Flipping to Open Access

  • Joshua Nicholson
  • Mike Taylor
  • Glenn Hampson
  • Michael Eisen
  • Angela Cochran
  • Rebecca Kennison
  • Caroline Black
  • Danny Kingsley
  • Alicia Wise
  • Ivy Anderson
  • Kim Barret
  • Joyce Ogburn
  • Roxanne Missingham
  • Jeff Mackie-Mason


OSI is a global collaborative effort between major stakeholders in scholarly publishing to improve how research information gets published, shared and accessed. The intiative is a 10-year series of annual meetings. The following dialogue is adapted from the listserv of the Open Scholar Initiative (OSI) and is intended to broaden the reach and transparency of the conversation.   

Laurie Goldman, Editor-in-Chief of GigaScience: 

I’m sure you’ve all seen this,  but if not,  pretty dramatic.


If publisher’s do get on board, we really need to consider affordable APCs. I’ve heard many researchers comment that it used to be they couldn’t afford to read articles, but now they can’t afford to publish.

Perhaps one of the discussion groups next year should be developing affordable APC models. 

There are definitely solid, reasonable business models for publishers to use that support their processes and make some reasonable amount of profit  (hopefully to be used for improving scientific communication… ah but, share holders…)

Mike Taylor, Software Engineer, Index Data and Research Associate, University of Bristol

There’s no mystery or challenge about APCs. It’s well established that a for-profit publisher can do just fine on APCs in the $500-$1000 range. The reason legacy publishers like Elsevier and Wiley charge $3000 instead is (A) a third of that is siphoned straight off into shareholders’ profits, and (B) another third is spent on negative-value stuff like paywalls, lawyers and PR.

Those legacy publishers CAN compete in the new world; but they will need to make radical changes to do it. Maybe more radical than they are prepared to make. I can certainly see why they would choose to keep collecting rents on the irrationally popular brands that they own, instead of facing up to these challenges.

Angela Cochran, Director of Journals, American Society of Civil Engineers (ASCE)

For many publishers, having an APC at $1,000 would mean a complete change in workflow and vendors and a decrease in services currently provided. It also requires building in an ecommerce where none currently exists. It’s not impossible but it’s only going to happen if we know it will be supported by authors (and in our community that is not happening).

I can’t publish a paper for $1,000.

Rick Anderson, Associate Dean of Libraries at the University of Utah and President-Elect, Society for Scholarly Publishing (SSP)

Nor can PLOS, apparently: http://blogs.plos.org/plos/2015/09/plos-publication-costs-update/

Jeff Mackie-Mason, Dean of Libraries, University of California Berkeley 

Much of the discussion about whether authors can “afford” APCs assumes the world as it is today, not the world as it would be after adjustment to a flip from post-publication (subscription) to pre-publication (APC) financing of publishers.  In a primarily Gold OA world, if it comes to pass, the gazillions that libraries currently spend on subscriptions would be freed up, and as funding flows adjust, those funds (which currently support the publishing industry) could be redirected to flow to the publishing industry through authors – that is, the *sources* of funds for authors to pay APCs in the future will be different than they are today.  Systems adapt to changes in their environment.

Danny Kingsley, Head of Scholarly Communication, Cambridge University

The question is not whether an article can be published for $1000.  APCs for fully open access journals are more or less reasonable as far as I can tell (as someone who processes thousands of the things). These APCs are set to cover the cost of the whole journal production and made open access.

What I object to is transporting that cost onto a hybrid article where the work involved in producing the article is covered by the subscription*. There is a minimal amount of work involved in subsequently making the article open access - changing the license (sometimes), possibly (but not usually) indexing it as open. https://aoasg.org.au/not-all-hybrid-is-equal/

*Hybrid* APCs are far too high. That is what has to change. We need to accept that hybrid has just been a wonderful gravy train for publishers and has failed as mechanism for flipping to gold. http://www.slideshare.net/DannyKingsley/be-careful-what-you-wish-for-unexpected-policy\n-consequences

 (*I know Elsevier claim that they use their APC to produce the article in hybrid journals but let’s leave that to one side and talk about every other publisher).

Kim Barrett, Dean of the Graduate Division, University of California San Diego (UCSD)

At least at my institution, I feel it is highly unlikely that monies currently devoted to subscriptions will be provided to investigators to pay for APCs.  It will simply provide a way to pay bills elsewhere than the library.

Joyce Ogburn, Dean of Libraries, Appalachian State

Don’t forget that with more open content comes the need for tools to take advantage of this content. Some tools may be free while others will not.  We have to look ahead to think about what scholarship will require to be created, delivered, mined and curated most effectively.  The trade offs in costs may shift in many different ways. Systems of perpetual access and open discovery may prove to be quite costly but we can hope that there will at least be completion in these arenas.

Jeff Mackie-Mason, Dean of Libraries, University of California Berkeley

Kim, you think that after years of supporting the research of its faculty through providing funding for the publication of journals (through paying subscriptions), a top research university like UCSD will decide it’s going to opt out and stop providing that support?  And that the faculty will stand for that?  And that UC San Diego won’t care when it has trouble recruiting top scientists because it decided to cut support for scholarly communications?  I doubt that: provosts and chancellors lose their jobs for less.  

But this is more complicated – and the world more likely to adjust – if there is a whole-scale flip to gold OA.  First, we’re talking a system change that will take years, and administrative and funding process will adjust over time.  Second, it won’t simply be up to the chancellors, presidents and provosts of individual campuses to decide.  Most of the money that pays the cost of scholarly publishing, ultimate, comes (in the US) from federal funding agencies, through indirect payments.  (The cost of running the research library – including on your campus – is one of the costs that is included in the calculation of campus overhead for the purpose of setting the indirect recovery rate.)  If the payment for the publishing industry switches from subscriptions to APCs, the NIH and NSF etc will be very involved in deciding, for example, whether their funds distributed to support scholarly communication will continue to flow through indirects or through direct payments as line items on grants to support the costs of publication, APC.  

