The Economics of Climate Change: a Primer/Chapter 4

by the Congressional Budget Office

Trade-Offs Among Policy Options

Governments may respond to the challenge that climate change poses by adopting a “wait-and-see” approach or by pursuing research programs to improve scientific knowledge and develop technological options, regulating emissions, or engaging in a combination of research and regulation. Should policymakers decide to act, they can choose from among a wide variety of approaches to regulate emissions and encourage the development of low-emissions and emissions-removal technologies.

Several characteristics of greenhouse gases make it possible to lower the costs of regulation by allowing for a great deal of flexibility in controlling emissions. Different greenhouse gases, measured in metric tons of carbon equivalent, have essentially the same effect on climate; they mix uniformly throughout the atmosphere and will only gradually affect the climate as they accumulate over time. Consequently, which gas is controlled and where—and, to some extent, whether a given reduction in emissions occurs this year or next—are immaterial. That principle is often referred to as “what, where, and when” flexibility in discussions of climate policy.

Governments could control greenhouse gas emissions in a variety of ways. Under direct command-and-control regulation, the government could specify the types of equipment and technology that may be used, or it could specify energy-efficiency or emissions standards for buildings, vehicles, and equipment. Alternatively, the government could impose emissions taxes or fees, which would discourage emissions by increasing their cost. It could also directly control emissions through a system of emissions permits, or allowances, that would strictly limit the total quantity of emissions.

Another option combining elements of taxes and permits would be a hybrid permit system under which the government allocated a fixed quantity of permits but sold an unlimited number of additional permits at a set “trigger” price. In such a system, if the cost of reducing emissions rose above the trigger price, emitters would simply buy additional permits rather than reduce emissions further. The system would thus cap the incremental cost of emissions at the trigger price—acting, in effect, like a tax.

Although U.S. environmental regulations are largely of the command-and-control type, most economists agree that as a general rule, taxes or permits—loosely termed “market-based” systems—can control emissions while offering greater flexibility and lower costs. In contrast to direct controls, market-based systems give firms and households stronger incentives to find low-cost ways to reduce emissions through behavioral changes and innovative technologies.

In the case of carbon dioxide emissions from the burning of fossil fuels, the most direct approaches would involve taxes or permits based on the carbon content of fossil fuels.^[1] Under either system, fossil fuel suppliers—producers and importers of coal, oil, and natural gas—would have to pay taxes or acquire permits in proportion to the carbon content of the fuel they sold. Such systems would be relatively simple to administer, monitor, and enforce if they were applied at the point of import or first sale because relatively few companies actually import or produce fossil fuels. The system would impose price increases or restrictions on output that would filter down the distribution chain, but it would avoid the administrative difficulties of a system that directly taxed “down-stream” retailers and consumers.^[2]

The relative ease of regulating energy-related emissions contrasts with the difficulties of regulating emissions from most other sources, particularly the substantial fraction that originates from forestry and farming. Because those other emissions come from many different kinds of mainly small sources under highly variable conditions, they tend to be much more difficult to track and measure. Although some such emissions could be regulated in a cost-effective manner (Reilly, Jacoby, and Prinn, 2003), controlling those sources would generally require different and varying approaches. That could complicate the regulatory process and might easily swamp the relatively low engineering costs of controlling some non-energy-related emissions.

For instance, carbon emissions from fossil fuels could be partly offset by paying landowners to plant trees to absorb and sequester carbon, thus reducing net emissions. Some tree planting is already supported for other purposes, such as soil conservation, and expanding those policies would be relatively straightforward. However, for the purposes of carbon sequestration, such policies are complicated by issues that do not arise in regulating fossil fuels. They include the costs of monitoring tree growth to determine how much carbon is absorbed and the difficulty of determining whether landowners would have grown the trees anyway. Another complicating factor is how sequestration activities might ripple through markets and affect carbon flows on agricultural and forest land not dedicated to sequestration. For instance, a decision to set aside a certain amount of forest for sequestration might lead to another area of forest being cleared that otherwise would have remained untouched. In that case, the carbon sequestered in the set-aside area would simply be offset by clearing elsewhere.

