Friday, February 4, 2011

I was forced to start this blog because Pablo Triana wrote the world's most ridiculous book about quantitative finance

Pablo Triana is the author of 'Lecturing Birds on Flying: Can Mathematical Theories Destroy Financial Market?  He has discovered that "demonstrably flawed and lethal models should be banned from the land, and the real reasons for their original embracement intrusively inquired". Triana suggests we all "engage in a healthy dose of loud name-calling and unabashed denunciation" because "the math has had its chance". Furthermore we "urgently need a Mathematical Finance Council of Nicaea" because the discipline is "in tatters". Mr Triana has dined with some of the high priests so as to get things underway.

These are quite significant findings backed up, one might imagine, with some thorough analysis. I owe an apology to Mr Triana for my hasty review of his work wherein I insinuated some curious features of his methodology. I may have hit a sore point and this drew a surprising, angry response. Here is my review of Triana's book:
In Triana’s case it is “excitingly obvious” that many “thinkers and doers (most of the time thinker-doers)” have “influenced the ideas and contents of this book”. New to quantitative finance however Triana was forced to identify the “thinker-doers” at break-neck pace and “immerse myself for a few months in equation-laden papers” while also having to “research the activities of theoreticians and quantitative professionals”. On the other hand, the author admits that he “didn’t contact (bother) too many third parties when penning this work” and therefore “It was, in the most explicit sense of the term, a solo effort”. Some urgency is understandable, given the “insultingly unpostponable” relevance of the topic which “seridipitously but sadly, was made exponentially more urgent by the neverending, ever-escalating unfolding of the credit crisis”. It was inevitable that a “book like this would require from the author a larger-than-desirable number of hours in the (intellectual) company of dogmas and dogmatics” but we should not presume inevitability of the author’s findings.

Or perhaps we should. “I do not particularly enjoy theoretical adventures”, writes the brave volunteer, laying it on just a little thick. “I can think of hundreds of better (and more productive, let alone enjoyable) things to do than read or, worse, compose a theoretical treatise. I don’t believe much in the power of theory when it comes to finance, either. And (orthodox, dogmatic) financial economists most likely would not be among my first choices for companionship on a deserted island, or a restaurant table”.

Triana relies on verbal bread crumbs like 'gaussian' to weave his way to lunacy. But by all means take this remotely seriously ... 

As you can see, about 80% of my review was Triana's own words. What led me to pen this was the mistaken belief that professional conduct in financial inquiries might somehow resemble other fields, like medicine or air safety. For instance, if I began a medical report about alcohol with "listen chaps, I've always loved a good binge drinking session and I really can't stand those white lab coat guys so I didn't bother to ask them any questions"... then I would not expect anyone to take me terribly seriously. Finance is different, apparently.

Mr Triana's reply has set me straight. "This was supposed to be a debate for grown-ups", he responded, before listing his credentials. Mr Triana has read not one but two biographies of quants (both famous) and "almost understood" a book written by Nassim Taleb.  In fact Triana might, if we are to believe him, easily be confused with an "uber expert" who can "successfully muse" on various topics in mathematical finance. So understandably it was necessary for him to quickly rush his newly formed opinions to print "once it became clear".

It is possible that Mr Triana missed my serious, adult point (related to his strong prior beliefs, determination not to change them, and adoption of research methodology designed to avoid the slightest temptation to do so) and got a little defensive about the old CV. I am sorry Mr Triana, for I was merely trying to point out that knowledge of that nerdy stuff was, it would appear, quite scant before this careful investigative journalism began. For example:
I have never produced a mathematical paper or taught a math-heavy course. My direct experience with the quanty universe is limited to my student days, to my days as a derivatives marketer, and to the brief period when I dated a New York girl who had just dated a (rich) quant.
Yes you read that right. Triana's knowledge of my field was sexually transmitted. (I'm now mildly concerned about what constitutes adult debate and what I might catch). Fine. But let me make a literary analogy. I favor a 'formalist' approach: examining the 'text' as it were, absent the Marxist, post-colonial or feminist interpretations. And I'm not sure I'm finding that in Triana's book:
To all those eager to break free from the Normal dictatorship, history may provide a comforting message. Prohibition and [commun]ism eventually did, of course, spectacularly fail. Why? Simply put, because they run dramatically counter to human nature.