In aggregate, the research system (funding agencies, universities, research institutes, etc) are currently willing to spend $X to support scholarly publishing.  Why does it make sense to think that if the last step in the process is re-organized (payments to publishers are made pre-publication as APCs rather than post-publication as subscriptions) that this will drastically change our (the research system’s) willingness to pay?  (Yes, many people think total payments will increase and campuses will refuse that – I disagree that payments will increase, in fact, after the system adjusts there should be more competition and thus lower prices (see http://goo.gl/tfQs7o for discussion of why) – but set that separate hypothesis aside for this discussion: assume the total industry cost of publishing is the same in the pre-publication payment world: why would the major decision makers decide that a change in when the checks are written should drastically change their willingness to write the checks?)  

Suppose the Army Corps of Engineers decided to build a dam upstream of New Orleans, stopping the flow of the Mississippi into the delta.  Do you think the upstream Mississippi would simply say, “OK, time to stop flowing downhill” and that would be the end of it?

Kim Barrett, Dean of the Graduate Division, University of California San Diego (UCSD)

I certainly hope you are right.  I don’t know how long you have been at Berkeley, but if you were not around for the last CA budget crisis you should ask Brian what happened to the library’s budget at UCSD.  When faculty salaries have to be paid, it is one of the obvious things to cut.  the option of providing APCs as line items in grants is highly problematic too.  Many people have modular budgets so money spent on APCs takes funds away from the research itself.  And even if you don’t, budgets are invariably cut before you get the money.  The Wellcome Trust put its money where it’s mouth is by providing extra funds for APCs.  The NIH said you “may” use funds for APCs but never gave an extra penny.

Ivy Anderson, Interim Executive Director and Director of Collections, California Digital Library (CDL) 

The problem with the Wellcome Trust policy is that, because they’ve committed to paying APCs over and above grant amounts and naturally don’t want to tell authors where to publish, there are no economic controls or market forces at work. An APC economy will only work if there are trade offs that force someone or some entity to make economic choices.

Alicia Wise, Director of Access and Policy, Elsevier

APCs vary in accordance with the quality of the journal that the OA article is published in.  Average APCs are higher for hybrid journals than they are for fully gold journals. However the reason for this is around the value researchers place in hybrid journals viewing them, rightly, as publications with high editorial standards, higher impact factors, greater levels of citation, higher standards of peer review (and therefore higher rejection rates), and higher production standards.  The link between APC price and quality is clear from the independent review commissioned by the Universities UK OA Monitoring Group - see pages 58 & 59 in Monitoring the Transition to Open Access available online at http://www.researchinfonet.org/wp-content/uploads/2015/09/Full-report-FINAL-AS-PUBLISH\nED.pdf

Hybrid journals support open access and are a significant reason why the UK is now ahead of the global average in the take-up of immediate OA models.  I would encourage members of this list to have a look at this paper by the Publishers Association: http://www.publishers.org.uk/EasySiteWeb/GatewayLink.aspx?alId=20877.  It notes that “uptake of the hybrid option grew at an annual CAGR of 61.7% between 2012 and 2014, while the uptake of the fully gold option charging APCs grew at just 15.4%”. It also explains that one of the reasons why it is assumed publishers are double dipping is because institutions are seeing an increase rather than a decrease in their overall spend. However, rather than being due to publishers charging twice for the same content – as Danny wrongly suggests - this is due to the fact that institutions are paying two types of cost (as opposed to paying twice for the same thing): institutions are paying to continue to subscribe to a growing number of subscription articles, and also separately paying APCs to cover the cost of publishing research on an immediate open access basis.

Mike Taylor, Software Engineer, Index Data and Research Associate, University of Bristol

“APCs vary in accordance with the quality of the journal that the OA article is published in.”

Now, let me think. What would be a more polite word for “bullshit”?

This paper looks at (among other things) the relationship between APCs and Impact Factors (which we all know are appallingly flawed, but they’re out there):


Figure 1 shows only an extremely weak correlation between APC and IF, with extremely wide spread. It also shows that hybrid journals have much, much higher APCs that similarly prestigious true-OA journals.

It’s clear what’s going on here: the same incumbent publishers that have been price-gouging on subscriptions for decades are now price-gouging on APCs. Simple.

Ivy Anderson, Interim Executive Director and Director of Collections, California Digital Library (CDL) 

Alicia, respectfully - I know  we’ve had this conversation before - this explanation of non-double-dipping is completely fallacious in my view. In the pre-hybrid era, article growth was also a standard publishing phenomenon, yet publishers managed their costs in such a way that there was no direct correlation between article growth and subscription costs. So to claim that this growth should not be offset against the subscription price is, to my mind, another way of saying “we used to absorb growth in article output as part of the business of running a journal and making decisions about the amount of growth we were willing to sustain, but now we no longer have to absorb that growth because we can obtain additional revenue for it.”  That sure sounds like double-dipping to me. 

Alicia Wise, Director of Access and Policy, Elsevier

I am quite confident that article volume is factored into pricing and long has been.  Ours certainly – as explained in our pricing policy at https://www.elsevier.com/about/company-information/policies/pricing.  I suspect this is true for other publishers too, but they would need to speak for themselves on this point.  Thinking back to my library consortium days, however, I remember conversations with publishers (plural!) regularly included such topics as content volume, content quality, usage, and the like. 

There are undoubtedly real cost pressures on libraries who wish to play a role in funding APCs. It is however untrue that we add to this pressure by charging twice for the publication of the same article, and the continuing allegations of double-dipping are both unfair and unhelpful.  

Roxanne Missingham, University Librarian (Chief Scholarly Information Services), The Australian National University

Many interesting thoughts.  I have considerable concerns about the flipped model. 