Taxes and Permits: Similarities and Differences

Taxes and permits affect a regulated activity in similar ways as long as people can buy and sell the permits on open markets. A tax on the carbon content of fuels directly raises the price of those fuels for the end user; a strict permit system indirectly raises the price by reducing the quantity of fuel that suppliers can sell. (As noted earlier, a fixed-price permit system works like a tax.) Either way, higher prices lead people to reduce their fuel consumption and thus their emissions. So for any level of emissions restrictions in a permit system, there is a corresponding tax level that will achieve the same purpose. In principle, both approaches should lead to identical levels and prices of emissions.

In practice, however, uncertainty about the costs and benefits of restricting emissions can greatly influence the relative cost-effectiveness of the two approaches. The government could impose a tax, expecting some level of reduced emissions; but emissions could end up higher or lower than it expected. Likewise, the government could impose a permit system with a cap on emissions and expect a given cost for meeting the cap; but that cost could end up being much higher or lower. And in either case, the price might not be consistent with the uncertain benefits of mitigating climate change. Which system is preferable depends on which type of uncertainty is the greatest and how rapidly costs rise—and benefits fall—as the government tightens restrictions on emissions.

Some research indicates that climate-related uncertainty gives an emissions tax (or fixed-price permit system) significant economic advantages over a system of strictly fixed permits, or allowances. Those advantages stem from two factors: both the costs and benefits of reducing carbon emissions are uncertain, and the incremental costs can be expected to rise much faster than the incremental benefits fall as regulation becomes more restrictive. Because climate change will result only from the long-term buildup of gases over many years, incremental benefits are essentially flat in any given year; that is, the incremental benefits of the millionth ton of carbon reduced are essentially the same as those of the billionth. In contrast, the incremental costs of reducing emissions are likely to rise sharply the more emissions are constrained.^[3] Thus, choosing to strictly limit the quantity of emissions could prove very expensive compared with the potential benefits, but choosing to impose a tax whose level reflected the expected benefits probably would not. A pricing system—of either taxes or fixed-price permits—is therefore likely to constrain emissions more cost-effectively than will a system with fixed limits on emissions.^[4]

The Distributional Effects of Regulation

Regulatory systems generally create winners and losers, even when the benefits of less pollution are ignored. Balancing the distributional effects of such systems can be more complicated and controversial than balancing their costs and benefits. Economic analysis provides several useful insights about the distributional issues involved in regulating greenhouse gas emissions.^[5]

Regulations Impose Costs

Regulations come with a price: one way or another, someone ends up paying for the environmental benefits they may generate. Households and firms, for instance, may have to make do with less energy, paying higher prices either directly or indirectly (in the form of lower wages, salaries, and profits)—or both.

Some analysts have argued that the regulation of energy markets might not be costly because energy conservation pays for itself. According to that point of view, people fail to use energy efficiently, either because they do not make sensible decisions about energy use or because they are poorly informed, or because they face a variety of market failures or barriers that deter them from making more-sensible decisions or becoming better informed. Proponents of that view believe that the government may be able to regulate energy use and emissions at a net savings to the economy by providing information, overcoming market barriers, and correcting market failures—for example, by including energy-efficiency requirements in standards governing buildings and appliances—and that the resulting energy savings may more than pay for the additional costs of more-efficient equipment.^[6]

Although energy markets do not necessarily function with textbook perfection—for instance, energy use produces pollution, the electricity distribution system is largely composed of regulated monopolies, and the electricity industry remains heavily regulated—neither are governments always able to correct energy market failures without imposing other costs that offset or even exceed the savings that the corrections might achieve. For example, inefficient electricity consumption is sometimes the result of regulations that are intended to prevent monopoly behavior by utilities. Nevertheless, governments may be able to intervene in some circumstances—for instance, by setting standards in markets in which reliable product information is hard to obtain or in situations in which specific regulatory failures may constrain businesses and households from making the most cost-effective capital investments. Economists find it difficult, however, to determine the circumstances in which standards clearly induce people to reduce their use of energy at no cost or with net savings.^[7]

