Did the Supreme Court’s 2009 BNSF Decision Change CERCLA Cost-Recovery Practice?

The U.S. Supreme Court’s decision in Burlington Northern Santa Fe Railway Co. v. United States, 129 S. Ct. 1870 (2009), examined two unsettled areas of CERCLA: (1) the proof necessary to establish whether a PRP has “arranged for the disposal or treatment… of hazardous substances…”; and (2) CERCLA apportionment, i.e., whether a PRP is jointly and severally liable for an entire site, or rather only severally liable for a portion of the site. 

In an article on Lexology titled, “Four Years Later: How Has BNSF Changed CERCLA Practice?”, posted on November 20, 2012, Anthony G. Hopp and Colin O’Donovan of Edwards Wildman and Paul S. Kline of Three Rivers Management, Inc., explore the ramifications of the BNSF decision.

In summary, the authors conclude that the Supreme Court’s decision has made it more difficult to establish “arranger” liability by tightening the proof requirements. At the same time, however, most trial courts in the post-BNSF era have failed to embrace that portion of the decision dealing with “divisibility.” This article, which provides an excellent survey of CERCLA jurisprudence, is required reading for lawyers involved in cost recovery litigation. 

The BNSF facts are straightforward. B&B operated an agricultural chemical distribution business that purchased pesticides and other chemicals from Shell Oil Company. Shell’s products arrived in tanker trucks and were transferred to storage facilities. Leaks and spills were apparently “common place” during B&B’s handling and transferring of Shell’s products. B&B eventually became insolvent and the Government sought cleanup costs from Shell as an “arranger” under § 9607(a)(3) of CERCLA. The Ninth Circuit affirmed the trial court’s ruling that Shell was liable as an “arranger.” 

The Supreme Court reversed the Ninth Circuit, holding that a PRP is an arranger only when it takes intentional steps to dispose of a hazardous substance. The Court found that “Shell’s mere knowledge that spills and leaks continued to occur is insufficient grounds for concluding that Shell ‘arranged for… disposal…’.” and that the evidence at trial did not support the inference that Shell intended such spills. In so holding, the Court effectively overruled twenty years of “arranger” jurisprudence, which had created a much more liberal standard for establishing liability.  Therefore, defendants that sell useful products and/or have no role in the actual spill are more likely to find BNSF helpful in avoiding CERCLA liability.

The second half of the BNSF decision was devoted to apportionment. As the authors observe, apportionment is different from allocation in that apportionment deals with whether a defendant is jointly and severally liable for an entire site, or rather only severally liable for a portion of the site. Allocation, by contrast, deals with how courts calculate a defendant’s share of liability after it has been determined that the defendant is, in fact, jointly and severally liable. 

In Yankee Gas Servs. Co. v. UGI Utils., Inc., No. 10-cv-580, 2012 WL 1067644 (D.Conn. March 30, 2012), the district court analogized the distinction between apportioning and allocating costs to several guests splitting a dinner check. “To apportion is to request separate checks, with each party paying only for its own meal. To allocate is to take an unitemized bill and ask everyone to pay what is fair.” 

But does the Yankee Gas court’s analysis blur the line between apportionment and allocation? Yankee Gas suggests that § 107 defendants may be able to reduce their overall exposure by taking certain costs off the table – those which a PRP can demonstrate it did not cause. Following the court’s analogy, if the guests each ordered their own entrees and a few bottles of wine, everyone would pay for their own entrees and equitably split the wine, with those guests who did not drink paying only for their entrées. Yankee Gas, therefore, suggests that there may be some interplay between apportionment and allocation.  

This debate is far from being merely academic. Only by demonstrating divisibility of harm is a cost recovery defendant able to defend against a joint and several liability finding. Many CERCLA courts have acknowledged that the universal starting point for the divisibility of harm analyses in CERCLA cases is §433(A) of the Restatement (Second) of Torts. But post-BNSF, the Restatement (Second) of Torts takes on new significance in the divisibility determination.  Thus, trial courts are empowered to look beyond CERCLA case law to other federal decisions interpreting Section 433(A) to determine what showing is required to establish divisibility. 

 Inasmuch as the Supreme Court has clarified the law with regard to allocation, the Edwards Wildman authors ask why not a single post-BNSF trial court has accepted this defense?  Shortly after the issuance of BNSF, it was widely predicted that "divisibility" would emerge as the new frontier in CERCLA litigation.  However, this has not yet occurred.  As difficult as establishing a “divisibility” defense remains in the courts, the article provides excellent strategies for attempting to establish the defense.

Pitfalls In Proving CERCLA Divisibility Of Harm

In a stinging decision, the Hon. Lonny R. Suko, a federal district court judge sitting in the Eastern District of Washington, ruled on April 4, 2012, that PRP Teck Cominco Metals, Ltd. failed to prove that contamination at a CERCLA site was divisible and, as a result, will be subject to CERCLA 107 joint and several liability at an upcoming September 2012 bench trial (Pakootas v. Teck Cominco Metals Ltd., E.D. Wash., No. 04-cv-256, 4/4/12).

