Tuesday, March 28, 2017

EU report on Open Access

The EU's ambitious effort to provide immediate open access to scientific publications as the default by 2020 continues with the publication of Towards a competitive and sustainable open access publishing market in Europe, a report commissioned by the OpenAIRE 2020 project. It contains a lot of useful information and analysis, and concludes that:
Without intervention, immediate OA to just half of Europe's scientific publications will not be achieved until 2025 or later.
The report:
considers the economic factors contributing to the current state of the open access publishing market, and evaluates the potential for European policymakers to enhance market competition and sustainability in parallel to increasing access.
Below the fold, some quotes, comments, and an assessment.

The report acknowledges that this is a strange "market":
It has long been recognised that the scholarly publishing market has, in economic terms, an unusual relationship between sellers and buyers. Essentially a supply-driven market, scholarly publishing serves the needs of researchers engaged in:
  • Conducting research
  • Writing publications
  • Reviewing the quality of other research, and
  • Constituting the main readership for scholarly work.
However, while researchers are both producers and consumers of scholarly publications, their purchase is typically undertaken by academic libraries. Under the dominant subscription model (variously described as 'reader-pays' or 'toll-access'), this results in an 'intermediated market' which weakens the price sensitivity of consumers, be they authors or readers.
Their estimates of the current total market and the open access part are:
The market for scholarly journals alone is worth some $10 billion per year, with scientific, technical and medicine (STM) publications accounting for the vast majority of this figure.
...
The global open access market is approaching $500 million in size, but accounts for only 5% of the journals market. The proportion of immediate open access content is substantially higher, at almost 17% of global articles in 2014. The wide discrepancy between open access's share of revenues and articles reflects both the use of non-market based mechanisms to deliver open access content, and the lower cost of open access publication.
Between a quarter and a third of this, $2-3 billion, is siphoned off into the profits of the publishers which, as the report underlines, means that resistance to change is fierce.

