Green taxes: can we protect the environment and improve the tax system at the same time?

Oates, Wallace E.

I. Introduction

On a cursory inspection, public economics appears to provide a straightforward and compelling answer of YES to the question posed in the title to this paper. As economists have long argued, appropriately designed taxes can, in principle at least, efficiently restrain levels of polluting activities. And, in addition, such taxes will generate revenues so that we can reduce rates on other forms of taxation that distort the functioning of the economy. Green taxes, in short, promise a "double dividend": they can both reduce excessive levels of pollution and increase the efficiency of the overall tax system. At least, so several economists have argued recently [17; 20].

But is this really so? With the growing recognition of the enormous revenue potential of certain environmental taxes (especially carbon taxes to address the problem of global warming), public economists have returned to this question with a new round of research efforts directed both to understanding the properties of pollution taxes as revenue sources and to estimating their revenue potential and the magnitude of their effects on the economy. And the newly emerging literature is producing some quite astonishing and very troubling findings.(1)

In this paper, I want to review this work and explore its implications. When viewed in relation to some other research, it appears to suggest that most forms of regulatory activity (at least those that increase the costs of the regulated activities) may have hitherto unappreciated costs of a stunning magnitude. What many of us would have taken to be "second-order side effects" of environmental and other regulatory policies can be of a magnitude that compromises to a significant extent their welfare-enhancing properties. Having reviewed this body of work, I wish also to explore some issues in the political economy of green taxes. It is helpful to begin by briefly putting the issue in historical perspective.

II. Some Background

The early work on Pigouvian taxes or charges essentially ignored the revenue issue. The assumption, either explicit or implicit, was that the revenues would somehow be returned to the economy in lump-sum fashion so as not to cause any distortions in economic activity. The one qualification in this literature is that the revenues must not be used to compensate victims for the damages they suffer from polluting activities, for such compensation would compromise the incentives for efficient levels of defensive activities by victims. But, for our purposes here, the essential point is that the early literature (with the important exceptions of Agnar Sandmo [22; 23] and Yew-Kwang Ng [16]) gave little serious attention to Pigouvian taxes as a source of public revenues.(2)

A later strand in the literature took up this issue. David Terkla [27] actually estimated the potential efficiency gains from substituting revenues from a hypothetical set of nationwide taxes on particulate and sulfur oxide emissions for revenues from federal income (or alternatively corporate income) taxes. Using an optimal tax framework, Dwight Lee and Walter Misiolek [14] explored the determination of pollution tax rates to maximize the efficiency gains. Drawing on such work, observers advanced the seemingly compelling intuitive case for a "double dividend" from pollution taxes. Not only can such taxes raise social welfare by reducing polluting activities to (or at least toward) their efficient levels, but they can give us an improved revenue system by reducing reliance on income, sales, and other distorting taxes. Pollution taxes, in short, can give us both enhanced environmental quality and a better tax system.

III. Is There a Double Dividend?

But things do not turn out to be quite this simple. As often happens when we enter the murky waters of the second-best, our economic intuition is not a very reliable guide. In this instance, we find that, in the presence of existing distortions, the introduction of pollution taxes can itself exacerbate these distortions with a resulting increase in the level of excess burden. And the striking result in the new literature is that this effect can easily outweigh the efficiency gains from recycling the revenues so as to reduce rates on existing taxes. Thus, there may be no double dividend.(3)

There are certainly instances where a double dividend exists. Sandmo [24] cites an especially transparent case. Suppose that the revenues from pollution taxes set equal to marginal social damage provide sufficient funds to finance the entire public budget. Then, of course, all distorting taxes can be done away with. And we will achieve both efficient levels of externality-generating activities and a revenue system with no excess burden. Here there is clearly a double dividend from the introduction of environmental taxes. But, more generally, where pollution taxes must exist alongside distorting levies, the various economic linkages between the demands for different goods and in their production will typically be the source of additional excess burden.

