A year ago, I predicted that the losses of US financial institutions would reach at least $1 trillion and possibly go as high as $2 trillion. At that time, the consensus among economists and policymakers was that these estimates were exaggerated, because it was believed that sub-prime mortgage losses totaled only about $200 billion.

As I pointed out, with the United States and global economy sliding into a severe recession, bank losses would extend well beyond sub-prime mortgages to include sub-prime, near-prime, and prime mortgages; commercial real estate; credit cards, auto loans, and student loans; industrial and commercial loans; corporate bonds; sovereign bonds and state and local government bonds; and losses on all of the assets that securitised such loans.

Indeed, since then, the write-downs by US banks have already passed the $1 trillion mark (my floor estimate of losses), and institutions such as the IMF and Goldman Sachs now predict losses of over $2 trillion.

But if you think that the $2 trillion figure is already huge, the latest estimates by my research consultancy RGE Monitor suggest that total losses on loans made by US financial firms and the fall in the market value of the assets they hold (things like mortgage-backed securities) will peak at about $3.6 trillion.

US banks and broker dealers are exposed to about half of this figure, or $1.8 trillion. The rest is borne by other financial institutions in the US and abroad. The capital backing the banks' assets was only $1.4 trillion last fall, leaving the US banking system some $400 billion in the hole, or close to zero even after the government and private-sector recapitalisation of such banks.

Another $1.5 trillion is needed to bring banks' capital back to pre-crisis level, which is needed to resolve the credit crunch and restore lending to the private sector. So, the US banking system is effectively insolvent in the aggregate. Most of the British banking system looks insolvent, too, as do many continental European banks.

There are four basic approaches to cleaning up a banking system that is facing a systemic crisis: recapitalisation of the banks, together with a purchase of their toxic assets by a government "bad bank"; Recapitalisation, together with government guarantees - after a first loss by the banks - of the toxic assets; private purchase of toxic assets with a government guarantee (the current US government plan); and outright nationalisation (or call it "government receivership" if you don't like the dirty N-word) of insolvent banks and their resale to the private sector after being cleaned.

Of the four options, the first three have serious flaws. In the "bad bank" model, the government may overpay for the bad assets, whose true value is uncertain. Even in the guarantee model there can be such implicit government over-payment (or an over-guarantee that is not properly priced by the fees that the government receives).

In the "bad bank" model, the government has the additional problem of managing all the bad assets that it purchased - a task for which it lacks expertise. And the very cumbersome US Treasury proposal - which combines removing toxic assets from banks' balance sheets while providing government guarantees - was so non-transparent and complicated that the markets dove as soon as it was announced.

Thus, paradoxically nationalisation may be a more market-friendly solution: It wipes out common and preferred shareholders of clearly insolvent institutions, and possibly unsecured creditors if the insolvency is too large, while providing a fair upside to the tax-payer. It can also resolve the problem of managing banks' bad assets by reselling most of assets and deposits - with a government guarantee - to new private shareholders after a clean-up of the bad assets (as in the resolution of the Indy Mac bank failure).

Nationalisation also resolves the too-big-to-fail problem of banks that are systemically important, and that thus need to be rescued by the government at a high cost to taxpayers. Indeed, the problem has now grown larger, because the current approach has led weak banks to take over even weaker banks.

Merging zombie banks is like drunks trying to help each other stand up. JPMorgan's takeover of Bear Stearns and WaMu; Bank of America's takeover of Countrywide and Merrill Lynch; and Wells Fargo's takeover of Wachovia underscore the problem. With nationalisation, the government can break up these financial monstrosities and sell them to private investors as smaller good banks.

Whereas Sweden adopted this approach successfully during its banking crisis in the early 1990s, the current US and British approach may end up producing Japanese-style zombie banks - never properly restructured and perpetuating a credit freeze. Japan suffered a decade-long near-depression because of its failure to clean up the banks. The US, the UK and other economies risk a similar outcome - multi-year recession and price deflation - if they fail to act appropriately.

• The author is professor of economics at the Stern School of Business, New York University, and chairman of RGE Monitor.

© Project Syndicate, 2009,www.project-syndicate.org

Sign up to our free newsletters

Get the best updates straight to your inbox:
Please select at least one mailing list.

You can unsubscribe at any time by clicking the link in the footer of our emails. We use Mailchimp as our marketing platform. By subscribing, you acknowledge that your information will be transferred to Mailchimp for processing.