From an economics point of view, what they are essentially proposing is to modify the banking system so that it resembles their models.
Their models are stylised, highly simplified, and have to be massively "callibrated" in order to match economic observations at all, and even then can only usually do so between financial crashes. (The part of economics that analyses business cycles hasn't discovered the Nyquist theorem yet.)
The best guess at what would result of this, is all the problems of centralised control, and none of the advantages of the existing system. Another way to think about it is a bunch of complete maniacs being allowed to run experiments on the live production system without any testing.
Although hopefully it won't come to that.