Mercury Emissions from Electric Power Plants: An Analysis of EPA's Cap-and-Trade Regulations

CRS Report for Congress
Mercury Emissions from Electric Power Plants:
An Analysis of EPA’s Cap-and-Trade Regulations
Updated January 13, 2006
James E. McCarthy
Specialist in Environmental Policy
Resources, Science, and Industry Division

Congressional Research Service ˜ The Library of Congress

Mercury Emissions from Electric Power Plants:
An Analysis of EPA’s Cap-and-Trade Regulations
EPA studies conclude that about 6% of American women of child-bearing age
have blood mercury levels sufficient to increase the risk of adverse health effects
(especially lower IQs) in children they might bear. Thus, there was great interest in
the agency’s March 15, 2005, announcement that it was finalizing new regulations
to control mercury (Hg) emissions from coal-fired electric power plants — power
plants account for 42% of total U.S. mercury emissions, according to EPA.
In announcing the regulations, however, EPA stated that most mercury in the
atmosphere comes from non-U.S. global sources. Even if regulations could reduce
power plant mercury emissions to zero, the agency concluded, there would be little
change in the mercury health effects it has identified. (This point is contested; many
argue that power plants make significant contributions to mercury emissions in local
hot spots, and that reducing power plant emissions would significantly reduce risks.)
Instead of stringent, plant-specific requirements, EPA promulgated “cap-and-
trade” standards that rely heavily on co-benefits from sulfur dioxide and nitrogen
oxide controls installed under a separate agency rule, the Clean Air Interstate Rule
(CAIR). This approach minimizes costs for electric utilities: by 2015, less than 1%
of coal-fired power plants will have installed equipment specifically designed to
control mercury under the Hg rule, according to EPA. By 2020, only 4% of plants
will have such equipment. Sixteen states have filed suit to overturn the agency’s
action, arguing that EPA is required by the Clean Air Act to impose more stringent
Maximum Achievable Control Technology standards at each individual plant.
Beginning in 2010, the cap-and-trade rule limits total power plant mercury
emissions to 38 tons annually (a 21% reduction vs. 1999 levels). A second phase
caps annual emissions at 15 tons, starting in 2018. According to the agency, trading
and banking of emission allowances will result in lower than required emissions in
the early years, but will delay achievement of the 15-ton cap to at least 2025.
EPA has sent contradictory signals regarding the importance of controlling
mercury emissions. Its January 2004 analysis estimated that the indirect benefits of
proposed regulations would be $15 billion annually, outweighing compliance costs
by a factor of at least 16 to 1. Direct benefits (although unquantifiable) were said to
be “large enough to justify substantial investment in Hg control.” The analysis of the
final rule, by contrast, concludes that quantifiable direct and indirect benefits of
mercury control are just $43 million per year. EPA’s calculations did not include
consideration of an academic study that it had funded, a factor contributing to the
calculation of smaller benefits. This decision was one of several irregularities in the
regulatory process alleged by the agency’s Inspector General, GAO, and others.
In addition to EPA’s regulatory effort, at least six bills that would regulate these
emissions have been introduced so far in the 109th Congress, with more expected.
S. 131, the Clear Skies Act, has many points in common with the EPA regulatory
approach. This report will be updated if developments warrant.

Background ..................................................1
Sources of Emissions/Status of Regulations.........................2
Regulation of Non-Utility Sources............................4
Electric Utilities and Mercury....................................4
Statutory Requirements.....................................4
EPA’s Utility Regulations...................................5
Costs and Benefits of the Final Rule...........................8
Availability of Technology.................................13
Hot Spots...............................................14
Effects on Eastern and Western Coal..........................16
The Regulatory Process....................................17
Legislation in the 109th Congress.................................19
Conclusion ..................................................20
List of Tables
Table 1. Mercury Emissions Estimates and Current Regulatory Status........3
Table 2. Mercury Caps and Projected Emissions Under the Mercury
Cap-and-Trade Rule............................................7
Table 3. EPA Projections of Installed Pollution Controls at Coal-Fired
Electric Generating Units, 2010-2020, in gigawatts (Gw)..............10
Table 4. EPA Estimates of the Costs and Benefits of MACT vs. Cap and
Trade for Utility Mercury Controls...............................11
Table 5. Estimated Changes in Coal Use from Imposition of the Mercury (Hg)
Cap-and-Trade Rule, 2003-2020, by Region

Mercury Emissions from Electric Power Plants:
An Analysis of EPA’s
Cap-and-Trade Regulations
Mercury is a potent neurotoxin that can cause adverse health effects at very low
concentrations. Concerns about public exposure to mercury have grown in recent
years as research has indicated its presence at significant levels in numerous species
of fish, and as analyses of dietary intake and resulting blood levels have pointed to
potential health risks from mercury ingestion, particularly for women of child-bearing
age and developing fetuses.
According to the Environmental Protection Agency (EPA), as of December
2004, 44 states had issued fish consumption advisories due to mercury. Twenty-one
states (primarily in the Midwest and Northeast) have issued statewide advisories for
mercury in all their freshwater lakes and/or rivers. In all, these advisories cover more
than 13 million acres of lakes and roughly 765,000 river miles.1 Twelve states in the
Southeast and New England, have statewide advisories for mercury in their coastal
waters, and Hawaii has a statewide advisory for mercury in marine fish.
Mercury reaches water bodies from naturally occurring sources, from past uses
(many of which, such as fungicide application to crops, are now banned), from
disposal of mercury-containing products, and from current activities — principally
combustion of fuels containing mercury in trace amounts. Mercury released to the
atmosphere can circulate for up to a year before being deposited on land or in water.
Thus, it is widely dispersed, and often is transported thousands of miles from the
emissions source.2 According to EPA, U.S. sources contributed only 3% of the 5,5003
tons of mercury emitted to the atmosphere on a global basis in 1995. Of the mercury
deposited in the United States, however, about 60% comes from U.S. sources.4

1 U.S. EPA, Office of Water, “2004 National Listing of Fish Advisories,” Fact Sheet,
September 2005, p. 4, at [].
2 U.S. EPA, Office of Air Quality Planning and Standards, 1997 Mercury Study Report to
Congress: Overview, December 1997, p. 1, available at [

112nmerc/me rcover.html ].

3 Ibid.
4 U.S. EPA, Office of Air and Radiation, Mercury White Paper, p. 1, available at [http://
www.epa.go v/ttn/oarpg/t3/memoranda/whtpaper.pdf].

