Reflections on trends in library big deals, consortiums and how it might apply to Singapore?
The news that University of California system cancelled their deal with Elsevier seemed to have caused a bit of a stir all over the world, including here in Singapore and I was asked to do a talk to brief faculty on the latest trends in this area.
The more I researched the more interested I got. What were institutions and consortiums doing to better their bargaining position with the big publishers? Was it all about reducing costs for the "big deals"? How successfully were they? Lastly to really negotiate with any credibility, you had to be prepared to walk away from the bargaining table and cancel and indeed some consortiums and institutions have done so, how were their users coping?
A somewhat embarassing lack of knowledge on my part
I must confess until then my knowledge of these areas was pretty mediocre. It was not that I was totally ignorant, it was hard to miss the deadlines and read about how consortiums in Germany and a few Scandinavian countries like Sweden, Norway (up to until 1 week ago!) as well as Hungry were in many standoffs or had cancelled deals with Elsevier and other big publishers.
I also read about how some of them have signed "transformative deals" like PAR (Publish and Read) contracts with Wiley but how it worked was a bit hazy to me.
This is despite reading Scholarly Kitchen blog almost religiously. While it's true that posts from Scholarly Kitchen can be pretty biased , depending on the "Chef" who is posting, it is *the* place to get an insight into what publishers think and keep track of issues that impact Scholarly Communication.
In particular, the blog covers Library negotiations with publishers in depth like few other places,
but I always tend to skim articles covering such topics, rather than read in depth.
Part of this lack of interest was because I have always been told, Singapore simply lacks the scale necessary to have any bargaining power in as a country because we are too small, which is worsened by the fact that unlike countries like Hong Kong, Australia, Singapore doesn't have a history of library consortiums.
Add the fact that libraries in Singapore (like a lot of asian countries I suspect) have traditionally made a stance to favour Green OA (self archiving in repositories) over APC based Gold OA (paying a charge to the journal for the paper to be made open), this makes it mostly pointless I felt to read about "transformative agreements" (see more about that later) that aim to shift payment from paying to read to paying to publish.
Does Singapore really favout Green OA route over Gold?
Just for fun, I ran an analysis using the nifty open tool Lens.org on articles published from 2014 to 2018 to see if it was really true Singapore OA output is mostly done using the green open access route (via self archiving in repositories) rather than Gold open access (via Gold Journal or Hybrid Journals).
I would take these results with a pinch of salt, since I didn't take a lot of time to validate the results, but let's just have a look shall we?
Country Output by OA color - via Lens
UK does seem to have a lot of hybrids. Of all the OA articles in the UK, 34% are hybrids and 56.6% overall are Gold OA (including hybrid). Of all OA articles in US, 23.4% are hybrid and 50% gold.
In terms of total output UK has 18.3% in Hybrid, 30% Gold (including hybrid). US has 8.8% of all output in hybrid and 19.0% Gold (including hybrid). How does Singapore compare?
Country Output by OA color - via Lens
Singapore actually has a slightly higher hybrid rate of 27.7% (of all OA) compared to the US with a 60.4% Gold OA rate (of all OA).
Of course the overall level of hybrids in Singapore is still relatively low at 8.3% of all output from 2014 to 2018 and 18.1% Gold OA overall. Still one can imagine this can be quite a big chunk of money paid to publishers if they are all APC bearing. Is anyone keeping track of this?
As a sidenote, you might be wondering why I'm not using Scopus or Web of Science as a data source. Part of the reason is as a courtesy/ convenience for any readers who have no access to these paid databases, while Lens.org is open to anyone to use.
Also Scopus makes it difficult to get the statistic I need because it lumps all types of OA including hybrid, OA Journals, "promotional OA" into one.
Using Web of Science as a sanity check, I got 17.7% of all Singapore output as Gold which matches the 18.1% from Gold OA, though I get a only 3.5% of hybrids from all output as compared to 8.3% in Lens.org. Differences can be attributed to the difference in the journal coverage of course.
Does the relatively high % of hybrid articles and Gold OA means Singapore now favours Gold vs Green? Not necessarily.
Firstly Gold OA, is not all APC carrying. So it might be the higher Gold OA rate including a ton of publishing on non-APC Gold.
But hybrids almost certainly are APC carrying and here Singapore's rate of 27.7% is pretty much similar to the rate in the US.
That said Singapore research tends to be pretty international, of the 63,054 papers attributed to Singapore affiliated authors, only 24,769 have all Singapore affiliated authors. In other words over 60% are international collaborations with at least one author from a non-Singapore affiliated institution and possibly the APCs were paid by those author. (In particular if they came from a institution that had a PAR (Publish and read) deal such as Germany and Wiley)
Reflections on developments internationally
Given the larger than expected share of Gold OA (including hybrid) in Singapore output this made me even more keen to really start reading and thinking about the deals struck by other institutions in the US and Europe and to wonder about the implications these developments would have for Singapore. For instance how much would we save if we managed to get one of such deals signed? Should we follow the lead of countries in Europe to switch to a different transformative type of deal?
