Ancillary Benefits of Reduced Air Pollution in the United States from Moderate Greenhouse Gas Mitigation Policies in the Electricity Sector Dallas Burtraw, Alan Krupnick, Karen Palmer, Anthony Paul, Michael Toman, and Cary Bloyd December 2001 • Discussion Paper 01–61 Resources for the Future 1616 P Street, NW Washington, D.C. 20036 Telephone: 202–328–5000 Fax: 202–939–3460 Internet: http://www.rff.org © 2001 Resources for the Future. All rights reserved. No portion of this paper may be reproduced without permission of the authors. Discussion papers are research materials circulated by their authors for purposes of information and discussion. They have not necessarily undergone formal peer review or editorial treatment. Ancillary Benefits of Reduced Air Pollution in the United States from Moderate Greenhouse Gas Mitigation Policies in the Electricity Sector Dallas Burtraw, Alan Krupnick, Karen Palmer, Anthony Paul, Michael Toman, and Cary Bloyd Abstract This paper considers how moderate actions to slow atmospheric accumulation of greenhouse gases from fossil fuel use also could reduce conventional air pollutants in the United States. The benefits that result would be “ancillary” to greenhouse gas abatement. Moreover, the benefits would tend to accrue locally and in the near term, while benefits from reduced climate change mostly accrue globally and over a time frame of several decades or longer. The previous literature suggests that changes in nitrogen oxides (NO ) would be the most important consequence of moderate carbon policies. We x calculate these changes in a detailed electricity model linked to an integrated assessment framework to value changes in human health. A tax of $25 per metric ton of carbon emissions would yield NO related x health benefits of about $8 per metric ton of carbon reduced in the year 2010 (1997 dollars). Additional savings accrue from reduced investment in NO and SO abatement in order to comply with emission X 2 caps. These savings sum to $4-$7 per ton of carbon reduced. Total ancillary benefits of a $25 carbon tax are estimated to be $12-$14, which appear to justify the costs of a $25 tax, although marginal benefits are less than marginal costs. At a tax of $75 per ton carbon, greater health benefits and abatement cost savings are achieved but the value of ancillary benefits per ton of carbon reductions remains roughly constant at about $12. Key Words: climate change, greenhouse gas, ancillary benefits, air pollution, co-control benefits, nitrogen oxides, sulfur dioxide, carbon dioxide, particulates, health JEL Classification Numbers: H23, I18, Q48 Contents I. Introduction........................................................................................................................1 II. Background.......................................................................................................................3 Emissions............................................................................................................................3 Health Effects......................................................................................................................4 The Baseline........................................................................................................................5 III. The Models.....................................................................................................................10 IV. Results.............................................................................................................................14 V. Previous Estimates..........................................................................................................22 VI. Uncertainty.....................................................................................................................30 VII. Conclusion....................................................................................................................32 References..............................................................................................................................36 Ancillary Benefits of Reduced Air Pollution in the United States from Moderate Greenhouse Gas Mitigation Policies in the Electricity Sector Dallas Burtraw, Alan Krupnick, Karen Palmer, Anthony Paul, Michael Toman, and Cary Bloyd ∗ I. Introduction A number of actions to slow atmospheric greenhouse gas (GHG) accumulation from fossil fuel use would also tend to reduce various "criteria" air pollutants (as defined in the Clean Air Act). The benefits that result would be "ancillary" to GHG abatement. Moreover, these benefits would tend to accrue in the near-term as does the cost of abatement, while any benefits from reduced climate change mostly accrue over a time frame of several decades or longer. In addition, ancillary benefits accrue largely to those countries undertaking mitigation action, in contrast to the benefits of reduced climate change risks that accrue at a