Skip to main content

Three Pollutants and an Emission: A Playbill for the Multipollutant Legislative Debate

For the first time since 1990, Congress may be poised to enact major clean air legislation. Proposals now before Congress would affect electricity generation and large industrial facilities only and, as such, fall short of a full Clean Air Act reauthorization. But in other ways, the proposals are dramatic. They address multiple pollutants, broadly applying market-based approaches and aggregate emission caps, and would mandate deep emission cuts.

All the Congressional proposals and that of the Bush administration would address emissions of three pollutants—sulfur dioxide, nitrogen oxides, and mercury—from large stationary sources. The affected plants account for two-thirds of total U.S. emissions of sulfur dioxide, a quarter of nitrogen oxides, and a third of mercury. The proposals differ in whether they address carbon dioxide, even though the sources in question give off 40 percent of all U.S. emissions of carbon dioxide, a key greenhouse gas. Although carbon dioxide regulation is destined one day to play the leading role in clean air policy and thereby shape controls on the other three pollutants, for now it is backstage. Paradoxically, however, even from backstage carbon dioxide plays a critical role in the proposals’ prospects for enactment.

The stage was set in spring of last year, when President Bush used previous clean air legislation to distinguish the three conventional pollutants from carbon dioxide. Because carbon dioxide was not identified as a pollutant in the Clean Air Act, the president termed it an “emission,” not a “pollutant” and thus attempted to keep carbon dioxide out of the debate. The president’s Clean Skies Initiative, announced this past Valentine’s Day, excludes mandatory controls on carbon dioxide until at least 2012. Instead, the proposal embraces the goal of simply maintaining the modest rate of reduction in the intensity of carbon emissions per dollar of gross domestic product that was achieved in the last decade. Champions in Congress are divided mostly along partisan lines. Some proposals, including most notably that of Senator James Jeffords (I-VT), have kept carbon dioxide in the script.

All the proposals would make important—and costly—cuts in emissions of sulfur dioxide, nitrogen oxides, and mercury. But costs of the proposals that include carbon dioxide go far higher—so high that they would pretty well swamp the costs of controlling the other pollutants and thereby determine the direction of investments and technological choices for the future. Including carbon dioxide in the upcoming legislation would lead industry to begin to change its choice of fuel for power production. But if carbon dioxide emissions continue unregulated, coal—which now accounts for about 51 percent of electricity production—will continue to fuel power plants at about the same level. As recent studies by the Energy Information Administration and Environmental Protection Agency show, three-pollutant legislation that excluded carbon dioxide would necessitate retrofitting coal-fired facilities with ambitious controls for the conventional pollutants, but would have little effect on coal use and carbon dioxide emissions.

Legislation that also mandates big cuts in carbon dioxide emissions, however, would require many plants to switch from coal to natural gas. The switch to natural gas, which burns more cleanly than coal, would also mean sizable cuts in the other pollutants.

Will the Bush administration’s proposal to leave out carbon dioxide and target only the three conventional pollutants resolve the future of carbon dioxide regulation, or will the issue be revisited? Power producers who invest hundreds of millions of dollars to comply with cuts mandated by three-pollutant legislation will not expect to see major new carbon dioxide controls just around the corner. Passing three-pollutant legislation would thus seem implicitly to commit the nation to a long-term delay in carbon dioxide policy and to an energy policy with a large coal component for many years. Indeed, for a future Congress to consider regulating carbon dioxide right on the heels of passing three-pollutant legislation could be perceived by power producers as highly unfair. Anticipating that reaction, environmental advocates may have little incentive to compromise over some pretty dicey issues involved with controls on the conventional pollutants.


A key goal of the proposed multipollutant legislation is to improve on the current regulatory framework by giving industry greater regulatory predictability and certainty about compliance requirements. Paradoxically, however, achieving such certainty, as well as taking into account the interests of industry, may require giving carbon dioxide cuts at least a small role in the multipollutant strategy.

