Is there something obvious I'm missing as to why journals can't have reviewers on staff? Charge the cost as a submission fee and lower the cost for submissions from developing countries. You might end up with a fee of $200 per article, but that would easily be worth not having to wait months for a response. Plus, there would presumably be better feedback, too.
I will submit a (fairly long) paper sometime this month. For this paper:
1) I don't think there are 100 people in the world who are competent to review this paper. A generalist journal in my field would need at least several hundred reviewers to cover everything.
2) Reviewing this paper is at least a full week of work. If the reviewer makes a salary comparable to that of a mid-career academic at a decent research university, we are talking over $1500 for a week, and that's for a fairly poorly paid academic and not counting benefits or various indirect costs.