w
         OFFICE OF INSPECTOR GENERAL
Evaluation Report
        Additional Analyses of Mercury Emissions
        Needed Before EPA Finalizes Rules for
        Coal-Fired Electric Utilities

        Report No. 2005-P-00003

        Februarys, 2005

-------
Report Contributors:
Rick Beusse
Carolyn Blair
Hilda Canes
Sarah Fabirkiewicz
James Hatfield
Erica Hauck
James Huber
Abbreviations
CAA         Clean Air Act
CAIR        Clean Air Interstate Rule
CHPAC      Children's Health Protection Advisory Committee
EPA         Environmental Protection Agency
FACA       Federal Advisory Committee Act
OIG         Office of Inspector General
ICR         Information Collection Request
IGCC        Integrated Gasification Combined Cycle
IPM         Integrated Planning Model
OMB        Office of Management and Budget
MACT       Maximum Achievable Control Technology
NTEC       National Tribal Environmental Council
NOX         Nitrogen Oxide
OCHP       Office of Children's Health Protection
SO,         Sulfur Dioxide
Cover photo:    Virginia Electric Power Company's coal-fired plant at Mt. Storm, West Virginia.
                Source: http://www.oag.state.ny.us/press/2000/nov/nov!6a_00_pictures.html

-------
                    U.S. Environmental Protection Agency
                    Office of Inspector General

                    At  a  Glance
                                                            2005-P-00003
                                                          February 3,2005
 Why We Did This Review
               Senate .
 Environment and Public
 Works Committee requested
 thai we r evlew BF A,' s
 
-------
   *- '^'^
                     UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                                    WASHINGTON, D.C. 20460
                                                                     THE INSPECTOR GENERAL
MEMORANDUM

SUBJECT:



TO:
                                   February 3, 2005
                    Evaluation Report:  Additional Analyses of Mercury Emissions Needed
                    Before EPA Finalizes Rules for Coal-Fired Electric Utilities
                    Report No. 2005-P-00003

                    Jeffrey R. Holmstead
                    Assistant Administrator for Air and Radiation
This memorandum transmits the results of an Office of Inspector General (OIG) evaluation
regarding the Environmental Protection Agency's (EPA) development of the proposed rule for
regulating mercury emissions from coal-fired steam generating electric utility units. This report
contains findings that should help EPA in its efforts to develop the final rule.  Also, the report
contains corrective actions the Office of Inspector General (OIG) recommends. This report
represents the opinion of the OIG and the findings contained in this report do not necessarily
represent the final EPA position. Final determinations on matters in this report will be made by
EPA managers in accordance with established procedures.

Action Required

In accordance with EPA Directive 2750, as the action official, you are required to provide this
Office with a written response within 90 days of the final report date. The response should
address all recommendations. For the corrective actions planned but not completed by the
response date, please describe the actions that are ongoing and provide a timetable for
completion.  Where you disagree with a recommendation, please provide alternative actions for
addressing the findings reported.

We appreciate the efforts of EPA officials and staff in working with us to develop this report. If
you or your staff have any questions regarding this report, please contact me at (202) 566-0847
or Kwai Chan, Assistant Inspector General for Program Evaluation,  at (202) 566-0827.

                                              Nikki L. Tinsley

-------
Attachment
cc:     Steve Johnson, Acting Administrator
       William Farland, Acting Deputy Assistant Administrator for Science, ORD
       Ann Klee, General Counsel
       Pete Cosier, Audit Followup Coordinator, OAR
       Kwai Chan, Assistant Inspector General for Program Evaluation, OIG
       Mark Bialek, Counsel, OIG

-------
                      Table of Contents
At a Glance

 Chapters



   1   Introduction 	   1

            Purpose	   1
            Background	   1
            Scope and Methodology  	   8
            Results in Brief 	   10

   2   Mercury MACT Development Compromised  	   11

            Requirements for MACT Standards	   11
            EPA's Process for Addressing Variability in Computation
                 of Mercury MACT Floor  	  12
            EPA Staff Instructed to Develop MACT Floor That Would Result in
                 National Emissions of 34 Tons	  13
            Conclusions	  16
            Recommendations	  16
            Agency Comments and OIG Evaluation   	16

   3   Cap-and-Trade Option Can Be Strengthened	  18

            EPA's Proposed Cap-and-Trade Approach	  18
            Proposed Cap-and-Trade Program Needs to Further Address
                 Certain Issues  	  19
            Proposed Emissions Trading Rule Should Also Address Tribal Concerns ...  24
            Conclusions	  25
            Recommendations 	  25
            Agency Comments and OIG Evaluation   	26

   4   Rule Development Process Not Consistent with
      Expected and Past Practices  	  27

            Description of Rulemaking Process	  27
            Some FACA Members Considered Job Unfinished	  28
            Intra-Agency Review Limited	  30
            Requirements for Cost-Benefit Analyses Not Fully Implemented	  32
            Required Children's Health Analysis Not Completed	  33
            Scope Limitations: Inter-Agency Review  	  34
            Conclusions  	  35
            Recommendations 	  35
            Agency Comments and OIG Evaluation   	35

-------
Appendices

  A  Timeline of Events Related to Development of Mercury Rule	37

  B  OIG's Request for Documents Related to Development
     of Utility MACT	 39

  C  Different MACT Floor Proposals  	43

  D  EPA's Rule Development Process	44

  E  Agency Comments to the Draft Report and OIG Evaluation	46

  F  Distribution 	 54

-------
                                 Chapter 1
                                 Introduction
Purpose
             The Office of Inspector General (OIG) initiated this review based on a request
             from members of the Senate Environment and Public Works Committee. In their
             written request, the Senators expressed concerns with the process used to develop
             the Environmental Protection Agency's (EPA's) January 2004 proposed rule for
             regulating mercury emissions from coal-fired steam generating electric utility
             units. The proposed rule included two different options for regulating mercury
             emissions. One approach was a Maximum Achievable Control Technology
             (MACT) standard that would establish emission limits applicable to all coal-fired
             utility units.  The other approach was a mercury cap-and-trade approach that
             would establish a national cap on mercury emissions and allow individual utilities
             to trade emissions allowances in a market-based system. The objectives of our
             evaluation were to determine:

             •  Do the data and analyses in the docket demonstrate that the proposed MACT
                option reflects the maximum achievable reductions from coal-fired steam
                generating electric utility units?

             •  Is the mercury cap-and-trade option, as proposed, sufficient to ensure public
                health protection?

             •  What process did EPA follow in developing the proposed rule, and was this
                process consistent with applicable statutes, regulations, policy, guidance, and
                past Agency practice?
Background
             Mercury is released globally into the environment through natural processes, such
             as volcanoes, and also from human activity. Man-made releases of mercury are
             primarily due to the burning of mercury-containing fuels and wastes, and through
             industrial manufacturing processes. Man-made mercury emissions from the
             United States are estimated to account for roughly 3 percent of total global
             mercury emissions. Mercury from lead smelters, municipal waste combustors,
             hospital waste incinerators, manufacturing operations, and other sources are
             largely already regulated by EPA.  In the United States, the largest source of
             airborne mercury emissions is the coal-burning electric utilities industry,
             representing an estimated 40 percent of total U.S. man-made airborne mercury

-------
emissions.  EPA has estimated that one-third of all U.S. emissions of mercury are
deposited within the contiguous United States, while the remaining two-thirds
enter the global cycle. The January 2004 proposal is the first attempt to regulate
mercury emissions from these utilities at the Federal level.

Airborne concentrations of mercury are generally considered to be small and not a
serious health concern while still in the air. However, once mercury enters fresh-
water and salt-water bodies, either directly or through air deposition, it can
bioaccumulate in fish and other animal tissues in its more toxic form,
methylmercury. As mercury bioaccumulates in the food chain, its concentration
becomes increasingly higher in animals at the top of the food chain (such as larger
predatory fish) that consume smaller contaminated organisms.  Because of the
bioaccumulation of methylmercury, the primary route of human exposure to
mercury is through the consumption of fish, both salt water and fresh water.
Excessive human exposure to mercury has been associated with severe
detrimental neurological and developmental health effects.  Depending on the
dose, human health effects from exposure to mercury can include subtle losses of
sensory and cognitive ability, tremors, inability to walk, and death. The
developing fetus may be particularly sensitive to the detrimental effects of
methylmercury, thus, exposure to mercury by women of child-bearing age is of
particular concern.

From a global perspective, mercury accumulation in salt-water fish is a public
health concern. EPA and the Food and Drug Administration have cautioned that
young children, as well as women who might become pregnant, are pregnant, or
are nursing should limit their consumption of certain salt-water predatory fish.
Mercury bioaccumulation in U.S. water bodies is also a public health concern,  and
45 States issued fish advisories for mercury in 2003. Many of these fish
advisories caution that women and young children should limit their consumption
of certain types of fish.

EPA Reference Dose for Methylmercury

Based on studies showing adverse health effects from exposure to methylmercury,
EPA set a reference dose for methylmercury that was designed to protect the most
sensitive subgroup (i.e., developing fetuses). An EPA reference dose reflects the
estimate of daily exposure to the human population, including sensitive
subgroups, that is not likely to cause harmful effects during a lifetime.  The
current EPA reference dose for methylmercury - which was included in EPA's
1997 Mercury Study  Report to Congress - is 0.1 micrograms per kilogram of
body weight per day.

-------
             Subsequent to EPA's 1997 Mercury Study Report to Congress, Congress directed1
             EPA to request the National Academy of Sciences to perform an independent
             study on the toxicological effects of methylmercury and to prepare
             recommendations on the establishment of a scientifically appropriate exposure
             reference dose. The National Academy of Sciences completed its review in 2000,
             and concluded that the EPA reference dose of 0.1 micrograms per kilogram was a
             scientifically justifiable level for the protection of health.

             The most recent results from Centers for Disease Control and Prevention's
             ongoing National Health and Nutrition Examination Survey show that mercury
             blood levels of most children and women of childbearing age were below levels of
             concern corresponding to the EPA reference dose. However, 5.66 percent of
             childbearing-aged woman had blood mercury levels at or above the reference
             dose. The survey also questions participants about their fish consumption. For
             the 1999-2000 survey period, tuna and shrimp were the two most frequently cited
             types of fish/shellfish consumed. These results, and other studies, suggest that
             seafood (as opposed to fresh-water fish) is  the predominant source of mercury
             exposure in the United States.  However, some subpopulations in the United
             States consume more fish, including fresh-water fish, than the general population.
             These groups may be at increased risk from mercury exposure. For example,
             studies have shown elevated blood levels of mercury in some Native American
             tribes that consumed fresh-water fish.

             Statutory Requirements for Controlling Mercury Emissions from
             Utility Plants

             The Clean Air Act (CAA) requires EPA to regulate emissions of 188 air toxics
             (also known as known as hazardous air pollutants), including mercury. EPA was
             to identify and establish emission standards for major source categories emitting
             these pollutants. Specifically, section 112(d) of the CAA requires EPA to
             establish emission limits for major source categories emitting air toxics,
             commonly referred to as MACT standards. The MACT standard is to require the
             maximum degree of reductions achievable for the source category, taking into
             consideration cost and any non-air quality health and environmental impacts.

             A key requirement of section 112(d) is that emission standards for existing
             sources in a category or subcategory shall not be less stringent than the average
             emission limitation achieved by the best performing 12 percent of the existing
             sources for which the Administrator has data.  The emission limitation achieved
             by the best performing 12 percent of sources is referred to as the "MACT floor."
       1 H.R. Rep. No. 769, 105lh Cong., 2d Sess. at 281-282 (1998). This is the Conference Report to accompany
H.R. 4194, Octobers, 1998.

                                           3

-------
              The CAA also established specific requirements with respect to air toxics
              emissions from utilities.  Section 112(n)(l)(A) requires EPA to perform a study of
              the hazards to public health that are reasonably anticipated to occur as a result of
              air toxics emissions from electric utility steam generating units.  This study was to
              develop and describe alternative control strategies for emissions that may warrant
              regulation under section  112.  Further, with respect to regulating emissions from
              utility plants, section 112(n)(l)(A) states:

                  The Administrator shall regulate electric utility steam generating units
                  under this section, if the Administratorfinds such regulation is
                  appropriate and necessary after considering the results of the study
                  required by this subparagraph.

              EPA published its Final Report2 with respect to utilities in February 1998, but
              deferred making a determination as to whether regulation of these units was
              appropriate and necessary. However, the Final Report concluded that:

              •   Mercury from coal-fired utilities was the air pollutant of greatest potential
                  concern to public health from utilities;
              •   Coal-fired utilities are estimated to emit about one-third (51 tons based on
                  1994 emissions) of U.S. anthropogenic (man-made) mercury emissions per
                  year;
                  Ingestion of contaminated fish is the most important route of exposure to
                  mercury; and
              •   Modeling in conjunction with the available scientific data provides evidence
                  for a plausible link between emissions of mercury from utilities and the
                  methylmercury found in soil, water, air, and fish.

              In its Final Report, EPA listed a number of research needs related to mercury
              emissions. These included obtaining additional data on mercury emissions, such
              as the amount emitted from various types of units; the proportion of divalent
              versus elemental mercury;3 and how factors such as the control device, fuel type,
              and plant configuration affect emissions and speciation.
        Study of Hazardous Air Pollutant Emissions from Electric Utility Steam Generating Units - - Final Report
to Congress, EPA-453/R-98-004a, February 1998.

       3 Airborne divalent mercury is adsorbed onto particles or bound to other compounds and is deposited
sooner and mainly in the vicinity of the emissions sources (local to regional distances), while elemental mercury
(vapor) remains airborne longer and is transported on a hemispherical/global scale.
                                             4

-------
              Information Collection Request

              Based on the research needs outlined in the Final Report, the then-EPA
              Administrator concluded that obtaining additional information from
              owner/operators of coal-fired electric utility steam generating units was needed to
              determine whether regulation of electric utility steam generating units was
              appropriate and necessary. Accordingly, EPA used its authority under CAA
              section 114 to collect data from all domestic coal-fired electric utility steam
              generating units.  The resulting information collection request (ICR)  consisted of
              three phases of data collection:

              •   Phase I collected general information on every coal-fired electric generating
                 utility unit and was completed in January 1999.