Reducing Multipollutant Levels

Debate over the legislative proposals involves three main questions. The first is what level of control to impose for each of the three pollutants. The stakes here are high, for the leading proposals would impose large costs—several billion dollars—and reap large benefits as well. But the costs are spread roughly equally over the three pollutants, while the benefits are not.

Reducing sulfur dioxide emissions will likely yield the lion’s share of quantifiable benefits. Sulfur dioxide as a gas is recognized as a potent health threat, but it is more noteworthy in economic terms for its role in forming secondary particulates and acidifying soil and water.

In the public policy world, sulfur dioxide hit marquee status with passage of the 1990 Clean Air Act Amendments, which mandated a 50 percent cut in aggregate emissions from electricity generation. At the time, Paul Portney, of Resources for the Future (RFF)—the only analyst bold enough to offer an estimate of the benefits and costs of the amendments—predicted that benefits of sulfur dioxide cuts would about justify the costs. Subsequent analysis by the EPA and RFF suggests that two unexpected developments caused benefits to exceed costs by far. First, since 1990, a cascade of new research—most recently, a study published in the Journal of the American Medical Association early in March—has linked secondary particulates from sulfur dioxide emissions with premature mortality, raising the quantifiable economic benefits of sulfur dioxide cuts many times over what was expected. Second, cutting sulfur dioxide proved less costly than anticipated.

In 1990, although the primary benefits were ultimately found to stem from improvements in health, arguably the main justification for the sulfur dioxide cuts was acid rain. Since 1990, sulfur deposition in sensitive ecological areas has fallen dramatically, but recovery has been slower than expected, leading many advocates in the northeast to call for further sulfur dioxide cuts of as much as 75 percent. Meeting that target will mean more than simply switching to low-sulfur coal at more facilities. Compliance will require the widespread installation of scrubbers—and will impose large capital costs on industry. Even so, from a benefit-cost perspective, the proposed sulfur dioxide cut is an easy pill to swallow.

Nitrogen Oxides

Pending multipollutant proposals would reduce nitrogen oxides from electricity generation to some 50-75 percent of current baseline emissions. The cuts have multifaceted and important benefits but are harder to justify on a cost-benefit basis than the cuts in sulfur dioxide.

Nitrogen oxides have effects similar to those of sulfur dioxide. They contribute to secondary particulates but are usually thought to be less potent in affecting health. They are also instrumental in acidification, but again most analyses attribute to them a lesser role. Both pollutants impair visibility. Nitrogen oxides also contribute to forming ground-level ozone, another pollutant identified in the Clean Air Act, but health scientists and economists assign lower benefits to reducing ozone than to reducing particulates.

The 1990 Clean Air Act Amendments first imposed modest cuts in nitrogen oxides emissions at power plants. In complying with the 1990 law, plants have already made the least expensive cuts. Dramatic further reductions will require widespread installation of post-combustion controls (mostly, selective catalytic reduction) and will be quite costly.

In sum, viewed from the basis of each ton reduced, the benefits of cutting nitrogen oxides appear to be equal to (or perhaps a little less than) those of cutting sulfur dioxide. But the cost per ton of the nitrogen oxide reductions is greater than the cost per ton of cutting sulfur dioxide. Thus, benefit-cost considerations would argue for greater relative reductions in sulfur dioxide—at least as long as they continue to yield the biggest bang for the buck. But current proposals would reduce the emissions of both by equal percentages.

Mercury Levels

Proposed mercury cuts are even larger, in percentage terms, than cuts in the other two pollutants. Though highly toxic and with potentially profound ecological and human health effects, mercury is emitted in much smaller quantities. And the benefits of reducing mercury emissions are hard to quantify. The Clean Air Act would normally require maximum achievable emission reduction for mercury as a hazardous air pollutant, leading to more than 90 percent removal, but the 1990 amendments gave mercury a temporary special exception. Mercury may, however, come under strict control in the next few years within the existing regulatory process unless exempted by legislation or blocked by litigation. Whichever ending is played out—implementation or exemption—will determine the baseline against which the multipollutant legislative proposals are measured. Environmental advocates seem to expect a baseline that reflects tightening controls on mercury. Realistically, however, long delays in implementing mercury controls on electricity generation are likely unless multipollutant legislation is passed.

