United States v. Allegheny Ludlum

Opinions of the United 2004 Decisions States Court of Appeals for the Third Circuit 4-28-2004 USA v. Allegheny Ludlum Precedential or Non-Precedential: Precedential Docket No. 02-4346 Follow this and additional works at: http://digitalcommons.law.villanova.edu/thirdcircuit_2004 Recommended Citation "USA v. Allegheny Ludlum" (2004). 2004 Decisions. Paper 735. http://digitalcommons.law.villanova.edu/thirdcircuit_2004/735 This decision is brought to you for free and open access by the Opinions of the United States Court of Appeals for the Third Circuit at Villanova University School of Law Digital Repository. It has been accepted for inclusion in 2004 Decisions by an authorized administrator of Villanova University School of Law Digital Repository. For more information, please contact Benjamin.Carlson@law.villanova.edu. ROBERT H. MILLER PRECEDENTIAL JOHN SITHER KATHRYN E. KOVACS (Argued) IN THE UNITED STATES COURT OF U.S. Department of Justice APPEALS Environment & Natural Resources FOR THE THIRD CIRCUIT Division, Appellate Section _________________________ P.O. Box 23795 L’Enfant Plaza Station Washington, DC 20026 NO. 02-4346 _________________________ KERRY NELSON LORI G. KIER UNITED STATES OF AMERICA U.S. Environmental Protection Agency Office of Regional Counsel v. 1650 Arch Street Philadelphia, Pennsylvania 19103 ALLEGHENY LUDLUM CORPORATION, Counsel for Appellee Appellant H. WOODRUFF TURNER (Argued) __________________________ JOHN E. BEARD, III THOMAS J. SMITH On Appeal from the United States TODD R. BROWN District Court Kirkpatrick & Lockhart, LLP For the Western District of Pennsylvania Henry W. Oliver Building (D.C. No. 95-cv-00990) 535 Smithfield Street District Judge: Honorable Robert J. Pittsburgh, Pennsylvania 15222 Cindrich __________________________ Counsel for Appellant Argued December 16, 2003 ____________________________ Before: ALITO, FUENTES and OPINION OF THE COURT BECKER, Circuit Judges ____________________________ (Filed April 28, 2004) BECKER, Circuit Judge. This is an appeal from an order of the THOMAS L. SANSONETTI District Court granting judgment for the Assistant Attorney General plaintiff United States and against defendant JOHN T. STAHR Allegheny Ludlum Corporation (“ALC”) in NANCY FLICKINGER an action brought for violations of the Clean Water Act (“CWA” or the “Act”) at five of ensure that polluters will take responsibility AL C’s We ste r n P e n n sy lv a n ia for ensuring the correct and precise manufacturing facilities. The judgment is measurements of their waste (which they are multifaceted, flowing from: (1) pretrial obliged to certify), we do not believe that a legal determinations by the Court; (2) a laboratory error defense—where the error jury verdict on a number of liability issues; resulted in overreporting—is inconsistent and (3) determinations by the Court with this regime. Rather, inasmuch as the following a penalty hearing. The jury penalty imposed is for an unlawful verdict was mixed; each side prevailed on discharge and not for faulty reporting, we a number of issues, and ALC’s appeal think that deprivation of the defense would leaves unchallenged significant portions of not advance the purpose of the CWA and the judgment against it. However, the that it would be grossly unfair, especially in appeal does challenge major aspects of the view of the presence of companion judgment and also of the civil penalty provisions of the CWA imposing liability assessment leveled against ALC for the for monitoring and reporting violations. We alleged violations in the sum of will therefore vacate the judgment in part $8,244,670. and remand so that the laboratory error defense can be considered and adjudicated The first important question with respect to the affected claims. presented by the appeal concerns the viability of the so-called “laboratory error The appeal also requires us to defense.” The CWA operates under a self- determine whether the District Court made monitoring and reporting system whereby either a mistake of law or abused its the discharger of toxic waste measures and discretion in calculating the economic reports to the Environmental Protection benefit that ALC obtained from those Agency (“EPA”) the volume of its violations that are unchallenged on appeal. discharge. ALC maintains that the EPA Section 1319(d) of the CWA requires that predicated certain aspects of the violation the District Court, when determining the upon reports submitted by ALC that were amount of a civil penalty under the CWA, tainted by laboratory error caused by a consider “the economic benefit (if any) contaminated reagent resulting in resulting from the violation,” so as to “level overreporting of the amount of the toxic the playing field.” The District Court’s zinc discharge. The District Court calculation here was an agglomeration, declined to allow the laboratory error based on a number of factors. The largest defense on the grounds that it had not been single factor was the 12.73% interest rate recognized in the Third Circuit, and that to used by the government and the District allow such a “new defense” would Court to compute interest from the date of contravene the CWA. violation to the date of the judgment so as to calculate the total economic benefit to ALC. Although the CWA operates under This rate was predicated largely on a a regime of strict liability, designed to 2 calculation of ALC’s weighted average District Court’s application of the other cost of capital (“WACC”). Noting that it legally required factors to calculate ALC’s was uncontested at trial that ALC had an economic benefit—the least costly method actual rate of return on capital that was of compliance and the periods of non- less than half the 12.73% rate used by the compliance—were supported by the record. District Court, ALC contends that the In the course of this determination, we 12.73% rate is excessive. clarify that the proper method for determining economic benefit is to base the We conclude that the application of calculation on the least costly method of the 12.73% rate may so vastly overstate compliance. On the issue of economic the economic benefit to ALC of its benefit, we therefore vacate and remand improper discharges, that it does not “level with respect to the interest rate issue. the playing field,” and that it constitutes an abuse of discretion. As a prelude to Finally we must decide whether, in making this determination we explore the compiling the number of violations for the potential ramifications of the notion of purpose of assessing a penalty, the District economic benefit under § 1319(d). We Court erred by counting violations of conclude that there are two possible monthly averages as violations for each day approaches to calculation of economic of the month. We, of course, follow our benefit: (1) the cost of capital, i.e., what it precedent in Natural Resources Defense would cost the polluter to obtain the funds Council, Inc. v. Texaco Refining & necessary to install the equipment Marketing Inc., 2 F.3d 493 (3d Cir. 1993), necessary to correct the violation; and (2) that the daily average limit is computed by the actual return on capital, i.e., what the averaging effluent levels only for days on polluter earned on the capital that it which the facility operated. Although some declined to divert for installation of the Courts—most notably the Fourth Circuit in equipment. Because these factors are so Chesapeake Bay Foundation, Inc. v. varia ble, depending upon market Gwaltney of Smithfield, Ltd., 791 F.2d 304 conditions and the financial soundness of (4th Cir. 1986), vacated on other grounds, the polluter, we leave it to the District 484 U.S. 49 (1987)—have held that a Court, in the sound exercise of its violation of a monthly average parameter discretion, to decide which approach to constitutes a violation for each day of the apply and how to apply it (there are a month, we find this approach incomplete. variety of models). However, we explain We adopt Gwaltney insofar as it establishes why the District Court’s application of the an absolute upper bound on the penalty that WACC in this case was, at a minimum, can be assessed for a monthly average unsupported by the evidence, and needs to violation. However, permit limits can be be recalculated should the District Court exceeded in many different ways, both by on remand elect to pursue that approach. very large, isolated discharges and by m o d e r a t e c o n t i n uo u s d i s c h ar g e s . In contrast, we conclude that the 3 Furthermore, daily and monthly average a considerable amount of pollution. ALC’s limits are designed to avoid distinct steel-making process uses water from environmental harms. As a result, in some adjacent rivers. The water is used as cases a violator’s wrongful conduct will process water and as non-contact cooling merit punishment for both daily and water. Process water is used directly in the monthly violations, while in others, the process of making steel, and makes contact conduct will have been sufficiently with steel or steel-making equipment. Non- punished by penalties for daily violations contact cooling water cools the steel-making alone. We hold that district courts have equipment without actually touching the discretion to determine, on the facts of steel. ALC operates six on-site wastewater each case, how many violation days should treatment plants (“WWTPs”) at these be assessed for penalty purposes for the facilities. The three WWTPs at the violation of a monthly average limit, based Brackenridge facility discharge to the on whether violations are already Allegheny River, pursuant to a National sufficiently sanctioned as violations of a Pollution Discharge Elimination System daily maximum limit. In this case, the ( “N PDES”) perm it issued by the District Court did not have the benefit of Pennsylvania Department of Environmental this standard, so we will vacate its penalty Protection (“PADEP”). The two WWTPs at assessment and remand for further the West Leechburg facility discharge to the proceedings. Kiskiminetas River pursuant to another NPDES permit. The Vandergrift WWTP We will therefore affirm in part, discharges treated process waters to the vacate in part, and remand for further Kiski Valley Water Pollution Control proceedings consistent with this opinion. Authority (“KVW PCA”) pursuant to I. Facts and Procedural History permits with it. After applying further treatment, KVWPCA discharges to the ALC manufactures steel and owns Kiskiminetas River. and operates five plants comprising three specialty steel manufacturing facilities in The United States filed this action Western Pennsylvania: the Brackenridge against ALC on June 28, 1995. The Facility (the Brackenridge and Natrona Complaint, as amended, alleged three types plants); the West Leechburg Facility (the of violations: (1) discharges at each of West Leechburg and Bagdad plants); and ALC’s five facilities containing discharges t h e V a n d e r g r if t F a c il it y. T he in excess of ALC’s permits as shown by the Brackenridge Facility conducts melting, Discharge Monitoring Reports (“DMRs”) continuous casting, rolling, and finishing submitted to the EPA; (2) discharges from operations. The West Leechburg and the Vandergrift facility that interfered with Vandergrift Facilities are finishing the operations of the Kiski Valley WPCA; operations. and (3) ALC’s failure to report violations as required by its permits. The parties filed The steelmaking process generates 4 cross motions for summary judgment. In that the defense had not been recognized in response, the District Court ruled that ALC this Circuit, and that it would not adopt such could not raise several defenses to the a new defense, “especially since the Act can reported violations, includin g the be interpreted as creating an obligation to “laboratory error” defense by which ALC insure that the self-monitoring of pollutants contended that its reported violations is accurate, assigning the risk of inaccuracy resulted from erroneous laboratory to the company.” The Court thus granted analyses—later discovered to be caused by partial summary judgment to the United a contaminated reagent— which overstated States on that issue. The Court denied the zinc pollutant levels.1 The Court opined governm ent’s motion for summary judgment on the reporting failure and interference claims, finding that ALC had 1 ALC’s overreporting of zinc provided sufficient evidence to create triable exceedences was based upon effluent issues of fact. sample analyses performed by ALC’s The District Court held a jury trial on Technical Laboratory which turned out to liability from January 5 to February 2, 2001. be flawed. ALC allegedly tried to The jury found in favor of ALC on all of the determine the cause of the zinc interference and reporting failure claims, but exceedences, without success. In in favor of the government on half of the February 1996, it started to examine its remaining reported violations claims. In own laboratory’s performance. ALC total, the violations for which ALC took samples of effluent and had part stipulated to liability, those for which the analyzed at the ALC Laboratory and part court granted summary judgment, and those analyzed by two outside laboratories, a for which the jury returned a verdict against protocol known as “split sampling.” ALC added up to 1,122 days of violations According to ALC, the zinc results obtained by its laboratory were significantly higher than those obtained and analyzed by the outside laboratories at the outside laboratories, while the showed significantly lower zinc results outside laboratories’ results were than the corresponding results from ALC’s consistent with each other. In this split laboratory. However, the sample sets sampling, each laboratory performed its analyzed by the outside laboratories after own digestion of the samples. In March, the samples were predigested by ALC’s 1996, ALC again split samples with the laboratory were as high in zinc as the two outside laboratories, but this time results from ALC’s laboratory, leading provided each laboratory with two ALC to the conclusion that it was the sample sets, one undigested and one digestion process in ALC’s laboratory that predigested by ALC’s Laboratory. was causing zinc values to be overstated. According to ALC, the values generated Once ALC switched to a different reagent, by the samples that were both digested it no longer reported zinc exceedences. 