On November 3, 1999, U.S. Environmental Protection

Transcription

1 Ten Years of New Source Review Enforcement Litigation Margaret Claiborne Campbell and Angela Jean Levin On November 3, 1999, U.S. Environmental Protection Agency (EPA) Administrator Carol Browner and Attorney General Janet Reno jointly announced what they called an unprecedented action. On that day, EPA filed separate civil enforcement actions against seven electric utility companies, as well as an administrative compliance order against the EPA sister agency the Tennessee Valley Authority (TVA). EPA s unprecedented action alleged widespread violations by the utilities and TVA of the federal New Source Review (NSR) program under the Clean Air Act (CAA). The allegations spanned violations at thirty-two power plants across ten states that, in some cases, dated back more than twenty years. So began the largest, most contentious industry-wide enforcement initiative in EPA history. While EPA has undertaken similar NSR enforcement actions against other industry sectors including the wood products and petroleum refining industries no other initiative has produced so much litigation and new law like EPA s actions against the electric utilities. Not since the Seventh Circuit s decision in Wisconsin Electric Power Co. v. Reilly (WEPCO) and the First Circuit s 1990 decision in Puerto Rican Cement have so many NSR issues been addressed by the federal courts in the history of the thirty-five-year-old program. The implications of the more recent decisions reach well beyond the electric utility industry. In recognition of the ten-year anniversary of EPA s unprecedented action, this article provides background on some of the major legal issues at play in EPA s electric utility NSR enforcement initiative and reviews the major decisions to date. The NSR enforcement actions against the electric utilities all involve virtually identical allegations. In each case, EPA alleges that the replacement of parts, typically boiler components or portions of components, at existing electric generating units amounted to major modifications of those units, triggering new source permitting and regulatory requirements. According to EPA, failure to obtain preconstruction permits constitutes a continuing violation, rendering ongoing operation of the units unlawful. To remedy the alleged violations, EPA seeks injunctive relief installation of state-ofthe-art emission controls and civil penalties on a per-day, per-violation basis. As a result, the companies face potentially significant exposure. Ms. Campbell is a partner and Ms. Levin is an associate with the Troutman Sanders Environmental & Natural Resources Practice Group in Atlanta. Both are actively involved in the defense of companies in EPA s electric utility NSR enforcement initiative. The NSR rules define major modification as a physical or operational change that results in a significant net emissions increase. 40 C.F.R (b)(2). The rules exclude a number of important activities from this definition, however, including routine maintenance, repair and replacement (RMRR), increases in hours of operation, and increases in the production rate. 40 C.F.R (b)(2)(iii)(a) (k). As a result, most of the key legal battles in the electric utility enforcement initiative have centered on the major modification definition from the scope of the RMRR exclusion to the proper test for determining emissions increases and the role of causation. In addition, the actions to date have addressed integral tenets of regulatory jurisprudence, including the applicability and effect of the fair notice doctrine, the role of agency deference in the context of an enforcement action, as well as the threshold question of the applicable statute of limitations. Because the CAA does not contain its own statute of limitations, the general federal five-year statute of limitations at 42 U.S.C serves to limit EPA s ability (as well as that of state agencies and private citizens) to collect penalties for alleged NSR violations. In addition, the concurrent remedy doctrine can prohibit untimely efforts by private citizens, and perhaps the federal government, to obtain injunctive relief in an NSR case. See, e.g., National Parks Conservation Ass n v. TVA (NPCA I), 502 F.3d 1316, 1327 (11th Cir. 2007) (holding that five-year statute of limitations bars equitable relief in NSR citizen suit); see also, Fed. Election Comm n v. Williams, 104 F.3d 237, 240 (9th Cir. 1996) (holding that federal government s claims for injunctive relief were barred because equity will withhold its relief... where the applicable statute of limitations would bar the concurrent legal remedy ). Because many of the alleged electric utility violations involve projects performed one to two decades earlier, the applicability of the five-year statute of limitations has become an important threshold issue in many cases. Early district court decisions disagreed on this issue. Some concluded that the failure to obtain a preconstruction permit under the NSR program constituted a one-time violation. See, e.g., United States v. Southern Indiana Gas and Electric Co. (SI- GECO), 2002 U.S. Dist. LEXIS 14040, *10 (July 26, 2002); see also, United States v. Westvaco, 144 F. Supp. 2d 439 (D. Md. 2001). Others held that the regulations create an ongoing obligation, making each day of noncompliance a new violation. See, e.g., United States v. Duke Energy Corp., 278 F. Supp. 2d 619 (M.D.N.C. 2003). For a more complete discussion of these decisions, see Ivan Leiben, Catch Me If You Can The 16 NR&E Spring 2010