To get a better sense of this, consider a case where there exist taxes on labor income (as in the seminal paper by Bovenberg and de Mooij [3]). There is thus a distortion in the work-leisure choice (assuming that the compensated labor supply function is not perfectly inelastic). A tax on polluting waste emissions in the production of various goods will raise the price of these goods, thereby reducing the return to work effort and inducing a substitution of additional leisure for consumption. Lawrence Goulder [10] calls this the "tax-interaction effect." And as Ian Parry [19] shows, so long as the output of polluting industries exhibits a (roughly) average degree of substitution for leisure, this effect will operate to increase the excess burden associated with the undersupply of work effort.

Note that although the induced change in work effort (employment) may be small, the incremental distortion in the labor market can be quite large because the welfare loss per unit of employment at the margin is large reflecting the large difference between the gross and net wage.(4) In terms of our usual diagrams, the point here is that the incremental distortions take the form of "tall" rectangles so that their area can be substantial even though their base is relatively small. These incremental distortions can be offset to some extent by using the revenues from the pollution taxes to reduce the rate of taxation of labor income (the "revenue-recycling effect"), but the conditions required for the revenue-recycling effect to offset fully the tax-interaction effect are stringent as Parry [18; 19] shows in his simple and very helpful analytical models.(5) Existing distortions in the labor market, incidentally, are by no means the only source of additional excess burdens. Programs, especially those that affect the prices of intermediate goods, can produce a wide range of incremental excess burdens in other markets.

We find, more generally, in the large CGE models that are being used to simulate hypothetical carbon taxes, that the incremental distortions from pollution taxes can be substantial [2; 13]. The distortions introduced by these taxes often exceed the reductions in excess burden from revenue recycling with the implication that the optimal Pigouvian tax is something less, not more, than marginal social damage. Bovenberg and Goulder [2], for example, in their seminal work on this issue find that their model suggests rates of pollution taxation significantly below Pigouvian levels. Likewise, in his simpler but more transparent analytical model, Parry [19] turns up tax rates on pollution that, for certain cases, are roughly 70 percent of marginal environmental damages.

These results are admittedly quite sensitive to the structure of the models, the values of key parameters (including the elasticity of labor supply, the elasticity of substitution between polluting goods and leisure, the tax elasticity of waste emissions, etc.), and the assumptions concerning the form of revenue-recycling. It is certainly possible (although not typical in these models) for there to exist a double dividend - that is, instances where the gains from revenue-recycling exceed the losses from the distorting effects associated with the introduction (or increase) in pollution taxes. But the general thrust of the findings seems to run in the opposite direction, thus undermining the case for a double dividend.

The cases in which we find a double dividend are typically those where there is some highly distorting tax in the system (as taxes on capital may be under certain circumstances) so that the revenues from the green taxes result in major reductions in existing excess burden (as, for example, in Jorgenson and Wilcoxen [13]). But this is a somewhat contrived result in that such efficiency gains could presumably be achieved through tax reforms of other kinds.

I don't wish to suggest that the double-dividend issue is resolved at this juncture. The outcomes vary significantly across different models and different assumptions concerning key parameter values. And they axe especially sensitive to the form that revenue-recycling takes. But the basic point seems well founded; in particular, the simpler analytical models (like those of Bovenberg and de Mooij [3] and of Parry [18; 19]) generate outcomes with no double dividend in a comprehensible and compelling way. The case for the double dividend appears shaky at best.

IV. Some Further Implications of the New Literature on Pollution Taxes

The insight that first-best policy measures may have their efficiency-enhancing properties compromised by existing distortions is obviously not a new idea; the second-best has been with us for many decades. In the field of environmental economics, for example, Albert McGartland [15] among others has shown that measures to control pollution in agriculture will have very different effects in the presence of U.S. farm programs than in an unregulated farm sector. But such cases involve policy measures with direct effects on distorted markets.

