Wednesday, April 22, 2009

IMF and Table 1.8 gate

There was a predictable media flurry on Tuesday night as the IMF mysteriously backtracked on its £200bn estimate on the cost of bailing out the UK’s banks, as quoted earlier in its Global Financial Stability Report.

On the box, Newsnight’s Paul Mason exclusively revealed the IMF had withdrawn the original figure at 9pm on Tuesday after the British Government informed him personally that the IMF had “made a drafting error”. The £200bn figure, reflecting 13.4 per cent of UK GDP, was actually the top end of an IMF range of between 6 and 13 per cent of economic output - the government’s argument being that it would have been rather more accurate to give the full range as a reflection of potential cost.

Basing the retraction on the word of the British government ahead of the UK budget, Mason went on to say that the IMF would not answer his calls. But in response the IMF had removed the original graph showing the 13 per cent figure from the PDF file on their website, “so there’s currently no table.”

All of which caused Newsnight host Jeremy Paxman to exclaim:

This is incompetence of a staggering order from the IMF!

Now, while there is obviously some embarrassment for the IMF here, the mistake might say rather more about the UK Treasury’s sensitivity to big bad numbers, rather than the competency of the IMF.

Firstly, here in Excel form is the original untampered IMF table (still available here):

IMF stabilization costs original

And here’s how the table changed overnight:

IMF Table 1.8 - GFSR

As can be seen, the IMF now puts the UK figure at 9.1 per cent, nicely in the middle of the government’s 6-13 per cent range. But note too how the estimates for the US and Canada have changed too.

The BBC’s business editor Robert Peston aka ‘Pestowire’ soon took up the case, issuing the following treasury release on Pestowire:

Specifically on the UK, the IMF estimates that the costs to taxpayers (or us) of bailing out our big banks will be 13.4 per cent of GDP or around £200bn, rather more than the Treasury has been estimating or will factor in to tomorrow’s budget. On the IMF’s figures, only Ireland will suffer greater taxpayer costs as a proportion of GDP.

And then an update before the fresh table above was published:

UPDATE 00:05 The Treasury has shouted very loudly at the IMF. And the IMF has tonight withdrawn from the online version of its Global Financial Stability Report the table showing the costs to the British taxpayer of the bank bailout as being 13.4 per cent of GDP. That table is now, according to the IMF, “embargoed” - whatever that means.

The Guardian offered some clarity a little later:

Europe and Japan between them account for $1.3tn of the write-downs, with UK banks facing losses of up to $316bn (£216bn). The Treasury last night disputed the UK figure, saying the IMF had offered a range of costs between 6% and 13% of GDP, and that £216bn was at the highest end of this range.The real point here is that the Treasury simply resented the fact that a top-of-the-range figure was being treated in headlines as the likely cost. But that doesn’t mean the top-end IMF estimate is wrong. Indeed, given the alarming rise in forecast losses elsewhere across the globe, such a range is probably quite conservative. (By way of illustration, examine table 1.3 here - expected losses on loans and securities in the US have all but doubled in the space of six months.)

Perhaps chancellor Alistair Darling will offer some further clarity when he presents the UK budget at 12.30 BST. Then again, perhaps not.

It’s worth noting though that the IMF has as yet not explained the reasoning behind the adjustments.

No comments: