Hilzoy makes a market based case for the nationalization of banks that would otherwise vanish into insolvency without government intervention (entire post well worth the read):
In the case of the large banks, I assume that we do not want them to go bankrupt not because it would hurt their shareholders, but because their bankruptcy would have broader systemic effects that we find unacceptable. That's fine. But in figuring out what to do about that fact, we need to try to preserve the incentives that bankruptcy normally provides.
To my mind, this means that we should proceed as follows. First, figure out exactly what it is that makes letting these firms declare insolvency such a bad idea: what effects we are trying to avoid. Second, try to craft a policy that avoids this particular bad consequence, while leaving the other disincentives to go bankrupt (or to invest in firms that are at risk of bankruptcy) in place. Third, if we can't do that, try hard to create incentives that mimic the operation of the normal market incentives that our actions are preventing. (E.g., if we prevent banks from declaring insolvency, we need to provide some other disincentive to becoming insolvent, in order to avoid moral hazard.)
This is the main reason why I tend to favor nationalizing those banks that are insolvent, clearing up their balance sheets, recapitalizing them as needed, and sending them back into the private markets as soon as is prudent. I am not, in general, in favor of the government controlling individual banks. But in this case, if we don't want to let the large banks declare bankruptcy, we need to provide some serious disincentives to their managers, investors, and bondholders. (I exempt depositors since I think that they should be insured, given the systemic value of avoiding bank runs.)
Nationalization would accomplish that. It would wipe out the shareholders and holders of unsecured debt, which is what the market would have done if left to its own devices. It would allow us to replace the senior management at the banks, which would give them every incentive to avoid needing to be nationalized. We would need to own the banks in order to do what needs to be done, and to do it as quickly as possible. This would mimic the market by treating the government as an owner in those cases in which it is, in fact, putting up the money: anyone else who provided this sort of capital would get ownership, and making an exception for the government would make government money more attractive than private capital. This would, I think, be a bad thing.
Nationalization would, in short, accomplish what my market principles tell me we should do: specify exactly what the bad consequence is that we want to avoid, and craft a policy solution that avoids this particular bad thing while either leaving other market signals intact or (where this is impossible) mimicking them. It would also allow us to return to what I take to be the right state of affairs (in which banks are private, and privately funded, and the government regulates them) as quickly as possible. (If you don't like excessive government involvement in banking, it's not clear why you'd prefer a long, drawn-out period of heavy government involvement over a shorter period of outright nationalization.)
This isn't rocket science. Models exist which can provide a blueprint for how to proceed. The only limiting factors in all of this are political and ideological. Our current crop of politician's fear the word "nationalization" almost as much as they fear openly contested elections.
