Bad bank

Let us join with the evangelist’s celebration of heaven’s joy at the sinner who repenteth. Even since the financial crisis kicked in, we’ve been banging on about the government’s approach to banking regulation, variously irrelevant as the bonkers Vickers proposals, self-contradictory as schemes simultaneously to recapitalise and to promote lending (help to buy, anyone?), and generally panic-stricken throughout.

Four years ago we argued for “the formation of a bad bank and subsequent work-out, which often proves profitable”. Well at last George Osborne has got the email, a few days ago telling reporters in China that he’s asked Rothschilds to make proposals. This would pave the way for RBS to be floated, with a sale of Lloyds signalled for the off following Royal Mail and the Government’s belated rediscovery of the feel-good properties of privatisations.

The most successful example of bad banking came after the Swedish banking bubble of the early nineties. Duff assets were transferred to a couple of asset management companies which in due course sold them off, often at a profit. The Swiss took a similar approach in 2009, as did the NY Fed which created “Maiden Lane” to take over assets from Lehmans and AIG. Indeed we already have something of a bad bank in “UK Asset Resolution”, which recently sold its portfolio of Northern Rock’s dodgy loans to an American consortium including debt recovery specialists.

So why so long to apply this tried and tested approach to RBS? Maybe it’s scale: the egregious Fred Goodwin was celebrated for his commitment to buying the worst business at the top of the market. Maybe it’s politics: the rediscovered merits of privatisations mentioned above, or the prospect of the Scottish independence referendum next September. In which case, why did Osborne rule out selling a stake in RBS this side of the UK’s general election in 2015? We couldn’t say, other than to note that once sinners take up this repenting habit, it becomes far easier for them to change their minds about other errors.