Raising Gas Taxes: The “Pigou Club” vs. the “Coase Club”

October 24, 2006

There’s been a flurry of discussion lately about the merits of raising the federal gas tax. Unlike the current gas tax which is mostly designed to raise revenue for transportation, recent proposals aim to impose what economists call a “Pigouvian” tax on gas, aimed at curbing gas consumption and correcting for “negative externalities” like pollution and threats from hostile foreign countries.

Harvard professor N. Gregory Mankiw has been at the center of calls for a Pigouvian gas tax, promoting what he calls the “Pigou Club”, and advocating higher gas taxes in a recent op-ed on the editorial page of the Wall Street Journal:

With the midterm election around the corner, here’s a wacky idea you won’t often hear from our elected leaders: We should raise the tax on gasoline. Not quickly, but substantially. I would like to see Congress increase the gas tax by $1 per gallon, phased in gradually by 10 cents per year over the next decade. Campaign consultants aren’t fond of this kind of proposal, but policy wonks keep pushing for it. Here’s why:… (Full piece here.)

Mankiw goes on to list seven distinct arguments in favor of higher gas taxes, only three of which can be plausibly classified as “Pigouvian”—gasoline pollution, traffic congestion, and foreign threats. However, putting aside the other non-economic arguments for higher gas taxes, is Mankiw’s recommendation of Pigouvian gas taxes on solid ground?

In theory, using Pigouvian taxes to correct for what economists call “market failures” is simple. But in practice, it’s anything but. One important problem often ignored by advocates of Pigouvian taxes is what might be called the “measurement problem.” That is, if gas taxes should be raised purely to offset the social costs of gas use, how high are those social costs?

Surprisingly, the question is rarely addressed in discussions of Pigouvian taxes. However, it’s central to their application to tax policy. While it’s easy to identify negative externalities in theory, Pigouvian taxation goes beyond that. It demands that those costs be empirically measured, not just identified, and that the tax rate be set equal to the per-unit external cost of gas that “spills over” onto others in society.

In this way, Pigouvian taxes place extremely high information burdens on policymakers. Clearly, the practical difficulty of compiling data and estimating social costs is not trivial. For one, current estimates in the literature vary widely.

For example, the figure below illustrates the wide range of external cost estimates for gas, oil and other methods of power generation. These figures are compiled from various studies throughout the 1980s and 1990s, and estimates range from very large external costs, to trivially small effects. (For a detailed discussion, see http://www.handels.gu.se/econ/seminar/Article1.pdf).

Estimates of “External Costs” Vary Widely for Various Fuels

Source: Thomas Sundqvist and Patrik Soderholm, “Valuing the Environmental Impacts of Electricity Generation: A Critical Survey,” Journal of Energy Literature 8, no. 2, December 2002, p. 19.

This lack of a basic scientific consensus on the social costs of gas is a serious problem faced by advocates of Pigouvian gas taxes, which has largely been ignored in recent discussions.

One common reply to this is that forcing policymakers to estimate social costs of gas consumption imposes too high of a burden on advocates of Pigouvian taxes. For example, it’s argued that we don’t require exact specifications of the “proper” income or sales tax rate before those taxes are implemented. So why require it for Pigouvian gas taxes?

This response ignores a fundamental distinction between broad-based income or consumption taxes and Pigouvian gas taxes. The goal of income and consumption taxes is primarily to raise revenue to fund a predetermined budget of spending programs. In contrast, the goal of Pigouvian taxes is not to raise revenue, but to provide federal lawmakers with a mechanism to fine-tune markets toward a higher level of efficiency than the free market could achieve without their guidance. While ordinary taxes only burden lawmakers with the task of setting tax rates high enough to meet revenue needs, Pigouvian taxes require that lawmakers set precisely the right tax rate that maximizes the overall welfare of more than 300 million individuals in the U.S. economy. That’s no simple task—and it’s one that some observers of the messy world of tax policy in Washington might dismiss as impossible.

Pigouvian taxes present a much more ambitious policy objective than other taxes that simply aim to raise revenue for programs. And as with all extraordinary claims, extraordinary policy interventions require extraordinary evidence to justify them. To date, economists who’ve advocated for Pigouvian gas taxes without satisfactorily answering the question of “how high” simply haven’t done the homework that Pigouvian taxation demands. And that’s just bad policy.

In his Nobel Prize acceptance speech, Ronald H. Coase derided what he calls “blackboard economics”—the practice of assuming simple economic models can be easily implemented in practice, without regard to practical considerations. It’s no surprise Coase spent much of his career debunking the naive application of the theory of Pigouvian taxation.

Recent advocates of the old idea of Pigouvian taxes would benefit from a careful reading of Coase’s criticisms. If they did, it’s likely they’d be forming the “Coase Club” rather than the “Pigou Club.”



Related Articles