Who pays when free academic journals change the rules?

A study of 440 academic journals that stopped being free for authors and readers raises questions about commercial incentives, publishing costs and who gets to share their research.

A journal can be free to read and still be too expensive to publish in. For researchers without a publication budget, that distinction matters. Removing a reader’s paywall offers little comfort if the author faces a bill they cannot cover.

Diamond open-access journals avoid that problem by charging neither side. Institutions, libraries and other organisations support the costs of publishing, allowing researchers to submit work without finding money for an article-processing charge.

Yet hundreds of journals have moved away from this arrangement. A preprint by Lisa Matthias, Juan Pablo Alperin and Mikael Laakso identified 440 titles that abandoned diamond open access between 2009 and 2026. Most introduced publication fees for authors.

The findings, reported by Nature Index, might sound like a familiar story about journals struggling to pay their bills. But the pattern suggests something more complicated. Many of the titles changing their model were established journals with years of publishing behind them.

Of the 440 journals identified, 369 moved to an open-access model funded through article-processing charges. Another 52 adopted a hybrid arrangement, combining subscriptions with fees to make individual papers openly available. The remaining 19 moved entirely behind a paywall.

The author charges ranged from US$8 to US$5,300. Professional publishers generally introduced higher fees than research organisations and scholarly societies.

For a researcher choosing where to submit, those differences can determine whether a journal remains a realistic option. A publication charge has to come from somewhere, whether a grant, an institutional fund or another source willing to cover it. For someone finishing a paper after their funding has ended, even identifying who to ask can be difficult.

Established journals are changing course

The team recorded journals leaving diamond open access in every year from 2009 onwards. Most of the changes occurred from 2020, with 2025 producing the highest annual count in the dataset.

There are important limits to what those numbers show. No comprehensive global register of diamond journals exists, so the researchers assembled evidence from several sources. Their list is incomplete, and the study cannot establish what proportion of diamond journals changed model or how many continued successfully.

It also did not investigate the reasons behind each journal’s decision.

Nevertheless, several findings challenge the assumption that these were mostly small, struggling publications forced to charge to survive.

Some 88% remained with the same publisher. Acquisitions therefore cannot explain most of the changes. More than half, 57%, had been operating for at least nine years before leaving the diamond model.

Speaking to Nature Index, Matthias suggested that an established reputation might make a journal commercially attractive to convert. Years of editorial work, published papers and recognition could create an opportunity to introduce charges.

That interpretation needs to sit alongside another result: although commercial publishers accounted for 78% of the changes, publication volumes across the dataset did not rise markedly afterwards. The study does not support a simple account in which every journal introduced fees and immediately began accepting more papers.

Prestige may also play a part. Saray Córdoba, a retired information scientist at the University of Costa Rica, suggested to Nature Index that some journals might regard a more commercial publishing model as a route to greater standing. That remains a possible explanation, rather than a motive established by the study.

What Latin America can tell us

The geographical pattern offers another reason to resist declaring diamond publishing unsustainable.

Relatively few of the identified changes occurred in Latin America, where diamond open access has a strong presence. Universities and publishing platforms such as Redalyc and SciELO provide support that may help journals retain their model.

The researchers suggest that this infrastructure could offer protection against pressure to introduce fees. Whether it can also sustain new journals remains uncertain.

Institutional support brings its own vulnerabilities. Universities and public bodies can reduce spending, and journals dependent on them are exposed to those decisions. But charging authors is not the only way to build a dependable publishing operation.

Nick Shockey, associate executive director of the Scholarly Publishing and Academic Resources Coalition, made that point in Nature Index: the absence of author and reader fees does not mean a journal lacks a robust plan to sustain itself.

A cost researchers need to check

For early career researchers, the immediate lesson is practical. Check a journal’s current charges before submitting, even if a colleague published there for free a few years ago. Establish whether funding is available and what conditions apply before committing to that route.

For funders and universities, the findings pose a harder question. How much support are they prepared to give publishing arrangements that keep costs away from individual authors?

The study documents hundreds of departures from diamond open access. It does not show that the model is destined to fail. Its more uncomfortable possibility is that a journal can build a reputation through years of supported, freely accessible publishing, then become valuable enough to start charging.

By that point, the journal may be well established. The researcher trying to publish their first paper may have far fewer options.

Based on research by Lisa Matthias, Juan Pablo Alperin and Mikael Laakso, available as a preprint on arXiv, and reporting by Dalmeet Singh Chawla for Nature Index, published on 22 September 2026. Expert comments above were reported by Nature Index.