This percentage is estimated to be “even higher in certain regions (e.g., northeast
U.S.),”5 because mercury emissions are concentrated in specific areas, and because
of variations in precipitation patterns. The highest deposition rates, according to
EPA, “occur in the southern Great Lakes, the Ohio Valley, the Northeast, and
scattered areas in the Southeast.”6
Of particular concern for aquatic organisms and human health is mercury in the
form of methyl mercury. Nearly all of the mercury that accumulates in fish tissue is
methyl mercury, an organic compound formed by a microbial process, often in
wetland environments. Once formed, methyl mercury tends to bio-accumulate in
aquatic organisms, increasing concentrations at each level of the food chain. “As a
result, top predators in a food chain, such as largemouth bass or walleye, may have
concentrations of bioaccumulative contaminants in their tissues a million times
higher than the concentrations found in the waterbodies.”7
Children born to women with fetal cord blood concentrations of mercury above

5.8 parts per billion (ppb) “are at some increased risk of adverse health effects,”8

according to EPA. These health effects include delayed development, neurological
defects, and lower IQ. Recent data reported by the Centers for Disease Control and
Prevention indicate that about 6% of women of child-bearing age had blood-mercury
levels above the 5.8 ppb threshold in 1999-2002.9 Because mercury levels in fetal
cord blood are likely to be higher than levels in maternal blood, the percentage of
babies that might be at risk is likely to be greater.
Sources of Emissions/Status of Regulations
As shown in Table 1, U.S. air emissions of mercury come from eight principal
sources. Of these, the largest source is coal-fired utility boilers (i.e., coal-fired
electric power plants). These accounted for an estimated 52 tons of mercury
emissions per year in 1994-1995, about one-third of total U.S. mercury emissions at
the time.10

5 U.S. EPA, “Regulatory Finding on the Emissions of Hazardous Air Pollutants from
Electric Utility Steam Generating Units,” 65 Federal Register 79827, December 20, 2000.
6 Ibid.
7 National Listing of Fish Advisories, previously cited, p. 4.
8 EPA, Office of Children’s Health Protection, America’s Children and the Environment:
Measures of Contaminants, Body Burdens, and Illnesses, 2nd edition, February 2003, at
[], p. 59.
9 “Blood Mercury Levels in Young Children and Childbearing-Aged Women — United
States, 1999-2002,” Centers for Disease Control and Prevention, MMWR Weekly, Nov. 5,

2004, v. 53, n. 43, pp. 1018-1020.

10 EPA does not have current data for all sources of mercury emissions. Since the mid-
1990s, mercury emissions have been reduced substantially from the three waste
combustor/incinerator categories, and marginally from electric utilities. In its regulatory
impact analysis for the proposed electric utility rule, EPA used 1999 data showing utility
emissions of 48 tons as the baseline against which to compare reductions.

Table 1. Mercury Emissions Estimates
and Current Regulatory Status
Emissions% ofCurrent Status of
Source(tons/year)Total Mercury Regulations
Coal-fired Utility5233%Rule finalized 3/15/05 would set a
Boilers38-ton cap on emissions in 2010 and
a 15 ton cap by 2018 (69%
Large (>250 tons per3019%Regulated: current emissions
day) Municipal Wasteestimated at about 95% below 1990
Combustors (MWC)levels.
Coal-fired Commercial2113%Rule promulgated 9/13/04 would
/Industrial Boilersreduce emissions by 1.9 tons by


Medical Waste1610%Regulated: current emissions
Incineratorsestimated at about 94% below 1990
levels, mostly through closures.
Oil-fired Commercial/85%Rule promulgated 9/13/04 would not
Industrial Boilersrequire reductions.
Mercury Cell Chlor-7*4%Rule promulgated 12/19/03 will
alkali Plantsreduce emissions 74% by 12/19/06.
Hazardous Waste74%Regulated: current emissions estima-
Combustorsted at about 50% below 1990 levels.
Portland Cement Plants53%Rule promulgated 6/14/99 reduces
hazardous air pollutant metal
emissions 24%; remanded by U.S.
Court of Appeals, D.C. Circuit to
require specific standard for mercury
and two other pollutants.
Other (at least 15128%Generally not regulated
categories of minor
Sources/Notes: Pre-regulation emissions data from U.S. EPA, Mercury Study: Report to Congress, Volume
II. RPS-452/R-97-004 (December 1997), p. ES-6; current status of mercury regulations from U.S. EPA,
Mercury White Paper (not dated), with information updated by CRS telephone communications with EPA.
Emissions estimates are for 1994-1995. Percentages do not add because of rounding.
* Data for the chlor-alkali sector are subject to considerable uncertainty. Chlor-alkali plants produce chlorine
by subjecting liquid mercury and saturated brine to an electric current. The mercury binds with potassium or
sodium in the process, but later is separated and reused. In the year 2000, 65 tons of mercury that the plants
consumed could not be accounted for.

Regulation of Non-Utility Sources. As of March 2005, EPA had made
regulatory promulgations for all major sources of mercury emissions. The authority
for most of these regulations is Section 112 of the Clean Air Act (42 U.S.C. 7412),
which requires National Emission Standards for Hazardous Air Pollutants. In
amending Section 112 in 1990, Congress included a list of 188 hazardous air
pollutants to be regulated — mercury among them. EPA was directed to identify
sources of these pollutants and impose Maximum Achievable Control Technology
(MACT). Section 112 defines MACT for new facilities as an emission standard no
less stringent than what is achieved in practice by the best controlled similar source
(i.e., the best demonstrated technology). For existing facilities, it allows a somewhat
less stringent standard, setting the average emissions of the best performing 12% of
units in the category as a minimum in most cases, but giving EPA discretion to set
a more stringent standard. Sources of mercury emissions, including coal-fired
commercial and industrial boilers, chlor-alkali plants, and Portland cement plants, are
subject to regulations promulgated under this authority.
Separately, Section 129 of the Clean Air Act requires emission standards for
solid waste incinerator units, including municipal and medical waste incinerators.
These standards, which were promulgated in the mid-1990s, limit 11 categories of
pollutants, including mercury. Under the standards, municipal and medical waste
incinerators, which together accounted for 29% of total U.S. mercury emissions
before regulation, have achieved emission reductions of 95%, and together emitted11
only 2.2 tons of mercury in 2000, according to EPA. As a result, coal-fired utilities
— the last major source to be regulated — now account for at least 42% of U.S.12
mercury emissions.
Electric Utilities and Mercury
Statutory Requirements. Electric utilities were singled out for special
consideration by the 1990 Clean Air Act Amendments. Under Section 112(n), EPA
was required to undertake two studies of mercury emissions and other hazardous air
pollutants from electric utility steam generating units, and to report to Congress
before deciding whether to impose MACT standards. One study was to characterize
emissions from utilities, municipal waste incinerators, and other sources, determine
their health and environmental effects, identify the technologies available to control
emissions, and estimate the costs of such technologies. The other study was to

11 “Major Reductions in Toxics, Metals Seen from Controls on Incinerators, EPA Says,”
Daily Environment Report, June 25, 2002, p. A-3.
12 The 42% estimate comes from Table VI-2 of “Revision of December 2000 Regulatory
Finding on the Emissions of Hazardous air Pollutants from Electric Utility Steam Generating
Units and the Removal of Coal- and Oil-fired Electric Utility Steam Generating Units from
the Section 112(c) List,” 70 Federal Register 16018, March 29, 2005 (hereafter cited as
“Revision of Regulatory Finding”). It refers to emissions in 2001. In the agency’s Toxic
Release Inventory (TRI) database [http://www,], electric utilities
accounted for 65% of total air emissions of mercury and mercury compounds in 2003 (latest
available year as of January 2006). The TRI database somewhat overstates the utility share
of the total because it excludes waste incineration and all sources that emit less than 10
pounds of mercury.

