoting and advancing pharmacology. Nearly £800,000 of this came from our publishing activities. Without this income we should either have to raise funds in a different way or cease to provide most of our current activities".[315] Similarly, the Biochemical Society reported that it generated between £500,000 and £750,000 from publishing. This surplus, "the so-called 'science dividend' supports its other charitable objectives and it is not clear how these vital activities could continue to take place if the Biochemical Society was unable to continue funding them".[316] These concerns are based on the perception that the author-pays model would generate a reduced publishing surplus. Indeed, it is one of the stated objectives of the Open Access movement that publisher profit margins should be reduced. ALPSP told us that "one estimate is that surpluses in excess of 12% would be likely to be reduced. In the above survey, median surplus was 17%, representing 33% of society income".[317] On the basis of these figures, learned societies could lose up to 10% of their income in a move to the author-pays publishing model.

179. The transition to an author-pays publishing model involves substantial one-off costs and continuing subsidies until the journal in question reaches revenue stage. IoPP told us that, since 1998, it had been subsidising its New Journal of Physics at a rate of £100,000 per annum. It also stated that, at $560 per article, it had not been charging authors full costs: "we have found authors reluctant to pay. Even if all authors paid, we would need a far higher page charge in order for the journal to cover its costs".[318] Smaller societies publishing less-prestigious journals might find it difficult to attract authors with a more substantial publication charge. They would also be more likely to struggle with the levels of investment committed by IoPP to its author-pays project.

180. Learned societies are greatly valued by the academic and wider research community. It is of concern to us that learned societies could stand to lose a substantial portion of their income in a move to the author-pays publishing model. This is another key issue that proponents of the author-pays model need to address.

181. Journal subscriptions are frequently provided as one of the benefits of membership to a learned society. The Geological Society told us of its concern that "if free or discounted subscription to these journals is one of the benefits of society membership, open access will negate this reason for membership - if you can read the journal for free, why join the Society? The likely decline in Society membership would be a fatal blow for many societies resulting in their closure".[319] Just as, in the commercial sector, it is intended that author payments would replace subscription charges, the loss of some learned society members would in part be compensated for by the publication charge levied on authors. Moreover, it is clear that free subscriptions are a major, but not the only benefit given to members by learned societies. We heard from Professor Fry that "it is absolutely vital for science in Britain and the world today to have strong learned societies". He added that "conferences, which are largely supported by the learned societies, are absolutely vital".[320] If learned societies are valued by their communities, which we believe to be the case, members are likely to remain loyal irrespective of the publishing model employed by their society.

182. The author-pays model would yield some benefits for learned societies. By facilitating a wider dissemination of their journals than is currently possible, it would raise their international profile. The Society for Experimental Biology also told us that a change in publishing model would encourage societies to innovate: "an author pays open access model will put pressure on societies to keep publication costs of primary research to a minimum and provide incentives to explore other sources of income. Adding value to the sections of the journals which remain under subscription control such as reviews and special issues will not only provide income to support the activities of the society but also fulfil the society's objectives in promoting and communicating science".[321] Some learned societies also house a library. Like academic and public libraries, these libraries are adversely affected by rises in subscription prices. The Institution of Civil Engineers stated that "premium prices for e-journals are making it very difficult for specialist libraries such as ourselves to maintain our subject coverage. […] Permitting learned societies to accede to academic networks which are essentially publicly funded would help us to continue to fulfil our role as laid out in our charter to promote engineering science".[322] These would all be encouraging developments, but further analysis will need to be carried out to assess to what extent they would offset losses elsewhere.

Government action

183. We received many calls for publishing in an author-pays journal to be made mandatory under the terms of research grants awarded from public funds. The World Summit on the Information Society (WSIS), for example, argued that "concerning the policy that should be adopted by funding entities, it is clear that because of sheer inertia, or because of fear of retribution from publishers, a compulsory policy must be adopted".[323] Whilst we do not believe that there is sufficient evidence to support the implementation of a Government mandate, we do not see this as an excuse for the current lack of any coherent Government policy on author-pays publishing.

184. In oral evidence, the DGRC told us that a decision in favour of the author-pays model would be "a pretty brave decision" for a government to take at this stage.[324] Some Government agencies have been less cautious, however. The Food Standards Agency was recently reported to be "moving towards actively supporting open access publication of the research it commissions".[325] Similarly, JISC has implemented two schemes to encourage experimentation with the author-pays publishing model. In December 2003, it announced a funding programme to offer short-term funding, or seed money, of £150,000 to publishers to make journals freely available on the internet using open access models. In a press release it announced that "the money is to encourage them to switch 'from the traditional subscription method to a model where authors have to pay to have papers published'. The result of this is that research journals will become freely available on the web for everyone in further and higher education and beyond to benefit from".[326] In June 2003, JISC announced a membership deal with BioMed Central. Under the terms of the deal, from 1 July 2003, article-processing charges were waived for all UK higher education staff when publishing in any of BioMed Central's journals. The deal cost £85,000 in the first year, and will cost £80,000 next year.[327] We strongly support further experimentation with the author-pays publishing model. In the short term Government may need to provide limited financial assistance to encourage publishers and institutions to take part in what, for them, may be an expensive process. We applaud the Joint Information Systems Committee for providing funding for this purpose so far and hope that it will continue to do so.