2 Misapplication of the Federal Statute of Limitations to Clean Air Act PSD Permit Program Violations, 38 Envtl. L. 667, (Summer 2008). A recent decision by the Eleventh Circuit may, however, have laid this debate to rest. In NPCA I, the Eleventh Circuit upheld the district court s holding that the five-year statute of limitations barred NSR claims brought by NPCA and the Sierra Club against TVA for a project performed in 1982 at one of TVA s coal-fired power plants in Alabama. 502 F.3d at The Eleventh Circuit focused on the language of the CAA and, significantly, the language of Alabama s State Implementation Plan (SIP), which represents Alabama s incorporation of the CAA and associated regulations into its state law. Under the Alabama SIP, BACT for a source is not an independent, ongoing requirement. It does not exist until it is determined by the Director on a case-bycase basis for a particular project before construction begins. See Ala. Admin. Code r (2)(l). Thus, the court held that provisions of the Alabama SIP compel our conclusion that TVA s obligation to install Best Available Control Technology was to be met at construction time in 1982, and was not an ongoing duty. 502 F.3d at Therefore, the citizens groups claims for civil penalties, which came more than five years after construction, were barred by the five-year statute of limitations. In addition, the court agreed with TVA that the claims for injunctive relief were also barred under the concurrent remedies doctrine, which provides that where a party s legal remedies are time-barred, that party s concurrent equitable claims generally are barred as well. Id. at In arriving at its conclusion, the Eleventh Circuit directly addressed a seemingly contrary conclusion reached by the Sixth Circuit in a parallel case, also entitled National Parks Conservation Ass n v. TVA (NPCA II). There, the Sixth Circuit evaluated similar allegations by the same plaintiffs against TVA involving a plant in Tennessee but concluded that the five-year statute of limitations did not bar the plaintiffs claims. 480 F.3d 410, 419 (6th Cir. 2006). Although superficially in conflict, both circuit courts focused on the provisions of the respective state SIP in arriving at their conclusions. In this context, the Eleventh Circuit reconciled its conclusion with that of its sister circuit by recognizing that the Tennessee SIP imposed continuing obligations on entities that failed to obtain a preconstruction permit, where Alabama s SIP did not. 502 F.3d at Specifically, unlike Tennessee, Alabama limit[s] the obligation to apply [Best Available Control Technology] to proposed modifications, with no caveat continuing the obligation for the operating life of the source if it was not met during the construction phase. Id. at Consequently, future courts addressing this issue likely will focus on the terms of the relevant state s SIP to evaluate the applicability of the five-year statute of limitations to alleged NSR violations. Unlike the case law regarding applicability of the five-year statute of limitations, where the courts decisions may be facially inconsistent but are substantively consistent, in the RMRR context courts have come to widely divergent conclusions. Remarkably, the courts generally identify the same relevant standard. From there, the devil lies in the details. Each court has had to confront the issue whether the named projects were properly excluded from NSR as routine maintenance, repair, and replacement projects. The seminal decision from which all courts begin their analysis is WEPCO, in which the Seventh Circuit affirmed EPA s reliance on a four-factor test that analyzes a given project s (1) nature and extent, (2) purpose, (3) frequency, and (4) cost. Courts generally agree that this inquiry is fact-specific and must be evaluated on a case-bycase basis. See, e.g., U.S. v. East Kentucky Power Coop., 498 F. Supp. 2d 976, (E.D. Ky. 2007) (EKPC). A fundamental dispute between EPA and utility defendants over the RMRR defense has been the proper context for evaluating the named projects. EPA has argued that the WEPCO four-factor test centers on what is routine at the particular emissions unit involved. The electric utilities have noted, on the other hand, that EPA historically interpreted the exclusion with reference to what is routine in the utility industry as a whole, thus requiring each factor of the WEPCO test (nature and extent, purpose, frequency, and cost) to be measured against an industry-wide yardstick. In general, courts that have focused on EPA s historical interpretations No other initiative has produced so much litigation and new law like the Environmental Protection Agency s actions against the electric utilities. and pronouncements prior to the 1999 