It provides an economic incentive to produce as many articles as possible, not an incentive to produce quality articles.  The tension between these goals will rub scholarly publishing with potentially dire consequences.

I have not seen any research that suggests that keeping APCs at their current level would be sustainable.   Some suggest that the flipped model would result in radically increased charges just to fund the current system – for example http://bjoern.brembs.net/2016/04/how-gold-open-access-may-make-things-worse/

I do agree that based on analysis of publications and quality the hybrid, open and full subscription with APC models do not produce differences in quality of journals.  From our perspective, quality relates to assiduous work by the editors, reviewers, authors and the discipline to ensuring that the scholarly outputs are of high quality.  I do think that when we look at the issue of quality it is just as important to consider journals produced outside the major publishers as inside that group.  In some cases, the small open access journals have outstripped their commercially published peers in quality. And of course quality is not simply measured by citations – it is a very complex area, differentiated by discipline and the nature of scholarly communication. I can think of a couple of articles in my field that are cited regularly as examples of dreadful scholarship and of course that means they get cited and the metric illusion is that they are measures of success.

I question the impact of the flipped morel on early career academics – shifting payments has provided significant barriers to their publishing – we regularly receive feedback from postdocs and early career academics about this issue.

If the flipped model needs to not replace current subscriptions but provide revenue for publishers the likelihood is that there will be significant increases in costs to quality. 

For our university the cost increase in the flipped model would be that, on the basis of the current APC charges (without any increases), it would cost more than  double the budget currently allocated to purchasing material by the library.  It is difficult to see that as an appealing solution for university managers.

Jeff Mackie-Mason, Dean of Libraries, University of California Berkeley

Kim, I agree completely…but these are two different things.  That is: yes, university library budgets have been declining (pretty much everywhere, but especially at public research universities since the public has been so drastically disinvesting in higher education).  But the change in university library budgets is not driven by any of the (modest so far) shift towards OA publishing, nor would a change toward pre-publication payment (APC) versus post-publication payment provide any rationale for reducing campus funding to support scholarly communications.  Scholarly communications funding may continue to decline, but let’s not conflate two different issues.  (Indeed, if anything, gold OA, by giving authors meaningful incentives to think about publishers’ prices, and thus by increasing competition among publishers, offers what I think may be the best chance we have of *lowering* the cost of publishing to universities, which will help our shrinking budgets).

Caroline Black, Editorial Director, BioMed Central (SpringerNature)

Agreed that this is an interesting discussion. The APC model doesn’t necessarily lead to a publish-everything approach, however. We publish selective journals with 85% plus rejection rates. Even those journals that offer to publish all studies conducted and reported soundly (to ensure the availability of useful results and avoid research waste - surely a benefit of the APC model) have rejection rates around 40-60%. A journal that doesn’t have a reputation for quality won’t attract submissions. 

Rick Anderson, Associate Dean of Libraries at the University of Utah and President-Elect, Society for Scholarly Publishing (SSP)

Another problem with the Wellcome Trust policy (and all others like it) is that it inevitably shifts funding away from research and towards dissemination. You can’t spend the same $1,000 on both research and APC subvention. If Gold OA is supported by funders, less research will be done. Whether or not this is a worthwhile tradeoff — it may well be — is a discussion worth having, but it’s a tradeoff that we need to acknowledge and deal with.

Mike Taylor, Software Engineer, Index Data and Research Associate, University of Bristol

This problem is only a problem if, while counting the $1000 per article that you spend on the APC, you ignore the $5000 per article that the world doesn’t pay for subscribing to the article. (That’s $10bn pr year subscription revenue to publishers, while 2 million papers are published. Figures from the STM Report.)

Rick Anderson, Associate Dean of Libraries at the University of Utah and President-Elect, Society for Scholarly Publishing (SSP)

No, the problem is a problem either way. The money that world would have paid for access is not money that the world would otherwise have paid in direct support of research if it had free access.

Pointing out what other players in the system save by having free access is an argument to make in support of the tradeoff, but it doesn’t change the fact of the tradeoff when it comes to funder expenditures.

Mike Taylor, Software Engineer, Index Data and Research Associate, University of Bristol

It is true that money spent on dissemination (whether subscriptions or APCs) can’t also be spent on research.

What is not true – which I read your earlier message as implying – is that there will less money for research if we spend money on APCs. The reason that is not true is that when we spend money on APCs, we can stop spending (more) money on subscriptions. And the savings can be put into doing more research.

Unless publishers get away with double-dipping, of course.

Michael Eisen, co-founder of PLOS and Professor of Genetics, Genomics and Development, U Cal Berkeley

It is as unreasonable to say that spending money on APCs “inevitably shifts funding away from research and towards dissemination” as it is to assume that money currently spent on subscriptions would automatically be used to increase research spending. What we can say is that:

a) The research system currently spend ~$10b/year collectively on publication

b) If we paid APCs of $1,000 on all of those articles that number would be $2b/year

Obviously, the $8b/year difference isn’t automatically going to go to into a research kitty - there are countless examples at my university of how money saved by introducing economic efficiencies in one place get applied to completely different areas. But that’s not really the thing we should be thinking about. Rather we should recognize that the massive savings under this model constitute a powerful argument to the institutions that collectively spend this money (primarily funders and research institutions) to support this shift. 

That is, the path here is not to shift everything to OA and pray that the money doesn’t come from research funds, but rather to collectively go to funders and research institutions and get them to realize that, by committing to spend money in support of APCs/open access, they will save money. 

Jeff Mackie-Mason, Dean of Libraries, University of California Berkeley


First. *most* of the funding for dissemination already is flowing from government funding agencies (through indirect cost payments).  And if the cost of dissemination in a gold OA world goes down much (which it will if gold OA is implemented in a way that increases competition), then this founders are likely to have *more* money for research.  (Unless legislatures say, “ah, the cost of dissemination is down, let’s move those savings over to spend on job training” – this conversation frequently forgets we’re in a dynamic adaptive system and that if the publishing industry is changed, other practices can adapt rather than remain fixed).

