This Can't Be Happening!
After all it's not often one lives to see an event that our mathematical models predict can only occur once in 73-603 trillion billion years!
http://www.eurointelligence.com/article.581+M5f21b8d26a3.0.html
Gives new meaning to the saying "You don't see that everyday!" So now it is safe & right to sneer at the disgraced "quants"
http://www.wilmott.com/blogs/paul/index.cfm/2008/11/17/Actuaries-Versus-Quants
The math got very very intricate and also very very wrong.
Currently my own wacky little model is predicting an upswing in Q3 2009. But maybe that's the outcome that won't happen for another few hundred billion years!
I'm going long on pencils and apples.
(Dedicated to Ellen and Quinn)
Advertisement
Perhaps they should have read this:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=970480
But hindsight is always 20-20.
December 12, 2008 1:59 AM | Reply | Permalink
Lux, when I try to open the pdf from this abstract, both Adobe and Foxit tell me it's corrupted. So what what was the bottom line (so to speak) of what we don't know?
Interestingly, Greenspan had the opposite notion of risk and derivatives and their effects on the portfolio than Wilmott:
Turns out that Greenspan didn't know a lot. Wilmott's theory intuitively makes more sense.
Quinn beat me to the black swan. Can it be safely assumed that Wall Street never met one while actuaries have had the occasional sighting and learned from them?
Other questions. Did (does) the US have the lead in Basel II, based on our risk management systems? Is Basel II dead yet?
Good post!
December 12, 2008 7:35 AM | Reply | Permalink
Hi Seashell; if you select the SSRN site for downloading you should get a good file. I don't use the other sites. I believe you have to have either a private or institutional subscription, but if you do, SSRN is a world of wonderful articles with a semi-searchable database.
I can save you the trouble though as it is a fairly short (8pp) paaper with very little technical passages. The authors tested several groups including portfolio managers, assistants, quant/analysts, ivy league financial engineering majors (read quantitative finance) and bank employees specializing in investments. What was being tested was something on the order of STAT 101 awareness of the difference between standard and absolute deviation especially when you sum over the business year.
"Our suspicion that there would be considerable confusion was fed by years of hearing options traders make statements of the kind, 'an instrument that has a daily standard deviation of 1% should move 1% a day on average.'"
Wrong, but this confusion showed up in the test results which showed an alarming confusion and even incomprehension on the part of the respondents. Uniformly everyone underestimated market volatility even though they understood the mathematics at an abstract level but they acted on a gut sense that was inaccurate. The authors concluded:
"Either we have the wrong intuition about the right volatility or the measure of volatility is the wrong one...."
Why I included this citation is that while it doesn't bear on the complex financial instruments, the hall of mirrors that mathematics created in the financial world, but it does show how we are not masters of our own mathematical apparatus. We understand notionally our models, but our intuition often leads us astray. How could our models be so wrong? The equations don't lie: its perhaps what we think they mean.
December 12, 2008 1:27 PM | Reply | Permalink
Thanks for the executive summary, Lux. Usually I have no problems downloading from SSRN. It took awhile today, but duh, I tried dl'ing with IE 7 instead of Firefox 3 and it worked just fine. Other papers would dl today from SSRN with FF3, just not that one. Don't ask me why, I have no idea. (Besides, you're the quant!)
So are you in favor of the risk managers being at fault vs the risk models that fail to foresee catastrophic losses much past a day or a week or two? Or a combo of the two?
Have you seen the newly discovered (but apparently not new) Default Recovery Swaps? Here is the Bloomberg article on them.
December 12, 2008 10:49 PM | Reply | Permalink
Hi Seashell, Let me clear this up: LUX IS NOT A A QUANT!!! He uses mathematical models in his research, yes. They can be applied to markets, right again. BUT HE IS NOT A QUANT! I only told that evil bully Quinn when he teased me with a "schoolyard chant of Quant Drone, Quant Drone!" On behalf of nerds everywhere I had to stand up for the numerically un-challenged! Kind of like Ich bin ein Berliner (tasty that)!
NOR DOES LUX HAVE AN MBA! I couldn't even pass ECON 201 much less the stuff they shovel down the students throats in the 600 level FIN courses.
Dickday had me pegged accurately, I am a bit of a simpleton who has idiot-savant level knowledge of one microscopically small academic sub-specialty that happens to study mathematical artifacts that can model various sorts of phenomena. If you know the difference between a Moore neighborhood and a Star of David neighborhood, you know where I hang my hat.
Finance has become a latter interest and since the math is either comprehensible or dispensible I have been pursuing my autodidactic course in coming up to speed on what blindsided the global economic structure.
So my reputation for financial sagacity is, like the value of a cds, fairly derivative and unhinged from reality. I have leveraged it beyond belief but the time for the crash has come! Notional sagacity!
