To reduce environmental damages from electricity generation, public utility regulators are proposing that utilities incorporate environmental costs into their planning and investment decisions. The success of any such environmental costing program will depend on its breadth and on the accuracy of environmental cost estimates. Researchers at Resources for the Future recommend valuing environmental damages directly and find that regulatory regimes that do not apply to both new and existing electricity generating units could lead to higher levels of emissions than would occur in the absence of environmental costing.
Growing public concern about environmental damages associated with electricity generation has led to proposals for incorporating environmental costs into the planning and investment decisions of electric utilities. Regulators are proposing that, when choosing among potential new sources of electric power, utilities should select the sources that will lead to electricity generation at the lowest social cost. In this context, social cost is usually defined as the sum of private production costs and external costs, both environmental and nonenvironmental.
In 28 states and the District of Columbia, public utility regulators require or are considering the requirement that the costs of environmental damage be formally scrutinized in the utility planning process. Through this requirement, regulators hope to reduce the amount of pollution produced per kilowatt hour (kWh) of electricity generated. However, the success of environmental costing (EC) will depend on accurate estimates of environmental costs and on whether EC regulations are limited to new electricity generating units or are applied to both new and existing units.
Researchers at Resources for the Future (RFF) have considered how different methods of valuing environmental damages and different designs for EC regulatory programs would affect the efficiency of environmental costing in reducing pollution resulting from the generation of electricity. In simulating a hypothetical utility's response to various EC programs, it was found that some of these programs might actually result in increased levels of pollution.
Estimating environmental costs
Three approaches have been or are being taken to estimate the environmental costs of electricity generation: arbitrary increases in the private costs of utilities to reflect negative effects on the environment, increases in electricity prices based on the cost of abating pollution (referred to as the abatement cost approach), and increases in those prices based on the value of avoiding damages to the environment (referred to as the damage cost approach). Several states, recognizing the difficulties in valuing damages, have simply decreed that the private costs of certain types of generation technologies will be increased by a given amount to reflect negative environmental effects. Others have looked to abatement costs—that is, the costs of reducing pollution—as a reasonable and measurable proxy for the benefits of such reductions. While no states currently use damage costs—or the benefits of avoiding environmental damages—to set environmental costs, several are funding studies to provide estimates of these costs.
Which of these approaches is preferable? It is easy to dismiss the first approach for its arbitrariness, although it can be rapidly implemented and avoids the false specificity and sense of complacency that an analytical approach could create. Arbitrary private cost increases could be replaced fairly easily by approaches based on credible abatement or damage cost estimates.
Distinguishing between the abatement cost and damage cost approaches requires consideration of what should be and is being measured. Every kilowatt of electricity generated produces pollution that can affect the environment. Generally, utilities do not pay a fee for using the environment to dispose of pollutants. Therefore, the price of electricity fails to reflect environmental damage, creating what is called an "externality." However, to the extent that government regulates the amount of waste produced, causing the utilities or their suppliers to incur some costs for pollution abatement, these externalities may be partly or wholly (or even excessively) reflected in the price consumers pay for electricity.
Assuming that environmental damages have not been fully internalized despite regulation, the idea is to value this externality and to incorporate it into utilities' decision making. To obtain the value of an externality, economists suggest finding the aggregate value that individuals place on reductions in their welfare—that is, the amount individuals would be willing to pay (or would require as compensation) to leave them indifferent to the welfare loss. Changes in individual welfare can then be measured by a variety of techniques, some straightforward (as when effects on marketed goods, say damages to crops, are of concern), and others less so (as when effects on nonmarketed commodities—such as human health, recreation, and aesthetic beauty—are of concern).
Those who advocate use of the abatement cost approach argue that direct damage costs can never be adequately measured. Assuming that regulatory standards are set efficiently to equalize environmental benefits and abatement costs at the margin, they further argue that compliance costs are a reasonable proxy for damage costs.
Placing too high a value on environmental damage can be just as undesirable as placing too low a value or even a zero value on such damage.
But are these arguments sound? With respect to the measurement of damage costs, progress in the area of environmental damage and benefit estimation since the early 1970s has been remarkable. Benefit analysis has evolved from an afterthought in cost-effectiveness analyses to an integral part of environmental economics and government decision making. Environmental benefit analysis has been institutionalized in the federal Regulatory Impact Assessment process, the natural resource damage assessment provisions of Superfund, and the 1990 Clean Air Act amendments. Although the amount and success of research efforts have been unevenly distributed, there is a large body of information and studies that can be used to value environmental damages.
