Supporting Open Access for 20 years: Five issues that have slowed the transition to full and immediate OA
The following article by Robert Kiley, Head of Strategy at cOAlition S, was originally published in the book titled “We so loved Open Access“, on the occasion of the 25th anniversary of SciELO.
For the last 20 years, Open Access (OA) has been at the centre of my professional life.
My first encounter with the concept of Open Access (OA) was during a Wellcome Library Advisory Committee meeting in 2002, where academic librarians expressed concerns about the escalating costs of journal subscriptions. This prompted discussions about Wellcome’s role in addressing the issue, particularly as their funded content was part of what libraries were paying to access.
Fast forward to 2023, and OA has become mainstream, global and, I believe, ultimately irreversible.
Current estimates suggest that more than 50% of the world’s research articles are published open access and that there are around 20,000 fully OA journals. Data also indicates that publishing OA is, on average, cheaper than publishing in subscription journals. For example, an analysis by Delta Think shows that around 45% of all scholarly articles were published as paid-for open access in 2021, but this accounted for just under 15% of the total journal publishing revenue.
However, after two decades of discussions, advocacy, policy development and strategy, can this level of OA be considered a success, particularly when half of all research articles published today is hidden behind a paywall? I think not.
In this blog post, I will reflect on the work I led – both in my role at Wellcome in supporting the implementation of its OA policy and more recently as Head of Strategy at cOAlition S – and highlight five issues that have hindered the transition to OA. On a more positive note, for each issue discussed, I will address how things are changing.
Issue 1: Reforming researcher assessment has been too slow
As early as 2005, when Wellcome became the first research funder with a mandatory OA policy, it was made clear that when funding decisions are taken what counts is the intrinsic merit of the work the researcher has undertaken and not proxies, like journal name and impact factors.
Despite this explicit statement, which was backed up by subsequent initiatives, such as Wellcome signing the San Francisco Declaration on Research Assessment (DORA) and implementing changes to the grant application form and review process, many researchers continued to believe that their future success depended upon having an article published in a high impact factor journal.
This fear not only discouraged some researchers from publishing in new, fully Open Access issues, as discussed in Issue 4 below, but it also resulted in some researchers not fully complying with OA policies. This non-compliance was in part due to high-impact factor journals being slow to adopt OA-compliant publishing policies.
What is changing to address this issue?
The recent development of initiatives such as the Coalition for Advancing Research Assessment (CoARA) and the Higher Education Leadership Initiative for Open Scholarship (HELIOS) – backed by a combination of research funders and higher education institutions – demonstrates a serious commitment to reform researcher assessment.
Issue 2: Business models based on APCs and subscriptions are highly inequitable.
The dominant business models that underpin scholarly publishing – namely subscriptions and Article Processing Charges (APCs) – are highly inequitable. It would be a significant policy failure if, in the transition to OA, we simply swapped one form of inequity – where subscription paywalls prevented a researcher from accessing research – with another, where APC barriers prohibited researchers from publishing their research.
Although in the early days of the OA APCs were seen as a credible strategy to deliver OA –the PLOS founders described APCs as “a natural alternative to the subscription model” – evidence now suggests that this model tends to exclude authors of particular career stages, genders, and institutions, in addition to also excluding those from certain regions in the world.
What is changing to address this issue?
Aware of the challenges discussed above, multiple stakeholders are currently exploring alternative models which prioritise global equity.
One approach that meets this condition is the Diamond publishing model, defined as one in which neither the reader nor the author pays, but where the costs of publishing are met by academic institutions and/or communities. Although Diamond journals have been in existence since the late 1980’s, they have gained momentum since the publication of the Action Plan for Diamond Open Access in 2022. At the time of writing, over 140 institutions and funders have publicly endorsed this plan, including agencies such as the French National Research Agency (ANR), the German Research Foundation (DFG) and the Swiss National Science Foundation (SNSF).
In addition, we have seen the development of the Subscribe to Open (S2O) model, described as a “pragmatic approach for converting subscription journals to open access—free and immediate online availability of research—without reliance on either article processing charges or altruism”. Currently, 168 titles from 20 publishers are published under this model.
Other publishers, such as PLOS, are experimenting with non-APC or volume-based models, as seen, for example, in their Global Equity Model. Furthermore, cOAlition S has commissioned a study to explore the feasibility of using the Purchasing Power Parity (PPP) as a potential model for devising an equitable pricing model for academic publishing.
Issue 3: Research funders and institutions worked independently, rather than collaboratively
Transforming a global, academic publishing system – which generates over $10bn in annual revenue and substantial profits for commercial publishers – was always going to be challenging. The task was made more difficult by the researcher assessment system (Issue 1) and the inherent conservatism of some researchers, who saw how the publishing system had worked for their senior colleagues and saw no reason not to emulate them and follow their publishing practices.
However, given that much of the funding to support academic publishing comes from academic institutions and funders, it is disappointing that these two stakeholders did not work more closely together to drive change.
Looking exclusively at funders, although there are a number of dedicated funder groups – such as the Global Research Council, the Heads of International Research Organisation, Science Europe, and the Open Research Funders Group – there was no attempt to align funder policies. Indeed, in 2005 when Wellcome developed its OA policy, it allowed an embargo of 6 months, believing this would align with the NIH policy. In the event, the NIH adopted a 12-month embargo period, which persists to this day.
What is changing to address this?
The launch of cOAlition S in 2018 was the first attempt to get research funders and performing organisations to align on a common OA policy. Over the past five years, membership has grown from 12 to 28, including funders in Australia, South Africa, Europe and the United States. In addition, even where funders have not joined cOAlition S – such as the federal agencies in the US – there has been significant progress towards broad policy alignment on key issues, such as zero embargoes and rights retention.
