A comment over at Brad DeLong's blog, in his discussion of this earlier post, suggested that my criticism of CAFE standards is off the mark:
Samwick is simply wrong.
CAFE standards, historically, are extremely effective at increasing efficiency.
The amount of tax you'd need to begin approaching the efficiency under CAFE would be pretty steep. While it could be done in principle, as commenter Rob points out above, it's very unlikely the necessary gas taxes would ever be passed.
I suspect we are not writing about the same exact concepts (e.g., efficiency versus usage), but it did prompt me to recall some academic work on the subject. The best paper I have ever seen presented on this topic is by Penny Goldberg, now at Yale. As far as I can tell, it is still state of the art in the economics literature. The full citation is:
Goldberg, Pinelopi Koujianou. "The Effects of the Corporate Average Fuel Efficiency Standards in the US" The Journal of Industrial Economics, Vol. 46, No. 1. (Mar., 1998), pp. 1-33.
And for those of you with access to JSTOR, you can get it here. She's got a pretty slick model of both automobile demand (including the new and used markets) and the non-competitive elements of the automobile supply side. Using data from 1984-1990 to estimate the model, she does some simulations that compare the effects of CAFE (as it existed in 1989) with a gasoline tax. First, she simulates what would happen if the CAFE standards were eliminated:
The results from these calculations ... indicate that CAFE standards lead to approximately 19 million gallons fuel consumption savings per year. This figure can only be interpreted as a one year ahead effect ... a rough idea about the total impact of the 1989 standard can be obtained by calculating the annual savings in the "steady state," after the current vehicle stock has been completely turned over and replaced by vehicles subject to the 1989 regulation. Assuming that there were no further changes in the standard after 1989, and that the total number of cars on the road remained constant ..., the 1989 CAFE standard alone would lead to approximately 400 million gallons of annual fuel savings.
Even thought this figure is much larger than the 19 million gallons estimated for the first year, it still represents only a small fraction of the approximately 130 billion gallons of gasoline consumed in the US every year.
Okay, so 0.4/130 = 0.3%. Not a very big dent in usage at all. But this doesn't mean that a gas tax is much better, as she shows in her next simulation:
[W]e can compute the gas tax increase necessary to achieve the same fuel savings as with the CAFE regulation. According to our results, this increase would have to be 780% or 80 cents per gallon, implying almost a doubling of gasoline prices.
The gas tax at the time was 10 cents per gallon. So this suggests that neither policy, as they have been implemented to date, have generated much of an impact on total gasoline utilization. So we'll give the commenter 1-for-2: CAFE standards have not been particularly effective, but the gas tax required to mimic their effects on utilization would have to be pretty steep.