As for the idea that Congress and the U.S. Environmental Protection Agency set efficient regulatory standards, it is demonstrably false. For instance, the Clean Air Act specifically precludes consideration of abatement costs in the setting of air quality standards, and the process of estimating and valuing environmental damages is given little weight because the act requires that standards be set to protect sensitive individuals with a margin of safety. This involves estimating risks for only a small segment of the population, but does not involve estimating the risks and damages (or the benefits of avoiding damages) to the general population that would be needed to set "efficient" air quality standards. Even if costs and benefits were considered in setting standards, efficiency would not be the only criterion used to set them.
Given that environmental standards are not set to yield the most efficient level of emissions, a defect in the abatement cost approach becomes apparent. Its use implies that the value of reducing emissions can be the same in locations where benefits from such emissions reductions are clearly different.
It is not even true, as is sometimes claimed, that abatement costs represent a lower bound to the benefits of pollution control. This view requires the assumption, on which serious doubt has been cast, that regulators always err by making regulations too weak. Perhaps the most potent argument against an EC approach that focuses on abatement costs is that it could harm the environment or protect the environment at higher economic costs than are necessary. These outcomes might result because utilities would be led to select a less than optimal mix of generating capacity. Indeed, placing too high a value on environmental damage can be just as undesirable as placing too low a value or even a zero value on damages.
Finally, it should be understood that, irrespective of the method used to monetize environmental (and nonenvironmental) damages, simply increasing the price of electricity to reflect these environmental costs would not, in general, be appropriate. Rather, the price should be increased only to the extent that such costs have not been internalized. This determination requires analysis of the government policies to control pollution and other activities that would potentially create externalities.
Environmental costing regimes
Proposed and existing EC regulatory regimes—both of which focus primarily on the selection of new sources of generating capacity—require utilities to invest in the project(s) with the lowest social costs of electricity supply, but do not require them to actually pay the environmental costs of electricity generation. Ranking potential sources of new supply by social cost per unit of electricity produced is known as "ranking with grandfathering." This is because new generation facilities, but not existing ones, are covered by this form of EC regulation.
Another approach to EC regulation would be to impose the environmental costs of power production at new generating units on a utility by means of an emissions tax. The tax rate would be set equal to the value of marginal damages associated with an additional unit of pollution. This form of EC regulation, referred to as "taxation with grandfathering," would pass to consumers the full social cost of generating electricity at new generating units.
Under a third form of EC regulation, known as "complete emissions taxation," a utility would be required to pay taxes on emissions from both new and existing generating units. This regulatory regime would govern not only investment behavior, but also the dispatch, or order of use, of all generating units. When a utility has to pay the full social cost of each kWh of electricity generated, it will dispatch its generating units in order of unit social cost, a practice known as "environmental dispatch."
Modeling utility decision making
To explore the effects of each of these three EC regulatory regimes on a utility's investment and dispatch decisions, RFF researchers developed and simulated a model of utility planning and dispatch. The hypothetical utility created for this exercise is representative of utilities in the mid-Atlantic region of the United States. In the simulation, the objective of the hypothetical utility—which uses nuclear, coal steam, oil steam, gas steam, gas turbine, and hydro technologies to generate electricity—was to build and operate generating units in a way that minimized the cost of satisfying a given expected level of demand for electricity plus a 20 percent reserve margin. The model allowed for demand to vary over time and specified a realistic load duration curve, which depicted the duration of electricity demand at or above a particular level. In the simulation, the utility had to meet increased demand by deferring retirement of old generating units and by constructing and operating new units.
The utility could invest in generating units that utilize one of three forms of renewable energy—wind, solar heat, and water. It could also invest in four kinds of units that utilize fossil fuels—gas turbine combined cycle (GTCC) units, which use natural gas; pulverized coal with wet flue gas desulfurization (PC/FGD) units, in which a traditional coal burner is fitted with desulfurization equipment to reduce emissions of sulfur dioxide (SO2); atmospheric fluidized bed coal (AFBC) units, which use a coal burner that captures SO2 during combustion; and integrated gasification combined cycle (IGCC) units, in which coal is converted to gas, resulting in extremely low SO2 emissions. In terms of private cost per kWh of electricity generated, hydroelectric units are the least expensive, followed by IGCC, GTCC, PC/FGD, AFBC, wind, and solar units.
To simulate how the hypothetical utility would respond to EC regulation, estimates of the environmental costs associated with SO2 and two other pollutants emitted when fossil fuel is burned to generate electricity were inserted into the RFF model of utility planning. These other pollutants are nitrogen oxides (NOx) and carbon dioxide (CO2).
