Who wins after a divorce?… or how to interpret the DEAL-Elsevier new agreement

Imagine that you are a young researcher in Germany, having started your thesis in September 2018. For the last 5 years, you have had no legal access to articles published by the world’s largest publisher, Elsevier. Your institution has saved hundreds of thousands or even millions of euros, but you don’t really know where that money has gone. By contrast, on a day-to-day basis, then as a PhD student, now as a post-doc, you tinker with your access by writing to authors, asking your colleagues abroad if they can send you this article, requesting your library to buy that crucial paper, scanning preprints, using the unpaywall button or, late at night from home, typing the full combination of letters and signs to reach the platform whose name you must never utter or write.

To my knowledge, this divorce between a major publisher and a national consortium, DEAL, folowed by a reconciliation, has been the longest for a very rich country,. This post analyses how the separation happened, what is known of a long period of divorce in which no German institution had a subscription to ScienceDirect, and finally moving on to the reconciliation agreement published on September 6th, 2023 and validated in January 2024.

From harsh talks to full divorce (2016-2018)

Indeed, it was not for the lack of money that DEAL did not sign with Elsevier, but because the conditions of a signing were not met. By contrast, reading DEAL’s agreement with Springer-Nature, analysed at length 3 years ago, shows what was expected: an agreement including subscription and open access publication, all at a cost deemed reasonable by the German consortium. So how did they get to a “no deal”? As often when trying to rest on past information with institutional sites and changing policies, I shall say that most documents cited below have disappeared from the DEAL website and, therefore are captures made by the Internet archive.

“At Elsevier, serving research is our paramount goal. We have therefore chosen to continue providing access to Elsevier journals for dozens of German institutions that cancelled their individual subscriptions at the end of 2016. They did so anticipating that a new Germany-wide license agreement would be in place by January this year, which we regret so far has not been achievable. We strongly believe that access to high-quality research is important for German science. The continuing access for the affected institutions will be in place while good-faith discussions about a nationwide contract carry on. This reflects our support for German research and our expectation that an agreement can be reached.”¹

I hope one day some colleagues will systematically study the rhetoric of big publishers PR. Anyway, the one above is typical of a service industry which makes believe its aims are totally aligned with the ones of its clients. Imagine the reverse situation, where DEAL would state : “at DEAL, assuring service providers profit is our paramount goal…”. Back to our main topic: the unconditionnal reconnection decided by Elsevier is not something unusual: at the same time, it happened for example for Taiwanese institutions in a similar situation². But Elsevier hopes for a soon-to-be new German agreement would not be fulfilled. Indeed, after these back and forths, the negociations stalled, leading to a full divorce by mid-2018, as stated by the German Rectors Conference, which had “no choice”:

“The excessive demands put forward by Elsevier have left us with no choice but to suspend negotiations between the publisher and the DEAL project set up by the Alliance of Science Organisations in Germany.” That was the verdict of the lead negotiator and spokesperson for the DEAL Project Steering Committee, Prof Dr Horst Hippler, the President of the German Rectors’ Conference, speaking in Bonn, where the last discussion took place this week.”³

At this point, we shall note that all cited documents are written in English, while negociations surely happened in German. DEAL had the clear intention of making its moves very public and widelly known beyond the Federal German space and Mitteleuropa.

Learning to work without simple legal access (2019-2022)

So Elsevier pulled the plug in July 2018 and everything went quiet after almost two years of turmoil. That was not a given: you could think that protest letters, petitions or lobbying from unsatisfied lay researchers would multiply as a whole nation of scientists were cut from at least a fifth of the published literature. To lift the veil on the actual frustrations and losses resulting from the switch-off, it was… Elsevier, which commissioned a survey in the summer of 2019, the summary results of which can still be seen on the pages of one news agency.

