Editor’s Note: Today’s post is by Christos Petrou, founder and Chief Analyst at Scholarly Intelligence. Christos is a former analyst of the Web of Science Group at Clarivate and the Open Access portfolio at Springer Nature. A geneticist by training, he previously worked in agriculture and as a consultant for A.T. Kearney, and he holds an MBA from INSEAD.

As several Chefs at The Scholarly Kitchen have noted (Ask the Chefs Part 1 and Ask the Chefs Part 2), the recently announced US Office of Science and Technology (OSTP) policy (aka Nelson Memo) has the potential to drive a significant volume of US, federally-funded papers toward the gold Open Access (OA) path, whereby authors typically pay an article processing charge (APC). For example, Michael Clarke notes that ‘the implications of removing the post-publication embargo will likely be a shift not to Green OA, however, but to Gold’, and Alison Mudditt suggests that ‘there is a not insignificant risk of a default to gold’.

Here, I analyse data from the Web of Science (journal indexes of the Core Collection: SCIE, SSCI, AHCI, and ESCI) in order to (a) assess the current gold-OA landscape across regions and (b) quantify the potential impact of the policy.

The policy affects about 31% of US papers and 7% of papers published globally. Some of the papers in scope are already published in a gold-OA format (33%). Converting all underlying papers to a gold-OA format will be a significant contribution toward the global transition to OA (my working assumption for these estimates posits that ¾ of the papers that are not currently gold-OA, will turn gold-OA as a result of the policy). Yet the true impact of the policy may be greater than these numbers imply, given that several of the leading, mostly paywalled scholarly titles (Nature, Science,Cell, PNAS) get more than 40% of their papers from US, federally-funded research.

Europe leading the way to OA, while the US and China are dragging their feet

As shown on the chart below, the EU and the UK publish 50% of their papers in a gold-OA format. These papers are in journals that typically charge APCs. Europe’s performance places it well ahead of the USA and China, which publish 30% and 33% of their content, respectively, in a gold-OA format.

Europe’s OA transition has been accelerated by pro-OA policies such as Plan S, which was launched in 2018 and led to the proliferation of OA, transformative deals. Between 2017 and 2022, Europe’s gold-OA uptake increased by 23 percentage points, whereas in the absence of similar, centralized policies, OA uptake in the US and China increased by 8 percentage points in the same period.

While the US and China have been OA laggards, the chart shows that their attitudes toward OA may differ. US-based authors have the lowest fully-OA uptake across regions (22%), indicating a preference for established journal brands and a corresponding avoidance of fully-OA titles, while also illustrating that the US is not yet a prime market for publishers such as MDPI, Frontiers, and Hindawi.

China-based authors have a fully-OA uptake that is as high as that of Europe (30%), indicating that they may not have qualms about publishing in fully-OA journals. However, they have the lowest hybrid-OA uptake (3%) indicating that they will avoid paying for an APC when their journal of choice offers them that option.

Figure 1. Gold-OA uptake by region. Articles and Review Articles per the Web of Science (SCIE, SSCI, AHCI, and ESCI indexes)

Up to 132k papers to be unpaywalled

The US published 630k papers in 2021 according to the Web of Science, based on the affiliations of all authors on an article. The chart below shows how these papers can be narrowed down to those in scope of the OSTP policy (197k federally-funded papers), then those that are currently paywalled and will be affected by the policy (132k papers), and finally those that are likely to be published in a gold-OA format (99k papers based on a working assumption that ¾ of papers will be published gold-OA as a result of the policy).

Figure 2. US paper output split by funding and OA status. Articles and Review Articles in 2021 per the Web of Science (SCIE, SSCI, AHCI, and ESCI indexes); all federal agencies with >1,000 papers and some with fewer papers were identified via the ‘Funding Agencies’ filter of the Web of Science.

OSTP estimates that between 195k and 263k articles were federally funded in 2020. It arrives at these results by multiplying the US paper output of 2020 (626k per OECD/Scopus) by ‘the 42% of research that the United States government funds as a proportional basis’. It is not clear where the report found the ‘42%’ multiplier or whether it employs it correctly. In any case, given that databases such as the Web of Science do a good job in indexing funding acknowledgments, OSTP should have explored the numbers in more detail.

Another Plan S

It is unlikely that the policy will achieve 100% compliance, meaning that fewer than 21% of US papers will turn to an OA format in the first instance, and depending on the interpretation and implementation of the policy by the various agencies, even fewer papers will become gold-OA. Here, I assume that ¾ of the paywalled papers in scope will turn to a gold-OA format, which represents 16% of all US papers. This is similar to the effect of Plan S and the pro-OA policies in Europe, which got an additional 15% of papers in a gold-OA format from 2017 to 2022 in comparison to the US and China.

