Springer Nature¡¯s chief publishing officer, Steven Inchcoombe, recently set out in these pages three things the research community was looking for from UK Research and Innovation¡¯s open access (OA) policy review.

Those were a clear prioritisation of the gold version of OA, a bespoke approach for OA books and a future-proofed funding model.

The new policy, , has much to commend. Placing transformative agreements at the heart of its policy demonstrates UKRI¡¯s confidence in these deals, which allow subscription expenditures to be repurposed to cover OA article processing charges. They are playing a key role in enabling the transition to full (gold) OA. A full 90 per cent of UK-authored articles to be published via this route by the end of next year, and our own experience bears this out. In each of the eight countries where we had a national transformative agreement live in 2019, between 70 and 90 per cent of research articles published with us were published full (gold) OA, against a global average of 30 per cent.

It is also to be applauded that funding is being allocated to support UK universities in reaching these agreements, including some ring-fenced specifically for monographs and books. Shifting from a system that saw publishers remunerated by journal subscriptions from libraries to one where publishing costs are met upfront via APCs requires this support. We hope other countries take note and adopt this enabling, progressive approach.

³ÉÈËVRÊÓƵ

Both these elements are important because of their role in delivering immediate access to the final published version of the research via full (gold) OA. Unfortunately, other elements of UKRI¡¯s policy appear to cut across this objective, with unintended consequences for researchers.

First, how does UKRI¡¯s decision to place subscription-tied (green) OA on a par with full (gold) OA deliver access to the version of the article that researchers want?????

³ÉÈËVRÊÓƵ

The final published version ¨C known as version of record, or VOR ¨C is not some artificial construct of publishers. We know from with 1,400 researchers, as well as an analysis of article usage, that it is overwhelmingly the VOR that researchers want to read and cite ¨C and it is also the VOR of their own research that, as authors, they want others to read and cite. They find the VOR easier to read, more reliable, and more authoritative and credible because of the reassurance provided by peer review and the stamp of credibility provided by proof of publication in a recognised journal.

Researchers also highlighted the value added to the VOR through the publication process, compared with earlier article versions (the submitted manuscript or the accepted manuscript), including copy-editing and typesetting. Critically, VORs include figures and links to relevant open data, open code and open protocols. This facilitates for the whole research system ¨C which is the main goal of making research articles OA in the first place.

Green OA typically revolves around posting the accepted manuscript, but the cost of creating these is, in essence, borne by library subscriptions given that they are created as part of the process of being published in paywalled journals. This is a problem in itself: OA should be about removing paywalls, not becoming dependent on them. Attempts to make accepted manuscripts more widely available do not reflect researchers¡¯ needs and could set back the transition to full (gold) OA and the realisation of the benefits of open science.

Second, as good as transformative agreements are, they have their limits. The industry-standard contract stipulates that a paper¡¯s eligibility for gold OA depends on whether the corresponding author¡¯s institution is part of the agreement. But the UKRI OA policy applies to all co-authors it funds in whole or in part. This is significant. We estimate that between 30 and 40 per cent of papers that have at least one UK author do not have a UK corresponding author and therefore wouldn¡¯t be covered by existing transformative agreements. Those co-authors risk of being left without a viable funded OA publishing route.

³ÉÈËVRÊÓƵ

We look to Jisc to utilise ¨C subscription or hybrid journals that have committed to transitioning to full OA ¨C or explore upgrading transformative agreements to cover these UKRI-funded authors, and ensure a full (gold) OA route to compliance.

Ensuring authors can publish in their journal of choice is something UKRI has said it wants to protect. But some publishers ¨C including some university presses and learned societies ¨C don¡¯t yet have the money or infrastructure to offer transformative agreements. Moreover, some UKRI-funded authors are not affiliated with institutions covered by such agreements. And some journals are so selective that many UK institutions don¡¯t publish enough articles in them to make transformative agreements practical.

We hope that Jisc ¨C which negotiates publishing deals on behalf of UK institutions ¨C will work with transformative journals over the next few years to enable a successful transition. We stand ready to work with it to achieve this and all the above.

We urge UKRI and Jisc not to get distracted by the false promise of subscription-tied (green) OA. Instead, they should see how these unintended consequences can be addressed. That way, we can all move forward together and deliver on the promise of an open future for all research.

³ÉÈËVRÊÓƵ

Carrie Webster is vice-president for open access at Springer Nature.