The report identifies four types of open access:
  • Open access archiving ('Green OA') - the practice of archiving a version of an article for free public use in an institutional or subject repository.
  • Gold-Hybrid - peer-reviewed articles within a subscription-based journal are made immediately open access, typically on payment of an article publication charge (or APC) to the publisher or through an offsetting agreement.
  • Gold-APC - publication in journals that make all of their content OA via payment of an APC, and do not rely on subscriptions.
  • Gold no-APC - publication in fully open-access journals which do not charge an APC.
Any large-scale move to these models would reduce the $2-3 billion significantly:
The commercial incentives for subscription publishers to move to APC-based open access remain weak. Mean APCs are approximately €1,500 (Gold-APC) and €2,500 (Gold-Hybrid), while average subscription revenues are double this, at €4- 5,000 per article. Open access would also jeopardise licensing revenues and corporate subscriptions, estimated at some 20% of STM publishers' current income. Recent initiatives aimed at repurposing existing library subscription budgets for open access, such as the Open Access 2020 movement, assert that there is sufficient money in the system to make the transition. Publishers have also recognised the opportunity for OA to generate additional revenue streams. However, unless the gap between per article revenues under the OA and subscription models narrows significantly, or threats to the sustainability of the subscription model increase, progress towards a large-scale transition is likely to remain slow.
Thus, absent some major intervention, it won't happen. What does the world get in return for the $10 billion per year?
In most cases, researcher-authors freely transfer copyright in their work to publishers, or grant them an exclusive right to publish the final version of their manuscript, also known as the 'version of record'. In return, journals perform four auxiliary but essential functions of scientific communication:
  • Registration: establishing the author's precedence and ownership of an idea
  • Dissemination: communicating the findings to its intended audience
  • Certification: ensuring quality control by managing the peer review process
  • Archival record: preserving a fixed version of the paper for future reference and citation.
The report recommends six main areas of intervention::
Our work has identified six main roadblocks to open access that should be addressed through appropriate policies and measures. These are:
  1. Weak author incentives for open access: The single greatest barrier to wider uptake of open access is cultural resistance within the academic community. Until there are sufficient incentives for researchers to actively choose open access publication and archiving, demand will remain muted - and publisher support for open access will mirror this. If Europe is serious about increasing access, then its mechanisms for research assessment, grant funding, academic promotion, and institutional funding need to reflect this.
  2. Unclear route to transition for subscription publishers: The gap between per article revenues under a subscription model and those available under an APC or Gold no-APC model still appears unbridgeable for many commercial and society publishers. More must be done to show that the transition can be made without irreparable damage to publishers' business models - whether through offsetting mechanisms, acceptance of higher APCs, or increased adoption of Gold no-APC models like FAIR OA. However, this must be accompanied by increased expectations of the service provided by publishers, including licensing and machine-readability.
  3. Lack of transparency in the market: The lack of transparency in the subscription market compounds the problem of journal non-substitutability, and results in a dysfunctional market which serves neither researchers, institutions nor the public interest effectively. Piecemeal attempts to improve transparency through Freedom of Information requests, often undertaken only by students and grassroots activists, must give way to a concerted policy-led effort to deliver transparency and improve competition in both the subscription and pure open access markets.
  4. Disparate national and disciplinary contexts: There is no single pathway to open access that finds support from a clear majority of stakeholders. Therefore the adoption of a pathway should not preclude also adopting other OA strategies and pathways. Different approaches are needed depending on the national and disciplinary context, and policy interventions must therefore promote and enable flexibility. Funders and institutions need to proactively support not only the entry of new players into the market, but also their development at scale.
  5. Suboptimal infrastructure: The administrative burden associated with open access models remains too high for all stakeholders - whether authors, institutions, publishers or funders. Improved processing, payment, deposit and reporting mechanisms, built on common standards and infrastructure, are needed to allow open access to scale rapidly and efficiently.
  6. Inadequate monitoring and reporting: Europe's ability to track progress and assess the effect of interventions in the market is greatly inhibited by fragmented and underdeveloped monitoring and reporting mechanisms. Investment is needed in standards and tools to track compliance with open access policies, reliably determine the aggregate proportion of the scientific literature which is available in open access form, and monitor sustainability on both the demand and supply-side.
Despite all this valuable content, in my view the report suffers from three significant flaws, one fairly minor and two really major.

The minor one is the issues of what "open access" actually means and whether Gold OA delivers it. The report writes:
The European Commission has defined open access (OA) as 'the practice of providing on-line access to scientific information that is free of charge to the user and that is re-usable'.

Legally binding definitions of 'open access' and 'access' in this context do not exist, but authoritative definitions of open access can be found in the Budapest Declaration and the Berlin Declaration. These definitions define 'open access' as including not only basic elements such as the right to read, download and print, but also the right to copy, distribute, search, link, crawl, and mine. This is commonly achieved through application of a Creative Commons Attribution License (CC-BY).
The report fails to observe that, while these may be the definitions of OA for the EC and OA advocates, these are not the definitions of OA in use by the oligopolistic publishers. As Poynder has observed some are using their own bogus "OA" licenses which are far more restrictive than CC-BY and specifically exclude most of the rights of the Budapest and Berlin Declarations, those necessary for the EC's goal that information be "re-usable".

Even when they do label content CC-BY, they can place technological barriers in the way of attempts to "copy, distribute, search, link, crawl, and mine". Publishers' ability to place such barriers is about to be greatly increased by the advent of Web DRM.

In addition, the Wellcome Trust and others have identified many cases in which an APC was charged but the article remains behind a paywall, illustrating the problem that neither the author nor the publisher have any interest in policing compliance. The report's economic analysis thus skewed, because the APCs paid to publishers using poor compliance, bogus "OA" licenses or technological protection measures do not obtain the benefits they assume. Not to mention the many tricky legal issues with the use of CC-BY that Poynder identifies.

