Submitted by Econbrowser

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.

Figure 1: Gasoline tax, in cents per gallon (blue) and 1982-84 cents per gallon (red, green), as defined by CPI-all and CPI-ex energy/food. Sources: ARTBA, FRED II, and author’s calculations.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 5.6% 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…

Posted by Menzie Chinn at April 21, 2008 10:01 PM

digg this | reddit

Trackback Pings

TrackBack URL for this entry:
http://www.econbrowser.com/cgi-bin/mt-tb.cgi/846

Comments

This is an election year (as most of us are aware)
and before November we will be treated to all manner of suggestions from politicians and their familiars.

Most will be downright dreadful or simply stupid but all will be carefully crafted to sound like someone is not just standing there but doing something, when in most cases the opposite is the correct action.

Posted by: esb at April 21, 2008 10:14 PM

If supply and demand were both perfectly inelastic over some range of prices, we’d have overlapping vertical supply and demand curves over that range, and therefore multiple equilibria. Either suppliers would price at the top price of that overlapping vertical segment if they wanted to maximize short-term profit, or they’d price somewhere below that price due to objectives other than maximizing short-term profit (e.g., wanting to slow the development of substitutes or of conservation, or reducing the risk of governmental intervention such as a windfall profits tax or price controls). If it’s the former, then removal of the tax will simply lead to suppliers raising the price by the equivalent amount, resulting in no retail price change and increased profits for suppliers. If it’s the latter, suppliers would allow the retail price to decline by some or all of the amount of the (removed) tax.

Posted by: Brooks at April 21, 2008 11:38 PM

I believe that the federal tax bite is 100x larger than shown in figure 2. By my calc its 5.6% in March 08.

Posted by: Bulldog Bond at April 22, 2008 12:11 AM

It’s 18.4 cents per gallon. At $3.28 per gallon, you get the 5.6%. I think we have a decimal problem…

Posted by: AJ at April 22, 2008 04:25 AM

First, the Federal tax is quite low, either in real or in relative terms.

Yes and no.

Barackstar Obama keeps mentioning Exxon-Mobil’s profits, which were roughly $40B on $440B revenue, or a profit rate of 10%, or 35 cents on a $3.50 gallon of gas.

Now, I’m not entirely sure how much the federal gas tax is, but there’s a sticker on the gas pump where I fill up “informing” me that the gas tax I’m paying is 44.5 cents per gallon.

So the combined state, federal, and local gas tax I’m paying is roughly 50% more than Exxon-Mobil’s profit from extracting, refining, and dispensing the product.

Keep in mind that Exxon-Mobile is more or less the best run company on earth, not just the best run oil company. None of the other companies are anywhere near as profitable, as any BP stockholder will tell you.

With that said, I agree with Menzie that the “tax holiday” is bad policy, and gas taxes are very low by historical standards.

Posted by: Buzzcut at April 22, 2008 06:36 AM

Gas prices rise in the summer due to leisure driving, correct? Doesn’t that imply that the short-run summertime demand should be relatively elastic?

I have no reference to back this claim, just intuition…

Posted by: Nemo at April 22, 2008 06:46 AM

Bulldog Bond and AJ: Thanks for catching that. Corrected now.

Posted by: Menzie Chinn at April 22, 2008 06:52 AM

dear buzzcut and others
Here is the data on Exxon Mobil from the most recent 12-month reported results:
Revenue - $361.71 billion
EBITDA - Earnings before Interest, Tax, Depreciation and Amortization - was $73.1 billion.
Net Income - $40.61 billion
Federal excise tax per gallon - 18.4 cents
State taxes (as of July 1007) range from adding:
8 cents in Alaska
to 44.4 cents in California.
Oregon, Kansas and Idaho are 43.4 cents total
and
Georgia is 44.9 cents total. If buzzcut does not live in one of those 4 states the sticker he reports MIGHT be out of date or the gas tax has been changed since the report I read.