Wednesday, May 9, 2012

On models

Last week, I wrote a couple of posts that touched on the usefulness of economic models (and other models) in our daily lives. I essentially argued that we need to take responsibility for our own decisions, rather than hiding behind spreadsheets and models that "tell us" what to do (and then asking for bailouts when the model gets things wrong). A quick reminder:
We all use models in our daily lives, because they help us to make sense of what are often very complex problems. Models simplify, organize, and categorize the variables in an uncertain world so that we can better understand the impacts of our decisions. But they DO NOT, ever, have the power to tell us what to do. You don't even need to know a thing about Black-Scholes (and trust me, a lot of people who should know a lot about it... don't) in order to accept that assertion as fact. 
The intelligent person knows to use a model only as a guide to confirm (or refute) what our intuition tells us. Very often, our painfully simple heuristic models (which you can learn or hear more about from Gerd Gigerenzer's speech, if you're a nerd like me) actually outperform very elegant statistical models. 
I still like what I said, but I came across an old line this week from statistician George Box that was significantly more efficient and eloquent than my rambling polemic. Said Box,
All models are wrong, but some are useful.
Yeah... that's what I meant to say. I think I'm gonna tack that one up on my wall, just in case I ever forget it.

No comments:

Post a Comment