Tales From the Financial Crisis Conference - Psychokillers, Bad Math & Burn Baby Burn

On Friday there was a conference on the financial crisis. It was held at Columbia University. Some very good interviews came from Economists Stiglitz, Volcker, Phelps and Bhide.

Below is an interview with Economist Joseph Stiglitz clearly stating the United States is putting good money after bad money. This is a fact we, those insignificant regular folk, have been saying for some time. Stiglitz also reviews what should be done with insolvent banks, discussion of Treas. Sec. Geithner's stress test and why a stress test based on a bad model or wrong assumptions is a problem.

 

At the same conference, Economist Paul Volcker acknowledges the frightening overall global economic indicators occurring. This is no ordinary crisis, the mother of all financial crises.

 

At least Volcker is acknowledging the issues and also he poked fun at people believing financial systems follow normal distribution patterns. (He is referring to a Gaussian distribution system which is commonly used in probability models and of course is another bad math assumption/model flaw. The world is clearly not a bell curve in time).

Volcker also pokes fun at those conspiracy theorists who think the Federal Reserve is a serial killer psychopath. He then refers to business cycles and how these are inequalities and imbalances in the system which can become like standing waves (hey Volcker, here's some more physics fer ya!) and if not prescribing corresponding corrections...can become a tsunami.

While Volcker, a very funny guy, talks about financial engineers in a disparaging way, he finally acknowledges what one really had with these financial models and financial engineering. Truly these derivatives are a more glorified Enron accounting method siphoned through these structured financial instruments. The only real differences are these guys used advanced mathematics to cook the books.

 

An even more damning interview on handling toxic assets with Joseph Stiglitz:

 

Additional videos are Economist Edmund Phelps calling for a new type of bank in which loans, invests, promotes innovation in the business sector. I agree with this although I am not sure if he is talking about a new type of Venture Capital or revamping commercial banking. He says the financial sector has lost the know-how on how to invest in the business sector. That one I believe since the financial sector has seemingly been so active in preying upon the consumer and middle class.

 

Finally here is Amar Bhide, from the Columbia Business School, referring to derivatives and the insanity thereof. I question whether the probability models were precise at all, for packaging up a series of securities and claiming they will vary with probability p might be a completely incorrect model, say implying each entity is an independent event, or who knows!

I believe at this point we need all of the math geeks to obtain the specific models of these CDOs in question so we can all take a look and see if they suffer from bad math.

In other words do we have a bunch of mathematical voodoo and jargon thrown out there which actually, by the mathematics itself, is completely invalid?

Who knows but finding the mathematical models themselves....discussion of these derivatives structures by the mathematics...well, someone put a link in the comments or a paper title if you know where to find the models, equations and assumptions themselves. Even code at this point....just give me the formula, give me the formula!

So for now, let's just leave it at this:

Ya all know Black Jack, Craps, Poker and Roulette also have probability distributions and models right?

Meta: 

Comments

This is a great collection of remarkable statements

Our current Secretary of the Treasury, Timothy Gethner,said this in early 2007:

As head of the New York Federal Reserve Bank, Geithner gave an interview to Jenny Anderson of the New York Times in Feb. 2007. When asked about the high risk credit derivatives market, a risk he claimed that he'd addressed, Geithner said: "The fact that the banks are stronger and risk is spread more broadly should make the system more stable. We can’t know that with certainty though. We’ll have a test of that when things next threaten to fall apart." Will anybody mention that we've had Mr. Geithner's anticipated "test" and things did "fall apart" because the banks were weaker not "stronger." Link

He sounds like he's conducting a project for the science fair.  At the very least, Geithner is stunningly wrong on the credit derivatives market.  Not the type of guy in charge who gives you a sense that somehow we'll survive this crisis.

Excellent collection of very useful video.

What is the point of the stress test?

AISI, it's just cover, it can either be used to justify more bailout

1. Well, we checked them out, things are grim but not irretrievable, and if we give them another bailout we'll avoid a complete crash

or nationalization

2. Dead banks walking. Receivership time.

Everybody already knows the answer is 2, but apparently many find this difficult to accept. Is it just process cover?

