Summary: What does a GPS have to do with the recession?Like many, I use a GPS app on my phone when I drive to a new place. Normally, this is a convenient way to get from A to B—no fussing with maps and obscure directions: “Turn left a half mile after the second railroad crossing where the 7/11 used to be, I think they replaced it with a Benny’s.” But when the GPS malfunctions, or has the wrong map coordinates, it can literally lead me into trouble: one-way roads, missing landmarks, or other difficulties.Part of the problem is when we rely on a tool, some of our other skills get rusty. With the GPS, the effect has been significant. It’s hard to find detailed maps any more. And I’m not as familiar with the landscape as I used to be, since I now have that endless road scrolling in front of me on my smartphone.What does this have to do with the mortgage crisis? Banks developed a new risk-management tool in the ‘90s: Value at Risk (VaR). It used financial theory and empirical inputs to measure the how risky a bank’s portfolio might be. The tool was fine. But some of the inputs were wrong—or even backwards. Experience told the risk managers that housing values don’t fall nationwide. Oops. Also, their volatility measures were about half of what they should be.As a result, when things got dicey, senior management seemed out-of-touch. They had come to rely on a tool that told them that their risk was only a fraction of what it actually turned out to be. Incidentally, a similar thing happened with the London Whale trade. Bad VaR inputs led to terrible trading results.The problem isn’t the tool, it’s data input. In the early days of computing we had a phrase for this: garbage in, garbage out.
