Nerds on Wall Street: Math, Machines, and Wired Markets

12.4. Take Them Out and Shoot Them

The designers of these complex derivatives created a vast edifice for derivatives analysts who had the task of building models to price them. There were many such programs, but one that seems to deserve the booby prize for being central to the crisis, and for inspiring equally dreadful imitators, made its home at Fannie Mae:

Fannie constructed a vast network of computer programs and mathematical formulas that analyzed its millions of daily transactions and ranked borrowers according to their risk.

Those computer programs seemingly turned Fannie into a divining rod, capable of separating pools of similar-seeming borrowers into safe and risky bets. The riskier the loan, the more Fannie charged to handle it. In theory, those high fees would offset any losses.

With that self-assurance, the company announced in 2000 that it would buy $2 trillion in loans from low-income, minority and risky borrowers by 2010.^[]

As it turned out, if this was a divining rod, it was held by a suicide bomber. Fannie (and others) built an edifice of enormous complexity and then fed it garbage. Noted systems analyst Alan Greenspan elucidated the old garbage in, garbage out maxim for a congressional committee, to share the blame with the monster models of Fannie Mae:

It was the failure to properly price such risky assets that precipitated the crisis.... In recent decades, a vast risk management and pricing system has evolved, combining the best insights of mathematicians ...