Follow the money – what really matters when choosing a journal (UPDATED)

I was asked to highlight the core link to data on journals so it is here at the top

As academics we make choices about journals all the time. We choose where to submit papers. who to review for, eventually who to edit for, who to subscribe to, who to recommend to libraries etc. As in all decisions both ethics and practicality weigh in. We both want to choose a journal that is “good” in the way it treats people and to choose a journal that will help our careers (after all that is a big part of why we do things involving journals). Different people will put different weights on good vs helping our career, and that is OK and not the topic of this blog piece to comment on further, but I don’t think anybody is 100% at one end or the other. What I want to do here is unpack the “good” part in some detail (so I am explicitly not talking about all the career-oriented reasons to chose a journal like if its readership will be interested in your paper, time to publication, acceptance rate, etc). Here I am exclusively focused on what makes a journal a “good” citizen of the research community and then you can (and should!) decide for yourself how to weight that in a manner appropriate to your career stage and goals. This is a complex story and I’ve put a lot of research into so I hope you’ll bear with me through this admittedly long post.

If you are attuned to the blogosphere/twitter-ville you will notice there has been an enormous amount of energy championing the idea that good=open access and emotions on all sides (e.g. comment section here and a strong reaction to the methods used for pushing OA here – HT Smallpond) (I have a brief dissection of different types of Open Access in a footnote, but I don’t want to digress here since it is not my main point). But I want to start from first principles. In the famous film “All the President’s Men” which is a loosely fictionalized account of the corruption in the administration of the US president Nixon, the phrase “follow the money” was used. I suspect, though, that this basic idea has been around since the Roman Empire (sequere pecuniam?), and probably the dawn of modern city-states with Sargon and Amenhotep. So where does following the money take us with academic journal publishing? For starters we know three basic things about the publishing industry (Figure 1):

• Revenues=expenses – this is a basic fact of all business. Unless you are a reserve bank you cannot create or destroy currency.
• Revenues for journals primarily come from four sources: library subscriptions, individual subscriptions, author charges and advertising. The first two represent payment by the reader (indirectly in the case of libraries). The third represents payment by the writer. The fourth represents payment by 3rd parties. For most journals library subscriptions are the dominant source of revenue (to the point that some publishers like Springer don’t even bother with individual subscriptions anymore), although the well-known open access journals (eg PLOS, BMC) have the effect of shifting to a 100% author pays, 0% reader pays model. Society journals (e.g. Ecology) have the most balanced mix across the four sources.
• Costs – include many factors including circulation costs (paper, printing, mailing, and now web pages), marketing and admin (including the pitiful stipends paid to editors in chief) but by far the two biggest are the labor involved in production (i.e. copy editing, typesetting, figure correction) and profits (in which I include money taken out of a journal to fund a society even thought this is not technically a profit). As an example Rosenzweig (editor of EER)  documents that the first category is less than 10% of total costs. Its hard to imagine that marketing and admin needs to be more than 10% either. Note that there is a giant subsidy by academics of contributing content and review for free (also not a topic for this post but see this if you’re interested: HT Meg). So we’re left with around 80% of the costs centered in labor-intensive production of individual articles and profits.

Figure 1 – Revenues have to balance expenses and the main factors on either side are shown (with font size) indicative of size in dollars.

I now want to pursue the revenue side a little bit. The best data I can find comes from American university libraries (which has an organization that surveys its members every year). The basics are shown in Figure 2:

Figure 2- trends in publishing. Data from American Research Librarians statistical summary 2004-2005 and 2010-2011 (no link – I actually had to use a physical library to look this up!)

Figure 2 shows several things. Start with books (in blue). The total $ spent by libraries on books and the cost per book (triangles and squares respectively) have tracked the rate of inflation closely (yellow) and as a result the number of books purchased has remained relatively constant (although the baseline year of 1986 represents a high point for libraries that we are only now returning to). Now look at journals. The cost per journal (red squares) skyrocketed from 1986-2000. To this point the cost per journal went up 250% while inflation only went up 50% – i.e. costs increased at a rate of 5x inflation. Total expenditures on journals (=revenues for publishers) did basically the same thing until 2000. In 2000 the $/journal plummeted while the total expenditures continued on basically a straight line increase.

