Top Diaries

aren't we back to the back alley where I steal your watch and wallet, then invest the proceeds with a vague promise that if I make enough profit I'll pay you back later?

I'll give you two hamburgers tomorrow for one hamburger today...

The difference between theory and practise in practise ...

by DeAnander (de_at_daclarke_dot_org) on Sat Jun 23rd, 2007 at 12:18:29 AM EST
[ Parent ]

except that I read last week that 13 million people died in 06 from "environmental causes" (which is kind of vague, but I think the report meant "from pollution and climate-related hardships, hunger from desertification, disease and thirst from aridity and falling river levels").

how do we pay them back?

The difference between theory and practise in practise ...

by DeAnander (de_at_daclarke_dot_org) on Sat Jun 23rd, 2007 at 12:20:02 AM EST
[ Parent ]

Re: Here is his response to mine... (4.00 / 6)

You'll notice he's not criticising Stern's claims about global warming or his damage estimates, but Stern's estimation of the present value of the future damage. It's all about accounting, not about the science of global warming.

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Fri Jun 22nd, 2007 at 05:09:30 AM EST
[ Parent ]

Re: Here is his response to mine... (4.00 / 6)

Stern uses an 0.1% year-on-year discount rate. You can calculate what this does to the 1000 rand/dollar banknote. What the discount rate does is downgrade damages and benefits in the future. Now, while the banknote might become worth less, presumably people, wildlife, ecosystems, real estate, etcetera won't become worth less. So the use of the discount rate in cost-benefit models of climate change is a highly contentious issue.

See this old NYT article on the topic.

by nanne (zwaerdenmaecker@gmail.com) on Fri Jun 22nd, 2007 at 05:10:51 AM EST
[ Parent ]

That's just a 10.5% rate of inflation after a century.

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Fri Jun 22nd, 2007 at 05:22:55 AM EST
[ Parent ]

Thanks :-)

Now if you look at the areas where climate change will cause damages, you have to ask whether these are subject to inflation at all, or if they'd sooner become worth more.

The discount rate, in other words, needs to be deconstructed.

by nanne (zwaerdenmaecker@gmail.com) on Fri Jun 22nd, 2007 at 05:44:28 AM EST
[ Parent ]

I will say this (4.00 / 8)

in this reply:

Vaclav Klaus: I ask myself several questions. Let's put them in the proper sequence:

Is global warming a reality?

If it is a reality, is it man-made?

If it is a reality, is it a problem? Will the people in the world, and now I have to say "globally", better-off or worse-off due to small increases of global temperature?

If it is a reality, and if it is a problem, can men prevent it or stop it? Can any reasonable cost-benefit analysis justify anything - within the range of current proposals - to be done just now?

Surprisingly, we can say yes - with some degree of probability - only to the first question. To the remaining three my answer is no. And I am not alone in saying that. We are, however, still more or less the silent or silenced majority.

Klaus outlines the problem which remains the problem in Climate Science - but then spins it out of control.

I strongly feel it has not been possible to convincingly show how CO2 relates and scales to other climate forcings and hence how large its sensitivity is. The knowledge on the system as a whole is to this day incomplete.

But: Anthropogenic CO2 leads to warming. End of discussion. Yet because Klaus lodges to the above problem, he spins all of the global warming into other contributing factors - it makes him a classic denialist. Here's my thing: It does not matter anymore to what level CO2 scales - we know it is a greenhouse gas, the current warming is causing enough imbalances on this overcrowded planet, let's not even risk another contributor to forcings especially if we don't understand what it exactly does.

Perhaps I should have poured that into a question...

by Nomad (Bjinse) on Fri Jun 22nd, 2007 at 05:14:45 AM EST
[ Parent ]

Re: I will say this (4.00 / 4)

[2.] If it is a reality, is it man-made?

"If it is a reality, is human activity a contributing and aggravating factor?" is the right question.

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Fri Jun 22nd, 2007 at 05:20:40 AM EST
[ Parent ]

Re: I will say this (4.00 / 2)

All those questions can be applied to his beloved "science" of economics too...

by Metatone (metatone [a|t] gmail (dot) com) on Fri Jun 22nd, 2007 at 02:10:14 PM EST
[ Parent ]

Re: I will say this (none / 1)

I strongly feel it has not been possible to convincingly show how CO2 relates and scales to other climate forcings and hence how large its sensitivity is. The knowledge on the system as a whole is to this day incomplete.