Adding 16% to the gold-OA uptake of the US, increases the global uptake by 3.2 percentage points to 40%. The direct contribution is not a milestone for the global OA transition, but it is another building block to its irreversible path.

bar chart showing percentages of OA uptake in USA and World overall
Figure 3. Gold-OA uptake globally and in the US before and after the implementation of the OSTP. Articles and Review Articles in 2022 (Jan-Aug) per the Web of Science (SCIE, SSCI, AHCI, and ESCI indexes).

Biological Sciences affected the most, while Engineering & Technology and HSS hold out

Some subject areas will be affected more than others from the OSTP policy. The chart below shows the three largest Web of Science categories for the research areas of Biomedicine, Biological Sciences, Engineering & Technology, Physical Sciences, Social Sciences, and Arts & Humanities. It assumes ¾ of the federally funded papers that are not currently gold-OA will become gold-OA thanks to the OSTP policy.

Biological Sciences are expected to be impacted the most from the policy. For example, the category of Cell Biology may increase its global gold-OA uptake by 7.1 percentage points, meaning that 61% of Cell Biology papers globally will be gold-OA.

Biomedicine and Physical Sciences come next in terms of impact. Half of their categories will increase their gold-OA uptake by more than 3.0 percentage points. Some categories such as Medicine General Internal appear to change little (+1.1 percentage points); that’s because the gold-OA uptake for federally funded papers in that category is already as high as 57%.

chart showing relative amounts of OA content in different fields and subfields
Figure 4. Paper contribution by region and OA uptake before and after the implementation of the OSTP for selected categories. Papers can be assigned to more than one region; assumes ¾ conversion to gold-OA for non-OA federally-funded papers as a result of the OSTP; Articles and Review Articles in 2021 per the Web of Science (SCIE, SSCI, AHCI, and ESCI indexes).

The areas that appear to be affected the least are Social Sciences and Arts & Humanities. None of their underlying categories are likely to grow their global gold-OA footprint by more than 2.0 percentage points. This reflects that few papers in these areas are federally funded. According to the Web of Science, about 100 papers were funded by the National Endowment for the Humanities, the National Endowment for the Arts, and the Institute for Museums and Library Services in 2021.

Another subscription holdout is Engineering & Technology. The gold-OA effect from the OSTP policy is limited, and the ensuing global gold-OA uptake is unlikely to exceed 33% in any of the three categories shown. The global leader in terms of paper output in Engineering & Technology is China, which contributes more papers than Europe and the US together in all of the categories shown above. Plan S and OSTP will not kick-start the OA transition in this space until China decides to jump on the OA bandwagon.

Quality over quantity

The main OA benefit of the OSTP policy may be less so about the volume of papers that are expected to turn OA and more so about the venues where these papers are published. Specifically, four of the most prestigious, mostly paywalled, scholarly journals (Nature, Science, Cell, PNAS) get more than 40% of their papers from federally funded research, with a minority of these papers currently published as hybrid-OA.

column chart showing funded paper amounts for Cell, PNAS, Nature, Science, The Lancet
Figure 5. Papers by funding status and OA status for leading scholarly titles. Articles and Review Articles in 2021 per the Web of Science (SCIE, SSCI, AHCI, and ESCI indexes).

It is difficult to imagine some of these journals maintaining their hybrid status if more than half of their papers are published as hybrid-OA (as many as 79% of papers in the case of Cell). Journals like Nature and Science include significant amounts of non-research paper content, and may retain appeal for subscribers despite shifts in the status of funded research papers. Should any of these journals flip to fully-OA (which is not straightforward – see Tim Vines’ comment about submission fees, and Angela Cochran’s comment about invited content), it will set an example for the scholarly industry to follow.

Catalyzing OA growth, up to a point

The transition to OA is an iterative process: journals are likelier to flip to fully-OA when there is a high OA uptake, and when journals flip, the OA uptake further increases. In that sense, the OSTP policy catalyzes the OA transition beyond the 3.2% gold-OA papers that it adds directly to the OA pool. Moreover, it further accelerates the transition by driving toward OA some of the most prestigious scholarly titles.

As with Plan S, the impact of the OSTP policy won’t be unlimited. As shown in the case of Engineering & Technology, an OA transition is likely to be very slow unless Chinese state funders decide to join their European and US counterparts in supporting OA. For the time being, China-based researchers will benefit from publishing for free in areas where they are a volume leader (e.g., Engineering & Technology) and reading for free in areas where they are a volume laggard (e.g., Biomedicine).