determine the hazards to public health anticipated as a result of emissions of all
hazardous air pollutants emitted by electric utilities after imposition of other
requirements of the act, and describe “alternative control strategies for emissions
which may warrant regulation under this section.” After considering the results of
this study, “the Administrator shall regulate electric utility steam generating units
under this section [Section 112], if the Administrator finds such regulation is
appropriate and necessary....”
Having submitted the required reports to Congress under this section in 1997
and 1998,13 EPA Administrator Carol Browner did find such regulation appropriate
and necessary, and issued a formal finding to that effect in December 2000.14 The
finding set in motion the development of MACT standards. The standards were to
be proposed by December 15, 2003, and a final MACT rule was to be signed by
March 15, 2005. MACT standards require compliance for existing facilities three
years after promulgation, with the possibility of one-year extensions.15
EPA’s Utility Regulations. In the March 15, 2005 final mercury rules, EPA
did not promulgate a MACT standard. Instead, it revisited the December 2000
analysis and reversed its regulatory finding. In its revised analysis, regulating
mercury from utilities under Section 112 is neither “appropriate” nor “necessary.”
It is not appropriate, the agency concluded, because the health effects of mercury
from utilities remaining after imposition of other controls “do not result in hazards16
to public health.” It is not necessary, in the agency’s reasoning, because mercury
could be regulated under other sections of the act, specifically Section 111.17 Thus,
the agency proposed a New Source Performance Standard (NSPS) for mercury from
utilities under Section 111(b), and a cap-and-trade system to control emissions from
existing and new utility sources under Section 111(d).
Of the two Section 111 rules, attention has focused primarily on the cap-and-
trade system under Section 111(d), rather than the NSPS. Very few new coal-fired
units have been built in recent years, and only a handful of such facilities are

13 U.S. EPA, Office of Air Quality Planning and Standards, Study of Hazardous Air
Pollutant Emissions from Electric Utility Steam Generating Units — Final Report to
Congress, February 1998, 2 vols., available at [
utoxpg.html#TEC]; and U.S. EPA, OAQPS and Office of Research and Development,
Mercury Study Report to Congress, December 1997, 8 vols., available at [http://www.epa.
gov/ttnatw01/112nmerc/mercury.html ].
14 Regulatory Finding on the Emissions of Hazardous Air Pollutants from Electric Utility
Steam Generating Units, 65 Federal Register 79825, December 20, 2000.
15 These dates on which regulations were to be proposed and promulgated were fixed in a
modified settlement agreement filed November 17, 1998. The case is Natural Resources
Defense Council, Inc. v. U.S. EPA, No. 92-1415 (D.C. Cir.). Originally, the deadline for
promulgation was December 15, 2004. In late April 2004, NRDC offered to extend the
deadline 90 days in order to allow for additional analysis of regulatory options. EPA
accepted the offer.
16 U.S. EPA, Revision of Regulatory Finding, 70 Federal Register 16004, March 29, 2005.
The full discussion begins on p. 16002.
17 Ibid., p. 16005.

currently moving through the permitting process. While some think the rise in
natural gas prices will spark construction of a new wave of coal-fired plants, the
Department of Energy projects essentially no change in coal-fired electric generation
between now and 2015.18 Thus, it is expected that the NSPS will affect far fewer
units than the cap-and-trade rules. Furthermore, under the promulgated cap-and-trade
system, new units are not given any emission allowances: for each ounce of mercury
that they emit, they would need to purchase an allowance from an existing unit that
has controlled emissions more or sooner than required. Thus, in the overall scheme
of things, new units would not increase total emissions of mercury above the
nationwide cap, no matter how weak or strong the NSPS might be.
The cap-and-trade system, on the other hand, has received significant attention,
much of it highly critical. Some of the criticism concerns the choice of Section
111(d) as the legislative authority for mercury regulation. Section 111(d) has rarely
been used until now, and has never been used to regulate a hazardous air pollutant
listed under Section 112. EPA staff say that it has previously been used to regulate
sulfur emissions from pulp and paper mills and fluoride emissions from aluminum
smelters, neither of which are controlled elsewhere in the act.19 EPA’s March 29,
2005 Federal Register notice contains a long discussion of the legality of the use of
Section 111, as opposed to the MACT authority in Section 112. The legality of this
approach was immediately challenged in a suit filed by New Jersey and 15 other

18 See Energy Information Administration, U.S. DOE, Annual Energy Outlook 2005,
February 2005, Table 9, at [].
19 Sulfur and fluoride are not listed as hazardous air pollutants under Section 112(b) and are
not criteria pollutants listed under Section 108(a).
20 The case is New Jersey v. EPA, No.05-1097 (D.C. Cir.) filed Mar.29, 2005. The original
plaintiffs were New Jersey, California, Connecticut, Maine, Massachusetts, New
Hampshire, New Mexico, New York, and Vermont. Seven Eastern and Midwestern states
(Delaware, Illinois, Michigan, Minnesota, Pennsylvania, Rhode Island, and Wisconsin) have
since joined them in challenging the rule, while seven mostly Plains states (Alabama,
Kansas, Nebraska, North Dakota, South Dakota, and Wyoming) have intervened on EPA’s
side. For a discussion of the legal issues, see CRS Report RL32203, Legal Analysis and
Background on the EPA’s Proposed Rules for Regulating Mercury Emissions from Electric
Utilities, by Michael John Garcia. Interestingly, neither Section 111 nor Section 112
actually mention cap-and-trade programs. Section 111 requires “standards of performance,”
defined as a standard that “reflects the degree of emission limitation achievable through the
application of the best system of emission reduction [emphasis added] which ... the
Administrator determines has been adequately demonstrated.” EPA states that this language
allows a cap-and-trade system. Section 112(d) also uses broad language, referring to
“measures, processes, methods, systems, or techniques,” but in elaborating on this definition
for nearly a page, the statute provides numerous examples and specifics, without mentioning
cap-and-trade systems.

The specifics of the cap-and-trade system — its timing and its level of
stringency — are also controversial.21 The system will cap nationwide emissions
mercury from coal-fired electric generating units at 38 tons beginning in 2010 and
at 15 tons beginning in 2018. If achieved, the caps would provide a 21% reduction
from the 1999 level of emissions in Phase 1, and a 70% reduction in Phase 2.
Table 2. Mercury Caps and Projected Emissions Under the
Mercury Cap-and-Trade Rule
Y e ar 2010 2015 2018 2020
Cap38 tons38 tons15 tons15 tons
Actual31.3 tons27.9 tonsn.a.24.3 tons
Source: U.S. EPA.
The caps will be implemented through an allowance system similar to that used
in the acid rain program, through which utilities can either control the pollutant
directly, or purchase excess allowances from other plants that have controlled more
stringently or sooner than required. Early reductions can be banked for later use, if
they are not sold. The agency says the banking provision will provide incentives for
early reductions. As shown in Table 2, it projects emissions of 31.3 tons in 2010,
nearly 7 tons less than the cap.22 If this happens, it will allow utilities to delay full
compliance with the 70% reduction until well beyond 2018, as they use up banked
allowances rather than install further controls. The agency’s analysis projects actual
emissions to be 24.3 tons as late as 2020 (less than a 50% reduction compared to
baseline 1999 emissions).23 It appears that full compliance with the 70% reduction
might be delayed until 2030.24

21 A variety of views have been expressed by interested parties regarding EPA’s mercury
rule. For additional views on the rule, the reader may wish to read Charles River Associates
study for the Edison Electric Institute (EEI), Projected Mercury Emissions and Costs of
EPA’s Proposed Rules for Controlling Utility Sector Mercury Emissions, available from
EEI. For environmental views, the reader may wish to visit the website of Clear the Air at
[] or see the statement of the Natural
Resources Defense Council at [].
22 U.S. EPA, Office of Air Quality Planning and Standards, Regulatory Impact Analysis of
the Clean Air Mercury Rule, March 2005, Table 7-3, p. 7-5, available at [http://www.epa.
23 Ibid.
24 EPA has not provided an estimate of the year in which the 70% reduction will be attained.
The Integrated Planning Model, which the agency uses to calculate regulatory impacts, runs
to the year 2030 and assumes that all allowances will be used by the end date. Discussions