No, that wasn't Joseph McCarthy back from the dead but an example of Triana's successful musing, from Lecturing Birds on Flying, on the topic of the Bell Curve. The Bell Curve, or normal distribution as it is known undoubtedly has something to do with the normal copula model, more recently known as The Formula That Killed Wall Street, because it has "normal" in the name, you see.  Nowhere in Lecturing Birds on Flying, however, do we learn precisely what the connection is because, to state the obvious, Triana has no idea.

I'll leave the pop-finance surrounding this for another time noting only that Triana's critical, discerning interpretation of his mentor's rants is quite touching, and yes, if you are wondering who started the red scare look no further than Nassim Taleb's The Black Swan where all this soft nonsense is "grounded". It is the sort of thing that has a long history, and might have led Richard Feynman to do some musing of his own in one of his last BBC interviews. The physicist noted that philosophy of science is about as useful to practicing scientists as ornithology is to birds, if I recall his words correctly. Curiously, Feynman is absent from the credits for Triana's deviously clever title, Lecturing Birds on Flying.

                         On a slightly more serious note ...

Of somewhat greater concern are the opinions expressed by Professor Steven Shreve in responding to Triana. Shreve is one of the great contributors to the field of financial mathematics but risks giving credence to the 'quants blew up the system' meme:
In some cases senior managers and even quants themselves did not appreciate the limitations in the models on which they based their risk analysis
Really? Professor Shreve would know, I am sure, just how naive the industry standard model was (and is) and for anyone with an ounce of mathematical intuition the limitations were more than evident a long, long time ago. I happen to have some insight as to the reasons why the normal copula model was popular at one major investment bank though, because I was responsible for creating it (the feeble mathematical tricks that made it practical, its physical embodiment in the form of a trading calculator, and the choice of colors for buttons - that is). Yet to suggest the limitations were not apparent is just a little beyond the pale and, after wrestling with management for several years I became so incensed I wrote up an internal polemic on the topic. 

To avoid accusations of hindsight I repeat here, verbatim, a small section. This was written in late 2003 or 2004 (I don't quite recall)   
The general absence of progress in this field is not likely explained by an absence of interest, experience or financial intuition amongst those who have contributed to a rapidly expanding literature. A more likely explanation is that most reasonable approaches channel into hard problems of mathematical or algorithmic nature for which no solutions are known - or circulated. It is not surprising that avoiding these problems leads to rather silly models, and we predict that firms not directly confronting these issues will rapidly asymptote to a rather poor level of quantitative risk management if they have not already done so.
Explaining this opinion to a lay audience will take some time (I'm trying to get around to it) and in the meantime the empty vessels are making quite a lot of sound. The Financial Times in particular appears intent on publishing Triana's every waking thought, and elsewhere the uptake has been almost entirely uncritical. The rather more adult Financial Crisis Enquiry Report appears to be an exception, but the reasons why non-experts are suckered time and time again by Triana are myriad. In parts, Lecturing Birds reads like a serious study and I suppose if it was about paleontology I might go for it. But the arguments comprise a black box treatment and lengthy observations along the lines of "the model doesn't fit the market" or "the model doesn't let you hedge perfectly". Well no shit, Sherlock. Show me the orthodox, sterile, in-the-box academic theory that says it should.

 Of course the object of Triana's fury is not so much the normal copula, which he evidently doesn't understand at all, but fanatical Gaussian Copulists. Triana is unable to locate any fanatical Gaussian Copulists, however, so must carry on the conversation with himself. Therein his attempts to finesse the question of advocacy are downright hysterical, at least to those of us who lived the unbearable lightness of being a quant. The "dealers" were "considering a move away" he writes ambiguously, referring either to institutions or someone with the ability to steer something. Perhaps management had convened a meeting, worked through the inconsistencies of the industry standard approach, and then reported their findings to crestfallen quants enamored by the normal copula? If you can believe that, you can with practice believe anything and in the new world of quasi-quantitative critique (or whatever this is) that's a useful skill to cultivate.