BTW, good catch on the conference.

Citigroup now negotiating

for more money in exchange for stock.

There was a great panel with Krugman, Roubini on ABC this morning and it's really looking like they plain need to deploy the Swedish plan.

The stress test might be a method to nationalization as well as a method as you say. Or it might be useful for responsible banks who really are fine.

Good point

It's not just Citi, BOA, WF ... getting cynical in my old age.

It's not just the math

it's the basic axioms behind the math:
- just because everybody doesn't *normally* remove their money from the bank all on the same day at the same minute doesn't mean it can't happen.
- just because you can insure something doesn't mean that the insurance company knows what they are doing or that your insurance policy will be any good if something happens.
- just because you can diversify by investing in several contracts of the same type on the thought that they won't all go bad at once, will not prevent them from all going bad at once.

In other words, your models should always assume the worst, not the best.

-------------------------------------
Maximum jobs, not maximum profits.

that is the math

every model has a set of assumptions, limits, boundary conditions. One can have the actual equation being incorrect but more often it is the assumptions that are incorrect or ignored.

I think it's time to install LaTeX and go through the actual models and point out, in mathematical language, the problems. Maybe on the upgrade.

I see no reason

To add any more assumptions than the basic math axioms. Any model that does so should be inherently suspect of support of fraud.

But then again, I'm one who thinks that fractional reserve banking and stock markets are a form of fraud.

-------------------------------------
Maximum jobs, not maximum profits.

just a little reminder

in mathematics, an axiom is a universally held truth.

i.e. 1+1=2 (base 10), or prime numbers are only divisible by themselves and 1. Logic would be (NOT A) OR (NOT B) = NOT(A AND B) and so on.

what is being referred to in the above is two fold, firstly are structured financial instruments and then mathematical modeling. Mathematical modeling is a system level technique, always has a series of assumptions and must be crafted to empirical evidence and analysis.

Then these various derivatives, were based on stochastic modeling as well as often truly not simulated (it appears) and (it appears) no analysis in terms of a system model.

I'm not sure yet on these structured finance details for I am reading up, but axioms is not applicable to the issues raised by these experts.

The point is

That you shouldn't bet the bank on *ANY* assumption. Axioms notwithstanding, because they are universally held truths.

If your assumptions aren't based on evidence, that makes them a little less useful than an assumption that is based on evidence. But even an assumption based on evidence can yield to the outliers in the data for which you have no experience at all.

To my way of thinking, anything beyond the rock solid guarantee goes outside of good science and business- and into the realm of fantasy and gambling.

-------------------------------------
Maximum jobs, not maximum profits.

by this reasoning

no bridge, no building, no road, should ever be built. You want to know why? Because in all of those models, which are systems in essence, have scenarios where they would break down. Examples are a 100 foot tsunami, say Yellowstone erupts, a meteor hits the Earth, suddenly 10,000 trucks all loaded with heavy metals cross a bridge in 100 mph winds at the same time...

list goes on and on.

That is the point of proving a model, it must correlate to statistical evidence, which is the inputs, the data, that it is a law. One can prove that some variables will result in a different output.

That is the point of the original commentary, it is implied and we know they did it, banks were leveraging out 40;1, 50:1, 70:1 instead of the accepted 12:1 previously...
but the main point is these corporations created a shadow banking system with some obscure advanced mathematical models that know one understand and I implied they need to get the people who do understand, the mathematics community to look at these things and expose the mathematical fiction they probably are. i.e. bad math.

I don't want to sit here on EP reviewing mathematical definitions, concepts, etc. and if the mathematical concepts are not understood, well, we have wikpedia and many textbooks to help out.

The origin of the problem.

I will give you this link which, I think, answers your question. At least, it is the best explanation I have found to date (h/t Angry Bear).

thanks

more details on the specifics can only be good. We have a very large database and on the upgrade I'm specifically looking for web tools to display complex mathematical equations, graphs, spreadsheets better.

(if anyone sees these types of tools anywhere, let me know for I'm on the hunt to incorporate any tools to display economic data, graph it, pie chart it, etc. easily that I can find or write).

Gaussian Equation

I read that, too.:-)