Karl,Thanks for the reply. I don't disagree that gas taxes "work" from an efficiency standpoint (disregarding distributional issues - though I would argue distributional issues could also effect long-run efficiency...see previous post on the subject).
I suppose my biggest problem with gas taxes is that elasticities DO matter once you look beyond the simple assumptions and if you understand that your goal is not just efficiency, but to better the environment. You say: "Suppose the elasticity of gas was zero. So that people just had to have it no matter what. Then there would be no change in consumption...no change in behavior...no deadweight...optimal."That makes sense to me but I would think we need to look beyond that...for the fact that you have to compare that to what actually would happen.
Gas demand is not perfectly inelastic (though close) and is not a perfect extreme. So, a gas tax WOULD in fact reduce consumption and reduce people's happiness - at minimal effect to the environment and at a cost to both consumers and producers.
The 'cost' would then be collected as government revenue.
Let's assume, oppositely for example, that gas demand were very elastic, then that same gas tax would have a much bigger effect on the environment, though it would raise less revenue for the govnernment.
So you see by those two examples, when gas elasticity is approaching perfect elasticity, for a GIVEN gas tax, the environment is better of. Since some policy maker has to 'fix' the tax hike at some level - that fixed level won't be as effective than if demand were more elastic - though it equally as efficient. So while from an efficiency stanpoint, a gas tax MAY make sense (IF you can determine costs etc - which I have doubts), it doesn't make sense if your policy goal is to better the environment - and that's regardless of distributional issues.
I should mention that the effect of the tax can vary as well not just with the elasticity of private demand, but the relative elasticity of private demand compared to social demand - they may not necessaily be parellell curves (they may have different elasticities )!
Once the government has to decide what to do with the revenue, that opens up a whole 'nother can of worms I would think....
I suppose the overall point is that textbook "efficiency" is based on a whole host of assumptions, not the least of which is a static model. Once dynamics are taken into account (environmental effects on efficiency in the future), that makes such models very limited in determining a policy goal... this is yet another example of econ trying to dominate decision making with its narrow view of things.