July 21, 2008
Gasoline Taxes
Should US Taxes on Gasoline be Higher? Becker
Gasoline prices have increased rapidly during the past several years, pushed up mainly by the sharply rising price of oil. A gallon of gasoline in the US rose from $1.50 in 2002 to $2 in 2004 to $2.50 in 2006 to over $4 at present. Gasoline prices almost trebled during these 6 years compared to very little change in nominal gas prices during the prior fifteen years. The US federal tax on gasoline has remained at 18.4 cents per gallon during this period of rapid growth in gasoline prices, while state excise taxes add another 21.5 cents per gallon. In addition, many local governments levy additional sales and other taxes on gasoline. Gasoline taxes have not risen much as the price of gasoline exploded upward.
The price of gasoline is much lower than in other rich countries mainly because American taxes are far smaller. For example, gasoline taxes in Germany and the United Kingdom amount to about $3 per gallon. Some economists and environmentalists have called for large increases in federal, state, and local taxes to make them more comparable to gasoline taxes in other countries. Others want these taxes to rise by enough so that at least they would have kept pace with the sharply rising pre-tax fuel prices. At the same time two presidential candidates, Hillary Clinton and John McCain, proposed a temporary repeal during this summer of the federal tax in order to give consumers a little relief from the higher gas prices. We discuss the optimal tax on gasoline, and how the sharp increase in gas prices affected its magnitude.
Taxes on gasoline are a way to induce consumers to incorporate the "external" damages to others into their decision of how much to drive and where to drive. These externalities include the effects of driving on local and global pollution, such as the contribution to global warming from the carbon emitted into the atmosphere by burnt gasoline. One other important externality is the contribution of additional driving to road congestion that slows the driving speeds of everyone and increases the time it takes to go a given distance. Others include automobile accidents that injure drivers and pedestrians, and the effect of using additional gasoline on the degree of dependence on imported oil from the Middle East and other not very stable parts of the world.
A careful 2007 study by authors from Resources for the Future evaluates the magnitudes of all these externalities from driving in the US (see Harrington, Parry, and Walls, "Automobile Externalities and Policies", Journal of Economic Literature, 2007, pp 374-400). They estimate the total external costs of driving at 228 cents per gallon of gas used, or at 10.9 cents per mile driven, with the typical car owned by American drivers. Their breakdown of this total among different sources is interesting and a little surprising. They attribute only 6 cents of the total external cost to the effects of gasoline consumption on global warming through the emission of carbon into the atmosphere from the burning of gasoline, and 12 cents from the increased dependency on imported oil. Perhaps their estimate of only 6 cents per gallon is a large underestimate of the harmful effects of gasoline use on global warming. Yet even if we treble their estimate, that only raises total costs of gasoline use due to the effects on global warming by 12 cents per gallon. That still leaves the vast majority of the external costs of driving to other factors.
They figure that local pollution effects amount to 42 cents per gallon, which makes these costs much more important than even the trebled cost of global warming. According to their estimates, still more important costs are those due to congestion and accidents, since these are 105 cents and 63 cents per gallon, respectively. Their figure for the cost of traffic accidents is likely too high –as the authors' recognize- because it includes the cost in damages to property and person of single vehicle accidents, as when a car hits a tree. Presumably, single vehicle accidents are not true externalities because drivers and their passengers would consider their possibility and internalize them into their driving decisions. Moreover, the large effect of drunk driving on the likelihood of accidents should be treated separately from a gasoline tax by directly punishing drunk drivers rather than punishing also sober drivers who are far less likely to get into accidents.
On the surface, these calculations suggest that American taxes on gasoline, totaling across all levels of government to about 45 cents per gallon, are much too low. However, the federal tax of 18.4 cents per gallon is almost exactly equal to their figure of 18 cents per gallon as the external costs of global warming and oil dependency. To be sure, a trebled estimate for global warming would bring theirs up to 30 cents per gallon. However, the federal government also taxes driving through its mandated fuel efficiency standards for cars, although this is an inefficient way to tax driving since it taxes the type of car rather than driving. Still, the overall level of federal taxes does not fall much short, if at all, from the adjusted estimate of 30 cents per gallon of damages due to the effects of gasoline use on global warming and oil dependency.
Any shortfall in taxes would be at the state and local levels in combating externalities due to local pollution effects, and to auto accidents and congestion on mainly local roads. Here too, however, the discrepancy between actual and optimal gasoline taxes is far smaller than it may seem, and not only because single vehicle accidents are included in their estimate of the cost of car accidents, and accidents due to drunk driving should be discouraged through punishments to drunk drivers. One important reason is that congestion should be reduced not by general gasoline taxes, but by special congestion taxes- as used in London and a few other cities- that vary in amount with degree of congestion (see our discussion of congestion taxes on February 12, 2006). Congestion taxes are a far more efficient way to reduce congestion than are general taxes on gasoline that apply also when congestion is slight.
In addition and often overlooked, the sharp rise in pre-tax gasoline prices has partly accomplished the local pollution and auto accident goals that would be achieved by higher gas taxes. For higher prices have cut driving, just as taxes would, and will cut driving further in the future as consumers continue to adjust the amount and time of their driving to gasoline that costs more than $4 a gallon. Reduced driving will lower pollution and auto accidents by reducing the number of cars on the road during any time period, especially during heavily traveled times when pollution and accidents are more common.
