Who’ll Pay for Public Access to Federally Funded Research?

The status quo for the last decade in publishing the results of federally funded research—sequestering it behind a paywall for up to one year—was from the beginning a compromise that fully satisfied no one. It created modest changes in publishers’ business models but also prevented scientists and citizens from accessing the most recent developments advancing environmental justice, cancer research, clean energy technologies and more.

The U.S. government acknowledged as much in August when it announced a new policy for free, immediate public access to federally funded studies, scheduled to take effect by 2026. Open-access advocates celebrated the news, while others, including the Association of American Publishers, expressed concern about the economic impact on members. But details concerning how much the new policy would cost and who would pay were left for another day.

“It is almost as if there was a decision to kick the hornets’ nest and just see what happens,” Tim Vines, founder and project lead of DataSeer, an artificial intelligence–based tool that helps authors, journals and other stakeholders with sharing research data, told The Scholarly Kitchen. “In this day and age, who needs to consult with stakeholders to see what sustainable open policies should look like?”

The new mandate, which applies to all federal agencies funding research, removes a 12-month embargo on article sharing implemented as part of a compromise in 2013 that has been key in sustaining publishers’ subscription-based business models. In short, as of now, the government is no longer willing to compromise. Federal agencies that fund research must develop plans that permit immediate public access to research and the underlying data. This includes peer-reviewed research articles and final manuscripts published in scholarly journals and may also include book chapters, editorials and conference proceedings born from federally funded research.

Under the new policy, researchers will be allowed to “include reasonable publication costs and costs associated with submission, curation, management of data, and special handling instructions” in the budgets they submit to federal agencies in applications for research grants, according to the White House memo that announced the policy. An economic impact statement that the Biden administration submitted to Congress about the policy offers rough, incomplete financial details that some argue relies on flawed data. It also makes no promise to fund the mandate. That has left many scholars wondering whether publishers will adapt their business toward more open-access models or whether authors will be on the hook to pay for the unfunded mandate.

“The path of least resistance is for scientists to write into their grant proposals the open-access article-processing charges that the publishers would like to generate,” said Roger Schonfeld, vice president of organizational strategy at Ithaka, a nonprofit focused on improving access to knowledge and education. “Some scientists may be happy to do that. Other will ask, ‘If I can get $1 million from a grant maker, do I use $100,000 [of that sum] to pay for another postdoc or to pay for open-access publishing charges?’ A lot of researchers would prefer to have that postdoc.”

The White House memo acknowledged that the federal policy change would likely result in “some downward pressure on publishers to adapt their business models.” This expectation or suggestion (depending on one’s point of view) has precedent in the 2013 policy that required federal departments and agencies with more than $100 million in annual research and development expenditures to provide public access for publications resulting from federally funded research, with an optional 12-month embargo.

Notably, publishers adapted their business models to be less reliant on subscriptions. In 2012, 7.9 percent of all scholarly articles were published as paid-for open access (sometimes called “gold open access”), according to Delta Think, a consulting and advisory company that works with publishers. This compares with approximately 36 percent in 2020, according to Delta Think. Many publishers now offer open-access or hybrid journals with different copyright, payer and access options.

The landscape of scientific publishing has also adapted in other ways in the past decade. Plan S, a European public-access initiative that took effect in 2020, impacted scholarly publishing and access worldwide. Also, transformative agreements, such as the 2020 open-access deal between the University of California system and publisher Springer Nature or the deal between more than 140 U.S. institutions and Cambridge University Press, accelerated the public-access trend. Google Scholar returns results with author-accepted papers. And librarians are better equipped with tools to determine which journals their patrons really need, according to an analysis provided by Clarke & Esposito, a management consulting firm that focuses on academic publishing, among other topics.

“The era of the journal subscription may be nearing its end,” the firm wrote, without forecasting a timeline. “A rapid decline in subscription journals would require federal agencies to adjust their budgets more rapidly, either by requesting additional funds (funds that the U.S. Congress may or may not be willing to provide) or by issuing fewer (or smaller, once publication fees are factored in) research grants. A longer window of subscription viability would be easier for agencies to plan for and manage.”

