Transformative agreements (TAs) were designed to bring about a global transition from paywalled content to open access. Significant time and resource has been invested by university libraries in understanding their value. Here Paul Cavanagh, Senior Librarian Resource Acquisitions and Lucy Veasey, Senior Research Librarian, from University of Nottingham Libraries explain why it’s so difficult.

THE key long-term priorities for universities and research funders in relation to open access publishing are to achieve savings, improve compliance and increase publishing opportunities.

Paul Cavanagh and Lucy Veasey

Measuring whether these priorities are being met is difficult because the sector is still in the middle of changing from the traditional subscription (pay to read) model to an open access publishing model in which authors pay to publish.

If this supposedly transformative phase passed quickly the complexity and uncertainty around their pricing would be short lived. But they don’t appear to be going anywhere, and institutions have to use increasingly scarce resources to understand them.

The change from pay to read to pay to publish is primarily being achieved through Transformative agreements (TAs) or ‘read and publish’ agreements between universities and publishers. As a research-intensive global university, University of Nottingham has invested time, resources and funding in reviewing, implementing and managing Transformative Agreements (TAs). We currently have 33 live TAs, combining both subscription payments (the ‘read’ part of the agreement) and article processing charges (APCs, the ‘publish’ part).

The new element in these deals is paying to publish and this is where data and understanding of value is most scarce. Some work has already been done by universities across the sector in partnership with Jisc to understand Value For Money (VFM) on the publishing side. For libraries working with TAs, it’s essential to understand the value of both Read and Publish, and how each individual TA is structured for both elements, to maximise VFM for readers (students, academics and researchers) and authors.

Read

The read side of these deals is relatively transparent in comparison to the publish side, but it clearly shows that prices are unsustainable with significant annual increases above inflation of around three to six per cent.

The Read element has established indicators for costs and use (i.e. statistical usage reports from publishers or standardised information harvested via usage data platforms) so it can be quantified and compared against historical subscriptions, cost and usage data.

However, assessing the value is complicated by other benefits an agreement may offer like additional access to content for all users (i.e. Transnational education access for University of Nottingham China and Malaysia students, staff and researchers), offer value within the wider collection (e.g. supports a strategic research or teaching area), and comply with sectoral standards (i.e. Accessibility standards and features).

There may be opportunities to make small costs savings in “big deals” through bundling but Read & Publish deals should be flipping hybrid journals to fully open access.

Bundling makes it difficult to understand the costs over time, comparing the value of discounts and weighing them up against individual subscriptions costs and benefits of having a more flexible collection. If leaving a TA, you need to confirm any Post Cancellation Access (PCA) to subscriptions you have paid for as part of the agreement, and model additional costs (i.e. Inter-Library Loans) required to ensure continuity and access for readers. Further complications to staying on top of a TA’s VFM for both the read and publish is new iterations of deals, where titles are dropped in and out of agreements by publishers.

Some titles are dropped from agreements without converting to fully open access, and it’s not always necessarily clear why (i.e. has the title moved to a different publisher?). Libraries must ensure that they check the licence for each agreement thoroughly throughout its duration (i.e. each year of a multiyear deal) to ensure that academics and authors know which titles are eligible for publishing, and readers know what content is accessible.

Moving journals in and out of bundles can also disrupt authors who are planning their manuscript submission, where changes to eligible titles can result in authors finding that the costs associated with open access publishing are met one year but not in the following year.

Publish

Although complex, there are established data sources, methodologies and indicators that are used when analysing value for money in Read.

Attempts to replicate or provide an equivalent to these principles when reviewing VFM in relation to publishing is challenging and complex, not least because Publish data is not as consistent or established as Read data.

Methodology

Reviewing an agreement’s Publishing value begins by establishing which methodology was applied to its pricing. However, there is no consensus among publishers about which methodology to base publishing prices on. For most agreements, publishing prices of TAs have been based on two methodologies:

One is previous APC spend (total or average spend on individual articles over a given number of years based on the corresponding author’s affiliation – more on this below).

The other is publication rates, specifically the total or average number of articles published over a given number of years, based on the corresponding author’s affiliation.

The lack of agreed methodology and data standards for the Publish element means that any analysis requires significant investment in terms of staff time spent on data collation and confirmation. There is also still a risk that the analysis will be inconsistent with the publisher’s methodology.