There is also greater coordination with institutions on OA issues, as evidenced by the close collaboration with members of the OA2020 initiative, the (time-limited) financial support for institutionally negotiated transformative arrangements and the development of institutional rights retention policies (IRRP), which mirror the rights retention strategy of cOAlition S.
By joining forces, funders and institutions can play a crucial role in supporting the transition to OA.
Issue 4: Unbalanced focus: commercial publishers receive too much attention over new publishing models
In the push to make research articles OA, much attention (and funding) was spent encouraging existing subscription, and in the main, commercially published journals, to develop OA publishing options and eventually to flip to OA. While this approach was reasonable, since many researchers were continuing to publish in subscription/mixed model journals, it likely slowed down the adoption of fully OA journals and emerging scholar-led models, such as pre-printing and publishing platforms that supported a post-publication peer review model.
Ultimately, the goals of commercial publishers, with a fiduciary duty to maximize profits for their shareholders, were invariably at odds with those of funders and institutions. This misalignment is manifestly evident in the response from publishers to cOAlition S’s Journal Comparison Service, where just 28 publishers have provided information on the prices they charge for their services, despite community calls for greater transparency.
What is changing to address this issue?
The COVID-19 pandemic demonstrated the value of publishing research findings ahead of peer review, with researchers – and not third-party suppliers, like publishers – deciding when and where to publish their research results.
Crucially, preprint servers like bioRxiv, and medRxiv, offer a much faster publication process, typically publishing in-scope submissions within 2-4 days, compared to the estimated 125 days in traditional publication venues. Although the immediate urgency of the pandemic may have subsided, researchers are continuing to make use of preprints, with bioRxiv reporting 3477 new articles published in March 2023.
At the same time, we are also witnessing the emergence of scholar-led, journal agnostic peer review models, such as Review Commons and Peer Community In (PCI), where preprints can be formally reviewed, without having to submit to a journal. As funders are increasingly recognising refereed preprints as equivalent to articles published in a peer reviewed journal, this should help to diminish the impact of journal names, and allow readers and those engaged in researcher assessment exercises to focus on the intrinsic merit of the published research.
Finally, the Council of Europe’s draft conclusions on High-quality, Transparent, Open, Trustworthy and Equitable Scholarly Publishing, highlight the importance of “not-for-profit, scholarly open access publishing models that do not charge fees to authors or readers and where authors can publish their work without funding/institutional eligibility criteria” and “stresses the importance of supporting the development of such models founded on led by public research organisations”.
Though it would be a brave commentator to suggest that commercially published journals will cease to exist in the near future, I think it is likely that their ability to continue to generate significant profits from the public purse will reduce. Moreover, I anticipate that they will focus their efforts on service provision – for example, research integrity services, copyediting and so forth – but will no longer have control over the content created by researchers.
Issue 5: Rights Retention should have been at the core of OA from day 1
Intrinsically linked to the matter discussed above is the issue of rights retention.
Rights retention ensures that researchers own and control the content that they created, based on the discoveries they made. Typically though, researchers hand over the copyright to their work to the publisher, and all subsequent use of that work (including the author being able to re-use their own content) is controlled by the publisher.
Rights retention was initially considered as a solution to encourage publishers to develop Open Access (OA) publishing options, with the idea that it would be used as a last resort when no other options were available to make research publicly available. In hindsight, as one of the architects of this approach, I believe that this was a mistake.
It gave too much power to the publishers, enabling them to set an APC at whatever price they deemed suitable. Consequently, this led to the emergence of APCs that exceed $11,500.
To be clear, I believe publishers should be fairly compensated for the services they provide, though this should not give them ownership (copyright) of the work, and the prices levied must be commensurate with the services provided. As most publishers have not shared their price and service data through the Journal Comparison Service, determining whether the price is fair and reasonable is impossible.
What is changing to address this issue?
Although rights retention is not a new concept – Harvard’s Faculty of Arts & Sciences voted unanimously to give Harvard a nonexclusive, irrevocable right to distribute their scholarly articles for any non-commercial purpose back in 2008 – in recent years it has become more mainstream. Specifically, several research funders include rights retention in their open access policies (and provide templated language to researchers to assert their rights). By way of example, those in receipt of a Wellcome award are required to “grant a CC-BY Public Copyright Licence to all future Author Accepted Manuscripts (AAMs)”, and that all submissions of original research to peer-reviewed journals must contain the statement: “for the purpose of open access, the author has applied a CC-BY public copyright licence to any author accepted manuscript version arising from this submission”.
Many universities are also implementing institutional rights retention policies in support of funder policies. In the UK alone, some 16 universities now have such policies, and more are expected to follow.
Several publishers are also supporting rights retention by explicitly stating in their publishing agreements that authors retain the right to share their author accepted manuscripts at the time of publication with a CC-BY license. The Royal Society and the Microbiology Society are two such examples. The scholarly publishing trade association (Open Access Scholarly Publishing Association – OASPA) has also signaled its support for rights retention, commenting that the proposal from the NIH to develop language to support authors in retaining their rights and bring clarity to the submission process for both authors and publishers “is welcomed”.
These combined approaches will help to ensure that, in the future, authors – not publishers – will retain control over how and when their research articles are openly shared.
Some 20 years after the publication of the Budapest, Bethesda, and Berlin declarations, OA continues to be a topic that generates a huge amount of interest, discussion and passion. And, though progress has not been as fast as I would have hoped, I am confident that it will not take another two decades before full and immediate OA is realised.