5 from July 1990 through February 1997. II. The Laboratory Error Defense From February 5 to 8, 2001, the A. Overview of the Clean Water Act Court conducted a bench trial on the The Clean Water Act (“CWA”) was penalty amount. To save time, the Court enacted by Congress in 1972 “to restore and allowed the experts to give their direct maintain the chemical, physical, and testimony in the form of written proffers, biological integrity of the Nation’s waters.” and allowed live cross-examination. 33 U.S.C. § 1251(a). In order to achieve Following the penalty trial, the parties this goal, the CWA prohibits the discharge submitted proposed (judicial) opinions, of any pollutant into waters of the United and on February 20, 2002, the Court filed States except as expressly authorized under an opinion and entered judgment against the Act. See 33 U.S.C. § 1311(a). In order ALC in the amount of $8,244,670. ALC to discharge pollutants into navigable filed a motion under Fed. R. Civ. P. 59(e) waters, one must obtain a National Pollution to alter or amend the judgment, which the Discharge Elimination System (“NPDES”) District Court denied. On November 26, permit. 33 U.S.C. § 1342. Discharges that 2002, ALC filed a notice of appeal from comply with the limits and conditions in an the District Court’s summary judgment NPDES permit are deemed to comply with order of September 28, 2000, the the Act. 33 U.S.C. § 1342(k). The CWA reconsideration order of November 28, requires NPDES permittees to test their 2000, the final judgment of February 20, effluent and report the results to the EPA in 2002, and the Rule 59(e) order of October Discharge Monitoring Reports (“DM Rs”). 8, 2002. 33 U.S.C. § 1318(a); 40 C.F.R. §§ 122.41(j), The District Court had jurisdiction 122.48. Section 307 of the CWA authorizes pursuant to 28 U.S.C. § 1331. ALC’s the EPA to promulgate regulations appeal is timely under Fed. R. App. P. prohibiting the discharge of any pollutant 4(a)(1)(B), and we have appellate into a Publicly Owned Treatment Works jurisdiction pursuant to 28 U.S.C. § 1291. (“POTW”) that “interferes with, passes Our review of the grant of summary through, or otherwise is incompatible with” judgment is plenary. See Shelton v. Univ. the POTW. 33 U.S.C. § 1371(b)(1). The of Med. & Dentistry of N.J., 223 F.3d 220, Act prohibits discharges to POTWs that are 224 (3d Cir. 2000). We review the in excess of those pretreatment standards. imposition of a penalty under Section 33 U.S.C. § 1317(d). The EPA has issued 1319(d) of the CWA for abuse of general pretreatment standards and national discretion, see Tull v. United States, 481 categorical pretreatment standards for the U.S. 412 (1987), but our review of the iron and steel manufacturing industry. See legal construction of Section 1319(d) is 40 C.F.R. Pts. 403, 420. plenary, see Public Interest Research The Act authorizes the EPA to bring Group, Inc. v. Powell Duffryn Terminals, civil enforcement actions for injunctive Inc., 913 F.2d 64, 80 (3d Cir. 1990). 6 relief and penalties, at times relevant, to up government then argues that, consistent with to $25,000 per day for each violation. See the Act’s requirement for accurate self- 33 U.S.C. § 1319(d). A violation of the reporting, courts should treat DMRs, which Act can be established by showing that the must be certified by the discharger, as defendant is a person who discharged admissions that are sufficient to establish pollutants from a point source into liability under the CWA. navigable waters in violation of the terms The government relies in this respect of the applicable NPDES permit or into a on Sierra Club v. Union Oil Co., 813 F.2d POTW in violation of a pretreatment 1480, 1491-92 (9th Cir. 1987), vacated on standard. See 33 U.S.C. §§ 1311, 1317(d). other grounds, 485 U.S. 931 (1988), where In assessing a civil penalty for a violation the Court of Appeals held that a CWA of § 1311 or § 1317, the court must defendant could not escape liability based consider: “the seriousness of the violation on alleged sampling violations. The Court or violations, the economic benefit (if any) no ted that the “N PD ES p rogra m resulting from the violation, any history of fundamentally relies on self-monitoring” such violations, any good-faith efforts to and that Congress deemed accurate DMRs comply with the applicable requirements, “critical to effective operation of the Act.” the economic impact of the penalty on the Id. It opined that allowing CWA permittees violator, and such other matters as justice to impeach their own DMRs “would be may require.” 33 U.S.C. § 1319(d). sanctioning countless additional hours of B. The Government’s Contentions NPDES litigation and creating new, complicated factual questions for district The government argues that the courts to resolve.” Id. at 1492. The Court CWA establishes a scheme of strict further reasoned that if permittees could liability aimed at facilitating enforcement. impeach their own reported violations with It first notes that Congress gave the EPA claims of laboratory error, it would “create the “authority to require information, data, the perverse result of rewarding permittees and reports, as well as establish monitoring for sloppy laboratory practices” and requirements,” recognizing that such an “undermine the efficacy of the self- authority is a “necessary adjunct to the monitoring program.” Id.; accord Conn. establishment of effective water pollution Fund for the Env’t, Inc. v. Upjohn Co., 660 requirements and the enforcement of such F. Supp. 1397 (D. Conn. 1987). requirements.” Government Br. at 16 (citing S. Rep. 92-414, at 62 (1971)). Relying on this reasoning, the Furthermore, it points out that Congress government submits that we should reject intended “these new requirements” to ALC’s laboratory error defense. Because “avoid the necessity of lengthy fact the regulations require dischargers to amend finding, investigations, and negotiations at their sworn DMRs whenever they discover the time of enforcement.” Id. (citing S. an error in their reporting, and because Rep. 92-414, at 62 (1971)). The failure to do so constitutes a criminal 7 violation in and of itself, the government between the defendant’s results and those contends that allowing dischargers to from outside laboratories, though no contest their own DMRs conflicts with the consistent pattern could be detected in those statute and the applicable regulations. See discrepancies (sometimes the defendant’s 40 C.F.R. §§ 122.41(k)(2), (1)(8). The results were higher, and sometimes they government further argues that allowing a were lower than the outside laboratories’ laboratory error defense would frustrate results). “congressional intent, would reward The Court explained that “if a companies for inaccurate monitoring defendant wishes to contest the accuracy of practices, and would give them an its DMRs, it ‘has a heavy burden to incentive to wait until they are sued to establish faulty analysis.’” Id. at 1178 ensure the accuracy of their DMRs.” (quoting Student Pub. Interest Research C. The Authorities Relied upon by ALC Group, Inc. v. Georgia-Pacific Corp., 615 F. Supp. 1419, 1429 (D.N.J. 1985)). “The ALC first cou nters th e ‘defendant must present direct evidence of government’s arguments by citing a reporting inaccuracies’ and ‘may not rely on number of cases from district courts within unsupported “speculation” of measurement this Circuit that have recognized—either error.’” Id. (quoting Georgia-Pacific Corp., explicitly or implicitly—the availability of 615 F. Supp. at 1429). The fact that “no the laboratory error defense. While no court in this district ha[d] thus far found a defendant in these cases has actually made defendant to have met this heavy burden,” it past the summary judgment stage based id., however, did not preclude the possibility on the laboratory error defense, that lack of the defense as a matter of law. of success has been due to district courts finding that the defendants failed to raise a The Elf Atochem Court, in discussing genuine issue of material fact as to the the reasoning in Upjohn, quotes Chesapeake existence of a laboratory error, and not Bay Foundation v. Bethlehem Steel Corp., because the defendants were precluded 608 F. Supp. 440, 452 (D. Md. 1985), which from raising the defense as a matter of stated that “‘[g]iven the heavy emphasis on law. accuracy in the Act and the clear Congressional policy that DMRs should be In Public Interest Research Group, used for enforcement purposes, the court Inc. v. Elf Atochem North America, Inc., will not accept claims of inaccurate 817 F. Supp. 1164 (D.N.J. 1993), a similar monitoring as a defense.’” Elf Atochem, 817 case of potential overreporting came F. Supp. at 1179. The Elf Atochem Court before the District Court. The defendant agreed that “the Act places the burden of claimed that errors in its laboratory testing accurately monitoring the levels of had resulted in the overreporting of toxic pollutants in their effluent squarely on the discharges. Split sampling over a six- shoulders of permit holders, and that we month period revealed large discrepancies must hold them to that obligation,” but it 8 ultimately held that ultimately granted summary judgment against the defendant— based on the fact while we agree with the that the cover letters the defendant [Upjohn] court that it is submitted to the Court were too speculative incon sis t e n t with th e in that they merely asserted that the structure and purpose of the defendant “felt” and “believed” that Act to allow permit holders laboratory errors had occurred—it clearly to escape liability altogether implied that had the factual situation been on the basis of laboratory different, Yates could have survived a error, we find it more summary judgment motion based on a accurate, where laboratory laboratory error defense. See id. error has been shown, to hold a defendant liable for a D. Discussion monitoring violation rather We find the reasoning of the Elf than a discharge violation. Atochem Court persuasive. The violations Elf Atochem, 817 F. Supp. at 1179 at issue here alleged that ALC discharged (emphasis added). pollutants in violation of the terms of its permit. In order to prove these violations, it Similarly, in Public Interest was necessary for the government to Research Group, Inc. v. Yates Industries, establish that ALC did in fact violate the Inc., 757 F. Supp 438, 447 (D.N.J. 1991), permit terms. If a permittee reports that it the Court expressly recognized the has violated a permit limit, the report is laboratory error defense, noting that sufficient to discharge the government’s “DMRs may be deemed admissions when burden of production, but neither the CWA establishing liability in summary judgment itself