My colleague Ginny Steel, UCLA University Librarian, published an “open letter” on 19 October 2016 opposing proposals to “flip” scholarly publishing from post-payment (subscription, or paywall) funding to pre-payment (article processing charge, or APC) funding, as advocated by, for example, the OA2020 initiative. While I admire UCLA’s commitment to advancing open access and affordability of scholarly publishing (which are two separate questions, though often entwined), I disagree with almost all of her arguments against pre-payment (“flipping”). They are based largely on unsubstantiated fears, fears that will remain unsubstantiated because they are a result of fundamental flaws in economic logic.
The basic argument is rather simple. Changing the funding model from post-payment to pre-payment does not increase the costs of production journal publications, thus publishers will not have higher costs that they “need” to recover. And changing the funding model from post-payment to pre-payment does not increase publisher market power, so they will not be able to collect more money from research institutions than they already collect.
In fact, the most likely outcome is that publishers will be forced to compete economically for article submissions, and this will lead to research institutions paying less to the publishing industry than it would in a continued world of post-payment (subscriptions).
Opposition such as Ginny’s – and she is far from the only person expressing these fears – makes me especially sad because they are expressed without any constructive, specific, coherent alternative for achieving the two goals: open access, and lower costs of scholarly communication.
Ginny offered many “concerns” about the economics of pre-payment publishing, before stating her opposition. Most of her arguments are variations on misunderstanding the same basic economic issues, but since each is stated as a separate concern, I respond to each. Block quotes below are all from her open letter.
“Certain titles carry more prestige and have a higher impact factor, which allows them to set higher article processing charges. The model does nothing to resolve this disparity in pricing, nor to address the factors that drive authors to pursue publication in these journals, such as tenure and promotion processes. It remains an unknown, and possibly quite expensive, factor in determining what the model would actually cost.”
Yes, some titles carry more prestige, and objects in higher demand can charge higher prices (nothing special about publishing in that). But this is true of the very same publications when the price they charge is a subscription. Nothing about pre-payment (article processing charges, or APCs) rather than post-payment (subscriptions) increases their market power – that is, the economic opportunity to extract even more profit from us than they already. Thus, changing from post-payment to pre-payment does not make this situation any worse.
In fact, pre-payment through APCs offers an economically sound path to decrease publisher market power and to reduce prices. The source of prestige – and market power – is copyrighted content. Subscription buyers (primarily academic libraries) have no control over the copyrighted content, and thus very little economic power in this market. However, authors do have control over their copyrighted content. Authors, if they have a direct economic stake in the outcome, are in the best – position to reduce publisher market power and slow the growth in prices.
As I have explained elsewhere, open access through pre-payment does not require that authors pay APCs themselves – libraries can pay the APCs just as they now pay for subscriptions: publisher market power won’t increase and we will get OA for the same total cost that today we get paywall (subscription) publishing. That is a huge step forward, opposing OA through pre-payment because it won’t also solve the pricing problem doesn’t make sense.
Yet pre-payment also offers the opportunity to reduce publisher pricing power if we let authors have an economic role. I’ve explained that author payment of (some of) APCs does not require that authors use their current scarce funds – we can provide (some or all of) the funds library currently pay for subscriptions to authors to spend instead on (some or all of) APCs, but let the authors – who control copyright, the source of economic power in publishing – choose between comparably prestigious but differently priced journals when submitting their articles.
“Given that their perceived status (or careers; see previous point) depend on it, authors will continue to be driven more by journals’ prestige than by the relative affordability of article processing charges.”
This argument is based on a common fallacy. Almost never is there just one standout journal for a given article submission. Thus, prestige generally is not the only thing on which submission decisions are based. (Other factors include those that affect acceptance likelihood – such as fit to the editorial board’s preferences – and expected time to publication.) All that publisher competition for submissions based on price (APC) requires is that they care somewhat about cost: that is, for some articles, some of the time, authors choose the lesser priced of roughly equally prestigious journals.
“Thus, the total cost to an institution of this model must include not only library journal subscriptions and grant funding allocated for article processing charges but also some unknown level of funding each author is likely to request from his/her department or take from his/her own discretionary funding.”
I don’t know why so many librarians have an irrational fear that changing the flow of funds will lead to publishers being able to extract even more money from universities…but there is no economic basis for that fear. If switching to a pre-payment (APC) model would enable them to make even more money…why haven’t they gone ahead and done so already? Publishing pre-payment journals does not use up more economic resources (labor, server time, data storage, paper and ink), so there will be no greater production costs publishers need to recover. And pre-payment does not increase publisher market power (in fact, more likely decreases it) so pre-payment doesn’t lead to higher profit extraction either. If the publishing industry can thrive on the funds we currently fork over, then those funds, directed to APCs rather than subscription payments, will continue to suffice.
“There is currently no definitive information about the true costs to publish a journal article. Though article processing charges vary widely, there is some evidence that many would need to be raised considerably to cover the actual costs of publication. Without reliable data, it is impossible to predict the true costs to research institutions of the “flipped” model.”
This repeats an error above. Changing the flow of funds to publishers from post-payment to pre-payment does not increase the cost of publishing journals. If the cost of publishing the journal does not change, then to earn the same profit as they do with subscriptions, publishers need the to obtain the same amount of money from APCs as they currently get from subscriptions. Total revenues do not need to be raised. 
“The model would not result in any fundamental changes to the current balance of power in journal publishing.”
Yes! Or almost yes – on this we basically agree, and this is why most of the arguments against pre-payment (APC funding) are flawed: switching from post-payment to pre-payment does not, to first approximation, change the balance of market power. And that is why it won’t lead to higher prices…but it will achieve the long-sought goal of open access. Sounds like a pretty good deal! So why are so many librarians scared of it?