You can probably guess the story here. Its when online publishing came home to roost. The publishers will tell you that eliminating print saved a lot of money driving down costs, but that is hooey – print and mailing is a small fraction of the total costs (see piece by EER editor Rosenzweig). The real story is that the publishers started “bundling” – libraries instead of picking their journals were sold bundles of 100s of journals they had not subscribed to before. This was essentially free to the publishers since the marginal cost of delivering an extra online journal to a library is zero. In principle libraries could still buy single subscriptions but the costs of single subscriptions were so expensive the bundles were the only practical option. You might say (and the publishers did say) “what a win for libraries” – more journals at less cost. But that statement is false – look at total expenditures on journals – it is still going up at 4-5x the rate of inflation. Even during the Great Recession of 2008-2010 when university and library budgets were decimated the publishers managed to keep revenue flat (the only such period of time in the 25 years of data). When you’re extracting increasingly more dollars at 5x the rate of inflation you are not helping libraries win! As evidence, the proportion of money libraries spend on books vs journals has dropped from 43% in 1986 to 23% in 2005. I very much doubt that reflects the librarians view that journals are now 3x (1/2 to 1/2 vs 3/4 to 1/4) as important as books. Bottom line – publisher revenues are going up at a rate of 4-5x the rate of inflation at a cost to other functions in the library and university.

What does the power of journal publishers to extract increasing amounts of money mean? It means journals are essentially monopolies. In the language of foraging theory, journals are non-substitutable. There is not a low-cost alternative to Nature (or to a specialized journal like Theoretical Ecology or Landscape Ecology). Or to be precise there are lower costs journals on these topics, but they don’t have the papers people want to read. In a monopoly world, the monpolists can extract much higher than inflation cost increases and this is just what has happened. And as predicted by economic theory and observed empirically, journal prices have gone up 4-5x the rate of inflation year after year.

Where has this extra revenue gone? Well the balance must hold – revenue equals the sum of all the expense centers. The main cost is labor for production which almost by definition cannot have gone up much faster than inflation (in the long run wages go up by inflation and improved productivity but not more or you’re in an inflation spiral). In fact this same period has seen a great deal of outsourcing to India and other countries and speaking from personal experience, the overall quality of production services like typesetting and proofing have gone down. So I suspect these costs have actually decreased. The other costs are too small to really matter and in any case have mostly gone up at the rate of inflation at most (paper and web programming are all highly competitive markets). That leaves one great big screaming slot for all the extra revenue to go – profits!

And indeed the profits of scholarly publishers are obscene. Since most academics are not experienced readers of the business pages, let me give a little background. On the whole 10% is considered a reasonable profit level. Low margin businesses like groceries (see Jeremy’s excellent summer post on his family roots in this business and lessons learned for science) or highly regulated businesses like utilities make a bit less (5-7%) and booming, fast moving businesses like computer technology make a bit more (15%). And businesses with monopolistic non-substitutable products (certain technologies or pharmaceuticals) make 20%. Now I’ve already tarred academic publishers as monopolists so you might be thinking they’re making 20%. You would be sorely underestimating. The large for-profit academic journal publishers are making 35-40%!!!! (as this well researched piece on the Sauropod vertebra picture of the week blog shows but I’ve seen the same numbers in other sources like this) This means greater than 1/3 of every dollar spent by a library, subscriber, or author on a for-profit journal goes in the pockets of shareholders with no direct benefit returned. (Economic theory says profits are a reward for risk and innovation but ask yourself how much risk and innovation you see in academic publishing from large for profit publishers?)

