Charles Darwin photograph
Charles Darwin. Photograph by Herbert Rose Barraud, thought by Darwin photograph scholar Gene Kritsky to be the last photograph taken of Darwin before his death in 1882, via Wikipedia

Open Access Millennialism is the belief that the world of scholarly publishing has a purpose and is moving toward the fulfillment of that purpose: at some point (but when? when?) all scholarly material will be open access (OA), and it is only the foot-dragging of self-interested publishers and the innate conservatism of academics that is holding it back. The signs are everywhere (“‘We would see a sign!’”): in the growing number of OA services, the mandates for OA from funders, and the general utility of locating content with no more than a Web browser and a mouse-click. I would have thought that the idea that history has a purpose was tossed out the window in 1859, but you can have OA without a sense of historical destiny. Just bide your time and let the utopians party. When you wake up the next morning, as I’ve said before, it’s just business.

We do have a number of divergent figures, however, that are likely to be reconciled in some way in the years ahead; these are in fact the two metrics that I personally keep an eye on. The first is that the average income to the publisher per article is about $5,000. You derive this number with great mathematical sophistication, by dividing the number of articles published each year (about 2 million) into the size of the journals market (about $10 billion). Voila! We now have a benchmark against we can measure OA pricing and services.

The reason that $5,000 figure (an average, I repeat) has captured my attention is that PLOS ONE charges $1,350 and Peer J even less (though if you can figure out the Peer J business model from the public utterances, you will indeed have to engage in mathematical analysis far more complicated than long division). The $3,650 difference represents an imbalance that may, through entropy if not as a matter of destiny, force down the size of the journals market. This is the argument for subtractive change, about which more in a minute.

The other metric that I keep my eye on is another awkward relationship, the relative size of the OA market measured in dollars (about 2.3% of the total) and the number of articles available as OA, which is variously reported as in the range of 15-20%. (I believe the percentage is much higher, but much of Green OA is hidden in institutional repositories that do not see the light of day or the tender mercies of SEO.) This is another argument for the subtractive scenario: the number of scholarly articles continues to grow (by 3-4% a year) as does the proportion of those articles that are OA, but the money received for these articles is much less than that for traditionally organized publications. Suppose the entire journals market, measured in dollars, were only 3-4% of what it is today? Or even 30%?  What then? What would happen to the established publishers as they tried to rethink their cost structure to align with the emerging market realities?

A digression, if I may, into the world of books, where I keep my eyes peeled for trends in library circulation. It’s often said that 40% of books in libraries never circulate, a figure that, at best, is imprecise, but it points to another gap, that between the books in a library’s collection and the actual demand generated by patrons. Many libraries are seeking to align collections with patron demand, which can’t be good news for book publishers. What we see here are publishers benefiting from inefficiencies in the publishing ecosystem, whether for book circulation or OA growth and pricing. And here is the subtractive argument once again: as the inefficiencies are worked out of the system, the total cost of scholarly communications will drop.

So the subtractive scenario is that the total market will eventually shrink because of the economic disruption brought about by OA. The additive scenario argues that the market will be bigger, that is, revenue from OA will be additive to the revenue from traditional journals. And the substitutive scenario has the market remaining at essentially what it is today, adjusted for inflation, but with the mix skewing toward OA.

I made the case for the additive scenario in 2004, and I was right: the market is significantly bigger today. But how about ten years from now? And after ten more years have elapsed, will we still be asking what will be happening in ten more?  This begins to sound like the Peak Oil argument, which is utterly convincing, but it is wrong. Peak Oil is not a function of how much oil there is in the ground but of how much human imagination you can bring to that oil. The market for scholarly communications is likely to have similar characteristics.  As OA grows (as I think it will), its downward pricing pressure will be more than offset by the overall increase in the amount of research and the ability of publishers to tease new value out of that research.

Let’s test the proposition. A show of hands, please, among the librarians reading this. How many of you have gone to your provost to say that you want to reduce your materials budget by 30% because the widespread availability of OA makes your current level of expenditure unnecessary? How many expect to have this conversation in three years? In five?  Yes, let’s now ask the big question: What about ten years, don’t you think OA will have made your traditional collections budget irrelevant by then?

