Higher gasoline taxes encourage work, and when this effect is taken into account, the optimal gasoline tax is substantially higher than previous research has suggested.
Gasoline taxes are a hotly debated topic in both environmental and economic policy settings. Previous studies have examined the effects of various gasoline taxes on the supply and demand for gasoline. But in Empirical Estimates for Environmental Policy Making in a Second-Best Setting (NBER Working Paper No. 10330), authors Sarah West and Roberton Williams examine the effect of gasoline taxes on work decisions. Do people work more or less when gasoline prices go up? The authors find that higher gasoline taxes encourage work, and when this effect is taken into account, the optimal gasoline tax is substantially higher than previous research has suggested.
If gasoline taxes are set based only on their effects on gasoline use, then the best governmental policy would be to set the gas tax equal to marginal damage: the value of all of the negative externalities that result from using a gallon of gasoline, including pollution, accidents, noise, and traffic congestion. Since these costs are imposed on others, people don't have enough incentive to conserve gas. Taxing gasoline forces drivers to take that cost into account when making driving decisions. If the gas tax equals marginal damage, then the cost of gasoline to the driver is the same as the cost to society, thus providing the proper incentives.
"First-best" government policy would provide the proper incentives for all decisions. But this isn't the case. People make work decisions based on the perceived return to them, working so long as the value of making more money is greater than the value of having more free time. Because part of what they earn goes to pay income taxes, people don't have enough of an incentive to work. Since the government needs a certain amount of tax revenue, though, the best it can do is the "second-best" optimum: the ideal policy, given the fact that people don't have sufficient incentives to work. The second-best optimal gas tax thus depends both on how gas prices affect driving decisions and on how they affect work decisions.
Taxes on specific consumer goods often discourage work by even more than the income tax does, thus exacerbating the disincentive to work. If the same were true for gasoline, then this would reduce the optimal gas tax. But the effect could be just the opposite. Whether a specific tax increases or decreases work effort depends in part on how it affects the costs of various leisure activities. Since a gasoline tax increase makes leisure driving more expensive, it may reduce the time spent in such activities, thus encouraging people to work more. The authors' empirical work suggests that this is indeed the case: raising gasoline taxes and lowering income taxes would cause people to work more, not less.
The authors find that a 10 percent increase in gasoline prices would decrease gas consumption by 4.3 percent, or roughly 37 gallons per household per year. That same increase in gas prices would also increase hours worked by 0.07 percent, approximately 2 hours per household per year. Raising the gasoline tax thus has the triple benefit of lowering fuel consumption, decreasing pollution, and providing an incentive for people to work at a more socially optimal level.
This change in work is tiny compared to the total number of hours worked, but still substantially increases the optimal gas tax, because the labor market is far larger than the gasoline market. Ignoring effects on work decisions, the optimal gas tax is equal to marginal damage, which other researchers have estimated at 88 cents per gallon (in 2003 dollars). Using this estimate, West and Williams find that the optimal gas tax is $1.19 per gallon (also in 2003 dollars), with the difference arising because higher gas prices encourage work.
Controlling for various other effects, the authors find that the work effect is not attributable solely to consumers paying more for gasoline. Even if other taxes are lowered to compensate for the increased price of gas, people still work more. This effect may be because decisions regarding non-work-related driving are more flexible than decisions about work-related driving, such as commuting. In other words, a person's choice to work longer or shorter hours does not affect how much gas he uses to commute. However, that person may choose to make fewer non-work automobile trips, thus leaving more time available for work.
-- Les Picker