Information Today, Inc. Corporate Site KMWorld CRM Media Streaming Media Faulkner Speech Technology DBTA/Unisphere
Other ITI Websites
American Library Directory Boardwalk Empire Database Trends and Applications DestinationCRM Faulkner Information Services Fulltext Sources Online InfoToday Europe KMWorld Literary Market Place Plexus Publishing Smart Customer Service Speech Technology Streaming Media Streaming Media Europe Streaming Media Producer Unisphere Research

Magazines > Online > Mar/Apr 2005
Back Index Forward

Online Magazine

Vol. 29 No. 2 — Mar/Apr 2005

Open Access or Differential Pricing for Journals: The Road Best Traveled?
by David Stern

Open access (OA) is becoming a reality, with new cost models under development. The various cost models will have serious short- and long-term implications for libraries and dangerously impact the scholarly communication network. I believe that the adoption of the OA model for journals will create serious instabilities within the existing scholarly publication industry. OA, as a business model, is neither necessary nor desirable. With or without the often-discussed author charges approach, it would be almost impossible to obtain the same amount of total revenue through selected libraries as now exists from the much larger base of library subscriptions. Tiered or differential pricing (and services) among the existing subscribers would be a far more logical approach to supporting a modified scholarly journal distribution network.


The OA model now being explored and promoted by selected publishers is actually the result of intentional corporate misdirection. Originally, publishers used this tactic in hopes of appealing to our altruistic tendencies, thereby deflecting attention away from unreasonably inflated prices [1]. The "more people than ever have access" position was used to rationalize high prices. This OA position was never proposed as a viable alternative business model, and the publishers that offered the OA rationale never wanted to actually create a different pricing model. In retrospect, this OA movement has become a danger to the very people who used it as a diversionary tool.

OA has now been adopted primarily by selected prestigious publishers and endorsed by certain start-up journal publishers. Meanwhile, many of the traditional publishers see this as a long-term danger that will significantly reduce the number of organizations making up the present revenue base.

From a business point of view, OA creates serious concerns. It reduces the stability of the system without offering a viable substitute for the existing revenue stream. A business analysis of the OA model will demonstrate the flaws within the assumptions and will revise the approach before it becomes a de facto option. It is the responsibility of publishers, librarians, serials agents, and authors/editors to review this OA business model. This article will attempt to place this review within the context of both the current and future scholarly information networks. It will compare OA possibilities with both existing subscription models and alternative, revenue-generating journal cost models.

Until now, many of the conversations about OA journals have resembled conflicting religious opinions rather than evidence-based conversations. Great confusion and emotion have been evident as people within the industry run toward and away from this new paradigm at meetings. Unfortunately, the ideas of author charges and OA have become synonymous. The separation of a general theory and a particular implementation is the first problem that must be addressed when discussing this new approach. In addition, serious evaluation must be based upon long-term and seriously analyzed business practices and economic realities.



Often, the facts inherent in the journal publishing industry are not clearly identified at the start of the conversation. Here are a few of the key issues to remember:

• Peer-review is important and must be maintained for the scholarly system to remain viable.

• There is a real cost for peer review, based upon the
required infrastructure.

• There is a cost for distribution, both in online and paper methods.

• There is a cost for the online infrastructure support (editing and archiving).

• There will be costs for enhanced linking options in the online environment.

• There is degree of revenue redirection in the present system for most publishers (commercial profit and societies subsidizing member dues).

• Technology makes possible less-expensive peer-review and networking options.

• It is now possible to create integrated scholarly networks for peer review and for the distribution of other materials (teaching tools, electronic communities, etc).

• Networks would be less expensive if they were shared and unnecessarily redundant features were removed.

• Subject-based centers of excellence would be logical and cost-effective network hubs.

• These costs must be supported by participants with a variety of interests and reasons for maintaining this
information network.

Once all of these elements are taken into consideration and a logical enhanced online network is planned for subject communities, one must compare these possibilities and realities to the business considerations of the OA model.


Open access journals do not provide a significant solution to current scholarly information distribution concerns. While OA publication may seem like a great idea at first, there are reasons why OA for journal articles is not the right approach to distributing scholarly information.

First, there is a fatal flaw in the pricing model of OA. Interest in maintaining the same support for peer-reviewed material is not equal within the existing revenue base. The majority of authors willing to support some form of scholarly publication reside in academia. However, many readers are located in commercial organizations having little interest in altruistic support if the information can be obtained freely through OA. If possible, these readers will simply drop direct support and expect additional taxpayer dollars to subsidize this process. If OA were adopted, the entire revenue stream would then need to be channeled through far fewer institutions—requiring a much higher budget for academic libraries. Given the current financial situation and existing allocation patterns, plus the obviously outrageous inflation and profit by commercial publishers, obtaining higher budgets for academic institutions would be very difficult to accomplish. Even if accomplished, it would not correct the existing problems of commodity trading within the academic publication world. It would merely delay addressing the problem and shift more of the problem to the individual taxpayers rather than to the corporate world. Clearly, there would be a sharper divide between the haves and have-nots over time as taxpayers located in different research communities supported such higher prices at differing levels.