              •   Phase II consisted of obtaining information on the amount of coal received on
                 a per shipment basis for the 1999 calendar year for every facility.  In addition,
                 the mercury and chlorine content of the coal was reported for every sixth
                 shipment.

              •   Phase in consisted of emissions testing  at 80 units,4 which were selected to
                 represent a cross-section of boiler and control device types.  For each of the
                 80 units selected, testing for mercury was conducted at the inlet and outlet of
                 the last pollution control device on  the unit. Each unit was to conduct three
                 separate test runs and  to also sample and analyze the coal used during each of
                 the three separate runs.

              December 2000 Findings and Determination

              In a December 20, 2000, Federal Register Notice, EPA published its  finding that
              regulation of mercury emissions from coal-fired utility plants was appropriate and
              necessary. The notice described four primary sources of information for the
              finding:

              •   EPA's February 1998  "Study of Hazardous Air Pollutant Emissions from
                 Electric Utility Steam Generating Units — Final Report to Congress."
              •   An ICR to all coal-fired electric utility steam generating units requesting coal
                 data for 1999 and a request to certain units for stack test results to evaluate air
                 toxics emissions.
              •   An evaluation of the mercury control performance of various emission control
                 technologies currently in use to control other pollutants or that could be
                 applied to such units to control mercury emissions.
       4
        Emission tests were actually conducted at 79 different units with 2 tests conducted at 1 unit for a total of
80 tests.
                                            5

-------
              •   An evaluation of available health data related to mercury conducted by the
                 National Academy of Sciences.

              The Notice concluded that, "... during the regulatory development process,
              effective controls for mercury and other HAPs (hazardous air pollutants) can be
              shown to be feasible." The Notice recognized the considerable interest in using
              economic incentive programs, such as emission trading, to achieve emission
              reductions. However, in its December 2000 notice, EPA cited concerns about the
              potential local impact of emissions trading and noted that any trading program
              must be constructed in a way that assured communities nearest a source were
              adequately protected. The Notice stated:

                 Thus, in developing a standard for utilities, the EPA should consider the legal
                 potential for, and the economic effects of, incorporating a trading regime
                 under section 112 in a manner that protects the local populations.

              After issuance of these findings and its determination that regulation of utilities
              was appropriate and necessary, EPA began to develop a MACT standard for
              mercury emissions from coal-fired electric utility units. Additionally, a
              workgroup was established in August 2001  under the Clean Air Act Advisory
              Committee to provide EPA with input regarding Federal MACT regulations for
              coal-fired electric utility steam generating units. Appendix A provides a timeline
              of events associated with the development of the MACT rule.

              Clear Skies Proposal

              Concurrent with EPA's initial efforts to develop a MACT for utility units,
              legislation was proposed in Congress5 to establish a multi-pollutant approach for
              addressing mercury, sulfur dioxide (S02), and nitrogen oxide (NOx) emissions
              from utilities. This legislation, referred to as Clear Skies, proposed a cap-and-
              trade approach to controlling emissions of these three pollutants.  With respect to
              mercury, the initial Clear Skies legislation called for an interim cap on total U.S.
              mercury emissions of 26 tons per year by 2010. Based on modeling done in
              support of the Clear Skies Proposal, EPA estimated that some facilities would
              install mercury-specific technology by 2010 in order to meet the  26-ton cap.
              Clear Skies proposed a final cap of 15 tons  on mercury emissions by 2018, and
              EPA analysis projected that additional sources would choose to install mercury-
              specific controls to meet the cap.

              When Clear Skies legislation stalled in Congress, EPA decided to propose a cap-
              and-trade approach for controlling mercury emissions as an  alternative to a
        Clear Skies was proposed in both the U.S. House of Representatives and Senate in July 2002, and
reintroduced as the Clear Skies Act of 2003 on February 27,2003.
                                            6

-------
              MACT standard. EPA proposed these regulatory alternatives in the January 30,
              2004, Federal Register Notice.  In addition to the proposed mercury rule
              alternatives, EPA on January 30, 2004, also proposed new air rules for reducing
              S02 and NOx emissions. This proposed rule, now known as the Clean Air
              Interstate Rule (CAIR), would establish a cap-and-trade program for 29 States in
              the Eastern United States and the District of Columbia whose S02 and NOx
              emissions significantly contribute to fine particle and ozone pollution problems in
              other downwind States. Together, the CAIR and mercury proposals would create
              a multi-pollutant approach to controlling emissions from utilities similar to what
              was originally proposed in the Clear Skies legislation.

              Proposed Mercury Rule

              As a result of a prior court settlement6, EPA had agreed to issue proposed power
              plant mercury emission standards by December 15, 2003. In the January 30,
              2004, Federal Register Notice, EPA proposed its rule for regulating mercury
              emissions from coal-fired steam generating electric utility units. This proposal
              includes two different approaches for  controlling mercury emissions from utilities:
              a MACT standard or a mercury cap-and-trade program.

              EPA's Proposed MACT Standard Approach:  EPA proposed separate emission
              limits to be achieved by 2008 for five  subcategories: three subcategories for
              different coal types (bituminous, sub-bituminous, and lignite); one for coal refuse
              or waste; and one for a specific type of combustion process known as Integrated
              Gasification Combined Cycle (IGCC).7  Table  1-1 shows the specific per unit
              emissions limits for existing units in the proposed rule.

                 Table 1-1: Proposed MACT Emission Limits
Subcategdi'y
Bituminous
Sub-bituminous
Usntte
Coal-Refuse
JGCC
Eittte&t&t* luttit (IfaAfTBht)* ;
2.0
5.8
9.2
0.38
19.0
              * = pounds per Trillion British thermal units.

              These emission limits were based on what EPA determined to be the MACT
              floor.  EPA proposed that the MACT standard be based on the MACT floor as
        Under a settlement agreement reached in 1998 with the Natural Resources Defense Council, EPA agreed
to issue a proposed rule for regulating mercury from power plants by December 15, 2003, and a final rule by
December 2004. (Natural Resources Defense Council v. EPA.D.C. Cir., No. 92-1415, 4/15/98). Natural Resources
Defense Council later agreed to extend the deadline for the final rule to March 15, 2005.

        The IGCC process converts coal into gas and uses the coal gas as fuel for generating electricity.
                                            7

-------
              opposed to a beyond-the-floor8 level because it concluded that technologies for
              reducing mercury emissions were not commercially available and, thus, beyond-
              the-floor emission standards were not achievable.  EPA estimated that total
              national mercury emissions would be reduced from 48 to 34 tons per year if the
              proposed MACT rule was implemented.

              EPA's Proposed Cap-and-Trade Approach. In lieu of adopting a MACT
              standard to regulate mercury emissions from utilities, EPA presented an
              alternative proposal that would regulate mercury emissions from utility units
              under a national cap-and-trade program implemented under section 111 of the
              CAA.9  The cap-and-trade proposal included an unspecified interim cap on
              mercury emissions in 2010 and a final cap of 15 tons by 2018. Though EPA did
              not specify an interim cap level, the Agency proposed that it be based on the
              maximum amount of mercury reductions that could be achieved through
              implementing the controls necessary to reduce S02 and NOx emissions, i.e., the
              mercury co-benefit of these controls through implementation of CAIR.  The
              preamble to the rule states that EPA modeling indicated an  expected co-benefit
              level, which is the result of implementing the CAIR rule, resulting in mercury
              emissions of 34 tons per year. EPA also took comment on  administering the cap-
              and-trade approach under CAA section 112 instead of section 111.10 The primary
              difference between these two approaches is that a section 112 cap-and-trade
              program would be administered centrally by EPA while the section 111 program
              would be administered individually by States.
Scope and Methodology
              We conducted our field work from May 2004 through December 2004, and did so
              in accordance with Government Auditing Standards, issued by the Comptroller
              General of the United States. We performed field work in EPA's Office of Air
              and Radiation locations in Washington, DC, and Research Triangle Park, North
              Carolina.  We interviewed staff from EPA offices and outside organizations to
              gain an understanding of the rule as developed, other options considered, and the
              rule development process. We interviewed officials from EPA's Office of Air
              and Radiation; Office of Research and Development; .Office of Enforcement and
              Compliance Assurance; and Office of Policy,  Economics, and Innovation. We
              also contacted environmental and utility industry representatives, and State, local,
              and tribal organizations interested in the development of this proposed rule, to
       8 A MACT standard more stringent than the floor is referred to as "beyond-the-floor."

       9 Concurrent with this approach, EPA proposed to revise its December 2000 finding that regulating utilities
under section 112 was necessary and appropriate.

         This approach would not require EPA to revise its December 2000 finding, but would require EPA to
"de-list" utilities as a source category requiring a MACT standard.
                                           8

-------
obtain their views. We reviewed data and analyses developed in support of the
rule, and public comments included in the rulemaking docket. We also reviewed
related information provided by both EPA and non-EPA officials contacted.

The Government Accountability Office is conducting a review of technology-
related issues for the proposed mercury rule, which is an important consideration
in determining whether the MACT standard can be set a level that is more
stringent than the floor. The Government Accountability Office report was not
available in December 2004 for consideration in the OIG report.

Limitations

Our evaluation was conducted and completed before the Agency had completed
the rulemaking process. Accordingly, our observations and characterizations
about the process reflect the status of the rulemaking process at the time we
completed our  review. Issuance of the final rule is planned for March 15, 2005,
and the final rule may consider additional information or analyses not available at
the time we completed our review. For example, EPA  released a notice of data
availability for the proposed Clean Air Mercury Rule on December 1, 2004. The
notice requests additional public comment on issues addressed in this report, and
solicits further comment on new data and information to help EPA evaluate which
regulatory approach will best reduce mercury emissions from power plants.  We
did not specifically consider the notice because it was released after we had
completed our  review and analyses.  However, the notice includes information
available previously in the public comment docket for this rule, and  it is possible
we had considered some of that information during our review.

The OIG was not provided with several important documents it requested from
the Agency; therefore, that information was not available for consideration in this
report. Our memorandum detailing the requested  information, as well as specifics
on what information was provided by the Agency, are provided in Appendix B.
Consideration of the inter-agency review process was limited to information from
EPA staff and information available in the docket only. We were not able to
discuss the inter-agency review process with Office of Management and Budget
(OMB) staff who were responsible for coordinating the inter-agency review
process. The OIG did not independently analyze the databases or computer
modeling programs that EPA used in developing the proposed rule.  With respect
to the development of the MACT standard, the OIG did not attempt  to
independently calculate the MACT floor.

-------
Results in Brief
             Evidence indicates that EPA senior management instructed EPA staff to develop a
             MACT standard for mercury that would result in national emissions of 34 tons
             annually, instead of basing the standard on what the top performing units were
             achieving in practice.  Also, we determined that EPA's mercury cap-and-trade
             proposal - a nationwide emissions trading program for an air toxic - can be
             strengthened to better ensure that human health is protected and anticipated
             emission reductions achieved, should this approach to reducing mercury
             emissions be adopted.  Further, although EPA rulemaking procedures are not
             consistently applied, Agency staff told us that they would have expected greater
             adherence to the guidance for mercury rule development due to the significance of
             this particular regulatory action, but this did not happen.

             We recommend that EPA re-analyze mercury emissions data collected, and
             conduct a revised cost-benefit analysis for the updated MACT that takes into
             account the impact of mercury co-benefits achieved through the proposed CAIR.
             We also recommend that the Agency strengthen its cap-and-trade proposal.
             Further, we recommend that the Agency conduct an integrated analysis with
             respect to whether emissions reductions under either of these proposals are the
             most child-protective,  timely, and cost-effective.

             The Agency disagreed with certain aspects of our draft report, and offered
             suggested changes or revisions. The Agency's response did not specifically
             address our recommendations.  We made changes to the final report based on the
             Agency's comments, as appropriate. See Appendix E for the full text of the
             Agency's official comments to our draft report and our response to these
             comments.
                                          10

-------
                                Chapter 2

         Mercury MACT Development Compromised
^^^^^^JJy^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^HS^^^fct^fcfcfcfchMJ^^M^**^^^^^^^^^^^^^^^^^^^^^

             Evidence indicates that EPA senior management instructed EPA staff to develop a
             MACT standard for mercury that would result in national emissions of 34 tons
             annually, instead of basing the standard on an unbiased calculation of what the top
             performing units were achieving in practice. The CAA requires that a MACT
             standard should, at a minimum, be based on the emissions levels achieved by the
             top performing 12 percent of units, not a targeted national emissions result. The
             34-tons-per-year target was based on the co-benefits expected to be achieved from
             implementation of NOx and S02 controls under the proposed CAIR. EPA noted
             that this target was based on extensive analysis and, in EPA's judgment,
             represented the lowest level of mercury emissions that it could reasonably expect
             the utility industry to achieve.

             Because the results of the MACT standard were prescribed and prior estimates
             were lower than what was proposed, we believe it likely that the standard
             understates the average amount of mercury emissions reductions achieved by the
             top performing 12 percent of power units. Some Agency officials told us that,  in
             their opinion, the true MACT floor would result in lower mercury emissions than
             the 34 tons estimated from current MACT floor limits. Therefore, if this proposed
             MACT standard was adopted,  it would not achieve  the maximum emission
             reductions achievable and the associated health benefits.  Further, this MACT
             standard, as proposed, does not provide a reasonable basis for comparison in
             determining which of EPA's two proposed regulatory alternatives (i.e., the MACT
             standard or the mercury cap-and-trade program) provides the better cost-benefit.