The RFF study considered four sets of illustrative estimates of the unit environmental costs of NOx and SO2 emissions. The first set of estimates, based on environmental damage costs, valued NOx emissions at $0.25 per pound and SO2 emissions at $0.50 per pound. The second set of estimates multiplied these values by a factor of 10. The third set of estimates, based on abatement costs, valued NOx emissions at $3.25 per pound and SO2 emissions at $0.75 per pound. The fourth set of estimates multiplied these values by 10. The costs of CO2 emissions were not based on estimates of environmental damage or abatement costs, but on two tax rates proposed for limiting these emissions: $25 per metric ton (MT) and $100/MT. The RFF study considered the effects of EC regulation both with and without valuing CO2 emissions.
Results of the model simulations
Initially, the RFF study simulated the hypothetical utility's investment decisions, its dispatch of generating units, and the related social costs under a base case scenario with no environmental costing. With no EC regulation, the utility builds new IGCC and GTCC generating units to produce 27 percent of its total electricity output. The utility builds no new solar or wind-powered units and runs its existing hydro unit at capacity during the periods of highest demand. It does make some life-extending investment in its aging coal-fired units, but allows its expiring oil-fired unit to shut down. The utility uses existing coal boilers to generate nearly 40 percent of the electricity it produces.
Not surprisingly, the RFF study found that the social costs of the hypothetical utility's investment and operating plan for the base case depend on the environmental cost assumptions adopted. The ratio of social costs to private generation costs consistent with damage-based costs and a zero value assigned to CO2 emissions is 1.5. The high level of environmental damage cost implicit in this ratio reflects the utility's heavy reliance on relatively dirty existing coal units in the base case. Assigning a value of $25/MT to CO2 emissions raises the ratio to 1.6. With a value of $100/MT, the ratio is 2.1. At 10 times the basic damage-based cost levels per unit of emission, the ratio is 5.7. Using abatement-based costs, the ratio of social costs to private costs is 5.8 in the absence of a carbon tax, 6.0 with a tax of $25/MT, and 6.4 with a tax of $100/MT. At 10 times the basic abatement-based cost levels, the ratio rises to an extremely high 48.9.
Under an EC program in which new generating units are ranked by the social cost per unit of electricity produced (ranking with grandfathering), the utility's dispatch of generating units is barely affected, but its investment behavior may shift relative to the base case. In particular, if carbon emissions are valued at $100/MT, the utility builds a new 200 MW wind farm. However, this unit accounts for less than 3 percent of the electricity generated. Due to its marginal effect on utility dispatch, ranking with grandfathering leads to only a 3 percent increase in the private costs of electricity generation. Under any environmental cost assumptions, it does not decrease the social costs of electricity generation.
A hypothetical utility's sources of electricity generation under an environmental costing program of complete emissions taxation
As compared with ranking, an EC program in which emissions from new generating units are taxed (taxation with grandfathering) has a greater effect on the utility's dispatch of generating units. The impact on dispatch is especially pronounced when carbon emissions are taxed, leading the utility to increase significantly its reliance on existing fossil-fueled units. This tendency to rely more heavily on existing units, which are generally more polluting than new units that use cleaner technologies or less polluting energy sources, may lead to an increase in social costs greater than that under ranking with grandfathering, given identical environmental cost assumptions. Social costs increase because the emissions tax applies only to new generating units, creating a strong incentive to substitute existing sources of electricity for new sources. The social costs of electricity generation never fall under taxation with grandfathering.
Under an EC program in which emissions from both new and existing generating units are taxed (complete emissions taxation), the utility's investment and dispatch decisions vary depending on the environmental costs adopted. With damage-based costs, the utility's investment and operating behavior is not changed substantially from the base case except in two instances. If CO2 emissions are taxed at $100/MT, new units that use renewable energy are built and existing units that utilize turbines fired by natural gas are employed more intensely. If SO2 and NOx emissions taxes are set at 10 times the basic damage-based cost level, the utility shuts down all of its existing coal-fired boilers, and relies heavily on new coal-burning units (see figure, p. 4).
With abatement-based costs, the utility shuts down its existing coal-fired boilers at the lowest levels at which SO2 and NOx emissions are taxed and increases its reliance on new coal-fired units. If a carbon tax is imposed, the utility increases its reliance on units that use renewable energy and reduces its use of existing oil steam units.
Taxing emissions from all generating units never leads to an increase in social costs. Under all of the illustrative environmental cost assumptions, the utility chooses generating resources that result in lower social costs than in the base case. As the estimates of assumed marginal environmental cost increase, the social cost savings from the imposition of complete emissions taxation rise. However, RFF's analysis suggests that even comprehensive EC regulation based on air pollution damage costs may have only a small effect on the investment and dispatch behavior of utilities.