What is striking here is that apparently secondary effects operating through linkages with other markets can be of such magnitude. This same result is emerging in other sorts of studies. Edgar Browning [8], for example, has re-examined the issue of the welfare losses from monopoly. The traditional Harberger results based on the distortion in output markets suggest quite small losses. But when Browning integrates this analysis into a setting with a large existing distortion at the margin in the labor market, he finds that the welfare losses from monopoly are from 5 to 15 times higher than when the effect on the labor market is ignored!

This suggests that regulatory programs that raise prices of regulated outputs may have hitherto unimagined costs resulting from the incremental distortions they generate in other markets. Returning to environmental policy, consider the case of systems of tradeable emissions permits where the permits, rather than being sold to sources at auction, are simply distributed without charge (perhaps under a grandfathering scheme - like that of the acid-rain provisions in the 1990 Clean Air Act Amendments). Such programs generate no revenues; hence, there is no revenue-recycling effect to offset (at least in part) the tax-interaction effect. Parry [18] finds that for such programs, it is quite possible for the distortions from the tax-interaction effect to exceed the entire welfare gain from pollution control!

The findings in this newly emerging body of research must certainly be taken as preliminary at this juncture. Much remains to be understood. But this work suggests strongly that we must be cautious in drawing policy conclusions from partial-equilibrium analyses - or from general-equilibrium analyses that ignore pre-existing distortions in the system. The effects operating through the linkages between distorted markets in a general-equilibrium system appear to have far greater implications for policy design than we have appreciated. And I suspect that we have just scratched the surface of this issue.

V. On the Political Economy of Pollution Taxes

For the remainder of this paper, I shift the focus to a set of important issues in the actual design and implementation of pollution taxes. In the course of legislative debate and compromise, it is easy for environmental tax programs to become emasculated in ways that undermine their basic economic rationale. In this section, I want to comment briefly on what I see as some central matters in the actual design of pollution tax legislation.

The Tax Base

Economic theory is clear on this issue: taxes on externalities should be levied directly on the activity that generates the external cost. If the process of production of some commodity involves the emissions of damaging wastes, it is the emission of the wastes that should be the subject of a unit tax, not the commodity itself (aside from the unusual case of fixed coefficients of production). A tax on the output or profits of a polluting industry is not, in general, a good substitute for a tax on the offending activity itself.

There are, of course, cases where because of difficulties of monitoring waste emissions, we may have no alternative but to tax an activity closely related to the emitting of the damaging wastes. But wherever possible we should design the tax to address directly the polluting act. This may in some instances require some ingenuity. To take one example, automobile emissions are a primary source of urban air pollution in many countries. It is tempting to regard such emissions from a particular vehicle as beyond our monitoring capacities and to settle for a tax on gasoline. But such a tax, while perhaps discouraging driving to some extent, fails to provide needed incentives to purchase automobiles with desirable emissions characteristics and, equally important, to maintain them in ways to keep pollutant emissions at low levels. But there are taking place important advances in monitoring technology. It may soon be possible through periodic inspections (or perhaps even remote detection devices) to measure the levels of emissions from individual vehicles and then, with some measure of miles driven, to fashion tax bills that reflect reasonably accurately actual emissions.

Auto emissions is obviously a difficult case. It is clearly easier to levy pollution taxes on sulfur emissions from power plants or the carbon content of various fuels, but my point here is that in designing these measures, it is important to be sure that, as closely as possible, we are taxing the right thing.

Setting the Tax Rate

As Sandmo's [23] earlier paper and the ongoing research makes clear, the determination of the second-best optimal tax rate on polluting activities is a complicated enterprise. In particular, it involves elements that reflect both the environmental costs associated with the pollution and the way the tax measure interacts with the rest of the revenue system. In practice, this suggests that the actual design of environmental taxes will involve two distinct sets of policy makers: environmental regulators and fiscal managers. And the setting of the rates of pollution taxes will entail$some sort of compromise reflecting environmental objectives and revenue needs.