“Most German researchers agree that losing access to ScienceDirect made their research activities less efficient (61%) and delayed the production of the research output (54%). High-quality research further required access to current, international research results. However, the survey shows that 49% of the scientists surveyed believed that the lack of access to new research findings leads researchers to miss current developments or to become aware of them only with a delay. 44% of respondents fear that this will have a negative impact on the quality of their research. All in all, 84% of researchers surveyed think ScienceDirect was important or somewhat important while 76% supported or strongly supported the restoration of full access to ScienceDirect in Germany.“

Of course, no raw data has been published and the study itself has not been shared beyond this PR. Nevertheless, in the body of the text, Elsevier mentions another ‘independent’ study carried out by the University of Münster. Like the previous one, this is not an actual academic study, but a library survey, published only on their blog, in German. Despite its limitations (size, a single institution), it presents some interesting, and most probably unique, results on the representations of German researchers one year after the cut. In particular, the following graph should be highlighted:

The orange answers indicate respondents’ agreement, and the statements have been ranked in descending order of positive responses. They show a mixed picture in terms of opinions, both across the population as a whole and for many respondents themselves. from one question to the other. Though the vast majority, namely two-thirds (66%), agreed with the statement “I need more time to get the literature” and 58% thought that the right thing to do was to put pressure on Elsevier to give in, also the option with the fewest disagreeing votes (5%). That does not imply support for the shutoff: in fact, 55% agreed that “No deal is no option – negotiations should be resumed as soon as possible”, and 46% that the lack of access was “a serious competitive disadvantage”.

While 43% agreed that “Elsevier as a profit-orientated company would only harm science”, and only 11% disagreed, only 29% would “refrain from writing or review articles for Elsevier journals” against 40% who would still perform it. After some questions on the importance of Elsevier journals and the use of spared funds, the last question shows another divisive view on the resuming of negotiations, with only 16% in favour of it – which of course was not addressed in the Elsevier PR mentioned.

These two surveys are the only public manifestations of a debate in Germany during this period. If opinions remain relatively unpublic, what about practices? Does the impossibility of immediate legal reading actually have an impact on the way German academics write, their choice to publish in Elsevier journals or their productivity? To my knowledge and through the extensive use of Matilda, only two academic articles have addressed these issues The first is counterfactual, in that it looks at the behaviour of affiliated authors in Germany in chemistry for Springer and Wiley with which DEAL has signed an agreement. Published in 2021 in economics, it only considers the first year of the agreement (2020), in comparison with the previous period and with a control group with no agreement of this type. Nevertheless, the authors are already measuring some effect :

“researchers’ submission behavior in the field of chemistry has changed to some degree, as eligible researchers have increased their publications in Wiley and Springer Nature journals at the cost of other journals. While the effect is not overly large yet, it is statistically significant, and it may increase over time, as the agreements become even more well-known among scientists. Hence, journals covered by the DEAL agreements appear to have a competitive advantage in attracting authors”.⁴

If agreements signed raise attractivity, then unsigned ones shoud diminish it. The second one deals with the latter by considering the evolution of publication and referencing activities of the whole population of German authors in Elsevier journals, with no control group.  Published in 2023 in scientometrics, it is based on more than 400,000 articles and more than 33M references:

“We also observe year-on-year decreases in the proportion of citations, although the decrease is smaller. We conclude that negotiations with Elsevier and access restrictions have led to some reduced willingness to publish in Elsevier journals, but that researchers are not strongly affected in their ability to cite Elsevier articles, implying that researchers use other methods to access scientific literature.”⁵

The two studies therefore show that the structure of publications is affected by the agreements; whether signed or not, but only marginally, at least over a short period. Furthermore, reading seems to be remarkably unaffected by the lack of legal and rapid access to the literature. To enable simple and legal reading, It is likely that other internal work has been produced by the consortium or that self-support systems have been put in place, similar to what the Swedish libraries deployed during their own breakup with Elsevier⁶. Beyond this study, there is anecdotal evidence, given by colleagues, but also an interview of a member of the negociation team, Dr. Bernhard Mittermaier, head of Forschungszentrum Jülich’s Central Library, which tends to show that they were following the rate of publications:

“The option to publish with Elsevier was not affected. Some scientists, however, asked me whether a publishing boycott would make sense in view of the fact that many editors from Germany – including Prof. Wolfgang Marquardt – had discontinued their work for the publisher with reference to the stalled DEAL negotiations. In fact, Elsevier’s share of all Jülich publications decreased from 26 % in 2018 to 18 % in 2022. Across Germany, there was a decline from 19 to 15 %. This may also be a reason why Elsevier returned to the negotiating table.”

In the end, it is reasonable to consider that German researchers have adapted to a life without ScienceDirect over the long term, still reading articles published by Elsevier, but publishing less in journals disseminated by it . What the French and British did not dare to attempt after lengthy negotiations, the Germans did, with very substantial savings and a diminished dependance to the biggest commercial publisher. But what happens afterwards, when the time comes for one or other of them to consider recontracting?

