Environmental tax policy using a two-part instrument.

Fullerton, Don

One important goal of tax policy is economic efficiency. In some cases, this requires raising revenue and avoiding changes in relative prices that may distort taxpayer behavior and create "excess burden." In other cases, however, economic efficiency might require changes in relative prices: for example, taxing the "negative externalities" from alcohol, tobacco, and disposal of household or industrial waste. (Negative externalities include injuries, second hand smoke, and aesthetic costs.)

A second goal of tax policy is administrative efficiency. This is often best achieved by taxes on market transactions, for which the tax base can be measured and verified most easily. Taxes can apply to wages paid by an employer, interest paid by a bank, dividends reported by a broker, and the sale of cigarettes and alcohol as reported by retail establishments.(1)

But what about disposal of household and industrial waste? To achieve economic efficiency, these activities should be taxed, but they are often not market transactions that can be verified by a third party. In such cases, a "two-part instrument" might resolve the conflict.(2) Instead of directly taxing waste, a two-part instrument would raise the relative price of waste indirectly through both a tax and a subsidy on other activities that are market transactions. This policy combination can change relative prices in the same way as a tax on waste, but each tax or subsidy can be verified by invoices. Thus, the two-part instrument might better achieve both economic and administrative efficiency.

In the next sections, I clarify the theory behind this idea and provide a few examples. The following sections consider interactions with other taxes and the issue of scarcity rents.

Any Tax Can be Set to Zero

Taxpayers long have known that government can tax them both when they earn and again when they spend; most economists recognize that one such tax is redundant. Generally speaking, a tax that takes half of your gross paycheck is equal to a tax that doubles the price of everything you buy. As a consequence, for any system of tax rates on different commodities, any one tax can be set to zero. Revenue can be raised by a tax on all forms of income. Then all the desired relative prices of the different commodities can be achieved by a set of taxes and subsidies on goods other than the untaxed good.(3) One simple example is a political promise not to tax cigarettes, which can be circumvented by a tax on all income and a subsidy to all goods except cigarettes.

The best actual example of a two-part instrument is a deposit-refund system. A tax is first paid at the store on some item(s), and then returned if and when the item (or its container) is recycled. The result is a tax that remains on the good when it is not recycled. But this idea can be applied much more generally. Even the U.S. income tax operates on such a principle, using a withholding tax collected by employers that may exceed the tax due. If so, a refund is paid if and when the taxpayer files properly.

Other Examples

Suppose that a government wants to tax all household waste disposal in order to reduce landfill costs and negative externalities (like truck noise, odor, and groundwater contamination). A tax per bag of garbage might be difficult to implement, administer, and enforce.(4) It also can induce illegal dumping. Under some conditions, however, the jurisdiction can: tax everything bought at the store, through a general sales tax; provide a partial subsidy to all regular garbage, through free curbside collection; and provide a higher subsidy to all recycling.(5) This combination leaves a partial tax on garbage, but it leaves the highest rate of tax on anything not put into regular garbage or recycling - that is, anything dumped illegally.(6)

Second, suppose policymakers want to tax some polluting emissions from a factory, and cannot measure those emissions through such devices as the "continuous emissions monitoring" (CEM) equipment used on large power plants. Ease of measurement and enforcement may vary for toxic or nontoxic emissions that are gaseous, liquid, or solid. (The emissions can be viewed as a necessary input to production with its own downward-sloping marginal product schedule, since additional emissions are successively less crucial to production.) The desired substitution in production from this "dirty" input to other "clean" inputs then can be achieved by a subsidy to all clean inputs. This subsidy tends to reduce the equilibrium price of output, which might encourage more purchases of this good. That effect can be avoided by a simple excise tax on the output. The result is a two-part instrument. The tax on output is equivalent to a tax on all inputs at the same rate. This tax is refunded on clean inputs, leaving an implicit tax on the dirty input. Each tax and subsidy applies only to market transactions which have invoices to verify the tax base.