This is a tricky legislative and administrative issue. While effective cooperation between environmental and fiscal agents is certainly possible in some settings, it is also easy to think of cases where one authority or the other will determine the levels of pollution taxes. Taking a public-choice perspective on this matter, suppose that the public revenue authority [e.g., Treasury officials or a tax committee in the legislature] sets tax rates. To take an extreme case, Geoffrey Brennan and James Buchanan [6] have argued that we can view the public sector as a revenue-maximizer, a "Leviathan" that seeks to extract the most it can in public revenues from the economy. For this polar case, the authority would seek tax rates on polluting activities that maximize the inflow of revenues- rates that would get us to the top of the Laffer curve.

What would be the environmental consequences of such behavior? Some environmentalists have expressed the concern that tax rates are likely to be set too low from an environmental perspective; such rates they fear would not provide an adequate incentive to reduce waste discharges. Is this a legitimate concern or would Leviathan set rates that are too high? Or put differently, how would the revenue-maximizing rate compare with the Pigouvian tax rate? There appears not to be a general answer to this question, for it depends on the elasticity of the tax base. Suppose that given an existing set of other taxes, the environmental authority, following the dictates of some environmental economists, were to set pollution taxes such that marginal abatement costs equal marginal social damages (although this is admittedly not, in general, optimal in a distorted setting). Next, suppose that rate-setting authority were transferred to a Leviathan fiscal agent. Would this agent raise or lower the tax rates on polluting activities? The answer is that it could go either way, depending on the elasticity of the tax base (at the Pigouvian rate). If the base is revenue elastic, our Leviathan agent would lower the tax rate and vice-versa. Moreover, as Lee and Misiolek [14] suggest, there is no strong presumption here; their estimates of the elasticities of the tax base for a number of pollutants in the United States exhibit considerable variation, some well below unity and others above unity. Thus, Leviathan could turn out to be either a friend or foe of the environment!

Note further that for the Leviathan case, there would be no welfare gains from revenue-recycling; the added revenues would simply go to swell the size of the public budget. This is admittedly a contrived case. More realistically, we might expect even a revenue authority to give some weight to the environmental and other economic effects of these taxes in the setting of rates. But there are potential problems associated with the locus of tax authority. If, incidentally, it comes to a choice of placing pollution taxes under the aegis of either a fiscal authority or an environmental regulator, my preference would go to the latter. Pollution taxes are a potentially quite powerful tool for environmental management, one of a quite limited set of efficient policy instruments. To place such taxes in the fiscal domain is likely to discourage environmental regulators from shifting away from their traditional command-and-control instruments for pollution control.

On Environmental Trust Funds

Environmental taxes have not been widely employed. Moreover, as Robert Hahn [11] and others have pointed out, environmental authorities that have used such taxes have typically employed them not so much as instruments for regulating levels of polluting activities, but rather as devices to raise funds for environmental projects. Rates have often been set at levels that will raise the requisite monies for certain environmental programs, not to regulate waste flows.

This is an important issue in the political economy of environmental taxes. Various proposed bills in the United States for pollution taxes have contained within them special provisions for the creation of a "trust fund" into which the revenues would flow. The funds would then be allocated to certain kinds of environmental projects. Such earmarking of funds has a certain appeal. And in the political process it can be an effective way to coalesce the support of environmental (and other interested) groups in support of such measures.

But such inclinations, I want to argue, should be resisted. If revenues from pollution taxes are siphoned off into trust funds, they will result in increased spending, and there will obviously be no opportunity to realize the welfare gains from revenue-recycling. Moreover, it will mean that certain environmental projects are likely to be undertaken simply because there are unused monies in the trust fund. Environmental projects should have to meet the same economic and budgetary tests as other projects; they should not be undertaken simply by virtue of the availability of some earmarked funds.

