An interesting story in a recent Business Week, with the apt title “Financial Models Must Be Clean and Simple” discusses financial modeling in light of the economic meltdown. Along the way, it presents a good explanation of CDOs (Collateralized Debt Obligations – read: bundles of toxic mortgages) and how they work. For the uninitiated, it’s a good entry point into understanding what the hell happened and why the turmoil will continue through at least the mid-term future.

What was interesting to me was the discussion about how financial modeling works, and how it failed. The authors correctly point out that modeling is extremely useful in the scientific world – indispensable, in fact. But modeling physical phenomena is very much different from modeling human behavior – which is essentially what financial models are designed to do. In the physical world, there are rules. We may not know the rules or exactly how they work, but they are consistent for the most part. Contrast that to the human world, where results hinge on the actions of governments, financial institutions, and masses of individuals – none of whom react in a strictly rational manner. Or, more precisely, they react in manners that are rational to them, but that rationality is definitely a product of their own world view.

Wall Street firms have long spent their money acquiring the best scientific brains they could find. When I was in graduate school (University of Chicago Graduate School of Business), some of the most highly sought-after candidates were people who had deep math skills in hard sciences (physics, biology, etc.). While I was sweating out grades in the finance classes, these folks were reading magazines and just waiting around for the math to get hard enough to hold their interest. When they found their homes in the financial capitals of the world, their skills were immediately put to use in building models to figure out the behavior of hideously complicated financial instruments under a wide variety of economic conditions. It takes a lot of computing power to do this, but it also takes a lot of brains to figure out and test the data interrelationships and to make the models sing.

So, fast forward to today. What the hell happened? Why didn’t anyone see the potential for this huge meltdown coming? Were the models wrong, or were they just ignored? What the Business Week authors bring out is that the models used to value CDOs were able to figure out a current value for almost any situation – including interest rate fluctuations, housing price futures, economic growth or contractions, currency effects, etc. The factors that they DIDN’T have a good handle on turn out to be the most important factors of all, including the actual quality of the underlying mortgages. And, more importantly, the probability that a default on one mortgage may be indicative of future defaults on others. This number, referred to as the ‘default correlation’ seemed to be, according to the authors, almost a plugged-in number rather than the result of rigorous analysis. This shocks me. I can certainly understand how these relationships are difficult to quantify and model, but damn it, they are the most critical factors of all – crucial to understanding whether you’re investing in a house or a house of cards. These relationships aren’t impossible to figure out; just very hard. The historical data is out there, and it’s not impossible to regress it against other data and discover at least some of the appropriate variables. For everyone – including Freddie Mac/Fannie Mae, the large banks, and institutional investors – to miss this is almost criminal. What, in my opinion, is truly criminal is how the rating agencies glossed over these risk factors as they gave A+ or AAA ratings to these incredibly risky investments. They need to be held accountable, and there need to be repercussions, but I don’t see that happening anytime soon. The loan originators also deserve scorn, but they were merely sales shills and should never have been trusted.

What happens from here is anyone’s guess. I’ve seen a few recessions in my time, but this one developed very fast and is very deep. However, there are real assets out there backing up the toxic mortgages, and while the value of these assets is certainly down quite a bit, it is almost assured that the values will rise over time – how long a time it will take is the real question. This process can’t begin until some stability is restored and we get to a point where there is confidence in the our basic institutions. That will also take some time, but once it happens, we can begin to rebuild for the future.