So what does this mean for you, the academic making decisions about which journals to read, write and volunteer for? Let me follow the money for two more paragraphs, and then I want to turn to some topics other than money. It appears a good estimate of the total cost (including overheads for management, etc, but NOT profits) of publishing an article is about $3,500-$4,000 ON AVERAGE. So one thing you might want to think about is in choosing journals is who is paying that $3,500-$4,000. Do you want to pick a journal that is 100% based on charging libraries (possibly with some advertising) like journals from Elsevier such as Ecology Letters TREE or journals from Springer like Landscape Ecology, or journals based 100% on charging the author like PLOS one/PLOS Biology (i.e. all gold-standard open access journals) or journals that have a mixed revenue stream like most society journals (e.g. Ecology) or journals that are mixed but have 0% from the reader (aka the hybrid model where you pay an open access fee to unlock your specific article but the journal still gets revenue from other sources such as library subscriptions as well – e.g. the open access option on Elsevier or Springer or Wiley journals) . Or 0% from the author (mostly for-profit publishers, definitely not open-access or society journals, although many society journals have waivers for students). You might have strong opinions about which of these is the “good”, ethical route to go. Personally I am pretty indifferent. The $3,500 is constant – it is just moving the burden around for who pays. The only thing that matters to me is that people in developing countries who can’t afford access get access but nearly all models address this. Beyond that, this is not ultimately an ethical issue, but a practical issue. Do I have a grant that has funds that let me go to a 100% author pay journal. Do I not have a grant or am a student and would benefit from a 0% author pay model? I know other people have hugely strong opinions, and I respect those, but I don’t (cue certain specific vociferous commenters telling me why I’m wrong in 3, 2, 1 …)

Something I do have a strong opinion about is total cost. Notice that I emphasized “on average” on that $3,500-$4,000 figure. There is a wide variation in cost per article (or slightly more accurately cost per page since article lengths vary across journals). Some publishers, especially not-for-profit and newer online-only publishers can get an article out for $500-$2000. Other publishers are less efficient (say $5,000) and then add a 40% profit on top ($7,000). There are even some extremes like Nature that appear closer to $30,000/article (but recall they are doing a giant paper print run of full color and are able to defray these costs with large circulation and advertising revenues). For a great resource on journal costs (and much else discussed here) see this website by Ted Bergstrom. The total cost matters much more to me than who pays it. It matters for both practical and ethical reasons. The ethical reasons are obvious, the cheaper something is the more available it is to the non-rich and the less resources diverted from other worthy goals like teaching and research. The practical is not that different – the cheaper it is, the more people that read it. So in the end, I care much more about the total cost than how that cost is divided up.

There are other things to care about too (which again largely align on the ethical and practical dimensions as they involve how many people can read it). These include:

• Self-archiving – in this day and age the single biggest factor in accessibility is whether Google Scholar can find a PDF that is not behind a firewall. Open access (specifically gold or hybrid forms*) are one way to achieve this. But a very functional alternative is what is known as self archiving (also called green open access* but I find this is not what most people are talking about when they just say an unqualified “open access”). Namely taking the PDF the publisher sends you and posting it on your website. That’s all it takes. And I would say right now when I use Google Scholar, I have to go through a pay wall about 50-60% of the time, and the remainder of the time I can get a PDF without the paywall with about 80-90% of this being thanks to self archiving (and <10% because of gold and hybrid open access). There is enormous variation in whether publisher contracts allow for self-archiving. Some flat out prohibit it, others allow it, and some are in between allowing you to post on a “preprint” (i.e. your Word file). A great resource is SHERPA/Romeo which tracks these policies for all journals. [ASIDE: And then there are authors who take matters into their own hands and post them on their websites whether the journal policies allow this or not (I’ve never heard of somebody being punished for this but for obvious reasons I’m not going to start naming such people here). I’m not in advocating this here in this public space, but if you wanted to take an aggressive, protest, slightly “risky” stance this makes more sense to me personally than refusing to review, etc.]
• Value-added – as mentioned the main cost is the expert labor involved in typesetting, proof-reading, copy editing, figure clean-up, etc. Different publishers have squeezed this function to different degrees. In my experience this function ranges from little more than turning your submitted word document into a PDF all the way to still adding significant value. Several publishers have sent this function offshore which is clearly driven by cost-savings and not quality improvements. At the other end Science and Nature (and PNAS) do very substantial editing and enormously improve figures. Among the smaller journals, the three best value-added experiences I have had in recent years were at American Journal of Botany, American Naturalist and Evolutionary Ecology Research. Those experiences contrasted so strongly with my typical experiences with big publishers of most journals, that it made me realize just what I was missing. The editors greatly improved flow and clarity (not just typos) and actually caught instead of introduced errors in equations.
• Availability of older issues – Have you ever tried to pull a classic article from 1970 or 1940? Unless you are at a rich university, your experience probably varied drastically depending on which journal it was published in. At one end are the big publishers who, having captured libraries into the bundle, are now shrinking the bundle by only including articles from say 1980 to present, and then making libraries pay extra for the older issues. I went looking for Volterra’s classic 1926 predator prey paper in Nature but my library didn’t have online access. I wrote a somewhat snippy note to the library complaining and got a nicer than I deserved note pointing out that getting full back issues of Nature (1869-1996) cost $84,000/year! (for my medium sized, medium wealth university and you better believe they charge bigger libraries more than that).That’s a lot of classroom microscopes! (and I was politely reminded that old-fashioned though it might be it might be cost effective to please walk over and use the print copy still on the shelves – what a concept!). Meanwhile if you want to look at most society journals, including Science, the publishers have given rights to these back issues to JStor a not-for-profit archiving service that delivers back issues to libraries at cost and to the developing world for free. Big difference!
• Open data standards – an increasingly important issue in ecology is whether journals require the data used to be published and made available (usually as supplemental material or in a database). This has been standard practice with genetic sequencing and GenBank for years, and is slowly coming to ecology but there are still big differences in policies and enforcement of policies. (and I still know people who prefer to publish and people who prefer NOT to publish in journals requiring open data).