There....!
But I would say that a whole culture or as Kuhn would say, a whole paradigm was at fault. I think CT's link to the Michael Lewis article does point out that some people knew the increasing intangibility of risk assessments, so there is blame there, but on the whole, the pack of followers were content to try to imitate the ways of the alpha's and there was always a mathematical legitimization providing a green light justification if one had to answer to one's boss or one's board on buy decisions. I also think there is an intense culture of competition especially among investment managers, and one was under pressure to bring one's own fund up to the profitability of those that had taken big positions in hyper-leveraged securities. Fuddy-duddies who worried about overhangs on tranches that were probably only BBB at best would soon get the reputation as not being "players" and everyone wanted to be a player.
We also have to look at the entire academic drive to quantize and establish on a firm mathematical footing all human endeavor (well maybe except the fine arts)from WWII onward. Nothing was really science unless it had some mathematical underpinning. Physicists are to blame for that methinks! Finance was given the full treatment and it led to the mathematics that was behind the fall.
Default recovery swaps! The ingenuity of the gambling..er, I mean financial world! They seem a little like a form of reinsurance or risk splitting. Kind of like a two-way cat bond?
Seashell, you and Tom and TheraP (with her multiple regressions) are the true QUANTS. I am just an information aggregator.
December 12, 2008 11:48 PM | Reply | Permalink
*quant drone quant drone*
I looked up that Moore stuff, Lux. And then quickly looked away. Thought Brain #3 might throw a rod. Damn.
While we're at it, can I give up that "Economist" label? I kept getting thrown out of orthodox Economics Depts. Ended up working in & around Economics, but not in the full-on math or number-crunching sense they seem to feel is required. And anytime the Finance guys show up, I just bolt. Let my friends handle them. I'm happier pitching pols or the public or greens or, hell, ANYBODY but the finance guys.
Now about these Moore neighborhoods..... *Explosion.* Oops.
December 13, 2008 12:04 AM | Reply | Permalink
Oh come now Quinn!
I googled you intensively and drew a blank, but Freddy is abler than I, and in about a femtosecond he had you correctly identified.
Holder of the Roger Darlington Chair in Finance, London School of Economics!
Fellow of the Royal Society for Economic Research!
Winner of two, TWO Fields medals, and knighted by the Queen Herself!
Pull the other one, SIR Quinn!
December 13, 2008 12:17 AM | Reply | Permalink
Are you stalking and exposing our dear Quinn?
Shame on yez. This best be a crush, luv. ;-)
December 13, 2008 12:24 AM | Reply | Permalink
Well actually, LisB, my toaster/supercomputer Fredkin did the stalking and I have to admit that ever since Quinn sent a chunk of Canada's military forces down to destroy Freddy and his brother Toff, Freddy has been somewhat erratic and perhaps the other choice he presented me was the true one.
Sitter in the Ethan Allen Chair, backwoods Canada,
Fellow of the local Royal Saloon and Grill,
Knight of Columbus (ret.)
December 13, 2008 12:31 AM | Reply | Permalink
What are the odds that one of the lowest-scoring defencemen in NHl history (averaging 2 goals per season) should score the winning goal, in Overtime, of the Cup Finals, against the Red Wings?
While playing with a broken leg?
Ask Bobby Baun. Did it in '64.
What are the odds that a kid emigrates to America in the overhead luggage bin of a flight out of Heathrow, keeps breathing, lands on a farm inhabited entirely by blond-haired, blue-eyed, carnivorous cousins standing 6' 3" to 6' 9", survives, comes to in a car driven by Desmond Tutu's kid, hysterical, surrounded by a rockin' mob on the first day of the Soweto Riots of 1986, crawls out, and ends up wearing an original Bobby Baun-sweater while posting semi-demi-insane rants at TPM?
Ok, zero chance of that one. Be serious. NOBODY'S got an original Bobby Baun sweater.
People. They never met a mathematical formula they didn't wanna break. Personally, I think humanity just did this to show off. I think it's been long odds all along, Lux. The whole human story. From the physical constants of the universe, through the primordial goo, from the egg-eaters beating the brontosaurus, from Herod missing the kid to Bobby Baun's busted leg, and now, up to & through those lunatics shoving dollars in & out of the market.
Hey, look! A black swan! ;-) Thanks for the post, and that first link in particular. And thank you Mr Dearman, for teaching Grade 9 Math so badly I fell completely out of love with it. Double ;-)
December 12, 2008 3:12 AM | Reply | Permalink
FYI - http://cgi.ebay.com/BOBBY-BAUN-Toronto-Maple-Leafs-SIGNED-Hockey-JERSEY_W0QQitemZ330286701802QQihZ014QQcategoryZ27284QQssPageNameZWD1VQQrdZ1QQcmdZViewItemQQ_trksidZp1638Q2em118Q2el1247
Rec'd
December 12, 2008 3:36 AM | Reply | Permalink
Dude. My life... complete.
