Harvard Law Review Harvard Law Review Harvard Law Review

Administrative Law

The Clean Power Plan

EPA Interprets the Clean Air Act to Allow Regulation of Carbon Dioxide Emissions from Existing Power Plants.

On August 3, 2015, President Obama announced the finalization of the Clean Power Plan1×1. Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric Utility Generating Units, 80 Fed. Reg. 64,662 (Oct. 23, 2015) [hereinafter Final Rule] (to be codified at 40 C.F.R. pt. 60). (CPP), a regulation he described as “the single most important step America has ever taken in the fight against global climate change.”2×2. Remarks Announcing the Environmental Protection Agency’s Clean Power Plan, 2015 Daily Comp. Pres. Doc. 546, at 2 (Aug. 3, 2015). The CPP marks the Environmental Protection Agency’s (EPA’s) first foray into regulating carbon dioxide emissions from existing power plants and, with the simultaneous enactment of regulations setting carbon dioxide emissions standards for new power plants,3×3. Standards of Performance for Greenhouse Gas Emissions from New, Modified, and Reconstructed Stationary Sources: Electric Utility Generating Units, 80 Fed. Reg. 64,510 (Oct. 23, 2015) (to be codified at 40 C.F.R. pts. 60, 70, 71, 98). represents an unprecedented exercise of the EPA’s jurisdiction over the energy sector.4×4. See Jody Freeman, Why I Worry About UARG, 39 Harv. Envtl. L. Rev. 9, 12, 17 (2015) (predicting that CPP opponents will frame the EPA as acting more like “an energy regulator rather than a pollution regulator,” id. at 17). While some argue that the changes the EPA is asking states to make are beyond the bounds of its statutory authority,5×5. See, e.g., Laurence H. Tribe & Peabody Energy Corp., Comment on Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric Utility Generating Units, at 3 (Dec. 1, 2014), http://www.masseygail.com/pdf/Tribe-Peabody_111(d)_Comments_(filed).pdf [http://perma.cc/4QSQ-ARR5] (criticizing the proposed version of the CPP). the EPA’s first fight appears to be over whether the agency has the authority to regulate greenhouse gas emissions from existing power plants at all. Whether the EPA has such authority will come down to the reading of one provision of the Clean Air Act: section 111(d).6×6. 42 U.S.C. § 7411(d) (2012).

In the year between issuing its proposed (Proposed Rule) and final (Final Rule) versions of the CPP, the EPA substantially altered its approach to interpreting section 111(d). While both readings reach the same conclusion and result in the EPA regulating the same universe of sources, the new legal analysis reveals the EPA as an agency responding to King v. Burwell,7×7. 135 S. Ct. 2480 (2015). the Affordable Care Act (ACA) case decided some five weeks before the Final Rule was announced. As one of the first major pieces of regulation released post-King, the CPP may serve as an early example of how agencies will justify their statutory interpretations in a time of uncertain relations between courts and agencies.

On June 25, 2013, President Obama issued a memorandum directing the EPA “to use [its] authority under sections 111(b) and 111(d) of the Clean Air Act to issue standards, regulations, or guidelines, as appropriate, that address carbon pollution from modified, reconstructed, and existing power plants and build on State efforts to move toward a cleaner power sector.”8×8. Power Sector Carbon Pollution Standards, Memorandum for the Administrator of the Environmental Protection Agency (June 25, 2013), in 78 Fed. Reg. 39,535, 39,535–36 (July 1, 2013). Section 111(b) of the Clean Air Act regulates pollutants from new sources. See Clean Air Act § 111(b), 42 U.S.C. § 7411(b) (2012). Section 111(d) regulates pollutants from existing sources. See id. § 7411(d). On June 18, 2014, the EPA proposed a rule in the Federal Register to regulate emissions from existing power plants.9×9. See Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric Utility Generating Units, 79 Fed. Reg. 34,830 (proposed June 18, 2014) [hereinafter Proposed Rule] (to be codified at 40 C.F.R. pt. 60). The Proposed Rule aimed to reduce carbon dioxide emissions from the power sector thirty percent from 2005 levels by 2030 and to achieve net climate and health benefits of between $48 and $82 billion.10×10. See id. at 34,832. These benefits were to be achieved through the enforcement of “[s]tate-specific emission rate–based CO2 goals” as permitted by section 111(d), with states free to use an array of policy tools to meet their targets,11×11. Id. at 34,833. Thus, although the CPP establishes emission performance rate standards for power plants, states need not rely solely on directly reducing emissions from their power plants to achieve their reduction goals. including energy efficiency and renewable portfolio standards.12×12. Renewable portfolio standards “require retail electricity suppliers to supply a minimum percentage or amount of their retail electricity load with electricity generated from eligible sources of renewable energy.” Id. at 34,849.

