Proponents of raising the gas tax — and the chattering class is littered with them these days — have a simple, central argument: gas taxes create a market signal that pushes all vehicle consumers in the direction of fuel efficiency.

Indeed, some conservatives (and car companies) go further: they say that CAFE standards are bad policy because they force automakers to create products for which there’s no demand. It’s no good making fuel-efficient cars if nobody wants to buy them! (Americans love big, powerful cars. “Everybody knows” that.) Higher gas taxes should replace CAFE, because they create create demand instead of forcing supply.

A moment’s thought reveals a serious flaw in these arguments. Fuel costs are a relatively low portion of total vehicle costs — maybe 10-20 percent. There’s maintenance, insurance, parking, but most of all, the price of the car.

And when the time comes to buy a car, people don’t behave like the rational interest maximizers of economic myth. They rarely calculate out future costs like fuel. They consider the number on the price tag in front of them: the price of the car.

It follows that if you want a market signal, you should put it where it will have the most effect: on the price of the car.

As it happens, we have a policy like that! Let’s hand the mic to John Heywood, who has headed the Sloan Automotive Laboratory since 1972:

I think we need a purchase tax, a feebate system, like the French have instituted fairly recently. Fees for high-consuming vehicles and rebates for low-consuming vehicles. That will help reinforce consumer response to CAFE requirements by providing a market incentive.

There you go. A clear price signal, applied at the point of maximum effect, supplementing rather than replacing fuel efficiency standards. CAFE standards push automakers to make fuel-efficient cars; feebates push consumers to buy them. (Oh, and unlike gas taxes, feebates aren’t regressive.)

How is this not a preferable policy, both economically and politically? What am I missing?