I'm still digesting everything as a lot of this is new to me, but what follows are some of my reflections of the current state of affairs and things that surprised me as I learnt more.
1. Quite a few libraries have cancelled "big deals" with Elsevier and other big publishers
Developments occuring faster than I can change my slides. As of April 2019 Norway has signed a RAP deal with Norway
As I was doing my research, I found that quite a few institutions and national consortiums have cancelled deals with big publishers.
SPARC (Society of Publishing and Resources Consortium) has a big deal cancellation tracking spreadsheet that tracks such cancellations. A quick filter of it shows as many has 50 cases of cancellations of "big deals" with big publishers (typically Elsevier, Wiley, Taylor & Francis, Springer Nature and Sage who collectively dominate >50% of journal publishing business)
In case you have never heard of the concept of "big deals" , how it works in simplistic terms is like this. Unlike in the old days where we selected individual titles to subscribe, these days most academic libraries would instead sign on to big bundles of journals with the big publishers. By doing so an academic library could get the benefit of getting access to thousands of journals belonging to one publisher rather than paying for each individual title.
Of course most of those titles would be titles that you didn't really want in the first place together with the "core" titles that you did want, but overall it was of great convenience to your users who would have access to thousands of titles.
It also helped reduce adminsrative burden for the librarian (who can know in advance what journals users would access in the future?), but that was of lesser importance.
For a taste of how it works, I found that for one of our big deals, you needed only the top 10% of journal titles by downloads to make up 75% of the total downloads. This sounds like we should have just subscribed to the 10% of titles right? Still if you cancelled the big deal and negotiated only for the core titles, it might cost just a bit less and your users would have to wait if they wanted something in the remaining titles of the big deal while you tried to buy the article they needed via Document Delivery.
2. The disruption on cancelling big deals can be less drastic than expected.
For a librarian like myself who wants to please our users as much as possible, the idea of cancelling the journal "big deals" which consists of bundles of thousands of journals at one go seems to be impossible, as it would mean losing access to some of the most important journals for many disciplines.
Obviously this means that any "negotiations" we can do with publishers is extremely limited, as the parties on the other side of the table know that we have really no option but to sign the big deal. This is of course why the big deal is so dangerous and that most institutions have been paying at least 3-5% annual increases for at least a decade if not more. This is of course not sustainable.
So how did institutions that walked away from the negotiations survive when the big deals were cancelled? Some like Florida State University cancelled the big deal but switched to subscription of a linked core set of journals, but there are others like the researchers in the Germany and Sweden consortia (who represent over 300 institutions) have had no access to Elsevier journals on Sciencedirect since Elsever cut access in July 2018 . In fact as of time of writing that access has still not being restored!
So in effect these Germany researchers have had no access for 9 months! According to an article I read while some researchers are unhappy to be inconvenienced there is no full scale "faculty revolt".
This seems to be possible due to a variety of reasons
1. A lot of the instititutions have "perpetual" access to back issues. So say you cancel the big deal from Jan 1, 2018. Your researchers might still be able to access all the older articles (possibly only from core titles) before that date. So it is only the newer ones going forward that access will be lost.
2. Strong ILL/DDS networks. It seemed there is at least one institution in Germany that still has access to Elsevier so DDS from that institution helps provide some access. Still a librarian from Germany was quoted as saying the number of ILL/DDS transactions did not really surge as much as expected indicating people were getting access from alternate sources.
3. What other sources you might say? Well it helps the level of Open Access is much higher today, but there is also the presence of Sci-hub....
Many of the institutions that have cancelled seem to be providing very similar guidance to faculty on alternative access. Below shows a pamplet provided by University of California (even though as of time of writing access has not being cut yet).
Alternative Access to Elsevier Articles - Guidance by UC to researchers
Notice the cheeky mention of not endorsing use of Sci-hub at the bottom...
3. A lot of details such as prices are increasingly becoming open
One of the tried and tested truisms as a librarian I was told was that most publisher deals had non disclosure agreements included so that prices could not be made public. Things were so bad that researchers had to use Freedom of Information Act to get the figures. Obviously keeping big deal prices secret benefits publishers...
However things seems to be changing and many of the deals announced recently (as you will see) have the prices announced. Granted quite a few are leaks but some are officially released.
One stunning example is the University of Virginia Library. On their collection disclosure page, they list detailed price information on the major vendors such as Elsevier, Sage, Wiley and SpringerNature
Collection disclosure page of University of Virgina
It's unclear to me how widespread this trend is e.g. some States in the US have laws against NDAs so libraries in those states can and often do disclose prices openly.