enforcement initiative have agreed with utilities. See, e.g., U.S. v. Alabama Power Co., 2008 U.S. Dist. LEXIS 58866, *53 (N.D. Ala. July 24, 2008) ( [T]he court does not believe that EPA can, in an enforcement action filed in 1999, ignore what it said and did back in the 1970 s, 1980 s, and 1990 s up until the filing of this action, in concluding that the four-factor WEPCO test must be applied with reference to the industry as a whole, not just the particular... unit at issue ); see also U.S. v. Duke Energy Corp., 278 F. Supp. 2d 619, 637 (M.D.N.C. 2003) ( Through the EPA s statements in the Federal Register, its statements to the regulated community and Congress, and its conduct for at least two decades the EPA has established an interpretation of RMRR under which routine is judged by reference to whether a particular activity is routine in the industry. ). In EKPC, the court noted that when an agency has interpreted one of its regulations in a consistent manner, that interpretation is controlling, unless plainly erroneous, but the court found that EPA s interpretation had been anything but consistent; therefore, the routine test with reference to industry would be applied. 498 F. Supp. 2d at 993. Those courts that have ignored EPA s previous pronouncements, however, have all but NR&E Spring

3 written the exclusion out of EPA s regulations. See, e.g., U.S. v. Ohio Edison, 276 F. Supp. 2d 829 (S.D. Ohio 2003) (finding that the RMRR provision is an exemption that must be construed narrowly and that broadly construing the exclusion to include those projects that are commonly undertaken in the utility industry would swallow the rule and is contrary to the statute s broad definition of modification ); see also New York v. American Electric Power Service Corp., 2007 U.S. Dist. LEXIS 10700, *8 *9 (S.D. Ohio, Feb. 15, 2007) (rejecting the argument that industry practice is dispositive and stating [to] consider only industry practices would be to ignore the plain language of the statutory regulations by judicially erecting a curious scheme in which the regulated determine the applicability of the regulations without an indication of supporting legislative intent ). These courts either rejected the argument or failed to understand that each factor of the WEPCO test Most of the projects named in the complaints are undertaken with such frequency that EPA s allegations effectively create industry-wide noncompliance. must be applied with reference to the industry standard. This determination is not without consequence. Although no court that has adopted the industry-focused version of the RMRR exclusion has yet to apply that standard on the merits, the courts focusing heavily on the unit at issue have uniformly found the RMRR exclusion inapplicable to the projects at issue. See, e.g., U.S. v. Cinergy Corp., 495 F. Supp. 2d 909, 936 n.14 (finding that none of the projects at issue fell within the RMRR exclusion and stating that [e]ven if Cinergy s exhibits could be construed as showing frequency in the industry for any of the projects at any of the plants under scrutiny... the Court would still conclude that all of the projects... were infrequently performed at the unit... and, more importantly, that consideration of all of the other factors together would still clearly demonstrate that none of the projects were routine ). Another issue surrounding the RMRR exclusion is the procedural question of which party bears the burden of proving the exclusion s applicability. The only decision to date that has placed the burden on EPA is the district court in Duke Energy. 278 F. Supp. 2d at 639 (stating that to place the burden on the utility would sanction an almost de facto presumption of a [NSR] violation whenever a utility performs any type of work.... Congress, however, did not provide a presumption or anything approaching a presumption ). Every other court that has addressed this issue has placed the burden on the utility. See, e.g., EKPC, 498 F. Supp. 2d at (summarizing case law to date, including Duke Energy, and concluding that there is no reason that the general rule should not apply.... Once EPA has proven that a physical change and net emissions increase has occurred, the burden shifts EKPC to prove that its activities are exempt from the definition of modification because they were routine ). The courts have also considered the proper test for determining emissions increases under the NSR program. The initial debate over the proper method for calculating emissions increases centered on whether increases should be measured in the first instance in terms of an emission rate (lbs/hour) as required under the New Source Performance Standards (NSPS) program (Section 111 of the CAA) or simply in total annual tons. In 2003, Judge Frank W. Bullock, Jr. in the Middle District of North Carolina agreed with Duke Energy that because EPA had defined modification under the NSPS program to mean increases in the maximum hourly emission rate, and because Congress had later used the term modification in the NSR preconstruction permitting program, Congress must have intended for modification to have the same meaning in the NSR program as EPA had given it in the NSPS program. 