Revenue-Neutral Tax Reform and Equity

A more appealing approach to the introduction of pollution taxes is as part of a revenue-neutral tax package that realizes the potential gains from revenue-recycling. This approach has considerable political potential, for the new taxes on pollution can be combined with welcome reductions in other taxes that will generate support for the proposed reform.

The Swedish Tax Reform of 1991 is instructive in this respect [1; 26]. Among other elements, the reform act included the introduction of taxes on carbon-dioxide and sulfur emissions. The revenues from these new levies were part of a package involving reductions in other tax rates. In particular, to reduce distortions in the labor-leisure choice, the top marginal income tax rates were dramatically reduced: these rates were reduced to about 30 percent for 80 to 90 percent of income earners and to about 50 percent for the remaining high income earners (from a top marginal rate of 73 percent in 1989 and 85 percent a few years earlier).(6)

Such revenue-neutral reforms can also address equity issues. There is concern that broad-based pollution taxes are likely to be somewhat regressive in their pattern of incidence. And some evidence [21] exists to support this contention. But in a revenue-neutral tax reform, it is possible to increase progressivity elsewhere in the tax system so as to offset (or more than offset) the regressivity inherent in the new taxes on pollution. Reductions in the lowest tax rates under the income tax or perhaps tax credits for low-income households can address this objective.

VI. Some Concluding Remarks

Pollution taxes can play a significant and constructive role in the revenue system. And the work described in this paper should not, in my view, be interpreted as a serious challenge to this basic claim. Not only do such taxes serve to reduce levels of polluting activities, but, as environmental economists have emphasized repeatedly, compared to their commend-and-control counterparts, they provide important incentives for research efforts into new and improved abatement technologies.

The theme of this paper is that the design of these measures must be undertaken with care. The tax base, wherever possible, should be the polluting activity itself- not some related activity. Moreover, there is a strong case to introduce these measures in (approximately) revenue-neutral ways so that the recycling of the revenues can be used to reduce some of the more damaging features (in both efficiency and equity terms) of the existing tax system. The ongoing research on this issue has much to contribute to our understanding of how best to do all this.

1. There is much current research addressing this important issue (most, at the writing of this paper, still in draft form). For a small sampling, see Bovenberg and Goulder [2], Bovenberg and de Mooij [3; 4], Boyd, Krutilla, and Viscusi [5], Goulder [9; 10], Johansson [12], Jorgenson and Wilcoxen [13], Parry [18; 19], and Schob [25].

2. I am somewhat puzzled why the work on optimal taxation in the presence of externalities by Sandmo [22; 23] and Ng [16] was overlooked. It certainly did not appear in obscure places. Ian Parry has suggested that the rather complicated optimal tax formulas may have masked the interpretation of the results and that their work had no empirical counterpart to give some sense of its likely importance. But for whatever reason these papers had no real impact on the environmental economics literature.

3. See Goulder [10] for a careful treatment of the concept of the double dividend. The literature employs this term with somewhat different meanings.

4. See Edgar Browning [7] for an illuminating treatment of the sources (both tax and non-tax) and the magnitude of the "wedge" at the margin between the gross and net wage.

5. Note that the revenue-recycling effect is itself offset to some extent by a loss in revenues from the tax on labor resulting from the reduction in work effort induced by the pollution tax. This is a kind of "tax-erosion" effect.

6. It is hard to believe that the efficiency gains from revenue-recycling were not quite sizeable here!

References

1. Bohm, Peter. "Environment and Taxation: The Case of Sweden," in Environment and Taxation: The Cases of the Netherlands, Sweden, and the United States. Paris: OECD, 1994, pp. 51-101.

2. Bovenberg, A. Lans and Lawrence H. Goulder. "Integrating Environmental and Distortionary Taxes: General Equilibrium Analyses." Unpublished paper, 1994.

