Puzzled by the Proposal for a Tax Holiday

I have been puzzled by the proposal for a tax holiday for gasoline purchases running from Memorial to Labor day (see (0), (1), (2)), with the objective of spurring the economy. First, the Federal tax is quite low, either in real or in relative terms. Second, the benefits that would accrue to consumers are probably pretty small, under reasonable assumptions.

Going to the first point, consider the total Federal tax on a gallon of gasoline.

As one can readily verify, in inflation adjusted terms, the tax has been eroded over time to levels not seen since the early 1990's. This is true regardless whether one deflates by the CPI-all or the CPI-ex. energy and food.

In relative terms, the total Federal tax has been shrinking as a share of gasoline prices, as gasoline prices have headed north (March = $3.293, all grades, inclusive of taxes). As of March 2008, the Federal tax accounted for 0.056% of that price.
Figure 2: Total Federal gasoline tax as a share of total price of gasoline. Sources: ARTBA, Energy Information Administration/DoE (a), (b), and author's calculations.

So this is the measure to jump start the economy? I think this measure would give relief to somebody. But I also think it's a pretty blunt instrument by which to provide assistance to the driving public, or consumers, for that matter.

Now, I'm not a microeconomist by training. Nor do I play one on TV. But it seems to me that if the supply of gasoline is price inelastic, and the demand is similarly price inelastic, then the incidence of the current Federal gas tax must be about evenly balanced. A tax holiday is then a holiday to both consumers and producers.

Do we know what these price elasticities are? Well, we know from this 2003 CBO report by David Austin and Terry Dinan that in the long run a plausible supply elasticity is 2, and plausible demand elasticities are in the -0.3 to -0.9 range. So the incidence of the tax on consumers is bigger than that on producers, for that time horizon.

But these elasticities are relevant for the long run (in this case, used to evaluate the relative merits of CAFE standards versus a gasoline tax). The proposed gas tax holiday was for a short duration of months. In this case, the short run price elasticity of supply is near zero, and the demand elasticity is plausibly near zero as well.

Assume both supply and demand are equally price inelastic, and this means the incidence of the Federal tax is about 50-50. Eliminating the gasoline tax for a short duration gives a windfall to both consumers and producers, of about equal proportion. (By the way, this conclusion is not true of state gasoline taxes; see Chouinard and Perloff (2004)). Now, giving a windfall to refiners and providers of feedstock for gasoline production might be a worthy goal, but I don't believe that was the stated goal. If those corporations get a windfall then either it gets stored away to be spent on investment in a new refinery or addition to an old refinery sometime in the future, or it leaks out to overseas oil producers.

Oh, and by the way, to the extent the lower price spurs gasoline consumption, this should increase the petroleum and petroleum products component of U.S. imports, and thence putting further upward pressure on the price of oil...