278 F. Supp. 2d at 644 (stating that the explicit reference by Congress incorporating the concept of NSPS modification into the [NSR] concept of construction compels the result that [NSR] is triggered only by an increase in a unit s hourly emissions rate ). Thus, Congress must have intended preconstruction permitting under NSR similarly to be triggered solely by increases in the maximum hourly emissions rate rather than increases in the annual rate of emissions from a facility. Id. On appeal of Judge Bullock s decision, the Fourth Circuit went even further, finding that EPA had no discretion to adopt a different meaning of the term modification in two different portions of the CAA. U.S. v. Duke Energy, 411 F.3d 539, 547 (4th Cir. 2005) (stating that because Congress mandated that the [NSR] definition of modification be identical to the NSPS definition of modification, the EPA cannot interpret modification under the PSD inconsistently with the way it interprets that term under the NSPS ) (emphasis added). Controversy over this decision quickly ensued, however, when the D.C. Circuit arrived at the opposite conclusion a mere nine days later in New York v. EPA, 413 F.3d 3, 13 (D.C. Cir. 2005) (upholding EPA s interpretation of modification in the NSR program as referring to annual increases in emissions); see also U.S. v. Cinergy, 458 F.3d 705, 710 (7th Cir. 2006) (stating that [t]he same word can mean different things in the same statute and agreeing with the D.C. Circuit s decision in New York v. EPA). The U.S. Supreme Court granted certiorari in Duke Energy to settle the dispute. Numerous articles have discussed the Supreme Court s decision in great detail. See, e.g., Benjamin E. Edwards, Environmental Defense v. Duke Energy Corporation, 33 Colum. J. Envtl. L. 197 (2008). In short, the Supreme Court concluded that while the statute did not require EPA to use a different test for purposes of NSPS and NSR permitting purposes, the regulations currently provide for different 18 NR&E Spring 2010

4 tests. According to the Court, the regulatory language simply cannot be squared with a regime under which hourly rate of emissions... is dispositive. Environmental Defense v. Duke Energy, 549 U.S. 561, 578 (2007). Early cases resolving alleged NSR violations glossed over all other legal issues related to emissions increases (likely in part due to the fundamental dispute between the parties regarding whether annual or hourly emissions increases were the relevant inquiry). For example, in Ohio Edison, once the Southern District of Ohio ruled in EPA s favor that annual emissions were the correct standard by which to measure the utility s projects, the court appeared wholeheartedly to adopt EPA s expert witnesses methods of calculating emissions increases. See, e.g., Ohio Edison, 276 F. Supp. 2d at 869 (quoting extensively from the report of EPA s expert, Dr. Rosen, and stating that [i]ncreased availability leads to more hours of operation. In turn, more hours of operation results in increased emissions, as Dr. Rosen s calculations show ). To prove that the utilities should have projected annual emissions increases resulting from the challenged activities, EPA has relied on expert testimony (as illustrated by Dr. Rosen s testimony in Ohio Edison) that reflects an ever-changing method of calculating emissions increases. Indeed, it has become clear that the methodology used by EPA s experts was developed during the enforcement initiative and refined with time. While one or two courts simply accepted EPA s approach without question, more recent case law has begun to question EPA s approach. The Western District of Pennsylvania said it best: [E]ven if [EPA s expert s] retrospective methodology demonstrates that [the defendant] might have projected a significant net increase, [EPA s expert s] opinion is insufficient to establish that all reasonable methodologies must have projected a significant net increase such that Defendants failure to obtain a permit at the time was unreasonable. Commonwealth of Pennsylvania v. Allegheny Energy, Inc., 2008 U.S. Dist. LEXIS 93800, *21 *22 (W.D. Pa. Nov. 18, 2008). Indeed, in Cinergy, the court adopted a more commonsense, reasonable-power-plant-engineer test to determine whether a project should have been expected to result in an emissions increase. Cinergy, 458 F.3d at 710 (7th Cir. 2006) ( What is required for determining whether a construction permit must be sought for a planned physical change in the plant is not prescience, but merely a reasonable estimate of the amount of additional emissions that the change will cause. ). Applying the reasonable engineer test to the facts of that case, only six out of numerous projects named by EPA were found to be modifications subject to NSR permitting requirements. Nonetheless, EPA continues to rely heavily on expert testimony and retrospective projections based on increased unit availability following the projects to prove emissions increases in each case. In addition to the substantive NSR legal issues, the courts have also addressed more fundamental regulatory issues, including fair notice and the role of