The public response was immense and immediate. The EPA received more than 4.3 million comments on the proposal.13×13. Final Rule, supra note 1, at 64,704. What is more, industry members “champing at the bit to challenge EPA’s anticipated rule” sought to prevent the rule from being finalized by bringing a claim in the D.C. Circuit, which dismissed the claim as premature.14×14. In re Murray Energy Corp., 788 F.3d 330, 333 (D.C. Cir. 2015); see also id. at 333–34. These claims gave the EPA a rather unusual opportunity to preview challenges to the CPP before it had actually been finalized. The EPA’s final brief in In re Murray Energy shows the agency beginning its shift toward the analysis of section 111(d) found in the Final Rule. See Final Brief for Respondents at 34–54, In re Murray Energy, 788 F.3d 330 (Nos. 14-1112, 14-1151). When the EPA announced the Final Rule on August 3, 2015, it took only ten days for a group of states to petition the D.C. Circuit for an emergency stay of the rule, even before the rule had been published in the Federal Register.15×15. See Emergency Petition for Extraordinary Writ, In re West Virginia, No. 15-1277 (D.C. Cir. Sept. 9, 2015) (per curiam). This petition was also dismissed as not meeting the “stringent standards” required for extraordinary relief.16×16. See In re West Virginia, slip op. at 1.

The final Clean Power Plan differs in many substantive respects from the Proposed Rule, including, notably, its removal of consumer-side energy efficiency standards as a factor in the calculations used to set target emission rates.17×17. See Jody Freeman, How Obama Plans to Beat His Climate Critics, Politico, http://www.politico.com/agenda/story/2015/08/how-obama-plans-to-beat-his-climate-critics-000186 [http://perma.cc/QN57-93TA]; see also Fact Sheet: Clean Power Plan Key Changes and Improvements, U.S. Envtl. Protection Agency, http://www2.epa.gov/cleanpowerplan/fact-sheet-clean-power-plan-key-changes-and-improvements (last updated Aug. 13, 2015) [http://perma.cc/KM9F-HBLM]. The Final Rule also foresees a slightly greater decrease in carbon dioxide emissions — now thirty-two percent below 2005 levels by 2030 instead of thirty percent. Compare Proposed Rule, supra note 9, at 34,832, with Final Rule, supra note 1, at 64,679. In the Final Rule, the EPA also “reconsidered its previous interpretation” of section 111(d)18×18. Final Rule, supra note 1, at 64,710. : instead of finding section 111(d) contradictory, and therefore ambiguous, as argued in the Proposed Rule, the Final Rule attempted to reconcile the potentially conflicting texts. While the EPA reached the same conclusion about the breadth of its jurisdiction as in the Proposed Rule, the analytical shift between the Proposed and Final Rules is pronounced.

Section 111 of the Clean Air Act regulates air pollution from stationary sources.19×19. See 42 U.S.C. § 7411 (2012). The section provides the EPA with a logical progression: First, the Administrator is required to create a list of categories of stationary sources that “cause[], or contribute[] significantly to, air pollution which may reasonably be anticipated to endanger public health or welfare.”20×20. Id. § 7411(b)(1)(A). Next, the Administrator must develop “standards of performance for new sources” within the listed categories.21×21. Id. § 7411(b)(1)(B). Then, and of particular interest here, the Administrator must establish regulations for states to submit their own “standards of performance for any existing source for any air pollutant,” subject to a few key restrictions.22×22. Id. § 7411(d) (emphasis added). Whether those restrictions preclude the EPA from regulating carbon dioxide emissions from power plants is the interpretive question addressed in both the Proposed and Final Rules.

