Report on the Proposed Changes to the Federal Mercury and Air Toxics Standards (MATS)

Environmental economists from RFF, Harvard, Yale, and other leading research institutions say an Environmental Protection Agency (EPA) proposal that could eventually allow more mercury pollution from power plants relies on a flawed cost-benefit analysis.

View Report


Dec. 4, 2019


Joseph E. Aldy, Matthew Kotchen, Mary Evans, Meredith Fowlie, Arik Levinson, and Karen Palmer



Reading time

8 minutes

Please visit the report page (above) for full text and citations.

Executive Summary

In 2012, the Environmental Protection Agency (EPA) promulgated the Mercury and Air Toxics Standards (MATS) to regulate the emissions of mercury and other air toxics at electricity generating units (EPA 2012). The agency argued that this rule-making is “appropriate and necessary” because: (1) electricity generating units are the largest domestic source of mercury emissions, and they emit other hazardous air pollutants; (2) these emissions pose a hazard to public health; and (3) effective emission controls are available. In 2015, the Supreme Court ruled that EPA must consider costs in making an appropriate and necessary finding and remanded the regulatory finding to EPA, but permitted implementation of the regulation to proceed (Michigan v. EPA). Coal- and oil-fired power plants began demonstrating compliance with the MATS in April 2016. In the same month, the agency responded to the Supreme Court decision by issuing a Supplemental Cost Finding (EPA 2016) based on several cost metrics, including the rule’s original benefit-cost analysis showing the benefits far exceeding the costs. The 2016 Supplemental Cost Finding reached the same conclusion as in 2012, namely that regulating mercury and air toxics at electricity generating units is appropriate and necessary.

In February 2019, EPA published a regulatory proposal to revise the Supplemental Cost Finding for the MATS (EPA 2019). In the proposal, EPA finds that it is no longer appropriate and necessary to regulate hazardous air pollutants (HAP) from coal- and oil-fired power plants under Section 112 of the Clean Air Act (EPA 2019). EPA states that in evaluating the appropriate and necessary finding, “the most appropriate basis for comparison is the relative size of the target pollutant benefits, both quantified and unquantified, relative to the costs imposed by the rule” (EPA 2018, p. 5). While the EPA acknowledges the co-benefits associated with reductions in PM2.5 and SO2 emissions, it claims that co-benefits should not be given equal weight to the target pollutant benefits (HAP emissions reductions) in making the appropriate and necessary finding.

By proposing to rescind the appropriate and necessary finding that serves as the premise for the MATS, EPA sets the stage for revising a rule that the agency initially estimated would impose annual costs of $9.6 billion and yield annual benefits ranging from $33 to $90 billion. Accompanying EPA’s 2019 proposal, the agency issued a six-page memorandum that includes a reinterpretation of the cost and benefit estimates from the MATS regulatory impact analysis (RIA) completed in 2011 (EPA 2018). The only change between the estimates included in the RIA (EPA 2011) and the memorandum is the exclusion of co-benefits.

Cost estimates and the benefits of direct mercury emission reductions remain unchanged, with the consequence being that the costs of MATS now appear to exceed the benefits.

In July 2019, the External Environmental Economics Advisory Committee approved this committee’s proposal to evaluate a series of economic questions raised by the EPA’s 2019 proposed rulemaking on the appropriate and necessary finding. We address two major topics: the accounting of co-benefits in regulatory impact analyses, and the scope for updating the analysis to reflect the most recent data, research, and understanding of the market subject to the regulation. We include in Appendix 1 the initial proposal with questions that provide the structure of our report. Here in the executive summary, we summarize four key findings.

(1) The EPA’s 2018 cost-benefit memo does not follow best practices for economic analysis with its omission of co-benefits of the MATS rule.

When determining whether a policy promotes economic efficiency, properly estimated direct benefits and co-benefits (or costs) should count on an equal footing when making benefit-cost calculations. We provide a simple conceptual framework that shows how and why accounting for co-benefits—such as those associated with human health effects from reduced exposure to particulate matter less than 2.5 micrometers in size (PM2.5) resulting from the MATS rule—are important. We conclude as a matter of best practices for benefit-cost analysis, that the EPA’s proposed revision to the supplementary finding is not consistent with the generally accepted understanding of how to quantify the net benefits of changes in co-pollutants.

