Jan Velterop schreibt in liblicense zum Übergang auf 100% Open Access: OA wird für stark forschungsorientierte und publizierende Universitäten wie z.B. Cornell teurer als das bisherige Subskriptionsmodell, für eher lehrorientierte Universitäten dagegen deutlich billiger:
Assuming that the total amount of money involved in the aggregate remains the same, redistribution of costs has the important academic and societal benefit of enabling full open access. Given that the funders (mostly governments) inject this money into the system anyway, this could be a winners-only game, the funders, academia, and society as a whole being the winners.
In derselben Post zitiert er die William H. Walters-Studie Institutional Journal Costs in an Open Access Environment:
Because institutional disparities in publishing productivity are far greater than institutional disparities in library holdings, the shift from a subscription-based model to either Open Access model would bring dramatic cost savings (greater than 50%) for most colleges and universities. At the same time, a small number of institutions — the top research universities — would pay a far higher proportion of the total aggregate cost.
Das hatte ich auch schon mal anhand von Nucleic Acid Research berechnet.