Implications of environmental costing for electricity prices
The effect of EC regulation on electricity prices varies with the illustrative environmental cost assumptions and the scope of the regulatory regime. For the hypothetical utility, the average cost of generating a kilowatt hour of electricity with no EC regulation is $0.052. This unit cost estimate includes the embedded capital cost of using existing generating units but does not include the overhead costs of operating the utility.
Assuming that demand for electricity does not change, the findings of the RFF study suggest that the unit generation cost and the consumer price for electricity would rise between 0 and 0.1 cents under an EC program of ranking with grandfathering. This virtually nonexistent price effect reflects how little a requirement to invest in new sources of generating capacity with the lowest social costs of electricity supply might affect a utility's employment of generating resources.
Under an EC program of emissions taxation with grandfathering, the size of the electricity price effect depends on whether environmental costs are based on damage costs or on abatement costs. With damage-based costs, the price would not be affected unless CO2 emissions are taxed. At a tax of $100/MT, the price could rise by as much as 0.5 cents per kWh. With abatement-based costs, the price could rise 0.1 cents per kWh if only NOx and SO2 emissions are taxed; however, it would likely rise 0.5 cents per kWh when CO2 emissions are taxed at $100/MT.
Under an EC program of complete taxation, the size of the price effect again depends on whether environmental costs are based on damage costs or abatement costs. With damage-based costs, the consumer price would rise by 1 to 3 cents per kWh. With abatement-based costs, the increase in price would be smaller because the utility significantly reduces its environmental tax bill when it shuts down its existing coal-fired units. However, with environmental taxes set at 10 times the basic abatement-cost level, the price could rise as much as 4 cents per kWh.
Limitations and results of the RFF analysis
The findings of the RFF simulation study represent neither a universal prediction of utility behavior under environmental costing regulations nor a characterization of the expected behavior of a particular existing utility. Instead, the RFF study provides an empirical example of the response of a representative utility in the mid-Atlantic region to various EC regimes based on illustrative en-vironmental costs. The applicability of the findings of this study to either the general class of electric utilities or to a specific utility is restricted by the limited range of environmental externalities that are included in the analysis and by the specification of environmental costs.
With respect to externalities, the study focused only on emissions of a small number of air pollutants and, therefore, ignored other environmental externalities, such as emissions of water and soil pollutants or environmental damages associated with the construction of various types of power plants. Moreover, a constant emission factor for each pollutant at each type of generating unit was used to estimate the levels of those emissions that were considered. While this assumption is accurate for some pollutants, NOx emission rates vary with the level of capacity utilization at each generating unit.
With respect to environmental costs, the illustrative damage-based and abatement-based cost estimates adopted in this study were intended to indicate a range of possible environmental costs that might be used for implementing EC regulation. However, these costs do not reflect state-of-the-art estimates for either environmental damage costs or abatement costs for the selected pollutants. If the studies of environmental damage costs of electricity generation currently under way estimate damage costs that lie outside of the ranges included in this study, then a particular utility's response to an EC regime based on these estimates might be at variance to the responses suggested in the simulation study.
Keeping these limitations in mind, the RFF study indicates that the outcome of an EC program depends on the breadth of the program and on the environmental cost estimates on which it is based. The most effective environmental costing regime is emissions-based taxes on all sources of electricity generation. Such a regime causes utilities to internalize the environmental damage costs of all the electricity they generate and, thus, leads them to reduce the social costs of electricity generation. A less comprehensive regime, such as ranking with grandfathering or taxation with grandfathering, could lead to higher emissions levels than those that would occur in the absence of environmental costing. Under partial environmental costing, utilities might have little incentive to invest in cleaner technologies as long as the lives of existing generating units that use more polluting technologies could be extended with no regulatory penalties. This incentive is particularly weakened under a regime of taxation with grandfathering in which only those emissions from new generating sources are taxed. Thus, according to RFF's empirical analysis, social costs never decrease under partial environmental costing, whereas they never increase under complete environmental costing.
Karen L. Palmer is a fellow in the Quality of the Environment Division at RFF. Alan J. Krupnick is a senior fellow in the division. The discussion of the RFF study's simulation results is based on research conducted by Karen Palmer and Hadi Dowlatabadi, an associate professor in the Department of Engineering and Public Policy at Carnegie Mellon University and a former RFF fellow.
A version of this article appeared in print in the October 1991 issue of Resources magazine.
A version of this article appeared in print in the October 1991 issue of Resources magazine.