We also note how the relevant government agencies themselves have already weighed in with similar answers to the question. The Office of Management and Budget (OMB) states that, when conducting

a benefit-cost analysis, agencies should, “[i]dentify the expected undesirable side-effects and ancillary benefits of the proposed regulatory action and the alternatives. These should be added to the direct benefits and costs as appropriate” (OMB 2003, pp. 2-3). EPA guidance states that, “An economic analysis of regulatory or policy options should present all identifiable costs and benefits that are incremental to the regulation or policy under consideration. These should include directly intended effects and associated costs, as well as ancillary (or co-) benefits and costs” (EPA 2014, p. 11-2).

Moreover, to the specific question of health co-benefits for the MATS rule, OMB states as recently as 2017 that “particulate matter ‘co-benefits,’ make up the majority of the monetized benefits, even though the regulation is designed to limit emissions of mercury and other hazardous air pollutants. The consideration of co-benefits, including the co-benefits associated with reduction of particulate matter, is consistent with standard accounting practices and has long been required under OMB Circular A-4” (OMB 2017, p. 13).

(2) The 2018 EPA benefit-cost memo underestimates the public health benefits of reducing mercury emissions.

In the original RIA for MATS, the quantified target HAP benefits, which range from $4 to $6 million in the first full year of compliance, derive exclusively from one category of benefits: increased IQ among children exposed to methylmercury (MeHg) from self-caught freshwater fish. At the time, EPA acknowledged several other categories of HAP benefits, but did not quantify them due to a lack of scientific consensus, data, or methodological limitations. The same approach and estimates were carried over into the EPA’s reconsideration of the benefits and costs in 2018.

We note substantive advances in the peer-reviewed research on the health impacts of mercury exposure that occurred between 2011 (the year the original RIA was released) and 2018 (Rice et al. 2010; Drevnick et al. 2012; Hutcheson et al. 2014; Cross et al. 2015; Giang and Selin 2016; and Sunderland et al. 2018). We highlight two advances in particular. First, scientists better understand the process by which mercury emissions from U.S. power plants disperse and deposit in fresh, coastal, and international waters, along with the implications of this process for exposure to MeHg through the supply of seafood in the United States. Second, recent studies that provide evidence on the health benefits in the United States of reduced MeHg exposure and incorporate cardiovascular impacts find that these effects dominate those from neurologic effects (i.e., IQ). More specifically, the monetized benefits of the MATS rule through mercury-related cardiovascular risk reduction (primarily fewer heart attacks) are estimated to be on the order of billions of dollars per year.

(3) The 2018 EPA benefit-cost memo fails to account for significant power sector changes since 2011.

During the eight years since the publication of the EPA’s original RIA for MATS, the power sector has experienced significant changes. Anticipating such changes is inherently difficult in ex ante analyses, especially for an industry that is now in the midst of unprecedented transition with respect to both technological innovation and slower demand growth. While not fully anticipating these shifts in the

2011 RIA is to be expected, not acknowledging or accounting for them in a reevaluation of the benefits and costs in 2018 provides an incomplete account of the MATS impacts, which in turn might inform the reconsideration of the appropriate and necessary determination.

Shifts in the electric power sector, for reasons apart from MATS implementation, have been significant enough to materially affect the estimates of the MATS benefits and costs. For example, EPA predicted in 2011 that just under 50% of electricity generation in 2015 would come from coal and that just under 18% would come from natural gas. In fact, by 2015, coal’s share of generation had declined to roughly one-third and natural gas generation had increased to approximately the same share. Underlying these differences between ex ante predictions and ex post realizations are lower natural gas prices, lower electricity demand, and greater renewable electricity generation. Recent peer-reviewed, retrospective studies have found that only a relatively small fraction of these shifts was due to the implementation of MATS. Specifically, two recent studies estimate similar impacts of MATS on coal-fired power plant retirements – about 5 GW of capacity, or 14% of the total retirements – and these estimates are approximately in line with the original EPA projection in 2011.

