Slashdot Mirror
The Formula That Killed Wall Street
We recently discussed the perspective that the harrowing of Wall Street was caused by over-reliance on computer models that produced a single number to characterize risk. Wired has a piece profiling David X. Li, the quant behind the formula that enabled the creation of such simple risk models. "For five years, Li's formula, known as a Gaussian copula function, looked like an unambiguously positive breakthrough, a piece of financial technology that allowed hugely complex risks to be modeled with more ease and accuracy than ever before. With his brilliant spark of mathematical legerdemain, Li made it possible for traders to sell vast quantities of new securities, expanding financial markets to unimaginable levels. His method was adopted by everybody from bond investors and Wall Street banks to ratings agencies and regulators. ... [T]he real danger was created not because any given trader adopted it but because every trader did. In financial markets, everybody doing the same thing is the classic recipe for a bubble and inevitable bust."
An interesting article, for sure. The issue with the Gaussian Copula model for pools of mortgages in CDOs is how sensitive they are to the assumptions of the model. If, for example, the annual growth rate of home prices is 2% instead of 10%, things look tremendously different. If correlations between housing prices in different cities is 50% instead of 10% -- disaster. The lack of stress testing of these models (checking what the results are for different inputs into the model) was a huge issue. Even if a model is decent (which in principle, copula models are), if they are too sensitive to inputs, then the prices it produces are not trustworthy. If the proper uncertainty was taken into consideration, then perhaps everyone would have been a little less gung-ho about CDOs.
Like the (worthless) Value-at-Risk figure, the (also pretty worthless in the end) Gaussian Copula was "easy" to understand. Given that the dynamics of financial markets are not simple and easy to understand, reliance on simple models that are easy to explain to the MBAs is probably not the best idea.
With respect, classical economics and Austrian economics are not quite the same thing, and the Austrian school of economics explains this quite well.
Notice any similarities here? No, it's not a perfect fit, but it's the best I could do on short notice.
No one is saying that these models have nothing to do with malinvestment, but it's likely the inputs to the model are also obfuscated by distorted monetary signals
You didn't LOOK at the graph, did you? That's my citation.
This is the last 2 years, with that "almost completely VERTICAL" drop.
This is a 2 year span at 1929, that little tiny blip on the left of your zoomed out chart. Notice how it's actually more vertical than the current drop?
This is your chart, redrawn to have a log scale vertical axis instead of linear. It looks like "now" is roughly comparable to 1938 or the early 1970s.
Here's a link to Taleb's views on the financial crisis:
http://www.edge.org/3rd_culture/taleb08/taleb08_index.html
It's an easy read with nice quotes like " The banking system (betting AGAINST rare events) just lost > 1 Trillion dollars (so far) on a single error, more than was ever earned in the history of banking."
and "I have nothing against economists: ... But beware: they can be plain wrong, yet frame things in a way to make you feel stupid arguing with them. So make sure you do not give any of them risk-management responsibilities."
I can't find the quote (I think it is in "The Black Swan" or "Fooled by Randomness") but I'm pretty sure that Taleb's comment on Li's Cupola is that it is a pretty piece of mathematics whose essential problem is that it never worked for what people were trying to use it for.
You can't see ANYTHING from a car, You've got to get out of the goddamned contraption and walk...Edward Abbey
This is why you can't build a model by looking at a list of numbers. You have to actually understand the source of the data. For example, to go back to the weather example: you can't forecast temperature by looking at a temperature log. You have to actually know something about the sun and oceans and wind and stuff. ;-)
It is foolish to look at investments abstractly. They're not just numbers. They're businesses (or houses or whatever) and they exist in the real world.
Some people say if you diversify enough, then you add so much noise that the sum becomes abstract, and you can start to treat it as a statistical problem rather than an intell problem. *sigh* Yeah, I guess you might get away with that.
For a while.
As copyright owner of this comment, I authorize everyone to defeat any technological measure which limits access to it.
Maybe you should get the facts before opening your mouth. Less than 5% of the mortages failed.
The banks however over extended themselves with the hope of using future profit to pay past due debt. Think of it this way. Balance your budget so you can pay all your bills. Now go max out your credit cards, take a second mortage and buy a couple more cars. Does it make sense? If so you have a future in banking, or government.
i thought once I was found, but it was only a dream.
Might want to check your numbers again. Bush ran up more debt that every President before him COMBINED. He came in with around $4 Trillion in debt and left with around $12 Trillion in debt. Obama has a long ways to go before he gets into Bush territory. In what fucking fantasy land did Iraq and Afghanistan cost $100 Billion? Shit we flew $125 Billion in cash in on pallets to hand out to contractors, most of that money is completely unaccounted for.