185. A wholescale transition to the author-pays publishing model would have profound implications for current funding structures, necessitating the transfer of funds between DfES and OST. Much of the evidence agrees that the author-pays publication model will become more prevalent, even if it does not replace the current publishing model, in the UK and globally. It is therefore likely that the UK STM publishing market will have to sustain a mixed economy for some time. This would be likely to prove costly for publishers, research funders and libraries alike. Given the scale of the financial impact of the new publishing model, whether or not it prevails, we expected that the Government would have conducted impact assessments and formulated a strategy. In oral evidence, Government agreed that this was a necessary step. The DGRC told us that "there ought to be a policy put together jointly between DfES, OST/DTI and DCMS".[328] We agree wholeheartedly. There is, however, scant evidence as yet of the departments involved working together on this issue. Author-pays publishing is a growing phenomenon. Its implementation on any scale will have important consequences for current funding structures and the UK publishing industry. So far the Government has shown little inclination to address this issue.

186. Government has not shown much evidence of a joined-up approach to the challenges posed by changes to the model for scientific publishing. Whilst the central departments have been slow to respond to the author-pays publishing model, at least two Government-funded bodies have given public support to it. This creates unnecessary confusion. We recommend that it formulate a coherent strategy as a matter of urgency.

187. This inquiry took place against the backdrop of an already lively debate. It soon became apparent to us, however, that discussions mostly tended to take place within separate interest groups. The relative lack of discussion between these groups has hindered any assessment of the potential impact of the author-pays publishing model and has prevented any sort of consensus view from emerging. Government has a role to play in bringing together policy-makers, funding bodies, publishers, librarians and academics to discuss the issues in a constructive way.

Can the UK act alone?

188. The UK writes more articles than it reads. As a consequence of this, Reed Elsevier claims that a Government mandate for open access would be against the UK's financial interests: "while Britain's spending on journal subscriptions currently amounts to 3.3% of the world's total, UK researchers contribute a much higher 5% of all articles published globally. As a result, we estimate that the UK Government, foundations, universities and researchers could together pay 30-50% more for STM journals in an Open Access system than they do today".[329] This statement can, of course, be inverted. Compared to its share of global STM research, the UK is currently paying proportionally less to access the research findings of other countries. Countries with lower research outputs, particularly in the developing world, are, on the other hand, paying proportionally more to access the same findings, a phenomenon which acts as a brake on further development. One of the main advantages of an author-pays system is that publication costs would scale with research costs: countries which invest greater amounts in research would pay more to see that research translated into outputs. Author-pays publishing would ensure that countries were paying an amount for publication that was commensurate with their total research spend. We are satisfied that, by scaling publication with research costs, the author-pays publishing model would ensure a fairer global distribution of the costs of publishing research findings.

189. A more immediate problem is the impact on the UK of a hybrid global publications market. Author-pays and subscriber-pays publishing models are alternatives: costs are met through publication or subscription charges. If the UK were to adopt a policy in favour of author-pays publishing, it would be at a financial disadvantage unless others acted in the same way. By acting alone, the UK would assume the full costs of publication for all its publicly-funded researchers. The resulting articles would be exported abroad where they could be read free of charge. This would increase the impact of UK research. Yet UK scientists also need access to global research. Thus the UK would still have to pay for subscriptions to journal articles originating abroad. We asked PloS how it would respond to this problem. It answered that "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".[330] This answer does not alter the fact that, by paying both to publish and to read articles, the UK would be investing greater sums in the publishing process than is currently the case. The UK would put itself at a financial disadvantage internationally if it were to act alone in mandating publicly-funded researchers to publish in author-pays journals.

Conclusions on the author-pays publishing model

190. The arguments for the author-pays publishing model are in many ways attractive despite some difficulties which require resolution, and we believe that its implementation would yield many benefits for the global research community. The endeavours of author-pays publishers such as PLoS and BioMed Central to widen access to scientific publications are admirable and they are to be commended for the vigour with which they have pursued their aims. We have recommended that funds be made available from Research Council budgets for those authors that wish to do so to publish in journals that impose publication charges.

191. We are satisfied that the implementation of an author-pays publishing model would not compromise peer review at the higher end of the market because it would not be in the interests of the publishers concerned to allow this to happen. We do, however, have concerns about free-riders and the potential impact of the new publishing model on learned societies. For this reason, we have recommended that the Government conduct a study and facilitate further experiments to ascertain what the likely impact of the author-pays publishing model will be. We hope that, subject to the resolution of these problems, the conditions can be created so that the author-pays model and other new publishing models can be allowed to flourish, to demonstrate whether or not they are sustainable and to confer their anticipated benefits. This aspiration should be viewed in the context of the current situation, which, as we have explained, is unsatisfactory. We have recommended that Government prepare for change in the formulation of a comprehensive strategy.

262 Q 164, Vitek Tracz Back

263 Ev 270 Back

264 As above. Back

265 Ev 275 and Wellcome Trust, Costs and business models in scientific research publishing, p 2 Back

266 Q 328 Back

267 Oxford University Press, "Experimenting with Open Access publishing", www3.oup.co.uk Back

268 Wellcome Trust, Costs and business models in scientific research publishing, p 2 Back

269 The British Medical Journal, 8 May, 2004;328:1094 Back

270 Q 164, Vitek Tracz Back