This is gonna replace CD’s soon; guess I’ll have to buy the White Album again. — Agent Kay, “Men in Black”

One of the ideas that has been relatively unexamined pertaining to the purchasing and pricing of scholarly content is the notion that publishers should no longer engage in “double-dipping” — that is, charge once for something, not twice, and never again. This is stated especially as regards open access (OA) articles in hybrid journals, but also is reflected in the preference some have for the CC-BY license.

SPARC Europe defines double-dipping thusly:

The term “Double Dipping” is used because publishers receive money twice: the subs[c]riptions are still paid by the universities and authors pay for Open Access

Peter Suber has complained that double-dipping isn’t prohibited by RCUK policies regarding Gold OA, and the fact that the policy allows it:

. . . is a waste of taxpayer money and a giveaway to bad actors.

But does double-dipping really make you a “bad actor”? Is it really a waste of money? Or is double-dipping merely what content businesses do? Does it make things affordable? Could it be that OA publishers themselves double-dip?

In the “no double-dipping” theory of hybrid OA journals, publishers who have 5% of their content published as Gold OA should reduce their subscription prices by 5%. Otherwise, they are double-dipping. For instance, here is Elsevier’s policy on double-dipping:

Elsevier amends its journal list prices to account for each and every sponsored open access article. Our subscription customers are not charged for open access articles – we do not doubledip.

While the concept of “no double-dipping” has obvious charm and a feeling of fairness, the real world poses challenges both to its utility and to its fairness in practice.

UTILITY PROBLEM — The “one to many” problem. The asymmetry between the subscription model (where many readers pay for content) and the Gold OA model (where an author pays once) creates a mathematical problem around “double dipping” constraints. Gold OA is predicated on the notion that an author can pay an APC once, and the article is free from that point forward. If the article is also published under a CC-BY license, the opportunity for secondary revenues based on selling the content to secondary markets is reduced even further, theoretically to zero. For argument’s sake, let’s say one publisher’s Gold OA fee is $1,500. Then, let’s say that publisher publishes 5% of the articles in a hybrid journal as OA articles, publishes 500 articles in that journal in total, and has to decrease the price of each of its site licenses by 5% to avoid the perception of double-dipping. And let’s say this publisher has 2,000 site licenses, which it sells for an average price of $750 each. Based on this math, the publisher makes $37,500 in APCs, but has to eat $75,000 in site license discounts to avoid double-dipping. The net loss to the publisher? $37,500. This is the problem of counting one article transaction against multiple licensing transactions — each license benefits from the reduced fee, but the publisher only receives the fee once.

UTILITY PROBLEM — Too many moving parts. Now, in the scenario above, presumably the publisher could discount the site license by only 2.5%, but this isn’t likely to satisfy librarians watching for double-dipping, especially if the publisher had also planned a 7% price increase and then tries to argue that it only increased price by 4.5%. There is likely to be a backlash if the discount rate doesn’t match the percentage published as OA. There are just too many moving pieces in pricing to isolate the influence of hybrid OA inside a pricing This becomes even more difficult when you include the wildcards of bundling and negotiations.

UTILITY/FAIRNESS PROBLEM — The effect on APCs. Let’s assume that despite all this complexity, somehow a “no double-dipping” approach becomes workable. A publisher is going to have a rational reaction — keep revenues up or increasing slightly, and look for another way in the same model to find more money. The most immediate option is then to increase the fee for the related Gold OA. Hence, “no double-dipping” creates upward pressures on OA fees. This makes the benefits of enforcing a “no double-dipping” policy unclear.

FAIRNESS PROBLEM — The deeper-dip at larger institutions. If we assume that Gold OA APCs have to increase in order to support not only the discounting required in a “no double-dipping” world, but also to support some share of normal annual price increases, we should begin to see more regular upward pricing dynamics in APCs — after all, why not increase prices elsewhere to avoid controversy? If APCs rise, who will pay for this? For good logical reasons — they publish more research, namely — larger institutions are most likely to pay the lion’s share of Gold OA APCs, shifting more costs to them, making them effectively the subsidizers of discounts across the board. The double-dip turns out to be an industry-wide deeper-dip into the coffers of larger institutions, and a de facto subsidy to smaller institutions.

FAIRNESS PROBLEM — Fully OA publishers are allowed to double-dip. While subscription publishers are reminded not to double-dip when they publish a hybrid OA journal, OA publishers are not reprimanded about double-dipping with APCs and institutional memberships. At BioMed Central, institutional and corporate members receive discounts, but there is no guarantee their membership fees will be recouped through these discounts. The risk is shifted to the institutions as soon as the membership is paid — will you get your money’s worth? There is no mention of refunds for unused funds, which means BioMed Central is essentially double-dipping — taking money from authors and institutions, and keeping any unclaimed funding it receives from institutions. It is the OA version of the double-dip. At PLoS, researchers at member institutions receive a 10% discount on PLoS APCs. Again, there is no guarantee that all the money will be utilized through discounts. If there were a true dedication to no double-dipping at BMC or PLoS, you would think that any funds not used in the discounting — which is ostensibly an incentive to authors at the sponsoring institution — would be refunded to the institution. Otherwise, BMC and PLoS are living off the margins (the unused discounts), which is equivalent to double-dipping — making institutions pay for OA when they shouldn’t have to.

