Model Behavior


Physicist Emanuel Derman, who spent 17 years as a quantitative analyst at Goldman Sachs, is the author of Models. Behaving. Badly: Why Confusing Illusion with Reality Can Lead to Disaster, on Wall Street and in Life (Free Press). reason asked Derman for three ways overreliance on quantitative modeling causes trouble in economics.

1 Economic models are mostly statistical regressions or based on crude mechanical analogies with simplistic dynamics. Because they use advanced mathematics, these models look like physics, but the resemblance is shallow. Economists put too much trust in their Tinker Toy models. And governments put too much trust in economists.

2 Economics is about what's good for us and how to get it. In British universities, it used to be taught as a part of the triple concentration in politics, philosophy, and economics—the "moral" sciences of value. Nowadays it's misleadingly taught as part of a group that includes mathematics, psychology, neuroscience, and economics—as though it's a value-free quantitative science. 

3 The most misused tool in economics is mathematical optimization. Your car's GPS system can optimize over a choice of roads whose routes and lengths are accurately measured and remain fixed. But an economist optimizes over inaccurate current estimates of future economic scenarios, none of which will actually come to fruition, because they are guesses at future social behavior. Under such circumstances, optimization is a waste of time, lending false comfort.