Balance sheet sensitivities July 27, 2010 at 7:30 am
FT alphaville has an interesting post on the Spanish stress test results. They quote Deustche bank research as follows:
We regard the impairment assumptions (both on sovereign and the different credit buckets) as sufficiently severe and consistent. We are, however, less convinced by and have lower visibility on what is included in and how the regulator arrived at some of its forecasts on PPP [pre-provisioning profit], capital gains and other “impairment buffers”. This is particularly relevant as under a marginally tougher set of assumptions in this area, a larger number of institutions would have failed the test (nine instead of five, with another six below 6.5% Tier 1), although admittedly the incremental amount of capital is limited to EUR 3.5B.
The point is that the stress test results are rather sensitive to these PPP estimates. As Alphaville says:
Reducing the PPP forecasts by 20 per cent leaves nine banks below the magic 6 per cent Tier 1 capital ratio used in the tests
Now it’s easy to use this as evidence of stress test fudging, and I won’t bother to do that – the tests have been analysed enough elsewhere, and the markets have their own view. Instead I’ll make a broader point: wouldn’t it be nice in financial statements to get some sensitivity analysis of key variables? That is, instead of getting a static book of a financial statement, with all the numbers hard-wired, what I’d like would be a little model. The default conditions of the model would be the reported financials. But then you could vary a few of the key parameters – loan loss provisions by book, say, or funding cost – and see the impact on the financial statements. That would really give investors a much better idea of what the company’s risks were. For that reason, I suspect it will not happen in my lifetime.