The almost part – as I’ve explained above and elsewhere, pre-payment funding actually is the most hopeful (economically sound) route to decreasing publisher market power (if authors are allowed to make some of the economic decisions – which they don’t today).
So, depending on how APCs are paid by institutions, either we are no worse off economically (but we achieve open access), or, we are better off (publisher power decreases).
“The model would require institutions to spend significant time and resources to coordinate and execute it, both during the transition from the current system to a “flipped” model, then afterward, to administer authors’ article publishing charges on an ongoing basis. These would increase the total costs for the model beyond simply the article processing charges.”
Let me first address the end-game: the post-transition world in which journals have already been flipped. No, there would not be significant time and resources to manage such a system. Every research institution already has a system for managing the flow of transactions related to research activity: the salaries and benefits for research assistants, the travel reimbursements, the computer and lab equipment purchases, etc., etc. The incremental cost of adding APC payment to the existing funds management system is negligible: we don’t need a new system, just to process a single additional transaction per article. Think of the many transactions that are already processed per article of research produced. Put another, way, the processing time for managing one more charge will be a tiny fraction of the cost of that charge itself, and the cost of publishing is estimated to be only about 1% of research expenditures, so at best we’re adding a tiny fraction of 1% of the cost of research. And even that negligible estimate ignores the substantial savings from not having to engage in protracted, costly license negotiations with publishers over subscription packages.
What about the transition? Here, I basically agree: there will be some additional costs during a transition (the funds management costs are still negligible, but there will be other more significant adjustment costs). Most new, improved systems require some up-front investment: if we weren’t willing to sometimes make those investments, library users would still be required to go to the card catalog. One reason to engage in a broad, international initiative to quickly transform scholarly publishing is that a short transition will minimize transition costs.
“SCOAP3 (Sponsoring Consortium for Open Access Publishing in Particle Physics) could be considered as a pilot, or use case, for this model. This partnership of more than three thousand institutions in forty-seven countries started with ten journals in 2014; in its next phase (2017-19), only eight journals will continue to participate.”
Umm, yes, some publishers may decide they prefer post-payment (subscriptions) to pre-payment (APCs). That hardly seems like a reason for research institutions to decide pre-payment is a bad idea – perhaps quite the opposite.
“The model is currently supported by very few organizations in the “global south”; paying article processing charges is difficult in those regions, where institutional budgets are extremely low. Thus, the model would further skew a scholarly publishing system that is already geographically unbalanced.”
The same common flaw underlies this argument: that APCs will cost us more than subscriptions. Overall, producing publications will not cost more, and market power will not increase, so the total revenue paid to publishers will not increase.
There is a question of whether some research institutions will pay more with pre-payment than subscriptions while others pay less. And the answer is probably yes, if prices don’t fall (or increase at a slower rate). (Since prices probably will fall, or at least increase at a slower rate, it is quite possible – even likely – that all institutions will pay less, as the market power of publishers decreases.) But this is almost surely a good thing for the global south. What matters is the relatively intensity of research consumption (subscribing to journals) versus research production (submitting articles to journals). Institutions that produce disproportionately more articles per researcher (of which mine is one) are the ones likely to face a possible increase in total cost. Those institutions that are not are at the high-end of research intensity (e.g., four-year liberal arts colleges, many global south institutions) will likely see a decrease in total cost: they will save more on subscriptions than they pay in APCs. This has been demonstrated in the Pay it Forward project.
We have been working towards an open access world for some 25 years now, with very little progress and many reasons to think that current efforts will not result in near-universal open access. Meanwhile publisher market power has been increasing, and current efforts offer no economic basis for reducing that power and the monopoly profits that many publishers extract from research institutions.
After cutting through the unfounded and economically illogical fears, switching the publishing industry from post-payment to pre-payment funding offers a path to near-universal access without raising prices (at least, not by more than they will increase in the post-payment or subscription model), and a path to reducing publishing market power and thus reducing the total payments they extract from us. Unless a more economically sound, practical approach emerges – and it hasn’t in 25 years – shouldn’t we be working together to figure out how to implement this win-win opportunity?
 I write this as someone deeply concerned that we achieve open access, and that we reduce the cost of scholarly communications: I am the university librarian at UC Berkeley. I also write with experience and expertise: I am an economist, and have spent most of my career doing research on markets, competition and incentives. I have published on monopoly, and on the economics of information production and distribution. I have been testifying in federal antitrust lawsuits as an expert witness since 1989.
 More than suffice! The leading for-profit journal publishers are earning profit rates of around 40% – that is, monopoly levels – and society publishers are earning enough surplus (the non-profit word for profit) to subsidize their other activities.
 Casual use often confuses “cost” with “price”. The cost is the underlying resource cost to produce a good or service: how many hours of labor, how many computers, how much paper and ink. It is what the producer pays to create a product. Price is what the producer charges its buyers to obtain the good or service once produced. The problem is that to buyers, the price is their cost, and so buyers (e.g., librarians) often talk about the “cost” of journals which is fine if all you are discussing is purchasing, but when discussing both the publisher and the buyer side, this creates confusion: do we mean publication “cost” will increase, or journal payment “price”?
 I think this fallacy may arise because some conversations are about whether future APCs will look like the current APCs for the few, non-representative journals that use this model. Probably not: it’s not an apples to oranges comparison. But that’s the wrong question if we’re trying to predict the overall cost of scholarly communications to research institutions: what matters is the total payment to publishers, and if the cost of producing the journals doesn’t change, then the total payment doesn’t change unless market power increases – and as I’ve argued above and elsewhere, pre-payment either leaves market power unchanged or decreases it (for publishers).