Requirements for MACT Standards

             In accordance with the CAA, EPA is to establish MACT standards that require
             the maximum emissions reductions the Agency believes are achievable for a
             major source category. At a minium, the MACT standard cannot be less stringent
             than the average emission reductions achieved by the  top performing 12 percent of
             units in a category (e.g., all coal-burning utilities) or subcategory (e.g., utilities
             burning bituminous coal) for which the Administrator has data. EPA has wide
             latitude in the types of emissions data used to determine the MACT floor,
             including the discretion  to select a reasonable method to estimate emissions
             achieved, and to address variability to account for the most adverse operating
             conditions reasonably foreseeable. If EPA decides to set a limit beyond the floor,
             it must consider the cost of achieving those reductions, any resulting non-air
             quality and environmental impacts, and energy requirements.
                                        11

-------
              In accordance with a court settlement, EPA had agreed to publish its final mercury
              rule by December 15, 2004. This date was re-negotiated with the court petitioner
              and the final rule deadline was extended to March 15, 2005.

EPA's Process for Addressing Variability in Computation of
Mercury MACT Floor

              As provided under CAA section 112(d), EPA first determined whether a MACT
              standard should be developed for all coal-fired units or sub-categories.  EPA
              analyzed the ICR data and identified the top performing units from all units for
              which emissions data were collected. Evaluation of the ICR data for the top
              performing units focused on coal type, plant processes, and control technology.
              EPA could not identify a common attribute that contributed to mercury emission
              reductions for all of the top performing units that would allow development of a
              single MACT emissions  limit for all units. Additionally, it was determined that
              no units had installed mercury-specific control technology, although controls
              installed to reduce emissions of other pollutants also helped reduce mercury
              emissions. When no single common factor was identified, EPA evaluated the
              data further and determined that sub-categorization by coal type, which is also a
              driving factor in plant design, was warranted to establish the MACT. One
              additional  sub-category was established for a particular plant type - Integrated
              Gasification Combined Cycle - because the plant burns gas from coal rather than
              any particular type of coal.

              For each sub-category, EPA identified the top performing units  based on emission
              tests collected during the ICR. However, EPA determined that  these emission
              tests alone did not sufficiently estimate the effect of fuel variability over time on
              the emissions of the best performing units. To account for this variability, EPA
              used coal composition data (i.e., mercury and chlorine content)  for coal shipments
              collected during the ICR to estimate emissions throughout the year for the top
              performing units in each subcategory.  This increased the number of emission
              points available from which to calculate the MACT limits. n

              The emission points for each of the  top performing units were ranked and then
              EPA selected one of the highest emissions points (i.e., the 97.5  percentile) for
              each unit.  According to  EPA, this emission point reflects the best performance
              under the worst foreseeable operating conditions for the unit. EPA took the
              average of these selected emission points for each sub-category and adjusted this
         Prior court cases have upheld EPA's right to consider variability in developing MACT floors. For a
discussion of the appropriateness of EPA's efforts to account for variability, see Cement Kiln Recycling Coalition v.
Envt 'I Protection Agency, 255 F.3d 855 (D.C.Cir. 2001), examining, Sierra Club v. Envt'I Protection Agency, 167
F.3d 658 (D.C.Cir.1999) and National Lime Ass'n v. Envt'l Protection Agency, 233 F.3d 625, 629 (D.C.Cir.2000)
("National Lime 11").
                                           12

-------
             average to further account for variability (i.e., the 97.5 percent upper confidence
             level of the average). This adjusted average was established as the MACT floor
             and the proposed standard for each subcategory.

             Unlike many previous MACT standards, the proposed utility MACT standard
             would not require the installation of a specific control technology since no
             mercury-specific control technology had been installed in utilities. EPA
             determined that emerging mercury-specific technologies were not yet
             commercially available for the utility industry.  The Government Accountability
             Office is conducting a study to assess the current state of mercury control
             technology.

EPA Staff Instructed to Develop MACT Floor That Would Result in
National Emissions of 34 Tons

             Evidence indicates that EPA staff were instructed to develop a MACT standard
             that would result in national emissions of 34 tons per year. Some staff told us that
             they heard these specific directions and others told us that they heard in different
             meetings during rule development that the application of the MACT floor to
             utilities should equal 34 tons per year (a 29-percent reduction from the present
             48-tons emitted nationwide). These statements were further corroborated by
             internal EPA e-mails, which specifically identified 34 tons per year as the number
             desired despite the fact that prior modeling results did not result in 34 tons.
             E-mails between EPA staff discussed various MACT emission limits by
             subcategory and modeling scenarios that could be used to get closer to the 34 tons
             target. For example, a November 2003 e-mail stated that:

                If the 14+K ofsubbitACI is using the 90% option and we restrict this to
                60%, perhaps we can get in the 34 tpy range.  I don't think that restriction
                would be considered inappropriate for a 2007 MACT analysis.

             EPA documents and an analysis of the process used to compute the MACT floor
             support EPA staffs statements that the MACT floor computations were
             developed to produce the desired national emissions of 34 tons per year.
             Documentation that we reviewed indicated that EPA conducted at least three
             Integrated Planning Model (IPM)12 runs in order to reach the pre-determined
             target for national mercury emissions of 34 tons. The initial IPM run to try to
             reach the 34-tons target yielded a national emission of 29 tons (i.e., the IPM
             model indicated that mercury could be reduced from 48 tons to 29 tons).  After
             changing the proposed MACT emission limits, a second IPM model yielded a
        EPA uses ICF Resources Incorporated's Integrated Planning Model for air emission modeling. The
model projects what decisions utilities would make for meeting air emission regulations based on economic
considerations.
                                          13

-------
              national emission of 27 tons.  While we were provided summary information
              about these two IPM model runs, they were not included in the EPA rulemaking
              docket.

              An Agency source indicated that these results were not acceptable to senior
              management because they were not close enough to the 34-tons target.  A third
              run performed, based on the proposed emission limits, showed 31 tons. EPA
              cited the 31-tons model results in the proposed rule, but explained in the preamble
              that 34 tons is the more probable emissions level because the model used to
              estimate emissions was underestimating the amount of mercury emissions that
              would occur.  EPA noted that the IPM model may have understated mercury
              emissions by 2.3 tons for units burning bituminous coal.13 Table 2-1 depicts the
              emission limits used in the three IPM runs and the resulting total national
              emissions:

                      Table 2-1: Results of Proposed MACT Scenarios to Reach 34 Tons
Goal type i
Bituminous
Sub-bituminous
Lignite
Total National Mercury
Emissions (tons-per-year)
Run#1
0.57*
6.46 *
18.45 *
29-30 "
Run #2
1.4*
5.06*
19.48 *
27.2-27.9**
Run #5 (Proposal ;
1.9679*
5.8*
9.2*
30-31"
               *  Proposed per unit mercury emission standard expressed in pounds per trillion British thermal
                  units (Ib/TBtu).
              **   Estimated tons of national mercury emission resulting from modeling the application of the
                  unit emission standard to all utility units.

              The emission limits shown in Run #3 above, ultimately proposed as the MACT
              standard, were based on a multi-variability analysis submitted by WEST
              Associates (a western utility consortium).14 However, EPA adjusted this
              approach, increasing the MACT floor emission limits for two of the three
              subcategories beyond those derived by WEST Associates. For example, WEST
              Associates used an upper confidence level of 95 percent of the mean of the best
              performing units to account for variability. EPA adjusted the confidence level to
         The 1PM model only allows Activated Carbon Injection technology, a mercury specific control
technology, to reduce mercury emissions at 60% and 90% levels. The inability of the model to address the full range
of reductions between these two levels means that the model may have understated mercury emissions by as much as
2.3 tons for bituminous-fired units.
       14
April 2002.
The analysis was submitted during the last Federal Advisory Committee Act meeting, convened in

                                    14

-------
             97.5 percent, which resulted in an increase in the emission limit for two of the
             three sub-categories.  According to EPA's variability analysis, this adjustment was
             made to account for EPA's interpretation of the number of units that should be
             included in the MACT floor analysis.15 These adjustments increased the MACT
             floor closer to  a national emission level of 34 tons per year.

             Relationship of the 34-Ton Estimate to Cap-and-Trade Proposals

             The 34-tons-per-year target is important because it is based on mercury emission
             modeling results used in two separately proposed cap-and-trade programs  for
             utilities - CAIR and the mercury cap-and-trade program - proposed as
             alternatives to the mercury MACT.  EPA has stated its intent to implement its
             multi-pollutant (mercury, S02, and NOx) cap-and-trade programs, originally
             included in stalled Clear Skies legislation, through the proposed CAIR and
             mercury regulations..

             EPA has also proposed that the mercury reductions gained from implementing
             CAIR should serve as the interim cap on mercury emissions in the mercury cap-
             and-trade program. According to the preamble to the mercury rule, the reason for
             basing the interim cap on the co-benefits from CAIR is that the Agency does not
             believe mercury control technology that has been demonstrated for all coal types
             is commercially available. In addition, Agency officials stated that the
             34-tons-per-year target was based on the co-benefits expected to be achieved  from
             implementation of NOx and S02 controls under the proposed CAIR. They noted
             that this target was based on extensive analysis and, in EPA's judgment,
             represents the  lowest level of mercury emissions that they could reasonably expect
             this industry to achieve by 2010.

             Additional Estimates of Mercury Emissions

             Interviews with sources both inside and outside the Agency suggest that if
             unbiased analyses of data were conducted, a range of possible MACT floor levels
             would most likely result. One EPA official stated that the true range of possible
             MACT floors  was probably as low as 8 to 10 tons per year up to the mid-20s, but
             that either end of that range would be a stretch.  Further, the source stated  that the
             real range is about 15 tons per year to the low 20s for this MACT, and that
             anything above or below those numbers was a stretch. This includes the 34 tons
             proposed by the Agency. These statements about the possible range of MACT
             floors are supported by results of different MACT floor limits and/or varying
             model assumptions used by some organizations providing comments to the
         For example, West Associates used 5 units for each sub-category, while EPA used 4 units for the
bituminous and sub-bituminous sub-categories and 5 units for lignite sub-category.


                                           15

-------
             proposed rule. For example, the Clean Air Task Force evaluated the ICR data to
             develop MACT floor limits that were different than those developed by EPA.
             Applying these limits to the same IPM model used by EPA resulted in national
             mercury emissions of 12 tons16 (i.e., a 75-percent reduction from 48 tons).
             Modeling by the Electric Power Research Institute and Edison Electric Institute
             used the MACT floor limits proposed in the rule and showed an estimated 32 tons
             of mercury emissions nationwide (i.e., a 33-percent reduction from 48 tons).
             Examples of varying modeling efforts and results can be found in Appendix C.

Conclusions

             EPA's current estimate of the amount of mercury emissions occurring after
             implementing SO2 and NOx controls, called for in EPA's CAIR, is 34 tons.
             Given (1) that EPA is attempting to implement the Clear Skies multi-pollutant
             approach through regulation; (2) the numerous modeling runs conducted to
             determine national emission resulting from different MACT emission limits;
             (3) the adjustments made in the accounting for variability; (4) the statements of
             EPA officials involved in the rulemaking process; and (5) EPA e-mails reviewed,
             we believe EPA's approach for developing the MACT floor was compromised.
             Further, it is unlikely that an unbiased calculation of the MACT floor would
             produce emission limits that would result in estimated national mercury emissions
             of 34 tons per year (i.e., EPA's current estimate of the co-benefit of SO2 and NOx
             proposed regulations).

Recommendations

             We recommend that the Assistant Administrator for Air and Radiation:

             2-1    Conduct an unbiased analysis of the mercury emissions data to  establish  a
                    MACT floor in accordance with the requirements of CAA section 112(d).

             2-2    Re-negotiate with the court petitioner for an extension of the final
                    rulemaking deadline sufficient to solicit and accept public comments on
                    the unbiased analysis of mercury emissions data in an open, public, and
                    transparent manner.

Agency Comments and OIG  Evaluation

             The Agency commented that the draft report incorrectly characterized the
             calculation of the MACT standard,  and that the Agency had calculated the MACT
        The Clean Air Task Force considered the effect of implementing the proposed CAIR rule on the mercury
MACT. EPA did not consider the impact of implementing CAIR in its MACT modeling efforts. More information
on this issue is found in Chapter 4 of this report.
                                          16

-------
floor in accordance with the requirements of CAA section 112(d). The Agency
also maintained that its extensive work, including development of the proposed
Clear Skies legislation, showed that, in the absence of immediately available
mercury control technology, the mercury reductions as co-benefits of SO2 and
NOx controls represent the lowest level of mercury emissions that the Agency
reasonably expects could be achieved. We believe our report accurately
characterized the MACT development process.  Our observations were based on
review of supporting documentation related to MACT development, and
interviews with Agency staff and stakeholders involved in the process, including
State and local, environmental, and industry groups. Although the MACT floor
was ostensibly based on data from the top performing 12 percent of units, this data
was analyzed with a final target already in mind, i.e., 34 tons. While the Agency
has conducted  analysis to determine the co-benefit of SO2 and NOx controls, we
do not believe  this meets the requirements of CAA section  112(d) in developing
the MACT standard. The Agency's complete response to the draft report and our
evaluation of its response are in Appendix E.
                             17

-------
                                Chapter 3

        Cap-and-Trade Option Cart Be Strengthened

             EPA's mercury cap-and-trade proposal - a nationwide emissions trading program
             for an air toxic - can be strengthened to better ensure that human health is
             protected and that anticipated emission reductions are achieved, should this
             approach to reducing mercury emissions be adopted. The cap-and-trade proposal
             could be strengthened by:

             •   Adequately addressing the potential for hot spots.
             •   Establishing an interim cap that would provide greater incentive for utilities to
                install mercury-specific control technology by 2010.
             •   Setting a reasonable safety valve provision.
             •   Clarifying conditions pertaining to exemptions for small emitting facilities.

             These changes could help ensure that the proposed mercury cap-and-trade
             program obtains the desired emissions reductions in a timely manner.