Cost-effectiveness is an important issue. Byron Swift of the Environmental Law Institute notes that cutting use of mercury from various small sources through pollution prevention—for example, by placing a declining cap on the use of mercury in products and process—would be far less costly than controlling mercury emitted in power production. That would argue for pursuing other, less costly options before pushing far in the power sector.

Furthermore, installing post-combustion controls for sulfur dioxide and nitrogen oxides would yield sizable ancillary reductions in mercury. Selective catalytic reduction controls for nitrogen oxides oxidize much of the available mercury, and sulfur dioxide scrubbers capture the oxidized mercury. Although technical experience is limited and prediction is thus uncertain, simultaneous controls for sulfur dioxide and nitrogen oxides probably cut more than half of mercury emissions. Making greater cuts would require installing new technologies, such as activated carbon injection, that are as yet unproven in widescale application. Their cost is especially high when viewed in marginal cost terms, because the incremental reductions (those over and above the ancillary benefits of sulfur dioxide and nitrogen oxides controls) are small.

In sum, moderate mercury emissions cuts are available as part of a multipollutant strategy that targets sulfur dioxide and nitrogen oxides comprehensively, and these cuts are reflected in the administration’s proposal. Making further modest cuts would encourage development of new technologies, but the dramatic reduction targets for mercury controls embedded in Jeffords’ legislative proposals come at a high cost.

New Source Review

A second question in the debate over multipollutant legislation is whether to revise requirements for the New Source Review (NSR) program, originally established by the 1977 Clean Air Act and now under scrutiny by the Bush administration. The program requires sources that add or modify equipment that could generate new emissions to use new technology to cut emissions. New source performance standards characterize allowable emission rates for either modified or new sources anywhere in the nation. Sources that increase emissions in areas of the country that do not meet federal air quality standards are held to even stricter emission levels and must cut emissions in other facilities to offset new emissions. To some, NSR appears to be a bargaining chip in the debate; to others it is the central issue.

The NSR program’s biggest problem is uncertainty. A firm that modifies a source does not know whether it will trigger NSR requirements and what those requirements might be until a determination is made by the EPA. The technologies that are acceptable in areas that do not meet federal standards are subject to change.

Differences in emission standards between facilities under NSR and those not under NSR lead to inefficient spending on pollution control. In addition, firms under NSR may try to avoid expense and uncertainty by delaying investments to modernize facilities. And new sources in areas that do not require emission offsets increase emissions in the aggregate.

Advocates vigorously defend NSR, arguing that it has led to dramatic emission cuts and that it identifies new technologies and innovations and brings them to application. But a key goal of the legislative debate must be to streamline and increase the predictability of the NSR process as it applies to electricity generation.


During the mid-1990s Washington sharply revised its view of market-based approaches to environmental regulation. After years of suspicion, the default question is now: why not use market-based approaches? All the pending multipollutant proposals would regulate sulfur dioxide and nitrogen oxides within an emission allowance trading program based on the trading program established in the 1990 Clean Air Act Amendments. But how the trading programs are designed and whether mercury is included are major issues still in play.

The existing regional cap and trading program for nitrogen oxides operates during the five-month summer ozone season in 11 northeastern states. It will be expanded to a 19-state program in 2004. Current three-pollutant proposals would make the program national and year-round. Recent research suggests that operating the program year-round in the 19-state region would improve its cost-effectiveness dramatically. But expanding it nationwide is harder to justify because population density is much less, and consequently the aggregate benefits of reducing exposure to pollution are much less. It would also lead to a divergence in the distribution of benefits and costs across the nation that could be its undoing in Congress.

