In the midst of debates about the justice and propriety of reader-funded research publication (aka, the subscription model) versus author-funded research publication (aka, open access [OA]), we might have lost sight of a very important set of fundamental dynamics — namely, the affordability in each model for each participant in the resulting economy.

Now, to address the first complaint some may find with the preceding paragraph, let me state that I believe that OA is a business model. Even Green OA is a business model, because, like it or not, “free” is a price, just as zero is a number. In an economy, the price “free” distorts markets just as much as, and perhaps more than, any other price. But Green OA can only exist as an exception price in an economy where the major production aspects are not free, which is why Gold OA has come on strong recently. Gold OA provides pricing that makes sense to pursue and compete over, and can support viable production systems and growth.

The conversation has matured accordingly, with nearly everyone in the scholarly publishing world acknowledging that publishing costs money, no matter the form. Claims that some high percentage of OA publications (some say 70%) charge nothing falter under close examination, as you can see other business models — sponsorship, institutional subsidy, institutional subscription/membership, or nascent subscriptions — at work under the blanket of OA that’s been thrown over them. Even if you were to accept the premise, many of these entrants are too new to have demonstrated they are sustainable or poised to succeed.

Besides, it’s actually good for OA to have a business model. It makes the main premise — that readers can have free access to scientific publications — much more likely to succeed.

But there is a problem with the OA business model which needs some serious attention.

As currently conceived, the payment system for Gold OA centralizes payments around far fewer sources than the subscription model. This leads to higher prices in fewer places, increases the risk to publishers as these payments consolidate, and puts more pressure on fewer points of failure.

Last year, we saw two interesting results of this.

First, at a publishing conference, a development officer with a large and previously OA-supportive philanthropy stated clearly that they no longer felt paying APCs was a good use of their funds. After all, most papers became free after 12 months, and the papers were reaching a scientifically powerless audience (i.e., people who couldn’t comprehend or take actions based on the information contained in the papers). These philanthropists were spending more than US$1 million on APCs, and were going to stop. Instead, they’d put it back into funding research. The danger of centralization is clear — a few people at one organization can change their minds, and US$1 million is taken out of the Gold OA market.

Second, many prominent OA journals published fewer articles in 2014 than in 2013. This suggests a couple of possible scenarios — either an overall slowdown in the support for OA publication, or a market that has increased the supply of OA outlets beyond the demand for them. In either case, it suggests that OA may have peaked. Perhaps it’s a blip, but it’s a troubling blip. Maybe other payers have changed their minds, too.

We also have seen plenty of cautionary signs. In the US, government funding of research has fallen 16% in constant dollars over just the past few years. In Europe, the Horizons 2020 research budget is being raided to fund other societal needs. In the UK, projections are that a fully OA future would be more expensive than subscriptions, and the costs of administering OA appear massive.

Taken one at a time, the consequences are sobering.

For instance, if funding for research were to decrease in a completely Gold OA environment, the funding for publication drops to some degree, and probably a higher degree than overall funding, as publication is a lower-priority function in a research economy. In fact, it’s entirely conceivable that funding agencies would pull back completely to preserve funding for research grants. Let’s assume the drop in available Gold OA funds is between 16% and 100% over a three-year period. What happens? Do publishers simply “suck it up” and deliver despite a gutting of their revenues? Or do they fail left and right, or leave the market before the market demands liquidation? Or does Gold OA fail, replaced by a rapid retrenchment in the subscription model?

And Gold OA is not the only form of OA that threatens research budgets. Centralized Green OA does, as well. The OSTP Memorandum specifically stated that no new funds would be made available to support public access initiatives, yet multiple agencies have opted to expand PubMed Central, which costs millions per year already, most of it spent converting author manuscripts. These decisions — all around a centralized solution, rather than a decentralized solution like CHORUS (which is not only more modern and scalable in design, but also free to taxpayers) — will only divert research and program funding to a redundant and unnecessary publishing environment.

In Europe, the same unreliable and politically susceptible funding mechanism needs to be carefully weighed. In general, dependence on politicians setting spending priorities seems a bad idea. Libraries have been suffering for three decades from inadequate funding thanks to decisions made by politicians in the 1980s and not addressed again. Research funding is not a priority. Anti-tax stances continue to cripple Western economies, leading to counterproductive austerity measures. It may not be wise to give a high percentage of the responsibility for funding scientific communication to these forces, which are both unpredictable and very insulated from the consequences.