The only issue for your concern is how much of dissemination is being paid for by sources who aren’t also paying for research *production* (e.g.  lawfirms that subscribe to journals but don’t spend any direct budget on scientific research).  I haven’t seen any hard evidence of how much this is.  And it still leaves open the question of whether there might be another adaptation so that these research consumers might still contribute in some way other than subscriptions to the cost of the research life cycle (production and dissemination).

Rick Anderson, Associate Dean of Libraries at the University of Utah and President-Elect, Society for Scholarly Publishing (SSP)

Folks, let me be clearer than I thought was going to be necessary: when any particular funder seeks to underwrite both research and APCs, the tradeoff lies in the fact that the funder will inevitably support less research. Arguments about felicitous effects of such a move elsewhere in the system are arguments in favor of the tradeoff, not arguments that the tradeoff isn’t real.

Glenn Hampson, Executive Director of National Science Communication Institute (nSCI)

Can I ask for a clarification? I think a lot of folks on this thread (including me) aren’t as familiar with the nuances of university funding streams Specifically:

1.       When a funder gives money to research, some money to publish that research is normally included in the grant—at least publishing activities like building a website and doing some outreach, but not necessarily paying for author publishing charges. Is your argument, Rick, that in a flipped world where authors are required to pay to publish, they would necessarily need to use more of their research funds to pay for APC changes because their grant funds would be insufficient to do so (except for Wellcome-funded research, since they are providing APC money on top of research grants)? How common is this situation and how much of a deficit are we talking about? Are there any facts here—research showing how much of a deficit would exist? And do some universities have a separate stream of funding to support this expense (or are these two streams somehow connected)? Which leads to question 2:

2.       When libraries receive funding to pay for journal subscriptions, can (or should) these budgets be flipped to pay for author charges under a gold system? It seems like an apples and oranges argument, where paying for subscriptions benefits the entire university and is therefore a common good, but paying for author charges benefits the author and is therefore an author-specific expenditure that should come out of the author’s pocket (or her department’s)—albeit you can argue, as has been suggested, that it’s a tradeoff for making everything open. But in the real world where budgets grow easily and shrink only under extreme duress, under a flipped system, wouldn’t library budgets shrink dramatically as the onus for publishing shift to authors and their departments? And if this is the case, wouldn’t these authors/departments be free to exercise their publishing discretion differently than libraries, and maybe choose to publish in different journals where price becomes a factor in their decision, or even forgo traditional publishing altogether (the market forces that Jeffrey is alluding to)?

So if I’m understanding this discussion correctly—and I know I’m not so please forgive my stupid questions here----a flipped world could lead to less research spending only if funders don’t provide an extra $2 billion in APC money (in aggregate) and/or libraries aren’t able to convince administrators to maintain budgets at a level that is adequate to help pay for APCs. Is this correct? If so, then the specific tasks we would face in a flipped world would be to:

1.       Make sure that funders include at least some modicum of APC money in grants (although, like Medicare, these below-market reimbursement rates would eventually create tiered levels of access for the haves and the have-nots); and

2.       Make sure that universities buy into the proposition that shifting library budgets from supporting subscriptions (demand-side) to APCs (supply-side) is a net positive, although faced with severe budget pressures, this approach may force a race to the bottom—finding the least-costly option to publish, regardless of quality, which could conceivably have a negative impact on research (although it could also introduce competitive pressures to increase efficiencies and lower publishing costs).

Sorry if I’m missing the thrust here—I think, though, that it might be helpful to sketch out this conversation in more detail in highlight where we’re making assumptions, where we’re modeling assumptions, where we need more hard facts, what the best case and worst case outcomes look like, what needs to happen to achieve the desired outcomes, and so on. This seems like an important conversation, and one you’ll keep having over the coming months, but it would be interesting to try to move beyond competing talking points to solutions. But to do this, I think it would benefit this group and this conversation if the experts among you could please get everyone better educated, in lay terms, about exactly what’s happening with the whole gold/APC flip situation.

Kim Barrett, Dean of the Graduate Division, University of California San Diego (UCSD)

We also need to keep in mind those disciplines that have little or no research funding in the first place, but still need to publish.  Likewise, their journal subscriptions are not being paid for by indirect costs.  There are also many companies (e.g., pharma) that currently pay for access but publish less.

Rick Anderson, Associate Dean of Libraries at the University of Utah and President-Elect, Society for Scholarly Publishing (SSP)

“Is your argument, Rick, that in a flipped world where authors are required to pay to publish, they would necessarily need to use more of their research funds to pay for APC changes because their grant funds would be insufficient to do so (except for Wellcome-funded research, since they are providing APC money on top of research grants)?”

No, my argument is simpler than that. If a funding body (like Wellcome Trust, say) includes money in its grants that is used for APCs, that money is not available to fund research. Such an arrangement creates broader access to the results of less research.  In 2014, Wellcome Trust reported that it had “spent approximately £3.9 million on APCs for 2126 articles in the year 2012-13.” That is money that did not go to the support of new research. It really is that simple.

What’s not so simple is trying to figure out whether the benefit produced by this approach (broader access to research, which is a very significant benefit) is worth the cost (which is also significant). That question is very much open to debate. What really isn’t open to debate is whether or not money spent on one thing can also be spent on another.

“When libraries receive funding to pay for journal subscriptions, can (or should) these budgets be flipped to pay for author charges under a gold system? It seems like an apples and oranges argument, where paying for subscriptions benefits the entire university and is therefore a common good, but paying for author charges benefits the author and is therefore an author-specific expenditure that should come out of the author’s pocket (or her department’s”

I would argue that this scenario is actually more complex than that. When authors publish high-quality science in prestigious venues, the whole university really does benefit. However, the tradeoff when libraries fund APCs is very stark: $2000 will make one or two locally-produced articles freely available to the world, or it could be used to buy a large number of articles (via journal subscription) and make those available to the campus community (but not to anyone else). There are pros and cons to both approaches, obviously.