nor any regulation of which we are motions,” but are not conclusive proof of aware makes such a report conclusive.2 The liability. The Court held that under some circumstances, a “defendant may avoid liability at the summary judgment stage on 2 the basis of inaccurate data in DMRs.” Id. The question before us is whether a While the Yates Court also recognized the permittee violates its permit if its heavy burden on the defendant to prove discharges in fact comply with the terms laboratory error, it stated that a showing of of the CWA but its reports erroneously “‘errors in the actual tests performed indicate the permit was violated. This is a which showed a permit violations [sic]’” pure question of law, and our review is may defeat a summary judgment motion. plenary. In its brief, the government did Id. (quoting Student Pub. Interest not argue that, in interpreting the relevant Research Group, Inc. v. Tenneco provisions of the Act or any relevant Polymers, Inc., 602 F. Supp. 1394, 1400 regulations, we should give any degree of (D.N.J. 1985)). Thus, while the Court deference to any formal or informal administrative interpretations of the Act or 9 trier of fact must still be convinced that the liability means that the CWA is violated if a permit was in fact violated. Evidence that permittee discharges pollutants in violation the reports inaccurately overreported the of its permit, regardless of the permittee’s level of discharge are certainly relevant to mens rea. Strict liability does not mean that show that no violation occurred.3 a permittee may be held liable for violating its permit even if it does not in fact do so. The government stresses the fact that the civil liability provisions of the While the government’s policy CWA create a regime of strict liability, but arguments are certainly forceful in the case this argument misses the mark. Strict of a permittee underreporting levels of toxic liability relieves the government of the waste and then claiming a laboratory error obligation to show mens rea, not the actus defense, we are unpersuaded that they prove reus. See, e.g., W. Fuels-Utah, Inc. v. Fed. compelling in a case like this where the Mine Safety & Health Review Comm’n, permittee alleges that the laboratory error 870 F.2d 711, 713-14 (D.C. Cir. 1989). In resulted in the overreporting of the levels of the context of the present case, strict toxic waste.4 From a public policy perspective, a polluter should not be given the opportunity to underreport levels of regulations. By failing to make any such toxic waste, thereby dumping in excess of argument in its brief, the government its permit, and then, when caught, cry waived any contention based on “laboratory error kept me from knowing that deference. Moreover, when counsel for I was in violation!” But in the case at bar, the government was questioned on this the opposite apparently occurred: ALC was point at oral argument, she did not call to conducting its sampling but a contaminated our attention any administrative reagent used in the ALC laboratory’s interpretation to which she claimed that analysis was causing the laboratory deference was owed. Nor has the systematically to overreport the amount of government brought any such toxic zinc that was dumped into the water. interpretation to our attention after the We fail to see what incentive ALC could argument. have had to overreport how much zinc it 3 was dumping into the river when it knew We use the term “laboratory error that such amounts would result in fines. We defense” in this opinion because the term do not believe that a scheme assigning strict has been used in prior cases and is used liability for discharge violations in the case by the parties here, but it is important to note that laboratory error is not an 4 affirmative defense to liability. Instead, In the underreporting situation, the evidence of laboratory error is simply permittee would be attempting to use evidence that is relevant to the question laboratory error to show that it lacked whether a violation of a permit mens rea, which is irrelevant under the requirement in fact occurred. civil liability provisions of the CWA. 10 of overreporting errors makes sense, nor regulations thereunder. Moreover the very do we infer from the CWA that such was circumstances that would support a Congress’s intent. laboratory error defense would also likely support the finding of a monitoring In citing United States v. Pozsgai, violation. See 40 C.F.R. § 122.41(j). In 999 F.2d 719, 725 (3d Cir. 1993), the light of these direct sanctions on inaccurate government correctly asserts that a DMRs, we find wanting the government’s discharge that is not in compliance with a argument that CWA provisions addressed to permit “is the archetypal Clean Water Act actual discharges ought to be made violation, and subjects the discharger to surrogate enforcers of the reporting strict liability.” But in Pozsgai, strict requirements. In sum, barring the assertion liability was imposed based upon an of a laboratory error defense seems unfair unlawful discharge, not the mistaken and at odds with the overall plan of the report of a discharge. The government CWA, especially in a case such as this seems to be aware of this difference when where the alleged laboratory error caused it argues that strict liability should be overreporting rather than underreporting. imposed on reporting requirements, as it writes about the conjunction of the We have considered the arguments of “CWA’s reporting requirements and the government and the Union Oil Court imposition of strict liability for permit that recognizing a laboratory error defense violations.” (emphasis added). So, while would reward sloppy practices and the CWA unambiguously imposes strict undermine the self-monitoring program by liability for unlawful discharges, it is by giving companies an invitation to wait until no means obvious that a similar strict they are sued. But these arguments do not liability regime has been imposed on faulty apply to overreporting, which is almost reporting. certainly involuntary. We also suspect that overreporting is rare, for only the most In fact, the existence of a penny-wise and pound-foolish of permittees mechanism to correct erroneous DMRs would expose itself to the cost of a decade suggests the opposite. See 40 C.F.R. § of litigation (as here) if it had any chance of 122.41(l)(8) (requiring a permittee who clearing the matter up with improved becomes aware of any inaccuracy in a laboratory testing and amended NPDES DMR to promptly notify the EPA). That reports. Conc omita ntly, we a re regulation was promulgated pursuant to underwhelmed by the government’s the Administrator’s authority under 33 argument that permitting the defense will U.S.C. § 1318(a) to impose reporting add time to NPDES litigation. At bottom, requirements. Since 33 U.S.C. § 1319 we do not believe that efficiency should authorizes administrative, civil, and even override fairness in administration. Thus, criminal penalties for violations of § 1318, while we do not gainsay the validity of the the failure to correct an inaccurate DMR is government’s argument that, consistent with an independent violation of the CWA and 11 the Act’s requirement for accurate self- defense argument in the proper light, it, not reporting, courts should treat DMRs, this Court, should consider the defense in which must be certified by the discharger, the first instance. We will therefore vacate as admissions that are sufficient to and remand so that the laboratory error establish liability under the CWA, we hold defense can be considered and adjudicated that the presence of certified DMRs does with respect to the claims that it affected. not preclude the laboratory error defense in III. The Penalty Calculation - Economic cases of overreporting. Benefit The government has argued that A. ALC’s Objections to the Penalty even if the laboratory defense is Assessment recognized, there is insufficient evidence in this record to support it. The District The assessment of civil penalties for Court did appear to endorse this position in these violations as sought by the United a post-trial opinion: “Nothing in ALC’s States is governed by 33 U.S.C. § 1319(d). proffer or testimony on this issue Section 1319(d) provides that the violator of persuades the court that these violations a permit issued pursuant to the Act shall be arise solely from laboratory error.” But subject to a civil penalty not to exceed that statement followed a trial at which the $25,000 per day for each violation. This laboratory error defense had been penalty provision further states that in excluded. More specifically, while the assessing the penalty, the court shall District Court did have available some of consider the following factors: ALC’s laboratory error evidence in the the seriousness of the penalty phase, having already determined violation or violations, the that ALC was liable for discharge claims, economic benefit (if any) this after-the-fact consideration of the resulting from the violation, evidence for penalty purposes does not any history of such violations, cure the error in precluding the laboratory any good-faith efforts to error defense in the liability jury trial. comply with the applicable Arguably the District Court’s evaluation of requirements, the economic ALC’s laboratory error evidence in the impact of the penalty on the penalty phase strengthens ALC’s argument violator, and such other that it was entitled to have the jury matters as justice may require. evaluate such evidence because, what the District Court was doing was to assess the credibility to that evidence (“[I]t is not Id. The District Court considered each of credible that laboratory error would persist these factors in connection with the penalty . . . .”, normally a jury function. determination. The Court found ALC’s Since the District Court did not violations of the CWA to be serious. It consider the sufficiency of laboratory error questioned the level of ALC’s commitment 12 to the obligations imposed by the Act. It Finance 932 (6th ed. 2002). 5 found the economic benefit to ALC to be ALC asserts that the District Court’s considerable, primarily in terms of the economic benefit calculation did not “level avoided cost stemming from reduced the playing field,” as required by law, but (inadequate) staffing at its wastewater rather imposed a severe penalty. ALC also treatment plants, its delay in a plant contends that the District Cou rt’s upgrade at the Vandergrift facility, and a calculations failed to apply other principles number of other smaller projects. The required by law, including that (1) Court totaled the economic benefit at expenditures made and included in the $4,122,335, and ultimately doubled it to economic benefit calculation must relate $8,244,670 as the final penalty. See infra directly to the violations; (2) the least costly note 6. method of compliance should be used in The imposition of a penalty under § calculating economic benefit; and (3) 1319(d) is subject to the exercise of a economic benefit calculations must be based district court’s discretion. See Tull v. only on periods of non-compliance. We United States, 481 U.S. 412, 426-27 reject the argument that the District Court (1987). In general, a district court abuses did not apply the proper legal precepts. its discretion when it “bases its opinion on Rather the question is the manner of a clearly erroneous finding of fact, an application, and whether the District Court erroneous legal conclusion, or an improper made clearly erroneous fact findings which application of law to fact.” LaSalle Nat’l skewed the calculations to ALC’s detriment. Bank v. First Conn. Holding Group, L.L.C. XXIII, 287 F.3d 279, 288 (3d Cir. 2002). B. The Economic Benefit Principle Many of the District Court’s As noted above, § 1319(d) requires findings are supported, and unchallenged the District Court to consider “the economic on appeal. The primary issue contested benefit (if any) resulting from the violation” here relates to economic benefit—i.e. the when determining the amount of a civil Court’s use of the government’s experts’ penalty under the CWA. ALC argues that computation of ALC’s weighted average the purpose of the economic benefit cost of capital (“WACC”) as the interest component of the penalty is to “level the rate to use to bring the money forward to the penalty judgment date. WACC is defined as “the average cost of capital on 5 Of course, this general definition is the firm’s existing projects and activities, only so useful; moving from the broad . . . . calculated by weighting the cost of definition to the actual numbers (in each source of funds by its proportion of particular establishing the “cost of ... the total market value of the firm.” funds”) can be extremely complex and Stephen A . Ross, Randolph W. subject to dispute as this case so aptly Westerfield & Jeffrey Jaffe, Corporate demonstrates. 