Consumers Bear Most of the Direct Costs in the Long Run

Producers who are required to pay a tax will not necessarily bear the burden of the tax if they can pass it on to others. Characteristics of the markets for fossil fuels would enable producers to pass on most of the costs of emissions taxes—or the burden of higher prices under a permit system—to consumers.^[8]

Nevertheless, producers would still bear some of those costs in the short run. And many firms and workers in the energy sector—coal mine operators and miners, oil companies, and electricity producers that rely on fossil fuels for generation—would bear a disproportionate burden in lost profits and wages. (In the oil sector, however, foreign oil producers would probably bear a significant portion of those losses.) So would companies and workers in energy-intensive industries such as petroleum refining, primary metals, chemicals, and paper.^[9] In contrast, alternative energy suppliers would tend to benefit from higher demand for their products, as would natural gas producers (since natural gas contains much less carbon per unit of energy than coal does).

Regulations and Taxes Have Substantial Distributional Effects

A third important insight is that the distributional consequences of pricing and permit systems can be very large compared with their costs and benefits. Whenever the government restricts something of value, people will bid up the market price in trying to obtain it. The difference between the supply, or production, price and the higher market price is known as a scarcity rent.

If the government restricts emissions by imposing a tax, it will receive the scarcity rent as tax revenues. By contrast, if it imposes a permit system and gives the permits away, the permits’ recipients will receive the scarcity rents as higher profits—because they can either charge higher prices for the fuel they sell or sell the permit. The income received as tax revenue or scarcity rents can be many times larger than the net efficiency loss.^[10] One important consequence of that fact is that efforts to restrict emissions may encourage the affected parties to seek regulatory provisions that provide them with tax exemptions or access to permits—that is, they may engage in rent-seeking behavior. For example, fossil fuel suppliers might advocate a system in which they were given emissions permits free of charge —so that they would receive the entire scarcity rent resulting from the emissions limits.

Distributional Effects Depend on How the Government Regulates Emissions

Under a system of taxes or auctioned permits, the government would receive revenues, and it could redistribute some of them in various ways—by cutting other taxes, reducing government debt, or funding new programs.^[11] Each method of "recycling," or returning revenues to the economy, would benefit different groups of consumers and suppliers in different ways. Some of those approaches could offset some of the costs of regulation but probably not all of them.

The case of permits is more complicated than that of taxes because permits can be distributed in different ways: the government could auction them and receive revenues, it could give the permits away, or it could use a combination of the two approaches.^[12] Auctioned permits are similar to emissions taxes in their distributional effects.^[13] In contrast, freely allocated emissions permits would greatly benefit their recipients, who could reap profits from the now-scarce right to sell fossil fuels (while passing on most of the costs to fuel-consuming businesses and households) or from the sale of permits to a fuel supplier. One possible approach to a permit system, known as grandfathering, would be to give all the permits to fossil fuel suppliers in proportion to their historical sales. Another method would be to distribute permits free to households and require that fuel suppliers buy them. Suppliers would then include the cost of the permits in the price of fuel. That approach would spread regulatory costs more evenly across the population but would also involve high transaction costs.

Alternative Uses of Revenues

Most government revenues are collected from income, payroll, and sales taxes, which tend to distort taxpayers’ behavior by discouraging people from working or saving. The government also uses tax incentives to encourage certain types of activities—for example, home ownership, through the home mortgage interest deduction. Such subsidies distort households’ and businesses’ behavior by encouraging greater spending on tax-favored goods and services, relative to spending on other items. In economic terms, taxes and tax incentives impose significant losses of economic efficiency.^[14]

In contrast, emissions restrictions are intended to correct existing market failures—and thus improve economic efficiency—by discouraging harmful emissions. (When those restrictions take the form of taxes, they are referred to as Pigouvian taxes.) Of course, the restrictions also discourage productive activity to some degree and so impose a direct cost on the economy. However, if the restrictions were set at an appropriate level, their cost would be balanced by the benefits of lower levels of emissions.

