I tend to agree with Greg Mankiw — more generally with Arthur Pigou — on his recommendation to increase certain taxes in order to cover the external costs that those products feed. However, there’s a flawed assumption in the analysis. Mr. Mankiw states,
[Pigovian taxes] raise revenue that the government can use to reduce other taxes, such as income taxes, which distort incentives and cause deadweight losses.
And then later,
The economists in the Treasury department are fully capable of designing a package of tax hikes and tax cuts that together internalize externalities and leave the overall distribution of the tax burden approximately unchanged.
I’m certain they are, too. Unfortunately, I’m also certain that the practical experience of government shows that taxes tend to go up, and that a new stream of revenue rarely leads to reductions in other taxes. Instead, it leads to the government spending, or redistributing if you prefer, the additional revenue.
So while I would prefer increased gasoline taxes and lowered income taxes, the chances of such are approximately equal to the likelihood of a snowball’s survival in Hell.
Or as Mr. Mankiw puts it,
I am eager for this tax only if the revenue is to be used to reduce other taxes, either contemporaneously or in the future.