The effects of high gas prices in reducing congestion, local pollution, and accident externalities could be substantial. These authors estimate the size of local driving externalities, aside from congestion costs, at 105 cents per gallon. Even after the sharp run up in gas prices, this may still exceed the 28 cents per gallon of actual state and local taxes, but the gap probably is small. It surely is a lot smaller than it was before gas prices exploded on the back of the climb in the cost of oil. In effect, by reducing driving, higher gasoline prices have already done much of the work in reducing externalities that bigger gas taxes would have done when prices were lower.
Should Gasoline Taxes Be Raised or Lowered? Posner's Comment
The economic study that Becker discusses treats gasoline taxes as a form of regulatory taxation, that is, taxation aimed at altering behavior rather than at collecting revenue. A gasoline tax is an excise tax, and excise taxes are a common method of raising revenue to pay for government. The best excise tax from a revenue-raising standpoint is one that causes minimum substitution against the taxed good or service, since (in the absence of externalities) such substitution distorts the efficient allocation of resources and reduces the revenues that the tax was supposed to generate. A regulatory tax aims at substitution because of the externalities caused by the taxed good or service, but complete substitution is rarely achieved (and indeed would usually be inefficient), and so a regulatory tax raises revenue as well as altering behavior. My guess is that the very high gasoline taxes in Europe, which are primarily responsible for the fact that the price of gasoline in Europe is on average almost twice the U.S. price, are intended and effective as revenue-raising devices, since those taxes antedate the current concerns with global warming, dependence on oil supplies from hostile or unstable nations, pollution, and acute traffic congestion. Whether from a revenue standpoint a stiff gasoline tax is an efficient tax, I do not know. But my guess is that it is. Since distances are shorter in Europe and public transportation far more extensive, Europeans can substitute against gasoline more easily than Americans can; nevertheless the very high price of gasoline in Europe, though for years it has been higher than U.S. prices are now, has not prevented demand for gasoline from growing, though in part this is due to extensive European construction of new non-toll highways and roads. An excise tax on a single commodity will not generate a great deal of revenue, because of its narrow base, but can be justified as part of a comprehensive system of excise taxes.
It is likely, judging from U.S. consumers' reaction to the recent increase in the price of gasoline, that a steep hike in the gasoline tax (I am treating the state and federal gasoline taxes as a single tax) would cause a further reduction in demand. Consumers would drive less (some of them by moving closer to work--and telecommuting would increase) and would switch at a higher rate to vehicles with better gas mileage. At some point, however, the fall in demand might cause the price of oil to decline. The reason is that the supply curve for oil is upward-sloping, meaning that a reduction in demand and hence in supply will reduce price. I say "might" cause the price of oil to decline because world demand for oil might continue to rise even if U.S. demand fell, in which event the world price would not decline.
I wonder, too, whether the recent decline in U.S. gasoline consumption doesn't represent to some degree an irrational panic reaction. To take a huge loss on the sale of your SUV in a market that is depressed because so many other people are doing the same thing at the same time is unlikely to be justified by the gains from the improved gas mileage of the car you buy with the modest proceeds of the sale. Likewise, driving a substantial distance to save a few cents a gallon on the gas you buy is unlikely to be worthwhile. A recent article suggests that people fixate on the price of gasoline because unlike most regularly purchased items, such as food, gasoline is purchased separately from other items so that its price is not buried in a bill for multiple items.
The economic study that Becker cites finds only modest externalities from gasoline consumption, and this argues for keeping our gasoline taxes low if we think of such taxes as primarily regulatory rather than revenue-raising. But except for its effect in reducing highway accidents by reducing the amount of driving, a gasoline tax is not an efficient regulatory tax. Congestion should be taxed directly, since people who travel on uncongested roads do not contribute to congestion. And the carbon emissions from the burning of fossil fuels (including gasoline) should be taxed, not gasoline, because a tax on gasoline does not create an incentive to produce lower emissions per gallon. Furthermore, taxing gasoline but not aviation fuel will increase the demand for air transportation, a potent source of both congestion and carbon emissions. Even the conventional pollutants produced by the internal-combustion engine do not argue strongly for a regulatory gasoline tax, because these pollutants, in the form of smog for example, reduce global warming by blocking sunlight. And from the standpoint of reducing our dangerous dependence on foreign oil, the proper tax is one on oil, rather than one on just one oil product.
Hence the case for higher gasoline taxes should rest primarily on the efficiency of such taxes as revenue-raising devices. Even if as I suspect they are efficient revenue-raising taxes, the time to impose this is when gasoline prices fall, not now when consumers are screaming. Once people adjust to a price of $4.50 per gallon of gasoline, any fall in that price can be offset by an increase in gasoline taxes. A complication is that a tax on carbon emissions will, depending on how stiff it is, retard any natural, market-drive reduction in the price of gasoline. A further complication is that the calculation of an optimal carbon-emissions tax is impossible because the costs of global warming and the benefits (in reducing those costs) from a tax on carbon emissions cannot at present be estimated with even minimal confidence.