But there is a catch: emissions controls—be they taxes, permits, or old-fashioned command-and-control regulations—also interact with the existing tax system and tend to aggravate its distortions. For instance, emissions restrictions would raise the prices of energy-intensive products, thus lowering real (inflation-adjusted) wages and further discouraging people from working. Through that sort of tax interaction effect, any regulation that raised the prices of products and lowered income would also impose additional, hidden costs by enhancing the distortions caused by the existing tax system. The more distortionary the existing system, the larger the interaction effect—and the higher the hidden costs—tend to be.^[15]

However, policymakers could offset at least part of the interaction effect by using the revenues from the emissions tax (or auctioned permits) to reduce the marginal rates—the rate on an additional dollar of taxed activity—of some existing distortionary taxes.^[16] Some analysts (for example, Jorgenson and Goettle, 2000, and Shackleton and others, 1993) argue that emissions taxes could even yield a "double dividend" of fewer emissions and more output if the revenues were used to eliminate particularly distortionary taxes in the current code—especially taxes that discourage saving and investment.^[17] The existing research on the question is not definitive, however.

Emissions restrictions that raised revenues, coupled with reductions in distortionary marginal tax rates, would impose significantly lower economic costs than emissions controls would in two circumstances: if the controls did not raise revenues (as in the cases of command-and-control regulations or freely allocated permits) or if they returned revenues to the economy in ways that did not reduce distortionary marginal rates. Policymakers would face a trade-off between using such revenues to offset some of the distributional effects of emissions controls (by making payments to affected producers and consumers) and using the revenues to offset some of the controls’ effects on economic efficiency (by reducing marginal tax rates). As a general rule, policymakers cannot fully achieve both goals.

These points are true for any sort of regulation, but they are particularly applicable to climate change policy because greenhouse gas regulations could involve so much money. The United States alone emits roughly 1.5 billion metric tons of carbon annually, so every dollar of tax per mtc would raise up to $1.5 billion per year. A carbon tax of $100 per mtc would raise about 15 percent as much revenue as the individual income tax and nearly 80 percent as much as the corporate income tax. Those large amounts suggest that some of the revenues from a carbon tax could be used to finance cuts in marginal income tax rates.

Emissions could also be reduced by eliminating subsidies and tax incentives that encouraged the production and consumption of fossil fuels or that encouraged deforestation. In the United States, such subsidies and incentives are fairly modest, and removing them would have relatively little impact on emissions (Congressional Budget Office, 1990).^[18] But many developing countries heavily subsidize energy use and land development. In those economies, the elimination of subsidies might lead to both reduced emissions and higher output.

Proposals for emissions taxes sometimes include a provision that the revenues be used for environmental purposes, such as an investment tax credit for energy-efficient equipment. Some studies suggest that such tax credits are considerably more effective than equivalent energy price changes in encouraging users to purchase such equipment, perhaps because purchasers focus more on up-front capital costs than on longer-term operating costs or because they are more uncertain about longer-term costs (see Jaffe, Newell, and Stavins, 2000, pp. 51-52 and 63).

However, such tax credits also have disadvantages. An emissions tax is intended to signal polluters to cut emissions; in effect, a tax credit for abatement distorts that message. Tax credits can cost the government a great deal per unit of reduced emissions, since purchasers who would have bought the equipment even without the credit receive it, too. Taken together, the literature on environmental taxation and revenue recycling suggests that using revenues from emissions taxes to finance a general reduction in taxes on all sorts of investment would be more cost-effective than using them to target investments for environmental purposes (Oates, 1992; and Baumol and Oates, 1988).^[19]

Regulation and Innovation

To a great extent, the cost of controlling greenhouse gas emissions and stabilizing atmospheric concentrations will ultimately depend on technological developments over the next century. Innovation that dramatically reduces the cost of producing energy from non-fossil sources or of sequestering carbon dioxide emissions will ease the process of controlling emissions; innovation that tends to reduce the cost of finding, extracting, and using fossil fuels will complicate it.

