Saturday 24 July 2010

Those Stress Tests. Technical View. Yawn

Well, they've laboured mightily and brought forth the expected mouse. For political reasons it wasn't possible for the Committee of European Banking Supervisors to run the only stress-test that counts right now, namely the shock to a bank of an actual sovereign default. So it's all as feeble as anticipated, though some good may come of there being a whole lot more numbers out in the open.

What they did instead was to impose 'haircuts' on the values of banks' bond-holdings. Another criticism that has been leveled at the process is that this only applied to trading accounts, not 'hold-to-maturity' assets. As it happens I don't necessarily agree with this line of attack: but the defence I could offer only serves to point out another weakness.

The reason why shocks should be used in stress-testing, rather than assumptions of (relatively) slow-time degradation of market / economic conditions - which is all a haircut represents - is because if an entity has a reasonable amount of time to defend itself, it (along with every other actor) will re-optimise its position as conditions evolve. It is impossible to model with any certainty what this would mean, but in most cases it will result in a much better outcome than passively taking the hit. That's how individuals, firms, sectors and whole economies adjust to, and mostly survive, tectonic shifts in conditions.

That could be given as a reason for omitting held-to-maturity assets from the haircut procedure.

But it simply highlights the need for proper shocks to be applied ! Amateurs ...



Sackerson said...

A more worldly view would be that the terms of the stress tests have been chosen to produce a result that will reassure an ignorant public.

Andrew B said...

A sort of relevant comment at the end of the buttonwood column about gold in the economist.


That prospect did alarm creditors in the 1980s when the real yields on government debt shot up. But it does not seem to now. America and Britain are paying only 3-3.5% to borrow for ten years. That may be because deflation seems the more immediate threat. It may be because bond markets are now dominated by other central banks, which are more interested in managing exchange rates than in raising returns. But it is not stable to combine low yields, high deficits and governments that are happy to see their currencies depreciate. Something has to give.

Combine that with the fact that over the last 10 years, you were better off buying gilts rather than equities.
Add in that from the same issue of the economist there is a criticism of the underlying DSGE based modelling tools used by central banks as they do not make allowance for bubbles / sudden corrections in unexpected places. I bet they used the same type of methods on the banks.

You can gain a sense that the real stress test is coming, just when and what it will stress and test may not be what you think.

James Higham said...

namely the shock to a bank of an actual sovereign default

Precisely. Those stress tests were orchestrated anyway, to produce a rosy picture of tentative rebirth but unfortunately, numbers are numbers and as you say, the sovereign test is the biggie.

Andrew B said...

I think it will be the things we are not looking at.

Tinfoil hat time.

Most of the USA's trade is internal - they sell to each other.

How about in 5-10 years time, China decides that they have enough of a middle class to largely support their own progressing industrialisation and they only really need the outside world for raw materials.

- just like that other country.

As part of that process they decide that there are a lot of benefits in their currency being a reserve currency - they denominate rare earths in rm, then later move on oil (no, they don't have to float the currency freely any more than they do now).

They simply do not repurchase US bonds as they come to maturity and reinvest in something else.

This is almost certainly wrong, but, all those who say 'Greece' or 'Spanish banks' are probably wrong too.

Nick Drew said...

always open to wordly views, Sackers, James, & the tests were certainly chosen politically (small 'p') but in this instance I'm not sure the public gives a toss

Andrew - a good scenario, if not a shock ... like you say, it'll be something we are not expecting: the problem with stress-tests is the testers' lack of imagination - they can only contemplate the worst that's happened so far

the Economist's summary of the ST's is:

they were needed to restore the delicate strands of trust without which the world’s financial system would not operate. And on this measure they are likely to be a qualified success, in that they remove the question-marks that the market had placed over many institutions. The success is, however, only qualified because a number of borderline banks passed so miserably that those are unlikely to regain the confidence of markets

removed question-marks - really ?

I'd say the benefit is extra data (which isn't to be sneezed at)

Sackerson said...

Nick: "I'm not sure the public gives a toss" - I disagree. They'll get the good news retailed to them by the middlebrow press and it'll reassure them that they don't have to "put all your money in a paper bag inside the zoo, the thing to do", as the Fab Four put it. And so it might help reduce the degree of cold fury still felt by large sections of the populace at the sight of fat cat bankers making everyone else pay while they still get their bonuses.