The first major flaw is that the report not merely fails to assess the value obtained in return for APCs, but more importantly fails to assess the value obtained in return for the much larger subscription payments. In each of the four "auxiliary but essential functions" the value obtained has been rapidly decreasing:
  • Registration: the long, variable and unpredictable delays in the publication pipeline mean that submission first is no guarantee of publication first. For a clear example, see DNA as a storage medium. A friend whose paper in Nature took 22 months from submission to publication wondered "whether one of the referees had been trying to delay it so they could get something out"; gaming the system in these and similar ways is endemic.
  • Dissemination: the ineffectiveness of the current system at "communicating the findings to its intended audience" is aptly illustrated by the large proportion of "convenience" accesses via Sci-Hub documented by Kramer:
    Overall, 75% of Utrecht Sci-Hub downloads would have been available either through our library subscriptions (60%) or as Gold Open Access/free from publisher (15%). In so far as these downloads were requested by academic users (i.e. affiliated with Utrecht University), use of Sci-Hub for these materials could be seen as ‘convenience’. For the remaining 25%, Sci-Hub use could be seen as a way of getting access to articles that are not freely available through publishers.
    As I wrote in Elsevier and the Streisand Effect:
    What the readers want is a single portal, preferably Google, that provides free, instant access to the entire corpus of published research. ... Sci-Hub is as close as anyone has come to providing what the readers want. None of the big publishers can provide it, not merely because doing so would destroy their business model, but also because none of them individually control enough of the content. And the publishers' customers don't want them to provide it, because doing so would reduce even further the libraries' role in their institutions. No-one would need "personal shoppers who moonlight as border guards".
    These are all problems even if you do not (as the EU does) believe that the public, which pays for the research, has no right to access the results.
  • Certification: It has been clear for many years that the publishers aren't "ensuring quality control by managing the peer review process". They have abdicated their role as gatekeepers by proliferating journals, so that the question is no longer whether a paper is published, only where it is published. The peer-reviewed research on peer-review shows that the journal in which it was published provides no information as to the quality of the paper. But the higher the bogus "impact factor" of the journal, the more likely the paper is to be retracted.
  • Archival record: As I've recently written, the idea that the owner of some content is best suited to preserve it is a counsel of despair. Not merely do paywalls make it impossible for archives to preserve journals without permission, the DMCA makes it impossible for them to preserve non-CC open access content, or content carrying CC-licenses protected by even ineffective DRM. The publishers will not allow archiving by entities over which they lack adequate control, because that risks their $2-3 billion per year profits. In addition, publishers currently personalize, watermark and re-format content that they do deliver, making it extremely difficult to establish that the content eventually received by readers is the same as that approved at the end of the review and editing process.
These deficiencies add another major economic argument in favor of OA, in that the vast sums being spent on non-OA content are not providing the benefits the recipients claim.

The other major flaw is that the report completely fails to address the fundamental role that copyright ownership and transfer plays in the multifarious problems of academic publishing. The sole mention of anything related that I could find was in Figure 8, where "Enable retention of copyright by author" was one of a laundry list of "Recommendations on promoting the transition to open access (sourced from 20 published studies)". The report fails to understand that it is precisely copyright ownership by authors that is at the root of the problems they describe, and that the problems will not be solved while authors retain copyright.