Costs and Benefits of the Final Rule. While acknowledging that it is
unable to estimate or express in dollars at least seven categories of potential benefits,25
EPA’s estimate of the direct benefits of the final cap-and-trade rule is close to zero.
The one direct benefit for which the agency made an estimate was the benefit from
avoided IQ decrements in children prenatally exposed to lower levels of
methylmercury, the value of which was estimated at $400,000 to $3.0 million
annually. In calculating this amount, the agency looked only at exposures due to the
mother’s ingestion of fish from fresh water recreational fishing.26
In addition, the agency found that the controls installed to reduce mercury would
slightly reduce emissions of fine particulate matter (PM2.5), saving up to 7 lives
annually. These co-benefits were estimated at $1.4 million to $40 million per year.27
The costs of the mercury rule were estimated to range from $157 million28

annually to $896 million, depending on whether a 3% or 7% discount rate was used.
24 (...continued)
we held with EPA staff indicate that some think the allowances will be used more quickly
(perhaps as early as 2025), while others think use of allowances will be stretched into the
2030s. The level at which the Phase 1 cap is set is a key determinant of when the rule
reaches its ultimate stringency. The final rule’s 38-ton Phase 1 cap generates 70.5 tons of
early reduction allowances by 2018 in EPA’s analysis. A cap that matched EPA’s estimate
of actual emissions in 2010 would generate only about 17 tons of allowances by 2018 if the
timing of industry compliance was unchanged. In general, the larger the number of
allowances generated by early reductions the later full compliance will be achieved.
25 Potential benefits that were not estimated included cardiovascular effects, ecosystem
effects, genotoxic, immunotoxic, reproductive, renal, and hematological effects. While
studies have suggested associations between exposure to methylmercury and each of these
effects, EPA found the evidence either inconclusive or insufficient to evaluate the impacts.
See U.S. EPA, “Standards of Performance for New and Existing Stationary Sources: Electric
Utility Steam Generating Units,” Final Rule, Preamble, pp. 186-187, available at
[] (hereafter, “Preamble to the
Final Rule”). In its analysis of the proposed rule, EPA listed 11 health and welfare benefits
of controlling mercury: reductions in neurological disorders, learning disabilities, and
developmental delays; impacts on birds and mammals, such as reproductive effects; impacts
on commercial, subsistence, and recreational fishing; and reduced “existence values” for
currently healthy ecosystems. It also listed as potential mercury control benefits reductions
in cardiovascular effects, altered blood pressure regulation, and reproductive effects in
humans. None of these benefits were quantified, but the agency stated at the time that it
believed that they “are large enough to justify substantial investment in Hg emission
reductions.” U.S. EPA, Proposed National Emission Standards for Hazardous Air
Pollutants; and, in the Alternative, Proposed Standards of Performance for New and
Existing Stationary Sources: Electric Utility Steam Generating Units (hereafter cited as the
Mercury Proposal), Preamble, Table 9, 69 Federal Register 4711, 4708, January 30, 2004.
26 U.S. EPA, Preamble to the Final Rule, previously cited, pp. 182-188.
27 Ibid., p. 187. A more thorough discussion of the co-benefits of particulate matter control
is found in Section 12 of the Regulatory Impact Analysis, at [
atw/ utility/ria_final.pdf].
28 The higher discount rate would imply greater uncertainty regarding future cost rather than

Cost factors estimated by EPA included capital investments in pollution control,
operating expenses of the pollution controls, and additional fuel expenditures, but not
the cost of monitoring, reporting, and recordkeeping.29
Several observations can be made regarding these cost and benefit estimates.
First, according to the rule’s preamble, “EPA modeling assumes no improvements
in the cost or effectiveness of Hg [mercury] control technology over time”; in the
next sentence, the agency notes, “In reality, by 2020, costs of Hg control are expected
to have declined.”30 The costs used in the modeling runs were estimated in mid-
2003, according to EPA. Since that time, mercury control costs have declined as
much as 75%, according to the Institute of Clean Air Companies (ICAC), the trade
association that represents manufacturers of pollution control equipment.31 If that is
the case, mercury control costs could be overestimated by as much as a factor of 4.
The discrepancy between ICAC’s estimate and EPA’s may be larger than
fourfold, however. At a January 2005 congressional briefing, ICAC and one of the
technology providers estimated current mercury control costs as low as $2,000 per
pound.32 EPA’s modeling runs estimated the marginal cost to be $39,000 per pound
in 2020.33
Second, while the estimated costs of the rule ($157 million to $896 million
annually in 2020) appear substantial in EPA’s cost-benefit analysis, there is little
additional detail provided as to what these costs represent. The analysis concludes
that almost no pollution control equipment would be installed specifically to control
mercury. As shown in Table 3, in 2010, only 2 gigawatts (less than 1% of coal-fired
capacity) is projected to have installed activated carbon injection (ACI) under the34
rule. An additional 1 gigawatt (Gw) is projected to add selective catalytic reduction
(SCR) to control mercury. (Total coal-fired electric capacity is currently 305 Gw,
and is expected to remain at that level in 2010.) By 2015, there is little change. By

28 (...continued)
the imposition of additional controls.
29 EPA estimates these costs elsewhere in its analysis at “less than $76 million” annually.
Preamble to the Final Rule, p. 190.
30 Ibid., pp. 198-199.
31 Institute of Clean Air Companies (ICAC), “Advances in Mercury Control Technology to
Meet Future Needs,” Congressional Briefing, January 31, 2005.
32 Ibid. Followup discussions with ICAC, one of its member companies, and EPA staff
produced a range of estimates regarding cost; but all three sources agreed that the cost of
mercury removal has been reduced substantially since EPA’s 2003 baseline. Personal
communications, week of April 11, 2005
33 See Regulatory Impact Analysis, previously cited, Table 7-8, p. 7-7. EPA’s modeling
results are expressed in 1999 dollars.
34 ACI is the only technology specifically designed for mercury control in EPA’s model.

