Open access is seen by many as the publishing model that is most in keeping with the egalitarian ethos of academia.
But a paper by two economists suggests that "gold" open access, under which the author pays for publication and the article is made freely available, could be costly for research-intensive institutions, while benefiting those that do little research.
Richard Watt, associate professor of economics at the University of Canterbury in New Zealand, and Frank Muller-Langer, senior research Fellow at the Max Planck Institute for Intellectual Property and Competition Law in Germany, claim in their paper "Copyright and Open Access for Academic Works" that under the current system, many institutions pay for access to a similar set of journals, and so have similar costs.
If this was replaced with a gold open-access model, the costs incurred would be in proportion to the number of papers published, so the most research-active institutions would pay a lot, while the less productive would pay little.
Despite this disparity, all would have equal, free access to the papers.
Analysing the costs for a set of economic journals, the authors estimate that the break-even point at which costs under both models are equal would be a fee of $72 (£44) per paper for Harvard University, but $246 for the London School of Economics and $2,523 for the University of South Carolina.
So, for example, an average fee of $1,250 per paper would be very expensive for Harvard and the LSE compared with the current system, but much cheaper for South Carolina.
The authors, who presented their findings at a conference in Germany last week, illustrate the point with a graph showing the break-even publication fees for 186 universities whose academics published at least six articles between 1991 and 2005 (a subset is reproduced above).
"The fact that active research institutions would cross-subsidise less-productive institutions under an author-pays model suggests the need to consider alternative pricing schemes for open-access publishing besides publication fees," the paper says.
The authors suggest several solutions, such as the option of a fixed fee so institutions could "insure" themselves against the risk of paying over the odds in publication fees.
A spokeswoman for the Million+ group of new universities in the UK argued against varying publication fees according to the number of papers published.
She added that since most research-intensive universities benefit from "huge" research grant income, publication costs could be factored in to funding bids, as is done with conference costs.
The European Commission has said that it sees open access to research findings as a key factor in supporting economic growth.
Other proponents argue for open access on the grounds that it reduces corporate profits made from the unpaid labour of academics, and promotes freedom of information.
Most higher education institutions and research institutes now have electronic repositories of their academics' papers (sometimes called "green" open access) freely available online.
However, the repositories are generally viewed not as a replacement for the current model but as a concession by publishers keen to maintain the status quo.
Partial solution: Open-access levels near 10 per cent
There are currently about 7,000 open-access journals, and the number is growing by several hundred a year, according to the Directory of Open Access Journals.
However, scholarly articles available via open access presently account for less than 10 per cent of the 1.5 million published each year, according to the Publishers Association in the UK.
Publication fees charged to authors by major open-access journals range from $1,250 to $3,000 (£763 to £1,832) per article, although lower fees apply to some journals with lower circulation.
The rising discontent with the status quo in the UK was underscored by the recent news that Research Libraries UK has told Elsevier and Wiley-Blackwell that it will not renew current "big deals", which give its members access to the publishers' entire portfolios, unless the member institutions receive price reductions.
Graham Taylor, director of educational, academic and professional publishing at the Publishers Association, told Times Higher Education last month that all publishers "aspire to universal access" but that it would take time to implement a "sustainable, scalable, funded solution" to achieve it.