Dealing again… on different terms (2023-2024)

2023 began, as in previous years, without ScienceDirect for German researchers. Im Westen nichts Neues, as a fellow economist lamented :

 

In fact, Elsevier had returned to the negotiating table in autumn 2022 and, after a four-year drought, seemed ready to make concessions that would have been unthinkable four years earlier. The negotiations took place behind closed doors, until the sudden announcement of their success at the beginning of September 2023, followed by the publication of the contract itself. Let’s dive into it, as DEAL has always been transparent on their agreements (nice PDF, full text and monetary information,…), published under a CC-BY-ND license⁷.

We will not delve into the details of the usual characteristics of this type of agreement (definition of the parties, services expected, users authorised to read, corresponding author limitations, etc.), but will instead focus on the most central elements and on some unique features compared to the bodies of agreements analysed elsewhere.⁸. This agreement is a “classic” Read & Publish, which includes in its core payment articles published in hybrid journals, but not articles in full open access journals, for which the fee is simply reduced by 15% or 20%. It also includes a back catalogue upgrade for all institutions, at a total cost of €10m. It is a “pay as you publish” agreement, with a PAR fee for each article, depending whether they are in a “regular journal” (2,500 €) or a Cell Press/The Lancet journal (6,450€), with an inflation rate of 3% and 4% respectively⁹.

This payment model has two consequences that are quite specific to this agreement. Firstly, with the exception of the back catalog, institutions have no front money to commit. Whereas in the past some agreements offered “tokens” or “waivers” for publication, the opposite is now true: you only start to pay after publication. Secondly, this provision would encourage free riding: as withalmost all agreements of this type, the corresponding author is offered, as a priority, to publish in open access under the CC-BY licence, but he or she can refuse. There is also a provision in the contract that prevents this refusal to publish in open access from being organised by counting all the publications:

“For the avoidance of doubt, the applicable PAR fee for Core Hybrid journals for the year of the acceptance date will be applied to both open access and subscription articles in these journals and to subscription articles published in Cell Press and The Lancet journals.”

So, despite the diminishing share of articles observed during the absence of agreement and the lack of front money, Elsevier has a certain guarantee of revenue as 18, 19% or 20% of the German research production will end in one of its disseminated journals. In exchange, the company had to accept very harsh conditions on the data generated by German users. A full page (section 7.6) is dedicated to Data Privacy in the agreement, with reminders of legal provisions derived from the GDOR European regulation. DEAL and Elsevier will co-supervise the whole data processing, the latter refraining from using any personal data without the consent of users. On this point, a loophole was anticipated by forbidding any general opt-in device: German colleagues will be able to fully use ScienceDirect without signing any consent. Of course, all data will be stored in one of the Member States of the European Union. The matter is so sensitive that a future workshop is planned during the first year of the contract, where part of the IP addresses would be automatically erased when IPs are not located in professional settings.

Without doubt, Elsevier’s transformation into a data company and the growing controversy surrounding its new business models on reselling user data¹⁰ has been closely observed in a country so keen on privacy. Still, despite these worries, DEAL signed the deal and did not include any fines in case these limits would be trespassed¹¹. But what about the signing of German iHER nstitutions?

Conclusion : which savings, for which uses?

In fact, there was still a little uncertainty when the agreement was unveiled, as a four-month period was about to begin during which the institutions would each have to indicate whether they would sign the agreement. It could only be ratified if at least 70% of the institutions approved it, and fees were lower if 90% did. On 15 January 2024, DEAL announced that this second threshold had been exceeded as “nearly all of Germany’s major universities and research institutions are now participating“. Elsevier has now joined Wiley & Springer in the DEAL family, with very similar agreements focused on hybrid open access. But what does it mean from the point of view of German HER institutions? Let’s go back to Dr. Bernhard Mittermaier’s interview, who talks about his own instiution costs and the global German ones:

“Taken together, Jülich institutes will now save around € 100,000 per year on fees for hybrid open access that were previously paid to Elsevier. For Forschungszentrum Jülich as a company, the costs for Elsevier will even decrease by about 40 % than was the case under the former agreement, assuming publication figures remain the same. This corresponds to about € 300,000 per year that can be saved compared to 2018, the last year of our previous agreement with Elsevier. Elsevier’s fees per article are now much lower than they were in 2018 and similar to those charged by Wiley and Springer Nature.”Compared to 2023, however, when hybrid open access, document delivery, and pay-per-view each cost around € 100,000, additional expenditure of € 200,000 will now be incurred.