Not everyone is convinced that a European-style public-access model is viable in the United States.

“In the U.S., that model is a hard one because of the structure of our higher education system,” Schonfeld said. “In Europe they tend to do national-level licensing. We don’t have national-level negotiating with publishers … Right now, lots of institutions pay to read things in the U.S.” Schonfeld sees possibility, if not certainty, in the fact that the 2013 mandate—the one that allowed for a 12-month embargo on article sharing—did not negatively impact publishers’ business models.

“Publishers … might say, ‘If a 12-month embargo didn’t ruin our subscriptions, maybe a zero-month embargo won’t, either,’” Schonfeld said. In such a case, he suggested that publishers might experiment with value-added subscription models offering, for example, enhanced editorial processing or links on their platforms that offer resources beyond the mandate.

The economic impact statement estimated that the government funded between 195,000 and 263,000 articles in 2020 that cost an estimated $2,000 to $3,000 per article to publish. That translates to an estimated cost between $390 million and $789 million for federally funded research articles. Though the statement noted that this figure was “less than half a percent of the billions of dollars spent on taxpayer-funded research each year,” it made no promise to pick up that cost. The impact statement also pointed to healthy publisher profits, given its estimate that publishers themselves spent between $200 and $1,000 to publish an article. The statement identifies factors that might support the mandate, including publishers’ shift from print to digital in recent years, the cost decline of hosting data online and the availability of cloud storage services.

“The ‘production’ cost of depositing a federally funded research article into a free public access repository can be, conservatively, as low as $15 and even lower under a federally owned and managed repository such as PubMed,” the economic impact statement said. That makes public access to federally funded research sound inexpensive and simple, but an official PLOS (Public Library of Science) blog offering transparency on finances suggests that sharing research requires more than posting a PDF.

“Many of our mission-driven publishing activities go well beyond peer review and production services,” the blog asserts.

The economic impact statement also failed to estimate the cost of the part of the new policy that mandates open data. That is, scientific data underlying federally funded research, including data “not associated with peer-reviewed scholarly publication”—an expansion in scope of current data-access standards—will need to be freely available and publicly accessible by default. This piece of the mandate will allow researchers to build on each other’s work and save time, but it will not come free.

Not everyone agrees with the government’s back-of-the-napkin calculations. Angela Cochran, vice president of publishing at the American Society of Clinical Oncology, and others have argued that the economic impact statement “heavily relies on flawed data provided by lobbyists for open access policies.”

“Article processing charge averages will not stay at $2,500-3,000,” Cochran told The Scholarly Kitchen. “Not because [those charges] will be artificially inflated, but because for highly selective journals, the [article processing charge] model doesn’t work at that price point.”

The impact statement may be incomplete and contested, but it makes clear that American taxpayers already support academic publishing in at least five ways: direct funding of the research, taxpayer-funded researchers can pay to publish, taxpayer-funded libraries pay for journal subscriptions, taxpayer-funded researchers peer review articles and taxpayers pay to access content behind paywalls. Still, some argue that supporting research, but not public access and data stewardship, is irresponsible.

“Funders hold the stick: they should disburse no further funding without a properly reviewed and budgeted data-stewardship plan,” Barend Mons, a biosemantics professor at Leiden University Medical Center and leading FAIR data specialist, wrote in a Nature opinion piece.

For now, the government is focused on how publishers might step up. For example, publishers that voluntarily made an exception to provide access to COVID-19-related research, according to the economic impact statement, “must” make timely “public access the new norm for all federally funded scientific research.”

That falls short of offering a detailed accounting for how the new mandate might be funded, but it does draw a line in the sand on behalf of federal agencies, including the six—the National Institutes of Health, the National Science Foundation, the Department of Defense, the Department of Energy, the Department of Agriculture and NASA—that account for more than 94 percent of the approximately $150 billion in federal research and development funds, according to the impact statement.

“The publishers are going to need to pay their bills somehow,” Schonfeld said. “They’re going to need to either charge subscriptions or pivot to a more open-access model. That’s the fork in the road … How many of the publishers can develop more mature open-access models?”