The more general trends that have shaped coal-fired generation mean that compared to ex ante predictions, fewer plants incurred capital expenditures associated with MATS compliance, and the cost of operations and maintenance of such equipment was lower. Indeed, coal-fired generating capacity is about one-fifth smaller today than in 2011, and output from generators still operating is significantly lower than in previous years (and significantly lower than EPA’s forecast). This means that not only are the costs of MATS smaller than expected, the anticipated impacts on emissions and associated health outcomes are smaller as well.

(4) A new retrospective and prospective benefit-cost analysis could better represent the impacts of the MATS rule.

In 2019, three years after power plants began complying with MATS, considerably more information is available to understand the impacts of the MATS regulation than was available in 2011. New and updated retrospective and prospective benefit-cost analyses of MATS would provide a more accurate evaluation of the rule’s economic impacts, in addition to a more fully informed basis to consider reevaluation of the EPA’s appropriate and necessary finding.

A new retrospective analysis of the MATS rule could build on several recent studies to assess the costs, emissions, and monetized benefits of the regulation. These could also leverage a richer understanding of the market factors influencing coal-fired power plant retirement, generation, and pollution control investment decisions. With respect to the important PM2.5 co-benefits of the MATS standard, such a retrospective analysis could also examine the extent to which MATS-related emission reductions induced a relaxation of other regulatory requirements on sources of PM2.5 emissions, such as State Implementation Plans required under the Clean Air Act.

A new prospective analysis should reflect the insights gained from such a retrospective analysis. Looking forward, the analysis could incorporate the most recent epidemiology and integrated assessment modeling of the public health benefits associated with reducing power plant mercury emissions. Likewise, the analysis should include the benefits associated with co-pollutant emission reductions, reflecting an updated assessment of how the choice of pollution control technology in practice influences the emissions of PM2.5 and SO2. The geographic location of reductions in PM2.5 and its precursors has important implications for public health benefits, and EPA could employ a richer approach for accounting for the regional variation in emission reductions in estimating such benefits.


This report was produced by the External Environmental Economics Advisory Committee (E-EEAC), an independent organization dedicated to providing up-to-date, non-partisan advice on the state of economic science as it relates to regulations of the U.S. Environmental Protection Agency (EPA). The E-EEAC was established following EPA’s dissolution of its own Environmental Economics Advisory Committee in 2018. That committee had provided guidance for over 25 years within the EPA’s science advisory board structure.

This is the first report by the E-EEAC and it focuses on EPA’s analysis that seeks to justify its proposed changes to the Mercury and Air Toxics Standards.

Specifically, the analysis was conducted by the following economists who are members of the E-EEAC's Mercury and Air Toxics Standards (MATS) Review Committee:

Joseph Aldy, Harvard University (co-chair)

Matthew Kotchen, Yale University (co-chair)

Mary Evans, Claremont McKenna College

Meredith Fowlie, University of California, Berkeley

Arik Levinson, Georgetown University

Karen Palmer, Resources for the Future


The E-EEAC receives funding from the Sloan Foundation; the Luskin Center for Innovation of the University of California, Los Angeles; and the Roberts Environmental Center at Claremont McKenna College. The authors appreciate support from JR DeShazo, Colleen Callahan, and Christian Zarate of the UCLA Luskin Center for Innovation. The authors also acknowledge helpful feedback and support from Ken Norris, Amanda Giang, Josh Linn, Kristen McCormack, Edward Rubin, Jim Stock, and the following members of the E-EEAC Executive Committee: JR DeShazo (UCLA, co-chair of the E-EEAC Executive Committee), Antonio Bento (University of Southern California), Dallas Burtraw (Resources for the Future), Reed Johnson (Duke University), Kenneth Gillingham (Yale University), Matthew Neidell (Columbia University), Steve Newbold (University of Wyoming), V. Kerry Smith (Arizona State University).


The analysis presented is those of the authors and not necessarily the authors’ affiliated universities/organizations nor of the funders/supporters.

For more information, contact the Co-Chairs:

Joseph Aldy, Harvard University at [email protected] and

Matthew Kotchen, Yale University at [email protected].


Related Content