Although technological innovations over the long run are impossible to predict with any reliability, relative energy prices have influenced the direction and pace of research and development (R&D). For instance, when energy prices rose in the 1970s, not only did people use less energy and install more energy-efficient capital goods but businesses shifted resources into the development of energy-efficient equipment, more-efficient ways of finding and extracting fossil fuels, and alternative energy sources (Newell, Jaffe, and Stavins, 1998; Jaffe, Newell, and Stavins, 2000; and Popp, 2001).^[20]

Emissions controls that raised the prices of fossil fuels would be likely to have somewhat similar effects, tending to redirect R&D efforts from finding more fossil fuels to improving energy efficiency, developing alternative sources of energy, and removing greenhouse gases from the atmosphere. Over time, those efforts would tend to lower the incremental cost of controlling emissions, reducing the tax (or permit price) needed to achieve a given emissions target and inducing more reductions at a given tax rate. Moreover, emissions controls are likely to induce more innovation if they exact a payment from emitters, as emissions taxes and auctioned permits do. In contrast, companies would have less incentive to innovate under a system of freely allocated permits—and even less under a command-and-control regulatory system.^[21]

Although the inducement effect would tend to lessen the incremental costs of controlling emissions, analysis suggests that such benefits would be offset to some extent by the costs of research and development (Goulder and Schneider, 1996). Some of the resources used to finance R&D projects would simply be redirected from the fossil fuel sector, but some would probably be redirected from other economic activities.

Basic research is often considered to be a public good. Private firms have relatively little incentive to undertake basic scientific research on the functioning of the climate or on the costs and benefits of averting climate change because they cannot easily reap profits from the wide-spread, long-term public benefits of learning about or averting such change. Nor does industry have sufficient incentive to develop low-emissions and emissions-free energy technologies as long as the prices of fossil fuels do not reflect the potential costs of the damages to the climate that fossil fuel use may cause. Because of the disparity between private incentives and public benefits, the government may be able to play a useful role by investing in—or encouraging private industry to invest in—basic and applied R&D projects that will yield widespread public benefits. However, constraints on emissions would tend to enhance private incentives to undertake such projects and weaken that rationale for government-sponsored research.

Ancillary Benefits of Greenhouse Gas Restrictions

By reducing the use of fossil fuels, restrictions on greenhouse gas emissions would also reduce emissions of other pollutants such as sulfur dioxide from coal burning and nitrous oxide from automobiles. Those reductions, in turn, could yield a variety of benefits such as improvements in health, visibility, and water quality. Thus, the costs of mitigating emissions of greenhouse gases would be partially offset by the ancillary benefits of reducing the problems caused by conventional pollutants.

In the United States, some economic studies (for example, Burtraw and others, 1999; and Burtraw and Toman, 1997) have found that ancillary benefits from modest restrictions on carbon dioxide emissions in the electric utility sector could offset a significant part of the restrictions’ cost. Those side benefits would include lower costs for complying with current and impending regulations that restrict conventional air pollution, and health-related benefits from reduced emissions of conventional pollutants that are not already strictly controlled. More-restrictive limits on greenhouse gases could also lead to reductions of conventional emissions beyond those already mandated, yielding further ancillary benefits. However, the more emissions were reduced, the smaller the share of total costs that the ancillary benefits would offset—mainly because the additional benefits from reducing conventional air pollutants would decline while the additional cost of reducing carbon emissions would continue to rise.

In developed countries that already control pollution, ancillary benefits from restricting greenhouse gas emissions are likely to be similar to those found in the United States. But in developing countries with extensive conventional pollution problems that remain unaddressed, ancillary benefits—such as improvements in people’s health—could be significant.

