Thursday, March 22, 2012

About the strange response of consumers to gas tax increases

When we think of price elasticity of demand, there is no reason to think that it would be different depending on the reason the price has changed. The quantity of apples you demand will be lower by the same amount whether the price has increased due a tax increase, a lack of supply or a typo by the grocer. To a large extend it is so because you generally do not know why the price has increased. But even if you knew, would that make a difference?

Shanjun Li, Joshua Linn and Erich Muehlegger
find it does for gas prices in the US. A five cent increase in gasoline taxes reduces consumption by 1.3% in the short-run. The same increase in the after-tax price reduces it by 0.16%. How can the difference be so large? First, you have to wonder whether the tax change covaries with something that has an impact on consumption, and that was not taken into account. To be honest, I cannot think about a good reason. Second, could these estimates be very imprecisely estimated and thus not significantly different from each other? The coefficients seem rather tightly estimated. Third, tax increases are widely announced and anticipated, and consumers thus can prepare themselves by filling the tank just before the change takes effects. That is not the case with market fluctuations. Yet, some of the difference persists in the long run. Do tax changes have a signaling effect that prompts households to change habits? The paper seems to allude to that in the sense that there is a change in fuel efficiency after tax changes that is largely absent after other price changes.


Rebecca Scot said...

I too have found consumers to react more strongly to tax-driven than to crude-oil-driven changes in the gasoline price, and I've tried to explain this difference in the context of rational (that is, forward-looking) habits: consumers have a greater incentive to adapt to price changes they anticipate to be permanent than to price fluctuations driven by the ups and downs of the crude oil price. See e.g.

Sylvain Weber said...

Studying gasoline consumption in Switzerland, we also find that a tax increase has a larger impact than a "natural" price increase of a similar magnitude.
The complete working paper can be downloaded at

Anonymous said...

Rebecca and Sylvain, sorry for missing your papers. It is great to know similar evidence has been found using different data sets. Will include them in the next revision. Thanks. Shanjun Li