Open access and learned societies

Will open access prove a blessing or a curse to learned societies?

Open Access has been acclaimed by many as the business model which will transform the scholarly publishing marketplace, rescuing libraries and academics from the 'evils' of commercial publishers. At present, most academic societies' publishing operations use the same business models as those commercial publishers. So will open access bring these learned societies real benefits or will it cause them financial hardship?

The role of learned societies
Learned societies exist in order to foster and disseminate knowledge about academic subject areas. Mission statements of learned societies tend to read along fairly similar lines. The Royal Society of Chemistry's aim is "to foster and encourage the growth and application of science by the dissemination of chemical knowledge", while the Society for Endocrinology aims to enable "the advancement of public education in endocrinology". Societies work towards these aims in a variety of ways, including publishing journals, running conferences and seminars, subsidising research funding, providing travel bursaries and assistance, and funding student scholarships. In order to finance these activities a range of revenue sources are tapped, including journal subscriptions, conference delegate fees and re-investment of surplus funds.

A recent survey from The Scientist found that around 80% of all scientists belong to at least one learned society. Many find that membership of more than one society is valuable, as these may cover a generalist society (such as the American Association for the Advancement of Science, publishers of Science and a more niche group such as the American Society for Cell Biology (ASCB) which publishes Molecular Biology of the Cell. Benefits of membership, according to this survey, include participation in conferences and meetings (mentioned by 67.4% of respondents), association with fellow scientists (65.6%) and free or reduced rate subscriptions to the society's journal/s (60.1%). These findings supported the eJUSt survey of 10,000 e-journal users conducted at Stanford in 2002, which found that "the most popular reason for joining societies was to support the society's mission but the second and third most frequent motivations given were economic benefits - receiving journals free or discounted with memberships and attending conferences at reduced rates".

Publishing and learned societies
Of the 21,000 peer-reviewed journals, monographs and book series listed by Ulrich's Periodicals Directory, at least 9,250 are published by not-for-profit publishers (learned societies, professional associations and university presses) according to the Association of Learned and Professional Society Publishers (ALPSP), the international trade association which represents learned society and other not-for-profit publishers. In addition, many societies outsource the production of their journals to commercial publishers, so it is fair to assume that more than half of all peer-reviewed journals are created by not-for-profit publishers.

A learned society's publishing activities serve the central mission of the society by enabling the dissemination of information about the society's subject area. In addition, publishing revenues are often used to subsidise the society's other activities. A straw poll of ALPSP members in February 2004 revealed that the vast majority of learned societies (87.5% of respondents) generate a surplus from their publishing activities.

The publishing activities of learned societies differ from those of commercial publishers. The reasons for publishing, for example, are different. Commercial publishers are in business to maximise profits from their activities, while not-for-profit publishers use publishing to further the society's knowledge dissemination aims and, if possible, to support other society activities. Profits for commercial publishers tend to be higher than for learned societies, as are the costs to the end user. A short survey of microbiological journals undertaken by the Society for General Microbiology's Ron Fraser found that commercially-produced journals cost between three to five times as much per printed page as those produced by learned societies. This corroborated research undertaken by Bergstrom and Bergstrom in 2002 which found that "in the fields of economics and ecology, the average institutional subscription price per page charged by commercial journals is about five times that charged by non-profit journals".

Despite these differences, the tasks undertaken by both commercial and not-for-profit publishers to produce peer-reviewed journals remain much the same. Both commercial and not-for-profit publishers accept submissions from researchers, arrange for the peer review of these submissions, copy-edit accepted articles, and then publish the material either online or in print (or, most commonly at the present time, in both print and online formats). In addition to this basic journal creation process publishers also:

• Launch new journals (which may need subsidising for several years, or indeed which may not succeed at all)
• Undertake marketing of their journals
• Invest in the creation, maintenance and development of online and other electronic services, including archives of historical content
• Provide customer support for online and electronic services
• Commission and pay for review articles.

The expected cost reduction benefits from publishing journals electronically rather than in print have not materialised to the degree originally anticipated. The reason is that most journals which began life in print are still produced in both print and electronic formats and the cost of creation, maintenance and development of digital services has in many cases been much higher than expected. Elsevier, for example, is believed to have spent £45 million on its ScienceDirect service over the past five years.