EPA's Proposed Cap-and-Trade Approach

             A cap-and-trade program could provide several benefits in terms of controlling
             emissions.  Trading programs generally provide regulated units with more
             flexibility to meet overall emissions reductions than do conventional command-
             and-control approaches because a unit may apply whichever control method it
             finds to be  most appropriate  and cost-effective to meet emission limits. This
             flexibility serves to minimize overall control costs in the market. Furthermore,
             cap-and-trade programs can provide greater environmental certainty by
             establishing fixed national emissions caps that cannot be exceeded. However, a
             cap-and-trade program's environmental benefits will depend on the adequacy of
             the cap.

             Under EPA's proposed mercury emissions trading program, units that cannot cost-
             effectively reduce emissions through controls may buy allowances from units that
             were able to reduce emissions beyond their established allowance limits and are
             willing to sell their extra allowances. Each unit is required to possess one
             emissions allowance per each ounce of mercury it emits. Units would be allowed
             to buy and  sell credits among one another in a national emissions market.  EPA's
             proposed cap-and-trade alternative proposes that the interim mercury emissions
             cap for 2010 be based on the amount of mercury reductions achieved solely as a
             co-benefit through implementation of SO2 and NOx controls under the proposed
             CAIR. As  noted in Chapter  2,  EPA's latest estimate of the mercury benefit from
             implementing CAIR is 34 tons per year.  The cap-and-trade proposal sets a final
             cap of 15 tons per year in 2018.
                                         18

-------
Proposed Cap-and Trade Program Needs to Further Address
Certain Issues

             The proposed cap-and-trade rule for mercury meets the three basic guiding
             principles of trading programs as defined by EPA: a cap on emissions,
             accountability, and simplicity of design and implementation.  However, we
             identified four issues with EPA's mercury cap-and-trade proposal that need to be
             further addressed. Details follow on each issue.

             Interim Cap Could Be Tightened to Force Earlier Development of
             Mercury-Specific Control Technology

             Although EPA has not yet set a specific interim cap  for 2010, the preamble to the
             proposed rule states that the interim cap will be based solely on the mercury
             emissions reductions achieved as co-benefits of regulating SO2 and NOx under
             CAIR, estimated by EPA to be 34 tons. Thus, it would not be necessary for units
             to install mercury-specific controls in order to meet the 2010 interim cap, and this
             would limit the effectiveness of the regulation to force new technological
             advances in mercury control. If the interim cap under this proposal is set at 34
             tons, utilities could delay consideration of installing new mercury-specific
             technology until meeting the more  stringent cap in 2018 is imminent. However,
             according to  EPA officials, if the banking provision of the cap-and-trade program
             operates as intended, some facilities would have the incentive to implement
             mercury-specific controls before 2018, which would reduce emissions beyond the
             interim cap level before the final cap becomes effective.  EPA officials also
             pointed out that experience under other cap-and-trade programs has shown that
             the largest emitters are typically the first to reduce emissions and will generally
             achieve the greatest level of reductions.  According to the preamble, the reason for
             basing the interim cap solely on the co-benefits from CAIR is that EPA does not
             believe mercury control technology that has been demonstrated for all coal types
             is commercially available.

             Further, the proposed rule does not address what would happen under the cap-and-
             trade approach if CAIR is not implemented.  Given that the 2010 cap is based
             solely on the co-benefits from CAIR, it is unclear what would occur under the
             proposed rule if CAIR is not implemented.

             An EPA official stated that although some EPA staff indicated they would like to
             see analyses on different cap levels for comparison purposes, no such formal
             analyses were conducted. EPA conducted one IPM run based on an interim cap of
             34 tons and a final cap of 15 tons (in conjunction with CAIR), but no runs were
             conducted using alternative caps for comparison.  Clear Skies analyses were made
             available in the proposed mercury rule docket, and according to an EPA official
             the mercury cap-and-trade IPM run is comparable to the Clear Skies IPM runs.
             According to this EPA official, one such run of Clear Skies had a different interim
                                         19

-------
cap (26 tons) and this run, while not exactly matching the modeling conducted for
the proposed mercury cap-and-trade program, provides an idea about the costs of
an alternative mercury cap.

Potential for Hot Spots Not Fully Analyzed

EPA did not fully analyze the potential for hot spots (i.e., areas of elevated
pollutant concentrations) to occur under its proposed cap-and-trade option. The
potential for hot spot formation under the proposed cap-and-trade rule has
generated a great deal of concern and debate among various stakeholders.
Modeling and projecting the  likelihood of hot spots under the proposed rule is
made difficult by the relatively high degree of uncertainty involved with mercury
transport and deposition patterns (i.e., when the airborne mercury is deposited
onto the ground or into water bodies), particularly local or near-field deposition.

Further complicating efforts  to use computer models to determine where mercury
deposition will occur is the fact that three different chemical forms of mercury are
emitted by utility units and each has varying deposition patterns.  For example,
oxidized  and particulate mercury are more likely to deposit locally or regionally,
while elemental mercury travels and is more global in nature.  Although air
emission-related hot spots are generally thought of in terms of high ambient air
concentrations near a source, this is not the only consideration with mercury.  The
main health risk associated with mercury is not its ambient concentrations, but
rather its deposition into water bodies and resulting bioaccumulation in fish.
However, the connection between air emissions and levels of mercury ultimately
found in fish tissue is not yet fully understood.

EPA's Clean Air Markets Division conducted  a Proximity Analysis to determine
"where, in relation to water bodies, emissions would occur" under the mercury
emissions trading provision of the Clean Air Interstate Rule. However, as noted
in the analysis, the issue of hot spots was not fully analyzed:

    This examination of projected mercury emissions has significant
    limitations and does not constitute an analysis of "hotspots." Such an
    analysis of hotspots would, in part, necessitate detailed assessments of the
    atmospheric fate, transport, and deposition of mercury from power
   generating sources, and assessments of the potential population  exposure
    to mercury contaminated Jlsh in water bodies due to generating and other
    sources.

Although EPA did not conduct the detailed assessment of hot spots described
above, EPA stated in the preamble to the proposed rule that it does not expect hot
spots to occur for several reasons, as follows:
                              20

-------
•  Modeling suggests that the largest emitters, which are more likely to produce
   local deposition, will be the first to implement control technology under a cap-
   and-trade approach and will reduce emissions by the largest amount.

•  CAIR would result in implementation of control technologies for SO2 and
   NOx that also provide the co-benefit of reducing emissions of the types of
   mercury (oxidized and particulate) that are likely to  deposit locally.

•  The Acid Rain program has not resulted in the formation of hot spots.

•  States have "the ability to address local health-based concerns separate from
   the mercury cap-and-trade program requirements," and under the proposed
   State-administered program would "retain the power... to adopt stricter
   regulations to address local hot spots or other problems."

•  The proposed final cap would be a 70-percent reduction in mercury emissions
   from current uncontrolled levels (from 48 to  15 tons).

However, potential problems arise with EPA's reasoning. For example, the Acid
Rain program controls for SO2 emissions, which are primarily deposited
regionally and globally, not locally, while mercury can deposit locally as well as
regionally and globally. Trading programs are generally thought to be most
effective for pollutants that do not deposit locally. Further, the Acid Rain
program co-exists with the National Ambient Air Quality Standards program,
which has established a minimum level of air quality for SO2, while no such
minimum standards exist as a back-stop in the mercury  cap-and-trade proposal.  In
addition, the Acid Rain program contains a provision stipulating that, in the case
of delayed implementation due to litigation, a more conventional command-and-
control approach would take effect, but the proposed cap-and-trade rule for
mercury lacks a similar provision.

While the preamble to the proposed rule notes that individual States have the
authority under section 111  to adopt stricter regulations than those set by EPA, it
does not address whether States would have this same authority under a section
 112 cap-and-trade program. Further, approximately one-third of States have laws
limiting "the ability of their regulatory agencies to adopt regulations that are  more
stringent than any federal environmental regulation." Thus, these States may not
be able to adequately address hot spots, should they arise.

EPA has recognized that additional information is needed to better understand and
address potential hot spots.  For example, in the preamble to the proposed rule,
EPA states its intent to reassess the hot spot issue by taking a "... hard look at the
Hg emissions inventory after full implementation of the first phase cap.. .," and
also requested comments on how it might address hot spots in a cap-and-trade
program. In addition, EPA suggested the use of trading ratios between regions as

                             21

-------
a way to address potential regional deposition differences.  The Agency also
requested site-specific data on areas where commenters believe hot spots would
continue to exist if a cap-and-trade program were implemented.
               o
Due to time constraints, the OIG did not fully evaluate potential environmental
justice implications resulting from a cap-and-trade program, nor did we fully
assess the extent of the Agency's analysis of these issues.

Safety Valve Provision May Not Encourage Reductions

The proposed safety valve price may be set too low to achieve the intended effect
of reducing mercury emissions through the installation of control technology and
the open-market trading of emission allowances. The safety valve provision in the
proposed cap-and-trade mercury rule provides a price cap on the cost of emissions
reductions, and was included in the proposed rule due to uncertainties associated
with future costs and the availability of mercury control technologies.  Under the
safety valve provisions of the proposed rule, if the price of allowances reaches a
certain level,  units will be permitted to borrow allowances from the future for a
fixed price. To help ensure that the overall cap on emissions is met over the long-
term, units can borrow only from their own bank of future allowances. The
provision is intended to "minimize unanticipated market volatility" and ensure
that "the cost of control does not exceed a certain level." Thus, in effect, units
may emit more in the current period, but would be forced to emit less  in the future
because they  are using future allowances.  However, we identified two concerns
with the proposed safety valve provisions.

Safety Valve  Price. For a safety valve provision to be used appropriately (that is,
only when market volatility makes it necessary),  the price should be set so that it
is higher than the market price of allowances or the actual cost of abatement
(emission reduction).  If this price is too low, it may be cheaper for the unit
operator to purchase future emissions allowance  at the safety valve price rather
than installing emission controls. Under the proposed rule, the safety  valve price
is set at $35,000 per pound, or $2,187.50 per ounce, adjusted annually for
inflation.  This figure was decided upon during development of the Clear Skies
Initiative, but new analyses have estimated that the actual cost of abatement will
be substantially higher than $35,000 per pound.

Although EPA stated in one of the rule's supporting documents that, "based on
current technological capabilities, the cost of mercury removal is expected to
reach the safety valve price ($35,000/lb) by 2010," it further stated that
"technological improvements could decrease the cost of mercury control over time
and cause prices to remain below safety valve levels." Staff within EPA indicated
that the current safety valve price of $35,000 was too low based on new analyses.
For example, 2003 and 2004 Department of Energy estimates show the "baseline
costs" of mercury removal to be $50,000 - $75,000 per pound, with cost
                              22

-------
reductions expected over time. However, senior EPA officials told us that they
did not believe the safety valve price would be reached because they expect the
cost of activated carbon injection, a mercury-specific control technology, to
decrease over time. According to these officials, the IPM does not account for
this variable and may be misleading since it shows the cost of activated carbon
injection remaining constant over time.

Safety Valve Borrowing.  The proposed rule stated that units may purchase safety
valve allowances from "following years," and the supplemental notice stated they
may be purchased from allowances available for allocation in the next control
period.  The supplemental notice also provided an example of how a State could
incorporate the safety valve provision into its cap-and-trade program.  However,
the proposed safety valve provision does not place a limit on the number of
allowances a unit can borrow under this provision. As the Clean Air Task Force
writes in its comments, a unit could, theoretically, continue borrowing indefinitely
from future years by buying safety valve allowances in lieu of installing controls
or buying allowances on the open market.  Such an approach would make
economic sense as long as the proposed safety valve price was set lower than the
baseline cost of controls. In the proposed rule, EPA acknowledges that its
"proposed approach may create implementation problems associated with the need
to 'reconcile' at some point in time the allowances borrowed from future
compliance periods," and requests comment on the issue.

Small Emitters Exemption Needs To Be Clarified

EPA has proposed that utility units emitting less than 25 pounds of mercury per
year be exempt from the cap-and-trade program, but has not completely addressed
how their exemption and the national emission cap will be impacted if their
emissions increase. EPA included this exemption because of concerns that new
mercury-specific control technologies  expected to be developed may not
practicably apply to these units. Based on EPA data developed for units
operational in 1999, 396 of the 1,120 units operational in 1999 were estimated to
have emitted less than  25 pounds  of mercury per year each.  These 396 units made
up 35.4 percent of the total operating units, but contributed only 3,742 of the
95,975 pounds of estimated mercury emissions, or 3.9 percent in 1999.
According to the proposed rule's preamble, EPA states there is reason to believe
that the 15-tons Phase  II cap can be achieved in a cost effective manner, even if
the lowest emitting 396 units are excluded from coverage under this cap. EPA is
soliciting comment on this proposal.

One commenter noted  that both capacity utilization and emission rate increases
could occur in small emitting sources after they have been exempted from cap-
and-trade requirements. EPA does not address this issue in the proposed rule.
Another commenter stated that EPA had done no analysis of the small emitter
exemption with respect to either costs  or impacts. According to this commenter, a
                             23

-------
             vast majority of the units emitting less than 25 pounds of mercury are part of a
             multi-boiler facility, and it is entirely likely that at some facilities all of the boilers
             are tied into common duct work for pollution control.  Consequently, these units
             should be considered as one unit emitting over 25 pounds and not eligible for the
             exemption.

             While we did not fully assess the impact of this, we believe the commenters have
             raised valid concerns.  Further, we noted that the relative significance of these
             small emitters increases as the cap-and-trade program progresses. For example, in
             2018, these emitters, based on their 1999 emissions, would represent 12.5 percent
             of the total 15 tons in emissions allowed under the final cap.  If EPA moves
             forward with its cap-and-trade proposal, the Agency can better ensure that
             anticipated emission reductions are achieved if it clearly addresses the
             circumstances under which small emitters would have to participate in the cap-
             and-trade program.