The idea of a two-part instrument is perhaps most important in a case where the emissions are difficult to measure and the tax is difficult to enforce. Therefore, a third example might be emissions from the millions of motor vehicles in this country that are owned by many individuals who might tamper with on-board devices or avoid remote sensing stations designed to measure the tax that each person owes. Even without tampering, measurement might be expensive. Preliminary findings indicate that all the desired incentive effects of an emissions tax can be achieved by the combination of a tax on each fuel and a subsidy at the appropriate rate on each abatement technology including methanol, compressed natural gas, or other alternative fuel vehicles.(7) Thus the measurement of emissions is unnecessary.

A fourth example involves the environment through common-property natural resources, such as water, that tend to be overused if not priced properly. Groundwater is hard to price explicitly, since a landowner can take as much as desired for free. Yet efficiency may require a price that covers the "scarcity rent" or any negative externality from depletion of the aquifer. The scarcity rent is the amount that others would be willing to pay for the water if they had the opportunity. An example of a negative externality is the reduction in springwater necessary for maintaining certain endangered species. If a farmer uses groundwater for irrigation along with other inputs in production, then a two-part instrument could tax the agricultural output, and subsidize all of the inputs other than water.(8)

When Government Needs Revenue

With no revenue constraint, or the availability of lump sum taxes, the "first-best" tax on a polluting input is equal to "marginal environmental damages." The firm is then faced with the full social cost of using that input. The two-part instrument taxes output and subsidizes other inputs, all at rates based on the same concept, marginal environmental damages.9

Now, suppose that revenue must be raised using distorting taxes that affect labor supply and saving decisions. Perhaps the two-part instrument could help to raise revenue by imposing a higher tax and paying a lower subsidy. This suggestion is related to the "double dividend hypothesis," that an environmental tax can help both to fix an environmental problem and to raise revenue for use in reducing other distorting taxes.(10) Some have inferred that this second-best pollution tax rate should exceed marginal environmental damages. Recent research finds the reverse, though: the pollution tax raises output prices and reduces the real net wage, so it distorts labor decisions as well as the consumption mix.(11)

What about the two-part instrument? Whatever the desired rate of tax on the dirty input, the same change in relative prices can be achieved by a tax on output that is returned on clean inputs. Thus the subsidy must match the tax. If second-best considerations reduce the desired-but-unenforceable tax on emissions, then they reduce both parts of the two-part instrument. Revenue considerations do not suggest raising the tax and reducing the subsidy.(12)

Scarcity Rents

The point about revenue and the double-dividend hypothesis relates to my recent research with Gib Metcalf.(13) The double dividend literature has suggested that a revenue-raising instrument, for example a pollution tax, can provide higher welfare than a non-revenue-raising instrument, such as quotas, permits, or command-and-control (CAC) restrictions on emissions. All of these policies can provide the same environmental improvement, and all raise the cost of production, but only the tax generates revenue that can be used to reduce distorting taxes on labor.

For example, the Clean Air Act Amendments of 1990 restrict emissions by using permits that are handed out to firms. The requirement to use these valuable permits raises the cost of production, and thus lowers the real net wage, but the scarcity rent goes to permit recipients. We show that the double-dividend debate should focus not on whether an environmental policy raises revenue, but on whether it creates scarcity rents that are left in private hands. Only if government sells all permits (or has a 100 percent profits tax) can it capture the scarcity rent and use that revenue to offset the reduction in the real net wage by cutting the labor tax.

Regulators can impose different kinds of CAC restrictions. If they simply restrict emissions, then they create scarcity rents that must be covered by a higher price of output. In contrast, regulators can require a reduction in emissions per unit of output. If it applies equally to all firms, and does not limit entry, then this policy does not create scarcity rents. For small changes, this "technology restriction" has no first-order effect on the cost of production. It does have first-order effects on the environment, however. So this nonrevenue-raising policy unambiguously improves welfare, just like the revenue-raising emissions tax.