---

1. The quantity of carbon dioxide emitted is directly proportionate to the carbon content of fuels and is therefore easy to measure. Carbon taxes fall most heavily on coal, which is composed almost entirely of carbon; they fall somewhat less heavily on petroleum products and least heavily on natural gas because those fuels also contain hydrogen. An emissions tax of $100 per metric ton of carbon equivalent translates to roughly $50 per short ton of coal, 25 cents per gallon of gasoline, and $1.50 per thousand cubic feet of natural gas. Other taxes on fuels—for instance, ad valorem (or value-added) taxes in proportion to sales prices or energy taxes in proportion to the energy content of fuels—would not be targeted specifically toward the carbon content and would therefore be somewhat less cost-effective in discouraging carbon emissions.
2. The characteristics of such emissions permit systems are discussed in greater detail in two studies published in 2000 and 2001 by the Congressional Budget Office. Another CBO study, published in 2002, discusses the relative merits of different approaches to regulating gasoline consumption, including a carbon tax.
3. In technical terms, price controls dominate when the marginal cost curve is steep or very uncertain and the marginal benefit curve is flat; quantity controls dominate when the marginal cost curve is flat or well understood and the marginal benefit curve is steep. A permit system would be more appropriate than a tax system if the unit cost of reducing emissions was relatively constant while the incremental damages from emissions increased very quickly with rising emissions levels. Weitzman (1974) and Pizer (1997, 1998, 1999) discuss these issues in more detail.
4. The balance could shift in favor of a strict permit system if technological advances made large reductions in emissions possible at a low unit cost that was more or less fixed.
5. Congressional Budget Office (2000) discusses the distributional impacts of different control policies for greenhouse gases.
6. Sutherland (2000, pp. 89-112) examines the differences between this “energy conservation” view and that of mainstream economics. A recent report from the energy conservation standpoint, prepared by five U.S. Department of Energy national laboratories, can be found in Interlaboratory Working Group (2000). The Energy Modeling Forum, in a 1996 report, offers a comprehensive discussion of the difficulties of identifying and measuring market failures and barriers in the energy sector.
7. In recent years, the U.S. government has tried to restrain the growth of emissions through a system of voluntary programs that attempt to identify opportunities for low-cost or costless emissions reductions and to promote them in the private sector. However, the programs have not been very successful in controlling emissions. For example, the Climate Change Action Plan developed in 1993 by the Executive Office of the President projected that voluntary programs would nearly stabilize U.S. greenhouse gas emissions at 1990 levels in 2000. (The plan is available at www.gcrio.org/USCCAP/toc.html.) In fact, emissions were roughly 12 percent higher in 2000 than they had been at the beginning of the decade. The plan’s failure was due in part to unexpectedly high levels of economic growth and low energy prices. Nevertheless, the voluntary programs’ successes are very difficult to evaluate because it is nearly impossible to determine what businesses and households would have done in the absence of the program. Welch, Mazur, and Bretschneider (2000) present a rigorous study that concludes that one such program had relatively little effect on emissions.
8. The supply of fossil energy is fairly elastic: for example, coal suppliers can easily raise or lower their production in response to small changes in coal prices. Moreover, demand for fossil energy is fairly inelastic because currently there are few cheap, plentiful substitutes for it. Inelastic demand and elastic supply together imply that energy producers can pass on taxes to consumers.
9. The Department of Energy (1997) notes that those four industries accounted for about 22 percent of manufacturing gross product originating in 1994 but 78 percent of manufacturing energy use. Yuskavage (1996) also treats this issue; updated data can be found at www.bea.doc.gov/bea/dn2/gpo.htm.