global level. A failure to adequately consider ancillary benefits could lead to an incorrect assessment of the "net costs" of mitigation policies--that is, the direct cost of climate policy less ancillary benefits that accrue from those policies--and an incorrect identification of "no regrets" levels of GHG mitigation. It also could lead to the choice of a policy that was unnecessarily expensive because of its failure to fully exploit potential ancillary benefits. This paper presents results from a model of the electricity sector called Haiku. The model calculates market equilibrium by season and time of day for three customer classes at the ∗ Burtraw, Krupnick, Palmer, Paul, and Toman, Resources for the Future; Bloyd, Argonne National Laboratory. The authors are grateful to Lawrence Goulder, Anne Grambsch, Peter Nagelhout and Joel Scheraga who shared co- authorship on an early draft of this paper and contributed significantly to its development. The authors are also grateful to Roger Dower and John Firor for comments on previous drafts, and to Ranjit Bharvirkar, Matt Cannon, Martin Heintzelman, Erin Mansur and Meghan McGuinness for outstanding assistance. This work was funded by the U.S. Environmental Protection Agency and by the U.S. Department of Energy. All errors and opinions remain the responsibility of the authors. Address correspondence to Burtraw, Resources for the Future, 1616 P Street NW, Washington DC 20036 (email: [email protected]). 1 Resources for the Future Burtraw et al. regional level, with power trading between regions. The model is used to simulate the effects of various moderate carbon taxes on investment, retirement and system dispatch for the year 2010, and on changes in emissions of nitrogen oxides (NO ) that result from these carbon taxes. We x model alternative baselines in the absence the GHG policy, all of which go beyond requirements of the 1990 Clean Air Act. In one case, we model full implementation of Title IV of the 1990 Clean Air Act in the electricity sector, coupled with Phase II of NO reductions in the X northeastern 11 state Ozone Transport Commission region. In another case, we include further reductions in the baseline by applying NO emission rates in an eastern 19 state region to comply x with standards expected to take effect in 2004, affecting the so-called “SIP Call” region associated the requirement that states revise their State Implementation Plans. In a sensitivity analysis, we vary the representation of the regulatory structure in the electricity industry. We find health-related ancillary benefits from further reductions in NO emissions under x a $25 carbon tax to be about $8 per metric ton of carbon reduced (1997 dollars). Aggregate reductions in sulfur dioxide (SO ) are not affected by the moderate carbon policies we model, but 2 additional savings accrue from reduced investment in NO and SO abatement in order to X 2 comply with emission caps. These savings sum to $4-$7 per ton of carbon reduced. Total ancillary benefits of a $25 carbon tax are estimated to be $12-$14. These compare to expected average cost of carbon reductions of about $12 for a $25 tax. Hence ancillary benefits contribute significantly to a justification for the moderate carbon tax of this magnitude, though the marginal ancillary benefits are less than marginal costs of a $25 tax. At a tax of $75 per ton carbon, greater health benefits and abatement cost savings are achieved but the value of ancillary benefits per ton of carbon reductions remains at about $12. These compare to expected average cost of carbon reductions of less than $37.5 for a $75 tax. In this case ancillary benefits are expected to be about one-third of the average cost per ton. These findings compare favorably with the most detailed models that have been used in the previous literature, reviewed in Section V, after accounting for the omissions in those models that have been explicitly captured in this analysis. Numerous uncertainties surround the estimates and the choice of assumptions in the parameterization of the models. Some of the previous literature has obtained relatively large 2 Resources for the Future Burtraw et al. estimates of ancillary benefits under assumptions that have been criticized. Therefore in this study we have tried to buttress the conclusions with assumptions that are well within the mainstream but may be likely to achieve smaller estimates than would defensible alternative assumptions. The main result survives this cautious approach. We find that ancillary benefits weigh importantly in the consideration of climate policy and provide near-term and local benefits that offset an important portion of the costs of the policies. II. Background Three types of methodological issues are important to the consideration of how GHG mitigation could yield ancillary benefits (Krupnick, Burtraw and Markandya, 2000). These include the characterization of changes in emissions, the characterization of health