To clarify further that raising revenue is not the crucial distinction, one can compare an environmental tax that raises revenue to an environmental subsidy that costs revenue. One might think that an environmental subsidy would provide less welfare, since it must be financed by raising other distorting taxes. Yet the environmental subsidy has exactly the same effects as the environmental tax! The tax on a dirty input raises the cost of production. This is turn raises the price level, and would reduce the real net wage except for the fact that the revenue can be used to cut the labor tax. Symmetrically, the subsidy to a clean input reduces the cost of production. This reduces the price level, and would raise the real net wage except for the fact that the subsidy needs to be financed by raising the labor tax. Either way the real net wage is unaffected, so labor supply distortions are unaffected.(14)

Fully specified, both of these policies are revenue-neutral. The two-part instrument (environmental subsidy financed by a higher labor tax) is equivalent to the emissions tax (with revenue used to lower the labor tax). The crucial distinction is not whether the environmental policy raises revenue, but whether it restricts the quantity of emissions in a way that creates a scarcity rent that is left in private hands, rather than captured by government and used to offset the effect of higher output prices.

1 For a discussion of tax systems that optimize administrative costs as well as tax rates, see J. Slemrod, "Optimal Taxation and Optimal Tax Systems," Journal of Economic Perspectives 4, 1 (Winter 1990), pp. 157- 78.

2 The terminology of a "two-part instrument" first appears in D. Fullerton, "Environmental Levies and Distortionary Taxation: Comment,"American Economic Review 87, 1 (March 1997), pp. 245-51.

3 These arguments are formalized in D. Fullerton and A. Wolverton, "The Case for a Two-Part Instrument: Presumptive Tax and Environmental Subsidy," NBER Working Paper No. 5993, April 1997.

4 A particular garbage pricing program is studied for a sample of households in D. Fullerton and T. Kinnaman, "Household Responses to Pricing Garbage by the Bag," American Economic Review 86, 4 (September 1996), pp. 971-84.

5 D. Fullerton and T. Kinnaman, "Garbage, Recycling, and Illicit Burning or Dumping," NBER Reprint No. 2024, January 1996, and Journal of Environmental Economics and Management 29, 1 (July 1995), pp. 78-91.

6 The equivalence is broken if the good is taxed upon purchase in one jurisdiction and recycled in a different jurisdiction without a matching policy. Thus the policy might need to be implemented by a state, rather than local government. Also, the subsidy need not be paid per item. A payment per ton can provide the recycling-plant with incentives to obtain more materials from households. Finally, a related set of taxes and subsidies can encourage firms to design their products to use the right amount of packaging and degree of recyclability. See D. Fullerton and W. Wu, "Policies for Green Design," NBER Working Paper No. 5594, May 1996.

7 D. Fullerton and S. West, "Two-Part Instruments for the Control of Vehicle Emissions: Derivation and Numerical Implementation," work in progress.

8 D. Fullerton and M. Mathis, "Achieving Optimal Use of Water," work in progress.

9 In a first-best model, without revenue constraints, explicit closed-form solutions for the necessary tax and subsidy rates are provided in D. Fullerton and A. Wolverton, "The Case for a Two-Part Instrument..."

10 A review of this literature is provided by L. Goulder, "Environmental Taxation and the Double Dividend: A Reader's Guide," International Tax and Public Finance 2, 2 (August 1995), pp. 157-83.

11 A. L. Bovenberg and R. A. de Mooij, "Environmental Levies and Distortionary Taxation," American Economic Review 84, 4 (September 1994), pp. 1085-9.

12 D. Fullerton, "Environmental Levies and Distortionary Taxation: Comment," American Economic Review 87, 1 (March 1997), pp. 245-51.

13 D. Fullerton and G. Metcalf "Environmental Controls, Scarcity Rents, and PreExisting Distortions," NBER Working Paper No. 6091, July 199 7.

14 Ibid.

Don Fullerton is Addison Baker Duncan Professor of Economics at the University of Texas at Austin and a research associate in the NBER's Program in Public Economics.