It remains to be seen whether mercury would be regulated through trading or through inflexible technology standards. Textbook treatment of pollution implies that the marginal costs of control among facilities converge as emission reduction targets approach 100 percent, because essentially every facility has to eliminate the emission. But Alex Farrell of Carnegie Mellon University has demonstrated that the marginal costs of mercury control among facilities fan out widely as controls approach 100 percent because of differences in coal types and in the other pollution controls that are in place. That argues for a trading program, which would be a departure from the usual controls on hazardous air pollutants.

One other design issue is a sleeper, but its importance will surface soon enough. How tradable emission allowances are allocated portends potentially large transfers of wealth and affects total costs as well. Because of expanded competition in the electricity industry, the value of the emission allowances may far outstrip the cost of compliance for power producers when allowances are allocated at zero cost. And prices for consumers may go higher than is justified by producer costs. When a similar scenario unfolded as industry complied with the 1987 Montreal Protocol to protect stratospheric ozone, Congress captured the “windfall profits” by taxing them. The same thing may happen here.

Looking Ahead

Compromise would enhance the cost-effectiveness of the multipollutant legislation and, at least from the view in the balcony seats, increase its chance of passage. Sulfur dioxide reductions of the size being proposed have apparent justification, but the nitrogen oxide reductions proposed for the entire nation are hard to justify economically and in any case face political opposition in the western states. A compelling alternative would be to transform the seasonal program in the 19 eastern states into an annual one and to impose somewhat less stringent controls in the other states. The case for maximum reductions in mercury emissions is the weakest. A better strategy would be to claim the mercury reductions that are achieved ancillary to imposed reductions on sulfur dioxides and nitrogen oxides as a victory. In fact, this approach may provide a justification for the cost-effectiveness of nationwide nitrogen oxide controls that is missing if those controls come on top of controls on mercury (and vice versa). One could also question whether the NSR program needs to exist at all if stringent caps are placed on emissions. But unless those caps decline continuously over time, the NSR program will be essential from the perspective of environmentalists.

The administration’s proposal embodies many of these suggestions with respect to levels of control (72 percent reduction for sulfur dioxide and 67 percent reduction for nitrogen oxides). It goes most of the way to meeting the Jeffords proposal with respect to sulfur dioxide and nitrogen oxides (75 percent reductions for both pollutants), but calls for mercury reductions that would be mostly ancillary to these (69 percent) rather than the maximum achievable (90 percent). The big difference between the proposals is timing. The Jeffords proposal would achieve these reductions a decade sooner than does the administration plan. Given the maturity of control technologies, the shorter timeline is practical and preferable.

The most important consideration, however, is what role is given to limiting carbon dioxide emissions. Large cuts do not appear plausible, but zero reductions do nothing to resolve the most important environmental uncertainty facing the electricity industry. A policy of modest cuts, beginning now and implemented through market-based policies, would stabilize the setting for new investment, provide incentive for innovation in technologies and institutions, and provide the opportunity to learn about the costs of climate change policy.

Every negotiation is shaped by expectations about what will happen if it breaks down. If the multipollutant debate collapses, what is expected is a return to the status quo—the so-called emission baseline. But like so much in this debate, the baseline itself is uncertain—so much so that the parties may find it more useful to compromise over legislation than to continue the familiar battles over a shifting baseline. Environmental advocates feel that regulatory inertia favors continuing reductions in conventional emissions. Regulatory calendars stretching to 2015 foresee the implementation of new particulate standards, mercury controls, and regional haze rules that will affect the various emissions individually. But advocates of multipollutant legislation have to feel their possible advances are at risk under the Bush administration.

The bellwether may be how the administration completes its ongoing review of NSR. If it backs away from proposed changes that would give power plants more freedom to expand or modernize without coming under NSR requirements, industry may feel it has the most to gain from legislation. If the administration review leads to substantial revisions or elimination of NSR, the baseline will be changed. Either fur may fly in Congress, or the stage may be set for an important compromise around multipollutant legislation. Or maybe both.

Get daily updates from Brookings