The votes are in and we see that not only do librarians expect to spend more money in three, five, and ten years, but the possibility that anyone would even consider sharp cuts to library budgets is something of a heresy. Librarians, in other words, are in an unholy embrace with the publishers they despise. If one goes, the other goes with it, and thus they are likely to continue to support each other. This does not have to be a friendly relationship; it just has to be an economically persistent one.

The problem with the subtractionist argument is that it is essentially a case of vulgar economics. It assumes that there is a law of supply and demand, that markets are efficient, that lower costs drive purchases. But of course none of this is true in institutional markets. The academy works via an economics of reputation, and the participants in this economy naturally look out for their own interests. We should not look for a $5,000-per-article publisher to embrace $1,350 without a whimper, and we should not expect a librarian to cheerfully make a case for fewer staff, smaller budgets, and a tiny building. The market is additive because it is made up of people.

For a publisher, which position you take on this matter is significant. If you believe in substitution, then your strategy is to replace toll-access revenue with OA revenue. You assume that you don’t have to resize your organization (though a lower cost structure is always desirable) and you spend your time on transition plans. On the other hand, if you believe that the OA Millennialists are right, you take out an axe; you have to get that cost structure down to where you can make money in a much smaller market. You look for a leaner operation and you strive to replace people with machines wherever possible.

The additive camp, though, begins with a different premise. This camp has as its bedrock foundation the belief that people make the difference. While reducing costs is part of operating every business, for the additive party it is not the strategic priority. To put this differently, it’s not the CEO’s job; it’s something to be delegated to someone further down in the organization.  For the additive party the emphasis is on the new: new products, new services, new markets. The company that has just cut 20% or more of its staff is at a disadvantage in this scenario. The machines that have replaced them are very good at what they do, but thinking up new products is not among them.

Having won my bet in 2004–ten years later and the market is much bigger–I could take my money off the table and go home, basking in self-satisfaction, but I am going to let the money ride. This is because the future belongs to the creative. Let’s check back in ten years.  For some of us, our optimism may exceed our life span.

Joseph Esposito

Joseph Esposito

Joe Esposito is a management consultant for the publishing and digital services industries. Joe focuses on organizational strategy and new business development. He is active in both the for-profit and not-for-profit areas.

View All Posts by Joseph Esposito


23 Thoughts on "Additive, Substitutive, Subtractive"

“The other metric that I keep my eye on is another awkward relationship, the relative size of the OA market measured in dollars (about 2.3% of the total) and the number of articles available as OA, which is variously reported as in the range of 15-20%.”

Joe, is that 2-3% which is gold but the 15-20% includes gold and green? If so, then the green money has been paid for in the traditional sense (ie licensing/subscriptions). So the gap might not be so large between what is actually paid for OA (Gold plus a percentage of subs) and the proportion of OA articles in total.

Indeed, for example PubMed Central posts most of its articles as 12 month green OA, based on NIH funding, which is roughly half of the $60 billion annual US basic research budget. That is a lot of articles, maybe 80,000 a year, but almost all are subscription based. It is not clear that this kind of OA has much to do with markets, unless it grows to the point where it supports cancellations. Then we might see a big forced shift to APC OA.

Perhaps we’re veering off into semantics, and the usual arguments about the definition of “open access”. Under the BOAI definition, most of those Green articles (including the majority of NIH funded research in PMC) don’t qualify.

Given that we’re talking about economics and the market for a given year, and that most Green articles become available after a 12 month or so embargo, would it be more effective to think in terms of “the market measured in dollars for this year’s articles” rather than looking at what will eventually happen to those articles at some point down the line?

Great post Joe and I’m sure that you’re right that there is enough economic incentive for all concerned in the current system of subscription based publishing to maintain it for many years to come.

One caveat that I might suggest though is that the role of the library is broadening over time. It’s changing slowly but there is a growing movement towards librarians acting not only to bring content into the institution but also to document and manage the scholarly output from the institution.

Some examples include the growing use of institutional repositories, particularly when they are tied into library publishing efforts, and the steady trend for University Presses, particularly in the States, to either have stronger ties with the library or be bought under control of the library. There’s also a growing role for libraries in working with the office of research management to collate and help analyze research output to inform research portfolio decisions and maximize funding for the institution.

While I think that you’re essentially correct in your assessment that many collection managers may not be in a hurry to see their job functions become redundant, I think that as a whole, librarians are thinking carefully about what their role will evolve into as the economic power shifts (even if only partially) from the readership side to the authorship side.