Proposed Emissions Trading  Rule Should  Also Address  Tribal
Concerns

             Although Executive Order 13175 requires EPA to develop an "accountable
             process to ensure meaningful and timely input by tribal officials in the
             development of regulatory policies that have tribal implications,"17 tribal concerns
             were not addressed during development of the proposed cap-and-trade rule. In the
             preamble, EPA states that the proposed rule may have tribal implications  because
             two coal-fired utility units are located in Indian Country. Representatives from
             the National Tribal Environmental Council (NTEC) informed us that neither they
             nor their approximately 180 member tribes had any involvement in the
             development of the proposed mercury rule.  This was confirmed by an EPA
             official at a March 2004 public meeting on the proposed mercury rule.

             Among NTEC's greatest concerns over the proposed mercury rule are:

             •   the absence of tribal involvement and/or consultation in the development of
                 the proposal;
             •   a  failure to adequately monitor mercury  deposition on tribal lands, which
                 means that the impact of mercury is unknown; and
             •   lack of consideration for American Indians and Alaska Natives'  dependence
                 upon fish and the terrestrial animals that feed on those local fish.

             The average tribal member and child eats much more fish than the typical
             consumer and the representatives explained  that tribes (especially children and the
             expanding youth population) are faced with  increased adverse health effects
       17 Consultation and Coordination with Indian Tribal Governments (65 FR 67249, November 6, 2000).
                                          24

-------
             caused by such exposure.

             NTEC does not support the cap-and-trade program and noted that, if the program
             is implemented, there is no mechanism currently in place for the tribes to enter
             into cap-and-trade allowance sales. In fact, allowances are only available to the
             States. NTEC cited the U.S. Government's trust responsibility, which includes
             looking after the health and survival of tribes.  This responsibility is met in part by
             conducting tribal consultation on a government-to-govemment basis.

             EPA officials noted that other organizations, including States, were not consulted
             during the development of the cap-and-trade proposal. Although States were not
             consulted, we noted that States were allotted mercury allowances while the Tribes
             were not.

Conclusions

             The cap-and-trade proposal can be strengthened to better ensure that the
             anticipated emission reductions are achieved, should this approach be adopted by
             EPA. First, the interim cap suggested under the current proposal  is set at a level
             that could be met without installing mercury-specific control technology, thus
             potentially delaying installation of mercury-specific controls until 2018. Also, the
             cap-and-trade option has not adequately addressed the potential for hot spots. In
             addition, EPA needs to ensure that it establishes a safety valve provision that will
             have the intended effect of encouraging unit operators to install controls or buy
             emission credits.  Further, EPA needs to ensure adequate tribal involvement for
             the proposed mercury rule to ensure that tribes are not negatively impacted by a
             cap-and-trade rule.

Recommendations

             We recommend that the Assistant Administrator for Air and Radiation:

             3-1    Re-assess the basis for the interim and final caps.  This analysis should
                    consider the results of the re-assessed MACT floor (see Recommendation
                    2-1).

             3-2   Further assess the risk of hot spots and, if CAA section 112 residual risk
                    requirements are not implemented, then section 111 cap-and-trade
                    regulations should specifically identify how EPA will meet its intention to
                    reassess the hot spots issue.

             3-3    Strengthen the safety valve provision so that the safety valve price is set at
                    a level whereby it is only used for its intended purpose of minimizing
                    unanticipated market volatility. Alternatively, EPA may stipulate other

                                           25

-------
                    controls over borrowing from future allowances, such as imposing a
                    greater than 1:1 allowance trading ratio; and allowances borrowed from
                    the future will be reconciled to ensure that facilities cannot borrow
                    indefinitely into the future.

             3-4    Reassess the necessity of a small emitter exemption, and if a decision is
                    made to exempt, explain in sufficient detail the reasoning for such a
                    provision and establish how small emitters will be handled within the cap-
                    and-trade program should they exceed emissions of 25 pounds a year.

             3-5    Address tribal issues by: developing a mercury emissions consultation
                    strategy with tribes, with the assistance of tribal representatives, that will
                    ensure the Agency fulfills its trust responsibility and conducts proper
                    government-to-government consultation with tribes; and establishing a
                    mechanism for coal-fired utilities located on tribal lands to participate in
                    the cap-and-trade approach.
Agency Comments and OIG Evaluation

             The Agency's comments expressed a concern that the report does not
             "comprehensively and accurately describe" how the proposed cap-and-trade
             approach would work. The Agency also expressed concern that we did not
             highlight the knowledge EPA has gained from modeling and past experience with
             cap-and-trade programs. We believe our draft report portrayed an accurate
             representation of how the proposed mercury cap-and-trade program would work.
             One of the objectives of our review was to evaluate whether the proposed cap and
             trade rule was sufficiently protective of public health. As a result, we highlighted
             certain concerns with the rule as proposed. We made revisions, where
             appropriate, based on technical comments made by Agency staff and officials.
             However, there are several important differences between the Acid Rain program,
             to which the Agency often refers when discussing past cap-and-trade experience,
             and the proposed mercury cap-and-trade program. The Agency's complete
             response to the draft report and our evaluation of its response are in Appendix E.
                                          26

-------
                               Chapter 4

     Rule Development Process Not Consistent with

	Expected and Past Practices	

            Although EPA rulemaking procedures are not always applied consistently, many
            Agency staff told us that they would have expected greater adherence to the
            guidance for mercury rule development due to the significance of this particular
            regulatory action, but this did not happen. When the Clear Skies legislation
            stalled, EPA decided to address the Clear Skies program in a regulatory manner
            instead. This led to EPA including a mercury cap-and-trade option, similar to
            Clear Skies, in its proposed mercury rule. As focus on the cap-and-trade approach
            increased, EPA began to de-emphasize the mercury MACT development process.
            This included:

            •   Cancelling the next scheduled Federal Advisory Committee Act (FACA)
                meeting and ending communication with FACA members.
            •   Abridging the normal intra-agency review process, particularly at the staff
                level.
            •   Failing to fully address the cost-benefit of MACT alternatives and not
                analyzing the potential impact of implementing CAIR on the proposed MACT
                option.
            •   Not fully analyzing the impact of the proposed mercury cap-and-trade program
                on children's health.

Description of Rulemaking Process

            EPA's Action Development Process: Guidance for EPA staff on Developing
            Quality Actions outlines steps EPA staff and management are to follow when
            developing Agency actions, such as rules, policy statements, and statutorily
            mandated reports to Congress.  The guidance suggests that EPA staff follow a
            prescribed set of steps beginning with tiering the action based on several of its
            characteristics. Once tiered, a standard process exists for developing the proposed
            action. As a Tier One action, the proposed mercury utility rule was considered a
            top action that would "... demand the ongoing involvement of the
            Administrator's office and extensive cross-Agency involvement on the part of the
            AAs/RAs (Assistant Administrators and Regional Administrators)."

            The Action Development Process guidance contains five key elements, which are
            summarized below. These include steps  for:

            •   planning sound scientific and economic analysis;

                                       27

-------
             •   developing and selecting regulatory options based on relevant scientific,
                 economic, and policy analyses;
             •   involving affected Headquarters and Regional managers early and continuing
                 involvement until the final action is completed;
             •   ensuring active and appropriate cross-Agency participation; and
             •   encouraging appropriate and meaningful consultation with stakeholders
                 through substantive consultative procedures.

             Appendix D describes the rule development process in detail.

Some FACA Members Considered Job Unfinished

             Within EPA, the creation of an advisory committee is not required for MACT rule
             developments, but such groups have been formed to advise the Agency in past
             MACT rulemakings and can provide a means of substantive consultation with
             stakeholders. An EPA official noted that for contentious rulemakings where a
             great deal of stakeholder involvement and public comment is anticipated, such as
             the mercury rule, it is not uncommon for an advisory committee to be formed.
             FACA allows for the creation of committees, boards, commissions, councils, and
             similar groups to furnish expert advice, ideas, and diverse opinions to officers and
             agencies in the executive branch of the Federal Government. The Act notes that
             the function of committees is advisory only, and decisions on how the advice will
             be used is determined by the official, agency, or officer involved.

             The FACA working group for this rulemaking, known as the Utility MACT
             working group, was formed within the Permits/New Source Review Air Toxics
             Subcommittee of the larger Clean Air Act Advisory Committee. Working group
             members consisted of representatives from State and local agencies;
             environmental organizations; industry; control equipment vendors; and coal
             interests, producers, and unions. Both co-chairs of the group indicated that they
             believed the working group had balanced stakeholder representation.18 The
             working group was formed for an initial period of 1 year and met approximately
             once per month starting August 2001.

             The working group was charged with providing input for the development of a
             MACT standard for utilities.  In a presentation given to the group by the EPA co-
             chair, the group was instructed that they were not to reconsider the Agency's prior
             finding that regulation of coal-fired electric steam generating units under section
             112 of the CAA was necessary and appropriate, nor were they to consider a cap-
             and-trade option. Although a cap-and-trade option was introduced in Congress in
             July 2002 in the Clear Skies legislation, this option was not considered by the
             working group.
       1R
        Although the working group did not include tribal representation, EPA solicited their participation.
                                          28

-------
In October 2002, the working group issued its final report, Recommendations for
the Utility Air Toxics MACT: Final Working Group Report, in which it identified
issues that "EPA must consider and resolve in its drafting of the utility MACT."
Some of the issues identified included:

•  sub-categories;
•  floor levels;
   beyond-the-floor levels of mercury;
   compliance method (monitoring); and
   compliance time.

The working group decided early that consensus among its various stakeholder
groups was unlikely, and did not attempt to reach agreement on specific
recommendations it could make to the Agency.  Instead, the report presented the
opinions of all the stakeholders on the issues.

Though the working group issued the final report in October 2002, it held another
meeting on March 4, 2003, just after Clear Skies legislation was re-proposed in
February.  Certain members  of the working group had requested that EPA conduct
additional analyses  using the IPM to further explore the cost-benefit of different
MACT proposals as presented by the working group members.  Members of the
working group did not have direct access to the  IPM,  as EPA contracts for its use
through a third party, and thus requested that EPA have the additional analyses
run and then provide the group with the results.  According to several members of
the working group we contacted, it was  expected that the working group would
receive the results of the additionally requested  IPM runs at the  March 4 meeting,
but were instead told the runs were not yet complete.  Another meeting was
scheduled for April 15, 2003, to provide the results of the IPM runs, but members
were notified by EPA of its cancellation via e-mail on April 1.

In July 2003, Administrator Whitman responded to Congressman Waxman's
request for the status of IPM runs for the working group. The Administrator
stated that it was the Agency's intention to convene an additional FACA meeting
when the IPM analyses were complete.  However, in March 2004, the Assistant
Administrator for Air and Radiation said the Agency would not provide the
additional MACT IPM analyses and would instead focus resources on developing
a cap-and-trade alternative, the administration's  preferred regulatory approach.

The working group has not met since its last meeting in March 2003 and has not
been officially contacted by the Agency since its planned April  15, 2003, meeting
was cancelled. A formal notice of termination has not been issued to the working
group and, according to some members, they were not given an explanation as to
why the working group ended. EPA has stated  on its web site that it began
proceeding with a cap-and-trade regulatory approach  in the absence  of
Congressional action on Clear Skies legislation. The FACA working group's
                             29

-------
             deliberations were stopped after Clear Skies was re-proposed and before EPA
             began developing its proposed cap-and-trade regulation. While some working
             group members indicated satisfaction with the work completed by the group,
             others considered the job unfinished due to the lack of opportunity to consider the
             additionally requested runs.

             According to senior EPA officials, the working group's original charter was for
             only one year.  One of the officials acknowledged that EPA had initially intended
             to conduct the runs requested by the working group but later decided that it would
             not be beneficial.  These officials further indicated that since the working group
             had not reached consensus, the Agency did not believe the working group should
             have been extended.
Intra-Agency Review Limited
             According to staff involved, the intra-agency work group review process followed
             in this rulemaking varied significantly from past Agency practice and applicable
             guidance for Tier One rules in that the group  only met two times and was not
             given an opportunity to provide meaningful feedback on the proposed rule.
             According to the Agency's regulatory development guidance, a work group is to
             meet frequently enough to ensure that all significant issues and options are
             discussed  and agreed upon.  Then, the significant issues and several options to
             resolve each issue are to be provided to senior management.  Senior management
             then selects those options they believe will best achieve the goals of the action for
             a Final Agency Review.

             The work group's first meeting was held on February 27, 2003, and the second
             and final meeting took place on August 7, 2004. In preparation for the first
             meeting, the work group chair e-mailed to the work group members a copy of the
             Utility MACT FACA working group's final report, along with a draft analytical
             blueprint for the rulemaking. According to EPA's Action Development Plan, an
             analytical blueprint is "a document that spells out a work group's plans for data
             collection and analyses that will support development of a specific action," and is
             intended to be developed as "a collaborative  effort." The draft blueprint stated,
             "the intent of the rule is to require that oil-and-coal-fired units achieve a MACT-
             level of control," and it listed the "minimum  analytical needs" for the rulemaking:

                 A regulatory impact analysis, assessing the economic impact on industry of
                 levels  beyond the MACT floor.
             •   Assessment of multi-pathway concerns.
             •   A regulatory flexibility analysis addressing small business concerns.
             •   Assessment of environmental justice concerns.
                 Children's health concerns.
             •   Unfunded mandate assessment, evaluating the impact of the rulemaking on
                 State/local/tribal governments, some of which own or operate coal-fired units.
                                          30

-------
•  ICR issues.

Although the above issues were identified for study in the draft analytical
blueprint, some were never fully addressed, such as the children's health study
and an assessment of environmental justice concerns.  The draft blueprint also
stated that:

    ". .. the EPA believes that emissions trading is prohibited under Section 112
   of the CAA.  However, industry, and to a more limited extent, some other
   stakeholders would like to explore emissions trading as an option (perhaps in
   beyond-the-floor analyses) for this rulemaking. "

Members of the work group, including the Office of Research and Development
and the Office of Policy, Economics, and Innovation, submitted comments to the
draft analytical blueprint via e-mail to the work group chair.  But work  group
participants we interviewed stated that they received no feedback or modified
drafts of any work products based on their comments and input.

