I'm a bit disappointed that Richard Murphy's answer to this problem hasn't attracted more debate on the Left blogosphere.
He carefully distinguishes between banking functions (like the operation of a national payment system and creation of credit) which have to be centralised, and therefore should be taken over by the State, and those functions, like deposit taking and investment, which might reasonably be franchised out to a number of operators to create some degree of competition and thus generate some consumer choice and 'alternate providers' in the event of any one of them crashing. Strong regulation would also be needed. In essence it is a similar kind of approach to the one New Labour have taken to much of the welfare state. I can't help laughing a bit at seeing it applied, even theoretically, to the banking industry.
But let's review the underlying arguments as to why New Labour use this kind of approach in the Welfare State: foremost amongst such justifications is the theory that without competition organisations are liable to 'provider capture'- that is to being run in the interests of their workforces, not their 'customers' or 'owners'(e.g. the tax payer).
Now this is mainly bollocks of course - but not quite entirely. The fact that in the Welfare State such arguments are used as cover for 'privatisation by devolution' and the creeping invasion of financial privatisation through PFI like arrangements does not mean that 'provider capture' never happens in any circumstances. My previous post waxed indignant on the subject of City bonuses - the size of these would seem to suggest provider capture has happened in that sector.( This would also fit in with Stumbling's ongoing worries about principal-agent problems, I think) In which case there would seem to be grounds for setting up a system designed to discourage it....Co-incidentally, it would be a very British form of socialism, but let's not go shouting that from the rooftops just yet, eh?
I'd be really interested in hearing others' take on this.