Let’s try to do the math (which does not add up), based on that paragraph in the following table, with three references, the last year of the former (local) agreement, the shut-off period and the first year of the new agreement.

Expenditures/Year	2018	2021	2024
Total	600,000€ 500,000€	100,000€	300,000€

The previous total cost is 500K if you follow the 40% reduction and 600K€ if you add the total savings mentioned. Whatever the case, the new deal is far below the older ones, in which German institutions were known for paying “much more” than similar institutions in Netherlands or France. Let’s now project the costs nationally:

Year	pre-2018	2021	2024
Expenitures	70M-100M€ in mostly reading agreements	5-10 M€ max in Hybrid OA publishing?	30-40M€ in P&R agreement

The first figure was never made public, but I have heard estimiations in between these two markings, The second one is very maximalistic as OpenAPC counts between, 1M€ and 1,3M€ for Elsevier in Germany for the years 2020 to 2022. The thrid one is based on the number of expected publications and the different fees defined in the agreement. So the savings have been huge during the shutdown and Elsevier lost probably at least 300M€ before resuming negiotiations. And despite losing probably around 50% of its 2018 revenue, the company prefered to sign rather than leaving almost all the money on the table.

While, for example, French institutions have made a major commitment to using some of the resources saved for OA initiatives and by replenishing the National Open Science Fund, this does not seem to be the case in Germany. The national research funder DFG has recently announced the launch of a Diamond OA publishing platform… with a maximum budget of 1.5M€ per year. I let you figure out what it would have been with just 30% of the money spared. So the German HER institutionswon won a lot, Elsevier stalled, but the dependence from big commercial publishers has not been halted, or even reinforced.

1. Harald Boersma, Continued Elsevier access in support of German science, 13th February 2017 [↩]
2. Schiermeier, Q., Mega, E. Scientists in Germany, Peru and Taiwan to lose access to Elsevier journals. Nature 541, 13 (2017). https://doi.org/10.1038/nature.2016.21223 [↩]
3. “DEAL and Elsevier negotiations: Elsevier demands unacceptable for the academic community”, 5 July 2018, German Rectors Conference press relase, https://web.archive.orga/web/20181219162556/https://www.projekt-deal.de/elsevier-news/ [↩]
4. Haucap, J., Moshgbar, N., & Schmal, W. B. (2021). The impact of the German ‘DEAL’ on competition in the academic publishing market. Managerial and Decision Economics, 42(8), 2027–2049. https://doi.org/10. 1002/mde.3493 [↩]
5. Fraser, N., Hobert, A., Jahn, N., Mayr, P., & Peters, I. (2023). No deal: German researchers’ publishing and citing behaviors after Big Deal negotiations with Elsevier. Quantitative Science Studies, 4(2), 325–352. https:// doi.org/10.1162/qss_a_00255 [↩]
6. Olsson, Lisa, et al. “Cancelling with the worlds largest scholarly publisher: lessons from the Swedish experience of having no access to Elsevier.” Insights-The UKSG Journal 33 (2020). 10.1629/uksg.507 [↩]
7. Elsevier B.V., & MPDL Services gGmbH, Max Planck Society (2023). Projekt DEAL – Elsevier Publish and Read Agreement. doi:10.17617/2.3523659 [↩]
8. Quentin Dufour, David Pontille, Didier Torny. Contracter à l’heure de la publication en accès ouvert. Une analyse systématique des accords transformants. [Rapport de recherche] 206 150, CNRS; Comité pour la science ouverte. 2021, pp.81. ⟨halshs-03203560⟩ [↩]
9. I won’t get here into some society journals excluded from the agreement, either because they won’t go hybrid or because they thought they won’t get paid enough by Elsevier. On the specific question of learned societies journals in such deals, see The Brief https://www.ce-strategy.com/the-brief/out-of-reach/ [↩]
10. Didier Torny. From paywall builders to data tracking moguls or… How the big publishers have put on a new super vilain costume. Politics of technoscientific futures, EASST, Jul 2022, Madrid, Spain. ⟨hal-03885480⟩ [↩]
11. Thanks to Björn Brembs for underlying this absence, see his plea for German institutions not to sign the new agreement https://bjoern.brembs.net/2023/09/no-evilsevier-deal/ [↩]

---

---