pun of the day

here’s the pun of the day from this great piece in mit’s technology review on the mortgage situation. this is the first article i’ve read which offers both a great background on how the CDO (and, in general, derivative) markets operate, as well as, describes the dynamics of what happened (is happening).

In fact, the summer might be described as a time when too many investors had purchased standard deviations that were too high for their means.

then there’s this bit on the increased complexity of modern derivative markets and their downside :

[Bookstaber] frets about complexity and what he calls “tight coupling,” an engineer’s term for systems in which small errors can compound quickly… “We have gotten to the point where even professionals may not understand the instruments,” he says. This was perfectly demonstrated this summer, when the subprime troubles touched off a reactionary wave of selling in equities that would nominally seem unrelated, or, “uncorrelated.”

“Nobody knew that what happened in the subprime market could affect what was going on in the quant equity funds,” he says. “There’s too much complexity, too much derivative innovation. These are the brightest people in the business. If it could happen to them, it could happen to anyone. No one could have predicted the linkage.”

Linkage is one of Bookstaber’s favorite topics. He believes that quants’ instruments have “linked markets together that wouldn’t normally be linked,” and that such linkages are dangerous because they are unforeseen.

This guy (Bookstaber) obviously knows far more about financial markets than me; however, I don’t fully agree with his assesment that this was unforseeable.

First, I think simply improving volatility measurements in the mortgage-trading sphere on subprime loans (as i said earlier), would improve the valuation which banks were giving them in the margin accounts. That alone would stabilize correlated perturbations between debt-obligation markets and the stock market. Secondly, banks require disclosure from the clients of their margin accounts - they need to by law (to secure their depositors) - and so - they could have predicted the new “linkages” by analyzing what instruments margin traders were purchasing with their CDO-backed margin funds. From there, simply analyzing projected behavior based on the same models the funds were using would allow the banks to infer the associated correlations. Of course, it’s a bit more complicated than that - both because different funds use different models and because funds split their margin accounts between multiple lenders to obscure their holdings… but, it’s a good first approximation. Secondly, to me, the idea that it’s all more complicated is no argument against pursuing more and more advanced, entangled and complicated financial markets. I think that ideally, financial markets should incorporate and encompass any transaction we can conceive and would want to execute.

Finally, this last bit (the last page or so) drove me nuts :

Emanuel Derman remembers dreaming of such a unified financial theory in the early 1990s, a little after he had made the leap from the university to the Street. But those dreams, he says, are dead. Quantitative finance “superficially resembles physics,” he says, “but the efficacy is very different. In physics, you can do things to 10 significant figures and get the right answer. In finance, you’re lucky if you can tell up from down.”

First, the idea that we “understand”, in some categorically different way, physical laws from financial laws is… um… absurd. Physics can solve *some* problems to within 10 sig-figures, but most it can only solve through a process of approximation - ie the brownian motion they discussed with reference to the black-scholes equation’s volatility metric. So… examples of intractable physics problems abound, in physics “you’re lucky if you can tell up from down”. Ok, one more digression - financial transactions in isolation are the function of individual, psychological decision-making processes - those are indecipherable. However, financial transactions in aggregate - and thus market dynamics - are inherently different phenomenon and don’t require complete (or any) psychological explanation - just like generalized crowd-behavior or traffic or an endless litany of other group-dynamic phenomenon. In fact - what psychological explanations do exist for such behaviors are often tacked on *post-hoc* as descriptive justifications of observed behavior - not implicit explanations of such. As such, financial sciences (in the realm of volatility and price-theory) are going to be akin to statistical mechanics & stochastic physics in general - not relativity, and it’s ilk (classical mechanics) - THEY are categorically different things. Both are theoretical models, but one is static and explanatory (as best as science can claim to be) and the other is reactive and a posteriori (if you want to be pedantic, it’s the deductive/inductive divide philosophers of science have been arguing about at least since hume). Anyway… if you’re interested at all… i’d recommend both the book - my life as a quant - which he recommends in the article and also - when genius failed - which is a great description of the events which led up to and resulted from the failure of long-term capital management.