Mileage

Eric Morris at Freakonomics asks what people know about their gas mileage.

Quick, how many of you can tell me:

1. Your cars’ fuel economy in miles per gallon or, even better, gallons per mile.
2. How much you drove in the last year.
3. The cost to fill your tank.
4. Your monthly and annual fuel expenditures.
5. How your cars’ fuel economy sits in relation to other cars in their classes.
6. What your fuel savings in gallons and dollars would be if you switched to a hybrid or other highly economical vehicle.

The most economically salient questions are probably 4, 5, and 6, but I found the others easier. I check my mileage frequently (~26 mpg) and I know that 1/26 is close to 0.04. I know how far I drive in a year and I know how much I spent to fill my tank last time but it fluctuates a lot. The others require calculation.

For most people, says Morris, fuel economy didn't even influence their purchase decisions. Well, it surely did influence mine, but hardly minutely. Even if I didn't do the arithmetic, I knew that I couldn't face putting a $100 bill in a Hummer's gas tank 6 or 7 times a month.

Would it have been better to do the math? Well sure, abstractly it would have been more precise. I can do the math. I do math for a living. But I just didn't feel like bothering in the case of my car purchase. Why not?

I don't exactly know the answer, but I can think of a few reasons: gas mileage was one factor, not the sole factor, possibly I feared that if I calculated it all out I would have bought some car I liked less. Most human calculations are done by satisficing - picking a "good enough" option rather than detailed searching for an optimum. It's a strategy that works well for lots of things and minimizes hard calculation, but sometimes it can fail big time. It's a giant hole in the theory of rational expectations.