In the Proposed Rule, the EPA described the somewhat odd status of section 111(d). In 1990, the Senate and House both passed amendments to the section, and both versions were included in the conference package, approved by both chambers, and signed into law without being reconciled.23×23. See EPA, Legal Memorandum for Proposed Carbon Pollution Emission Guidelines for Existing Electric Utility Generating Units 21–23 (June 18, 2014) [hereinafter Legal Memorandum for Proposed Rule] http://www2.epa.gov/sites/production/files/2014-06/documents/20140602-legal-memorandum.pdf [http://perma.cc/HPZ5-RCXJ]. The Senate version “would exclude the regulation of any pollutant which is ‘included on a list published under [Clean Air Act section] 112(b),’”24×24. Proposed Rule, supra note 9, at 34,853 (alteration in original) (citing Act of Nov. 15, 1990, Pub. L. No. 101-549, § 302(a), 104 Stat. 2399, 2574). the Clean Air Act’s hazardous air pollutant program. The House version “would exclude the regulation of any pollutant which is ‘emitted from a source category which is regulated under section 112.’”25×25. Id. (quoting Act of Nov. 15, 1990 § 108(g), 104 Stat. at 2467). In the Proposed Rule, the EPA argued that both versions were good law, but that the provisions were contradictory: the Senate version would allow regulation because carbon dioxide was not a pollutant listed under section 112(b), but the House version would preclude regulation because the sources in question, power plants, were already being regulated for other pollutants under section 112.26×26. See id. For a discussion of whether both provisions are in fact good law, compare Laurence H. Tribe, Why EPA’s Climate Plan Is Unconstitutional, Harv. L. Today (Mar. 20, 2015), http://today.law.harvard.edu/why-epa-climate-plan-is-unconstitutional [http://perma.cc/2VSH-V73J] (arguing that only the House version is law), with Jody Freeman & Richard Lazarus, A Rebuttal to Tribe’s Reply, Harv. L. Today (Mar. 21, 2015), http://today.law.harvard.edu/a_rebuttal_to_tribe_reply [http://perma.cc/HKE2-CTAD] (arguing that both versions are law). In justifying its Proposed Rule, the EPA found these conflicting texts created ambiguity and argued that choosing the permissive Senate version was “both reasonable and entitled to deference.”27×27. Legal Memorandum for Proposed Rule, supra note 23, at 12; see also id. at 26 (“[I]t is not reasonable to give full effect to the House language.”). Thus, the EPA found it had authority to regulate carbon dioxide from power plants, even though the plants were already regulated for other pollutants.28×28. See Proposed Rule, supra note 9, at 34,853.

In the Final Rule, though the outcome remains the same, the legal analysis is quite different. Instead of finding the two amendments contradictory and therefore ambiguous, the Final Rule asserts that each authorizes the EPA to regulate carbon dioxide from power plants.29×29. Final Rule, supra note 1, at 64,710–11. As in the Proposed Rule analysis, the Senate version is described as unambiguously permissive of the EPA’s rulemaking authority.30×30. See id. at 64,712. However, the Final Rule describes the House version, previously seen as unambiguously precluding this regulation, as “ambiguous and subject to numerous possible readings.”31×31. Id. at 64,713. The EPA then proceeds through a text-centric analysis of three of these possibilities,32×32. See id. at 64,712–14. including the reading from the Proposed Rule, before determining that each is unreasonable in light of the Act’s “comprehensive scheme.”33×33. Id. at 64,714. The reading the EPA settled on, which the agency found to be the only reasonable option, is that the exclusion precludes only duplicitous regulation: section 111(d) cannot be used to regulate hazardous air pollutants (which are enumerated elsewhere) from source categories described in section 112, but it can be used for the regulation of other pollutants from those same source categories, including carbon dioxide.34×34. See id. This reading, the EPA noted, was also endorsed by the Supreme Court in 2011. See id. at 64,778; see also Am. Elec. Power Co. v. Connecticut, 131 S. Ct. 2527, 2537 n.7, 2539–40 (2011) (finding that federal common law nuisance actions for greenhouse gas emissions were precluded by the EPA’s authority to regulate greenhouse gases from power plants under the Clean Air Act).

It is not unheard of for an executive agency to change its analysis of its own statutory authority.35×35. See, e.g., Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967, 981 (2005) (“Agency inconsistency is not a basis for declining to analyze the agency’s interpretation under the Chevron framework.”). And in fact, this isn’t even the first time the EPA has revised its interpretation of this particular section of the Clean Air Act.36×36. See Legal Memorandum for Proposed Rule, supra note 23, at 26 (citing Revision of December 2000 Regulatory Finding on the Emissions of Hazardous Air Pollutants from Electric Utility Steam Generating Units and the Removal of Coal- and Oil-Fired Electric Utility Steam Generating Units from the Section 112(c) List, 70 Fed. Reg. 15,994 (Mar. 29, 2005)). But the change here seems less an example of a statute being reread to fit new circumstances and more an act of defense in anticipation of future judicial review. After the Supreme Court decided King v. Burwell on June 25, 2015, some commentators predicted trouble for the CPP.37×37. See, e.g., Jonathan H. Adler, Could King v. Burwell Spell Bad News for the EPA?, Wash. Post: Volokh Conspiracy (July 3, 2015), http://www.washingtonpost.com/news/volokh-conspiracy/wp/2015/07/03/could-king-v-burwell-spell-bad-news-for-the-epa [http://perma.cc/NZL2-APUF]. Although Chief Justice Roberts’s opinion had upheld the Internal Revenue Service’s reading of the ACA, the Court offered no deference to the agency’s interpretation and foreclosed any alternative reading.38×38. See King v. Burwell, 135 S. Ct. 2480, 2489, 2496 (2015). While the King decision was released only some five weeks before the CPP was finalized, the Final Rule cites King four times.39×39. See Final Rule, supra note 1, at 64,712, 64,715, 64,769 n.519, 64,775. The CPP provides an early look at how agencies may justify their statutory interpretations in the shadow of King: with more careful legal analysis in the face of greater uncertainty over whether Chevron will apply, with an attempt to read texts to harmonize rather than conflict even in cases of presumptive statutory error, and with a resulting preference for definitive answers rather than retained agency flexibility.