One variation here is the increasing effort to set up OA monograph publishers on the basis of endowments, as at the new Amherst College Press, which relieves the press of significant market pressure and guarantees it the ability to endure indefinitely with a publishing output defined by the size of the endowment.

” I think that as a whole, librarians are thinking carefully about what their role will evolve into as the economic power shifts (even if only partially) from the readership side to the authorship side.”

Phill, I would only amend the above statement to read, “I think that some librarians are thinking carefully . . . .” Many are simply denying that these changes are taking place or that they will become pressing before said librarian reaches retirement age–which isn’t far off for much of the profession.

What would make your argument more persuasive is a link (of which you have many) to someone credible in the industry arguing that we will be spending less in the future. I’m sure this argument has been made — I’d just like to read the context myself.

Also, have expenditures increased by more than the 2-3% of revenue accounted for by OA? If so, then it sounds like expenditures aren’t tied to OA, so I’m not quite sure how this all ties together.

The revenue percentage and the article percentage are from different sources, but I believe both refer to Gold OA, not OA in total. My own observational research for 2013 shows a revenue figure that comes out right around that 2.3% for publication of some 403,000 articles, which should be right around 20-22% of all scholarly articles in 2013. In both cases, that’s for immediate (no-embargo) gold OA. Total availability of *current* articles, including embargoed green OA, should be much higher. A link to my research:

Thanks Walt, very interesting. Are these gold OA articles all APC or are a lot of them in subsidized journals that charge neither an author charge or a subscription fee? It is my understanding that there are a lot of journals that charge neither, but they are not really part of the market.

I’ll try again (my response seems to have disappeared): The Gold OA articles in journals that don’t charge fees are *certainly* part of the scholarly article marketplace (but the dollars come from elsewhere), and a majority of Gold OA journals in DOAJ don’t charge fees. (The free journals include roughly three-quarters of those outside of biology, biomed and medical fields.)

This is an interesting issue because while subsidized no fee journals play into some aspects of the publishing business, they may be irrelevant for other analyses. I can’t speak for Joe but he seems to be talking about the sum of subscription plus APC revenue, going up or going down. No fee journals may not be part of this equation. Likewise I think the $10 billion estimate does not include subsidies. In fact I am curious what the total of subsidy revenues looks like?

The reason for this is that while there are some active schemes for converting the subscription and APC journals to a subsidy basis, the basic business issue today seems to be subscription versus APC. I think very few subscription publishers are looking to convert to a subsidy model.

Walt, to your knowledge has anyone counted the number of OA articles (as distinct from the number of OA journals) that are published on an author-pays basis each year? It would be interesting to see whether the article percentages are different from the journal-title percentages.

My understanding is that the count of OA publications is by article, not by journal (is this what you were asking?). That is, in the past year something in the area of 400,000 articles were published that were OA. I must say it is odd that for an area that has attracted so much energy, there is very little in the way of hard data.

Actually, I was responding to Walt’s observation that “a majority of Gold OA journals in the DOAJ don’t charge fees.” That’s objectively true: search the DOAJ and you find that about 65% of the listed Gold journals charge no author fees. What I wonder is whether 65% of the _articles_ published by Gold OA journals listed in DOAJ are published without author fees. In other words, do the author-fee journals publish proportionally more (or fewer, or the same number of) articles as institutionally-funded Gold OA journals? I have no expectation as to what that number probably is, but I think it would be interesting to know.

Long answer appears below. Very short answer: APC-charging gold OA journals (especially in medicine) publish proportionally more articles than no-fee gold OA journals.

(Responding to Joe’s comment as well): My numbers were for articles, but not explicitly those in journals with APCs (“author-pays” involves an implication that’s probably false for many/most medical articles, for example). As for breakdown of free vs. fee, I have the data for 2011-2013 (for most English-language gold OA journals), but it’s spread across three articles, one not yet published. Joe’s final comment is a good one–but getting that hard data is not easy (we don’t have truly good data for the subscription sector and # of articles, do we?). A crude number, including lots of “journals” that have never published anything: for 2013, 4,800-odd “journals” with APCs published around 274,000 articles (but I’d consider only 2,200 of those journals to be real and worth tracking); some 3,200+ journals funded through other means (including 2,900 real and worth tracking) published around 100,000 articles; and 600-odd “unknowns”–journals that probably have APCs but don’t reveal them–published around 29,000 articles. But those are crude numbers for one year. The majority of real gold OA journals with reasonably transparent practices don’t charge fees, but the majority of articles are from the minority that *do* charge fees. I may be doing a more coherent combined writeup, but I’m not there yet, since this all started by examining the reality of Beall’s lists (which is: lots of smoke, relatively little fire–most of the “journals” don’t exist) and grew from there.

