How Models Caused the Credit Crisis

Ryan Chittum asks if we want to know what caused the global credit crisis, and suggests that if we do, we should "start here" [pdf file], with a 22-page report about subprime lender IndyMac (IMB) from the Center for Responsible Lending and former WSJ reporter Mike Hudson.

The report certainly manages to be both shocking and depressing at the same time. But by this point it's well known that subprime lenders often behaved in an irresponsible and predatory fashion. What's more, that behavior wasn't a cause of the global credit crisis, so much as a symptom of it.

As Wolfgang Münchau says in today's FT:

If this had been a mere financial crisis, it would be over by now. The fact that we are suffering its fourth wave tells us there might be something at work other than merely financial euphoria and bad regulation.

This is true, even if you don't buy Münchau's assertion that the real cause of the crisis was New Keynesianism and the dynamic stochastic general equilibrium model in particular. I would rather place the blame at the acceptance of models in general. Gillian Tett explains:

This decade, financiers have invented so many brilliantly clever mathematical tools to repackage risk that the industry has slipped, almost unthinkingly, into an assumption that "credit" is a collection of abstract equations, stripped from any human context.

Thus banks have become so dazzled with their powers that they have ignored how they interact with the rest of society - or how the tribal aspects of their own institutions can create dangerous traps.

Meanwhile, the cult of models has become so extreme that banks have believed them even when this collides with common sense. Yet, as any Latin scholar knows, the word "credit" hails from credere: "to trust". It is, in other words, also a social construct.

And bankers forget this human dimension to their cost.

Sam Jones has one striking visual of that cost: a bar chart showing the current ratings of 469 CDO tranches which were rated AAA at issue. Fewer than a quarter of them retain that top rating; a lot of them are now CCC rated, and a fair few even have a D rating.

The reason that IndyMac was writing so many horribly bad mortgages was that there was no shortage of investors willing to trust models telling them that the bonds secured by those mortgages were incredibly safe. They didn't need to look at the mortgages themselves, since they had bankers using models to do that for them.

In other words, IndyMac's behavior was certainly irresponsible, probably illegal, and also entirely predictable. Could we have known that IndyMac specifically would extend a loan against a stated Social Security income of $3,825 a month, even when the maximum Social Security income at the time was barely half that? No. But inevitably someone would, just because of the ease with which lenders were able to sell all their downside at a substantial profit.

Hudson concludes his report by calling for "rigorous oversight" of lenders, and "rules that will prevent such disasters from happening again". But that's only half the solution. The real art is to try very hard to design a financial architecture where rules and incentives work with each other, rather than in opposition to each other. Because when there are billions of dollars to be made by breaking the rules, you can be sure that the rules will end up being broken.

Comments

[bbzz24:]

and if you look fat and not sexy in victoria's secret underwear immediately blame marissa miller & gisele bundchen!

2008 Jul 07 10:16 AM

[Edward Janeck:]

I think that Münchau has a point but in my view the global derivatives caused a large portion of the problems. Nobody understands how they will act long-term.

2008 Jul 07 10:42 AM

[wholesaler:]

Interesting article, but very biased. Perhaps the weak link in this scenario could be shared by the originator/broker.

2008 Jul 07 11:16 AM

[pochovilla:]

Greed + stupidity make a dangerous potion.

2008 Jul 07 12:56 PM

[GMen:]

There's 2 sides to every story and the truth is probably somewhere in the middle. Seems as if the Center for responsible lending has an "Agenda". IMB's response to what they call a "Hit" piece below.

Indymac Bank has updated its corporate blog, theimbreport.com, with its response to a shoddy (at best) hit piece issued yesterday by an organization calling itself the Center for Responsible Lending. This report is essentially a regurgitation of unfounded statements made in lawsuits against Indymac or by those participating in those lawsuits. Indymac was not contacted in advance of its publication or given the opportunity to review or comment upon it or the sources it cites. While this hatchet piece bills itself as an objective case study about just mortgage lender (Indymac), the reality is that this organization has released similar “reports” about many other lenders, including Wells Fargo and Freddie Mac.

While every lender made mistakes (and you can see our 2007 Shareholder Letter for a listing of them), Indymac made over 1.7 million mortgage loans over the past 15 years and nearly 98% of those loans have been successful. Indymac feels that a truly responsible case study would present the facts about these matters, positive or negative, rather than just the unfounded and apparently unchallenged assertions of trial lawyers.

2008 Jul 07 01:13 PM

[Max Dama:]

Salmon,

You seem to think that a revised rating means that the model has failed. A rating represents the likelihood of default assuming normal market conditions. Since conditions are significantly worse than they were during the rating process, we should expect revisions beyond the historic standard deviation of frequency.

The latin root of a word does not signify a deep immutable meaning.

"Models in general" (such as mechanical models of a bridge, or electrical models of a circuit, or economic models of supply and demand) are NOT to "blame". It seems supersticious to even consider.

Regards,
Max Dama

2008 Jul 07 02:34 PM

[Matt Moscardi:]

I disagree with the premise that "models" were the smoking gun of the subprime crisis. The real issue is much more mundane - greedy people and corporations. IndyMac underwriting bogus mortgages to unqualified homeowners smacks of pure greed, not "the model said so." When the loans were repackaged, even then it wasn't "the model," it was a lack of oversight, foresight, and understanding by the eventual purchasers - institutional and retail investors alike - what it was they were buying. Whether by stretching the truth or just plain lying, the end purchaser was unaware of the process, because the middlemen, in this case the Street and the IndyMacs of the world, wanted the quick buck.

It's the same story all the time. In the end, it will take a massive ideological overhaul of the Street, one that promotes corporate accountability and sustainable business practice over short term gains, to fix it. So really, it will never happen. Every 10 or 20 years this will happen again (see the real estate bust in the 1980s).

(Here's where I take a moment for shameless self promotion) To that end, I say invest sustainable, green, and socially responsible (see my website). It's not a solution, but at least it's a start, and you can sleep a LITTLE better!

2008 Jul 07 02:37 PM

[THofler:]

A several years ago a famous Berkeley physics prof complained to me that few of the PhD students he encountered were interested in slaving in his lab for a few years because most of them were not interested in a career in physics. They were in a hurry to get to Wall St.

And before that crew hit the street, there was the crew of physics wash-outs (& wised-ups) that went there.

Having built a few computer models & seen a great many presentations of others, I'd cynically suggest that the modeler's confidence of the model's applicability to the problem at hand is proportional to the the number of man months or years invested in the model. Even in the area of physical science these models frequently miss major portions of the realities at hand.

I'd suggest to Max that the assumption of a "normal" market when applying a financial model is invariably a mistake. Maybe someday we will have models that include factors to compensate for a variety of possible market & economic shocks, but so far collapses of the LTCM type seem to be the norm.

Over reliance on models badly applied, is not the whole problem, but a big contributor. My guess is that the substantial contributors are: a very preferential tax structure for real estate; Greenspan easy money; Greenspan lax regulation of mortgage brokering; bad models applied badly by lenders; grotesquely irresponsible and ignorant ratings agencies; and financial illiteracy about debt in the general population.

2008 Jul 08 01:19 AM