I agree in theory, and there are many many many examples of this practice arrangement currently in play.
However...we should never, ever, underestimate the level of incompetence of organizations and people in charge can reach. Just...hall of fame level idiocy.
The "correct" way to play the game (in my opinion) is for a hospital to join up with an academic system in some kind of joint venture or true network integration. Whatever gives the hospital the leverage to negotiate better contracted technical rates, as well as enter into the "too big to fail" realm.
The downside to that arrangement would be if the individual hospital takes a "bad" deal in the JV, or if "the system" takes too big of a piece of the network pie, so even though they're getting better rates, the skim makes it not worth it.
The next best would be as
@RickyScott described - a mutated version of a PSA where the university medical group staffs a non-network hospital and pays a fee to do so. The university gets access to that patient population and the hospital has a presumably stable staffing arrangement with a deep bench for vacation and call coverage.
The dumbest arrangement, and hopefully uncommon, would be a hospital staffed by a medical group in a PSA arrangement but without "paying a fee" or some other benefit. Well...I guess it's not fair to make a blanket statement like that. It's the "dumbest" if the academic medical group and the hospital don't "actively manage" the partnership, because each side is constrained by organizational red tape that can lead to suboptimal outcomes.
Now...what's good for "the hospital" and "the university medical group" - those are VERY different things than "what's good for the individual doctor".
None of this is generally good for the individual doctor, unless you want to have zero responsibility on the practice management side and are OK with paying a tremendous cost in doing so.
(which is totally ok of course, many docs don't want to have to worry about any business things)