In preparation for the second intra-agency workgroup meeting, members were
asked to review and comment on four sections (approximately 42 pages) of an
early version of the draft (July 3, 2003)  preamble. However, intra-agency
workgroup  members received no modified work products that incorporated their
feedback. Additionally, no Final Agency Review meeting was held for the
proposed mercury rule whereby core intra-agency review participants had the
opportunity to concur or nonconcur with the proposed rule before it was sent to
OMB for review and final action.

Several EPA staff who were involved in the abbreviated intra-agency work group
review process told the OIG that it was  made clear to them by their managers, and
in the case of one work group representative, by the work group chair, that
decisions about this rule were being made at a "higher level." For example, in an
e-mail discussing intra-agency comments, a member of the work group was told:

   The decision was made at a much higher level than mine to
    "bypass " the normal EPA Work Group procedure prior to the
   proposal and we have been told that all the Office directors
   were contacted about both the process change and rulemaking.

Similarly, these officials told us that it became clear to members that their
feedback would not likely be  considered.  One Agency source said that, in
general, there was not a meaningful opportunity for EPA offices to comment on
this rule. Some Agency officials said they considered the intra-agency  review
process to have been conducted, but at a higher staff level and with less input than
usual from  lower staff levels. However, at least one office usually involved in the
intra-agency review process - the Office of Enforcement and Compliance
                             31

-------
             Assurance - was neither given the opportunity to review nor submit comments
             regarding the proposed rule before it was sent to OMB, according to former and
             current Office of Enforcement and Compliance Assurance officials contacted.

             According to senior EPA officials it is not unusual during the development of
             high-profile rules, particularly those under a tight deadline, for EPA to not strictly
             follow the Agency's prescribed rulemaking process.

Requirements for Cost-Benefit Analyses Not Fully Implemented

             Although EPA conducted certain required analyses, other analyses were not
             completed. For rulemakings with an annual economic impact of $100 million or
             more, Executive Order 1286619 requires that Federal agencies, in deciding whether
             or how to regulate, assess all costs and benefits of available regulatory alternatives
             and provide the reasoning for selecting the proposed regulatory action over such
             alternatives. This Executive Order also  directs that Federal agencies base their
             decisions on the best reasonably obtainable scientific, technical, and economic
             information concerning the need for, and consequences of,  the intended
             regulation.

             EPA staff told OIG that senior management instructed them not to undertake
             certain scientific and technical analyses  that they thought necessary. For example,
             staff were instructed during meetings not to conduct IPM runs (which could have
             been helpful in considering alternatives) until they were told the national mercury
             emissions per year desired for the MACT. As discussed in Chapter 2, EPA
             conducted analyses of various MACT floor levels, but presented only a 34-tons-
             per-year option to the public.  In addition, the Agency did not fully analyze a
             beyond-the-floor MACT alternative.

             EPA's cost-benefit analysis of the MACT proposal did not take into account
             mercury emissions reductions that would be gained as co-benefits resulting from
             NOx and SO2 controls installed  under the proposed CAIR.  However, the
             Agency's cost-benefit analysis of the cap-and-trade option did consider CAIR
             co-benefits. This prevents a balanced comparison of the two options.  EPA staff
             told us that a MACT-plus-CAIR alternative was not analyzed because, when the
             MACT floor was completed,  CAIR had not yet been proposed.  However, EPA
             issued a December 2004 Notice of Data Availability for the proposed rule, which
             included an analysis submitted by the Clean Air Task Force that estimates the
             impact (in terms of emission reductions) of CAIR in conjunction with the
             proposed MACT standard.  The notice did not include a similar analysis by EPA.

             The Agency did not monetize the health benefits of mercury reductions, though
       19 Regulatory Planning and Review, 58 FR 51735, October 4, 1993.
                                          32

-------
              Office of Air and Radiation staff have said the final rule will include quantitative,
              non-monetized endpoints as well as a qualitative discussion. EPA staff told us
              that they have ongoing efforts to develop a benefits analysis, but that it is slow
              moving and has not been completed. Since March 2004, when the Administrator
              stated the Agency would take a closer look at the issue, there has been a process to
              try and do a full benefits analysis, but the process is moving slowly.  While a
              benefits analysis should be based on scientific literature, staff told us that there
              had been pressure to base the analysis on public comment through the Notice  of
              Data Availability. The notice presents a methodology for determining the benefit
              of mercury reductions and requests comment on this methodology.

Required Children's Health Analysis  Not Comprehensive

              EPA did not adequately evaluate the environmental health effects of the proposed
              rule on children. Executive Order 1304520 requires such an evaluation because
              "[a] growing body of scientific knowledge demonstrates that children may suffer
              disproportionately from environmental health risks and safety risks." In prior
              MACT rulemakings EPA had determined that Executive Order 13045 and,
              therefore, a children's health evaluation, is not applicable because MACTs are
              technology standards and apply consistently to covered sources.  However, since
              the proposed rule includes  a cap-and-trade option, which is a performance
              standard that could result in an uneven distribution of emissions, it is covered
              under Executive Order 13045 and, therefore, an analysis of the rule's impact on
              children's health is required.

              Although the proposed rule states that EPA evaluated health and safety effects
              pertaining to children, our review of the proposal and docket did not show that
              EPA performed such analyses in accordance with Executive Order 13045.  We
              requested such analyses from EPA, but were not provided with any specific
              studies of the rule's impact on children's health.  Interviews with officials from
              EPA's Office of Children's Health Protection indicated they were not involved
              during the rule development. However, Office of Children's Health  Protection
              staff said their lack of involvement in such functions is not unusual due to limited
              staffing.21  Members of the Children's Health Protection Advisory Committee
              (CHPAC) told us that the proposed rule does not adequately take into account
              children's vulnerabilities. The  CHPAC outlined their concerns in a January 26,
              2004 letter to the Administrator, in which they made several recommendations,
              including that the Agency "[e]valuate the possibility that hot spots could result"
       20 Protection of Children From Environmental Health Risks and Safety Risks, 62 FR 19885, April 23, 1997.
       21
         A May 2004 O1O report found that there was no overall, coordinated strategy integrating children's
environmental health efforts into the Agency as a whole (The Effectiveness of the Office of Children 's Health
Protection Cannot Yet Be Determined Quantitatively; 01G Report No. 2004-P-00016; May 17, 2004).
                                           33

-------
              from the cap-and-trade program as proposed.22 In a subsequent June 8, 2004 letter
              to the Administrator, CHPAC additionally recommended that EPA "[e]valuate the
              relative health benefits of reducing mercury exposure for children and women of
              child-bearing age under the MACT and cap-and-trade regulatory options."

              EPA senior officials noted that prior studies on the health impact of mercury
              addressed the impact of methylmercury exposure on children and, therefore, the
              rule itself addresses children's health.  We recognize that current reference dose
              levels for mercury exposure are based on the impact to children's health.
              However, we were not provided any analyses assessing the extent to which the
              proposed rule may result in uneven distribution of mercury deposition that could
              increase some children's exposure to mercury. Office of Research and
              Development officials noted that regardless of the extent of any additional
              analysis, they do not know what the impact of reducing sources emissions by a
              certain percentage would have on deposition or in what timeframe. However,
              they noted that reductions in emissions will reduce atmospheric mercury, which in
              turn will result in less deposition, lower mercury levels in fish, and ultimately
              reductions in human exposure to mercury.  EPA officials stated that this type of
              extensive analysis had not been done for the proposed rule, but they hoped to have
              a more detailed assessment for the final rule.  They further explained that the
              Notice of Data Availability issued in December 2004 proposed a process for
              quantifying the proposed rule's impact on mercury deposition and the resulting
              bioaccumulation in the environment.

Scope Limitation: Inter-Agency Review

              Due to time constraints23 and the fact that OMB controls this process and not
              EPA, the OIG did not evaluate the inter-agency review process and EPA's
              response to the edits resulting from that process.  The inter-agency review process
              occurs under the direction of OMB after a proposed rule is submitted to the Office
              of Information and Regulatory Affairs in OMB for review, as stipulated in
              Executive Order 12866.  The process is typically informal and, according to one
              EPA official, details on the meetings between OMB and other agencies, as well as
              comments submitted to OMB during the review, often are not included in the
              formal docket.

              It is difficult to determine every agency involved in the editing process, which
              agency made specific edits to the proposal, or the timing of these edits based on
              inter-agency review documents contained in the docket.  We identified comments
              from at least four agencies or offices other than EPA and OMB: the Department
         Chapter 3 of this report recommends that EPA further assess the risk of hot spots.

         Our field work in some areas was limited in order to provide the results of our review to EPA
management in time for (hem to consider our recommendations in developing the final rule.
                                           34

-------
             of Energy; the Department of the Interior; the Small Business Administration; and
             the Council on Environmental Quality.

Conclusions

             The rulemaking process did not meet the expectations of some EPA staff and
             FACA work group members, and did not fully address certain Executive Order
             requirements to conduct cost-benefit and children's health analyses. These
             deviations from prior practice and Executive Order requirements appeared to have
             occurred, in part, because of the Agency's decision to include a proposed cap-and-
             trade option in the proposed rule, as well as a need to meet the deadlines for the
             proposed MACT rule reached in prior court settlements.

Recommendations

             We recommend that the Assistant Administrator for Air and Radiation:

             4-1    Ensure  that the Office adheres to the Action Development Process during
                   EPA's future rulemaking actions to include obtaining input from all
                   relevant Agency Offices.

             4-2   Conduct more in-depth cost-benefit analyses of the proposed mercury
                   options to determine the preferred approach.

             4-3   Conduct a more in-depth analysis of the  impact of the proposed options on
                   children's health.

Agency Comments and OIG  Evaluation

             The Agency stated that the draft report failed to  recognize the nature of the
             regulatory development process and incorrectly  stated that EPA did not
             adequately evaluate the proposed rule's impact on children's health. Further, the
             Agency stated that the draft report improperly characterized the process by
             suggesting that it had not been sufficiently inclusive. We believe the draft report
             accurately described the rulemaking process, and continue to believe that the
             Agency should have more comprehensively evaluated the proposed cap-and-trade
             rule's impact on children's health.  A cap-and-trade program, while reducing
             overall emissions, can result in geographically uneven distributions of emissions.
             The proposed rule did not include an analysis of where or how likely such varying
             mercury emissions and resulting depositions could occur, and what impact this
             may have on children's health. The OIG does not agree that the Agency review
             process was inclusive.  As we noted in our draft report, according to staff
             involved, the intra-agency work group review process followed in this rulemaking
             varied significantly from past Agency practice and applicable guidance for Tier

                                          35

-------
One rules. Given this rule's far-reaching national implications for human health,
the environment, and the economy, the OIG believes it was important for the
Agency to have been more inclusive of available Agency expertise and external
stakeholder input to develop this rule.  The Agency's complete response to the
draft report and our evaluation of its response are in Appendix E.
                             36

-------
                               Appendix A
Timeline of Events Related to
Development of Mercury Rule
Date !
November 15, 1990
December 1997
February 1998
July 11, 2000
December 2000
August 1, 2001
July 2002
August 28, 2002
October 2002
February 27, 2003
February 27, 2003
March 4, 2003
March 4, 2003
March 14, 2003
Eveflt
President signs CAA Amendments of 1990. Section 112 requires EPA
studies of mercury and Hazardous Air Pollutant emissions from utilities.
EPA issues "Mercury Study Report to Congress." Emissions trading
discussed as a control option.
EPA issues "Study of Hazardous Air Pollutant Emissions from Electric
Utility Steam Generating Units." Defers decision on whether regulation of
utilities is necessary and appropriate under CAA section 112.
National Academy of Sciences releases report, "lexicological Effects of
Methylmercury," which concludes that EPA's reference dose for
methylmercury is a scientifically defensible level. Estimates that 60,000
newborns a year could experience neurological damage due to mercury.
EPA Issues Federal Register Notice making final determination that
regulation of mercury from utilities under CAA section 112 is "appropriate
and necessary." Discusses cap-and-trade as an option but states that
such an approach must protect local populations close to a source.
First meeting of Utility MACT working group. Charge to the Group is to
develop a MACT standard. Explicitly directed not to consider trading.
Clear Skies Act of 2002 introduced in the Senate and House of
Representatives. Proposed a multi-pollutant approach to controlling SO2, i
NOx, and mercury emissions from power plants.
EPA contractor memo outlines options for developing proposed MACT
floor.
Utility MACT working group issues final report. Consensus not reached.
Additional IPM runs recommended based on MACT emission limit
proposals from stakeholder groups.
Initial meeting of intra-agency work group (one of two total meetings).
Analytical blueprint prepared for group addresses traditional MACT, not
cap-and-trade, and identifies minimum analyses needed.
Clear Skies re-introduced in House and Senate as Clear Skies Act of
2003.
WEST Associates issues white paper proposing multi-variability method
for determining MACT floor; presented at last meeting of Utility MACT
working group. Paper presented to FACA at its last meeting. Method
eventually adopted by EPA but with some changes.
Last meeting of Utility MACT working group. April meeting canceled by
EPA; group had planned to discuss results of recommended IPM runs.
Briefing provided to Administrator Whitman. Presentation states EPA will
continue to develop a section 112 MACT standard unless Congress
removes the requirement.
             37

