Bear Stearns...

[IdleRich]

"Any good? Did he talk about the financial crisis"

Yes, I enjoyed it, I haven't read the book so I don't know how much of it may have been redundant for those that had but I thought he spoke well. A lot of things he didn't really go into depth in and assumed a fair bit of knowledge (problem for one of my friends who had just come back from six months in South America) but I was fine with that. I thought he came across as quite likable and really came into his own with the questions which he answered well, it's just a shame that they seemed in a hurry to wrap it up when I would have happily sat through a load more questions. He didn't really speak about what was going on at the moment except to say that he preferred to speak about things before they happened. He claimed that lots of people had asked him to comment on the subject but that he always refused saying that everything he had to say on it had already been written prior to it occurring.

 

[vimothy]

Yeah, I'm quite jealous that you got to hear him talk. I think he's got a really nice voice. I think you'd probably enjoy this interview with GMU's Russ Roberts, where Taleb goes into his ideas in some depth but assumes no prior familiarity with his work. Really nice interview, really cool talk.

 

[IdleRich]

Thanks, I will give that a listen when I'm somewhere that I can hear it.
Taleb alluded a couple times to (failed) attempts by statisticians to discredit him - do you know to what he was referring?

 

[vimothy]

Taleb alluded a couple times to (failed) attempts by statisticians to discredit him - do you know to what he was referring?

'Fraid not, though I know he's debated famous statisticians and economists and made them very angry. Be great to hear recordings of those, but I've yet to find them, if they exist.

 

[DWD]

Thanks, I will give that a listen when I'm somewhere that I can hear it.
Taleb alluded a couple times to (failed) attempts by statisticians to discredit him - do you know to what he was referring?

He had a famous* set-to with Philippe Jorion around 97/98 on the subject of VaR.

It's also worth bearing in mind that Taleb loves himself quite a lot and isn't exactly averse to talking up his own importance in the history of risk management ideas.


*In certain circles.

 

[vimothy]

It's also worth bearing in mind that Taleb loves himself quite a lot and isn't exactly averse to talking up his own importance in the history of risk management ideas.

The man is clearly well up himself -- still a good book, though.

 

[IdleRich]

This is what you mean I guess....

http://www.derivativesstrategy.com/magazine/archive/1997/0497fea2.asp

 

[DWD]

This is what you mean I guess....

http://www.derivativesstrategy.com/magazine/archive/1997/0497fea2.asp

Blimey. I didn't know that stuff was still online. But yep - that's the one big, public debate I can remember Taleb getting involved in. I have a feeling it carried on in the academic journals for a while as well.

 

[IdleRich]

"Nassim Taleb, a veteran option arbitrageur, is the author of Dynamic Hedging: Managing Vanilla and Exotic Options. He holds an MBA from Wharton and is soon to defend a Ph.D. thesis in option pricing at Universite Paris Dauphine. His next book (coauthored with Helyette Geman) will be called Applied Option Theory. Aside from option theory, Taleb is interested in the philosophy of statistical inference."

Shame that when I was a market-maker/option arbitrageur using dynamic hedging I'd absolutely never heard of Nassim Taleb.

 

[vimothy]

Blimey. I didn't know that stuff was still online. But yep - that's the one big, public debate I can remember Taleb getting involved in. I have a feeling it carried on in the academic journals for a while as well.

I think you have to score this one in Taleb's favour.

DS: In a word, what bothers you about LTCM?

NT: It is the excuse they gave after the fact, and their attitude that their science is above observation. Meriwether wrote a letter to his investors as they were blowing up asking for more money and explaining that "the trades will converge," rather than saying "may converge if we have the right model." Obviously he had (and probably still has) the wrong model. Such blindness also manifests itself with their econometrician saying it was a 10-sigma event. 10 sigma is supposed to occur so infrequently that the probability of having the wrong model is far, far greater. Every time you hear "10 sigma," beware the pseudoscience.​

http://www.derivativesstrategy.com/magazine/bestof/VAR_Chief.asp

 

[IdleRich]

"10 sigma is supposed to occur so infrequently that the probability of having the wrong model is far, far greater. Every time you hear "10 sigma," beware the pseudoscience."

Very good point very neatly made.