4. Many institutions are aiming for more than reduction in price but also want a transformative agreement
Recently, the French consortium Couperin announced they have come to an agreement with Elsevier. Among other details leaked, the 4 year deal would have a 13.3% reduction in subscription fees over 4 years.
https://www.soundofscience.fr/1754
I believe and please correct me if I'm wrong, it works out to such that on the 4th year, the cost will be 13.3% discount off the subscription fees on 2018. For institutions that are used to 3-5% increases this is almost a miracle!
On top of that, the deal includes 25% off any APCs paid to Elsevier journal as well as a provision for setting up some sort of feed such that author accepted manuscripts submitted to Elsevier will automatically be setup (to be linked from?) HAL the open access repository of the french research council (after embargo of course). Reminds me a little of the Elsevier arrangement with University of Florida's institutional repository?
This feels like a great win to me if all you care about is subscription fees and some have speculated that Elsevier was willing to do such a good deal, due to a string of highly publicised cancellations in the weeks before.
While there are some people who think this was a good deal, quite a few do not.
Leaving aside some doubts on whether the percentage decreases are really over the 2018 rate, a lot of the disagreement as always comes from disagreement on goals.
My understanding is some believe what libraries should do is not just secure one off discounts on subscription fees but shift the business model to one that we achieves open access at a reasonable cost with a transformative agreement
In Europe, the existence of Plan S which requires all research funded by major european funders to be made open access immediately. This will push authors towards open access and makes this type of agreement even more critical.
I understand while it is possible for self archiving to repositories to be compliant with Plan S, most repositories as yet do not meet the technical specs stated in Plan S. So this means a lot of authors will be paying APCs (article processing charges) to hybrid or normal Gold Journals, which will allow publishers like Elsever to recoup any discounts from APCs on subscriptions (note :Plan S bans hybrids).
What is a transformative agreement?
Take University of California negotiations break down with Elsevier and subsequent cancellation that created waves both in the news and the stock market (RELX stock listed on NYSE dropped 7% in a day!)
RELX stock price dropped 7% greatest single day drop since Nov 2016
Why did negotiations breakdown with Elsever? Looking at this presentation entitled UC and Elsevier
A blueprint for publisher negotiations at CNI presented by UC staff in April just days after the news of the cancellation tells us quite a bit.
I'll be using the slides from the above talk liberally from here on.
UC and Elsevier - A blueprint for publisher negotiations
While cost saving is listed as the first point (also see later what this actually means), we see the goal of making default OA publication for all UC corresponding authors as well as "transformative agreement that intergrates publishing and reading with offsetting"
UC and Elsevier - A blueprint for publisher negotiations
A lot of these deals like to talk about "cost-neutral" or "revenue neutral". Of course even achieving a stay in annual increases in an industry where 3-5% is considered "industry practice" is a win.
But what transformative agreements that UC and other institutions are negotiating for goes beyond this. The idea roughly is that there is enough money in the system right now (aka the money we pay to publishers) to shift to a business model where we pay publishers for publishing our authors papers open access rather than access for reading.
So far, there seems to be two kinds of such agreements - Publish and Read (PAR) and Read and Publish (RAP).
The former only pays for publishing and has reading access bundled in for free while the latter has components in the bundle for both reading and publishing. There are some nuances on how those two types of deals will play out but the key point is money that institutions are paying for access will be shifted to APCs for publishing rather than just access.
For instance UC pays $10 million to Elsevier typically for subscription. In the new transformative deal they proposed, 1 million would be set aside for reading and the remaining 9 million would be set aside to pay for UC authors (who are estimated to publish 4,500 articles in Elsevier) to be made open access at the negotiated APC of $2000 per article.
Typically such agreements will take into account the typical output published with the publisher and then negotiate a price. Institutions will of course want to end up projected to be paying the same amount as before the deal, while publishers will want more. In fact, Elsever's stance that you cannot get both access for reading and open access of articles for the same price as it would be paying for 2 services at the price of one has I suspect always led to negotiations breaking down.
In case you are wondering this isn't a radical idea, such deals have been signed by the big publishers in the past before, including Wiley with German consortium, UC themselves signed a RAP deal with Cambridge University Press days after this talk.
More surprisingly, soon after this talk, Elsevier signed their first RAP deal with Norway! It's unclear to me if the agreement was projected to be "cost neutral" to the institutions, I suspect not.
In any case, transformative agreements are quite complicated compared to normal big deals for just access and I am still trying to understand the nuances. If you are interested in learning more about transformative deals, I highly recommend "Transformative Agreements: A Primer" by Lisa Hinchliffe.