10. Tax revenues equal total emissions under the tax times the tax, whereas the net economic costs from the tax (called the efficiency loss, or the deadweight loss) are roughly one-half the tax times the reduction in emissions. A higher tax raises more revenues, but it reduces emissions even more; so the income loss rises faster than the tax revenues, and the ratio of revenues to income loss declines. For example, based on the analysis of the potential costs of the Kyoto Protocol by Lasky (forthcoming), a reduction of 5 percent in U.S. carbon dioxide emissions in 2010 would involve direct costs of just over $1 billion but would raise almost $50 billion in revenues. However, a reduction of 15 percent would cost $12 billion (10 times as much) and raise over $150 billion in revenues (less than four times as much); a reduction of 30 percent would cost almost $60 billion and raise $330 billion in revenues. In those examples, tax revenues (or permit values) are between six and 40 times the direct costs.
11. Not all of the revenues from an emissions tax would be available for redistribution. The tax would curb economic activity, reducing other tax revenues and raising government spending for income-related programs. The tax would also raise the government’s costs for purchasing fossil energy and energy-intensive goods. As a consequence, some emissions tax revenues would be needed to cover higher spending and lost revenues from other taxes. However, emissions tax revenues would generally be greater than policy-induced increases in government expenditures and revenue losses from other taxes, so net government revenues available for redistribution would rise.
12. Regarding auctions of permits for greenhouse gas emissions, Crampton and Kerr (1998) show that a standard ascending-clock auction is the most effective system to ensure that all bidders pay a uniform price that reflects the market value of the standard emissions permit. Under that kind of system, the auction would begin at a low asking price, and in each succeeding round, the price would rise and bidders would reveal the number of permits they wanted to buy at that price. The process would continue until the number of permits demanded was exactly equal to the number being auctioned.
13. Even if the government gave permits away, it would collect some revenues because permit recipients would pay taxes on their higher income. However, the government would also lose revenues from other taxes and would spend more on transfers, fossil energy, and energy-intensive goods.
14. Congressional Budget Office (1996) and Gravelle (1994) examine the distorting effects of taxes on labor supply and on saving and investment, respectively.
15. For discussions of the tax interaction effect, see Congressional Budget Office (2001), Parry (1997, 2002), Parry and Bento (1999), Parry and Oates (1998), and Parry, Williams, and Goulder (1996).
16. That “revenue recycling” effect could be particularly strong in the presence of tax incentives such as the home mortgage interest deduction (see Parry, 2002). However, as discussed in Babiker, Metcalf, and Reilly (2002), if the existing tax system was sufficiently distortionary, some forms of revenue recycling might actually enhance the interaction effect, so that the negative economic effects of the emissions tax would actually outweigh the positive environmental benefits.
17. Proponents of the "strong" version of the hypothesis argue that substituting appropriately set environmental fees for existing taxes would more than offset the tax interaction effect and thus improve both the environment and the economy. Proponents of the “weak” version argue that such a substitution would offset at least part of the tax interaction effect. The potential for a double dividend depends mainly on the distortions of the existing tax system and is thus more a statement about the existing system than about the benefits of emissions taxes. In principle, policymakers could also reduce the existing system’s distortions by replacing it with other, less distortionary taxes. That alternative would tend to lower the potential for a double dividend from an emissions tax.
18. Using a conservative definition, the Department of Energy’s Energy Information Administration (1992, 2000) estimates that federal subsidies and tax incentives to the energy sector amounted to about $7.3 billion in 1992 and $6.2 billion in 1999 (both in 1999 dollars)—or roughly 1 percent of total energy expenditures. Applying a much broader definition, a study funded by the Alliance to Save Energy and reported by Koplow (1993) estimates that subsidies in 1989 totaled from $21 billion to $36 billion in 1989 dollars—or from 5 percent to 8 percent of total energy expenditures.
19. Gravelle (1994, Chapter 5) provides a broader discussion of the cost-effectiveness of investment tax credits.
20. Some research—for example, Nordhaus’s 1997 study—suggests that the innovation inducement effect of higher energy prices will not be very large, compared with the more basic inducement to substitute capital and labor for energy.
21. Under certain circumstances, which are discussed at length in Fischer, Parry, and Pizer (1998), freely allocated permits may induce more innovation than taxes or auctioned permits. However, the case of climate change does not involve such circumstances.