Short answer (hah!): Yes, the article percentages are different from the journal percentages–and PLOS One alone is almost enough to guarantee that. A considerably higher percentage of articles than journals are in the APC/fee category. Given the sheer bulk of bio/medicine journals and articles (including PLOS One), that is, around 177,000 articles in 2013 from more than 2,870 journals and “journals,” that’s scarcely surprising–take those away, and the percentages are very different.

the reality of Beall’s lists (which is: lots of smoke, relatively little fire–most of the “journals” don’t exist)

Not to digress here, but isn’t the nonexistence of many of these journals the very point that Beall is making? In other words, if by “don’t exist” you mean “if an author contacts them and tries to pay for publication, his email will bounce back unanswered,” then you’re right that their inclusion on the list constitutes smoke without fire. But if by “don’t exist” you mean “they are scam entities that take authors’ money and then never publish anything,” then that’s not smoke — it’s fire. (And, of course, it’s also possible that some of these nonexistent journals disappeared, like cockroaches, as soon as Beall shined a light on them, in which case their inclusion on today’s version of the list isn’t a matter of “smoke without fire” but a matter of the list needing to be updated.)

For the record, I have no idea what the answer is to that question.

Following the digression: I have no idea what the answer is either (and it may not be answerable), but so far I haven’t heard of the latter happening–that is, entities taking money and not publishing at all. That would be straightforward fraud. Beall’s claim is that these are entities who don’t carry out proper peer review and publish whatever gets paid for (and, let’s face it, publishing using OJS with author-provided PDFs is not a big deal)–thus, the vast majority never publishing anything is significant. Most of these are “journals”–entities that exist only as web pages, most of those created using templates (and using them badly). Only the most venal of authors would submit an article to such “journals”; they’re only preying on willing victims.

I really doubt that 403,000 articles is 20-22% of all scholarly articles. It depends what your universe of ‘all scholarly articles’ is. Nobody knows the total number of articles published each year, as there is no database that contains them all.

On the questions above of the proportion of articles each year that were published in open-access or hybrid journals, there has been done work done using Elsevier’s Scopus database (which does not include all scholarly articles, but is a start). I looked that up again on reading this thread, so thought I’d summarise here. I’d like to know what more there is, if anyone can add information. It’s just the backbone to Joe’s ’15-20%’ estimate, about which there’s been some questioning in the comment thread.

Bjork et al, BMC Medicine 2012, 10:124 found that 12% of articles in Scopus in 2011 had been published in immediate open-access journals – getting on for 200,000 articles. (That did not count free articles published in ‘hybrid’ journals, though this was estimated at 0.7% in that year). Interestingly, Bjork et al’s separate count from the Directory of Open Access Journals (DOAJ) gave 340,000 articles published in open-access journals in 2011, which shows you how selective a slice the Scopus database cuts through the market for scholarly open-access journals. (And reminds us, again, that we don’t know how many articles are published each year).

A separate analysis by Science Metrix (again using Scopus) tallied with the Bjork et al figure, and said the rise had been 4% in 2004 to 12% in 2011 – i.e. approximately an incursion of 1% each year, within a collection of articles that is itself growing each year. The latest analysis by Science Metrix put the same figure at 13% in 2012.

So the trend for articles published in open-access journals in ‘the universe of articles in the Scopus database’ seems clear. Approximately, a 1% incursion each year, and we’re probably at 14% last year and 15% this year. Or perhaps slightly more if the trend has accelerated. And note that many of the open-access articles published in DOAJ are not indexed in Scopus.

This does not include the growth in articles that were immediately free to readers, but published in subscription journals (i.e. hybrid). Getting hard figures on the growth of that side of things would be useful. Bjork et al, 2012, as noted, counted hybrid OA as an additional 0.7%, pushing the Scopus immediate OA proportion up to close to 13% in that year. I missed later estimates, if there were any. Perhaps we are at 2-3% hybrid, by now.