13 economic playing field.” We agree. See Putting aside the ultimate way in United States v. Mun. Auth. of Union Township, 150 F.3d 259, 263-64 (3d Cir. 1998) [hereinafter Dean Dairy]. In other employ the “bottom up” approach, in words, the purpose is to prevent a party which economic benefit is established, and violating the CWA from gaining an unfair the remaining five elements of § 1319(d) advantage against its competitors, and to are used to adjust the figure upward or prevent it from profiting from its downward. Dean Dairy, 150 F.3d at 265. wrongdoing. See Powell Duffryn In Dean Dairy, we held that the method Terminals, 913 F.2d at 80. The used in assessing the civil penalty is best government, on the other hand, submits left to the trial court’s discretion. See id. that CWA penalties are intended to In the case at bar, the District Court “promote immediate compliance” and followed the “bottom up” approach. “deter future violations” by the defendant Having arrived at a figure of economic and other regulated entities. Friends of the benefit totaling $4,122,335, the District Earth, Inc. v. Laidlaw Envtl. Servs., Inc., Court then conducted a detailed analysis 528 U.S. 167, 185 (2000). Therefore, of the remaining factors enumerated in § while the government agrees that the 1319(d) and found that, while the economic benefit analysis is designed to government advocated a trebling of the calcu late how much money was economic benefit, a doubling would be illegitimately gained by failing to spend more appropriate under the circumstances the appropriate amounts on environmental of the case for a total penalty of safeguards, it does not agree that the $8,244,670. In Dean Dairy, we approved assessment of a penalty need stop at that the doubling of economic benefit as a figure. In our view, the latter point possible method for assessing a penalty addresses a different aspect of the Act, as stating that, even after the doubling of explained in the margin.6 economic benefit, the “penalty was barely 9% of the maximum statutory penalty to which Dean Dairy was subject.” Dean 6 The CWA does not prescribe a Dairy, 150 F.3d at 265. In the case at bar, specific method for determining the statutory maximum penalty that could appropriate civil penalties for violations. have been leveled against ALC was In Dean Dairy, we noted that some $28.05 million, counting $25,000 for each courts use the “top down” approach in of the 1,122 days of violations. While which the maximum penalty is set $8,244,670 is approximately 29% of ($25,000 per day of violation at the times $28.05 million, a much larger proportion relevant here), and reduced as than the 9% approved in Dean Dairy, we appropriate considering the six are satisfied that the District Court was enumerated elements of § 1319(d) as well within its discretion to assess such a mitigating factors, while other courts penalty in this case. 14 which the result of the economic benefit The District Court derived this rate calculation might be employed, such a from the proffer of government witnesses calculation is intended, at its base, to Gary Amendola and Robert Harris who identify the benefit realized by a violator explained the three steps they took to from delayed expenditures to comply with calculate the WACC. First, they determined the CWA. The economic benefit that ALC had a debt rating of “A,” as calculation starts with the costs spent or assigned by Standard & Poor’s (“S&P”). that should have been spent, to achieve Then, they researched what the typical compliance. Once that figure is monthly interest rate was for A-rated bonds established, an appropriate calculation of in each relevant year and computed yearly economic benefit should also reflect the averages. This rate was adjusted to account time value of money. In order to make for the advantageous tax treatment of that calculation, a court must “apply an interest payments on corporate debt. interest rate to determine the present value Second, they calculated the cost of equity as of the avoided or delayed costs.” United follows: They started with a 30-year States v. Smithfield Foods, Inc., 191 F.3d treasury bond as a baseline. They next 516, 530 (4th Cir. 1999). Herein lies the looked up the company’s “beta,” which is a crux of the disagreement: ALC contends measure used to evaluate the relative risk of that the District Court used an interest rate a particular stock for an equity investor. so high that the effect was punitive rather Finally, they assumed a generic value for the than “leveling,” whereas the government market-risk premium— the premium that a contends that the interest rate used by the person would demand to invest in stock District Court was entirely appropriate and rather than in a (risk free) treasury yielded a result that was well within the instrument. At that point, they multiplied Court’s discretion. the beta by the market-risk premium, and added an “intermediate stock premium” for C. The Interest Rate Adopted by the the years before ALC merged with another District Court entity and became a bigger, “safer” The District Court, in arriving at its company. They then added this to the 30- penalty assessment, adopted the economic year treasury bond rate to arrive at an equity analysis proffered by the government. In cost by year. Third, they combined these that submission, the alleged economic cost of debt and cost of equity measures by benefit stemming from each violation was taking a weighted average of them, based on computed forward from the date of the relative proportions of debt and equity in violation to February 28, 2001 (roughly the ALC’s capital structure for that year.7 date of the judgment) at a rate of 12.73% annually, to arrive at a $4,122,335 total economic benefit at the time of judgment. 7 In its brief, the government mischaracterizes its own experts’ testimony and states that the WACC was 15 1. The Contentions of the Parties not achieved by ALC. ALC characterizes the 12.73% rate ALC submits that, instead of the as “a theoretical, risk-adjusted rate 12.73% rate, one of four alternative rates (denominated by EPA as the weighted should have been used: average cost of capital or ‘WACC’), based (1) the statutory interest rate on broad averages across the U.S. capital (6%) markets.” As the foregoing explanation suggests, this characterization is generally (2) the risk-free rate accurate. ALC contends that using such a represented by the short-term hypothetical rate of interest was an error of U.S. treasury rates during law because ALC had presented evidence the relevant time period of its actual rate of return on capital which, at the time of the penalty trial, showed that (3) the actual average rate of the average rate of return on capital for ALC’s return of capital from ALC and its parent company between 1990-2000 1990-2000 was 5.7%. This fact was (4) the actual average rate of uncontested, and thus ALC submits that ALC’s return of capital from the 12.73% rate did not achieve the legal 1990-2001 purpose of “leveling the economic playing field,” but rather was used to exact a Each of the rates suggested by ALC results severe penalty “reflecting not the time in approximately the same interest rate, value of money nor ALC’s benefits from hovering between 5.2% and 6%, which is retaining funds, but rather theoretical less than half the rate that the District Court investment averages that indisputably were actually used.8 ALC adds that the “theoretical WACC has been rejected consistently when applied to companies and calculated by “first determining the rate industries that are not achieving such at which ALC borrowed funds during the theoretical rates of return.” 9 relevant time period.” This representation implies that the government experts relied on figures that 8 Although the government maintains were much more ALC specific than was that the only alternatives to the WACC actually the case. As we have explained, preserved by ALC for appeal are the T-bill the experts seem to have relied primarily rate and the Pennsylvania statute rate, we on general market numbers for have examined the record and do not find companies situated similarly to ALC that ALC waived any of the proposed over a long period. We do not know the alternative rates. reasons for the government’s 9 mischaracterization, but we do note our We think that ALC overstates the disapproval. principle—if any—that may be drawn 16 The government responds with a from the cases it cites. The two cases it number of arguments. First, the government discusses are Chesapeake Bay correctly notes that the economic benefit Foundation v. Gwaltney of Smithfield, calculation need not be precise. In Dean Ltd., 611 F. Supp. 1542 (E.D. Va. 1985), Dairy, we recognized that economic benefit aff’d, 791 F.2d 304 (4th Cir. 1986), “ m a y n o t b e c a p a b l e o f re a d y vacated on other grounds, 484 U.S. 49 determination,” and the Court gave “the (1987), and United States v. Sheyenne district court’s award of a penalty wide Tooling & Mfg. Co., 952 F. Supp. 1420 discretion, even though it represents an (D.N.D. 1996). approximation.” 150 F.3d at 264 (citing In Gwaltney, 611 F. Supp. at Tull, 481 U.S. at 426-27). The government 1559, the Court held that “the actual couples this deference accorded to district interest rate Gwaltney itself paid on court awards with the suggestion that, since borrowed funds [] is a more accurate the statutory maximum penalty for ALC’s basis for determining Gwaltney’s violations was $28.05 million, the District economic benefit from delay” than “the Court gave ALC “a break.” The ten-year rate of return on equity earned government advocated taking other statutory by Smithfield Foods, Inc.—Gwaltney’s factors into account and trebling the parent corporation.” While this case economic benefit to yield a penalty of does not adopt WACC as a measure of approximately $12.3 million, see supra note economic benefit, it also does not 6, but the District Court only doubled the affirmatively reject it. economic benefit and ordered ALC to pay Likewise, in Sheyenne Tooling, $8,244,670. The government points to this 952 F. Supp. at 1426, the Court held that discrepancy between what it asked for and the principle of requiring that what the Court actually did as proof that the persons at fault must be held District Court really does have, and should to a ‘level playing field’ means have, a great amount of discretion in that the defendant must be held determining these types of penalties. to the conditions of his field, not that of larger or more wealthy The government also points to the players. And the economic decisions of other courts that have approved experts for the United States the use of WACC to discount economic used averages and generalizations benefit when calculating CWA penalties, which were not compatible with particularly Smithfield Foods where the the playing field in which the District Court, crediting expert testimony, defendant operated. used the WACC to discount the defendant’s This is doubtless a sound principle, but simply does not address what the appropriate measure is for determining economic benefit. 17 economic benefit. See United States v. the District Court.10 In this case, however, Smithfield Foods, Inc., 972 F. Supp. 338, 349 & n.17 (E.D. Va. 1997), cited with approval in Dean Dairy, 150 F.3d at 266. 