LOL. That hurt.
December 12, 2008 3:57 AM | Reply | Permalink
Less amusingly Brad deLong points to Michael Lewis' comment that the only problem with the quants approach to risk was that it ceased to work in risky situations.
Otherwise it was great.
December 12, 2008 4:04 AM | Reply | Permalink
Hey! My new seat belt design works great unless you're traveling over 25 mph. C'mon, pony up and invest in it.
December 12, 2008 4:12 AM | Reply | Permalink
Weather. We keep getting these 100 year floods and so on. I'm wondering if a lot of models are just wrong!
Thank goodness I had to take some statistics and 2 out of 3 of those courses were taught by fantastic teachers - so I "get" what's being discussed in some of these articles. I also read a lot about "stochastic" process once. But I wish someone with the understanding could provide a metaphor for what's gone wrong here. But I'm also keeping in mine that one of those excellent stats profs could not give me a metaphor for how to "think" about multiple regression - something which fascinates me.
Here's a question for the "experts" among us: I'm wondering if part of the problem is trying to predict the probability of huge swings is based upon just trying to see the math in terms of a "static" descriptor - like averaging along a presumed bell curve rather than working with a more dynamic descriptor like multiple regression (or something that would give a sense of dynamic factors working together.... like the butterfly effect in climate change).
Am I off my rocker here? Or in some kind of ballpark?
By the way, I adore the caliber of TPM people! And especially those who combine wit with wisdom and quirky self-revelation.
December 12, 2008 10:31 AM | Reply | Permalink
As I told Ellen last month, I flunked "Stochastic Methods" and dont know the difference between heteroskedasci.whatever.ity and heterozygotes!
So count me out! I just post hyperlinks!
December 12, 2008 11:55 PM | Reply | Permalink
I know nothing of stochastic "methods" - only the gist of randomness and then a process for "selection" from the randomness. (so structure emerging from chaos) I read about it from Gregory Bateson, who was an anthropologist and later taught medical students (long after he and Margaret Mead were divorced).
So maybe we're all building each other up into these TPM legendary polymaths. And if so, perhaps that's how the QUANTS built themselves up too and basked in it - like bunches of people in hot tubs! That's where they went wrong!
So I'm glad we here at TPM for the most part aren't full of ourselves and to the extent we stay that way, we may not go so wrong over time.
Seems to me chaos theory might do these QUANTS better now. The opposite of the stochastic thing - a kind of entropy leading to chaos taking over the entire financial world. That's probably taking it too far. But quinn's weasels might know. We need a process for consulting weasels.
December 13, 2008 12:09 AM | Reply | Permalink
Ughhhhhhhhh...I only passed high school algebra because my teacher had a crush on me. I am suitably in awe of all of you. I have no idea what you are talking about, but you sure sound smart! :-)
December 12, 2008 11:53 AM | Reply | Permalink
It's just a ruse! :-)
December 12, 2008 12:27 PM | Reply | Permalink
The only thing that made sense to me was the last two paragraphs:
December 12, 2008 3:02 PM | Reply | Permalink
Good post, comments too. I was kind of bewildered about something, Lux maybe you can help me here. I thought that a fundamental assumption of the normal distribution is that the events being distributed are independent of each other. Obviously, that assumption makes no sense at all with regard to volatility of stock prices. And if you assume that dependent events are independent you do get ridiculously low probabilities for combinations of events that are occurring together because they are influencing each other.
Am I missing something?
December 12, 2008 7:44 PM | Reply | Permalink
Tom, that's brilliant! That's part of what I meant when I was looking for a more "dynamic" model!
I think you've clarified the reason for the false predictions!
December 12, 2008 7:51 PM | Reply | Permalink
Thanks, but it's hard for me to believe that the quants could be that statistically naive. If that's the case, then it's really scary.
December 12, 2008 9:00 PM | Reply | Permalink
You yourself already deduced that! The scary part. I think more and more we can see these guys were deluding themselves bigtime!
December 12, 2008 11:10 PM | Reply | Permalink
Tom, when I first read the October market article I assumed it was a spoof or an attempt at morbid humor and looked for a "comments" that others saw it in the same light.
It then dawned that the authors may have been serious in which case they have some major problems with their grasp of elementary probability! Or I simply don't understand what they are saying.
If the authors were serious, then that was SCARY! And remarkable enough to warrant a post to bring it to others attention.
December 13, 2008 12:02 AM | Reply | Permalink