The tradition and practice, of somewhat dubious legality, that copyright in works by faculty and staff belongs to the authors rather than to the institution or the funder of the research ensures that the interests of the institution, the funders, and the public are never represented in the fundamental transaction of academic publishing, which is between one or more individuals and a corporation (the publisher):
  • The individual authors bear none of the economic or societal costs of their decisions. Even if they could not divert project funds to pay APCs, the career benefits to be obtained by publication would persuade them to pay out of their own pocket. They gain no benefit from negotiating with the publisher to refuse copyright transfer or exclusivity in licensing.
  • The publishers have massive incentives to, and vast resources to devote to, ensuring that they do have copyright ownership or license exclusivity. Even when they have been paid an APC, they can simply ignore the OA requirement, redefine it to mean exclusivity, or use technological means to implement exclusivity. Expecting authors, even if they had motivation to do so, to emerge victorious from such an unequal negotiation is futile.
There are only two effective means of ensuring open access to future works by publicly- or charitably-funded research. Either:
  • The author's institution must change the terms of employment for faculty and staff to make it clear that either their writings are works-for-hire, and thus the copyright belongs to the institution, or that these writings not copyrightable, as is the case for works by employees of the United States. Or:
  • Funding institutions must change the terms and conditions of their grants to make it clear that copyright in works derived from the grant belongs to the funder.
In neither case would authors' signatures on copyright transfer or license forms be valid. In both cases the institutions that bear the economic and societal costs would be across the table from the publishers, and would be both motivated to, and able to, prevent the publisher acquiring exclusivity. Without exclusivity, the publishers' extortionate margins could not be sustained.

7 comments:

David. said...

The results of the deal that the Dutch obtained after walking away from the table are revealed in How Elsevier plans to sabotage Open Access by Sicco de Knecht. They are instructive in this context:

"As with fossil fuels the parties controlling the resources will do everything in their power to draw every last drop of revenue from the old system before making the shift to a new system. This is essentially what Elsevier is doing with this so called ‘pilot’ for open access: giving in just enough to prevent any major alteration to the status quo.

Right now the same uphill battle for open access is fought in many other countries such as Finland, Germany and in the European Union in general. The recent publication of the Dutch contracts might change the conversation there, making parties reconsider whether the way forward for open access is with Elsevier and its competitors at all. At least we now have a glimpse of what a major publisher like Elsevier is willing to concede — not much."

Thomas Munro said...

Minor typo - the link given leads to a different report. The OA report is available at: https://zenodo.org/record/401029/files/OA_market_report_Final_13_March_2017_.pdf

Galen Charlton said...

To ask a naive question regarding the notion of academic employers unilaterally requiring work-for-hire contracts... what would be the incentive for faculty to acquiesce to what would surely be perceived by some or many as an administrative appropriation of historical privileges? Furthermore, doing so in the name of open access seems to me like it would likely convert many faculty who are indifferent to the notion of open access into outright opponents.

But, for the sake of argument, suppose a university could sell it to their faculty — or at least impose it without too much fuss. While it makes sense that holding the copyright on all future article output would strengthen the negotiating position of a university, what assurance would the institution's constituents have that the additional negotiating power would be used to further more open dissemination of research — as opposed to simply bargaining to keep a lid on pricing?

David. said...

Thanks, Thomas! My bad. I need to blog about that report too.

David. said...

Galen, that's one reason why I believe that funders retaining copyright is a more practical approach.

David. said...

The Economist last week had an editorial and an article arguing for immediate open access via preprint repositories.

David. said...

I'm late to linking to Making A Few Elsevier Predictions by Lisa Janicke Hinchliffe, which has more insightful thoughts on Elsevier's plan to capture the entire research workflow. In particular:

"The pivot to metrics and analytics underscores that Elsevier is on a trajectory to convert its journal portfolio to being open and no longer behind a paywall. Elsevier is already a leading open access publisher and initiatives to deliver author manuscripts to institutional repositories, among other projects, indicate a shift to making publications more and more discoverable and accessible. Discovery and access generate data that are critical for developing useful metrics and analytics."

This thought is reinforced by Elsevier's recent announcement that Cell Press Helps Authors Share Papers Under Consideration:

"Recognizing that science benefits from communicating reliable, vetted findings as quickly as possible, Cell Press now offers authors the opportunity to surface their submitted manuscript as soon as it has been evaluated by an editor and sent out for peer review. The service is called "Cell Press Sneak Peek." For authors who elect to participate, manuscripts under consideration by peer reviewers will be uploaded on the authors' behalf to the Sneak Peak group on Mendeley, an academic social network, where it will be freely accessible to all interested readers."

Yet another effort to prevent copies of papers ending up in institutional repositories that Elsevier doesn't control.