3. Bovenberg, A. Lans and Ruud A. de Mooij, "Environmental Levies and Distortionary Taxation." American Economic Review September 1994, 1085-89.

4. -----. "Environmental Tax Reform and Endogenous Growth." Unpublished paper, 1994.

5. Boyd, Roy, Kerry Krutilla, and W. Kip Viscusi, "Energy Taxation as a Policy Instrument to Reduce C[O.sub.2] Emissions: A Net Benefit Analysis." Journal of Environmental Economics and Management, forthcoming.

6. Brennan, Geoffrey and James Buchanan. The Power to Tax: Analytical Foundations of a Fiscal Constitution. Cambridge: Cambridge University Press, 1980.

7. Browning, Edgar K., "The Non-Tax Wedge." Journal of Public Economics, March 1994, 419-33.

8. -----. "The Welfare Cost of Monopoly and Other Output Distortions." Unpublished paper, 1994.

9. Goulder, Lawerence H. "Effects of Carbon Taxes in an Economy with Prior Tax Distortions: An Intertemporal General Equilibrium Analysis." Unpublished paper, 1994.

10. -----. "Environmental Taxation and the 'Double Dividend:' A Reader's Guide." Unpublished paper, 1994.

11. Hahn, Robert W., "Economic Prescriptions for Environmental Problems: How the Patient Followed the Doctor's Orders." Journal of Economic Perspectives, Spring 1989, 95-114.

12. Johansson, Olaf. "Optimal Indirect Taxation in a Second-Best Perspective with Regard to Externalities, A Public Budget Restriction and Distribution Effects." Unpublished paper, 1994.

13. Jorgenson, Dale W. and Peter J. Wilcoxen. "Reducing U.S. Carbon Emissions: An Econometric General Equilibrium Assessment." Unpublished paper, 1992.

14. Lee, Dwight R. and Walter S. Misiolek, "Substituting Pollution Taxation for General Taxation: Some Implications for Efficiency in Pollution Taxation." Journal of Environmental Economics and Management, December 1986, 338-47.

15. McGartland, Albert M., "Implications of Ambient Ozone Standards for U.S. Agriculture: A Comment and Some Further Evidence." Journal of Environmental Management, No. 2, 1987, 139-46.

16. Ng, Yew-Kwang, "Optimal Corrective Taxes or Subsidies When Revenue Raising Imposes an Excess Burden." American Economic Review, September 1980, 744-51.

17. Oates, Wallace E. "Pollution Charges as a Source of Public Revenues," in Economic Progress and Environmental Concerns, edited by Herbert Giersch. Berlin: Springer-Verlag, 1993, pp. 135-52.

18. Parry, Inn W. H. "Environmental Policy in a Second Best World." Unpublished paper, 1994.

19. -----, "Pollution Taxes and Revenue Recycling." Journal of Environmental Economics and Management, forthcoming.

20. Repetto, R., R. C. Dower, R. Jenkins, and J. Geoghegan, "Green Fees: How a Tax Shift Can Work for the Environment and the Economy." Word Resources Institute Document, Washington, D.C., 1992.

21. Robison, H. David, "Who Pays for Industrial Pollution Abatement?" Review of Economics and Statistics, November 1985, 702-706.

22. Sandmo, Agnar, "Direct versus Indirect Pigouvian Taxation." European Economic Review, 1976, 337-49.

23. -----, "Optimal Taxation in the Presence of Externalities." Swedish Journal of Economics, March 1975, 86-98.

24. -----, "Public Finance and the Environment." Unpublished paper, 1994.

25. Schob, Ronnie. "Evaluating Tax Reforms in the Presence of Externalities." Unpublished paper, 1994.

26. Sterner, Thomas. "Environmental Tax Reform: The Swedish Experience." Unpublished paper, 1994.

27. Terkla, David, "The Efficiency Value of Effluent Tax Revenues." Journal of Environmental Economics and Management, June 1984, 107-23.