10 We note a provision from the field of trusts that enables the District Court to 2. The Appropriate Interest Rate exercise its discretion in choosing the appropriate measure for assessing a The methodology used by the trustee’s liability in the case of a breach of District Court and those advanced by the trust. The choice to make here (i.e., both parties do not exhaust the possible cost measures and actual returns are interpretations of economic benefit under possible ways of valuing economic § 1319(d). It will be helpful to analyze the benefit, so which should be adopted?) options. There are, as we see it, two resembles that choice. The Restatement possible approaches. The first is the cost (Second) of Trusts § 205 provides (in the of obtaining capital— i.e., the interest rate disjunctive): necessary to acquire the capital with which If the trustee commits a to make the improvements (which were breach of trust, he is never made). The second is the use of the chargeable with: corporate offender’s actual return on its (a) any loss or depreciation capital, which, it is conclusively presumed, in value of the trust estate was not used to make the improvements. resulting from the breach of These are both highly variable factors, trust; or turning on the cost of money to the (b) any profit made by him company (which depends not only on the through the breach of trust; general market forces but also on its or financial strength and credit rating) or on (c) any profit which would the profitability of the company at a given have accrued to the trust time. estate if there had been no In view of this variability, we think breach of trust. that it would be inappropriate for us to As in Gwaltney, 611 F. Supp. at 1558-59 decree which methodology should be used & n.17, the choice is within the discretion since in any given situation, “leveling the of the District Court, and we are confident playing field” might be more readily that it will give due consideration to the achieved with one or the other. Therefore, equities involved in selecting an we think that the choice of methodology appropriate measure of economic benefit. should be left to the sound discretion of Indeed, we do not even hold that economic benefit is the sole permissible approach to assessing a penalty; there may well be other ways. Given this variability, we disagree with the dissent’s contention that 18 it is not clear that the District Court was figures specific to ALC’s bonds. 11 aware of or considered the range of The second problem is the options available. government’s application of the WACC. a. Economic Benefit as Measured by the WACC averages are constructed on the Cost of Capital basis of a company’s existing capital structure (that is, the relative proportions of As noted above, economic benefit debt and equity). A WACC figure based on can be measured by an entity’s cost of a company’s existing capital structure at a capital. In accepting the government’s given time is not, without further support, experts’ position, the District Court necessarily the same as a company’s adopted one such measure—WACC— but marginal or current cost of capital at that there are others. In commenting upon the time (i.e., what it would cost to obtain cost-of-capital measure adopted by the additional capital) because new capital District Court, we hope to provide some might come in a different mix of debt and guidance as to what constitutes an equity. See Aswath Damodaran, Applied appropriate cost-of-capital measure of Corporate Finance 108 (1999) (“In economic benefit. estimating [the current cost of capital using With respect to the cost-of-capital WACC], we have in a sense conceded the measure used by the District Court, we status quo in terms of financing mix, since conclude that both the calculation and we have estimated the cost of capital at the application are, at the very least, existing mix. It is entirely possible that a unsupported. The first problem is the firm, by changing its mix, could lower its government’s calculation of the WACC. cost of capital.”). Unless WACC is shown That calculation relied on values that were to be a good approximation for the marginal not ALC-specific. Instead of using the or current cost of capital, it sheds little light actual yield on bonds that ALC had issued, on how expensive it would have been for the government experts computed the the company to go to the market for its WACC by using the yield on Standard & capital, instead of diverting funds that Poors A-rated bonds. While using the should have gone to improving pollution S&P figure might well have been a reasonable approximation of ALC’s 11 bonds’ yield, a more accurate calculation In contrast, as far as we can tell, the could easily have been achieved by using cost of equity calculation was as ALC- specific as could reasonably be achieved: The value for beta seems to have been ALC-specific, and the other figures that entered into the computation (the market- our holding saps too much discretion risk premium and the “intermediate stock from district courts in cases under the premium”) are not by their nature CWA. company specific. 19 controls. established by a cost-of-capital measure, the measure to use is ALC’s marginal or current As noted above, the government cost of new capital in the years in question.12 and the District Court relied on Smithfield Some courts appear to have endorsed this Foods. But, upon closer analysis, approach. See, e.g., Gwaltney of Smithfield, Smithfield Foods does not help the 611 F. Supp. at 1559 (“[T]he actual rate government. There are reasons to suspect Gwaltney itself paid on borrowed funds . . . that in the food processing industry (in is a more accurate basis for determining which Smithfield operated), the WACC Gwaltney’s economic benefit from delay.”). may have been an entirely appropriate approximation of Smithfield’s economic It is of course possible that this benefit, whereas conditions in the steel approach might make an offender worse off industry (in which ALC operates) are than under the government’s WACC radically different. More precisely, it may proposal. For example, a company in dire have been that in Smithfield Foods that the financial straits may well have a marginal WACC was a good approximation for the cost of capital (offered by lenders who see it terms on which money could have as a high-risk investment) that exceeds its currently been raised; the food processing WACC. This is no anomaly. For industry is a stable industry where companies that are hard up for capital and companies probably attract new capital on cannot afford to raise it in the market, it is terms similar to their existing capital doubtless all too tempting to forego the structure. The steel industry, in contrast, sometimes costly improvements and has been highly volatile and rife with stiff pollution controls that are required by the foreign competition, dislocations, and CWA and EPA regulations. But such bankruptcies. Indeed, as the District Court companies must still be held to the law. To noted, the industry is going through a do otherwise is, in essence, to allow capital- “brutal restructuring,” and more than twenty-five United States steelmakers have 12 sought bankruptcy protection since 1997. This could be established by looking Thus, a company in ALC’s position may to, for example, the cost of any capital not have, at the times in question, been actually raised by ALC at the relevant able to raise capital on the same terms as times, or by the expert opinion of an its existing capital structure. We need not investment banker regarding the terms on (indeed, cannot) resolve this; but for our which ALC could have raised capital. Of purposes, it is enough that there was course, if expert testimony can establish to insufficient evidence for the District Court the District Court’s satisfaction that to say that ALC’s existing capital structure WACC is—in this particular case—a good was representative of the terms on which approximation for marginal cost of capital, new capital would be raised. Thus, if the then WACC could be accepted as a economic benefit to ALC is to be surrogate measure of the marginal cost of capital. 20 strapped polluters to take out low-interest its retained funds or the risk-free return it loans against the environment. might have enjoyed using those funds. We think that the return on capital is a quite We of course intimate no view on viable means of leveling the playing field, what a remand may develop respecting along with the marginal or (then) current ALC’s situation in the 1990s. The cost of capital. government’s experts’ proffer shows debt costs for S&P A-rated bonds were in the 3. Other Observations About the District 6.68% - 10.06% range in the 1990s. That Court’s Analysis is significantly lower than the 12.73% There are other potential problems WACC figure relied on by the District with the District Court’s calculation, which Court. Moreover, in recent years, which relied on the methodology provided by the would also figure in the calculations, government’s experts. It appears that the interest rates have been very low. The government’s experts computed annual record does not reflect ALC’s actual estimates of WACC for each of the years financial strength, and it may (or may not) 1990-1998, and came up with the 12.73% also have (or have had) a good credit figure by taking the arithmetic mean.13 rating throughout the relevant period. Since the savings from different violations b. Economic Benefit as Measured by accrued on different dates over a several Actual Return year period, it is questionable whether an average interest rate is appropriate, when We have so far been talking about year-to-year interest rate estimates are measuring the economic benefit of known and could be used with only minimal additional capital by the cost to obtain that additional effort by the experts.14 The capital elsewhere. But the other option is to use actual rates of return on capital to com pute economic benefit. The 13 government’s experts cited the importance To be clear, by “mean” we are of leveling the economic playing field “in referring not to WACC (which, as a the same industry.” It is obvious, for “weighted average” is a mean of sorts) but example, that ALC and the steel industry rather to the further step of taking the were not, at times relevant, enjoying stellar mean of a whole series of WACC figures returns. Indeed, as noted above, it was (one for each of the years in the relevant uncontested at trial that ALC had a return period). We have no objection, as the on capital that was less than half the dissent suggests, to the use of the WACC 12.73% rate used by the District Court. formula to assess economic benefit. On this view, any advantage that ALC 14 enjoyed over its competitors by avoiding While any correction will be slight, in the cost of CWA compliance is measured the interest of precision the District Court by the return that ALC actually realized on might also consider whether, if an average is to be used, the correct procedure would 21 potentially problematic practice of using a We are, of course, acutely aware that mean interest rate over a large time span is we review the District Court’s interest rate present in the government’s experts’ determination for abuse of discretion, and report.15 As it happens, this wound up that its determination need not be exact. See hurting ALC: The theoretical WACC Dean Dairy, 150 F.3d at 264-65. Our figures from the early 1990s (15.85% in deferential scope of review does not mean, 1990 and 1991, and 13.95% in 1992) are however, that we cannot intervene when a the highest of the group, but really have no District Court makes a finding that is bearing on the economic benefit conferred methodologically flawed, even if, under by post-1992 violations. Thus, the average such theory, the penalty figure it ultimately WACC was biased toward the less- arrives at is plausible. relevant higher WACC estimates from the In the dissent’s view “given our early 1990s. highly deferential standard of review, the Finally, we note that the District Court did not clearly err in crediting government is unquestionably correct in its the government’s witness over ALC’s assertion at oral argument that any witness and adopting the WACC to computation must use the same discount calculate economic benefit.” Of course, rate for both forward and backward when presented with two sound but computations during the same period. For conflicting expert opinions, a district court example, it would be clearly inappropriate has discretion to credit one over the other. to discount all econo mic benefit But this discretion is not a license to adopt backwards to a uniform date using one an opinion based on unsound methodology, rate, and then use a different rate to carry whatever its source. the value forward to the date of judgment. Based upon our analysis of the 4. Conclusion government’s expert’s methodology, we are unconvinced that the use of the 12.73% interest rate achieves the stated purpose of “leveling the economic playing field,” nor be to use a geometric mean (computed as are we sure that it bears much connection to the nth root of the product of n items), a meaningful measure of ALC’s cost of since the percentages involved are capital (much less its return on capital). applied in consecutive multiplications. We therefore must set aside the penalty See Damodaran, Applied Corporate calculation and remand for further Finance at 69-70. proceedings with respect to the interest rate, 15 Moreover, this