The initial flaw of creating a reduced revenue stream based only upon production-centered vested interest is even more problematic if revenue is also based upon author charges. With OA, online journals are available freely on the Internet to all libraries without a subscription. Decisions to maintain paid subscriptions will be based solely upon (1) a need to maintain a paper copy, or (2) significant savings obtained through reduced author charges. As few libraries will feel the need to maintain paper archives, and fewer will see any benefit from reduced author charges, there will be far fewer libraries remaining as a revenue base compared to the present revenue sources from academic institutions. Only a few institutions will generate enough author publications to see any benefit. In addition, there are currently no existing channels for blending joint revenue sources between library funds and direct author funding. This would be an auditor's nightmare to develop, is not desirable by libraries or authors, nor is it necessary in order to maintain a working revenue model.

OA will result in far fewer commercial revenue sources, and many lost academic library subscriptions. The remaining libraries will be asked to cover the entire revenue base—and this is simply not viable. As a business model, the reliance on fewer vested and altruistic supporters simply is not practical.

In addition, OA would also mean the loss of existing subsidies that support the maintenance and development of long-term archives such as the JSTOR consortia. OA would create a system more like HighWire, with reliance upon annual subscriptions to support OA infrastructures. There would once again be even greater reliance on fewer annual contributors. As a business model, the reliance on fewer vested and altruistic supporters to maintain archiving is also not practical.

In short:

• OA based upon author charges. OA based upon author charges will reduce the number of subscribers—creating complex changes in the budget processes and problems with maintaining the existing revenue base.

• OA without author charges. OA regardless of author charges will still reduce the number of subscribers—creating a serious problem in maintaining the existing revenue base.


Nucleic Acids Research is one of the first major titles to move to this OA model—and there will be a significant reduction in support unless (1) many libraries feel a burning need for paper, and/or (2) there are many libraries that expect to see and manage real savings from repeated author charges. Even with such a prestigious title, this seems to be a rash action reminiscent of the old Britannica OA attempt. Even if it were to work for the prestigious NAR, it certainly does not seem a viable model for the many less-prestigious titles with far fewer authors and institutions willing to support these less-valuable entities. Publish-or-perish pressure supporting unaccountable library expenditures can only reach down so far into the existing and exploding titles.

Nucleic Acids Rqueues

• Yale authors: 10 in 2002, 18 in 2003, estimated 22 in 2004

• Cost model of $1,500 per article (full cost without license) = $33,000 for the year

• $500 reduced institutional cost = $11,000 in author charges

• Savings = $22,000

At first this looks like a good deal, but the current Yale expenditure is only $2,855 for paper plus online without the OA option.

In purely economic terms, this model makes no sense for Yale. Are libraries willing to shift the burden to a few institutions, hoping these will find a way to discover new money? Are there enough institutions willing to pay based upon these "proposed" savings? Under this model, an organization would need to publish five articles to break even, and more than five to see any "savings." How many organizations would fit within that category? Not many, and not all of these could afford such a price increase in order to subsidize the rest of the world.

Other Popular Offerings

The following two well-known OA initiatives started with entirely new journals. Without an existing subscription base, a pricing model was developed that initiated institutional fees as a supplement or alternative to author charges. Both initiatives include differential pricing models for the institutional membership fee. Membership means that these organizations agree to support the archival infrastructure for all previous authors and memberships and for the rest of the world.

Public Library of Science (PLoS)

This initiative offers reduced author fees within membership organizations. As an institutional member, authors can publish at a 50 percent discount ($750) on the $1,500 2004 publication fee. Depending upon the number of papers published, this can become very expensive; further, it would be impossible to project costs and would require complex author payment support mechanisms. Compared to subscription plans with known annual costs and one-source payments, this does not seem like a preferred method for both logistical and budget planning reasons.

The Public Library of Science journals are PLoS Biology [] and PLoS Medicine, [].

BioMed Central

This initiative has decided to charge no author fees within membership organizations. As a benefit of library membership, the $500 article processing fee is waived for all articles published in one of BMC's 60 peer-reviewed journals. In addition, articles are immediately archived for permanent access in PubMed Central by the U.S. National Library of Medicine, which is part of the National Institutes of Health. Membership means that these organizations may not be supporting an archival infrastructure for all previous authors and memberships if an outside archive platform is trusted. At first, compared to traditional subscription plans with known annual costs and one-source payments, this does seem like a reasonable charging method; however, there is still the issue of significantly fewer organizations needing to pay, as well as much higher charges for the author-based organizations. This voluntary institution fee provides an adequate charging mechanism, but the OA altruistic vested interest contributor aspect still creates a problem in maintaining a broad and viable revenue stream compared to the current larger revenue base.