2020, mercury controls would include 13 gigawatts of ACI (about 4% of coal-fired
units), 3 Gw of SCR, and 1 Gw of scrubbers.35
The reason that so few units would install ACI in EPA’s analysis is that the
agency designed the rule to complement its simultaneous promulgation of the Clean
Air Interstate Rule (CAIR), a cap-and-trade program for utility emissions of sulfur
dioxide (SO2) and nitrogen oxides (NOx). Controlling SO2, NOx, and mercury
simultaneously reduces costs by allowing utilities to maximize “co-benefits” of
emission controls. Controls such as scrubbers and fabric filters, both of which are
widely used today to control SO2 and particulates, have the side effect (or co-benefit)
of reducing mercury emissions to some extent. NOx control technology (known as
selective catalytic reduction, or SCR) tends to oxidize elemental mercury, making it
easier for scrubbers and particulate controls to remove. Under EPA’s cap-and-trade
regulations, both the 2010 and 2018 mercury emission standards are set to maximize
use of these co-benefits. Thus, few controls would be required to specifically address
mercury emissions under the rule, and the costs specific to controlling mercury would
be low.
Table 3. EPA Projections of Installed Pollution Controls at Coal-
Fired Electric Generating Units, 2010-2020, in gigawatts (Gw)
level146125 — 177151 — 1981530.5
due to Hg — 122231312.5
Source: U.S. EPA, Regulatory Impact Analysis, Table 7-9. Total coal-fired electric generating
capacity, according to U.S. DOE, was 305.2 Gw in 2003, and is projected to be 304.6 Gw in 2010,
310.6 Gw in 2015, and 334.6 Gw in 2020.
No tes:
FGD = Flue Gas Desulfurization (scrubbers, principally intended for SO2 control);
SCR = Selective Catalytic Reduction (principally intended for NOx control);
ACI = Activated Carbon Injection (principally intended for Hg control).
Base Case equals continuation of control programs in place as of March 2004.
CAIR = Clean Air Interstate Rule signed March 10, 2005.
Hg rule = Clean Air Mercury Rule, signed March 15, 2005.
There are also two observations regarding the benefit side of the analysis. One
notable feature is the contrast between EPA’s final cost-benefit analysis and the
agency’s analysis of its proposed rule, which appeared in the Federal Register a year

35 FGD scrubbers and SCR are designed primarily to reduce SO2 and NOx, but have the co-
benefit of reducing mercury.

earlier.36 In the proposed rule, calculations of the overall costs and benefits supported
the imposition of a more stringent standard. The proposed MACT rule would have
taken effect two years earlier than the cap-and-trade rule, and allowed emissions of
34 tons annually. The agency projected compliance costs of the MACT proposal at
$945 million per year. Quantifiable benefits of MACT (essentially longer lives and
less illness from the reductions in fine particles achieved as a co-benefit) were
estimated at more than $15 billion annually (a 16 to 1 advantage over costs). If costs
to other sectors of society were added in, the benefit-cost ratio was still 9 to 1.
As noted earlier, none of the $15 billion in benefits the agency quantified in its
proposal was a direct benefit of mercury control. They were all benefits that resulted
from reduced particulate emissions. In addition to these co-benefits, however, the
agency concluded that the unquantifiable direct benefits of controlling mercury “are
large enough to justify substantial investment in Hg emissions reductions.”37
Table 4. EPA Estimates of the Costs and Benefits of MACT vs.
Cap and Trade for Utility Mercury Controls
Annual benefits
AnnualRatio of
Regulation/ef f e cti Em issions compliance benef i t sCo-
ve datecapcost to costsDirectbenefit
MACT200834 tons$945 millionn.q.>$1516:1
Cap and201038 tons$160 millionn.q.n.q. —
(phase 1)
Cap and201815 tons$750 million$0.4-$1.4 -0.0024 -
Trade $3.0 $40 0.0573:1
(phase 2)millionmillion
Source: U.S. EPA.
Note: For both the MACT and the cap-and-trade rules, there are varying estimates of cost in EPAs
Regulatory Impact Analyses and in the Preamble to the rules, depending on whether one chooses the
industrys compliance cost or the cost to society at large, and depending on the discount rate used to
adjust future costs and benefits. To facilitate comparisons, this table relies on compliance cost rather
than social cost. The range of cap-and-trade benefits includes both direct and indirect (PM2.5-related)
b e ne fits.
n.q. = not quantified.
As shown in Table 4, the analysis accompanying the final rule presents results
far different from the earlier analysis: whereas the earlier analysis showed benefits
far outweighing costs, the final analysis concludes that the costs of mercury control
outweigh the benefits by 17 to 1. The primary change appears to be a reassignment

36 U.S. EPA, Mercury Proposal, Preamble, 69 Federal Register 4651, January 30, 2004.
37 Ibid., p. 4711.

of the $15 billion in particulate matter co-benefits to the CAIR rule.38 By making
implementation of mercury controls simultaneous with CAIR, the co-benefits are
attributed to CAIR, instead of to the mercury rule. EPA’s analysis of CAIR now
shows benefits of $85-$100 billion annually by 2015, more than 25 times its cost.
Some of this change is simply a paper exercise: the co-benefits are taken from
one rule and given to another. But there is also a real difference in the agency’s
position regarding direct benefits. Rather than maintaining that the benefits of
mercury control — although unquantifiable — are large enough to justify substantial
investments in mercury control, the agency now concludes that even a complete
elimination of the mercury emissions from U.S. power plants would have little
impact on human health.39
In assessing this latter conclusion, though, we arrive at a second key observation
regarding the benefits analysis: EPA’s assessment did not include two peer-reviewed
studies that indicated stricter utility mercury rules would have yielded large benefits.
One study, by the Mount Sinai Center for Children’s Health and the Environment,
published in the National Institutes of Health’s Environmental Health Perspectives
February 28, 2005, concluded that between 316,588 and 637,233 newborn children
each year have blood mercury levels associated with loss of IQ. The study
The resulting loss of intelligence causes diminished economic productivity that
persists over the entire lifetime of these children. This lost productivity is the
major cost of methylmercury toxicity, and it amounts to $8.7 billion annually
(range: $2.2-$43.8 billion, 2000 dollars). Of this total, $1.3 billion (range:
$0.1-$6.5 billion) each year is attributable to mercury emissions from American40
power plants.
This report may not have been available to EPA in time to be considered in
developing the final rule, but a second study, funded by EPA and submitted to the
agency February 22, 2005, reportedly was available to the agency, at least in

38 If MACT were imposed in 2008, as would have been required by statute had the agency
implemented a MACT approach, the co-benefits of installing scrubbers and SCR (reductions
of hundreds of thousands of tons of SO2 and NOx) would be attributed to the MACT rule.
By moving implementation to 2010 in the final rule, the co-benefits estimated to result from
MACT (reductions of 2,200 premature deaths annually, 2,900 non-fatal heart attacks,
thousands of hospital and emergency room visits, and millions of work loss and restricted
activity days) are no longer attributed to mercury regulations.
39 “EPA, in its expert judgment, concludes that utility Hg emissions do not pose hazards to
public health.” U.S. EPA, Revision of Regulatory Finding, 70 Federal Register 16025,
March 29, 2005.
40 Leonardo Trasande, Philip J. Landrigan, Clyde Schechter, “Public Health and Economic
Consequences of Methylmercury Toxicity to the Developing Brain,” Environmental Health
Perspectives, posted online February 28, 2005, p. 6, available at [
me mbers/2005/7743/7743.pdf].