IPCC scientists feel somewhat differently about this since the last report considers as very probable (>90% certainty) that greenhouse gases are the cause of most of the warming incurred over the last 50 years. This of course doesn't mean that our knowledge of the forcings is complete.

Klaus, like all the other deniers, conveniently uses all the various stages of denial in the course of an argument, which shows they know nothing and yet, refuse everything. The comment about the silenced majority is pure demagoguery (besides being factually wrong): as if a "silenced majority" could be expected to have an opinion differing from that of the scientists on a topic as complex as climate change.

by Fete des fous on Fri Jun 22nd, 2007 at 08:45:45 PM EST
[ Parent ]

Re: Here is his response to mine... (4.00 / 6)

My criticism of Stern Report's conclusions - and I am not alone in it - is based on serious economic arguments, not on aprioristic statements.

The discount rate is an aprioristic statement about the future. Stern makes an aprioristic statement, and so does Klaus. It's anybody's guess whose guess is correct.

Klaus chooses to concentrate on the discount rate and ignores the value at risk, which has a lot to do with the uncertainty in the estimation of the appropriate discount rate.

Underestimating the discount rate is like taking the lower end of a confidence interval, which is the proper thing to do when calculating value at risk.

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Fri Jun 22nd, 2007 at 05:14:54 AM EST
[ Parent ]

BTW, Migeru - I guess you are working back office and do something like derivatives pricing? Did you hear anything about using large deviations theory in real-world finance?

by Sargon on Fri Jun 22nd, 2007 at 08:40:15 AM EST
[ Parent ]

No, I work for a Hedge fund on equity research.

I was at a seminar last week where large deviations theory was briefly mentioned as part of the discussion of estimating value at risk and the expected shortfall. The speaker is a senior researcher with a large international investment bank.

The problem with large deviation theory, according to this guy, is that it only works for the very tail, but that means in practice you don't have a lot of data to fit the tail distribution.

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Fri Jun 22nd, 2007 at 09:07:58 AM EST
[ Parent ]

Remember name of the guy?

Migeru and Pierre, thanks!

by Sargon on Fri Jun 22nd, 2007 at 09:39:49 AM EST
[ Parent ]

Even better: here's his website.

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Fri Jun 22nd, 2007 at 09:43:38 AM EST
[ Parent ]

I work on computing VaR for equity derivative portfolios in a bank. My neighbour is working on a prototype engine to tackle large deviations. Unfortunately, this is still based on Gaussian Copulas and "sum of gaussian approximations", because gaussians enable nice linear parametric VAR computations which is:

about the only one we can compute daily with existing computing power (I guess this is the same everywhere: it's just one giant matrix multiplication, and this is very efficient)
the only one that you can explain nicely by projecting the VAR on the risk factors (otherwise, you cannot have a "dashboard" to steer the activity of your traders in order to contain your risk within set bounds).

And he still has a lot of trouble with his models. Like mig says, there are few large deviations to calibrate the stuff, and Gauss does indeed work well day to day (as proven by backtesting over periods without krachs). It's only when things get real bad that you see your quantile was wrong (about to know pretty soon I guess).

Mandelbrot claimed that scaling laws where better than Gauss, can be calibrated, and incur manageable computing costs, but I don't think anyone has looked into it on the industry-scale. I'll dig a link

Pierre

by Pierre on Fri Jun 22nd, 2007 at 09:20:26 AM EST
[ Parent ]

gurus get risk all wrong

Specially impressive is the centenial chart of stock markt without 10 largest daily changes. Note they tend to be seriously down-biased.

Pierre

by Pierre on Fri Jun 22nd, 2007 at 09:27:04 AM EST
[ Parent ]

fantastic graph (4.00 / 5)

In the long run, we're all dead. John Maynard Keynes

by Jerome a Paris (etg@eurotrib.com) on Fri Jun 22nd, 2007 at 12:28:17 PM EST
[ Parent ]

I think you should be able to do "nice linear parametric [...] computations" with any member of the exponential family (of which the Gaussian is the best-known example).

