Nationalize, Already!

Given precedents, this isn't that hard to figure out. Japanese-style zombie-banks, bad! Swedish-style nationalization, good! We seem to, as Krugman and others keep having to point out, keep coming back to the SuperSIV idea, which is as bad then, despite all the recent acrobatics around it, as it was when the MLEC was first proposed.

I've seen no-one explain that any better than Felix Salmon, in his post Why Nationalization is the Best Alternative, at Portolio.com:

Kevin Drum is a bit like Joe Nocera: he's reluctant to nationalize, but he doesn't really say why.

It's wise to be wary of nationalization. It should be a last resort, and I've gotten a sense recently that a lot of people are talking about it awfully casually. Still, it's true that there are some benefits to nationalization, and one of them is that it allows us to avoid the problem of valuing and buying up toxic assets from troubled banks. If the government owns the whole bank, then the bad stuff can be easily hived off without any kind of valuation at all, and then left to sit for a while before it's sold off -- which is what the Swedes did.If we have to nationalize, then we have to nationalize. But we should understand the precedents before we do, and go ahead only if we have to.

The only argument I can find in here is an argument in favor of nationalization, not against it. Why should nationalization only be a last resort?

Let's work from an ex hypothesi assumption that a certain bank -- let's call it Citigroup -- is insolvent. This is not an unreasonable assumption, given what happened to the likes of Lehman Brothers and Washington Mutual. But I don't want to get into the details of Citi's balance sheet here: I want to ask what we should do if we've already determined that its assets, many of which fall into the "toxic" category, are significantly smaller than its liabilities.

Now the red-blooded American way of dealing with insolvent companies is bankruptcy: either Chapter 11, where the company continues as a going concern, or some kind of liquidation. For a bank, Chapter 11 is pretty much impossible, since you're not going to find anybody to provide debtor-in-possession financing to keep it going. Except the government. And if the government is in possession, then, hey, you've just nationalized the bank.

As for liquidation, that's not an option, because Citigroup is too big to fail. Dumping Citi's trillions of dollars of assets onto the market in a fire sale would depress asset prices worldwide so much that we'd enter a global depression, not just one in the US.

So what about the bad-bank option? The government buys Citi's toxic assets, taking them off Citi's balance sheet, and leaving behind a healthy bank. Sounds good -- except remember that, ex hypothesi, Citi is insolvent. If the government buys the toxic assets for what they're worth, then that doesn't help, since the amount of money that Citi gets in return isn't enough to pay off the loans that Citi essentially took out against those assets. In housing parlance, Citi's underwater on its recourse loan, and when you're underwater on a recourse loan, selling the house at its market price doesn't make you any less insolvent.

So maybe the government deliberately overpays for the toxic waste? That's a recipe for opacity, and it's very hard to systematize. If you're willing to pay 150% of market prices for Citi's bad assets, shouldn't you do that for everybody else's, too? Even perfectly healthy banks which don't need the money? Or do you just decide that Citi, because it's too big to fail, is going to get a big handout which no one else qualifies for? If you do make that determination, why not just go the whole hog and write a check to the bank outright, and put it straight into Tier 1 equity? Oh, wait, you can't do that, because that's called buying equity, and if you spent that much money on Citi's equity, you'd end up with a majority stake in the bank -- which is nationalization.

Essentially, any government purchase of toxic assets can be split into two components: the market price, and a subsidy. If the subsidy is greater than half the market capitalization of the bank, and the government doesn't end up controlling the bank, then there's something very fishy going on indeed.

It's worth bearing in mind here the first TARP proposal, which envisaged the government buying up bad assets at some kind of long-term value price which was greater than the distressed market price. That never happened, the bad assets stayed on the banks' balance sheets -- and then, in the fourth quarter, we saw some absolutely monster write-downs from those loans' end-September marks, including $15 billion at Merrill Lynch alone. You still think that the end-September marks were distressed bargain-basement prices?

Then there's the insurance proposal -- which is cropping up now in the UK after being rolled out in an ad hoc fashion with Citi and BofA here in the US. Robert Peston explains how it works:

Our biggest banks would identify their bad loans and foolish investments. And they would then pay a fee to a new state-backed insurer to protect themselves from losses over a certain level on these stinky assets.But the banks would retain these bad assets on their balance sheets. They would not be transferred to a new toxic bank. We as taxpayers wouldn't own the stinky loans - though we would be liable for losses on them over a certain level.

This has all the same problems of the create-a-bad-bank idea: the government still has to come up with a price (a/k/a expected default rate) for the bad assets, and there will still be a huge implicit subsidy, in many cases greater than the bank's market capitalization, for any institution which takes the government up on its offer. After all, the mark-to-market value of the insurer is certain to be massively negative, otherwise Warren Buffett would have set up something like this already on a for-profit basis.

Finally, the government could take the Irish approach, and target the banks' liabilities rather than their assets. Keep the assets on the banks' balance sheets, and simply guarantee all of their unsecured debts. After all, there's a government guarantee on a lot of the unsecured debt already, and there has been for years: it's called the FDIC deposit guarantee.

This is basically a massive bailout for all the banks' bondholders, who thought they were buying risky leveraged single-A bank debt, and who will suddenly find it backed by the full faith and credit of the US government. At this point, it doesn't matter if a bank is insolvent, because it can roll over its debt indefinitely, since that debt has a government guarantee. Indeed, it should be quite happy to lever up as much as it's allowed, and spend its cheap new funds on all manner of risky assets, since that gives shareholders the best chance of making lots of money and recovering some of the billions of dollars that they have lost. It's akin to taking a man with a large debt, pointing him in the direction of a casino, and telling him he has unlimited credit to try and pay that debt off. The best way for the government to avoid the obvious outcome in such a situation is for the government to take over and run the bank: nationalization. Since the government has an interest in protecting its own liabilities, rather than maximizing shareholder value, the chances of crazy gambles will be minimized. In any case, since the government is taking virtually unlimited downside, it should by rights have all the upside as well -- i.e.,
ownership.

Given how messy all of these alternatives are, why not simply go down the nationalization route? It's transparent and easy to understand: if a bank is insolvent (and the FDIC is good at making those determinations), then simply nationalize it. That's what the Swedes did, and that's what we should do too.

So I'm interested in what Kevin means when he talks about a situation where "we have to nationalize". Does he mean any situation where a too-big-to-fail bank is insolvent? Or are there further criteria he has in mind?