Search This Blog

Monday, September 24, 2007

Volatile Gas




I try to answer Mike Moffatt's post regarding the volatility of gas consumption (and hence revenue from taxes) compared to other goods. It appears gas is quite volatile esp. when you look at the mid/late 70s shock period.

I use BEA NIPA table data from their set called "Table 2.3.5. Personal Consumption Expenditures by Major Type of Product" to compute % change. The data is nominal.

I tend to agree with Mike though that this may not be that important in the long run since gas tax revenues are so relatively small. But could it be an issue year-to-year in certain circumstances? I think it could. Thoughts?

There are only 2 things that can cause this relative volatility: volatile changes in quantity consumed, or volatile changes in prices. Since the former is unlikely given the inelasticity of demand, the latter likely makes up most of the volatility. This makes sense theoretically since prices are dependent on behavior of cartels, and supply-chains that are tied to highly volatile conditions (like weather etc)



UPDATE 9/25: Mike correctly points out to me that my above graph doesn't really directly address the question of volatility of revenue of a gas tax since the above is likely mostly price volatility. Since a gas tax, unlike a typical sales tax, is a tax on a set quantity (40C per gallon) as opposed to a % of sales, a more appropriate graph would show the volatility of just quantities, not price changes.



In an attempt to do this, I use BEA's quantity index comparing the same items. Notice gas quantity is not really much more or less volatile than other goods (actually household goods is quite a bit more volatile swinging up and down from slightly negative growth from the previous year to as much as 15% growth). Though the point made above still holds; it is subject to huge spikes (supply shocks etc):

2 comments:

Unknown said...

i don't have anything intelligent to say, but will watch for more posts and comments...

Daniel Hall said...

The original post from Peter Dorman at EconoSpeak actually concerned carbon taxes rather than gasoline taxes. Thus the appropriate measure of volatility is total U.S. CO2 (or GHG) emissions, which have less annual variability than gasoline purchases.

Regardless, at any reasonable (read: politically feasible) tax level the revenues will be a small enough percentage of total U.S. revenues that volatility shouldn't matter much. I provide illustrative calculations in this post.