Key issues for Learned Societies
One of the problems which academic societies are now facing, according to John Willinsky in his article Scholarly Associations and the Economic Viability of Open Access Publishing¹ is that the benefit of receiving a subscription to a society's journal/s as a part of the membership fee is becoming less valid because of the wide variety of ways in which scientists and academics are now able to access journals.

Most society journals are subscribed to by many university and research libraries and articles from these publications can also be found in facilities such as PubMed Central, or through an open archives repository such as arXiv. Scientists have easy electronic access to this content, often direct to their desktop and, Willinsky argues, are beginning to find that while they were previously paying for exclusivity and receiving it, they are now paying for 'exclusive' access to articles which everyone else has access to as well. Journal subscriptions may therefore be less of an incentive for society membership than they have been in the past. This suggestion is disputed by the eJUSt research programme, which found that "half of all respondents reported taking journal subscriptions and society memberships specifically to gain access to online full-text articles". In addition, the recent straw poll of ALPSP members, mentioned earlier, found that for the majority of societies (68.75% of respondents) membership levels were rising or remaining static. Easier access to articles, therefore, does not seem to be having a negative impact on societies' membership levels.

For many learned societies, publishing generates a significant percentage of total revenue. For the Society for Endocrinology, for example, publishing represents 48% of the society's revenue base. However, Willinsky's study of 20 US not-for-profit scholarly associations found that, in 1999/2000, only six out of 20 were generating a surplus from their journal publishing activities. On average, these 20 societies were losing almost $200,000 each year through journal publishing, with publishing revenues covering only around 75% of publishing costs. This was not reflected in the ALPSP straw poll discussed earlier and there could be several reasons for this divergence. For example, Willinsky looked only at US societies, while the ALPSP poll covered learned societies worldwide. Experiences in different geographical regions could prove to be very different. Of course, sample sizes in both pieces of research were small - to obtain a clearer picture it would be necessary to undertake the exercise on a larger scale.

Enter Open Access
The business model of open access is coming into a world in which society memberships seem to be stable or rising and in which many societies are able to cover publication costs and generate revenue to support other society activities. But it is also one in which journal subscriptions are falling and library budgets are under enormous pressure. Jan Velterop, in his article Should scholarly societies embrace open access or is it the kiss of death ², says that three criteria must be met for a journal to be classed as open access:

• Free accessibility to all articles published in that journal
• The copyright owner must grant any third party the right to use, copy or disseminate articles, provided that the correct citation details are given
• Full texts of all articles are deposited in at least one widely recognised open access archive (the example given by Velterop is PubMed Central for life and medical sciences)

The open access concept has arisen for a number of reasons. Two of the core reasons are:

• the 'serials crisis': growing dissatisfaction amongst the librarian and university communities over the escalation of journal subscription prices at a far higher rate than increases in library budgets, which, being funded largely by the public sector, generally increase along with or just above the rate of inflation
• the idea that in order to encourage developments in science, scientific research should be freely available to anyone who expresses an interest

The economic argument behind the latter point is that since the majority of scientific research is funded by taxpayers it should be freely available to them. There are also the added benefits of providing access to this information to those who would find it valuable but would be unable to pay. This would include those in underdeveloped nations, although there are a range of initiatives such as HINARI and AGORA to ensure that access to scientific journal content is already available free of charge or at significantly discounted rates. There is also a suggestion that scientific development can only really flourish in an environment in which researchers are aware of other activities which they might be able to incorporate or build on, thus reducing duplication.

Willinsky quotes estimates that open access journals now make up 10-20% of all online journals, although this seems high given that the Directory of Open Access Journals lists 765 journals (as at 27 Feb 2004), and Ulrich's Periodicals Directory lists 21,000 serials (even though Ulrich's includes print-only journals as well as monographs and book series). Open access journals must necessarily be electronic - providing free access to print journals would be both economically unfeasible and impractical. When we discuss open access, therefore, we are really only talking about the online publication of journal articles.

Open access is also known as the 'author pays' model. Rather than users paying a subscription to access content, authors pay a publication fee to ensure the ongoing free availability of the material. All other publication operations (peer-review, copy-editing, production, marketing, investment in new journals and systems, customer support) are assumed to remain necessary and to continue to take place in an open access world. Author fees must, therefore, cover not only the direct costs of publishing a single article but also current and potential indirect costs incurred as part of the publishing process. In some cases, author fees are not charged and the journal is subsidised by a parent institution, charitable grants or both. This is not, however, a scalable model for the entire corpus of scientific information.