benefits, and the baseline against which these changes are measured. Emissions Recent comprehensive studies of electricity fuel cycles indicate that the lion's share of the quantifiable environmental and public health effects of fuel and technology choices in electricity generation stem from air emissions (Lee et al., 1995; Rowe et al., 1995; EC, 1995). In reviewing these studies, Krupnick and Burtraw (1996) find that 82% to 93% of all quantifiable damages (e.g. excluding climate change and species biodiversity) stem from the air-health environmental pathway. Other effects may exist but are not quantifiable at this time (Burtraw et al., 1998). The major component of quantifiable damage is attributable to the change in particulate concentrations. Previous studies that address only the electricity sector identify potentially significant reductions in NO that may result from policies aimed primarily at reducing CO emissions. The x 2 studies vary in their predictions about reductions in SO depending on their treatment of the 2 emission cap under the 1990 Clean Air Act Amendments, an important baseline issue we discuss below. Secondary pollutants (sulfates, nitrates and ozone) are treated in an inconsistent manner across previous studies, and often are not mentioned at all. 3 Resources for the Future Burtraw et al. In this study we focus on the reduction in emissions of NO that are ancillary to CO x 2 emission reductions achieved in the electricity sector, and which is the pollutant of greatest interest in previous studies. We focus on the effect of NO directly and through particulate x (nitrate) formation (but excluding ozone formation) on health effects. These limitations contribute to the view that our estimates may be a lower bound of the estimates that would be achieved if a complete analysis was possible. The focus on the electricity sector in not especially limiting. The sector is responsible for one-third of CO emissions presently, and the EIA projects that this sector 2 will be responsible for about three-quarters of CO emission reductions in the United States under 2 economy wide and cost-effective climate policies (USEIA, 1998). This sector will be especially important as the least expensive and likely first source of reductions under moderate reduction scenarios. Health Effects Many previous studies have attempted to calculate health benefits based on aggregated "unit values," i.e., uniform estimates of benefits expressed as "dollars per ton of pollutant reduced." These estimates do not incorporate information about geography and demography in valuing benefits. An alternative method, the "damage function approach," focuses on estimating the social cost of electricity generation from facilities examined on an individual basis. This approach has been used in recent analyses of environmental impacts of electric power plant siting and operation in specific geographic locations (Lee et al., 1995; EC, 1995; Rowe et al., 1995; Banzhaf et al., 1996). The damage function approach is more complex than the use of simple unit values. However, the results of detailed studies may be generalizable. Krupnick and Burtraw (1996) survey three major social cost studies and largely reconcile the differences in quantified damages from conventional pollutants based on measurable differences in technical parameters at the power plants and in the size of exposed populations, although atmospheric modeling remains an important source of unpredictable variation. 4 Resources for the Future Burtraw et al. It also is important to account for changes in population, especially since population trends have greatly outstripped energy prices over the last century.1 The United States’ population is expected to grow by 45% over just the next fifty years, which coupled with expected income growth, suggests that there will be greater exposure to a given level of pollution and consequently greater benefits from reducing that pollution (Krutilla, 1967). This demographic consideration suggests that the reported values for conventional pollutants in previous studies underestimate damage in future years, if all other things are equal. In this study we use a damage function approach that involves an atmospheric transport model linking changes in emissions at a specific geographic location with changes in exposure at another location. Concentration-response functions are used to predict changes in mortality and a number of morbidity endpoints. The model accounts for expected changes in population, and for expected changes in income that affect estimates of willingness to pay for improvements in health status. The Baseline An analysis of benefits requires a clear definition of a baseline against which the prospective scenario can be measured. In a static analysis the baseline can be treated as the status quo, but since climate policy inherently is a longer-term effort, questions arise about projecting energy use, energy regulation, technology investments, and