Derive This
May. 31st, 2006 08:33 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
peterbirksThe Bank of France said yesterday that the tools being used by banks to measure risk were now "obsolete".
Well, I've been wittering on about this for years, so I suppose that it's nice for at least one Central Bank to fall into line behind me.
Thinking poker players are, of course, well-versed in the land of what, for want of a better phrase, we shall call the "one-outer". So, let's compare the thinking banks with those players who know that, every so often, you will be hit by not one, not even two, but by something apparently impossible like three one-outers on the trot. Now, let's compare the vast majority of the smaller banks in the world with the vast majority of poker players, who seem utterly amazed that even one one-outer can occur. As soon as three appear on the spin, they will be shouting "impossible!" and "conspiracy" faster than their opponent can trouser the money.
Anyway, our good old Bank of France isn't talking about the one-outers -- not yet, anyway. It's talking about the fact that many of the investments held by the more innocent banks in mainland Europe entail risk which they haven't even measured or, if they have, have failed to measure properly. In other words, they have no idea whether their "opponent" is drawing to one out, two outs, or an open-ended straight flush draw. In particular, the Bank reckons that the derivatives in which a number of these innocents-by-night have invested are rather more dependent on low interest rates than they realized.
All that the BoF can do about it is urge the banks to tighten their stress test models, and then to publish the results so that the potential dangers can be seen.
Yeah, some chance of that happening.
Although the products that have been designed can be hideously complicated (deliberately so) the problems are fairly simple. They tend to come down to two things:
1) a lack of appreciation of correlation between two events, and
2) a lack of appreciation that an apparently liquid market can get very illiquid if everyone wants to do the same thing at once.
This last point is, it appears, finally getting appreciated by at least some people. The standard time frame for liquidating a portfolio under the "value at risk" measuring system is, wait for it, 10 days. That kind of timeframe would have Warren Buffett choking on his latest petty cash acquisition (say, $1bn). There are some portfolios out there (for example, SocGen's) which, if they had to liquidate them in 10 days, would send several derivatives markets into a tailspin reminiscent of the crashing plane in Lost.
In fact, if it happened, a desert island might be where you want to be. You never hear them worrying about a possible rise in interest rates, do you?
Well, I've been wittering on about this for years, so I suppose that it's nice for at least one Central Bank to fall into line behind me.
Thinking poker players are, of course, well-versed in the land of what, for want of a better phrase, we shall call the "one-outer". So, let's compare the thinking banks with those players who know that, every so often, you will be hit by not one, not even two, but by something apparently impossible like three one-outers on the trot. Now, let's compare the vast majority of the smaller banks in the world with the vast majority of poker players, who seem utterly amazed that even one one-outer can occur. As soon as three appear on the spin, they will be shouting "impossible!" and "conspiracy" faster than their opponent can trouser the money.
Anyway, our good old Bank of France isn't talking about the one-outers -- not yet, anyway. It's talking about the fact that many of the investments held by the more innocent banks in mainland Europe entail risk which they haven't even measured or, if they have, have failed to measure properly. In other words, they have no idea whether their "opponent" is drawing to one out, two outs, or an open-ended straight flush draw. In particular, the Bank reckons that the derivatives in which a number of these innocents-by-night have invested are rather more dependent on low interest rates than they realized.
All that the BoF can do about it is urge the banks to tighten their stress test models, and then to publish the results so that the potential dangers can be seen.
Yeah, some chance of that happening.
Although the products that have been designed can be hideously complicated (deliberately so) the problems are fairly simple. They tend to come down to two things:
1) a lack of appreciation of correlation between two events, and
2) a lack of appreciation that an apparently liquid market can get very illiquid if everyone wants to do the same thing at once.
This last point is, it appears, finally getting appreciated by at least some people. The standard time frame for liquidating a portfolio under the "value at risk" measuring system is, wait for it, 10 days. That kind of timeframe would have Warren Buffett choking on his latest petty cash acquisition (say, $1bn). There are some portfolios out there (for example, SocGen's) which, if they had to liquidate them in 10 days, would send several derivatives markets into a tailspin reminiscent of the crashing plane in Lost.
In fact, if it happened, a desert island might be where you want to be. You never hear them worrying about a possible rise in interest rates, do you?