The upside of Open Access for learned societies
The main benefit for learned societies in open access publishing is that enabling free-of-charge access to journal content supports the key mission statement of all societies: to disseminate knowledge and education about the society's subject area as widely as possible. Open access publishing supports this mission statement more effectively than charging subscriptions for access to content as the size of the potential audience is much larger. However, because many open access journals are newly-formed (rather than being conversions from existing products) it can be difficult to measure whether open access articles are cited more than those published in subscription-charging journals. ISI, which tracks citations and on this basis issues 'impact factors' must track a journal for three years before an impact factor can be issued, although citation information at article level is available as soon as it is entered into the database. In addition, impact factors of journals cannot meaningfully be compared across different disciplines because citation patterns vary greatly between disciplines. While research conducted in 2001 by the NEC Research Institute's Steve Lawrence found that "online articles are cited 4.5 times more often than offline articles"³, no similar figures yet exist to compare citation rates for open access vs. subscription-based material.

The potential downside of Open Access for learned societies
For those societies using journal publishing revenues to support other society activities, the foremost concern must be to identify revenue sources which can be used to replace the income lost by abandoning the practice of selling subscriptions. This could include increasing membership fees (which would probably not be a popular move given that the journal would no longer form part of the membership package), raising delegate rates for conferences and events or selling on other non-related services (some societies offer cut-rate car insurance, for example, on which they receive a commission). In the worst case scenario, losing a significant amount of revenue from journal publishing could de-stabilise the society as a whole. Implementing a new business model will be something the majority of societies will, therefore, approach with care and not a little trepidation.

Of course, some revenue can be generated by charging author fees but it is widely assumed that this will not generate the same levels of income that subscriptions have traditionally provided. BioMed Central currently charges $550 per published article but an analysis undertaken by the Open Society Institute in 2003 put the cost of publishing a single article much higher, at $3,750. ALPSP has published a benchmarking study in this area which gave first copy costs (i.e. peer review, editing, typesetting but not marketing or any overhead contribution) of £200-1200, with a median of £450. However, these costs do not simply reflect the cost for publishing an individual article, but for handling articles which are not accepted and for undertaking tasks such as marketing and investment in electronic services which benefit both authors and end users. In addition, Open Access publishers will have to invest time ensuring that articles are deposited with the relevant open access services and archives and tagged with the appropriate metadata. This is likely to become part of the marketing activity budget as it contributes towards raising awareness of the publication and of individual articles. Publishers can not rely upon individual authors to place their articles in institutional and subject archives for this purpose and self-archiving will also not be sufficient to ensure that this material is preserved for perpetuity -- another key function of a publisher's activities which is not paid for directly.

Willinsky claims that the use of open source journals publishing systems can significantly reduce publication costs for e-journals, and there are certainly tools and services available which societies can utilise for this purpose. However, many societies will wish to continue to publish in print, and while back-end publishing systems producing print and online output can be cost-effectively streamlined there remain unavoidable manufacturing and storage costs associated with print publication.

The majority of learned societies continue to publish their journals simultaneously in print and online and will in all likelihood need to continue charging for print editions, perhaps on a print-on-demand basis, even after access to the electronic edition has been made free of charge. At present, many learned society members continue to prefer to receive their edition of the journal in the print edition. Making a move directly to open access publishing and charging extra to receive a print copy could alienate a large percentage of the membership base The ALPSP found that while 75% of journals are available online, the percentage is significantly higher among not-for-profit publishers. No statistics are available for the percentage of journals which are now online-only but this is thought to be very low.

Another significant danger is that removing journal subscription charges may cause some members to cancel or fail to renew, their membership. If more than 60% state that receiving the journal free or at a reduced rate is a key reason behind their subscription, careful investigation into their likely actions if a move towards open access were made will be crucial.

One of the difficulties behind the introduction of open access models is that there is little evidence that the majority of journal authors are in any way dissatisfied with the present system. Their concerns are mainly to do with such matters as fairness of peer review and speed of publication. Consumer dissatisfaction amongst librarians and some academics has been enough to raise the issues but without author support it will prove to build a long term future for open access.