One response might be to say “Yes, but if all the libraries simultaneously pooled all their money and coordinated a flip with all the publishers, then the libraries’ money would make content freely available to all.” That might be true, but I have yet to see a model for doing so that shows serious promise.

Roxanne Missingham, University Librarian (Chief Scholarly Information Services), The Australian National University

Thanks Glenn.  The whole issue of funding is quite complex.  A major Australian funder, the Australian Research Council, regularly grants less funds that have been sought so we find that costs for dissemination of information are often the first to be cut from the research activities.  In a world where funding is often incomplete or costs may change (eg unexpected increases in equipment), there isn’t a simple ratio between funding for APCs and project funding.

Glenn Hampson, Executive Director of National Science Communication Institute (nSCI)

Of possible interest to this group, this 2015 report from Peter Suber looking at the pros and cons of various OA flip models (ask and ye shall receive). A link to this report just came through on the Science of Team Science listserv: https://osc.hul.harvard.edu/programs/journal-flipping/public-consultation/1/

Mike Taylor, Software Engineer, Index Data and Research Associate, University of Bristol

“There are undoubtedly real cost pressures on libraries who wish to play a role in funding APCs. It is however untrue that we add to this pressure by charging twice for the publication of the same article, and the continuing allegations of double-dipping are both unfair and unhelpful.”

That is surprising. It certainlty doesn’t tally with this analysis by Research Libraries UK: http://www.rluk.ac.uk/about-us/blog/the-costs-of-double-dipping/

 “Let’s take the first publisher listed: Elsevier. In 2013 the 20 institutions surveyed spent in total £14,259,959 on subscriptions and £937,531 on APCs in hybrid journals. It is clear that the UK’s embracing of gold OA brought to Elsevier an increase in their revenues from these institutions of over 6%. The ‘double dipping is impossible’ argument appears to be that these are two completely separate revenue streams. The OA papers are viewed by Elsevier as ‘additional’, over and above what a subscriber gets access to. However, if the UK had not gone for gold, these OA papers would still have been published as subscription-access papers, only available to subscribers. The payment of the APC takes the paper out of subscription-control. If no APC had been paid the total number of papers under subscription access would have been higher. And the subscription income? It would still have been £14,259,959. Without hybrid OA the total from these 20 institutions is £14,259,959. With hybrid OA it is £15,197,490. It is clear that this is additional revenue for the same content – i.e., double dipping!”

Also relevant: Cameron Neylon’s observation when testifying before the House of Commons NIS committee on open access in 2013: http://www.publications.parliament.uk/pa/cm201213/cmselect/cmbis/uc1086-i/uc108601.htm

“The detailed information of how much revenue is coming to a publisher from subscriptions, from article-processing charges, from advertising and from re-print revenue is the kind of level of information that is really required to ensure that you do not get into a situation of double dipping.”

I think this is the bottom line. We can all speculate and assert on whether double dipping is or is not taking place. But until we have fully transparent accounts from the publishers, we can never really know. And unfortunately, the publishers in question have always done all they can to obfuscate of even the most basic pricing information – hence the ubiquity of non-disclosure agreements, and the need for Tim Gowers to have used Freedom Of Information legislation to force universities to disclose how much they are paying publishers.

In the absence of these transparent accounts, which we are not going to be getting any time soon, the only way we can be confident publishers are not double dipping is to take their word for it. And frankly those publishers’ records of exploiting every possible source of revenue does not make me optimistic that they can be trusted on


Laurie Goldman, Editor-in-Chief of GigaScience:

I really agree that transparency on costs is needed. We’re switching publishers next year and trying to work in some sort of transparency of APCs into the contract. Our expectation is that WE can be transparent on what part of the APC we get, but e as an independent journal can’t dictate what a company does, but we think it’s fair to indicate what OUR costs are should we do any revenue sharing with the publisher. (So far we have not- so the entire burden of publishing costs regarding editorial salaries, office overhead (or when we the editors want discounts or waivers for authors, but the publisher doesn’t) are covered completely by BGI).

But, I think it’s fair for us at the journal (especially since we are independent) to be clear about what money, if any, we make on an APC and where it is spent. I think all publishers should do the same- though corporate issues can be quite complex, legally, competitively, and more. So… 

Glenn Hampson, Executive Director of National Science Communication Institute (nSCI)

Hi Mike,

I know you’re not the only one who feels this way. So speaking to you and these others, let me say this:

1. You are engaged in a new enterprise with OSI to be able to speak face-to-face on an at-will basis with a wide variety of high-level stakeholders. It’s a different dynamic than what some of you have been used to. If you’d like to see detailed information, just ask for it. If you take issue with certain points of view, ask for clarification and explanation. A big part of the problem to-date in the scholcomm reform conversation has been this dynamic of launching barrages across each other’s bows without really talking and without seeing each other as part of the same team. Please take full advantage of this forum to talk and think through solutions. Don’t assume that because it hasn’t been done before it won’t happen. And,

2. Labels aren’t helpful here. They’re just positions wrapped in semantics. There is lots of double-dipping that occurs everywhere—some could make the case that all of scholcomm is double-dipping since the university pays to play and pays again to subscribe. I don’t know if this is evidence of ill-will, but for the sake of our working partnership, let’s assume that it isn’t, and that more largely, it is evidence of a system that isn’t efficient and effective right now for anyone, and that’s what we’re all working together to fix. To paint our colleagues into a corner by accusing them of wrongdoing and then expecting them to work together with us for change is just continuing with the same dynamic we’ve tried for the past 15 years. 

