I suggest that smaller institutions pay much more per article usage (though they need much of the same research), than larger research institutions and larger institutions may have other reasons to support a system that costs them slightly more-- in order to increase exposure to research as a public good.
[SNIP]