Sean O'Grady: Big banks cannot fail because a depression would result – and they know it

Time was when the Governor of the Bank of England acted as a sort of ambassador for the City to the Government. Mervyn King has definitively drawn that era to a close. Indeed, it seems that Mr King has given up trying to be diplomatic to anyone. The three-way split of oversight between the Bank, the FSA and the Treasury, designed by Gordon Brown, was "inadequate", he says; the bankers cannot be trusted to "play with fire"; they have cost us very dear; and so on.

We have become mesmerised by the City and the contribution it makes to the economy. The calculus, at least as it applies to taxpayers, is less beguiling. In a good year, financial services of all kinds might contribute £120bn in tax to the UK exchequer; yet now the cost to us all of propping up the financial system is eight times that, according to Mr King – close to £1 trillion. And it could easily have been much worse, effectively bankrupting Britain.

So Mr King is right to focus his thoughts like a laser on what the jargon calls the "too-important-to-fail question". It means that banks can even now take almost unlimited risks with their depositors' money and the taxpayers' cash that also funds their activities because they know that there is an informal, but secure, guarantee that even if they collapse they will get bailed out. Most managers will keep their jobs and bonuses will have been spent in the usual conspicuous fashion.

Big banks cannot fail, because a depression would result. And they know it. After the chaos that ensued when Lehman Brothers was left to go under last year, everyone knows it. When things go right, as they have for the likes of Goldman Sachs lately, huge profits are generated and bonuses awarded, but with no payback to the taxpayers who have kept the system afloat and revived markets. Mr King, who sees his responsibilities as stretching across the whole economy rather than just the City and Canary Wharf, plainly finds it annoying.

What to do about this? One answer, to which Mr King seems sympathetic, is to separate the banks' "utility activities" – taking deposits, making simple loans – from their "casino" businesses – inventing derivatives that no one understands, buying and selling shares endlessly to each other or themselves. Such fun and games would not be backed by the taxpayer.

But even if we did this, some of the "casino banks" are so big that they still could not be allowed to fail, because they would owe so many other banks so much money. So we ought to break them up into smaller units, no one of which would threaten the system if it fell over – just like Barings and BCCI did in the 1990s, when no "systemic" damage was caused. That is the logical answer to the "too-important-to-fail question"; but can anyone see the US authorities breaking up Goldman Sachs or Mr Brown splitting Barclays?