The recent news analysis of Sci-Hub by Kate Murphy in the New York Times provides an opportunity to discuss some important dynamics between the funding of research and paying for research reports and related materials.

In a one section of her analysis, Murphy wrote:

. . . Elsevier, like other journal publishers, pays nothing to acquire researchers’ studies. Moreover, publishers don’t pay for the volunteer peer reviewers or editors. But they charge those same researchers, reviewers and editors, not to mention the public, whose tax dollars most likely funded the study in the first place, to read the resulting articles.

The shift from "pay" to "fund" is an important conceptual shift in economic and financial terms, and a key dynamic in the scholarly publishing marketplace.

Many editors are paid. Murphy simply gets this wrong. Publishers typically buy out the time of major academics in order to keep funding at the university in place so the editor can be paid. For example, if an editorial job is estimated to be a "20% time" position (requiring a few hundred hours per year), the publisher would pay the university 20% of the academic's salary. For larger and busier journals, editors are full-time, with full salaries, benefits, and so forth.

By saying that publishers charge "those same researchers, reviewers and editors, not to mention the public," Murphy misses a chance to actually bring some data and nuance to the discussion about funding and paying. Only a fraction of scientists publish in any given year (more precisely, only a fraction are funded to perform research). In many fields where practitioners and clinicians predominate, reading is the primary information modality. Therefore, our information economy has a core asymmetry -- a few funded researchers, and a broader audience of interested readers. This market reality is essentially why the subscription model still holds sway -- when a few want to reach many, having the larger group pay spreads the costs and lowers prices for each participant in the system. Open access (OA) publishing has a business model (Gold OA) predicated on the inverse, and there are legitimate concerns that if it were to become predominant, expenses would hit a few major players (major research universities, major funders, major governments) inordinately. The UK has already had a taste of this with the RCUK's efforts of a few years ago. Therefore, when it comes to funding and paying in the scholarly information ecosystem, a core asymmetry should be acknowledged, one which make a model of broadly shared costs more generally appealing.

Why don't journals pay their authors? Aside from clear concerns about motivations and further increases to the expenses within scholarly publishing, there don't seem to be any benefits to be had from doing so. One study of cash incentives based on publication in top-tier journals showed that while authors submitted more papers to top journals when given cash incentives, they published no more papers in those journals. The study also found that researchers receiving indirect incentives -- salary increases or career progression for published research, for example -- submitted more papers and had more published. A plausible explanation is that the indirect incentives led scientists to do better work, while the cash incentives encouraged scientists to submit papers in a way more akin to playing the lottery. Also, paying authors would only increase the overall costs (direct and administrative) of the system.

So why do researchers submit their papers without being paid directly to do so? There are two major incentives for researchers to publish: to get credit and to claim primacy. Only by publishing can scientists make their research their own, and gain primacy for it. Publishing is also an effective way to measure productivity and prevent shirking, as the economist Paula Stephan has noted.

We want funded research to lead to publications of findings. Therefore, the funding of research directly impacts what institutions and other purchasers pay in aggregate. Here, we get to a crucial relationship between funding and paying. Funding of research has tripled in the past 20 years when you compound the increases. This has led to a tripling of research outputs, which has led to three-times as many published papers. These related volume increases have led to a tripling of what many institutions pay to access. Yet, publisher prices have only increased 9% over that period vs. a 67% increase in the CPI. This is an average. In the UK, subscription prices have actually fallen quite a bit, adjusted for volume. To absorb this increase in volume, many journals publish more papers than ever. Many new journals have been launched so new research findings can be published. This all leads to higher overall expenditures (paying more because there is more to buy), despite prices (adjusting for volume) only increasing 9% over the past 20 years.

So, a more accurate statement would be:

Funding of science has tripled, generating three times as many papers, and this volume explains the increase in expenditures to access the bulk of the scientific literature. Publishers have actually been controlling their costs during this explosion of available research reports, and most of the increases in payments libraries and others are seeing can be explained by these volume increases, which is itself explained by funding increases.

It's a dynamic that we will continue to wrestle with, especially as we continue to fund more STEM education, encourage more children to pursue STEM careers, and push for a world with more scientific research.

Unfortunately, at the same time, library funding (which impacts institutions' ability to pay for all this additional research) is not being maintained. Even with triple the research funding (much of it going to universities) and 200-500% increases in tuition (what students and their families are paying), funding of libraries has been falling as a share of university budgets for three decades. This seems an abdication of economic responsibilities at the university level, yet one that is rarely called into question. Libraries are paying out of an eroding funding base, despite their institutions being generally better off than ever and the increasingly vital role of the libraries' scholarly collections.

Another important nuance is that scholarly publishers pay for the infrastructure that supports the communication of research results -- from online systems to standards like ORCID and the DOI, to initiatives like CHORUS and HINARI. Paying for infrastructure also means paying the cost of rejections, the work to publish the next papers, the hiring and training of new editors, the expansion of titles to support the growth of scientific outputs, the maintenance and migration of archives, and so forth.

Therefore, to say that a publisher "pays nothing" to acquire new papers is a misstatement. You'd never say that a newspaper pays nothing to cover a sporting event or a traffic accident. The same holds true in our field, as we pay the salaries, systems costs, and infrastructure costs to be ready to evaluate dozens of papers per day, including personnel costs for acquisition editors, scientific editors, and so forth. Competition for papers can be fierce, and creating the venues, systems, reputations, and processes to support scholarly research publication certainly costs money. Publishers take on these risks on behalf of researchers, paying for all of this and more, with no guarantees of success.

In summary, despite many profound shortcomings, the news analysis in the New York Times invites a discussion of some important facts, nuanced points, and interesting financial and economic realities:

government funding of research is often partial funding
increases in science funding across all sources will drive the volume of research outputs, which increases library and OA expenditures because volume is a major determinant of what scientists and their representatives spend in the scholarly information economy
stolen access is not open access, and cannot actually lead to the demise of the current system because you can only steal what exists
university budgets benefiting from research funding growth have been funding libraries at consistently lower rates for 30+ years
publishers are integral to the scientific and scholarly communication infrastructure, running an independent signaling marketplace that has low costs (0.5% of overall research funding, volume-based price increases well below the rate of CPI increases), and which appears to be fairly efficient

As long as more scientific research is funded, the scientific community can expect to pay more for well-edited, carefully curated, independently vetted, and competitively placed research reports, no matter the underlying business model. We may slow the already low rate of price increases, but as long as more funding drives a higher volume of researchers with valid scientific findings seeking outlets that boost their career prospects, aggregate spending seems to track aggregate funding fairly reliably.