That's only true if all parties have the option to back out: If Bob doesn't like the terms of Alice's loan, he doesn't have to take it. But if Alice grants Bo

That's only true if all parties have the option to back out: If Bob doesn't like the terms of Alice's loan, he doesn't have to take it. But if Alice grants Bob a loan and then later sells the debt to Charlie, Bob didn't get the opportunity to consent to the transfer. Depending on Charlie's behavior, it may suddenly be in Bob's best interest to default on that loan. So it's not necessarily mutually beneficial, in fact it's frequently exploitative. But suppose Alice sells the debt to 1000 people in tiny slices, 200 of which are depended on by Bob in some way, 600 of which are independent parties that neither help nor harm Bob besides profiting on his debt, and 200 of which are working with Charlie in a mining endeavor which is poisoning Bob's water or whatever. It creates an information asymmetry where Bob may not be capable of determining who he is harming or helping if he defaults on the loan (Charlie and his 200 cronies, however, they have accountants on staff and know exactly what they're doing). That's the second problem: Bob can't make an informed choice regarding the real-world consequences of his behavior, so the system becomes blind to those consequences. Bob and the rest of us who live nearby--whose retirement portfolios may have been unwittingly contributing to the problem--still have to deal with the consequences. Had the consequence of holding Bob's debt been known to us, perhaps we'd prefer to take the hit to our portfolios than to deal with the fallout of Charlie's actions.