preliminary form, before the agency’s deadline for inclusion in the final analysis.41
This study, conducted by the Harvard Center for Risk Analysis, concluded that the
benefits of reducing power plant mercury emissions to 15 tons per year range from
$119 million annually if persistent IQ deficits from fetal exposures to methylmercury
are counted, to as much as $5.2 billion annually if both IQ deficits and cardiovascular
effects and premature mortality are estimated.42 The Harvard report was not
referenced in the agency’s analysis of the final rule. The agency maintains that since
the final report on the study was not submitted until after a January 3, 2005 agency
deadline, it was ineligible to be considered.
Availability of Technology. Critics of EPA’s mercury rule argue that the
regulations should be more stringent or implemented more quickly. To a large
extent, these arguments and EPA’s counterarguments rest on assumptions concerning
the availability of control technologies. Specifically, EPA maintains that the
technology required for mercury control will not be adequately demonstrated until
after 2010,43 and that the technologies for SO2 and NOx — while available now —
cannot be implemented at a faster pace without causing “extremely high” costs and
overwhelming the capacity of equipment suppliers.44
Analysis by other experts came to a different conclusion. For example, a late
2003 paper co-authored by representatives of two power companies, the Electric
Power Research Institute, the U.S. Department of Energy, and ADA-ES, a leading
consultant on advanced mercury control technologies, concluded:
Recent full-scale field tests have proven the effectiveness of activated carbon
injection for reducing mercury emissions. This technology is ideally suited for
use on existing coal-fired boilers as it provides the following advantages:
!Minimal capital cost of equipment (<$3/kW);
!Can be retrofit with little or no downtime of the operating unit;
!Effective for both bituminous and subbituminous coals;
!Can achieve 90% removal when used with a fabric filter that has been
designed properly for carbon injection; and

41 See “EPA Draws Fire for Ignoring Harvard Study of Benefits of Reducing Mercury
Pollution,” Daily Environment Report, March 24, 2005, p. A-9.
42 Glenn Rice and James K. Hammitt, Harvard Center for Risk Analysis, “Economic
Valuation of Human Health Benefits of Controlling Mercury Emissions from U.S.
Coal-Fired Power Plants,” report for Northeast States for Coordinated Air Use Management,
February 2005, [],
pp. xvi-xix. This report appears to have two key differences from EPA’s assessment of
benefits. First, it estimated cardiovascular effects and mortality. Second, it looked at
impacts of mercury emissions on coastal Atlantic, Gulf of Mexico, and global fish, in
addition to non-commercial fresh water fish.
43 Mercury Proposal, Section IV.D.2., p. 4698.
44 Ibid.

!It can be integrated to enhance mercury capture with virtually every
configuration of air pollution control equipment including ESPs45
[electrostatic precipitators], fabric filters, wet and dry scrubbers.
The agency’s own Office of Research and Development (ORD), in a white paper
posted on the EPA website March 2, 2004, appears to conclude that technology is
more available and more effective than is maintained in the agency’s rulemaking.
ORD found that fabric filters, a relatively simple technology that is currently installed
on more than 12% of power plants, achieve a 90% reduction in mercury emissions
at bituminous coal plants and a 72% reduction at subbituminous plants. The addition
of a scrubber increased the emission reduction to 98% at bituminous plants,
according to ORD.46
The white paper further stated that, by 2010, activated carbon injection with a
fabric filter “has the potential to achieve 90% Hg reduction” on any rank of coal, and
could be installed within 1-2 years of signing a contract to do so.47 Since the white
paper was written, there have been reports that a European firm, Donau Carbon, has
begun offering commercial guarantees for mercury removal from coal-fired power
plants using ACI technology.48
ACI has also been used successfully, for the past decade, to control mercury
emissions from municipal waste incinerators. According to EPA’s Office of Air
Quality Planning and Standards, emissions of mercury from these sources were
reduced from 45.2 tons in 1990 to 2.2 tons in 2000, and, “The performance of the
MACT retrofits has been outstanding.”49 While the concentration of mercury in
incinerator emissions is higher than that in coal-fired power plants, requiring some
adjustments in operating methods, the basic ACI technology has proven itself to be
reliable and effective through a decade of experience.
Hot Spots. One of the main criticisms of the cap-and-trade approach is that
it would not address “hot spots,” areas where mercury emissions and/or
concentrations in water bodies are greater than elsewhere. EPA has developed data
on such hot spots: Environmental Defense released a report in December 2003 based

45 Michael Durham, et al., “Full-Scale Results of Mercury Control by Injecting Activated
Carbon Upstream of ESPs and Fabric Filters,” paper presented at PowerGen 2003, Las
Vegas, NV, December 9-11, 2003, p. 19.
46 U.S. EPA, Office of Research and Development, “Control of Mercury Emissions from
Coal-Fired Electric Utility Boilers,” undated, posted March 2, 2004, available at [http://
www.epa.go v/ttn/atw/utility/hgwhitepaperfinal.pdf].
47 Ibid., pp. 13-15.
48 Personal communication, U.S. EPA, Office of Air and Radiation, May 21, 2004.
49 “Emission from Large MWC Units at MACT Compliance,” memorandum from Walt
Stevenson, Combustion Group, Office of Air Quality Planning and Standards, U.S. EPA,
to Docket A-90-45, June 20, 2002.

on EPA’s data that concluded: “At hot spots, local sources within a state commonly
account for 50% to 80% of the mercury deposition.”50
That the local contribution to hot spot concentrations is this high is disputed by
utility sources, particularly for mercury emitted by power plants. Utility
spokespersons maintain that much of the mercury emitted by utilities is in the
elemental form, is non-water soluble, and is released from taller stacks. The result,
they say, is that it is less available to fish, and disperses over a wider area — with
much of it entering a global mercury cycle.51 EPA has adopted this point of view.
In supporting materials that accompanied the release of the final mercury rule, the
agency noted that only 1% of global mercury emissions comes from U.S. power
plants.52 EPA modeling in support of the final rule concluded that even if power
plant emissions were reduced to zero, there would be little impact on methylmercury
concentrations in U.S. lakes and rivers, and little consequent reduction in the
concentrations of mercury in fish-eating women of child-bearing age.53
The concern over hot spots, and the impetus to address them, were reinforced
in late 2003 by a study of mercury contamination in the Everglades. The study found
that concentrations of mercury in fish and wading birds in the area dropped around
75% after Florida imposed stringent controls on incinerators and other local sources
of mercury emissions in the 1990s.54 Backers of strong controls on utility emissions
have cited these results in arguing against a cap-and-trade approach. Cap-and-trade
programs are not, in principle, well designed to address hot spots, they argue. The
cap-and-trade approach allows facilities to purchase allowances and avoid any
emission controls, if that is the compliance approach that makes the most sense to a
plant’s owners and operators. Opponents of the rule maintain that if plants near hot
spots purchase allowances rather than install controls, the cap-and-trade system may
not have an impact on mercury concentrations at the most contaminated sites. By
contrast, a MACT standard or a very stringent cap would require reductions at all
plants, and would, therefore, be expected to improve conditions at hot spots.
EPA’s analysis of the cap-and-trade rule has addressed this issue.55 In the
preamble to the proposed rule, it noted that all states would remain free to establish
more stringent controls to address local health-based concerns separate from the

50 Environmental Defense, Out of Control and Close to Home, December 2003, p. 12.
51 See Electric Power Research Institute written statement, as quoted in “Backers of Utility
Rules Expect Florida Study of Effect of Mercury to Affect EPA Decisions,” Bureau of
National Affairs, Daily Environment Report, November 19, 2003, p. A-10. Also see EPRI’s
press statement, “Power Plants and Mercury,” available at [
attachme nts/296982_1011415_Power_Plants_and_Mercury.pdf].
52 [].
53 Revision of Regulatory Finding, 70 Federal Register 16029, March 29, 2005.
54 Florida Department of Environmental Protection, Integrating Atmospheric Mercury
Deposition With Aquatic Cycling in South Florida, revised November 2003, available at
[], visited January 11,