UTILITY PROBLEM — “No double-dipping” drives up prices. The double-dipping concept gets to one of the fundamental truths of content businesses, which is that a key survival technique is to sell the same content multiple times, but at lower prices in each instance than you could if you didn’t have other outlets. Not only does this help lower prices for various participants in the content marketplace, but it also protects publishers from dramatic shifts in single segments because these can be offset by dynamism elsewhere. Selling a movie in theatres, via DVD, on demand, via a tie-in book, and as a collector’s box set is not “double-dipping” — it is a sales technique that allows customers to differentiate what they want to buy, creates affordable prices at every sales point, and ensures enough funds for the next project. Selling filtered and refined scholarly content to relevant users, to licensing entities, to businesses, to information services, and to sell the audience’s attention via advertising is all about selling the same content multiple times — and it makes each piece viable if done correctly. Refusing to allow multiple instances or methods of commercial gain only puts more pressure on those that are allowed, driving prices up accordingly. The “no double-dipping” theory and the CC-BY license both stifle revenue diversification.

FAIRNESS PROBLEM — Publishers can’t “double-dip” fees but institutions can “triple-dip” the value out of subscriptions. One fundamental change with the IP-based site license was that all the value of a large institution’s content access suddenly went through the library budget. Departmental subscriptions withered away quite quickly. Add in proxy server access, and the institutional site license also helped erode personal subscriptions and began to whittle away at membership dues as well. In essence, the site license is “triple-dipping” in value, and is itself a great example of how pricing pressures visit when business diversity is limited. Prices go up faster than inflation because the underlying diversified business is eroding at a rate higher than inflation, and the site license has to bear the parts of the business it is absorbing. Again, we have an asymmetry — the site license is absorbing all sorts of revenue opportunities for publishers and non-profit societies, yet there are efforts to decrease the ability of these entities to find new ways, including OA, to make up the difference.

UTILITY/FAIRNESS PROBLEM — “Double-dipping” lowers taxes. The government itself double-dips, and many times over. Taxpayers pay for taxes to fund national parks, but also pay admission fees. Taxpayers support bonds to build airports and their parking structures, but still have to pay parking fees even after the bonds have been repaid. The same goes for toll roads. Schools charge taxpayers various fees for their students to participate in athletics, musical programs, and other activities, despite the taxpayers having paid taxes to build the buildings, pay the teachers and coaches, and buy and fuel the buses. All these “double-dips” factor into the overall pricing (tax rates) and sense of fairness of the system. If there were no double-dipping, taxes would increase for everyone. Double-dipping in the scholarly space may be just another way of making sure taxes don’t go up. For instance, it will be interesting to see how the UK government deals with the RCUK mandates, their effect on research funding, and their tax policies. I believe OA in the UK will lead to a general tax increase at some point unless other funding approaches come in to alleviate the funding diversion the RCUK mandate has created.

UTILITY PROBLEM — Even OA has carrying costs that require long-term payments to support. As I’ve noted elsewhere, online publishing certainly isn’t free, and in the scholarly space, isn’t cheap. Using one-time payments to support an ongoing online business creates a type of payment pyramid scheme — as costs rise, you need more payments to cover not only current costs but carrying costs for what you’ve done previously. This is why the bigger OA publishers have created membership options — they need the recurring, renewable revenues memberships can provide in order to attend to their infrastructure and keep APCs low. In essence, the long-term vitality of any content business requires some level of consistent support. If we want the OA literature to be around long enough to provide a meaningful archive, it needs a complementary source of support in addition to APCs.

FAIRNESS ISSUE — Academia itself double-dips. You don’t have to look far to see double-dipping in science or academia:

• Tuition and fees, both charged to the same students.
• Multiple grants supporting the same research project.
• Universities taking grants from governments and philanthropies and keeping the patented IP that emerges.

Essentially, “double-dipping” is a pejorative term for risk mitigation strategies in a business model and for a clever system designed to create multiple entry points and overall affordability. Individually and locally, we can afford things we couldn’t otherwise because all we’re paying for is one of the many “dips” a content or other business is making in a market. The site license is more affordable because advertising, reprints, licensing, or other lines of business allow it to be priced well below what the site license alone would cost. The DVD is $14.99 because the box office receipts were so good, the on-demand revenues are strong, and the sequel (paying it forward) is being primed. We should recognize “double-dipping” for what it is — risk mitigation, price sensitivity, and market awareness — and not condemn it or make policies that have an unnecessary chilling or unanticipated rebound effects.

Given all this — the pressures that single-dipping could place on APC prices, on subscription prices, on institutional budgets, on taxpayers — it seems we may actually want to support reasonable “double-dipping.” Done well, double-dipping makes things affordable.