I hope this doesn’t sound harsh. This is an interesting digression, but it takes us down a dark and unproductive path. We’re trying to chart a new, collaborative path toward new, collaborative solutions and not rehash old arguments and approaches. So drawing from what you’re asking here, a different approach might be to ask whether it would be possible to get more transparency in the system, what kind of transparency is needed and why, what precedents we can draw on to support this request, etc.

Rebecca Kennison, Principal, K|N Consultants

I likewise hope that we can have different kinds of conversations than some of the ones we’ve had in the past. I would particularly welcome coming as close to complete transparency as possible, as several others have also urged.

In that spirit, Alicia, could you offer Elsevier’s own analysis of the RLUK’s “double-dipping” figures? If hybrid payments are in fact offsetting subscriptions at Elsevier, could you tell us what the mechanism is to ensure that payments of APCs are truly covering the costs of producing the articles for which they are paid and are thereby offsetting journal costs in some way, even if (let’s all admit) we have not seen subscriptions decrease in any perceptible manner? If hybrid payments are not offsetting subscriptions, could you detail for us what the APCs are covering that are not already covered by that subscription revenue?

To Danny’s point – and I’m speaking here as a long-time journals production person – first-copy costs are first-copy costs. APCs, as the name implies, are designed to cover those costs. The non-double-dipping argument requires some proof that the APC (in the case of a hybrid journal) is being used to cover the first-copy costs of that article and that subscriptions are only covering the balance of the costs to produce the content within that particular issue or that particular journal. Is that the level of granularity with which you track APC expenditures vs. subscription expenditures? If so, could you share some insight into how that works within your billing, invoicing, and payments systems so that you can assert as confidently as you do that Payment A goes to cover the costs (or at least most of the costs) of that article within Journal A and that subscription revenues only cover the costs of publishing the non-hybrid articles within that journal? Or is it more complicated than that?

I can, in fact, imagine that any mechanism(s) you have to track hybrids must be particularly complicated since we’re talking about two completely different scales of payment that need to be reconciled: (1) the Big Deal bundles of journals that libraries usually purchase under the subscription model and (2) the cost-per-unit pricing of an APC for a single hybrid-journal article. So perhaps I should ask my initial question with a slightly different twist: What is the current mechanism in place to ensure that APCs paid for hybrid journals that are included in your Big Deal packages are applied in some way to those packages and that accordingly overall library subscriptions to those packages might someday decrease, rather than increase, as they do now? I know I would not be alone in wishing to see some hard numbers that show that money paid for hybrids is having the desired effect of offsetting subscription costs.

Mike Taylor, Software Engineer, Index Data and Research Associate, University of Bristol

Hi, Glenn. Thank you for this thoughtful and constructive intervention.

Unfortunately, I think it is based on a deeply mistaken premise – a

premise so misleading but so deep-seated in some quarters that your

message moved me to write a blog-post about it. You can find it here:


The TL;DR is this: the research community wants certain things (to own

their infrastructure, to make all research freely available

everywhere, to pay as little as possible for publication services) and

legacy publishers want the opposite. With the best will in the world,

it is simply not possible for any outcome to satisfy both groups.

See the blog-post for a fuller discussion.

Alicia Wise, Director of Access and Policy, Elsevier

The RLUK passage cited below makes a rather odd leap and this is the assumption that had UK open access papers in hybrid titles not been published open access then they would instead have been published under the subscription model and funded through already-paid subscriptions.  This does not follow at all.  If open access is important to the author, then more likely they would instead have submitted to a fully open access journal and, oddly, in this case no one would bat an eyelash that the very same APC being charged to cover the cost of publication.  Even if the article were published under the subscription model, the RLUK passage does not acknowledge that subscription prices do in fact change to reflect decreases in the subscription content of journals (for example, see http://www.slideshare.net/aliciawise/price-adjustment-slide-43429236.)  Nor does the passage take into account that both the numbers of subscription articles and open access articles continue to grow.

Glenn Hampson, Executive Director of National Science Communication Institute (nSCI)

Hi Mike,

I admire your voice and think you have a lot to bring to this effort if you stay with it. But from the sound of your email, you’ve checked out. I hope you don’t. I’m flattered that you took so much time to talk about me (even if you did end up calling me deluded)—I really don’t deserve this kind of attention—but I categorically reject just about everything you said in this article. No offense—I enjoy your writing at least.

In a nutshell, think of it this way: Since the dawn of the OA movement, many of the most vocal advocates have portrayed publishers as public enemy number one. Added to this, most serious discussions about how to move improve the publishing system have been one-sided and exclusionary, and the feverish pitch that this issue has been drawn to leaves no room for compromise, no admission that just maybe there might be different ways to arrive at the same objectives. So we’re trying something new here. Forget what has been tried before if you can, and try this instead—try finding common ground and you’ll see how easy this gets.

But most of all, stay patient. As a teaser, I’ve been reading through the paper drafts and I think you’ll be astonished at some of the proposals that are being put forward by teams filled with publisher reps—I know I am. You may be conditioned for disappointment after all your years of heartfelt struggle of fighting for open access, but trust me—if we stick this through, you’re going to see some very innovative approaches get started this year (or at least soon), with the full and necessary participation of our publisher colleagues, that will blow the doors off of the OA progress we’ve seen to date.

Keep up the good fight.

Rebecca Kennison, Principal, K|N Consultants

Thanks, Alicia, for the slide showing the decrease in pricing of these 26 journals based on increased open content – or, as the slide puts it, decreased subscription content. I’ve also reread the argument you at Elsevier make for why you do not double-dip. If I understand your pricing process correctly, no matter how many authors in any given journal pay an APC, that amount rarely has any bearing on the overall pricing of the journal because the OA content pricing and the subscription pricing models are completely disconnected in your calculations. I don’t think I had fully understood that before.