 

[Mr. Tea]

*unlurks*

Well considering that a five sigma event happens (in a standard distribution) in about 2 events out of every billion, surely a 10 sigma event is so unlikely as to be virtually impossible?

*relurks*

 

[vimothy]

Well considering that a five sigma event happens (in a standard distribution) in about 2 events out of every billion, surely a 10 sigma event is so unlikely as to be virtually impossible?

Seems to happen all the time, though... Didn't the credit crunch immediately get described as a "25-sigma event" in an FT report? (Trying to find it now).

Marshall, Allais and Coase used the term charlatanism to describe the concealment of a poor understanding of economics with mathematical smoke. Using VAR before 1985 was simply the result of a lack of insight into statistical inference. Given the fact that it has been falsified in 1985, 1987, 1989, 1991, 1992, 1994, and 1995, it can be safely pronounced plain charlatanism. The prevalence of between 7 and 30 standard deviations events (using whatever information on parameters was available prior to the event) can convince the jury that the model is wrong. A hypothesis testing between the validity of the model and the rarity of the events would certainly reject the hypothesis of the rare events.​

 

[Mr. Tea]

Haha, I can't be arsed to do the calculation, but I bet you'd have to wait like a trillion times the age of the universe before you'd expect to see an event like that.

 

[IdleRich]

"Well considering that a five sigma event happens (in a standard distribution) in about 2 events out of every billion, surely a 10 sigma event is so unlikely as to be virtually impossible?"

That's the exact point he's making, it makes more sense to infer that the event was not in fact a 10 sigma one ie the model was wrong. But saying "our model was wrong" doesn't play so well with investors as saying "We were defeated by an event that could only have happened once in the history of the world, if it wasn't for that billion to one occurrence everything would have been fine".

 

[vimothy]

Think it was this one:

“In a rare unplanned investor call, the bank revealed that a flagship global equity fund had lost over 30 per cent of its value in a week because of problems with its trading strategies created by computer models. In particular, the computers had failed to foresee recent market movements to such a degree that they labelled them a ‘25-standard deviation event’ - something that only happens once every 100,000 years or more.

“‘We are seeing things that were 25-standard deviation events, several days in a row,’ said David Viniar, Goldman’s chief financial officer.”​

 

[Mr. Tea]

That's the exact point he's making, it makes more sense to infer that the event was not in fact a 10 sigma one ie the model was wrong.

Oh sure, I got that, I just wanted to get an idea of how unlikely it was.

 

[vimothy]

Anyone watch the Senate Banking Committee hearing on Bloomberg last night? Bernanke said the US could be in a recession. Dimon and Schwartz gave testimony together... pretty fascinating.

 

[IdleRich]

I keep reading about how significant this "crisis" will be for the way banks operate in future. People are making a lot of noises about increased regulation as there is a lot of resentment about banks' perceived ability to privatise their gains but generously share their losses with everyone - but will this regulation ever materialise and if so how strong and sweeping will it be? I've also read that CDOs will not be traded any more - but is that really the case? Surely if no-one is trading them then someone will see an opportunity and offer a way to do this (at a higher cost to pay for the risk) and someone will undercut them and so on. Banks are supposed to be going back to their "old ways" of doing business, nice and simple and that everyone can understand but my guess is that as soon as things have stabilised a bit people will be looking for extra ways to make a profit and vanilla stuff just won't scratch the itch. Basically, what I'm asking is, are people exaggerating the significance of the credit crunch in terms of the way it will fundamentally effect the way banking is done? And if not will it be a good thing or a bad thing?

 

[DWD]

I've also read that CDOs will not be traded any more - but is that really the case?

No. I mentioned up-thread that the majority of the collateral in the higher-rated tranches of these deals is still performing. As soon as investors realise that (which might take some time) we're going to see a lot of mark-to-market losses reversed.

And you're right about the rest of it as well: securitisation makes far too much sense for the economy as a whole for it to be abandoned. No-one (including regulators and policymakers) wants to go back to a situation in which credit supply is constrained by the amount of capital in the banking industry. The problem was never with securitisation per se - it was primarily with the quality of a specific type of underlying asset. The other ancillary problems that contributed to the crisis - investor complacency, excess liquidity, rating agency incompetence, mark-to-market accounting, bank capital rules - should now change, but securitisation as a mechanism for distributing risk is not going to go away.