So Joe’s ’15-20%’ for immediate OA + hybrid OA would fit these estimates for the market and wouldn’t include any articles made free by other means (self-archiving, etc.)

It’s hard to measure additional papers made free to read, outside of those papers that were part of the open access market. But Science Metrix’s sampling of Scopus suggests the availability of free papers is getting at or above 50% in recent years, though I don’t think one can say much more about the mechanism by which these papers became free to read.

So, all in all, papers published in the open access market – that is, made immediately free in hybrid or open access journals – come to around 15-20% in the Scopus index. Any self-archiving or publishers’ archiving of papers in repositories probably more than doubles the availability of Scopus’s free-to-read papers.

And while you can get figures on the number of papers in open-access journals by sampling from DOAJ, I don’t see how you can make conclusions from that about the proportion of all papers that are published in open-access journals.

I hope this helps, given some of the questions above. I also hope it helps that I didn’t mention ‘green’ or ‘gold’ at all 🙂 I’m keeping track of this so I hope Joe or others can add any other quantitative research.

I should note that my numbers aren’t “sampling”; they’re based on visits to each of 14,000 or so journal sites. Journals are so wildly varied in number of articles that sampling isn’t very workable. On the other hand, my survey is limited to journals with English as the first language code, and excludes journals where I couldn’t sensibly count the articles and those I don’t consider open access (including hybrids and ones requiring registration). I agree that we don’t really know how many articles are published in all, but I believe “roughly 20% are Gold OA” is reasonably supported by the evidence–counts, not sampling. The growth from 2011 to 2013 was quite rapid: for the very large subset I studied, it went from roughly 247,000 in 2011 to roughly 403,000 in 2013. Note again that these are all Gold OA (with or without fees) and all immediately available in final published form.

What is not clear to me here includes:
a) At what pace is the number of articles increasing and how does this impact?

b) Is there a difference when one considers the various international sources for the articles and whether the respective funding sources, internationally following the US funders wrt journal article payments to print?

c) what is the increase in the number of journals being published and the number of articles available through or to these journals.?


Doesn’t what is going to “force down the size of the journals market” also include a growing realization that commercial publishers are unfairly charging exorbitant subscription rates?
For example (2012 data)… J. Comparative Neurology $30,860 for 232 articles vs J. Neurophysiology $1600 for 598 articles. The difference is attributable to the J. Neurophysiol. charging a $50 submission fee and $75/page charge. This seeming ‘relic’ from the past (i.e. reasonable page & subscription charges) would seem to be a much better solution than gold OA.

The open access mantra is: good for readers, bad for authors!

One should be ‘mad’ to pay $5000 when he could pay $1000!
One should be ‘crazy’ to pay $1000 when he could pay $100!
One should be ‘crazy’ to pay $100 when he could publish for ‘free’.

Open access is good for poor readers and rich authors but not for poor authors.
Because of excessive fees and commercial goals behinds, the Open Access would look line as a kitchen’s chef recipe:
“Eat but don’t cook”!
Eat (read) but don’t cook (do not write and submit to open access, particularly if you are poor!)

Science is getting sick with all the metrics and ‘prestigious’ notions… One should be ‘stupid’ to think that an article published in journal X should be better than article published in journal Y just because of a name or high publication fees!
The most important in an article is the ‘idea’, the content, not the journal, not the ‘prestige’, not the open access price, not, not, not…!
As readers, we can find here, in this ‘scholarly kitchen’ many valuable articles, much more valuable than many other articles published in so-called high “golden-journals”!
The key point in science in general is to find the right information whatever the journal, whatever the website, whatever the author, whatever the name, whatever the gender, whatever, whatever….etc.!

Knowledge should be free funded by the tax payer for readers and for AUTHORS, too.
Are tax payers only readers?

Merely one year ago, Science magazine was criticizing the open access and its peer-review caveats:

But paradoxically, less than a year later, Science has announced its new open business model:

Just the sum of contradiction!

I know a guy whose article was clearly rejected because he submitted it to a free section in an open access journal because that section was not generating money! I saw the email sent by the editorial office to the handling editor advising him to reject!
Scandalous, unbelievable but true.
So, the open access is just an open business nothing else!

Comments are closed.