In King, the Court provided two reasons for not applying Chevron deference to its analysis of the ACA. First, the Court noted that the provision at issue raised a major question of the sort involved in FDA v. Brown & Williamson Tobacco Corp.40×40. 529 U.S. 120, 160 (2000) (finding that the FDA did not have authority to regulate tobacco because “Congress could not have intended to delegate a decision of such economic and political significance to an agency in so cryptic a fashion”). — “a question of deep ‘economic and political significance’ that is central to this statutory scheme,” which requires a clearer signal from Congress than was present in King.41×41. King, 135 S. Ct. at 2489 (quoting Util. Air Regulatory Grp. v. EPA, 134 S. Ct. 2427, 2444 (2014)). The Court seemed to find this question particularly worthy of attention because of both the significance of the ACA itself and the significance of the particular interpretive question at issue to the fate of the ACA as a whole.42×42. See id. Second, the Court looked at the qualifications of the agency invoking deference, noting that “the IRS . . . has no expertise in crafting health insurance policy of this sort.”43×43. Id.

The Court’s jettisoning of Chevron based on its interpretation of King as an “extraordinary case[]”44×44. Id. at 2488 (quoting Brown & Williamson, 529 U.S. at 159). leaves agencies with real uncertainty about whether Chevron will apply to their statutory interpretations. In the case of the CPP, the King considerations point in different directions. The CPP is certainly a big undertaking — while the EPA estimates the regulation will generate substantial net benefits, the estimates for total compliance costs range from $5.1 to $8.4 billion by 2030.45×45. Final Rule, supra note 1, at 64,679. However, no reading of section 111(d) would have as significant an effect on the Clean Air Act as a contrary reading in King would have had on the ACA, where a different interpretation could have led to a “death spiral” in health insurance markets.46×46. See King, 135 S. Ct. at 2493. This difference seems significant. While a contrary reading of section 111(d) would clearly have enormous impact on the continued existence of the CPP, the King Court was concerned with the statutory, not regulatory, framework. The Clean Air Act has existed without the CPP for this long and presumably could continue to do so even if the CPP were struck down. Agency expertise would also seem to point firmly toward deference to the EPA: Congress specifically authorized the EPA “to prescribe such regulations as are necessary” under the Clean Air Act,47×47. 42 U.S.C. § 7601(a)(1) (2012). and has long recognized the EPA as its expert in the field of environmental regulation.48×48. See, e.g., Am. Elec. Power Co. v. Connecticut, 131 S. Ct. 2527, 2539 (2011) (“It is altogether fitting that Congress designated an expert agency, here, EPA, as best suited to serve as primary regulator of greenhouse gas emissions.”).

Still, with new uncertainty about when Chevron deference will apply, agencies may be encouraged to support their interpretive positions more thoroughly. The Proposed Rule relied almost exclusively on the existence of Chevron deference: the EPA found two conflicting readings and chose one that it found to be “both reasonable and entitled to deference.”49×49. Legal Memorandum for Proposed Rule, supra note 23, at 12. The Final Rule spends more time with the text, analyzing all possible readings as carefully as a court might, and finding ambiguity in the House version where none had been found before.