practice is not fully open to the possibilities that the record unique to the use of W ACC as a measure on remand will support a higher, lower, or of economic benefit; it is an issue substantially similar penalty. We will not regardless of the method used to derive choose among the alternatives we have the interest rate. suggested (or those suggested by ALC ) in 22 the discussion above; rather we shall leave question does not appear to have been it to the District Court, after receipt of addressed by any Court of Appeals. Those further submissions by both parties, to District Courts that have addressed the issue decide what alternative rate is best applied hold that the calculations should be based on to the circumstances developed in the the least costly method of compliance. See, record on remand.16 e.g., Gwaltney of Smithfield, 611 F. Supp. at 1563 n.25 (holding that economic benefit D. The District Court’s Determination of calculations could not be based on a more Costs of Compliance expensive, “permanent solution” when a 1. Introduction—The Least Costly less expensive “interim solution” had Method of Compliance already achieved compliance); United States v. WCI Steel, Inc., 72 F. Supp. 2d 810 (N.D. The second basis on which ALC Oh io 1999) (finding cre dib le the asks us to overturn the District Court’s defendant’s expert testimony regarding calculation of economic benefit is its poss ible com pliance measures and contention that the District Court erred in calculating economic benefit based on calculating the amount of money it would significantly less expensive method of have cost ALC to institute the changes that compliance than that proposed by the would have led to compliance with the government’s expert). We find these requirements of its permits. In brief, ALC decisions persuasive, and hold that argues that the numbers the EPA came up economic benefit analysis should be based with (which were adopted by the Court) on the least costly method of compliance. and the kinds of solutions it proposed were However, contra ALC’s contentions, it does cons iderab ly overpriced, es pecia lly not appear to us that the District Court took considering that, according to ALC, it had a different approach. already fixed the problems for much less money and could show that the solutions it 2. The Vandergrift Plant had implemented already worked. ALC cites to two main instances of The threshold question is whether, alleged miscalculation of benefit. The first, as a matter of law, the District Court must relating to the Vandergrift plant, stems from calculate economic benefit using the least the District Court’s calculation which costly method of compliance. This incorporated a $600,000 project that the government’s expert posited would have brought ALC into compliance with the 16 The District Judge who originally pretreatment permit issued for that site. heard and decided this case has resigned ALC, however, claims that, in October of from the bench. Accordingly, the parties 1993, shortly before the Vandergrift will doubtless have to develop a record violations ceased, it installed and began to for the edification of the newly assigned operate a diversion tank connected to the judge. discharge piping leading to the Vandergrift 23 facility WWTP outfall. This diversion at a time without reporting any violations. tank cost no more than $150,000 to buy The District Court chose to credit the and install. According to ALC, the testimony of the government’s expert that pretreatment violations stopped shortly the diversion tank “would not have been after the installation of the diversion tank adequate to prevent all violations.” A although there were two monthly average decision to credit the expert testimony of and four daily maximum violations in one expert witness over another is entitled to November and December 1993, which deference. See Gen. Elec. Co. v. Joiner, 522 ALC attributed to “start-up problems.” U.S. 136, 143 (1997) (holding that a District ALC contends that starting December 15, Court’s assessment of expert testimony is to 1993, not a single violation occurred. be accorded “the deference that is the ALC then argues that, in adopting the hallmark of abuse-of-discretion review”). government’s proposed $600,000 project Under these circumstances the District to solve the problem and bring ALC into Court’s findings of fact were not clearly compliance, the District Court made erroneous, and they must therefore be left to clearly erroneous findings of fact. stand. There is, however, another side to 3. Outfall 107 the story. As noted above, there were The next issue concerns a $476,090 several so-called “start-up violations” after project that the government’s expert posited the diversion tank was installed, and ALC was necessary to bring ALC into cannot claim a clean record until compliance for non-contact cooling water December 15, 1993. ALC claims that the violations at Outfall 107. ALC contends District Court should have used the that the District Court’s economic benefit December 15, 1993 date as the compliance calculation which adopted that figure was date because that is the last reported premised on clearly erroneous findings of pretreatment violation before the WWTP fact which led to misapplication of the least upgrade in August of 1994. However, the costly method of compliance legal principle. government’s expert, Gary Amendola, ALC argues that violations at the outfall explained that he chose to use August were limited to June though October 1994 1994 as the compliance date (as did the and that those violations were resolved District Court) because the diversion tank through various maintenance efforts, installed in October 1993 was not including repairing cracks in certain sufficient to address the problem at trenches and sumps. The government, Vandergrift. Amendola explained that the however, points out that ALC’s brief does fact that ALC had reported no violations not contain any record citation indicating during the first half of 1994 did not that it presented factual material to the establish that the diversion tank was a District Court at trial relating to its sufficient compliance measure because the maintenance efforts, and that to the extent facility had previously operated for months that there is such evidence in the record, the 24 evidence is limited to ALC’s own has a direct relationship to the calculation of pleadings. Furthermore, the government economic benefit: The longer the period of contends that ALC reported violations non-compliance, the greater the amount of long after it alleges that it cured them in economic benefit, and the higher the October 1994. penalty. ALC contends that the government miscalculated the period of non-compliance This difference of opinion as to and that, in adopting the government’s whether more violations occurred stems calculations, the Court calculated ALC’s from the fact that ALC identifies only one purported economic benefit on lengths of internal monitoring point, “Outfall 107,” time that bore no semblance to reality. associated with the $476,090 re-lining project. The government expert, We do not find it necessary to engage Amendola, however, opined that the re- in a lengthy analysis of the various lining project was required to cure contentions regarding the periods of non- violations associated with Number 90 compliance and will set forth some of the Anneal and Pickle Line, which discharged factual disputes only in the margin.17 through Outfall 007. Outfalls 107 and 207 are internal monitoring points that 17 discharged through Outfall 007. Since ALC relies on a table it has created ALC reported violations at Outfall 007 that purports to show the non-compliance through December 1995, long after ALC periods designated by the government alleges it cured those violations with were far greater than the actual non- maintenance efforts in October 1994, the compliance periods that occurred. In the government contends that the maintenance table, ALC challenges the non-compliance efforts at Outfall 107 are not enough to dates for the $476,090 relining project carry the day. discussed above. That project was completed in 1996 and was necessary to It is clear that the District Court cure violations at Outfall 007, at which decided to adopt the government’s ALC reported violations through framework regarding the monitoring and December 1995. However, government links between these different outfalls. In expert Amendola extended the non- view of the bona fide evidentiary dispute, compliance date back to the beginning of its findings were not clearly erroneous and the limitations period for this case because must be upheld. ALC reported violations associated with E. Periods of Non-Compliance the Number 90 Anneal and Pickle Line beginning at the time it came online in ALC’s final complaint relating to 1988. See App. 583-88; see also App. 991- the economic benefit analysis undertaken 92, 994-95 (ALC documents stating need by the District Court is the identification of for treatment upgrade to attain the period of non-compliance. Obviously, compliance). the length of the period of non-compliance ALC also appears to be repeating 25 Having thoroughly reviewed the record, Court held: we hold that the District Court’s findings Plaintiff’s Motion in Limine as to the periods of non-compliance are on Cou n t i n g D a y s of supported by the record, were not clearly Violation, Doc. No. 242, is erroneous, and must be left to stand. GRANTED. All violations of IV. Monthly Average Violations the monthly average parameters of defendant’s In a pretrial ruling the District NPDES permits shall be counted as violations equal in number to all the days in the its contention that the $150,000 diversion monitored month. See tank it installed in October 1993 cured its Atlantic States Legal Found’n pretreatment violations at Vandergrift, v. Tyson Foods, Inc., 897 but the District Court found that it was F.2d 1128, 1139 (11th Cir. not until the WW TP upgrade in August 1990). 1994 that the pretreatment problem at Vandergrift was solved, and we have ALC maintains that the District Court erred declined to disturb that finding. in so ruling, and in particular by improperly Additionally, ALC’s table excluding evidence that actual exceedences challenges the non-compliance date of occurred on fewer days. ALC relies December 1994 for the 24-hour staffing primarily on Texaco Refining & Marketing, the District Court deemed necessary to 2 F.3d at 507. The relevant portion of the alleviate pretreatment violations at holding of that case is that violations of the Vandergrift. The Court found that ALC daily average limits result in penalties only did not have 24-hour staffing in place for the number of days within the month that until “late 1994” or 1995. ALC’s the facility operated. That decision does contemporaneous internal documents not, however, resolve this case. confirm that 24-hour staffing was The leading authority in this area is necessary and was not in place before the Court of Appeals for the Fourth December 1994. Thus, like the WWTP Circuit’s opinion in Gwaltney. Gwaltney upgrade, the 24-hour staffing problem held that a violation of a monthly average was not solved until December 1994. parameter constitutes a violation of each day ALC’s table also challenges the non- of the month. The Court reasoned: compliance dates for 24-hour staffing at West Leechburg. The District Court While the statute does not rejected this challenge, finding ALC’s address directly the matter of analysis “misleading” because ALC monthly average limitations, committed 599 violations between 1990 it does speak in terms of and November 1993 that were the subject penalties per day of violation, of consent agreements with the State. rather than penalties per 26 violation . This moderate, long-term discharges are language strongly potentially harmful.” Id. at 315 n.17. The suggests that where a Court also observed that the statute merely violation is defined sets a maximum penalty; the District Court in terms of a time retains the discretion to assess a smaller period longer than a penalty where appropriate. Id. The Court day, the maximum stressed that counting average monthly penalty assessable violations as a violation of each day of the for that violation month is essential to providing a framework should be defined in that allows district courts “sufficient terms of the number flexibility to assess penalties that suit the of days in that time particular circumstances of each case.” Id. period. at 314. We find the reasoning of Gwaltney incomplete. A discharger who exceeds the 791 F.2d at 314 (footnote omitted). The monthly average maximum by a great Court of Appeals for the Eleventh Circuit amount will probably also have committed has followed Gwaltney. See Atl. States a number of daily violations, and the Legal Found., Inc. v. Tyson Foods, Inc., penalties for those violations will mete out 897 F.2d 1128, 1139-40 (11th Cir. 1990). at least part of the total punishment that the ALC contends that charging it with permittee’s conduct for the month merits. a month’s worth of violations based on the The penalty for violating the average excedence of a monthly average permit monthly maximum seems well suited to limit yields illogical and unfair results. punish a pattern of discharges that, with a For example, ALC claims that a single few exceptions, do not violate the daily upset caused the average of the four maximums but are nevertheless, in the samples for M ay and September 1991 to aggregate, excessive. However, we find exceed the monthly average limit, while problematic the proposition that the three of the months’ samples were within maximum penalty for such a course of the effluent limits. ALC submits that the conduct should be thirty times the maximum District Court’s ruling “automatically penalty for the worst daily violation converted a single event into 31 violation imaginable. days, despite evidence to the contrary.” Under 33 U.S.C. § 1319(d) a violator This was the justification rejected in is “subject to a civil penalty not to exceed Gwaltney. In that case the defendant $25,000 per day for each violation,” which presented the Court with hypotheticals means that a civil penalty of $25,000 may be similar to ALC’s contentions. 