Elsevier’s APC price points are based on “journal impact factor; the journal’s editorial and technical processes; competitive considerations; market conditions; other revenue streams associated with the journal.” In other words, APCs are priced to reflect what the market will bear, which may or may not having anything to do with actual cost, since the “journal’s editorial and technical processes” are only one factor in the overall pricing. Subscription pricing is also not based primarily on cost, but rather on “article volume; journal impact factor; journal usage; editorial processes; competitive considerations; and other revenue streams such as commercial contributions from advertising, reprints and supplements.” APC prices can be raised (or lowered) independently of subscriptions. Subscription prices can be raised (or lowered) independently of APCs. Because Elsevier’s pricing is not based solely or perhaps even primarily on editorial and production costs, any argument that they are double-dipping becomes moot. Double-dipping can only occur when first-copy costs are being paid for by both an APC and a subscription and the publisher is not offsetting or reducing subscription costs accordingly. If subscription pricing is not based on costs (first copy or otherwise), then there is nothing to offset by APCs.

That’s not what most of us think of when we think of hybrid journals. Instead, here’s how we think hybrid journals work. Let’s say we have Hybrid Journal A, which publishes 100 articles/year. Before it was a hybrid journal, it used to have a subscription price of $1000 and had 1000 subscribers, bringing in $1M per year. This year, 25% of the articles in this newly “hybridized” journal were published via APCs to the tune of $2500 each. Since for this particular journal subscription pricing is coupled with APC payments – as we commonly think is the case (or at least often hope is the case) for hybrid journals – that $62,500 in revenue now means for next year there will be a reduction to the subscription price of $62.50 per subscriber, making the journal subscription now $937.50. This sort of revenue balance via offsets only works for awhile, though. To maintain that $1M in revenue, the journal eventually needs either to increase its volume (to 400 articles per year at $2500/article) or increase the individual APC (to $10,000). The likelihood is that that $1M revenue cannot be maintained, but the hope (and, for some, the fervent dream) many OA advocates have is that the reduced costs of an all-OA world would result in reduced need for revenues at their current level. (In my example those cost reductions would need to be rather severe for the journal to be able to maintain a $2500 APC and stick with only publishing 100 articles per year. I’ll leave it to others to debate what and where cost cutting could happen to make up that $750,000 in lost revenue.) But this approach becomes moot when pricing is already not based on costs but rather on journal reputation and other competitive considerations. (Let me be clear in saying I’m not arguing for pricing journals like we do widgets, but merely observing that pricing based on value is considerably more difficult to evaluate than is pricing based on costs.)

As I now understand it, the above scenario is not how Elsevier works (although it does sound like on occasion Elsevier does sometimes reduce its pricing for journals with considerable APC-funded content, although when that happens it unclear). So I must say I actually agree with Alicia: Elsevier does not technically double-dip because it technically does not have hybrid journals. It has articles paid for by APCs and articles paid for by subscriptions. The idea of hybrid journal is as an entity that has two revenue sources – one from APCs and one from subscriptions. In the Elsevier model, pricing for both types of content happens on the article level, one paid for upfront (and individually) and the other at the backend (albeit as part of a bundle, whether a journal or a package). To rework my example above, it doesn’t matter how many articles in Journal A are paid for by APCs as far as subscription pricing is concerned. The $2500 APC goes to cover the article for which it has been paid, but the journal subscription can remain at $1000 because while the “subscription article volume” may have dropped by 25% the journal itself still has a high impact factor, strong journal usage (including usage from those OA articles), stringent editorial processes, and a robust reputation as the top journal in its field. The amount of OA content thus makes not a lick of difference in the pricing of the subscription – and if the journal is part of the Big Deal, it makes even less of a difference. As long as subscription revenues can be maintained, there is not likely to be any resulting decreases to those subscription at a large scale, no matter how much OA content may *also* be produced at the same time.

So what’s the takeaway? For me it’s what I’ve argued pretty vehemently for some time: If you’re going to pay an APC, then publish in a “pure” OA journal. If for whatever reason (journal reputation foremost of all) you choose to publish in a subscription journal with a hybrid model, don’t pay the APC. Instead, fight to keep your rights and then post the article anywhere you like. But the dream of flipping subscription journals to OA one APC at a time is probably just that – a dream – at least when it comes to Elsevier journals.

Alicia, please feel free to correct me if I’ve misrepresented Elsevier’s position in any of this …

Mike Taylor, Software Engineer, Index Data and Research Associate, University of Bristol

Thanks, Glenn, I appreciate your taking that post in the spirit that it was intended. I do feel bad for you, when you’re evidently very sincere is your perception the legacy publishers as colleagues, and you’re doing your best best to run this list as a collegial space where everyone is in it together. I can absolutely understand how frustrating it is for you to have me doing my best to bring that house of cards down. But I think it’s necessary. It’s better to face an unwelcome truth, however unpleasant the process, than to continue to live in its shadow.

You say “try finding common ground and you’ll see how easy this gets”. Again, I appreciate the sentiment. But to me, this is like being one of a herd of zebra that constantly loses members to a pride of lions. (Again: we don’t need to *blame* the lions for this; it’s just what they do.) I am saying “Er, have you noticed that the lions have been eating us for 15 years?” And you are saying “Try finding common ground with them” as though that will change the situation. You may be quite sure that the legacy publishers see us as a herd of zebra, and are delighted when we refuse to see them as what they are.

You say “from the sound of your email, you’ve checked out.” Not at all. This list is full of people representing groups whose interests are very strongly aligned: researchers, libraries, patient groups, all-OA publishers and so on. All of those groups can work together constructively and I deeply appreciate your having provided this space to help facilitate that. More: there are perfectly fruitful conversations still be had with legacy publishers, so long as we understand what the basis of those conversations is. (We can be quite sure that they do.) I plan to be around this list for as long as it’s running … and I hope that as you read that, you don’t regret having said you hope I don’t check out :-)

With all that said, I am very much looking forward to seeing the papers.

