In an earlier post, I took Kevin Drum to task for referring to partially state-owned, but publicly listed Nordbanken, as "the state bank". I noted then that I didn't know what percentage of the bank was state-owned.

Kevin Drum is not lazy. He's managed to dig that up. In an e-mail, he reports that at the time of Nordbanken's nationalization, it was 77% state-owned, and private shareholders were bought out at better than the market price. At 77% government ownership Kevin's characterization of Nordbanken as a state bank seems pretty defensible. (That the Swedes bought out the old shareholders doesn't interest me very much, as long as the bank's price had already collapsed. For example, paying Citi's shareholders a substantially above-market $4 per share — Citi closed at $3.11 today — after the shares have fallen in value by more than 90% wouldn't much change the incentives of pre-crisis shareholders.)

"Nationalizing" a bank already 77% owned by the state might not seem like a big deal. But, reading an insider's account of the Swedish crisis, it was a big deal:

The crisis continued. We had to make a big quick fix for Nordbanken. However, it was soon clear that the quick fix was not enough. So we decided to nationalize the bank and recapitalize it. Nordbanken was very large, as its asset base equaled 23 percent of GDP. The initial cost of recapitalizing Nordbanken equaled 3 percent of GDP. A few years later we were able turn it around at a profit for the taxpayers and that transaction, more or less, paid for the banking crisis.

The restructuring of Nordbanken was really important in that it served as a showcase for the rest of our work. It demonstrated the government's determination to address and resolve the crisis and it helped us to gain respect.

Britain is in a similar situation today. The British government already owns 68% of RBS, but the question of whether or not to "fully" nationalize the firm remains important. Why? What's the difference between a "full nationalization" and majority ownership? Does "full nationalization" matter?

I think it does matter, quite a bit. First, there is the obvious matter of unity of control. A fully nationalized bank can be reorganized in the public interest, e.g. by division into smaller firms, without minority shareholders complaining or even suing on the theory that a different structure would be more profitable.

More importantly, only full nationalization eliminates investors' incentives to concentrate capital in "too big to fail" banks before the crisis. Assume that a bank is insolvent, such that if it were not "too big to fail", regulators would insist it merge or wind down at a cost the the deposit insurance fund. But the bank in question is too big to fail, so regulators cannot wind it down. The government is then forced to become the capital provider of last resort. Existing shares would be worthless under this scenario, if the bank had no power to threaten a chaotic failure. However, after a recapitalization, the reorganized bank might become a very valuable. The bank retains its existing network of branches, benefits from the deposits and habits of its old customers, and may leap from sickest bank to safest bank with a single "bold" injection of government capital. If the old shareholders are permitted to ride along after the reorganization, they reap a large reward from having invested in a bank that was too important to fail.

Suppose that a bank whose true book equity is $0 has failed to mark down some assets, and shows a position of $10B. The bank receives a $90B capital injection, valuing existing shares at book. Then the old equity whose true value was precisely zero prior to the recapitalization suddenly has a real book value of $9B. That is, old shareholders reap an immediate windfall from the recapitalization, and the size of the windfall increases in direct proportion with the amount to which management had lied about the banks losses! Further, the market value of old equity should rise by much more than that $9B, since all of a sudden, the bank switches from a horseman of the apocalypse to a going concern with a bright future. Failing to exclude old shareholders from a post-recapitalization bank results in a transfer of wealth from taxpayers to shareholders in proportion to the degree to which i) they invest in "too big to fail" banks, and ii) encourage management to understate asset impairments in their books. Those are really bad incentives. The details of this story change a bit if, for example, the recapaitalization comes in the form of preferred equity with warrants, or if market values rather than book values are used to estimate the how much dilution old shareholders suffer. But the core bad incentives do not change.

By eliminating private shareholders entirely, full nationalization permits regulators to "do what needs to be done" to restructure the firm without having to hew to a fiduciary duty of profit maximization in designing the new structure. Full nationalization limits the ability of shareholders to extract windfalls from taxpayers by becoming "too big or interconnected to fail". Finally, full nationalization makes it possible to value assets ruthlessly, thereby eliminating market uncertainty about whether a bank is really fixed. Either to maximize their share of a recapitalized firm or to maximize the subsidy in a "toxic asset" purchase, legacy shareholders will always insist on optimistic asset values. But getting past a banking crisis requires working from an assumption of extremely pessimistic values.

So I do think that Nordbanken still "counts" importantly as a nationalization (and that AIG, for example, remains importantly undernationalized). But kudos to Kevin Drum for unearthing the public/private split. What do you think? Does a shift from 77% government-owned to 100% government-owned really matter?