Madoff made-off with billions...

Since we're sorta kinda discussing economic theory here, this thread might be a good place to put this bit in Financial Times that I came across: it's basically a scathing attack on macroeconomics as it's been (generally but not quite universally) practiced for the past several decades. In particular, it whacks the efficient markets hypothesis that's at the core of Classical, Keynesian, new Classical, and new Keynesian theory (and, FWIW, implicit in Austrian school economics that mudsling likes so much).

[ http://blogs.ft.com/maverecon/2009/...state-of-the-art-academic-monetary-economics/ ]

It's somewhat heavy reading, but if you're at all interested in how come so many so very smart people got it so wrong for so long, it's definitely worth it. (It's also gratifying to notice that some of the people I've personally found most enlightening all along are cited as positive exceptions to the general rule.)
 
Joined
Oct 19, 2006
Messages
8,540
If you add information from this artice: http://www.wired.com/techbiz/it/magazine/17-03/wp_quant?currentPage=2 a pattern seems to emerge: with Gaussian copula function and Efficient Markets Hypothesis economist thought that they have finally found The System which turns unpredictabe into predictable. And in both cases what was willfully disregarded (I can't believe that it was simply missed by so many) was the fact that (to adopt a quote from the atricle you linked PJ) "It excluded everything relevant to the pursuit of financial stability" or, to quote from Wired article: "In the world of finance, too many quants see only the numbers before them and forget about the concrete reality the figures are supposed to represent. They think they can model just a few years' worth of data and come up with probabilities for things that may happen only once every 10,000 years. Then people invest on the basis of those probabilities, without stopping to wonder whether the numbers make any sense at all."
 
Joined
Jan 10, 2008
Messages
4,721
They think they can model just a few years' worth of data and come up with probabilities for things that may happen only once every 10,000 years. Then people invest on the basis of those probabilities, without stopping to wonder whether the numbers make any sense at all."

Lunatics . . . . we only have to model to one in 200 for our capital, but even that we have to speculate quite a bit and err very much on the side of caution. And we've got a whole chapter of potential disaster scenarios beyond that level that speculate on totally ridiculous possibilities just to see whether there is anything that could break us.
 
Joined
Feb 2, 2007
Messages
2,351
Location
London
We're living in the Age Of Greed.
 
Joined
Nov 5, 2006
Messages
21,908
Location
Old Europe
Since we're sorta kinda discussing economic theory here, this thread might be a good place to put this bit in Financial Times that I came across: it's basically a scathing attack on macroeconomics as it's been (generally but not quite universally) practiced for the past several decades. In particular, it whacks the efficient markets hypothesis that's at the core of Classical, Keynesian, new Classical, and new Keynesian theory (and, FWIW, implicit in Austrian school economics that mudsling likes so much).

[ http://blogs.ft.com/maverecon/2009/...state-of-the-art-academic-monetary-economics/ ]

It's somewhat heavy reading, but if you're at all interested in how come so many so very smart people got it so wrong for so long, it's definitely worth it. (It's also gratifying to notice that some of the people I've personally found most enlightening all along are cited as positive exceptions to the general rule.)

Anytime there is a big downturn, we see articles like this. They aren't necessarily wrong, but they just have their biggest audience in these times.

I don't think we have it wrong, we just don't have it completely right, and not having it completely right can cause some very nasty times.

The main issue is that economics is the study of something that is fundamentally unobservable. Sure you can observe parts, but never truly the whole, and you can't model every possible input, it's just not possible. That causes certain frictions to be left out of the equation.

One of the issues with this is something that I deal with in risk management each day. I can tell you with a 99% confidence interval, what will happen in various situations given various events. The problem is that our understanding of those probabilities is based on historical data and things sometimes happen in ways that they haven't happened before.

I don't think we have it fundamentally wrong, but we do have a long way to go towards a better understanding of it.

One last note. Maybe I'm just an uneducated American, but damn, while grammatically correct, that guy's writing style hard to read!
 
Last edited:
Joined
Jan 10, 2008
Messages
4,353
Location
Austin, TX
We're living in the Age Of Greed.