Yale has additional information on BioMed Central and how to publish in a BMC journal [].

Other Sample Journal Costs

The following individual journal scenarios demonstrate the real and projected costs of traditional and OA models—using a "reasonable best guess" standard cost-per-article value.

Journal of Insect Science

An inexpensive OA journal based within the University of Arizona, with shared responsibilities among the faculty and the library. Estimates from the staff are $850/author per paper. There are no author charges now.

• Produced 25 articles in 6 months of 2004

• 50 articles per year @ $850/author = $42,500

• if 20 libraries subscribe, the annual cost for each is $2,125 per year

• if 60 libraries subscribe, the annual cost for each is $708 per year

• if 120 libraries subscribe, the annual cost for each is $354 per year

Clearly even a small and inexpensive journal is much easier to maintain by a larger number of institutions. This situation scales even more as the number of titles is expanded and the journals become more expensive. At $850, author charges would be outrageous for authors from small organizations. Continued absorption of these costs will need to be through either sharing among many organizations or through one larger, centralized funding source.

Physical Review B

An important and relatively inexpensive society journal based upon impact factor and price-per-page. We will assume the same estimates of $850/author per paper.

• 497 articles for year 2003 @ $850/author = $422,450

• OA author charges: Yale authors: 17 in 2002, 18 in 2003, estimate 20 in 2004 = $17,000

The current Yale expenditure is only $8,740 for paper plus online without the OA option.

• if 20 libraries subscribe, annual cost for each is $21,123 per year

• if 60 libraries subscribe, annual cost for each is $7,041 per year

• if 120 libraries subscribe, annual cost for each is $3,520 per year

• if 200 libraries subscribe, annual cost for each is $2,112 per year

Clearly, a large and inexpensive journal is much easier to maintain by a larger number of institutions. At $850, author charges would still be outrageous for many organizations. As the OA costs are far higher than current subscription costs or the proposed shared cost, continued stable revenue will be best provided through either sharing among many organizations or through one larger, centralized funding source.


Perhaps differential pricing might be a way to adjust more fairly the costs across the existing subscriber base. These differential pricing models could be developed in relation either to the level of access or types of customized services desired. The American Physical Society, the publisher of Physical Review B, already offers this type of tiered pricing model based upon research level of the organization. Other publishers use this model, and additional differential pricing models base costs upon actual use.

Using the same Physical Review B cost information to develop a simple differential pricing model, let's assume two rates: with 200 libraries at full rate and 200 libraries at half-rate. This yields the following numbers:

• higher = 75 percent = 316,838 = $1,584 per library

• half rate = 25 percent = 105,613 = $528 per library

This differential pricing model seems to be a far more practical way to share journal costs without endangering the entire scholarly distribution network. Differential pricing creates a larger but fairer revenue base, and a less expensive per-library model than OA. Why would libraries willingly want to move away from a model that includes many types of existing and reasonable revenue support streams toward one that is far more expensive and has questionable and potentially flawed business assumptions? Moving to an OA pricing model does not address or adjust the problems created by commercial inflation. Changes need to be made in order to develop a more powerful and less-expensive scholarly distribution mechanism, but OA does not seem to be a solution.


Clearly, there are also unintended consequences of accepting the OA model without significantly changing the underlying peer-review and commercial aspects of the existing journal publication system. It is therefore time to ask a few questions about these traditional publication features.

First, are we happy with what exists? Do publishers produce enough added value for their commercial surcharge? Is the added value what users and librarians want and/or need, and is it worth the cost? Technology allows for significant changes in copy editing and formatting, and these might change the expectations and costs. To be an important factor, OA models must adequately address these escalating costs for both traditional offerings and new R&D options. In terms of OA impact, it is likely that these improvements will only be supported by those interested in having their own published materials available, as opposed to being supported by all participants interested in enhanced navigation and distribution options. OA seems to provide a great challenge in terms of supporting revisions and enhancements to the existing systems, as it separates and places in conflict the interests in (1) least expensive publication costs and (2) expensive enhanced discovery and delivery options.

Here are some other related questions: What are the incremental add-on costs for an online distribution system? What are the best ways to support the new infrastructures, metadata, and enhanced linking options? Is there benefit to having a variety of branded products, or would users be better served with a standard set of options under one basic interface style? How will OA models handle these beyond-publication infrastructure costs? In the differential pricing subscription model, there could be add-on options for those interested and willing to pay for premium services. If the questions are about how publishers can generate revenue and whether there are other, better pricing models, then perhaps the differential pricing models offer a more appropriate mechanism. These models utilize the new technologies to provide a variety of customized services that are now possible and desirable for different user populations.

