A couple of weeks ago Heather Morrison had an entry in her blog on the economics of peer-reviewed repositories vs. subscription-based publishing. As many have noted before, the economics of the situation are obvious: the former approach would be much cheaper for institutions than the latter. However, I don't see it happening. Sixteen years ago, when I started Information Research it was so blindingly obvious that academics could create and publish their own journals at relatively modest cost that I assumed that in ten to fifteen years, open access would be the norm. It isn't, because a number of things get in the way.
First, few (and increasingly fewer) academics have the motivation and the time to start up new journals - and yet new journals are being created continually and edited by the same academics, with contributions reviewed by the same academics. In other words, they have time and motivation to work for publishers, but no time or motivation to work for their academic community. I don't see this changing since, before all else, humans are driven primarily by self-interest.
Secondly, it has so far proved impossible to get the message across to university administrators that the present system costs them money that could be redirected to better use. Essentially, the idea is too radical and if vice-chancellors, rectors, etc. are any one thing, it is not radical. They'll happily shuffle around departments and create new faculties or disband them, but ask them to take a really critical look at the present system of scholarly communication and its alternatives and they'll shuffle back into their holes.
Thirdly, governments everywhere are at the beck and call of business. If a business sector tells the minister that a move of OA will cause the loss of n thousand jobs, the minister will rapidly back off, whether the business proposition is true or not. Faced by a determined business lobby, ministers are wimps. In any event, certainly in the UK, none of them has any knowledge of the academic research process and scholarly communication.
So there we have it: no drive from below, no support in the middle, and apathy and capitulation to the forces of the market at the top.
The situation is not helped, of course, by the recent crop of dodgy 'publishers', relying on the desperation of those who cannot get their work published in the established journals to pay for inclusion in a new international journal of this and that. The number of OA journals as recorded in DOAJ may be ever increasing (standing at 7,431 today), but a significant proportion of those are author charging journals, and not 'open publishing' or 'platinum' model, and I doubt whether 10% of the remainder will be around in another ten years.
Ultimately, I believe, the present subscription/author charging model will collapse, but, sadly, it is unlikely to do so as a result of the actions of academics.