Agreed. I know that the US has been the worst on it, but the pursuit of material wealth at all costs has been very damaging to society as a whole.
 
Joined
Jan 10, 2008
Messages
4,353
Location
Austin, TX
I don't think we have it wrong, we just don't have it completely right, and not having it completely right can cause some very nasty times.

Yes BN: "But in the CDO market, people used the Gaussian copula model to convince themselves they didn't have any risk at all, when in fact they just didn't have any risk 99 percent of the time. The other 1 percent of the time they blew up. Those explosions may have been rare, but they could destroy all previous gains, and then some"
 
Last edited:
Joined
Jan 10, 2008
Messages
4,721
@bn, I don't entirely disagree. But the man has a point -- if meteorologists could predict the weather in Florida with uncanny accuracy not counting hurricanes, wouldn't you say that they're missing something rather fundamental?

To go on a bit of a tangent, back when I was at the university, I developed a strong sense of danger about theory. That's because I got to know lots of extremely smart people who believed completely absurd things, because that's what their pet theory told them to believe. I treat theories like tools: they can be highly useful for understanding and conceptualizing things, but they're always fundamentally limited. If you're not aware of what those limitations are, you run a risk of outrunning them, applying them in areas where they stop working, and then mistaking the result for reality. This is an especially big danger in disciplines without strong built-in reality checks: sociology, history, psychology, and, yes, economics. Not so much in "hard" sciences where the results are usually empirically verifiable with little ambiguity.
 
Joined
Oct 19, 2006
Messages
8,540
You also learn in university that theories never work. :p

Most of them are simply abstractions of reality. Models that human mind can understand. People find one pattern or theory and then go and apply it to everywhere. "When you have a hammer in your hand, everything looks like a nail."

Sorry to continue this thread more off topic. Agreed on the Age Of Greed. It was supposed to work well in theory I guess. ;)
 
Joined
Dec 13, 2006
Messages
233
Yes BN: "But in the CDO market, people used the Gaussian copula model to convince themselves they didn't have any risk at all, when in fact they just didn't have any risk 99 percent of the time. The other 1 percent of the time they blew up. Those explosions may have been rare, but they could destroy all previous gains, and then some"

A gaussian copula? AFAIK that's just based on a bivariate normal distribution so is a fairly symmetrical distribution. For something like this where the extremes are catastrophic market contagion they should have been using a far more skewed distribution & copula, like a gumbel or pareto.

They'd have been MASSIVELY understating the tail of the distribution using that.
 
Joined
Feb 2, 2007
Messages
2,351
Location
London
@bn, I don't entirely disagree. But the man has a point -- if meteorologists could predict the weather in Florida with uncanny accuracy not counting hurricanes, wouldn't you say that they're missing something rather fundamental?

Given that year in and year out scientists are off huge amounts in their predictions of named storms for the coming year, I'd say yes they are missing something fundamental.

I treat theories like tools: they can be highly useful for understanding and conceptualizing things, but they're always fundamentally limited. If you're not aware of what those limitations are, you run a risk of outrunning them, applying them in areas where they stop working, and then mistaking the result for reality. This is an especially big danger in disciplines without strong built-in reality checks: sociology, history, psychology, and, yes, economics. Not so much in "hard" sciences where the results are usually empirically verifiable with little ambiguity.

I completely agree with you.
 
Joined
Jan 10, 2008
Messages
4,353
Location
Austin, TX
Given that year in and year out scientists are off huge amounts in their predictions of named storms for the coming year, I'd say yes they are missing something fundamental.

It'd certainy be interesting if anyone could come up with some good predictions. They are getting a lot better all the time though, to take the two highest profile cases lately -

Katrina they got very wrong partly because they didn't make sufficient allowance for the risk of the levees breaking and partly because they didn't have a feedback loop in the model that recognised that when there is a really massive storm like that then the repair costs go through the roof because there's a ridiculous demand surge, along with things staying out of action for longer because nobody can be found to repair them so the business interruption / alternative acccommodation costs are far higher.

