The rhetoric of enforcement under the Comprehensive Environmental Response, Compensation and Liability Act (“CERCLA”), 42 U.S.C. §§ 9601-75, plays well in public. The government seeks “cleanup” and makes “the polluter pay.” Cf. S. Rep. No. 96-848 at 13 (1980); Atlantic Richfield Co. v. Christian, 140 S. Ct. 1335, 1345 (2020). Contrast that with mitigating climate change where, in the words of the Pogo Earth Day poster, “we have met the enemy, and he is us.” Four years ago, then-new EPA Administrator Scott Pruitt described the Superfund program as the “cornerstone” of EPA’s “core mission” of protecting the environment and human health. Memorandum on Prioritizing the Superfund Program (May 22, 2017). (To be sure, as things turned out, emphasizing Superfund may not really have been what the last administration was about. See, e.g., Rule on Strengthening Transparency in Pivotal Science Underlying Significant Regulatory Actions and Influential Scientific Information, 86 Fed. Reg. 469 (Jan. 5, 2021).) If this next administration succumbs to the temptation to feature Superfund, perhaps it and we might focus on some of the mess the statute as drafted and read by the courts has made of private contribution claims, or a small corner of that mess.
The rhetoric plays better when the costs are higher and the work is more impressive. So, Superfund has become a tool used repeatedly to clean up sediments in industrial waterways, mining sites, and similar “mega sites.” By contrast, it does not even have a straightforward way to approach voluntary response to common problems that cause human exposure to hazardous substances, like PCB contamination within buildings to be redeveloped. And, of course, the “polluters” who are paying are not really the human beings who caused the releases. Even if the business entity remains the same, the shareholders, directors, officers, and employees who were associated with a release decades in the past are not the same as those who must pay today.
Large, complicated sites and a capacious notion of “polluter” lead, naturally, to a desire to reallocate costs through contribution or private cost recovery actions. But sections 113(f) (dealing with contribution), 113(g) (dealing with limitations periods), and 107 (dealing with cost recovery), produce a maze of procedural incoherence.
A CERCLA contribution claim seeks reallocation of costs the contribution plaintiff has incurred in excess of its fair share of the total. Arguably, then, the claim is premature until the plaintiff can allege that it has incurred that excess. That requires the plaintiff to be able to allege something about the total costs. But there is confusion – sometimes so profound that it is not expressed – among lawyers and courts about what actions or costs count as the total.
One natural approach would be to consider the total to be all costs incurred or to be incurred with respect to that facility. Indeed, one ordinarily would think that costs incurred during various stages of response would net against each other when deciding whether a party had paid more than its fair share. If I buy the baseball tickets, you would naturally buy the beer (at least up to the price of a ticket); if one party overpaid on a removal, it might get a compensating discount on the remedial action. So, under this approach, a party cannot recover in contribution unless it can allege that it has incurred more than its fair share of all costs.
But that might be very hard to do before a large portion of the remedial action had been completed. Section 113(f)(1) allows a contribution claim “during or after” a claim for cost recovery, and the limitations period for a contribution claim runs from entry of the judgment or entry into the settlement. See 42 U.S.C. § 113(f)(1), (g)(3). Thus, arguably some level of legal undertaking to incur future costs may count as “incurring” those costs for purposes of satisfying the requirement that a contribution plaintiff incur more than its fair share.
But even that may leave timing problems. The limitations period is three years from a trigger: a judgment or judicially approved settlement for costs or damages, or an administrative order settling costs under section 122(g) or (h). An early settlement for RI/FS costs, for example, could be more than three years before a record of decision, let alone entry into any sort of commitment by anyone to implement the remedy. The remedy could be divided into operable units, and they could be separated in time by more than the limitations period.
Treating each separate action as a stand-alone total would deny the netting principle. It will always be true that a party (other than a solely liable party) pays more than its fair share of every dollar that that party pays. But if you don’t allow for tasks to be divided up, you make settlement of cases with the government, or even among private parties, quite difficult.
None of this would be overly troubling if one could have confidence that courts would treat separate actions as simply parts of a single whole. If one knew that, then allowing the limitations period to expire on a small, early set of costs would simply mean that the settling up for those costs would occur in the contribution action for later costs.
But courts have, in some circumstances, decided that contribution may not even be available for later actions if they are conducted under orders or agreements that do not meet the requirements of Cooper Industries, Inc. v. Aviall Services, Inc., 543 U.S. 157 (2004), and Atlantic Research Corp. v. United States, 551 U.S. 128 (2007). That is, costs incurred complying with a Unilateral Administrative Order, for example, are not costs incurred “during or following any civil action” under section 106 or 107, nor are they incurred under a settlement. 42 U.S.C. §§ 9613(f)(1), (3)(B). Those costs would be recoverable from a jointly and severally co-responsible party through a cost recovery claim under section 107(a)(1-4)(B). Although typically the defendant would have a section 113 counterclaim, it is not clear how that would play out if some of the defendant’s (or the plaintiff’s) costs were incurred in a separate response action as to which the limitations period had run and how the accounting of paying more than one’s fair share would be accounted.
Finally, note that most courts would take equitable account of payments by insurers, indemnitors, or other contribution defendants in determining whether a contribution plaintiff has paid more than its share. If separate actions are separate totals for this purpose, then one must allocate those payments by action when the insurer or indemnitor did not allocate, and even if the contribution defendant was allocating, it may also have been netting.
There isn’t any reason for all this complexity. It is all readily fixable. We just haven’t fixed it.
