What open access publishing means for libraries

Open access publishing means research papers are free to read online, but for university libraries it means the bill arrives differently: instead of paying subscriptions for journals, they increasingly pay article processing charges, known as APCs, that range from roughly 2,000 to more than 10,000 dollars per published paper at the largest commercial journals. Libraries now manage this transition, and the numbers explain why many call it a budget crisis in slow motion.

This site publishes information about scholarly publishing economics, not legal or procurement advice.

How did open access change who pays?

Under the subscription model, a library paid a publisher for access, and anyone outside the subscribing institution hit a paywall. Under gold open access, the published paper is free to everyone, and the cost moves to the author's side as an APC, usually paid from grant money or institutional funds. The argument, made influential by the 2003 Budapest-originated movement and formal declarations such as Berlin and Bethesda, was that publicly funded research should be publicly readable.

The shift accelerated when funders got involved. Plan S, launched by a European research-funder coalition in 2018 and effective from 2021, required that work those funders pay for be immediately open. In the United States, the White House science office's 2022 memo, known as the Nelson memo, directed federal agencies to make federally funded papers freely available without embargo, with requirements phasing in by the end of 2025.

What do APCs cost, and who pays them?

List prices published by major publishers in recent years show a wide spread: hybrid journals, subscription titles that also offer an open option, commonly charge several thousand dollars, and the largest multidisciplinary journals have listed charges above 10,000 dollars per paper. Pure open-access publishers often charge in the low four figures. For a university producing thousands of papers a year, even the low end compounds quickly, which is why libraries began negotiating.

What are transformative agreements?

Since around 2019, large consortia have signed read-and-publish deals that bundle subscription access with the right for their authors to publish open in the same publisher's journals at no separate charge. Germany's DEAL consortium struck an agreement with Wiley in 2019 after negotiations with Elsevier collapsed; the University of California system, after canceling its Elsevier subscriptions in 2019, reached its own landmark agreement in 2023 following a month-long journal boycott by UC authors. These deals move money, sometimes slightly more of it, from subscriptions to publishing, and they are meant to be temporary bridges toward full open access.

Why can't libraries just cancel subscriptions?

Because the incentives lock in. A handful of publishers, with Elsevier's parent RELX the largest, control more than half of published scientific papers by most estimates, and their most-cited journals remain places researchers need to publish for careers. Canceling a big deal saves money but cuts off the campus. The 2019 University of California cancellation was possible only because of a systemwide coalition and author pressure; individual libraries rarely have that leverage alone. Librarians describe the result as double dipping, paying subscriptions and APCs at once, during the transition years.

How do we know what the transition costs?

The main evidence comes from libraries' own budget disclosures, published price lists, and analyses by groups such as the SPARC coalition and researchers who study scholarly publishing economics. Studies published in the 2010s and early 2020s estimated that APC-based publishing at current list prices could cost the system as a whole as much as or more than subscriptions, depending on how hybrid charges are unwound. The limitations are real: publisher discounts are confidential, so public list prices overstate what large consortia pay, and open-access uptake differs sharply by field, with biomedicine far ahead of mathematics or the humanities. No neutral dataset of actual payments exists, which itself is a finding about market transparency.

What does this mean for readers and authors?

For readers, the change is mostly good news: a growing share of new papers, especially from federally funded U.S. work after the 2025 policy deadlines, is legally free to download. For authors at institutions without publishing support, APCs can be a barrier, a concern equity researchers have documented in studies showing authors from lower-income countries publish open less often when fees apply. Fee waivers exist but are patchy. The system's next question, already being debated by funders and libraries, is whether paying to publish can be replaced by diamond models and shared infrastructure, so that free-to-read stops meaning pay-to-publish.

What about green open access and institutional repositories?

Green open access sidesteps the fee question: authors deposit versions of their papers in repositories, subject to publisher embargoes that commonly run six to twenty-four months. Universities host thousands of such repositories, and the 2022 federal memo strengthened their role by requiring free access to submitted or accepted versions of funded work. Repositories are cheap relative to APCs, which is why library associations favor them, but they carry their own costs, staff, preservation infrastructure and compliance tracking, and the published version of record still sits with publishers. The Nelson memo's choice to accept accepted manuscripts rather than require published versions is widely read as a compromise that kept this two-track arrangement in place, to the visible frustration of open-access advocates and the visible relief of the publishing industry.