Christos Petrou

Founder and Chief Analyst at Scholarly Intelligence. Christos is a former analyst of the Web of Science Group at Clarivate Analytics and the Open Access portfolio at Springer Nature.


10 Thoughts on "Guest Post – Quantifying the Impact of the OSTP Policy"

(I just asked this on Twitter, too.)
Why use the WOS for humanities? I know humanities is relatively small $ and volume (but not in importance!) and yet in part because you use it here as a sharp contrast it is worth pointing out why and how these calculations don’t work.

The big question about the OSTP and any OA policy for humanists is what triggers the policy. Humanists don’t credit funders in the same way as scientists because funding is rarely singular even for a big project but rather individual authors get small grants (usually for time to write, but also for research time, sometimes to use particular research collections). So would any amount of funding from NEH trigger the policy? Or only a major NEH grant (for an individual those are usually a year of salary support– you can see the caps on that at neh.gov) –and even those are usually only a fraction of what’s required for research production.

The question of how well WoS identifies funding sources aside, how extensive is the AHCI as far as including HSS journals? Is the coverage too limited to get a sense of the journals in the various fields?

There has been a vivid conversation on Twitter (https://twitter.com/rschon/status/1569675006480732161) pointing out that the suggestion that about 100 papers were funded by NEH, NEA, and IMLS is likely an underestimate. According to Scott Weingart, it comes down to (a) the less comprehensive coverage of the Web of Science in A&H and (b) the practice of not acknowledging grant links in these areas.

Why are funding acknowledgments less comprehensive in A&H? Here is my best summary of the comments on the thread. It comes down to (a) the lag between the funding and the written output (A&H are slow), (b) the fragmentation of funding (multiple funds contributing to one piece – hard to decipher the contribution of each), and (c) the capturing of funding acknowledgments in a less standardised way that is harder to index. It was also pointed out that A&H folks do a better job at capturing such acknowledgments in monographs.

Finally, as Roger Schonfeld concludes, ‘the larger point holds: US humanities publishing is likely to be *far* less impacted by the policies that result from the new OSTP guidance as compared with STEM fields’.

(Thanks to Scott Weingart, Karin Wulf, Ted Underwood, and Roger Schonfeld for their input – might have missed someone)

Well this is a frustrating article.

“the US and China have been OA laggards”
This indicates a fundamental misunderstanding of the goals of publishers versus authors and libraries. Perhaps it’s not that the US or China aren’t “laggards”, but rather recognize that the current costs for the model business models perpetuated by the largest commercial publishers do not serve researchers or readers well.

“In any case, given that databases such as the Web of Science do a good job in indexing funding acknowledgments, OSTP should have explored the numbers in more detail.”
I would not say that WoS does a “good job” in indexing funding. It’s nearly impossible to do a “good” job of this, because authors are actually quite bad at properly crediting their funding correctly.

“Plan S and OSTP will not kick-start the OA transition in this space until China decides to jump on the OA bandwagon.”
Again, just because authors, libraries, and institutions haven’t “jumped on” the extremely expensive OA bandwagon that publishers prefer, indicates a problem with the publisher models.

I suspect it will depend on whether the relevant funding agencies require public access to the Version of Record of the article versus the Author’s Accepted Manuscript (which I believe is what AAAS allows to be deposited). Also, Science and Nature are outliers in that they have a broad subscription base for materials they publish other than research papers, such as their News reporting and their News and Views summaries of research. So I suspect they will be able continue with some success for the subscription model even if all of the research they publish is made freely available.

Thanks for sharing your thoughts on quantifying the impact of the OSTP policy, Christos. One aspect I’ve been wondering about for some time is: Will libraries truly be able to save up on funds with increasing OA costs (especially with Gold OA replacing subscription costs), or is it that increasing APC payments might balance out things eventually? If libraries are already saving up on funds (or are projected to save more funds in the Gold OA era), where would these funds be redirected to? I would appreciate any insights from the author, Chefs, and other readers of the Scholarly Kitchen in this regard.

Thank you for sharing your thoughts Cristos! I was wondering, what makes you think this policy will be retroactive, releasing prev papers to this mandate, rather than going forward?

I don’t think there’s anything in this post that suggests the policies will be retroactive. If they follow previous US funding agency policies (NIH 2008, Holdren Memo 2013), then they will include all grants issued after the effective date.

Leave a Comment