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Fri Jun 22nd, 2007 at 09:28:28 AM EST
[ Parent ]

Well, I'm not familiar with the generalized exponential distribution framework, but from the primer referenced by the Wikipedia (bold mine):

For most insurance applications, we would expect a decreasing unit hazard functton. That is, as we move to higher and higher layers, the chance of a partial loss would decrease. For instance, if we consider a layer such as $10,000,000 xs $990,000,000 we would expect that any loss above $990,000,000 would almost certainly be a full-limit loss. This would imply h (y) ~ 0.
The decreasing hazard function is not what we generally find in the exponential family. For the Normal and Poisson, the hazard function approaches 1, implying that full-limit losses become less likely on higher layers - exactly the opposite of what our understanding of insurance phenomena would suggest. The Negative Binomial, Gamma and Inverse Gaussian distributions asymptotically approach constant amounts, mimicking the behavior of the exponential distribution

The table below shows the asymptotic behavior as we move to higher attachment points
for a layer of width w.

Distribution Limiting Form of h (y) Comments

Normal lim h(y) = 1 No loss exhausts the limit

Poisson lim h (),) = I

Negative Bmomial lim h,(y) = I - ( I - p ) "

Gamma lim h(y) = I-e "''°'~'

Inverse Gausstan lim h(y) = I-e -''a°''~

Lognormal lim h(v) = 0 Every loss ts a full-limit loss

From this table, we see that the members of the natural exponential family have tail behavior that does not fully reflect the potential for extreme events in high casualty insurance. It would seem that the natural exponential distributions used with GLM are more appropriate for insurance lines without much potential for extreme events or natural catastrophes.

Sorry the layout is crappy, had to rework an OCR-ed pdf.
But the general idea seems to be that if it's linear, then you don't have a fat tail. I don't know if there is any kind of proof of this or of something heuristically similar (I'm by no mean a statistician)

Pierre

by Pierre on Fri Jun 22nd, 2007 at 09:58:04 AM EST
[ Parent ]

Re: Here is his response to mine... (4.00 / 2)

By the same two guys:
http://www.fooledbyrandomness.com/FT-1.pdf
http://www.fooledbyrandomness.com/FT-2.pdf

Pierre

by Pierre on Fri Jun 22nd, 2007 at 09:28:54 AM EST
[ Parent ]

about the only one we can compute daily with existing computing power (I guess this is the same everywhere: it's just one giant matrix multiplication, and this is very efficient)

I could suggest a correction to this in public, but then I'd have to kill you ;-)

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Fri Jun 22nd, 2007 at 09:34:38 AM EST
[ Parent ]

Now this gets interesting...

Do you suggest that there are places where the computing power is much greater than the average bank (that I know, my previous job had more computing power than meteo france and the french dod combined, my new place is a lot more frugal), or that your fund developed some non-gaussian models of risk that can be valued as efficiently as the gaussian with "reasonable" computing power ?

Pierre

by Pierre on Fri Jun 22nd, 2007 at 10:01:56 AM EST
[ Parent ]

No, I'm just saying the guy whose page I linked to upthread seemed enamored with "elliptical distributions" which are parameterised by mean, covariance, and a probability distribution for the Mahalanobis distance, which allow most of the multivariate "Gaussian" matrix algebra to be recycled, and are general enough to incorporate fat tails if necessary.

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Fri Jun 22nd, 2007 at 10:09:13 AM EST
[ Parent ]

Wow, sounds nice. Unfortunately, probably too unconventional for most statistics graduates recently hired in quant teams to pay any attention: considering that it will dog the front business, banks only want to look into something "incremental" to handle the fat tails (that is, some cheap ugly patch of the existing var system that you can recode with a couple of interns). Is the guy working in industry ?

Pierre

by Pierre on Fri Jun 22nd, 2007 at 10:15:57 AM EST
[ Parent ]

Yes, he is in charge of a big chunk of risk model development for a large investment bank.

Can the last politician to go out the revolving door please turn the lights off?

by Carrie (migeru at eurotrib dot com) on Fri Jun 22nd, 2007 at 10:19:32 AM EST
[ Parent ]