Lindsey Graham, the Republican senator, Alan Greenspan, the former chairman of the US Federal Reserve, and James Baker, Ronald Reagan's second Treasury secretary, are in favour. Ben Bernanke, current Fed chairman, and an administration of liberal Democrats are against. What is dividing them? “Nationalisation” is the answer.

In 1978, Alfred Kahn, an adviser on inflation to President Jimmy Carter, used the word “depression”. So angry was the president that Mr Kahn started to call it “banana” instead. But the recession Mr Kahn foretold happened all the same. The same may well happen with nationalisation. Indeed, it already has: how else is one to describe the actions of the federal government in relation to Fannie Mae, Freddie Mac, AIG and increasingly Citigroup? Is nationalisation not already the big financial banana?

Much of the debate is semantic. But underneath it are at least two big issues. Who bears losses? How does one best restructure banks?

Banks are us. Often the debate is conducted as if they can be punished at no cost to ordinary people. But if they have made losses, someone has to bear them. In effect, the decision has been to make taxpayers bear losses that should fall on creditors. Some argue that shareholders should be rescued, too. But, rightly, this has not happened: share prices have indeed collapsed. That is what shareholders are for.

Yet the overwhelming bulk of banking assets are financed through borrowing, not equity. Thus the decision to keep creditors whole has huge implications. If we accept Mr Bernanke's definition of “nationalisation” as a decision to “wipe out private shareholders”, we can call this activity “socialisation”.

What are its pros and cons?

The biggest cons are two. First, loss-socialisation lowers the funding costs of mega-banks, thereby selectively subsidising their balance sheets. This, in turn, exacerbates the “too big to fail” problem. Second, it leaves shareholders with an option on the upside and, at current market values, next to no risk on the downside. That will motivate “going for broke”. So loss-socialisation increases the need to control management. The four biggest US commercial banks – JPMorgan Chase, Citigroup, Bank of America and Wells Fargo – possess 64 per cent of the assets of US commercial banks (see chart). If creditors of these businesses cannot suffer significant losses, this is not much of a market economy.

The “pro” of partial socialisation is that it eliminates the risk of another panic among creditors or spillovers on to investors in the liabilities of banks, such as insurance and pension funds. Since bank bonds are a quarter of US investment-grade corporate bonds, the risk of panic is real. In the aftermath of the Lehman debacle, the decision appears to be that the only alternative to disorderly bankruptcy is none at all. This is frightening.

The second big issue is how to restructure banks. One point is clear: once one has decided to rescue creditors, recapitalisation can no longer come from the debt-into-equity swaps normal in bankruptcies.

This leaves one with government capital or private capital. In practice, both possibilities are at least partially blocked in the US: the former by political anger; the latter by a wide range of uncertainties – over the valuation of bad assets, future treatment of shareholders and the likely path of the economy. This makes the “zombie bank” alternative, condemned by Mr Baker in the FT on March 2, a likely outcome. Alas, such undercapitalised banking zombies also find it hard to recognise losses or expand their lending.