-------
Date
April 1,2003
August 7, 2003
November 4-5, 2003
November 26, 2003
December 15, 2003
December 2003
December 15, 2003
January 2004
January 28, 2004
January 30, 2004
March 16, 2004
	 Syfcntjt. 	
EPA cancelled last FACA working group meeting. E-mail indicates runs
not yet available, and meeting would be rescheduled at a later date.
Second (and final) intra-agency work group meeting held, reviewing draft
preambles. Several MACT emission limits proposed, none of which
match those in published proposed rule.
E-mails between EPA officials discuss efforts to establish MACT floor
resulting in mercury emissions of 34 tons per year, based on IPM runs
using various proposed MACT emission limits.
EPA memo to file explaining MACT floor (based on WEST Associates
method).
"Regulatory Flexibility Act Analysis" entered in Docket.
EPA contractor issued memorandum discussing beyond-the-floor
analysis.
Proposed mercury rule signed.
EPA Report on Benefit Analysis entered in Docket.
"Energy and Economic Impact Analysis" entered in Docket.
Proposed mercury rule published in the Federal Register.
Supplemental Notice issued to the original proposed rule providing
procedures for implementing cap-and-trade proposal.
38

-------
                                                                   Appendix B
          QIC's Request for Documents Related to
                  Development of Utility MACT
          UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                           OFFICE OF INSPECTOR GENERAL
                          OFFICE OF PROGRAM EVALUATION
          ',.              1301 CONSTITUTION AVENUE, N.W.(2460T)
       •-<-/ |                    EPA WEST BUILDING
 \ «f$&w»*                   WASHINGTON. DC 20004
  \
November 15, 2004

MEMORANDUM

Subject:     Document Request for Assignment Number 2004-1021 - Development of the
            Proposed MACT for Utility Units

To:         Jeffrey Holmstead,
            Assistant Administrator for Air and Radiation

From:       Kwai Chan, /s/
            Assistant Inspector General for Program Evaluation
This memorandum is a formal request to you and your staff cc'ed below for several documents
that we need in order to complete our work on the subject evaluation. The majority of these
documents have already been requested and are listed again herein. In addition, we are
requesting specific information (see item 7 below) not previously requested that is needed for us
to fully and comprehensively address our evaluation objectives.  We request that you provide us
with the following information by November 26, 2004, in order that this information can be
fully considered in our review:
      1.  Any and all statistical analysis and related internal correspondence for the two
         MACT IPM runs conducted in November 2003, including electronic records, that
         are not included in the docket.
      2.  Any and all written OGC analysis concerning use of Section 111 vs. Section 112,
         both for the December 2000 findings and determination and the January 2004
         proposed rule, including electronic records.
      3.  Any and all documentation showing final intra-agency concurrence (or equivalent)

                                          39

-------
          for issuing the proposed rule, including electronic records.
       4.  Any and all written comments resulting from the intra-agency review process,
          including electronic records.
       5.  The analysis related to children's health that was specifically referred to in the
          proposed rule's preamble on page 4715 of the Federal Register Notice.
       6.  The Agency analysis determining the origination of Latham and Watkins language
          that was included in the proposed rule's preamble, and
       7.  Any and all internal and external Agency correspondence or other written
          communications related to the development of the MACT floor that were developed,
          transmitted, and/or received during the period October 15, 2003 through December
          15, 2003, including e-mails meeting the definition of Federal Records.
We appreciate your prompt response to this request. Please contact Jim Hatfield, Assignment
Manager, at 919-541-1030, or Carolyn Blair, Project Manager, at 919-541-7702, to coordinate
the submittal of information related to this request. If any of the above information does not
exist please indicate that fact in your response.
cc: Robert Brenner, Deputy Assistant Administrator for Air and Radiation
   Bill Wehrum, Office of the AA for OAR
   Jason Burnett, Office of the AA for OAR
   Stephen Page, Director, OAQPS
   Sally Shaver, Director,  Emissions Standards Division, OAQPS
   Bob Wayland, Combustion Group Leader, BSD, OAQPS
   William Maxwell, Principal Rulemaking Contact, Proposed MACT for Utility Units, BSD,
   Nikki Tinsley, Inspector General
   Eileen McMahon, Assistant Inspector General for Congressional and Public Liaison
   Mark Bialek, Counsel, OIG
                                          40

-------
   Status of Agency's Response to OIG's Request for Documents
                Related to Development of Utility MACT
Item Requested
Status
1.  Any and all statistical analysis and related
internal correspondence for the two MACT
IPM runs conducted in November 2003,
including electronic records, that are not
included in the docket.
1.  We received limited information after the
draft report was provided to the Agency for
comment.  Specifically, we were provided
copies of Agency e-mails that discussed how
the information used in these MACT IPM
runs was developed.
2. Any and all written OGC analysis
concerning use of Section 111 vs. Section
112, both for the December 2000 findings
and determination and the January 2004
proposed rule, including electronic records.
2.  Since this was a legal issue before the
courts, we determined that we would not
address this, so the information was not
needed.
3.  Any and all documentation showing final
intra-agency concurrence (or equivalent) for
issuing the proposed rule, including
electronic records.
3.  No documentation provided.
4.  Any and all written comments resulting
from the intra-agency review process,
including electronic records.
4. No documentation provided.
5.  The analysis related to children's health
that was specifically referred to in the
proposed rule's preamble on page 4715 of
the Federal Register Notice.
5. Additional information in general was
provided after the draft report was issued,
but no analysis on children's health specific
to this rule was included.
6.  The Agency analysis determining the
origination of Latham and Watkins language
that was included in the proposed rule's
preamble, and
6. The Agency pointed out the information
in the docket related to this issue, but did not
provide specific Agency analysis. Since this
issue was related to inter-agency review
process, which is controlled by OMB, we
did not fully address this issue (See Scope
Limitation in Chapter 4 of this report.)
                                        41

-------
7. Any and all internal and external Agency
correspondence or other written
communications related to the development
of the MACT floor that were developed,
transmitted, and/or received during the
period October 15, 2003 through December
15, 2003, including e-mails meeting the
definition of Federal Records.
7. No documentation or response received
other than the limited information in the e-
mails provided for Request 1 above.
                                         42

-------
                                                                                                                            Appendix C
                                       Different MACT Floor Proposals
Source
EPA Proposed
Rule
FACA-
Environmental
FACA-
Industry
FACA- State &
Local - Option 1
FACA- State &
Local - Option 2
FACA -Clean
Energy Group
Clean Air Task
Force [4]
Emission Limits (Input base - Ibs/Tbtu)
: Process Sub-Categories
FBC
NA
0.1SQ
2.000 ;
NA
NA
NA
NA
FBC
(sub-Bit.
+ Bft)
NA
NA
NA
NA
NA
0,320
NA
FBC-
Lignite
NA
NA
NA
NA
NA
12,0
NA
Other
NA
0.210 (2j
NA
NA
NA
NA
NA
Bit. -
Hot
NA
NA
3.700
NA
NA
NA
NA
Bit. -
Wet
NA
NA
3.200
NA
NA
NA
NA
Bit-
Satu rated
NA
NA
2,200 :
NA
NA
NA
NA
IGCC
19.0
NA
NA
NA
NA
NA
4*
Coal Type Sub-Categories
Sub-Bit.
+ BH.
NA
NA
NA
0,600
0,400
1,223
NA
Bit
2,000
NA
NA
NA
NA
NA
0.420
Sub-
Bit
5.300
NA
,4,200.,
NA
NA
NA
1500
Lignite
9,200
NA
, 6,500
NA
NA
9,091
4.SOO
Coal
Refuse
0,4
NA
NA
NA
NA
NA
0,1
Total
Estimated
Emissions
34(1]
1.9 [3]
25-30 [3]
36 [6]
6.7 [3]
6.3 [3]
13.1 [3]
12 [5]
Abbreviations
Bit.:     Bituminous
FACA:   Federal Advisory Committee Act
FBC:    Fluidized Bed Combustion
Ib/Tbtu:  pounds per Trillion British thermal units
IGCC:   Integrated Gasification Combined Cycle
MACT:   Maximum Achievable Control Technology
NA:     Not Applicable
Notes:
1  Estimate based on Integrated Planning Model results.
2  Applied to all units except FBC
3  Based on estimates developed by Northeast States for Coordinated Air Use Management.
4  Clean Air Task Force preferred a MACT with no sub-categorization but re-computed a MACT floor based on
   EPA's proposed subcategories.
5  Based on Integrated Planning Model run and includes co-benefit reductions from Clean Air Interstate Rule.
6  Based on calculations performed by the industry group.
                                                                    43

-------
                                                                          Appendix D

                EPA's Rule Development Process
EPA actions are assigned to one of three tiers based on the nature of the anticipated issues and
the level of cross-Agency interactions needed to ensure a quality action. The proposed rule is a
Tier one rule and meets the following criteria.
 Tier t Criteria; Administrator's Priority Actions
 Thi$ B<£r wilt intludfe tap fc$fi0fl$ &$t d*fri$rtd th* Ongoing t
         cross-Agen«₯ irtvoivem$&t on tte partof fee Assisteflf/ftegion^ A^mfmStrators,
 Factors to consider in making a judgment about placing an action in Tier 1 are:

 •   major cross Agency or cross-media policy implications or precedents

 •   potential for major or precedent-setting implementation issues

 •   potential for major cross-Agency, cross-media, or inter-agency controversy
 •   potential for major economic impact on other levels of government or the regulated community
    highly controversial in terms of external interest

 •   ongoing, formal involvement of the Agency's highest level of management (Administrator,
    Deputy Administrator) is necessary or desired
    presents a significant opportunity for the Agency to advance the Administrator's priorities
 Action should be placed in Tier 1 if...
    science issue(s) are precedent setting and controversial
 •   economically significant per Executive Order 12866 (i.e., > $100 million), unless the program
    office can justify placement in Tier 2
    economics issue(s) are precedent setting and controversial
The program office develops the proposed rule, which may take months to years depending on
the complexity of the rule, priorities, and court/statutory deadlines.  Rule development follows
five major stages, as outlined in the Agency's Action Development Plan.  The first stage is
determining the proper tier for the action based on the criteria outlined above. The following
table describes the  five stages of an Action Development Plan.
                                           44

-------
fw« ftta|or Stafes of an Action Development Plan
Stage 1. Tiering the Action
    Understand tiering
•   Place action in the appropriate tier
•   Obtain tiering  approval
Stage 2. Developing the Proposed Rule or Draft Action
   Charter the workgroup
   Get the workgroup underway
•  Prepare the preliminary analytic blueprint and get early guidance from senior management
•  Prepare the detailed analytic blueprint
   Senior management approval of analytic blueprint
   Complete data gathering, consultation, peer review, analyses, and options development
•  Select Options
   Develop the proposed action by preparing preamble, rule, and supporting documents
   Conduct Final Agency Review to ensure senior management approval
   Office of  Policy, Economics, and Innovation review for rules deemed as "significant" under
   Executive Order 12866
Stage 3. Requesting OMB Review for Proposed and Final Actions (if necessary)
•   Determine if OMB review is necessary. Only those regulatory actions designated "significant"
    under Executive Order 12866, "Regulatory Planning and Review" are subject to review by OMB
    (e.g., actions having an annual effect on the economy of $100 million)
    Prepare regulatory action for submission to OMB
    Address OMB's comments
    Docket the OMB review process
Stage 4. Requesting the Administrator's Signature and Publishing an Action
    Request the Administrator's signature
    Publish the action in the Federal Register and open docket(s)
Stage 5. Developing the Final Action and Ensuring Congressional Review
•   Receive public comments
•   Consider and address public comments
    Determine next steps
    Submit actions to Congress under the Congressional Review Act or the Courtesy Copy Policy
                                          45

-------
                                                                       Appendix E
    Agency Comments to the Draft Report and OIG Evaluation
MEMORANDUM
SUBJECT:
FROM:
TO:
DATE:
Comments on the December 17, 2004 Draft Evaluation Report Entitled,
Additional Analyses of Mercury Emissions Needed Before EPA
Finalizes Rules for Coal-Fired Electric Utilities
Jeffrey R. Holmstead
Assistant Administrator for Air and Radiation
U.S. Environmental Protection Agency
William H. Farland, PhD
Acting Deputy Assistant Administrator for Science
Office of Research and Development
U.S. Environmental Protection Agency
Nikki Tinsley
Inspector General
U.S. Environmental Protection Agency
January 24, 2005
Thank you for giving us the opportunity to review the draft report referenced above and
to open dialogue with OIG staff. We have substantial concerns with the referenced draft
including several inaccuracies and flaws that we feel must be addressed before the report
is finalized. This memorandum briefly summarizes our major concerns.
Agency scientists and experts know a great deal about mercury: what are the sources,
both domestically and internationally; where does mercury in this country come from;
what is the chemistry that converts mercury deposited on the land and in the water into
mercury that becomes available to the food chain; what are the routes of exposure in this
country to mercury; what are the potential impacts of controls on that exposure; and what
is the status of the various technologies now being studied.
While some questions remain in our understanding of many of these linkages, this will
not prevent the Agency from regulating mercury from power plants, and it will do so as
effectively as possible, informed by the full body of knowledge it now possesses. The
Agency also recognizes that mercury emissions from facilities as complex as coal-fired
power plants should not be considered in isolation of the other efforts to reduce air
pollution; hence the Administration's strategy to further control SO2 and NOx while
instituting new, specific regulations for mercury. The Agency believes that such a
                                      46