The Confederated Tribes of the Colville Reservation, and the State of Washington as Plaintiff-Intervenor, filed a CERCLA action against Teck, alleging that the company had discharged slag and other hazardous substances into the Upper Columbia River, a Superfund site (the “UCR Site”) from its lead-zinc smelter across the border in British Columbia.

An earlier Ninth Circuit decision in the case discussed how Teck’s smelter had dumped slag waste into the Columbia River, ten miles north of the border, over several decades of operation, which resulted in pollution downstream in the United States. In 2003, the EPA placed the site on the National Priorities List. Thereafter, EPA issued a unilateral administrative cleanup order, which Teck failed to comply with. Initially, a lawsuit seeking enforcement of the order was brought. The State of Washington intervened in the action and amended its initial complaint to seek future CERCLA response costs and declaratory relief seeking natural resource damages.

A settlement between Teck and EPA followed, pursuant to which EPA withdrew the unilateral administrative cleanup order. Because enforcement of the order was then no longer at issue, there were no longer any pending claims by Plaintiffs Pakootas and Michel. Thus, what is at issue in the upcoming trial are the cost recovery and natural resource damages claims of the Tribes and the State.

Against this procedural backdrop, the motions before Judge Suko were the Tribes’ motion to dismiss Teck’s affirmative defense seeking to apportion liability and the State’s motion for partial summary judgment on Teck’s Divisibility Defense. At the outset, the court provided some helpful definitions of the technical CERCLA terms that would be discussed in the Opinion. For example, the judge explained that divisibility/apportionment is not a defense to liability itself. Rather, it is a judicially created defense to joint and several liability. Although “divisibility” and “apportionment” are terms often used interchangeably, what is potentially divisible is the harm, and if the harm is divisible, what it potentially apportions is liability, assuming a reasonable factual basis for apportionment. 

In opposing the motions, Teck argued that, even assuming it was liable under CERCLA, liability should be several, not joint and several, because the harm at issue is divisible. Pursuant to the Supreme Court’s landmark decision inBurlington Northern and Santa Fe Railway (“BNSF”) case, liability under CERCLA is generally joint and several unless a defendant meets its burden to prove the harm is divisible and capable of apportionment. Under the Restatement (Second) of Torts §433(A), the universal starting point for divisibility of harm analyses is “when two or more persons acting independently cause a distinct or single harm for which there is a reasonable basis for division according to the contribution of each, each is subject to liability only for the portion of the total harm that he has himself caused.” According to the Eighth Circuit’s decision in U.S. v. Hercules, “evidence supporting divisibility must be concrete and specific.”  

At the outset, it appeared that Teck had some probability of defeating the motions on the strength of an impressive array of expert witnesses. One environmental expert, Mark W. Johns, Ph.D., opined that there was no detectable release of hazardous substances from Teck’s barren slag and no evidence that dissolved metals from liquid effluent releases were located at the site. Using three different methodologies to apportion Teck’s liability for the harm at the UCR Site, Dr. Johns argued persuasively that Teck’s share of liability should be nothing or next to nothing 

Unfortunately for Teck, the court concluded that Teck had failed to present sufficient evidence to support its divisibility argument. Critical to the district court’s decision was its analysis of the term “harm.” Teck argued, based upon its reading of the BNSF decision, that the type of harm subject to apportionment was the alleged contamination from the leaching of metals traceable to the leaching of Teck’s slag and effluent. However, the Court ruled this argument missed the mark. The Ninth Circuit’s definition of harm, relied upon by Teck, was “for the purposes of determining divisibility,” not liability in the first instance, according to the district court. The court held that “the harm is the entirety of the contamination in the UCR site and what the Plaintiffs seek are recovery of costs to investigate and cleanup the entirety of that contamination…” The court continued:

"This contamination is not limited to metals which have been released or which threaten to be released from Teck’s slag and/or liquid effluent deposited in the UCR Site. None of Teck’s apportionment theories address the entirety of the contamination. Instead, they begin with the assumption that the only harm at issue is whatever metals were released from Teck’s slag and/or liquid effluent and the same metals which were released from non-Teck sources. This is a fatal flaw. Because Teck has not addressed the relevant harm in the first instance, it has failed to establish as a matter of law that the relevant harm is a single harm divisible in terms of degree".

In other words, Teck’s fatal flaw was in failing to account for all of the harm at the UCR Site. Because it did not do so, it would not prove that the harm it caused was divisible and thereby capable of apportionment.

For the CERCLA cost recovery practitioner, Pakootas makes for important reading, not only because of its cautionary holding, but because of its detailed analysis of other CERCLA cases, including BNSF, in which cases all of the harm at the respective sites was accounted for in determining that divisibility was possible. In short, Teck failed to consider the full range of environmental consequences at the UCR Site and, subject to post-trial appeal, may pay a steep price .