The analysis in King also injects uncertainty into the question of whether possible statutory errors create ambiguity worthy of deference. While the Court in King made only oblique reference to some “inartful drafting” within the ACA,50×50. King v. Burwell, 135 S. Ct. 2480, 2492 (2015). “the possibility of a drafting error is a cloud that hangs over the opinion.”51×51. Abbe R. Gluck, The Supreme Court, 2014 Term — Comment: Imperfect Statutes, Imperfect Courts: Understanding Congress’s Plan in the Era of Unorthodox Lawmaking, 129 Harv. L. Rev. 62, 101 (2015). By foreclosing Chevron deference from the outset in King, the Court sidestepped its fractured 2014 holding in Scialabba v. Cuellar de Osorio,52×52. 134 S. Ct. 2191 (2014). which did address how to handle interpretation in a case of possible statutory error. In that case, a plurality of the Court deferred to the Board of Immigration Appeals’s interpretation of a seemingly contradictory immigration statute.53×53. Id. at 2203 (plurality opinion). Chief Justice Roberts concurred in the judgment, but offered a strong critique of the plurality’s deference: “Direct conflict is not ambiguity, and the resolution of such a conflict is not statutory construction but legislative choice. Chevron is not a license for an agency to repair a statute that does not make sense.”54×54. Id. at 2214 (Roberts, C.J., concurring in the judgment).

In light of uncertainty over whether courts will give more or less deference to agency readings of conflicting texts, the EPA’s Final Rule reads the competing versions of section 111(d) as reconcilable rather than necessarily contradictory. In the Proposed Rule, the EPA read the two versions as mutually exclusive, ambiguous precisely because the EPA had to choose one version or the other.55×55. Proposed Rule, supra note 9, at 34,853 (“The two versions conflict with each other and thus are ambiguous.”). The Final Rule, though, reads the two versions as pointing in the same direction, citing not to Scialabba’s plurality, which supported agency deference in the face of conflict, but to Chief Justice Roberts’s concurrence and Justice Sotomayor’s dissent, both of which were able to find coherent readings of the provisions at issue.56×56. See Final Rule, supra note 1, at 64,713 (citing Scialabba, 134 S. Ct. at 2214 (Roberts, C.J., concurring in the judgment); id. at 2219–20 (Sotomayor, J., dissenting)). However, in the case of section 111(d), although it seems unlikely, Congress really could have been legislating in contradictory terms: this isn’t an issue of inconsistencies in a single document, but of two distinct amendments with distinct authorship.

One possible outcome of these post-King approaches to justifying a chosen interpretation is a shift away from asserted statutory ambiguity and toward a position that would foreclose the availability of a different interpretation. The EPA’s analysis in the Final Rule not only presents its reading as reasonable, but also attempts to show that no other reading could be reasonable — a marked shift from its earlier finding of two opposed but possible solutions. In fact, the EPA may even prefer not to receive deference here: if Chevron deference is dispositive of how the CPP is reviewed, future administrations could undo the entire basis for the CPP with a change in interpretation. At the oral argument for King, the Chief Justice asked if relying on Chevron “would indicate that a subsequent administration could change [the agency’s] interpretation.”57×57. Transcript of Oral Argument at 76, King v. Burwell, 135 S. Ct. 2480 (2015) (No. 14-114), http://www.supremecourt.gov/oral_arguments/argument_transcripts/14-114_1bo2.pdf [http://perma.cc/74B3-ZQRT]. By foreclosing alternative readings, the King Court prevented future administrations from fundamentally dismantling the ACA. In its final legal analysis, the EPA may be trying to achieve the same effect: preventing future backsliding by staking a position of no ambiguity that it hopes a reviewing court will endorse.58×58. Just one week after King, in Michigan v. EPA, 135 S. Ct. 2699 (2015), the Court, using legal analysis similar to that conducted by the EPA in the CPP Final Rule, determined that a different provision of the Clean Air Act was not ambiguous, see id. at 2707. While Justice Scalia’s majority opinion purported to give deference to the EPA’s interpretation of the statute, the Court found that only one reading (not the EPA’s) was actually reasonable. Id.

In the end, the fate of the CPP will almost certainly be determined by the Supreme Court. While it is impossible to forecast whether the Court will use the CPP to announce new sweeping theories of agency deference, or resolve the issue on a much narrower basis, in the meantime, concerned agencies will likely follow the EPA’s interpretive lead. That may not be cause for concern. If an agency more thoroughly analyzes its statutory text up front — not just looking for ambiguity and a reasonable resolution, but ambiguity and the best resolution — it may arrive at an interpretation whose textual grounding is more clearly articulated and better supported. Moreover, if one presumes that coherence is an important principle in interpreting the law,59×59. Ronald Dworkin, Law’s Empire 19–20 (1986). it seems preferable for agencies to read their statutes as such. Justice Sotomayor’s dissent in Scialabba, quoted in the Final Rule, discussed what she called “a fundamental tenet of statutory interpretation: We do not lightly presume that Congress has legislated in self-contradicting terms.”60×60. Scialabba, 134 S. Ct. at 2220 (Sotomayor, J., dissenting), quoted in Final Rule, supra note 1, at 64,713. These two considerations suggest that uncertainty about deference after King may motivate more robust agency interpretation.