agency deference in the context of an enforcement action. The utility industry has contended throughout the NSR litigation that EPA s interpretations of the RMRR exclusion and theories regarding emissions increases in this litigation constitute significant departures from the Agency s historical interpretations and public statements. As a result, the industry had no fair warning of how the regulations were going to be applied in the enforcement initiative. In fact, Duke Energy asserted counterclaims against EPA for its failure to undertake notice-and-comment rulemaking for its new interpretations. See U.S. v. Duke Energy Corp., 2004 U.S. Dist. LEXIS 8724, *10 (M.D.N.C. Apr. 14, 2004). Thus far, courts have generally been unreceptive to this defense. See, e.g., SIGECO, 245 F. Supp. 2d 994, 1011 (S.D. Ind. 2003) (finding that EPA s Clay Memorandum and the WEPCO decision put the regulated community on notice that the routine maintenance exemption was a multi-factor test, and that no one factor would have dispositive weight ); Cinergy, 495 F. Supp. 2d at 906 (stating that [a]s long as Cinergy was aware of the regulatory standards for determining whether a project may result in significant increases in emissions, its understanding of the exact mathematical formula is irrelevant ). While unwilling to apply the fair notice doctrine, courts have been willing to consider the level of deference due to EPA s litigating positions. Traditionally, courts grant EPA significant deference for its interpretations of its own regulations. Thomas Jefferson Univ. v. Shalala, 512 U.S. 504, 512 (1994). However, when, as here, the regulatory agency takes an inconsistent view of the regulations, makes inconsistent statements with respect to the regulation, and also enforces the regulation with no discernible consistency (which was the situation at least as of the time the work at issue in this case commenced), the weight to be given that position diminishes considerably in the Court s view. EKPC, 498 F. Supp. 2d at 993. The EKPC court made this statement in the course of concluding that the relevant focus for the RMRR exclusion was not, as EPA maintained during the litigation, solely the unit at issue, but rather, with reference to the industry as a whole. Id. Other courts similarly appear willing to evaluate and potentially refuse to grant EPA deference in this context. See, e.g., Alabama Power, 2008 U.S. Dist. LEXIS at *56 (stating that Alabama Power may offer evidence... [regarding the] amount of deference the EPA is entitled to in this action at trial). One might ask why the utility industry enforcement initiative has resulted in so much litigation as compared to other industries. As in any case, many factors influence the ultimate decision whether to litigate. An important factor in the electric utility enforcement cases, however, has been the fact that the vast majority of these cases involve full or partial component replacement projects that are commonly performed across the industry as part of regularly scheduled maintenance outages in an effort to maintain the reliability of existing units. Indeed, most of the projects named in the complaints are undertaken with such frequency that EPA s allegations effectively create industry-wide noncompliance. Moreover, these enforcement actions have come at a time NR&E Spring

5 when the industry is investing billions of dollars in its existing units ironically, to install new emission controls under other provisions of the CAA to reduce the very pollutants at issue in these cases. While a number of companies have settled, many others have simply found it too difficult to accept EPA s litigation interpretation of the rules because doing so could effectively prevent them from continuing to maintain their existing units without going through NSR anew in advance of almost every regularly scheduled maintenance outage. As we marked the tenth anniversary of the first enforcement actions in the electric utility industry initiative, EPA was issuing numerous new information requests and Notices of Violation and filing new complaints in district courts against electric utilities around the country. EPA s focus, however, is not limited to electric utilities. EPA is expanding the reach of its NSR enforcement efforts to other industries. At the American Bar Association Section of Environment, Energy, and Resources September 2009 Environment, Energy, and Resources Law Summit in Baltimore, the Director of EPA s Office of Civil Enforcement, Adam Kushner, reported that EPA sees rampant noncompliance with the NSR regulations. According to Kushner, the Agency is actively evaluating numerous industries, including PVC manufacturers, acid producers, cement plants, glass manufacturers, carbon black producers, oil and gas producers, polystyrene foam, landfills, industrial boilers, iron and steel manufacturers, natural gas transmission, aluminum smelters, municipal waste combustors, ethanol producers, wood products, and pulp and paper facilities. Future enforcement actions will be shaped and informed by EPA s experience and the law established in this decade-old utility-industry litigation. 24 NR&E Spring 2010