-------
a strategy can deliver significant overall health benefits to a broad segment of the American
public.
EPA strongly urges the IG to take the broad base of information we know about mercury, as
well as the outstanding unanswered questions, into consideration when developing the final
report.
1. The draft report criticizes the rulemaking process as being incomplete even before a
final rule is issued. This critique rings hollow given the iterative nature of rulemaking. The
rulemaking process consists of a proposed rule, a public comment period and often additional
information before final decisions are made. The IG characterized the process as incomplete
before the process had finished. For example, a number of the issues regarding benefit-cost
analysis raised in the draft report are issues that the Agency is working on as evidenced by its
Notice of Data Availability on November 30, 2004.
    OIG Response:  Our review was initiated at the request of seven U.S. Senators, who asked
    that we complete this review in sufficient time to allow the Agency to address any issues
    raised in our report. We have added information to the Scope and Methodology section in
    Chapter 1 of the Final Report explaining that our review was completed while the Agency
    was still in the process of finalizing the rule. Accordingly, our report reflects findings and
    observations about the status of the process at the time we completed our review.  We look
   forward to seeing the results of the Agency's additional cost-benefit analyses, as
    recommended in our report.
2. The draft report inaccurately suggests that US power plant mercury emissions
represent a large part of the human exposure problem. Most exposure to mercury conies
from eating fish from the world's oceans and the mercury in these fish comes from a variety of
sources released over many years, including natural emissions like volcanoes, and
anthropogenic emissions from many countries, representing emissions from a variety of
sectors, in addition to emissions from US power plants. It is because US power plants are part
of the larger problem that EPA has proposed, for the first time ever, to require reductions from
this sector.
Given the global nature of mercury exposure and the uncertainty in the time to realize benefits
from current emission reductions, the action to reduce mercury emissions from power plants
must be seen in the larger context of all the activities EPA and others in the international
community are implementing to reduce exposure to mercury.
                                         47

-------
    OIG Response: Our draft report did not suggest that mercury emissions from U.S. power
    plants represent a large part of the human exposure problem. Power plants are one of
    many sources of mercury emissions.  The primary objective of our review was to assess
    EPA's development of the proposed rule for regulating mercury emissions from coal-fired
    electric utility units, and we included information in our draft report on mercury emissions
    and mercury health effects for background purposes. Nonetheless, we have included
    additional information in Chapter 1 of the Final Report to put total U.S. mercury and U.S.
    power plant emissions in the context of global mercury emissions.  We understand that
    primary route of human exposure to mercury is through the consumption offish and that
    the Centers for Disease Control and Prevention surveys indicate that seafood is the
    predominant type offish consumed by women of child-bearing age and children.
    However, certain subgroups, such as Native Americans, eat more fresh-water fish and may
    be more susceptible to mercury exposure than others.  We added this information to the
    background section of our final report.
3. The draft report does not comprehensively and accurately describe how the proposed
cap-and-trade system would work, leading the reader with misimpression about what our
experience and modeling has taught us.  The draft report fails to recognize that a
cap-and-trade system requires emissions reductions on a concrete timeline of declining caps,
thus leading to continual reduction of emissions and promotion of new technologies. It also
fails to acknowledge that, under this system, the largest emitters typically will be the first to
reduce their mercury emissions and will generally achieve the greatest level of reductions.
The draft report criticizes the cap-and-trade proposal for not requiring the installation of
mercury-specific controls until 2018, but this is inaccurate and reflects a misunderstanding
about how cap-and-trade works. The report should recognize the fact that it is reductions in
mercury emissions that will lead to improvements in public health and these reductions will
occur much earlier than 2018. Moreover, neither the Maximum Achievable Control
Technology (MACT) approach nor the cap-and-trade approach would require any particular
technology for controlling mercury. Either approach would require power plants to meet
certain standards for mercury control, and then let individual plants find the best way to meet
those  standards.
    OIG Response: One of the objectives of our review was to evaluate whether the proposed
    mercury cap-and-trade rule was sufficiently protective of public health. As a result, we
    highlighted certain concerns with the rule as proposed. As such, we limited our focus of
    the mercury cap-and-trade proposal to concerns about the interim cap level, the potential
   for hot spots formation, the safety valve provision, the exemption of small emitters, and
    tribal impacts.
                                          48

-------
    Our draft report portrayed an accurate representation of how the mercury cap-and-trade
   program works.  While the proposed mercury cap-and-trade rule should ultimately result
    in emissions reductions, we do not agree that the proposal provides a "concrete timeline
    of declining caps. "  For example, the proposed rule provides an interim cap that is based
    on co-benefits from existing technologies and can be achieved without the implementation
    of mercury-specific controls. Since the interim cap for mercury emissions can be achieved
    without mercury-specific controls, the proposed rule may not adequately promote the use
    of new technologies.  Also, the only other mercury cap is the 2018 final cap, and EPA
    modeling indicates it may not be met in 2018 due to the banking provisions of the
   proposed mercury trading program. Finally, our draft report noted that neither the
   proposed cap-and-trade nor the MACT  option require the use of any specific technology.


    While EPA has experience with cap-and-trade programs such as the Acid Rain program,
    there are differences in the transport and fate ofSO2 and mercury emissions which need
    to be addressed in a cap-and-trade approach to controlling mercury emissions. For
    example, S02 emissions are primarily deposited regionally  and globally, while mercury
    can deposit locally.  Additional differences between these  two cap-and-trade programs
    were highlighted in Chapter 3 of the draft report.
4. The draft report incorrectly characterizes the calculation of the MACT standard. The
draft report did not independently calculate the MACT floor, but instead simply relied on
assertions made by critics of the proposal as the basis for their critique. The proposed MACT
floor was calculated in accordance with the requirements of CAA Section 112(d) by basing the
standard on what the top performing 12 percent of units were achieving in practice, taking into
account subcategorization and variability.
Contrary to the claims in the draft report, the Agency did investigate beyond-the-floor MACT
alternatives and did propose a beyond-the-floor standard where technology was found to be
available (i.e., Integrated Gasification Combined Cycle (IGCC) subcategory).
                                         49

-------
    OIG Response:  The OIG did not inaccurately characterize the calculation of the MACT
   floor.  Our analysis was based on discussion with a number of EPA stakeholders and EPA
    officials, and review of supporting documentation. We found evidence that although the
    MACT floor was ostensibly based on data from the top performing 12 percent of units, this
    data was analyzed with a final target already in mind, i.e., 34 tons.  As stated in the
    Agency's Comment 5 to our draft report, this "floor" of 34 tons was obtained during the
    Clear Skies legislative process. Accordingly, we do not consider this floor to be based on
    an unbiased analysis of what the top performing 12 percent of units were achieving.

    With respect to IGCC units,  our review focused primarily on the development of the
    standards for existing units.  Of the over 400 coal-fired power plants in operation in the
    U.S., two are IGCC plants. Although EPA did not propose a beyond-the-floor standard
   for existing IGCC units, EPA proposed an emission limit for new IGCC units that was
    below the calculated floor for IGCC units and was based on EPA's determination that
    mercury reduction of 90 percent could be obtained for this subcategory through the use of
    carbon bed technology.
5. The draft report suggests that the proposed rule was flawed because other regulatory
alternatives that would achieve emissions levels lower than about 34 tons per year were
not developed or proposed. In particular, the draft report makes much of the fact that the
MACT proposal was developed with the goal of achieving a nationwide emissions level from
affected power plants of about 34 tons per year. The report fails to consider the fact that EPA
had developed  extensive information about mercury emissions and control techniques in the
power sector during the MACT regulatory development process and during the development of
the Clear Skies initiative.  That work caused us to conclude that mercury reductions could, in
fact, be achieved in the power sector over the  3-4 year MACT compliance period specified by
the statute.  However, these reductions would not come for the most part from mercury-specific
controls (such as activated carbon injection).  Extensive work conducted by the Office of Air
and Radiation and the Office of Research and Development indicated that mercury-specific
controls will not become readily available for commercial application to this industry until
2010 or later - well beyond the MACT compliance period. Consequently, the proposed rule is
predicated on the assumption that virtually all mercury reductions during the MACT
compliance period would have to be accomplished as a co-benefit of installing air pollution
controls designed to remove SO2 or NOx.  As part of the Clear Skies effort, EPA had
extensively studied the capacity of the power sector to install S02 and NOx controls during the
period up to 2010.  That work showed that 34 tons per year was the lowest level of mercury
emissions that we could reasonably expect the power sector to achieve through the aggressive
application of SO2 and NOx controls up to 2010. Further, as a part of the FACA process
established for this rulemaking, industry submitted what they thought would be possible under
a true co-benefit approach (i.e., no mercury-specific controls). Their estimate was that 36 tons
per year of mercury would be emitted under a MACT approach. The EPA proposal is
grounded in careful analysis as to what levels of mercury control reasonably can be expected
over the MACT compliance period.
                                         50

-------
    OIG Response: Our draft report concluded that the MACT development process was
    compromised for several reasons.  This included the fact that several MACTfloor
    proposals were lower than the EPA's proposed MACT rule, including several proposals
    developed by EPA in trying to achieve a floor that would result in annual emissions of 34
    tons. This included two EPA IPM runs that showed national emissions of 29 tons and 27
    tons, that were not included in the  rulemaking docket or available for public comment.
    While the Agency has conducted analysis to determine the co-benefit ofSO2 and NOx
    controls, we do not believe this meets the requirements ofCAA section 112(d) in
    developing the MACT standard. For example, the co-benefit is based on an average
    performance of all units, not just the best performers.  We continue to believe the Agency
    should conduct additional analyses before finalizing the rule. As noted in the draft report,
    the Government Accountability Office is conducting a review of technology-related issues
    for the proposed mercury rule.
6. The draft report fails to recognize the nature of the regulatory development process
and incorrectly states that EPA "did not adequately evaluate the environmental health
effects of the proposed rule on children."  We have made it clear from the start of the
rulemaking process that the health effects of greatest concern are possible developmental
effects in fetuses and young children exposed to unsafe levels of methylmercury. Unlike most
other rules that EPA develops, this rulemaking is singularly directed at developing an
appropriate regulatory approach for addressing the potential impacts on children. Evidence of
this can be seen in EPA's first guiding principle in the development of a final mercury rule
which states that the rule will concentrate on the need to protect children and pregnant women
from the health impacts of mercury.
Consistent with this principle, EPA Office of Air and Radiation participated in an ongoing
dialogue with the Children's Health Protection Advisory Committee (CHPAC) and responded
to CHPAC's recommendations on mercury exposure in children. Further, EPA and others
have conducted extensive work on the health effects of mercury for the developing fetus and
young children, including a National Academy of Sciences review completed in 2000. The
Inspector General's draft report misses this key point.
                                         51

-------
    OIG Response: We do not believe we failed to recognize the nature of the rulemaking
    process.  Further, the Agency should have more comprehensively evaluated the proposed
    cap-and-trade rule's impact on children's health. A cap-and-tradeprogram, while
    reducing overall emissions, can result in geographically uneven distributions of emissions.
    The proposed rule did not include an analysis of where or how likely such varying
    mercury emissions and resulting depositions could occur, and what impact this may have
    on children's health.
    Children's Health Protection Advisory Committee members did not characterize their
    interaction with the Agency as an ongoing dialogue. Committee members  told us that the
    Agency's response to their concerns with the proposed rule did not satisfactorily address
    their recommendations.
7. The draft report improperly characterizes the process by suggesting that it has not
been sufficiently inclusive. EPA has held dozens of high-level inter-office and external
meetings on this rule. This inclusive process was needed both because the rule has
far-reaching national implications for human health, the environment, and the economy and
also because a well-informed decision on an issue this complicated requires hearing diverse
perspectives. While there is always room to improve communications within and with those
outside of EPA, there is little basis to fault the Agency in this case.
    OIG Response:  The OIG does not agree that the Agency review process was inclusive.
    As we noted in our draft report, according to staff involved, the intra-agency workgroup
    review process followed in this rulemaking varied significantly from past Agency practice
    and applicable guidance for Tier One rules. Specifically, the work group process
   followed in this rulemaking was unusual in its short duration, infrequent meetings, late
    start with respect to the final rule deadline, and overall lack of communication and
   feedback between the workgroup and Agency decision makers. Further, workgroup
    members were not given the opportunity to review and comment on an entire draft
    proposal before it was published in the Federal Register.  For example, staff from the
    Office of Enforcement and Compliance Assurance were never given a draft of the
    proposed rule to review or comment on, thus this office could not assess the adequacy of
    the proposed rule's monitoring, record keeping, or reporting provisions as it typically
    does for Tier One MACTs.  With respect to meeting with external stakeholders, tribal
    representatives told us  that they were not consulted during the development of the
    proposed cap-and-trade option. Given this rule's far-reaching national implications for
    human health, the environment, and the economy, the OIG believes it was important for
    the Agency to have been more inclusive of available Agency expertise and external
    stakeholder input in developing this proposed rule.
                                          52

-------
Again, thank you for the opportunity to review the draft report. We would be happy to work
with you and your staff to ensure that you promptly receive all the information and analysis you
need to finalize the report. The final report should include an improved discussion of (1) the
global nature of mercury exposure and the uncertainty in the time to realize benefits from
current emission reductions; (2) how a proposed cap-and-trade system would require emissions
reductions on a concrete timeline; (3) the approaches to calculation of the MACT floor; (4) the
substantial effort EPA devoted to evaluating the risk of mercury exposure on children; and (5)
the inclusiveness of EPA's process towards reaching a final rule.
    OIG Response: The Agency's comments have been included in the final report as
    appropriate. We appreciate the efforts of both the Office of Air and Radiation and the
    Office of Research and Development in working with us to clarify certain technical issues
    and in providing prompt input so that we could issue our report in a timely manner.
                                          53

-------
                                                                      Appendix F
                                Distribution
Acting Administrator (1101 A)
Assistant Administrator for Air and Radiation (6101 A)
Deputy Assistant Administrator for Air and Radiation (6101 A)
Acting Deputy Assistant Administrator for Science, Office of Research and Development
      (A101R)
Acting Deputy Assistant Administrator for Management, Office of Research and Development
      (A101R)
General Counsel, Office of General Counsel (4010A)
Agency Followup Official (the CFO) (2710A)
Agency Followup Coordinator (2724A)
Audit Followup Coordinator, Office of Air and Radiation (6102A)
Audit Followup Coordinator, Office of Research and Development (A102R)
Associate Administrator for Congressional and Intergovernmental Relations (1301 A)
Associate Administrator for Public Affairs (1101 A)
Director, Office of Air Quality Planning and Standards (C404-04)
Deputy Director, Office of Air Quality Planning and Standards (C404-04)
Audit Liaison, Office of Air Quality Planning and Standards (C404-2)
Inspector General (2410)
                                        54

-------