The Research Information Network (a very worthy organization which, consequently,will probably be axed under the new coalition government regime) has produced An introduction to open access. As is so often the case, the guide (which is intended for researchers) fails to distinguish effectively between true open access, where there are neither subscription costs nor author charges and partial open access, which involves author charging. Both of these go under the label of gold access, following a distinction suggested years ago by Steven Harnad. But these are not identical methods of achieving open access, which is why I describe true open access as Platinum access. So-called "gold" simply transfers the production costs to the author, from the subscriber; "platinum" calls upon neither of these sources for finance but relies upon either subsidies or voluntary work, or a combination of the two.
Governments, universities and research funding agencies around the world have been slow to see the potential of the platinum route - largely it seems (at least in the case of governments) to protect the publishers. The exceptions are countries in the smaller language groups, where publication of scholarly research has generally been through journals produced by universities, produced with subsidies, and exchanged around the world for other journals to reduce library acquisition costs. In such cases, the transfer to free, subsidised open access has been quite logical and simple to achieve.
The economics of scholarly publishing undoubtedly support the platinum route since social benefit is maximised in this way. It is possible, although I hold out little hope, that the current financial crisis in the UK will lead institutions to embrace the platinum method: individual universities, or, better, collaboration between universities would enable the publication of journals in specific disciplines which could be funded at very low costs, and universities could require researchers to publish in these journals, just as they mandate the deposition of papers in institutional archives.
How low are the costs? Well, taking Information Research as an example of the platinum journal, the only direct costs of production are borne by the University of Lund and those costs amount to whatever proportion of server maiintenance costs can be attributed to the journal: I imagine that these costs are rather low. All other costs: editing, copy-editing, reviewing, layout, production, are borne by the voluntary workers of the journal. Is there really any economic case to answer?