Open access will be more appealing to some learned societies than to others, depending on the discipline which the society covers. In some disciplines, particularly in the sciences, research is expensive to undertake (it may, for example, require the use of specialist equipment costing millions of dollars, such as a particle accelerator). In these areas 'author pays' publication costs are likely to make up a small percentage of research grants and are, therefore, likely to be seen as an acceptable cost. However, in other disciplines, particularly in the humanities, research is relatively inexpensive to undertake, research grants are much smaller and publication costs would make up a much larger percentage of this funding. Moving to an open access model in these disciplines will be much more difficult. Some of these difficulties could be eased by funding agencies and other bodies which support research financially. The Wellcome Trust, for example, has declared that it "supports open and unrestricted access to the published output of research, including the open access model, as a fundamental part of its charitable mission and a public benefit to be encouraged wherever possible". This includes "meet[ing] the cost of publication charges including those for online-only journals for Trust-funded research by permitting Trust researchers to use contingency funds for this purpose" but stops short of mandating the publication of Wellcome Trust-funded research in open access journals.

The current situation
Some learned societies and not-for-profit publishers have chosen to begin experimenting with open access already, although the vast majority have elected to wait until more research is available. The ALPSP is conducting research into the costs of open access publishing and the impact of this new model on the publishers which have taken it up. In addition, the UK's Science & Technology Committee is undertaking an Inquiry into Scientific Publishing and most learned societies, at least those based in the UK, are unlikely to take definitive action before the conclusion of this process.

Some society publishers have already taken up open access but they are very much in the minority. The American Society for Cell Biology (ASCB), which has 10,000 members, 80% of whom are based in the US, has gone part way to offering open access, providing free online access to articles published in its flagship journal, Molecular Biology of the Cell, two months after the original date of publication. Many other journals have done this but with a six or twelve month delay so as to retain as many members and subscriptions as possible.

The ASCB has found that subscriptions to the journal have actually continued to grow. But the Association is in the fortunate position of generating a significant surplus from its annual conference which is used to support other society activities. Experimenting with the journal may, therefore, have been easier for the ASCB than it might be for other society journal publishers.

Conclusions
The move to open access is still very much an untested concept, particularly where learned societies are concerned. While BioMed Central has been running for several years, it is a commercial operation and thus may have more in common in terms of commercial motives and intentions regarding profits with other commercial publishers. Nonetheless, learned societies also seek to generate profits (or 'surpluses') from their publishing operations, so watching the development of BioMed Central will prove of interest to all publishing operations. However, BMC's level of financial support is such that it can afford to run for several years while operating in the red. This will not necessarily be the case for many publishers, both commercial and not-for-profit, although most societies do have reserves, sometimes extremely substantial ones.

Commercial publishers have a duty to their owners or shareholders to maximise the profitability of the company in the long term. Businesses sometimes have to take risks in order to adapt to market change and experimentation with open access models could be seen as an acceptable risk for the long-term good of the business. By contrast, learned societies exist to foster and disseminate knowledge and undertaking an open access experiment which could threaten the economic survival of the society would be unlikely to be seen as an acceptable risk. Society publishers are, therefore, much more likely to continue to wait and see than to be at the vanguard of open access experimentation, although many are experimenting on a small scale.

It is likely that hybrid models will emerge, particularly in the short term and this may cause some confusion amongst libraries and readers because some material will be free to access immediately after publication, some after a few months and some will remain paid-for indefinitely. Differences in approach to the issue of open access will be seen between:

• Commercial and society publishers
• Large players and smaller players
• Publishers in different subject disciplines

Even within these divisions, publishers may use one business model with one publication and another with a different publication, either to compare experiences or because of the market situation in that discipline or surrounding that title. There is no single solution for all publishers. We are entering an age of experimentation and confusion and are likely to hear some stellar success stories but also some tales of disaster and collapse.

Kate Worlock

Kate Worlock is a director of EPS, a consultancy with offices in London and New York, specialising in the information industry.

This article originally appeared in appeared in imi Insights an EPS publication. We would like to acknowledge EPS's kind permission to republish it in this Focus.

1. Willinsky, J. Scholarly Associations and the Economic Viability of Open Access Publishing, Journal of Digital Information, 4,2,(2003)
2. Velterop, J. Should scholarly societies embrace open access or is it the kiss of death, Learned Publishing, 16, 167-169, (2003)
   
3. Lawrence, S. Free online availability substantially increases a paper's impact. Nature (Web Debates), (2001).