2006. See especially, pp. 56-59.

55 Mercury Proposal, 69 Federal Register 4702-4703.

mercury cap-and-trade program requirements. But it went on to state that the agency
does not anticipate hot spots, for two reasons. First, EPA’s modeling suggests that
larger coal-fired units, which have the highest “local deposition footprints,” are likely
to control emissions more than required and sell excess allowances achieved through
overcompliance to smaller units. Second, mercury emissions come in several forms.
The most difficult to control is elemental mercury, according to the agency, and it is
the most likely to be transported long distances from the generating units. Thus, if
plants focus on the more easily controlled forms of mercury, they will control
mercury that would more likely be deposited locally. Finally, the agency raised the
possibility that it could adjust the trading program to favor controls at units in
sensitive areas.56
Effects on Eastern and Western Coal. Whether imposition of controls
on mercury will affect the total amount and/or the types of coal consumed at the
nation’s power plants has been an important issue in the debate over power plant
mercury controls. The United Mine Workers of America, for example, were critical
of EPA’s proposed MACT on the grounds that the standards could be met by a
majority of Western subbituminous coals without the need for any emission control
technologies, whereas Eastern bituminous coals, representing roughly one-half of
domestic coal production, would have needed to meet an average emission removal57
rate of 75%. The proposed MACT rule’s costs and reductions in emissions came
almost entirely from controls on bituminous units, according to EPA.
Table 5. Estimated Changes in Coal Use from Imposition of the
Mercury (Hg) Cap-and-Trade Rule, 2003-2020, by Region
(in million tons)
Coal SupplyCAIR +CAIR + Hg
RegionBaseHg RuleBaseRule
Appalachia275 325 303 301 330
Interior135 161 169 173 224
Western526 603 589 714 572
National Total9361,0891,0611,1881,127
Source: U.S. EPA, Regulatory Impact Analysis, Table 7-12.
The cap-and-trade rule, because it sets a national cap on emissions rather than
imposing controls on specific units and coal types, appears more favorable to Eastern
and Interior coals than the MACT rule would have been. Eastern coal interests,
unions, and states continue to express concern, however, because of the way the rule

56 Ibid., p. 4701.
57 Comments of Cecil E. Roberts on behalf of the United Mine Workers of America to EPA
Docket ID No. OAR-2002-0056, April 30, 2004, p. 1.

will allocate allowances. Reducing an ounce of mercury from bituminous coal would
generate one allowance under the cap-and-trade mercury rule, whereas equivalent
reductions from subbituminous coal will generate 1.25 allowances. Reductions from
lignite-powered facilities will generate 3 allowances per ounce. The agency’s
rationale is that it is more difficult to remove mercury from lignite and
subbituminous coal emissions.58
Even with this allowance scheme, EPA projects greater growth in use of Eastern
and Interior versus Western coal under the rules. As shown in Table 5, EPA projects
20% growth in utility coal use overall, between 2003 and 2020, with imposition of
the CAIR and mercury rules, about 5% less than would have occurred without the
regulations. By region,59 Interior coal use is projected to grow 66%, Appalachian60

20%, and Western 9%, under the CAIR and mercury rules, according to EPA.

The Regulatory Process. The utility mercury rule is among the most
controversial rules in EPA’s 35-year history, not only because of its substance, but
also because of claimed irregularities in the regulatory process.
Shortly after the proposed rule was published early in 2004, the Clean Air Task
Force, a Boston-based environmental group, reported that significant portions of the
proposal had been copied verbatim from memos written by Latham & Watkins, a law
firm representing electric utilities. Both Jeffery Holmstead, the EPA Assistant
Administrator for Air, and his chief counsel, Bill Wehrum, had worked at Latham &
Watkins before joining EPA in 2001. Holmstead reportedly stated that he was
unaware how the memos had found their way into the regulatory proposal.61
Both the EPA Inspector General (IG) and the Government Accountability Office
(GAO) have released reports critical of the regulatory process. The IG’s report
concluded that the development of the rule was “not consistent with expected and

58 EPA’s rationale is discussed in U.S. EPA, “Standards of Performance for New and
Existing Stationary Sources: Electric Utility Steam Generating Units,” Final Rule, Preamble,
available at [], pp. 84-87.
59 The Integrated Planning Model, which EPA uses to assess the impact of proposed
regulations on utilities and the coal industry, divides the coal-producing states into several
geographic regions: Appalachian (including Pennsylvania, Ohio, West Virginia, Eastern
Kentucky, Tennessee and Alabama); Interior (including Illinois, Indiana, and Western
Kentucky); and Western (principally the Rocky Mountain states and North Dakota,
including the Powder River basin).
60 In the base case, without CAIR or the mercury rule, EPA projected that Interior coal use
would grow by 28%, Appalachian 9%, and Western 36%. In a sensitivity analysis, EPA
notes that DOE’s Energy Information Administration (EIA) assumes a higher electricity
growth rate and a bigger differential between natural gas and coal prices than does EPA.
With EIA’s assumptions, total coal use by utilities would grow to 1,371 million tons in 2020
even with the CAIR and mercury regulations. Each of the three coal-producing regions
would see at least 20% more coal production under EIA’s assumptions. See EPA
Regulatory Impact Analysis, Table 7-32 and p. 7-23.
61 See “Two Democrats Say Mercury Rulemaking ‘Improperly Influenced’ by Industry
Lobbying,” Daily Environment Report, February 13, 2004, p. A-1.

past practices” — specifically that “EPA’s rule development process did not comply
with certain agency and Executive Order requirements, including not fully analyzing
the cost-benefit of regulatory alternatives and not fully addressing the rule’s impact
on children’s health.”62 The IG was particularly critical of the process used to
establish the proposed mercury MACT limits — a point rendered moot by EPA’s
choice of the cap-and-trade option.63
GAO was critical of the analysis underlying both the MACT and cap-and-trade
proposals. For example, GAO found that “EPA did not document some of its
analysis or adhere to the principles of full disclosure and transparency as directed by
OMB, and it did not provide decision makers or the public with consistent
information on how changes in the proposed level of control would affect its
estimates of net economic benefits for each option.”64
As noted above, the agency did not consider in its cost-benefit analysis two
studies that would have indicated direct benefits of mercury control as much as three
orders of magnitude above those it used in determining that more stringent standards
were unnecessary. One of the unused studies was funded by the agency; an agency
staffer was one of its principal authors, and other EPA staff had peer-reviewed it.65

62 Office of Inspector General, U.S. EPA, Additional Analyses of Mercury Emissions Needed
Before EPA Finalizes Rules for Coal-Fired Electric Utilities, February 3, 2005, “At a
Glance,” at [
pdf]. A fuller description of these conclusions is found on pp. 27-36 of the report.
63 For example, the IG report stated: “Evidence indicates that EPA senior management
instructed EPA staff to develop a Maximum Achievable Control Technology (MACT)
standard for mercury that would result in national emissions of 34 tons annually, instead of
basing the standard on an unbiased determination of what the top performing units were
achieving in practice.... [T]he standard likely underestimates the average amount of mercury
emissions reductions achieved by the top performing 12 percent of utilities, the minimum
level for a MACT standard required by the Clean Air Act.” See ibid., At a Glance, and pp.