As for the UK, here we see some of the major problems of centralization of an author-pays payment model (or any payment model, for that matter) — that is, huge administrative costs for the central bureaucracies managing payments. The shift from individual to institutional subscriptions created new costs and bureaucratic burdens for libraries. The shift from libraries to funders and university administrations (who will lean on their libraries, it seems) simply makes matters worse. Instead of librarians dealing with themselves, in the future they may have to deal with hundreds or thousands of authors, each with multiple funder mandates for each paper.

Price sensitivity in a more centralized payment model is bound to be higher. This will matter, especially if recent trends in OA publication rates continue. Many sources are noting that there were fewer Gold OA papers published in 2014 overall, and that the CAGR for number of papers in major Gold OA outlets has dropped dramatically. The market may have even shrunk in 2014. If fewer papers are now being published via Gold OA, then the only way the model works is for prices to increase. However, with  centralized payers, price increases are very unlikely to be significant or sustainable.

On top of that, there is repeated evidence that OA does not prove to be cheaper overall than subscription models. This is a point we should not underestimate. If the total cost of “free” is greater than the total cost of “paid,” then we have created a paradox.

There is another point about all this — most OA publishing is currently subsidized by the subscription model to some extent. That is, most publishers dabbling in OA have strong subscription models that throw off surpluses currently, while also paying for the staff, infrastructure, and editorial teams making the OA publications work. If OA has to carry the burden itself, can it? Can it do it consistently? What level of actual expense to fewer payers are we talking about?

Is it just me, or does the centralization of payers, the deep dependence on political spending, and the increased expense of administering and achieving OA publication, give you pause?

The subscription model, while vilified, continues to operate well and offers many calming benefits. It has diverse inputs — from personal subscriptions (from household or business budgets) to site licenses (which are paid for by student fees, tuition, grant overheads, and other sources that institutions can refine and adjust over time). It is inherently cheaper for users because it has strong secondary revenue streams that make it more affordable (aggregators, licensing, advertising, and so forth). With the exception of grant overheads — and if you think doing away with the subscription model will prompt research institutions to decrease their overheads, you’re in fantasy land — all of these other aspects of the subscription model make it cheaper for the research community and let them spend their funds on research. Centralizing payments around government grants, major funders, and large academic centers means that researchers will be footing most of the bill. Not only that, but as the UK has discovered, those who pay often attract freeloaders. The UK’s taxpayers are providing free research to the rest of the world. Are large research institutions setting themselves up to become the sugar-daddies of OA, too?

A pillar in business is to diversify your risk. If you have a concentration of risk, you may fail because of one bad decision, one business calamity, or one vendor’s failure. Our entire industry seems to be drifting toward more centralized funding approaches, which concentrate risk. In addition to perhaps giving these payers more leverage over what gets published and reducing our independence, this trend concentrates risk to scholarly communication.

We need to be careful and thoughtful as we proceed. OA may constitute a structural and inadvisable redistribution of risk. The evidence is pointing to a future with fewer points of failure around the economics of scientific publication and communication. This has major implications. Failing to do a course correction soon could mean that we will be much more subject to power plays from funders and politicians.

Your business model matters. It can either preserve independence or make you vulnerable to influence. It can either increase or decrease risk. It can either concentrate power among your customers or make it more diffuse. The choices are yours, and they are important.

(Note: Seven years ago today, I launched the Scholarly Kitchen. At the time, I pledged privately to give it a year of dedicated effort to see if it would work. Now, after more than 2,500 days and 1,136 posts of my own, I’ve determined that it’s time for me to shift gears, making this my final post in the Scholarly Kitchen. I’ve learned a lot over the past seven years thanks to this blog, and I’ve been amazed to see what the Kitchen has become with the incredible help of everyone involved, generating a worldwide readership and posts of great quality. The Chefs (and especially Head Chef David Crotty) and the SSP have my eternal respect and gratitude. I have no doubt this will continue to be the place to visit and learn. The late, great Douglas Adams penned my favorite farewell for moments like this: “So long, and thanks for all the fish.”)