11/03/2011 06:16 pm ET | Updated Jan 03, 2012

Car Buyers Seem to Be Rational to a Fault

Here's the classic economic view of your car-purchasing behavior: You walk into a dealership, choose a car based on brand, color, cylinders, looks and general feel and then start comparing prices among different options. And you don't just look at how much you pay to drive home with the car, you also include all likely future expenses. You look at maintenance costs and might decide that paying a bit extra for a Mercedes is worth the upfront expense because it tends to break down less often than the Yugo next door. You also look at gas mileage, today's price per gallon, form an opinion about future gas price trends, attach probabilities to them, calculate expected total gas costs over the lifetime of the car, balance all of that information against expectations over future inflation rates and interest paid were you to just leave your money in the bank, take into account how your preferences for driving will evolve over time, make a few assumptions about how future buyers will perceive your choice when you are ready to sell the car, and do all that and probably a few things I'm missing, while the car salesman at the dealership explains to you the awesome industry-leading warranty and zero-down loan program offered through the end of the month.

To be fair, no sensible economist really thinks this is how you buy a car. Individuals make mistakes, focus on the wrong things, and manage to remember only some of the points that they should, as they are lulled by the dealer's promise of the free child seat with any car bought with the DVD player and nine-inch screen built-in for the little one. It's just that on average car buyers behave as if this model holds.

It's easy to make fun of the "rational" economic model. Of course, no one does these things in real life.

In my book, I cite a working paper by Hunt Allcott and Nathan Wozny, who try to measure to what extent people behave "irrationally." In theory, car buyers ought to be "willing to pay one extra dollar in vehicle purchase price to decrease the expected present value of future gasoline costs by one dollar." You'd expect drivers to think every dollar is worth 100 cents no matter where they find that dollar. They don't.

In their working paper, Allcott and Wozny found that car buyers only pay 61 cents for every dollar of gas saved. You can look at this number and talk about the failures of the rational model.

Or you can look at this and say that the zombie-economic, rational model gets us 60 percent to the true answer. Not bad for something whose assumptions no one believes.

Allcott and Wozny have since re-run the numbers. Their latest figure: 72 cents.