Ike they got very wrong because it was a far wider storm than anything that had come previously (normal enough strength and took a fairly conventional path), which meant that far more oil platforms needed to be closed or moved (some are mobile) and the business interruption went through the roof, especially since it coincided with the oil price being pretty well at its ridiculous peak which nobody had anticipated.

Anyway, leaving aside the fact that they are continually refining and improving their models, for something of this complexity they will always be wrong on a single year / single storm / single location basis. Beyond that though they aren't too bad, one can write a profitable reinsurance account over a 5+ year period using them.
 
Joined
Feb 2, 2007
Messages
2,351
Location
London
That's rather the point. Both economics and meteorology (and climatology) are statistical disciplines. The difference is that meteorologists can tell with a remarkable degree of accuracy how many cat 3, 4, and 5 hurricanes will make landfall over the next 10, 30, or 50 years -- but economists can't make any such predictions about cat 3, 4, or 5 recessions. What's more, meteorologists can also predict pretty accurately if, where, and when any given storm will make landfall, with sufficient warning to evacuate the population. Economists don't appear to be able to predict hurricanes even when they're right on top of them, Krugman, Roubini, and a few others notwithstanding.
 
Joined
Oct 19, 2006
Messages
8,540
:lol: link

One of the new toys on display at the 2009 Toy Fair in New York is "Smash-Me Bernie," a figurine of Bernie Madoff, the U.S. financier charged with running a $50 billion Ponzi scheme. "Smash-Me Bernie" wears a devil's red suit, carries a pitchfork and comes with a golden hammer for smashing the doll on the head. Madoff's former clients might not be in the mood to shell out the cost of the figure -- $99.95, plus shipping.
 
Joined
Feb 2, 2007
Messages
2,351
Location
London
Looks like Bernie isn't going the plea bargain route after all for some reason:

Madoff faces life in prison


The government said in court documents it wants Madoff to forfeit all of the money and property that can be traced back to the alleged fraud -- a sum it estimated at more than $170.8 billion....Madoff was charged with securities fraud, mail fraud, wire fraud, three counts of money laundering, making false statements and perjury among other charges, according to the court documents. He faces up to 150 years imprisonment, according to sentencing guidelines.

There is no plea agreement with Madoff, who remains under house arrest.
 
Joined
Oct 18, 2006
Messages
7,834
Interesting tidbit from discussion with some of our underwriters - some of our existing US fraud or directors liability claims are possibly going to go down quite a bit because the lawyers fighting them are keen to push them into a commercial out of court settlement so they've got free time to milk it on madoff & stanford court cases.

Lawyers, always acting in your best interests eh?
 
Joined
Feb 2, 2007
Messages
2,351
Location
London
That's rather the point. Both economics and meteorology (and climatology) are statistical disciplines. The difference is that meteorologists can tell with a remarkable degree of accuracy how many cat 3, 4, and 5 hurricanes will make landfall over the next 10, 30, or 50 years -- but economists can't make any such predictions about cat 3, 4, or 5 recessions. What's more, meteorologists can also predict pretty accurately if, where, and when any given storm will make landfall, with sufficient warning to evacuate the population. Economists don't appear to be able to predict hurricanes even when they're right on top of them, Krugman, Roubini, and a few others notwithstanding.

Well one problem with predicting out that far is simply that there are too many events you can't model. How do you model war? Or a terrorist attack?

You really can't compare the two. The weather involves highly complex patterns, sure, but they consistently follow a logic that can be broken down. Human behavior is often illogical and can't always be predicted.
 
Joined
Jan 10, 2008
Messages
4,353
Location
Austin, TX
Well one problem with predicting out that far is simply that there are too many events you can't model. How do you model war? Or a terrorist attack?

Heh.

We use external services like those provided by GLobalInsight or ExclusiveAnalysis that come up with country ratings in various categories, that we've built into a model.

As to whether it's right or not nobody will ever know because there just won't ever be enough data for it, but it's a system at least.
 
Joined
Feb 2, 2007
Messages
2,351
Location
London
Back
Top Bottom