For example, one publisher costed an agreement by using the date they approved the article for publication. The preferred date University of Nottingham Libraries use for our evaluation is actual publication date, identified via the Scopus database. This left seven articles within their methodology but outside of ours. Potentially small margins, but when agreements are costed per article with an APC value in the thousands of pounds, a few differentials can make a significant impact on cost.

Data

Methodological inconsistencies aside, we have found that data in Scopus has solved most of our problems in getting data. It has the granularity we need per article - published date, corresponding author affiliation and article funder. However, gaps emerge when a journal isn’t indexed in Scopus, which has relatively low coverage in the Arts and Humanities.

Publication forecast

To achieve our goal of understanding the VFM of the publishing element we are looking at historic rates of spend or publishing depending on the methodology, which whilst useful in establishing past patterns, is not an indicator of where our academics will publish in the future.

It is difficult to forecast Publish patterns accurately. On a fundamental level readers are predictable, they depend on a pool of existing literature while writers expand that pool. Also, researchers and authors are directly exposed to, and reacting to, the world outside the institution. And the relationship of authors to their institutions is not the same as readers, they can start and finish their employment throughout the year and often take or bring their research expertise with them.

A formal illustration of the difference between readers and authors is that, for authors, there is no equivalent of an advertised curriculum anchored with a student body. A curriculum brings relative commitment and assurance to make suitable review and purchasing decisions.

In practical terms, one approach to make purchasing decisions which support future publishing practices is to focus on the research priorities of the institution, alongside the external funding opportunities the university has bid for, ideally including both successful and unsuccessful bids. The risk is that for a large research-intensive university with broad research strategies such as the University of Nottingham, there are simply too many disciplines to support fully.

A reading list, once created, is a semi-fixed point for at least a full taught course cycle (i.e. a three-year degree). There may be new books or journal articles which are added to a reading list, but quality processes such as checking for new editions are undertaken as part of the review cycle. In contrast, research is developing and fluid, and dependent on different institutional factors such as specialisms, funding, and research centres.

Corresponding author

Another author-related VFM problem is already inherent in academia and ­academic publishing. In order to be eligible to publish under the terms of TA, the corresponding author must be affiliated to an institution that is part of the agreement. TAs also base their publishing costs on rates of corresponding authorship.

Authorship decisions are made during the research process based on discussions amongst individuals and are determined by and reliant upon the academic norms of the discipline(s) and local research culture at the time. In addition to creating an article pricing model determined by corresponding author (a responsibility previously unknown to authors), in practice predicting levels of corresponding authorship becomes unquantifiable for the university.

Whether or how this practice is skewing research more broadly is the bigger underlying question we have encountered when attempting to evaluate VFM for the Publish element.

Conclusions and opportunities

There is recognition across the sector that the TA model is not fulfilling its intended function, specifically as the mechanism that would bring about a global transition from paywalled content and collections to open access.

As noted in Jisc’s ‘A review of transitional agreements in the UK’ report, this needs to be balanced alongside the positive outcomes from TAs, such as extended access to subscriptions and OA publishing, increase in publishing and publishing venues for authors, and open access compliance with funder policies.

In the long term this may all mean stepping away from the article model entirely and Jisc, universities and publishers are having discussions about future equitable models. Whilst this is happening, we’re working out how to measure and achieve VFM within the current model. We would say that’s the crux of our work – how do we make the model work for readers and authors; and for us as librarians working with publishers, academics and funders?

Achieving transparency on the publishing side will only be the first step towards a solution when dealing with already unsustainable cost increases. Whilst TAs remain in place, we must keep exploring these questions to measure and understand their value.

We think establishing and implementing Publish data standards will improve the effectiveness and transparency of TAs; and allow institutions, funders, and publishers to be confident in the quality and reliability of the data used in decision making. We also see scope for institutions to integrate this data into their planning and decision making when focusing on the research priorities and external funding opportunities the university has bid for. This holistic approach will go some way to mitigate the risks and challenges we face in supporting our authors and promoting institutional research strategies, whilst continuing to work towards an open future for the sector.

This article first appeared in CILIP's free Annual Buyers' Guide.