64 U.S. GAO, Clean Air Act:Observations on EPA’s Cost-Benefit Analysis of Its Mercury
Control Options, February 2005, Report no. GAO-05-252, p. 4.
65 Economic Valuation of Human Health Benefits of Controlling Mercury Emissions from
U.S. Coal-Fired Power Plants, previously cited. Glenn Rice, one of the report’s two
authors, was on leave from the agency’s Office of Research and Development. The
Acknowledgments page of the report lists two EPA employees among the report’s 8 peer
reviewers, and acknowledges the assistance of 13 other EPA staff members. For additional
detail, see “New EPA Mercury Rule Omits Conflicting Data,” Washington Post, March 22,

2005, p. A1.

Legislation in the 109th Congress66
Although EPA has now finalized standards for electric utility mercury emissions
under Section 111, the Administration has also proposed that Congress amend the
Clean Air Act by passing multi-pollutant legislation for utilities — the Clear Skies
bill. A version of Clear Skies introduced by Senator Inhofe (S. 131) would establish
nationwide cap-and-trade programs for SO2, NOx, and mercury similar to those
established by the CAIR and mercury rules. In addition, it would eliminate or
suspend more than half a dozen specific regulatory programs for electric power plants
including New Source Review, Prevention of Significant Deterioration, New Source
Performance Standards, the NOx SIP call,67 nonattainment area requirements, Best
Available Retrofit Technology, and any mercury MACT. The Administration states
that many of these rules would prove redundant under a multi-pollutant cap-and-trade
regime, and that eliminating or suspending them would create a more efficient
regulatory structure. (For a discussion of the ways in which Clear Skies would
change the Clean Air Act, see CRS Report RL32782, Clear Skies and the Clean Air
Act: What’s the Difference?, by Larry Parker and James E. McCarthy.)
Other multi-pollutant or mercury bills also have been introduced. Most other
legislation would reduce mercury emissions more and faster than Clear Skies.68
Senator Jeffords’ S. 150, for example, would reduce utility mercury emissions to a
total of 5 tons (i.e., about 90%) by 2010. Representative Waxman’s H.R. 1451
would set comparable requirements, a reduction of at least 90% from 1999 levels by

2010. Senator Leahy’s S. 730 also envisions a 90% reduction. Under the Jeffords,

Waxman, and Leahy bills, there would be no allowance trading or banking programs
for mercury.
Two other bills remain under discussion. Senator Carper’s S. 843 (from the
108th Congress) and its 109th Congress House counterpart, Representative Bass’s
H.R. 1873, present a middle ground between Clear Skies and the Jeffords, Waxman,
and Leahy bills. Like Clear Skies, the Carper/Bass bill would establish a tradeable
allowance program to ease compliance. But it would also establish caps at individual
units, and would mandate sharper reductions sooner than Clear Skies — an 80%
reduction in mercury emissions by 2013.

66 This report focuses on mercury emissions to the air and on legislation to address such
emissions. Congress is also considering legislation to reduce the amount of mercury in
products and waste streams. For information on mercury in products and wastes, including
congressional and state actions on the subject, see CRS Report RL31908, Mercury in
Products and Waste: Legislative and Regulatory Activities to Control Mercury, by Linda
G. Luther.
67 The NOx SIP call refers to regulations under which State Implementation Plans in 22
eastern states and the District of Columbia must be revised to control NOx emissions in
order to improve ozone air quality in downwind states.
68 An exception is Representative Sweeney’s H.R. 227, under which the Clean Air Act’s
existing provisions for mercury are essentially restated, with EPA to promulgate regulations
for utility mercury emissions by March 15, 2005.

High concentrations of mercury in aquatic environments, and the resulting
advisories to limit consumption of fish in order to protect human health, have
focused attention on the role of mercury emissions from a variety of industrial
sources. Among these, coal-fired power plants are the largest, accounting for 42%
of total U.S. emissions; and they are the last major category of emission sources for
which EPA has considered regulations. Under a consent agreement, EPA agreed to
propose regulations controlling mercury emissions from this category by December

15, 2003, with promulgation by March 15, 2005.

In meeting this commitment, EPA promulgated cap-and-trade regulations rather
than the Maximum Achievable Control Technology regulations it had previously
found to be “appropriate and necessary.” In doing so, it revised its earlier finding —
a step that 16 states have sued EPA to overturn.
The switch from MACT to cap-and-trade means that the control technologies
installed will be primarily scrubbers and selective catalytic reduction — controls
designed to reduce emissions of SO2 and NOx to meet caps on those emissions under
a simultaneously promulgated rule, the CAIR rule. Mercury will be reduced as a co-
benefit of this other rule. During the next decade, EPA estimates that only 1% of
electric generating units would be likely to install pollution control equipment
specifically designed to capture mercury under these regulations.
The mercury cap-and-trade regulations are highly controversial. Read literally,
they offer a reduction of 70% in mercury emissions by 2018; but the agency’s
accompanying analysis indicates that, due to emissions banking and trading, the full
70% reduction might not be achieved until 2030. The agency projects actual mercury
reductions of 35% by 2010, and about 50% by 2020.
Many observers contend that EPA’s estimates of the benefits of mercury control
are understated. In its final analysis, the agency did not include two peer-reviewed
studies — one of which it funded — that concluded that annual benefits of mercury
control were as high as $1.3 billion or $5.2 billion. At least on paper, in order to
select the cap-and-trade mercury control option, EPA passed over a MACT option
that offered higher benefits for essentially the same cost.
In large part because of rapidly improving technology, the agency’s supporting
documentation relies on estimates of mercury control costs that are 4 to 20 times
higher than current projections by pollution control industry sources. The agency
disputes information suggesting that specific control technology is available now,
arguing that it will not be available until after 2010.
EPA states that the mercury to which Americans are exposed — primarily in
fish consumption — comes overwhelmingly from non-U.S. sources. From its
perspective, a more flexible, cap-and-trade approach has economic advantages,
allowing utilities to maximize the co-benefits of controlling several pollutants
simultaneously. In the rule it proposed a year earlier, however, the agency found that
more stringent, more quickly applied limits using Maximum Achievable Control

Technology would also deliver co-benefits, which would outweigh the costs of
compliance by an order of magnitude.
As promulgated, the utility mercury rule may raise equity concerns: other
combustion sources (municipal waste combustors and medical waste incinerators)
have been required to reduce mercury emissions more than 90% under existing Clean
Air Act authority, with considerably shorter deadlines than those in the cap-and-trade
regulation for utilities. Since similar technologies could be applied to coal-fired
power plants, critics argue that different standards are being applied to different
sources of the same pollutant. The agency has offered no explanation for its less
stringent approach to the regulation of mercury from utility sources, other than its
general desire to limit costs and provide flexibility.69
With the 1990 Clean Air Act amendments, Congress determined that electric
utilities would be treated differently from other sources of hazardous air pollutants,
before regulation. The amendments required that EPA report to Congress before
determining whether regulating utility emissions of these pollutants was appropriate
and necessary. This special treatment for electric power producers was motivated by
a number of factors, including a desire to preserve the use of coal as an energy
option, for both economic and energy security reasons. Whether these concerns
justify the agency’s approach to the utility sector’s mercury emissions, and the
balancing of those concerns against competing health and equity considerations, are
at the core of the continuing debate over mercury issues

69 Section 129, which addresses solid waste combustion controls, specifically does mention
using Section 111 and 111(d) to control incinerator emissions. This would appear to give
the agency a firmer statutory basis to provide flexibility to waste combustors in meeting
mercury reduction targets. Yet the agency imposed MACT-like standards on incinerators,
and achieved greater than 90% reductions in mercury emissions within three years of