791 F.2d at assessed for each day that a violation 314-15. The Court noted that the occurs. Under Gwaltney, a violation of the defendant’s hypotheticals ignored the fact monthly average maximum occurs on every that “both large, isolated discharges and day of the month, which could result in a 27 monthly penalty of roughly $750,000, but House and Senate Committees. But we that does not seem to be the most literal must still decide this case. We are not reading of the statutory language. That prepared to say that Gwaltney was simply said, we are fairly confident that no one in wrongly decided. Instead—and the best we Congress ever thought of the question that can do in view of the muddled state of is now before us, and it does not appear affairs—is to follow Gwaltney on the that there is any answer to be found in the question of the statutory maximum, and to text of the CWA or its legislative history. use it as a framework, but to give guidance Nor do we think that the structure or structuring the way in which a district court purpose of the Act yields any clear answer. is to exercise its discretion in setting an Certainly we can infer that Congress actual penalty. This is the course we follow. wanted to set an upper limit on the civil More particularly, in exercising its penalty that a district court can award; discretion, a district court should take into Congress did not want to leave this account the degree to which the polluter’s entirely to the district court’s discretion. conduct had already been punished by But without knowing Congress’s views on penalties for daily violations and to use the the relative severity of a violation of a maximum penalty for a daily violation as a monthly as opposed to a daily limit, it is basis for comparison. Thus a district court difficult to tell what sort of upper limit would not assess a daily penalty of more Congress wanted to propose. than $25,000 as a function of the monthly Given the opaqueness of the statute average violation unless it could say that the and the consequent muddle that we have permittee’s violation of the average monthly described, we urge either that the Congress maximum was as blameworthy (taking into amend the statute to clarify its intentions account the factors enumerated in 33 U.S.C. or that the EPA consider the matter and, § 1319(d) including environmental harm) as after notice and comment, promulgate a daily violation for which the $25,000 regulations that will give more guidance.18 maximum would be appropriate. This To that end we will direct the Clerk of exercise will not always be simple as there Court to send copies of this opinion, is a certain incommensurability between directing attention to this section, to the short, intense and prolonged moderate Administrator and General Counsel of the discharges, but we are confident that the EPA and to the counsel for the relevant district courts, in the exercise of their discretion, can do the job. Since the District 18 Court did not have the benefit of this Indeed, in a sense it is the EPA’s standard, we must vacate and remand so that regulations that have created the it may apply it to reconsider the penalty for quandary, because they inject the concept monthly average violations. of a monthly violation into a statute that authorizes penalties denominated only in Our modified Gwaltney approach days. must, however, be applied in accord with 28 Texaco. Under such a regime, there must United States v. Allegheny Ludlam, be excluded from the calculation days on No. 02-4346 which the facility in question did not operate. If there was evidence in this record that the plant did not operate on FUENTES, Circuit Judge, dissenting. certain days, this District Court would have to consider that as well. As best we I concur and join in Part II, Parts IIID can ascertain, however, there is no such and E, and Part IV of the majority’s well- evidence in the record. The closest ALC crafted opinion. I disagree, however, with comes is to represent that the Basic the majority’s conclusion that the District Oxygen Furnaces were not operating Court abused its discretion when it credited during the week of January 24, 1994, but the EPA’s expert economist and used that ALC makes no claim that non-functioning expert’s interest rate to calculate ALC’s furnaces establishes overall plant closure. economic benefit rather than the rate In fact, one ALC witness testified “all of presented by ALC’s expert. The majority our facilities typically operate 365 days a writes that the District Court committed year, 24 hours a day” and that “Allegheny clear error because, in applying the EPA’s Ludlum’s facilities generally operate 24 12.73% discount rate, the Court so vastly hours a day, 365 days a year.” At all overstated the economic benefit to ALC of events, no date other than January 24, its Clean Water Act (“CWA”) violations 1994, is identified as a date for (possible) that it failed to level the economic playing plant shut down. Additionally, we note field. In my view, in selecting the 12.73% that the argument maintained by ALC in rate, the District Court acted squarely within its briefs is not that the Court’s order its discretionary authority. deprived it of the opportunity of proving that plants were not operating on given days, but rather that it was not discharging I. or was in compliance during parts of the Before discussing the discount rate month. issue and the Court’s exercise of discretion, V. Conclusion I think it worth commenting on the proceeding conducted by the District Court. For the foregoing reasons, we will The $8,244,670 penalty imposed on ALC affirm the judgment on liability, except as came after a three-day penalty hearing to those aspects of the government’s during which the District Court heard claims that are affected by the laboratory testimony from 13 witnesses, 11 live and 2 error defense and the monthly average through depositions. These witnesses violations. We will vacate the assessment included experts on economic benefit, cost of penalty and remand for further avoidance and aquatic toxicology, ALC’s consideration in light of this opinion. Director of Environmental Affairs, and Parties to bear their own costs. 29 officials from the United States Coast that the District Court erred in using the Guard, the Pennsylvania Fish and Boat 12.73% discount rate. Commission and the Pennsylvania Department of Environmental Protection. Expert testimony was submitted by written II. proffer with live cross-examination. On the subject of economic benefit, the EPA As I see it, the central issue here is presented testimony from Robert Harris, whether the District Court abused its an economist, who explained how he discretion in crediting one expert over calculated the 12.73% WACC. ALC another when it determined the interest rate. presented testimony from Dr. Howard We have noted many times that abuse of Pifer, who proposed using the 30-day discretion is a highly deferential standard of treasury bill rate to determine the value of review. And, we have stated, on numerous the money going forward to the penalty occasions, that a decision to credit the payment date. In a 30-page opinion issued testimony of one expert witness over after the hearing, the District Court another is entitled to deference. See United credited the EPA’s expert testimony, States v. Universal Rehabilitation Services concluding that Dr. Pifer’s argument was (PA), Inc., 205 F.3d 657, 665 (3d Cir. not supported by the facts and that the 2000), quoting General Elec. Co. v. Joiner, WACC offered a reasonable approach 522 U.S. 136, 143 (1997), United States v. because it represented an average of Mathis, 264 F.3d 321, 335 (3d Cir. 2001), potential investments made by ALC during Laverdi v. Jenkins Township, 2002 WL the time it had use of the funds that it did 31108910 at *364 (3d Cir. Sept. 19, 2002), not spend on compliance. The District Matlin v. Langkow, 2003 WL 283164 at Court also followed Dean Dairy’s *382 (3d Cir. Jan. 22, 2003). The Supreme endorsement of the WACC, as used in Court has held that a district court’s Smithfield Foods. evaluation of expert testimony is to be accorded “the deference that is the hallmark The majority finds fault with the of abuse-of-discretion review.” General District Court’s analysis, noting that the Elec. Co. v. Joiner, 522 U.S. at 143. A government’s calculation of the WACC district court abuses its discretion when it “relied on values that were not ALC- “bases its opinion on a clearly erroneous specific.” Maj. Op. at 19. The majority finding of fact, an erroneous legal also believes that, rather than using an conclusion, or an improper application of average such as the WACC, the law to fact.” LaSalle Nat’l Bank v. First government should have applied the actual Conn. Holding Group, L.L.C. XXIII, 287 rate it would have cost ALC to raise F.3d 279, 288 (3d Cir. 2002). Indeed, we capital for the years when it was diverting have said that “[i]n order to justify reversal, funds that should have gone to pollution a district court’s analysis and resulting control. Therefore, the majority concludes conclusion must be “arbitrary or irrational.” 30 United States v. Universal Rehabilitation (1993); see also Kumho Tire Co., Ltd. v. Services (PA), Inc., 205 F.3d 657, 665 (3d Carmichael, 526 U.S. 137, 147 (1999) Cir. 2000), quoting In re Paoli R.R. Yard (extending Daubert’s gatekeeping obligation PCB Litig., 113 F.3d 444, 453 (3d Cir. to all expert testimony). 1997) (internal quotations omitted). Abuse of discretion requires a showing of clear error, not inappropriateness. In my Still, the majority conducts a view, given our highly deferential standard protracted survey of economic theories, of review, the District Court did not clearly considers treatises not specifically presented err in crediting the government’s witness by experts before the District Court, and over ALC’s witness and adopting the decides that it disagrees with the District WACC to calculate economic benefit. Court’s discretionary determination. Of course, there will always be disagreement among experts concerning scientific, or in Here, after considering all of the this case economic, theories. However, it is testimony, the District Court credited the for the District Court Judge, as fact finder, testimony of the government’s economic to resolve those disagreements by judging expert concerning the WACC, stating that the credibility of the expert witnesses, it “represents the rate of return a company resolving the conflicting evidence, and must earn annually to continue to attract its assessing the weight of the expert’s current investors and maintain its current testimony. There is nothing in the record levels of operations. It is a rate which is here to indicate that the government’s expert commonly used by companies in making did not use sound methodology and capital budgeting decisions.” Dist. Ct. Op. adequately support his opinion, and nothing at 22, quoting Harris Proffer at 6 (internal to show that the District Court was clearly quotations omitted); App. I at 47. The erroneous in crediting that opinion. District Court also credited the testimony of the government’s expert on avoided costs, noting that he had 30 years of The majority’s disagreement as to experience in the environmental field, which interest rate is more “appropriate” is including working for and as a consultant not enough to justify a remand.19 This is to the EPA and several major steel companies. Dist. Ct. Op. at 16-17; App. I at 41-42. The District Court was not 19 required to explore every possibility. As The majority states, for example, that the Supreme Court has stated, a district “[i]n commenting upon the cost-of-capital court need not have conducted an measure adopted by the District Court “exhaustive search” of all possible [i.e., the WACC], we hope to provide alternatives. See, e.g., Daubert v. Merrell some guidance as to what constitutes an Dow Pharms., Inc., 509 U.S. 579, 597 appropriate cost-of-capital measure of economic benefit.” Maj. Op. at 19. 