Another issue needs to be considered: How does OA affect the remaining journal package plans? Will a move toward OA result in better long-term costs/benefits for users and/or authors/editors? If the goal is to provide the best cost for the dollar, OA seems to significantly raise the cost to a few and reduce the cost to many—but in the process, it endangers the entire system. Package plans spread the cost out across the researcher population, exactly the opposite of the OA model. As we have seen before, these two approaches are diametrically opposed, and this entirely new approach to allocating support revenue would create an impossible strain on the few. The intention of package plans (for many librarians) was to allow not just access to a broader range of materials, but also to generate use statistics that would eventually allow us to develop more accountable use-based pricing models. This would, over time, provide differential pricing based upon local populations—a more fair distribution of costs. OA would allow for gathering use statistics, but without the ability to identify and allocate the appropriate and reasonable shared infrastructure costs. While an OA model can offer use-based analysis, it provides fewer possibilities for influencing reductions in commercial distribution costs based upon the findings. The incentive to reduce costs for the few is far less powerful than the leverage created by reducing the costs to many.

There is one final question buried within the OA charging model: Is it acceptable to include "hidden altruistic" OA to the world automatically through higher journal costs—or is it better to offer this OA subsidy as an obvious choice to taxpayers? I suggest that the average person in a state with serious researcher populations would not be happy to discover that by accepting the proposed OA model, they are unknowingly continuing to accept inflated prices, they are absorbing much of the cost previously paid by commercial firms, and they are heavily subsidizing the world's existing commercial distribution method. If the idea of OA is treated as a real cost/benefit issue, it will be clear that the costs are higher to the few and the benefits would go to the publishers rather than to the community.

Of course, this ignores a rather significant issue, one that I will leave for another time: Library subscriptions are still supporting a commercially controlled and somewhat misdirected peer-review process in need of serious revision. Quality control and distribution can be handled in many less-expensive ways using the latest technologies, but this will require significant modifications of author, editor, and administrative behaviors.


Beyond the simple economic concerns, open access to academic information should exist on a platform and infrastructure providing more than just peer-reviewed literature. More comprehensive OA networks should develop, using redirected current commercial revenues, and these networks should utilize new technologies in order to create powerful integrated scholarly networks. These centers of excellence can be discipline-based scholarly platforms that handle all materials: peer-reviewed journals, teaching materials, curriculum support software, simulations, etc. Scholar networks are springing up all over the Internet with little library support or integration. This is where OA should be focused, on support for the distribution of noncommercial materials within these newly developed networks. We need to maintain a fair and reduced revenue base for publications, supplemented by redirected revenue for innovative scholarly networks. Don't be misdirected by talk of altruistic OA of journal literature. We should be collaborators and players, redirecting our efforts and resources into developing enhanced scholarly networks, rather than discussing simplistic and flawed economic models for maintaining the limited and expensive journal distribution network of today.

In summary, the open access publishing model is not necessary, not desired by the key players, and a real danger to the stability of the existing scholarly publication network. As long as the publication of journals remains commercially tied to the underlying peer-review infrastructure, there will need to be a large revenue stream, regardless of the actual distribution mechanism. This expensive quality control and distribution system will need to be supported by a large population, and OA works in the opposite direction by attempting to shift the support from the large group of readers to the few information producers. The OA process is not based upon a realistic business model. Those with a vested academic interest in producing and disseminating results do not have the capital to subsidize the entire process. Those with a need for the information for profit would be exactly the readers that would no longer support the infrastructure. An appeal to altruism will simply not work, and the removal of corporate and less research intensive support will be exactly the opposite of what is required to make this system an enhanced and more comprehensive scholarly network. Perhaps differential pricing models would provide a much better alternative for shared support of a revised scholarly distribution network, composed of open access for portions of the contents held within subject-based centers of excellence information platforms.

Open Access Readings

Budapest Open Access Initiative FAQ []

ARL Framing the Issue site []

The Open Archives Initiative (DLF, CNI, NSF)— Santa Fe Convention []

Landmarks in the evolution of open access []

Directory of Open Access Journals (1096 journals) []

ARC demonstration—harvesting search engine []

Public Library of Science (PLoS) []

arXiv (high energy physics with no peer review) []

DSpace MIT Libraries and Hewlett-Packard (HP) []

Charleston Advisor special issue []



[1] Ewing, John H. Point of View article "Open access to journals won't lower prices" Chronicle of Higher Education October 1, 2004 [].

David Stern [] is Director of Science Libraries and Information Services, Kline Science Library, Yale University.

Comments? E-mail letters to the editor to

       Back to top