Felix Salmon (via Brad DeLong):
The damage was foreseeable and, in fact, foreseen. In 1998, before Li had even invented his copula function, Paul Wilmott wrote that “the correlations between financial quantities are notoriously unstable.” Wilmott, a quantitative-finance consultant and lecturer, argued that no theory should be built on such unpredictable parameters. And he wasn’t alone. During the boom years, everybody could reel off reasons why the Gaussian copula function wasn’t perfect. Li’s approach made no allowance for unpredictability.
From this and other scattered things I’ve read, this sounds like a fatal organizational problem, ascribable to managerial bad judgment culpably motivated by greed and toxic optimism.
The finance people couldn’t understand the math, or refused to, so they ignored the nitpicky warnings from the mathematicians. The quants may not have understood the economic and financial aspects, or maybe they also just didn’t care, since they were having the best paydays of their lives. In any case, there was an enormous pathology at the quant-finance interface.
Everyone involved saw the bonus checks rolling in, and maybe they just decided that the problems with the Gaussian copula formula probably really weren’t very important, because of complexity or emergence or disseminated intelligence or consilience or the cosmic vortex or the Aquarian Age or the singularity or some other reason which they also didn’t understand. Not their department. Nobody’s department.
And then BOOM! Half of everything was gone, and the joke was on us.
I don’t understand a bit of the math, but as it turned out, neither did they. Another joke on us!
I love the Three Stooges.
*****
February 24, 2009 at 3:00 pm
[…] (See also: A funny math joke just destroyed half of everything.) […]
February 24, 2009 at 4:54 pm
The funniest part of it, to me, is: “Investment banks would regularly phone Stanford’s Duffie and ask him to come in and talk to them about exactly what Li’s copula was. Every time, he would warn them that it was not suitable for use in risk management or valuation.”
Why did they keep calling? Was it like checking your Email? “Damn, I refreshed the thing again and still didn’t hear that little “bing!” sound that means that a proof of why Li’s copula is safe to use for valuation has come in.” Or perhaps like a child asking his parents over and over whether he can have the same thing. “No, Junior, you’re not going to get a proof of Li’s copula for Christmas. It’s too big for Santa to bring down the chimney. Just stop asking.” Or maybe they were like reporters looking for the perfect sentence fragment to misquote. “Maybe if we have him say all that again, he’ll say ‘Li’s copula can be used for valuation — you must be high’ and we can just quote the first half of the sentence.”
There’s a long history of bosses asking the same question over and over until they get the answer they want. If the person actually works for them, they quickly give the desired answer or get fired, and then a few years later the company goes out of business. If the person doesn’t work for them, you get this kind of disconsolate ritual repeat, and then a few years later the company goes out of business.
February 24, 2009 at 5:13 pm
Except … who got hurt? I don’t mean that as a rhetorical question. I’d really like to know if the individuals inventing these securities and making these trades are now in line in food kitchens.
I suspect that many of those who really got burned when these markets evaporated are now unemployed … and sunning themselves on beaches in the Caribbean, voting absentee for Republicans.
March 2, 2009 at 7:53 pm
The failure was both institutional and systemic. Institutionally there were multiple interfaces that were involved, in addition to quant-trader; for example, risk was broken into market, credit, and operational without recognition of the interplay, the opinions of supposedly disinterested independent rating agencies were taken without salt, there was gross overreliance on historical parameterization and on measuring only that which was measurable, satisfaction of regulatory constraints was paramount (beyond which risk was advisory not determinative for management, counterweighted by an apparently extraordinarily profitable delusion for the CEO crowd). The compartmentalization of information within investment banks also contributed in that no one with the requisite skills had access to the big picture.
But the problem with Gaussian copula itself was that it was stretched well beyond its capabilities, which with regard to corporate securities had weaknesses that were tolerable; the Gresham’s Law rush to incorporate higher-yielding assets (mortgage-backed ABS, CMBS and the like) had the quants playing catch-up with issuance of new structured securities while demand weakened loan-underwriting standards top to bottom.
March 2, 2009 at 10:36 pm
Lest the foregoing appear too apoditic a response to rage against cogs in the machine, I’ll amend with letters to the Corinthians:
I. Paul Wilmott:
http://www.wilmott.com/blogs/paul/index.cfm/2009/2/24/Copulas-and-Cults
II. Paul Volcker:
http://network.nationalpost.com/np/blogs/fullcomment/archive/2009/02/17/paul-volcker-the-banking-world-needs-more-canadas.aspx
excerpt:
“One of the saddest days of my life was when my grandson – and he’s a particularly brilliant grandson – went to college. He was good at mathematics. And after he had been at college for a year or two I asked him what he wanted to do when he grew up. He said, ‘I want to be a financial engineer.’ My heart sank. Why was he going to waste his life on this profession?
A year or so ago, my daughter had seen something in the paper, some disparaging remarks I had made about financial engineering. She sent it to my grandson, who normally didn’t communicate with me very much. He sent me an email, ‘Grandpa, don’t blame it on us! We were just following the orders we were getting from our bosses.’ The only thing I could do was send him back an email, ‘I will not accept the Nuremberg excuse.’
There was so much opaqueness, so many complications and misunderstandings involved in very complex financial engineering by people who, in my opinion, did not know financial markets. They knew mathematics. They thought financial markets obeyed mathematical laws. They have found out differently now. You know, they all said these events only happen once every hundred years. But we have ‘once every hundred years’ events happening every year or two, which tells me something is the matter with the analysis.”
March 2, 2009 at 11:17 pm
Ultimately I believe that this was a pervasive , society-wide, multi-pronged madness of crowds. I’ve made the prosperity theology / cornucopian freemarketer comparison many times. Throw in cocaine/amphetamine euphoria. Throw in various forms of transgressivity, mad-dog realism, and mad-dog skepticism. You just keep finding more layers. You never find the bottom, but you’ll keep on finding more and more self-serving pathologies of toxic optimism and purposive ignorance.
The various individualist “rational actor” models all assume the impossibility of “the madness of crowds.” And of course, that’s OK! Because Milron Friedman proved 60 years ago that you can make as many counterfactual assumptions as you want. Only small-minded populist Luddite Nazi Communists disagree.
Thanks for the links. When someone like Volcker makes criticisms of that particular kind, it’s damning. Like a lot of people, though, he should have spoken up much sooner and much louder.
May 17, 2009 at 6:17 pm
[…] The tweakers are crashing on us We are living in Greenspan’s experiment A funny math joke just destroyed half of everything) […]