31 especially true in light of Dean Dairy, the District Court’s calculation was where we stressed that economic benefit “reasonable,” we cannot find the Court to “may not be capable of ready have abused its discretion. Relying on determination,” and we accorded “the theoretical values rather than actual values district court’s award of a penalty wide to calculate the WACC does not render the discretion, even though it represents an District Court’s decision “unsupported,” as approximation.” 150 F.3d at 264, citing the majority contends. United States v. Tull, 481 U.S. 412, 426- The record shows th at the 27 (1987). Surely the choice to credit the government’s expert gave a satisfactory government’s expert over ALC’s falls explanation for his decision to use the within this wide discretion. Indeed, the WACC in this case instead of, for example, Dean Dairy Court went on to say that the the marginal or current cost of capital for “[p]recise economic benefit to a polluter the relevant years, as the majority suggests. may be difficult to prove” and that He stated: “[r]easonable approximations of economic benefit will suffice.” 150 F.3d at 264, [The WACC] is a rate that I consider proper quoting Public Interest Research Group of and represents a rate that falls between the N.J., Inc. v. Powell Duffryn Terminals, risk free rate and the equity rate. The reason Inc., 913 F.2d 64, 80 (3d Cir. 1990). As that I believe that the WACC rate is here when the District Court credited one appropriate is because a company’s cash is expert’s reasonable approximation of the fungible. That is, funds are not segregated economic benefit over another’s, it acted and used for specific purposes. Funds are well within its discretion. We ought not used in many different ways and the substitute our own opinion for that of the company receives different returns for each District Court’s. use. Some projects earn a high rate of return. Others earn a low or no rate of In its attempt to fault the District return. It is impossible to say exactly how Court’s calculation of the WACC for the funds that should have been spent in this “rely[ing] on values that were not ALC- example were used. Therefore, I believe the specific” [i.e., using theoretical yields on most appropriate rate to use is the average bonds issued rather than actual yields], the return the company earns on all of its majority, in fact, concedes that the District projects. In essence, this is the average Court’s analysis contained reasonable return for the company. approximations. Maj. Op. at 19. It states that while the bond-yield “figure might well have been a reasonable approximation App. IV at 1009. The record evidence of ALC’s bonds’ yields, a more accurate clearly shows that the District Court’s calculation could easily have been decision to use the WACC was supported by achieved by using figures specific to various considerations, including, as ALC’s bonds.” Id. However, as long as testified by the government’s expert, the 32 fungibility of a company’s funds and the figure to use when calculating ALC’s variable rates of return a company receives economic benefit during those years. I depending on how it uses those funds. disagree with the majority’s contention that “[t]he theoretical WACC figures from the Further, the Court’s exercise of early 1990s . . . really have no bearing on discretion is supported by the case law. the economic benefit conferred by post- Dean Dairy cites the Smithfield Foods 1992 violations” simply because they are the Court’s use of the WACC favorably, highest figures of the group. Maj. Op. at 22. indicating that the WACC is a perfectly The figures from 1990 to 1992 are equally acceptable interest rate for a district court as relevant as those from 1993 to 1998, as in this circuit to adopt when calculating CWA violations occurred in each of the economic benefit. 150 F.3d at 266, citing years from 1990 to 1998. There is no record United States v. Smithfield Foods, Inc., support for the majority’s assertion that the 972 F.Supp. 338, 349 (E.D.Va. 1997). WACC figures from the early 1990s are The majority’s failure to find clear “less-relevant” than those for later years. Id. error after combing the record is evident in Therefore, the majority’s suggestion that the several places. For example, the majority average WACC was unduly biased towards criticizes the government’s expert’s use of high numbers is inaccurate.20 Further, the the arithmetic mean (instead of the District Court pointed out that, in some geometric mean) to compute an estimate of instances, it credited the government’s the WACC for the years 1990-1998. expert in ways that wound up benefitting Although the majority admits that “any ALC. For example, in calculating the correction will be slight,” the WACC economic benefit that ALC enjoyed by comes to 12.71%, as opposed to 12.73%, spending less money to staff its facilities, when it is calculated using the geometric the District Court noted that the mean. Maj. Op. at 21-22 n.14. Surely, a government’s expert discretionary choice by a district judge that made two assumptions that were favorable results in an interest rate .02% higher than to defendant. First, he included in ALC’s an alternative cannot be viewed as clearly actual staffing costs time billed by erroneous. maintenance workers who stopped by the The majority also criticizes the government’s use of a mean interest rate at 20 all, asserting that it “wound up hurting The majority also overstates the ALC.” Maj. Op. at 22. I do not agree that degree to which the highest figures deviate this calculation unduly punished ALC. from the rest of those in the calculation. A Taking an average of the interest rates for figure of 15.85% would not be considered all of the years in which ALC was non- a statistical outlier when computing an compliant is a common and perfectly average, particularly when the same figure acceptable method for arriving at a single appears twice and the rest of the figures range from 10.53% to 13.95%. 33 facility, even though having a maintenance the WACC may not have been as worker stop by is not the same as having appropriate an approximation of economic full-time staffing. Second, [his] benefit for ALC as it was for the company calculations do not include money saved in Smithfield Foods because of differences by ALC at its West Leechburg and in the volatility of the industries in which Brackenridge facilities prior to entry of the each company operated. Again, the c onsent agreements w i th P aD E P standard of review is abuse of discretion, [Pennsylvania Departm ent of and not whether another decision might Environmental Protection]. have been more “appropriate.” Further, the majority cites no authority for the Dist. Ct. Op. at 16 n.7, citing Amendola proposition that using a theoretical interest Proffer at 17; App. I at 41. Also, in rate as opposed to an actual one in a calculating the least costly upgrade that particular industry is clearly erroneous. The would have brought ALC into compliance majority quotes Chesapeake Bay Found., at its Vandergrift facility before 1994, the Inc. v. Gwaltney of Smithfield, Ltd., 611 F. District Court noted that Supp. 1542, 1559 (E.D.Va. 1985), as stating the United States might have pointed to a that “[t]he actual interest rate Gwaltney $1.8 million upgrade considered by ALC itself paid on borrowed funds . . . is a more in 1988 and 1989, or the entire cost of the accurate basis for determining Gwaltney’s $5.7 million upgrade of the Vandergrift economic benefit from delay.” Maj. Op. at. WWTP [Wastewater Treatment Plants], 20 (ellipsis in original). When put into and argued that money should have been context, however, this case does not support spent in 1990, rather than 1994. But in an the majority’s position. In Gwaltney, the approach that is favorable to ALC, [the plaintiff’s calculation computed Gwaltney’s government’s expert] calculated the least economic benefit from delay using “the ten- costly upgrade in 1994 that would likely year rate of return on equity earned by have eliminated the violations, and Smithfield Foods, Inc.--Gwaltney’s parent provided a $600,000 alternative. corporation.” 611 F. Supp. at 1559. The Court went on to hold that “[a]t least in Dist. Ct. Op. at 19, citing Amendola these circumstances, the Court believes that Proffer at 12-13; App. I at 44. As with its 13%--the actual interest rate Gwaltney itself WACC calculation, the District Court paid on borrowed funds--is a more accurate exercised its discretion here and supported basis for determining Gwaltney's economic its decision with acceptable explanations. benefit from delay.” Id. The Gwaltney Here, however, it arrived at a figure that Court, therefore, held against the use of a benefitted ALC. The majority fails to parent corporation’s interest rate, but not the explain how this decision falls within the use of a theoretical interest rate per se. In District Court’s discretion while its addition, the record shows that the District WACC calculation does not. Court did consider the economic benefit The majority also hypothesizes that calculation in an industry-specific context, 34 stating that “[f]ailures to comply with the figure. Despite the majority’s contention to [CWA] can . . . result in indirect the contrary, the District Court demonstrated competitive benefits when compared with a proper application of the law in assessing companies in the same field that do the penalty and, therefore, did not abuse its comply with the [CWA].” Dist. Ct. Op. at discretion. 15; App. I at 40. In short, there is nothing in the record Finally, the majority asserts that the to show that the District Court committed District Court abused its discretion in clear error in its choice of the interest rate to choosing the WACC instead of a lower calculate economic benefit. After carefully alternative interest rate because using the weighing the evidence presented by experts WACC evidenced an effort to punish and on both sides during a three-day penalty deter when calculating the economic trial, the District Court exercised its benefit. However, the District Court discretion as the trier of fact and credited the clearly recognized that there are two steps testimony of one witness over another. The to the “bottom up” approach to penalty decision is supported by the expert assessment and it is the second step that is testimony as well as our case law. Because geared toward punishing and deterring the I do not believe that the District Court’s violator. The District Court stated: fact-finding was clearly erroneous, its decision is entitled to deference under abuse To achieve the goal of deterrence, of discretion review. an appropriate penalty must encompass both the economic benefit that the d e f e n d a n t o b t a in e d t h ro u g h i t s I would, therefore, affirm the District noncompliance, and an additional punitive Court’s decision as to the interest rate used component that takes into account the to calculate economic benefit. penalty factors listed in Section 1319(d). Without the second component, those regulated by the CWA would have nothing to lose by violating it. Dist. Ct. Op. at 29; App. I at 54. The District Court was clearly mindful of the two-step process to be used when assessing penalty, first calculating the economic benefit and then considering the penalty factors to increase that figure. The Court followed the correct analysis, only taking punitive measures in the second step when it doubled the economic benefit 35