Kudos to Bryan Caplan of Econlog, Victor of
Dead Parrot Society, and the Governor of Georgia for proving wrong my conjecture that lowering gasoline taxes in the present environment would not result in cost savings for consumers.
In a post written September 1, I characterized the U.S. gasoline market in the aftermath of Katrina as one that required a significant short-run drop in U.S. gasoline consumption. I argued that the only mechanism that could achieve that, short of rationing, would be an increase in the net price that consumers pay. I reasoned that cutting the gasoline tax in such an environment would not change this reality, and would simply result in higher profit margins being received by the sellers of gasoline. Others reaching a similar conclusion included Alex Tabarrok of Marginal Revolution, Robert Lawson at Division of Labour, and Steve Verdon of Outside the Beltway.
Bryan Caplan of Econlog argued that I was missing the fact that if, for example, Georgia lowered taxes and other states did not, any increase in the net receipts for gas sellers in Georgia would cause gas to flow to Georgia at the expense of other states, which added supply could reduce the cost to drivers in Georgia. The Governor of Georgia did not wait around to see how our academic discussion might evolve, but simply suspended Georgia’s gasoline tax effective midnight September 2. Victor of
Dead Parrot Society noted that the retail price in Georgia immediately fell relative to the price in neighboring Florida and South Carolina. In addition to Bryan’s mechanism, Victor raised the possibility of de facto rationing and sales out of inventories as two other possible explanations for why, contrary to my prediction, Georgia’s tax cut seemed to bring immediate relief to Georgia’s motorists.
To these explanations let me offer a fourth– I was wrong in the underlying premise that we were facing a severe disruption in which gasoline consumption had to be sharply reduced. Since September 1, the retail price of gasoline fell not only in Georgia, but in every other state I checked, with the exception of Hawaii, where theory would predict that their cap on wholesale prices could easily produce an increase in retail prices. Victor noted that the price drop in Georgia was bigger than the price drop in neighboring Florida and South Carolina. Here are comparisons with two other states, from the same neat site that Victor used:
Eyeballing this graph, it appears that since September 1, the price of gasoline has fallen even faster in Indiana than it has in Georgia. Nationally, the price of unleaded gasoline on the NYMEX October futures contract (which includes neither state nor federal gasoline taxes) has fallen 38 cents a gallon since September 1, about twice the value of the drop in the after-tax price in the three states shown. Thus my conjecture that the final price paid by consumers would have to rise sharply above the values on September 1 has, thankfully, proven to be false. The price of gasoline has fallen everywhere, and the only question is by how much. My underlying premise that gasoline consumption had to be radically and immediately curtailed proved to be off the mark, and that, in my opinion, is why the prediction about the effect of cutting taxes also proved to be wrong.
Although we turned out not to be in the situation that I thought we were in last week, the mild rebuke from Bryan and Victor reminds me that elasticities are never really zero, and are easy to underestimate if you don’t think things through. Notwithstanding, if we have the misfortune to find ourselves again in a situation where supply disruptions do require an immediate and drastic reduction in gasoline consumption, I am sufficiently pig-headed to offer the same policy advice then that I offered on September 1: I do not believe that cutting gasoline taxes is a sensible policy response to that kind of problem. Even if Bryan’s argument proves to be correct, that in such a setting, a state that cut its taxes could divert gasoline from others, I would not want to recommend that as an appropriate policy option.