So, I have argued there are many dimensions of trade-off to think about when choosing a journal that is a “good citizen”: who pays, how much has to be paid, and other features like self-archiving, value-added, and etc (for example). To those of you who have read my posts on statistical machismo this attitude of mine is probably not surprising. Aggressive pushing of a one-dimensional good/bad axis doesn’t float my boat. Acknowledging the complexity and the ensuing responsibility to be informed and think (and recognize others may come out at a different point for valid reasons) does float my boat. Choosing journals is no different than choosing statistics in this regard.

So in the real world where you have finite time and can barely manage to get a manuscript written and have no time to spend 5 hours researching journal policies and prices, what should you do? I have two suggestions. First, I have assembled a table of data on many of these issues for many journals of ecology (available here as a google document) which I offer as a service for quick and dirty research. I hope this lives on as a resource for people to use. If people want to research journals not on the list or send me data on new columns (example I don’t have a column on open data policies) I would be glad to add them. You can use this as a quick tool to see where journals rank on multiple dimensions.

If you’re really tight for time such that you only have time to think about one axis, I suggest you think about the “follow the money” axis. This runs from society journals published by the societies or by non-profit (e.g. university press) publishers, through society journals published by for-profit publishers to at the other end purely for profit journals published by for-profit publishers and without society affiliations. These three categories are spelled out in the aforementioned spreadsheet. But it is often easy to guess – Ecological Society and university press journals are non-profit, Wiley-Blackwell usually equals society sponsored but for profit published, while Elsevier and Springer=for profit (and yes that is how much the industry is consolidated that there are only two big publishers producing most of these journals). I want to make three points with regard to my “follow-the-money” handy dandy single axis dimension for the lazy or time-pressed journal chooser:

• There are a few that fall through the cracks of my big groups. Allen Press is I think technically a for-profit publisher but they were founded by university people to print academic journals for societies and they are small, service-oriented and reasonable cost (but currently only publish in fields outside of ecology at the present time to my knowledge). In the book world, Sinauer (and several others) clearly fall in this category too. Allen Press and Sinauer are both “good” publishers in my book (and many other academics I know). Big-for-profit and small-for-profit is an important distinction too (see here for a quantitative argument of how big for-profits are not representative of the whole industry or even most for profits), but it for the most part doesn’t play out much in ecology journals which has only the big (indeed mega-big) for-profit players.
• Nearly all of the features I identified as important correlate with this one dimension. As clearly demonstrated in a nice article by Bergstrom and Bergstrom 2006, cost per journal article strongly breaks out by these three groups, as does cost increases over time. As a quick glance at my spreadsheet will show you, so do for the most part self-archiving friendly policies, JStor friendliness, and although I don’t know how to measure it, my experience strongly tells me that ironically the not-for-profit cheaper presses provide the best value added in editing. Again my aforementioned list of best editing experiences are at American Journal of Botany (society published), American Naturalist (university press) and Evolutionary Ecology Research (small not-for-profit), all among the cheapest journals around. Its amazing what lack of profit motive can do for you.
• The one thing that is NOT correlated with this dimension is who pays (i.e. the conventional interpretation of open access). There are expensive for profit gold-standard open access publishers (OMICS and Hindawi presses come to mind with Hindawi speculated to pull down a 50% profit margin and there is a growing list of predatory gold OA journals by a variety of publishers) and there are very cheap, service-oriented not-for-profits (like Ecology) that have no open-access option at all (not even as a for fee add-on – although I quickly note that ESA has launched Ecospheres as a gold-standard open access journal before Don Strong gets mad at me). If you care about who pays you need to add this as a second axis as it is essentially orthogonal to the “follow-the-money” axis.

