Not everyone can afford open access monographs

There are many benefits to open research, and in recent years there has been increased funding made available for academics in receipt of grant funding to enable open access.

For example, the Wellcome Trust provides funding for both articles and monographs to be published, and UKRI has provided block grants since 2013 to assist research performing organisations in implementing open access changes and to cover costs for gold open access (the publisher’s pdf made freely available on publication).

Substantial progress has been made in the amount of scholarly research that is now freely available.

The role of the REF

The Research Excellence Framework used the 2014 open access policy to drive change – and while this was largely a success, there were some unintended consequences (not least the additional administrative burden). The difficulty now in REF further extending its policy, expanding from journal articles to include longform publications, is that there is no funding provided to support this outside of QR funding – and the “green” route is not readily available (with most publishers allowing one chapter from a monograph to be made available, not the whole monograph in accepted version format). Therefore research funded by larger bodies like UKRI or the Wellcome Trust will easily meet the proposed changes to policy, but unfunded research will struggle.

This is particularly challenging for universities that are not research intensive and not in receipt of the UKRI block grant, and for those universities that have a focus on the arts and humanities.

On 18 March REF published their proposed changes and consultation on the REF2029 open access (OA) policy, with current research outputs being required to meet the REF2021 OA policy. One of the areas that added the most administrative burden in REF2021 was the convoluted OA policy – which changed rules three times during the REF cycle.

In the early days we were told it could count as extra credit in the environment, and we were reassured at town hall meetings that no academic’s research would be forbidden from submission due to the open access policy. Then, with only two years to go before submission, it was announced that there would only be a five per cent tolerance for outputs that didn’t meet the OA requirements, and any additional outputs would be returned as unclassified.

This decision led to many universities not having their best research assessed. A five per cent non-compliance allowance may seem generous, but if you have a small unit of assessment, this only covers two (FTE up to 20) or even just one (FTE up to 11.5) non-compliant output. In our institute, this meant that we submitted outputs that were internally reviewed to be of 2-star quality rather than outputs that were internally reviewed as 3 or 4 star quality, hence REF2021 did not assess our most excellent research.

Compliance culture

REF2021 pushed forward with requirements for accepted manuscripts to be in institutional repositories within 3 months of acceptance, and were assessed as non-compliant if this criteria was not met, even if the journal published in had an embargo period of 24 months for the accepted manuscript. So an academic (sometimes on a 0.2 contract that included both teaching and research) took four months to put the accepted manuscript into our institutional repositories only to find their research output to be ineligible – even though it made no difference to access to the accepted version. Research intensive universities were quick to use part of their block grants from UKRI to increase staffing, improve workflows, and implement systems to ensure maximum compliance. Universities without UKRI block grants were left to muddle through alone, without the additional resources, and this often resulted in a higher number of “non-compliant” outputs.

REF2029 has now proposed that chapters and books must also be made available open access, and there will be no additional funding to cover these substantial increases in costs. UKRI’s own open access policy for longform publications came into force on 1 January 2024, with an investment of £3.5 billion towards covering costs (capped at £10,000 per monograph). It has been reported that even research intensive universities are estimating having to top this up by £60,000 to £80,000 due to the high costs of publishing monographs in this way.

There are some models and publishers that do allow for the published version to be made available at no extra cost, and a small number that allow the accepted manuscript to be made available after an embargo period of 24 months. But most only allow for a single chapter to be made available. To meet REF2029 proposed requirements this would either require academics to be extremely limited as to where they can publish, have a substantial budget for publication (in a time when every cost is under pressure) or only allow UKRI or other funded research to be submitted.

Counting costs

As a small institution, submitting approximately 160 staff to the REF, we submitted 21 books to REF2021. Ignoring any costs for chapters, and based on the minimum costing of £10,000 per monograph, this would cost our institute a minimum of £210,000! If we added our edited books (we submitted seven), that is an additional £70,000 – bringing the total extra cost to £280,000.

This is just for one small institute. In REF2021 overall the number of books submitted was 11,801 – multiply that by £10,000 and add the cost of 2,135 edited books and the costs are eye-watering. While REF2029 have proposed a similar route to that used with articles in allowing embargo periods unless publishers change their practices substantially the green route for monographs is not a viable consideration. And these costs do not include any staffing to check copyright, liaise with publishers, or process items within the repository.

The policy for journal articles, with a proposed change to articles being available after one month of publication, could lead to further administrative burden and a growing number of outputs that would potentially be non-compliant. Many publishers now publish immediately on acceptance, often without informing the author(s). Even if our researchers were to upload the accepted manuscript on the day of acceptance, this still requires repository workflows to be able to check and approve the records within a month. Journal articles are already given a higher priority for processing in institutional repositories, often at the expense of equally valuable research (for example non-text outputs) not being processed and made visible.

Should encouraging open research be a goal of REF2029? Absolutely. But this aim best fits in the research environment section and/or the outputs additional statement. There, institutes can talk about their approaches to open research, not just in terms of publications, but including other work like open educational resources, software, and datasets.

My plea to the REF is to not have an open access policy, and remove all the rules around outputs needing to meet set open access requirements to be eligible – is the REF not meant to assess our most excellent research? Research should not be excluded because of choice of publisher, an administrative error, lack or resources or – fundamentally – a lack of funding.