I know a lot of academics in the social sciences and humanities. Many of these pieces are handled by free labor in these disciplines as well. As paper copies have become increasingly disused, we are seeing that projects like sci-hub can handle the distribution - let's say that it gets a 500k grant for operating expenses and is in turn run on open source software contributed by academic professors who can count it as part of their community involvement for tenure considerations.
That means we're talking about the editors and copy editors themselves as the majority of the cost. Solve that by creating a university consortium. Instead of paying Elsevier, they pay a non-profit run at a set of universities who are responsible for a pool of editors.
Academia handles so much of this in-house already - there's nothing saying they couldn't do this as well.