Scott Delman, Director of Group Publishing & DL Sales, Association for Computing Machinery

While I appreciate the repeated posts related to Elsevier’s policies and [Mikes’s] position that we publishers are all the “enemy” and are all undoubtedly driven by the same profit motives, it is simply not true. There are many publishers out there and many different approaches being experimented with to lump us all into a single box. The world is simply not as black and white as Glenn portrays it to be.

Here is ACM’s policy related to Hybrid Open Access: http://librarians.acm.org/policies

In short, ACM gives all of the APC revenue it receives from authors back to its institutional subscribers (universities). We have committed to do this as long as we do not realize a decline in our subscription base as a direct result of OA, which to date we have not. We return this revenue as direct credits on next year’s licensing fees. For 2016, each university subscriber received a $98 credit against their 2016 licensing fees. This credit is still small but is growing each year. This guarantees that we are not double dipping, is totally transparent, and was a condition of our scientific publications board before approving the roll out of our hybrid OA option for ACM authors back in 2013.

I’m am not suggesting that all publishers do as we are doing. Many have done similar things and many approach this in a different way, but it is simply not the case that we are all the same and are fundamentally at odds with the best interests of the scientific community.

Mike Taylor, Software Engineer, Index Data and Research Associate, University of Bristol

Scott, I think you misread some of the earlier messages. Glenn is certainly not taking the position that publishers are the enemy – he is the one trying to persuade *me* that we’re all fighting on the same team.

And you say “it is simply not the case that [publishers] are all the same and are fundamentally at odds with the best interests of the scientific community” – which I don’t think anyone disagrees with. That’s why in my blog-post I was careful to refer throughout to the “legacy publishers” – the for-profit corporations that have traditionally dominated the scholarly publication market and charged very high and ever-increasing fees.

Quite where the ACM falls on that scale, I’m not in a position to judge. But I think you’re reacting to an idea (“all publishers are bad”) that simply has not been proposed – least if all by Glenn, who’s trying to be the peacemaker here!

Keith Webster, Dean of Libraries, Carnegie Mellon University

I almost hesitate to enter the fray here, but I did want to echo both Scott’s and Mike’s views.  Disclosure: I’m a member of the ACM Publications Board.  

There is a distinction to be drawn between for-profit corporations whose legal duty, in many jurisdictions, is to operate at a profit for the benefit of their shareholders.  A good company (in the legal sense) will maximize its profit, and it is in the hands of customers and suppliers (libraries, authors, and reviewers) to change that model.  

OTOH, there are publishers like ACM where the aim is not to make a profit, but to operate a good publishing program that comes as close to possible as breaking even.  As Scott said, in ACM’s case, incremental revenue from APCs is returned to subscribers.

Having been away for the weekend, I haven’t been able to read through all of the messages on this thread, so apologies if this has been covered already, but I did want to raise an issue about flipping to OA.  I acknowledge that this is a hypothetical model: if the total subscriptions universe was flipped from reader (via library) pays to an author pays model, we would see a substantial shift in the burden of cost towards research intensive universities where most of the content, I suspect, is produced.  Scott would be able to signal the impact of this from an ACM perspective on my university, Carnegie Mellon, where our faculty produce more papers than any other university.  Whilst a public good argument can be formulated, I wonder if anyone has computed how overall cost burdens would shift at the level of the university, and the shape of shifts per university from reader to author payment.  Of course, if profit were excised from the system, and it operated at a cost recovery level, the total burden would be greatly reduced, although the distribution of costs would still weigh most on those institutions where content is generated.

Neil Jacobs, Head of Scholarly Communication Support, UK Joint Information Systems Committee (JISC)

Keith, all, in terms of cost modelling at university level, we did this back in 2012, which might be helpful: http://repository.jisc.ac.uk/610/

It models costs and savings for four different UK universities in an APC-based model, using various average APC levels.

It’s certainly not perfect, but it gives a recognisable picture.  I would be interested to know of more up-to-date estimates, eg from the Pay It Forward project?

Scott Delman, Director of Group Publishing & DL Sales, Association for Computing Machinery

Thanks for this. I don’t have exact figures in front of me and I can provide them if helpful, but in general here are some useful stats:

* ACM publishes ~40,000 authors annually

* ~75-80% come from the top 500 research institutions worldwide (of which CMU is in the top 10 in computer science)

* ~3,000 research institutions (academic, government, and corporate research) subscribe to ACM’s version of the “big deal”.

* Usage follows a similar trend to authorship (same institutions account for vast majority of usage)

* ACM Hybrid APCs vary from $700 to $1,700 depending upon conference or journal publication or member or non-member author

* a complete flip to OA would mean a roughly 30-40x increase in fees for CMU when comparing subscription-based fees today against APCs from CMU tomorrow. 

* in ACMs case approximately 500 institutions would be supporting ACMs publication program vs the current 3,000 figure.

Please do not read the above in any way as anti-OA. I am actually very pro-OA as anyone within my organization can attest to. These are just the facts as they exist today for one mid-sized society publisher.

Keith Webster, Dean of Libraries, Carnegie Mellon University

Thanks - these figures reflect my own assessment of our ACM activity.  Like you, I am, I hope, known for my advocacy for a change in the system.  It is, however, difficult to advocate for additional funding, internally, when the cost to us grows by an order of 30-40 times.  It’s more difficult to gauge the impact when looking at major for-profit publishers as total subs revenue, and even number of articles published, are not readily available.

Scott Delman, Director of Group Publishing & DL Sales, Association for Computing Machinery

Very true and I don’t think any rational person would advocate for a complete flip overnight, unless their intention were to simply tear the current system down and completely start over, which despite what some might say is not in the best interest of any stakeholder, least of all the scientists themselves. 


This article and its reviews are distributed under the terms of the Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and redistribution in any medium, provided that the original author and source are credited.