From colleagues in the library profession, I regularly hear the assertion that the current scholarly communication system is “broken” – largely because it requires academic institutions to “buy back” their own professors’ work from commercial publishers, or because it requires taxpayers to “pay twice” for research they have already funded. For reasons I’ve enumerated before (and therefore won’t belabor here), I disagree.

I do, however, believe the system is broken — or at least fundamentally flawed — in at least one important way, and that is that its structure forces each participant to make choices in vacuum, or at best in a distortion chamber. To put it another way, each participant is a blind person with his or her hand on a different part of the elephant.

In an earlier posting, I alluded to the fact that the journal publishing marketplace is not traditionally competitive. Journals do compete for content suppliers (authors) because every journal provides authors a functionally similar service. However, journals don’t compete in the same way for buyers, because every journal provides unique content — two linguistics journals may look, on the surface, like similar products offering competing value propositions to the marketplace, but in fact each offers a completely unique value proposition, since the articles in each journal are unique, only broadly related to each other, and unavailable from any other source.

Choosing between two linguistics journals is not like choosing between a Ford and Chevrolet, both of which do almost exactly the same thing and compete for your business around the margins of the driving experience (smoother drive, better gas mileage, etc.). In this scenario, you only want to buy one car and will try to choose the best one you can get at the price you can afford. Two linguistics journals, on the other hand, compete for your subscription money in the same way that groceries and clothing do: what you really want to do is buy both, but if you don’t have enough money for both you have to choose between them. You can’t spend the same dollar twice.

But the more I’ve thought about this issue, the more I’ve become convinced that the weirdness of the scholarly communication marketplace goes far deeper than the question of competition for buyers. In fact, I think the frustrations that many constituents experience with the current structure arise from the fact that each participant in the system gets fundamentally incomplete signals from the marketplace in response to the inputs it contributes.

Here I’ll focus mainly on the journal marketplace, which accounts for the great majority of dollars spent by research libraries. Books pose a somewhat different set of issues.

To invoke another metaphor and provide some comparative context, consider the marketplace for office chairs. If I choose not to buy a chair from Herman Miller, that doesn’t mean I don’t get to have a chair; it just means I have to buy a different brand of chair. Herman Miller can only command a very high price for its chairs if, a) a great many of its potential customers believe its chairs are of very high quality, and b) those customers have sufficient money to pay for them. Herman Miller’s ability to command those high prices will be undermined to the degree that other chair makers enter the marketplace and offer high-quality chairs as well. If, however, Herman Miller were the only chair maker in the world, its position would be much different, and the market’s responses to its pricing signals would be distorted.

This is, arguably, the position in which journal publishers find themselves. So let’s start with their situation.

Publishers. Journal publishers sell two services in the scholarly-communication marketplace — one is a suite of services they sell to authors (who trade distribution rights and often their copyrights for the services of review, editing, certification, dissemination, and some degree of archiving) and one set of services to libraries and individual subscribers (who trade money for access to articles). Author services are provided in a more-or-less normally competitive marketplace, because authors have a meaningful choice of venues for their work. Services to libraries, however, are not. The moderating pressures that competition exerts on price increases in more conventionally competitive environments do not apply in the journals marketplace. If a publisher decides to double what it charges a customer for access, the customer does not have the option of going to a competitor for a functionally similar product. This fact severely distorts the signals a publisher receives back from the marketplace in response to any change in price, and thus makes it very difficult for the publisher to know whether it is charging the “right” price.

Authors. Authors also receive partial and distorted signals from the marketplace in response to their inputs. They submit manuscripts to publishers, and get feedback based purely on quality and relevance: if their submission is of high enough quality and good enough fit for the journal the article will be accepted, in which case the author will receive (in lieu of payment from the publisher) the benefits of editing, certification, and dissemination. The author’s choice of journal has varying impacts on the marketplace depending on whether she submits to a $500 journal or a $20,000 journal—but those impacts are absorbed by other participants in the system, and are not felt directly by the author. The only consequence directly experienced by the author is that of prestige and exposure (or lack thereof). An author who contributes to a journal seen by others as exploitative or “immoral” may get some negative feedback from colleagues, but this is quite unlikely—and that feedback constitutes a relatively weak signal compared to, say, seeing one’s name on the contents page of the top journal in one’s field. Similarly, when an author contributes to a more “virtuous” journal, she receives no direct benefit for doing so.

Libraries. Although they don’t usually characterize it this way, research libraries are selling a service in the scholarly-communication system as well — or, more accurately, a suite of services. In return for institutional funding (and, often, explicitly dedicated student fees), research libraries broker access to information resources and provide a variety of support services to students and scholars. They also provide archival hosting for print backfiles and, increasingly, digital archiving of scholarly products as well. In an academic environment, journal purchasing decisions are typically made by librarians — not by the readers who will benefit directly from those decisions. Since library budgets are limited, librarians have to cancel or forego some purchases; but since readers don’t experience the impacts of price, they put pressure on the library never to cancel or to forego purchase. If the library cancels a subscription, it experiences budgetary relief — but does not feel (thought it may hear about) the pain of its constituents who have suddenly been denied access to content that may be important to their work. Librarians do typically canvass their constituents in order to gather information about the impacts of their decisions, but information gathered in this way will always be partial and only partly reliable.

Readers. On an academic campus, readers experience little or nothing of price. It is no harder for them to access a library-provided $20,000 journal than a $1,000 journal. What they do experience are relevance (“Does the library give me access to journals in my field?”) and quality (“Does the library give me access to really good journals in my field?”). Since price impacts are outside the readers’ experience, they will tend to exert significant and constant pressure on libraries to acquire relevant and high-quality journals, no matter what they cost. Obviously, pricing levels or trends that might modify the reading behavior if readers paid the subscription bills themselves will have little if any impact on reading behavior when the bills are being paid by someone else. Similarly, readers (as faculty members) don’t care about cancellations except those that fall within their disciplinary purview. A mathematician, faced with the possibility of losing access to a $5,000 journal subscription due to insufficient library funds, will happily let the library cancel one or more linguistics journals in order to preserve the math title — and vice versa.

In summary, the biggest problem I see with the existing scholarly communication marketplace is not the fact that faculty provide “free” content to journal publishers and then ask their libraries to pay for the formally-published versions of those same articles. This is a natural consequence of the value-added services that the academy expects publishers to provide. The biggest problem is the fact that each participant in the system receives distorted and radically incomplete market responses to its inputs. There is virtually no competitive pressure on publishers to control journal prices; authors’ submission decisions have significant impacts on other players that the authors themselves never feel; librarians make selection decisions but do not experience directly any of the meaningful consequences of those decisions either; readers make requests of their libraries without regard to price, because they do not pay the bills.

Solutions? It’s difficult to see a solution to this problem that would not take scholarly publishing out of the commercial marketplace altogether. Some would hail such a development. I will share thoughts on this idea in my next posting.