I have no intention of being prescriptive to others. But just to give an example, I’ll describe my own philosophy. I care almost exclusively about the follow-the-money axis and its correlates with a strong bias towards the not-for-profit end. That said I am not absolutist.  My ultimate goal is to get my research widely read, and if somebody makes money off of it, I can live with that. But, all else being equal, I choose the not-for-profit end. This ends up looking like a mixture with a bias towards not-for-profit.  I have coauthors who will only publish in or review for not-for-profits, but that is not me. And if I have no grants the who-pays axis  (not me end) becomes really important but otherwise, that is one dimension too many to try to optimize on over and above just trying to get my work accepted!  Anybody who wasted the time to look at my CV to see where I publish or review would see that it is a broad mix including many for profit journals but that I broadly aim for non-profits leading to a biased sample. As one example, until recently I was an associate editor for two non-profit and one for-profit journal (i.e. a mix but biased to non-profit)l. As another example,  when I went looking for a book contract I went to three not-for-profit university presses and one small for-profit high service press – Sinauer – and if I had time I would have added Wiley as my 5th – again a biased mix is where I ended up (book publishers are a bit different – see Figure 1 – but still take a look at book prices vs publisher type sometime walking through the exhibitors at a conference). Final example:some decisions are easy. Say I have a nice paper on seed size vs number or any other evolutionary ecology topic. If its not quite good enough for a general ecology journal. I have two choices. One (Evolutionary Ecology Research) is not-for-profit, cheap, etc. The other (Evolutionary Ecology) is for profit, very expensive, etc. Granted the 2nd one has a bit higher impact factor but when you remember impact factors are mostly driven by a few highly cited papers in each journal amidst an enormous variability in citation rates across papers, the odds that choosing one over the other will influence my readership or citation rates is small. This to me is an easy choice. Especially if you know the history of these journals and the fact that the first was a leading rebel in taking on for-profit obscene cost increases. Here ends boring you with my own thought processes.

So, what should you do with all this? That is up to you. That is my point. But be informed. Realize that open-access is at its core (until publishing papers is free) a debate about who pays, not how much gets paid and who profits. And follow the money!

---

* Technically open access means anybody can access the PDF (i.e fully proofed and fancy formatted copy) legally without paying. In practice there are three kinds of open access (OA):

1. Gold OA – author pays 100% of the costs and article is available to anybody immediatley
2. Hybrid OA – journal as a whole is not OA (i.e. is behind a paywall), and journal revenues derive broadly from library subscription fees, but the author can pay an additional amount (very often around $3,000) to make their one specific article open access (not behind the pay wall and freely available)
3. Green OA – author has right to post article on their own website making it effectively OA with a little work. I called this self-archiving above.
4. There are many other permutations as well, many being driven by government funders like NIH and NERC requiring open access where open access is given after say 6 months (non-profit PNAS) or 1 year (some for profit journals, usually only if required by the author’s research funder).

Although all of these can be accurately called open-access, I find that most people when they just say open access mean #1/#2 (gold and hybrid) which are 100% author pays (gold) or library pays and then author pays even more (hybrid – I’m not a fan of this!). So for the most part when I say OA here I am following this convention and meaning gold/hybrid. I discuss green OA separately as self-archiving.

I think it is also fair to say that OA has sort of become a muddled umbrella that started with good intentions but now includes predatory publishers making 50%+ profits and excludes some really worthy publishers. Any debate of OA really needs to start with a clarification of terms. Some people who say they are “for OA” really mean “for non-profit” when you drill into details. I hope this piece can help clarify the discussion.

PS – Thanks to Jim Bird, science librarian and Deb Rollins Head of Collections Services at UMaine for help tracking down journal price trend data and to all the UMaine library staff for doing what I think is a great job navigating this mercenary world on all us users’ behalf.