emissions of GHGs and criteria pollutants with and without the GHG policy (Morgenstern, 2000). One potentially important aspect of the baseline is the regulation of the electricity sector. In this analysis we adopt a cautious assumption regarding the future regulation of the industry by assuming that traditional average cost pricing continues in effect for most of the nation over the study period. Seven subregions of the North American Electric Reliability Council (NERC) located in the northeast (New England and New York State), the west (California and the 1 In real terms, energy prices have been about constant for the last century. The price of oil in the U.S. has fluctuated between $15 and $20/bbl for about a 100 years, except for the period 1974-1985. The mean jumped slightly for the period after 1986 as compared to that before 1973. 5 Resources for the Future Burtraw et al. mountain states) and Texas, are modeled to have marginal cost pricing. The year in which restructuring is assumed to occur is reported in Table 1. In sensitivity analysis we explore an alternative scenario and describe the effect of electricity restructuring and marginal cost pricing at the national level. The issue of the baseline is confounded further because of ongoing changes in the standards for criteria air pollutants. If one proceeds on the basis of historical standards and ignores expected changes in the standards, one would fail to anticipate that there may be less NO emitted per ton of CO than there is today and the ancillary benefit estimate will overstate X 2 environmental savings. Historical emission rates may be ten times the rates that apply for new facilities. The recent tightening of standards for ozone and particulates and associated improvements in environmental performance over time imply that benefits from reductions in criteria air pollutants resulting from climate policies will be smaller in the future than in the present. The benefits of NOX reductions from current levels would have already been achieved, but the credit for the improvement could not be given to the climate policies. This underscores the general point that focusing on the ancillary benefits of climate policies is a partial view. Furthermore, the nature of the ancillary benefits varies directly with the structure of the environmental policy that is in place (Lutter and Shogren, 2001). For example, regulation that establishes uniform emission rates such as a performance standard for new or all sources would enable reductions in conventional pollutants at those sources as a facility is utilized less. On the other hand, a cap and trade program will prevent aggregate emissions from changing as long as the cap continues to bind under the carbon policy. A climate policy is likely to yield savings in avoided investments in abatement under each type of policy, though the magnitude of those savings will differ greatly. Hence, absent the promulgation of a specific policy or identification of a specific proposal for implementing future emission reductions, one cannot estimate the ancillary benefits of concomitant climate change policy. In this study we look as far as possible into the future with respect to regulation of conventional pollutants as far as specific proposals regarding the shape of the regulation have taken shape. 6 Resources for the Future Burtraw et al. Finally, it is also challenging to establish a baseline for technological change.2 The rapid introduction of new technologies such as fuel cells could change both the overall efficiency of energy use but also the fuel type, but the rate of penetration is difficult to anticipate. Since the end point of this study is 2010, the technology baseline uncertainties should be small. Table 1. Listing of NERC subregions, the year marginal cost pricing begins, and subregions covered by cap and trade NOX policies under modeled scenarios. Year OTC NO SIP NO X X NERC Geographic Area Marginal Trading Trading Subregion Cost Pricing Region Region Regime Begins ECAR MI, IN, OH, WV; part of KY - ECAR ERCOT Most of TX 2002 MAAC MD, DC, DE, NJ; most of PA 2000 MAAC MAAC MAIN Most of IL, WI; part of MO 2001 MAIN MAPP MN, IA, NE, SD, ND; part of WI - NY NY 1999 NY NY NE VT, NH, ME, MA, CT, RI 2000 NE NE FRCC Most of FL - STV TN, AL, GA, SC, NC; part of VA, MS, KY - STV SPP KS, MO, OK, AR, LA - NWP WA, OR, ID, UT, MT - RA AZ, NM, CO, WY 2001 CNV CA, NV 1998 In this paper, baseline controls include restrictions on NO emissions beyond Phase II of x Title IV of the 1990 Clean Air Act Amendments. These controls are modeled as cap and trade programs set to achieve an average emission rate of 0.15 lbs. per million Btu of heat input at all fossil-fired and wood-fired generation facilities. In one baseline, we model further reductions 2 For example, SO emissions in 2020 that were forecast in 1990 varied by a factor of two on the basis of 2 expectations of clean coal technology and plant life (USNAPAP 1991, p. 222). 7