5. A major benefit to transformative deals is transparency & elimination of double dipping
In the talk mentioned at CNI, UC produced this diagram.
My first thought seeing this is, what an amazing amount of APCs paid by UC. For every $4 paid in subscription fees , $1 is paid to publishers through APCs.
But I've been reminded this is UC , whose 10 campuses have authors that produce 10% of all research output so I can imagine the APCs add up.
A note in the slides says it is "our internal analysis calculation for what our authors are spending on OA based on publication volume and list price APCs, including both full OA and hybrid.", so not all $10 million is potential double dipping but it still shows the complete lack of transparency on how much institutions are paying big publishers.
One of the virtues of transformative deals is by explictly negotiating and putting such APCs cost in the deal along side any fee for reading access we will have more transparency on the total bill paid to publishers rather than what UC describes has having two seperate independent actors that might lead to "uncontrollable spending"
As a sidenote, I do notice that by making libraries the middle man and gatekeeper in the handling of APC it does secure a role for the libary even in a mostly OA world!
6. Do we have information on amount of APCs paid to publishers? Can we estimate it for Singapore?
Negotiating creditably means you need to be ready to walk away from a deal and of course UC detailed the amazing amount of support they have had. For example there were public letters of support from the President's Office, Faculty Senate and provost library advisory committees. You need to communicate a whole lot and have faculty on your side if there is a need to cancel.
But beyond that you also need information which will inform your goals, and the CNI talk included a whole statement on "The critical role of data analysis" by a "Open Access Collection Strategist, California Digital Library"
The part that made me most curious was, how much is Singapore paying to publishers in APC or other fees besides the subscription fees we are paying?
Singapore academic libraries are quite well funded , I would guess the 6 Universities combined with A*Star and medical hospitals would come close if not exceed 40 million USD subscription fees.
I highly doubt we are paying $1 APC for every $4 though but is it possible to estimate how much we are paying?
UC notes that their APC figure is based on "internal analysis" , which I suspect involves quite a bit of estimate. In Singapore, my guess is it might be even harder to get the figure as I suspect we have never tracked this formally.
Let me try a rough back-of-the-envelope calculation
My rough back-of-the-envelope calculation of how much APCs Singapore paid in 2018
The difficulty is that while you can easily get to see how many Singapore affiliated authors publish in Hybrid or Gold journals, for multiple country affiliated papers there is no way to know where the funds came from.
The method I came up with roughly works like this.
1. Find all Singapore institution affiliated author only papers in 2018 and filter to Hybrid and Gold Journals. Not all of Gold journals will be APC bearing of course, but you can match against a APC price list (there are scripts out there that do it) to only include APC bearing ones.
2. What about papers where there are both Singapore institution affiliated authors as well as non-Singapore institution affiliated authors? If they published in a Gold or Hybrid journal, I simply filter to papers where there is an acknowledgement to a Singapore based research funder (most of whom funds come ultimately from tax payers money just like the universities in Singapore).
The logic behind this heuristic is most likely the funding for APC came from the funder who was acknowledged in the paper
Of course this probably will be an underestimate since there will be papers which there is payment by a Singapore institution affiliated author who might not choose to acknowledge anyone.
Using both Lens and Scopus, I got between 1,000 to 1,200 articles published in 2018 in hybrid and Gold Journals (almost all APC bearing). Depending on the APC price paid, say an average of $2000 USD per article , that's about $2 million USD conservatively?
Conclusion
I will be first to confess I really don't fully understand all the details about transformative agreements. There is a double whammy in that I have zero idea on how library consortiums work and I can imagine part of the complexity of consortiums (even without the complexity of how PAR and RAP deals) is to agree how to allocate costs to different institutions, particularly if they are very different in size and scope.
It occurs to me if publishing fees is the main component of such details, suddenly the relative reseach output becomes the main cost allocator not FTE as per usual, so some institutions might actually pay a relatively bigger share despite being much smaller in terms of FTE. We won't know until we do the calculations.
One of the arguments I have always heard is that Singapore is too small to be worth doing and that publishers would ignore us, but then countries like Norway which has roughly the same output as Singapore in 2018 based on Scopus and Lens data, has consortiums that have signed transformative deals with Wiley and most recently with Elsevier and even if there is no interest in that, a Couprin type normal subscription deal with 13% discount looks good?
Of course the research landscape in Singapore is different from in Europe and the US and it's naive to think everything applies.
Still I wonder, if more and more countries in particular Europe and US, shift towards Open access due to such deals, the share of Open Access will sky rocket, this means that it makes it easier for institution in countries not in such deals to cancel or threaten to cancel subscriptions because it becomes less and less a loss of convenience when that happens.....
I guess this is what publishers fear?