Mankiw's other arguments are expressed in high-sounding terms as a case of "Pigovian taxes," named for economist Alfred Pigou. A Pigovian tax assumes politicians and their economists have the knowledge and motivation to discern when people are buying too much of something, because they fail to take account of the "social costs" their purchase imposes on others. Assuming such wisdom exists, the government can supposedly use selective sales taxes as a tool of behavioral modification. Yet the CBO noted that a 2002 National Research Council estimate of the Pigovian "external costs" of consuming gasoline amounted to just 26 cents a gallon -- less than the average federal-state tax of 41 cents.

The real motive behind high taxes on liquor, tobacco and gasoline is more plausibly related to "Ramsey taxes," named for philosopher Frank Ramsey. Pigou might have argued that we should tax wine to discourage excess drinking. Ramsey would argue that we should tax wine precisely because the demand for wine is relatively unresponsive (inelastic) to a higher price. Because a tax on wine, tobacco or gasoline does not have a strong effect on consumption, such taxes are "efficient" in the sense that they yield the most revenue with the least distortion of the way resources are used (unless they result in black markets).

Governments like to claim they raise these "sin taxes" to discourage drinking, smoking and driving -- as though driving to work is a sin. In reality, governments like these taxes because their effect on consumption is weak. And the real reason the federal government has not pushed this tax much higher is that doing so would pre-empt and reduce an important source of state revenue.

Mankiw wants to raise the gas tax twice as much as the CBO estimated, which might cut gasoline consumption 20 percent from where it would otherwise be a decade from now. But that would be only a 5 percent cut from current consumption. That couldn't make a noticeable difference in global warming because U.S. passenger vehicles account for only 20 percent of carbon dioxide emissions. Even a 20 percent cut in 20 percent is only 4 percent, and the United States is only part of the globe.

It would not make a huge difference in domestic oil consumption either, because passenger vehicles account for only 40 percent of U.S. oil demand. A 20 percent cut in 40 percent is only 8 percent. From such a trivial change, Mankiw imagines "the price of oil would fall in world markets." But that undermines his environmental arguments. If the world price of oil fell, then China and India would use more oil and global emissions would not decline.

Claiming to remedy social costs with higher taxes is a game with no clear rules. Using Pigovian logic, I would argue that wine purchases should be tax-free and tax-deductible, because wine is so beneficial to public health and sociability that private demand fails fully to reflect it social value. Nobody could prove me wrong because all such analyses of social benefits and costs are incurably opinionated. Yet I will never be asked to testify on the Pigovian merits of a tax break for wine because federal and state governments crave the money a wine tax brings in.

This, too, is all about the money: If you have some, the government wants it. But they aren't doing such a great job with what they have are they?