VERMONT JOURNAL OF ENVIRONMENTAL LAW
VERMONT LAW SCHOOL
Volume 11, Issue 2
Winter 2009
ARTICLES
Superfund at 30
Martha L. Judy & Katherine N. Probst.................................................... 191
The Superfund Cost Allocation Liability Conflicts Among the Federal
Courts
Steven Ferrey ........................................................................................... 249
Dis-Jointed? Several Approaches to Divisibility After Burlington Northern
Steve C. Gold ........................................................................................... 307
Restatement for Arranger Liability Under CERCLA: Implications of
Burlington Northern for Superfund Jurisprudence
Alfred R. Light.......................................................................................... 371
NOTE
Examining Contractual Models for Transferring Environmental Liability:
How They Work and Where They Are Headed
Kristin L. Hines ........................................................................................ 395
VERMONT JOURNAL OF ENVIRONMENTAL LAW
Vermont Law School
P.O. Box 96
South Royalton, Vermont 05068
(802) 831-1024
[email protected]
www.vjel.org
Cite to this Journal as: 11 VT. J. ENVTL. L. ___ (2009).
The views expressed in this issue are those of the authors and do not
represent the position or views of VJEL or Vermont Law School.
Submissions: VJEL welcomes the submission of unsolicited articles,
comments, essays, and book reviews. Manuscripts can be submitted to the
above addresses.
Subscriptions: The current bound volume can be purchased for $12.95.
Law libraries can subscribe through William S. Hein & Co. Others can
subscribe by directly contacting VJEL.
Copyright: © Copyright 2009 by Vermont Law School. All rights
reserved. Except as otherwise provided, the author of each article in this
issue has granted permission for copies of that article to be made for
classroom use, provided that: (1) copies are distributed at or below cost; (2)
the author and Vermont Journal of Environmental Law are identified on the
copied materials; (3) each copy bears the proper notice of copyright; and (4)
Vermont Journal of Environmental Law is notified in writing of the use of
the material(s).
The Journal is published by Western Newspaper Publishing Co. on
100% recycled paper and uses a soy-based ink.
VERMONT JOURNAL OF ENVIRONMENTAL LAW
VERMONT LAW SCHOOL
Volume 11, Issue 2
Winter 2009
2009–2010
EDITORIAL BOARD
EDITOR-IN-CHIEF
Kristin L. Hines
ADMINISTRATIVE EDITOR
Lillian Kortlandt
SENIOR MANAGING EDITOR
Andrew J. Klimkowski
SENIOR ARTICLES EDITOR
Joseph Barnett
SENIOR NOTES EDITOR
Jessica Scott
MANAGING EDITORS
Emily Cobb
Ferrell Ryan
SYMPOSIUM EDITOR
Genesis Wren Miller
HEAD NOTES EDITORS
Markell Ripps
Nathaniel Smith
Laurie Wheelock
WEB EDITOR
Sam Weaver
EVENTS EDITORS
Joseph Starnes
Louisa Yanes
ARTICLES EDITORS
Kevin Cooke
Shannon Sawyer
Andrew Schwartz
Abel Russ
PRODUCTION EDITORS
Daniel D. Burke
Christiana Cooley
Noel Hudson
Andrew Jacob Rouchka
Adam Sherwin
Tyler Soleau
INACTIVE MEMBERS
Caroline Jelavich
Ashley Santner
EDITORIAL STAFF
James Abraham
Scott Allen
Alice Baker
Addison Barnhardt
Mary Beth Blauser
Brent Bowden
Clare Cragan
Mark H. Foster, Jr.
Brandon Gillin
Crystal Heide
Ashley Hintz
Eric Hutchens
Kirby Keeton
Jenna Kennett
Melissa Krah
Dudley Loew
Brian E.J. Martin
Heather Marie McCarthy
Edalin Michael
Stephen Nadeau
John (Jack) Peters
Nathan Rectanus
Jessica Reiss
Shannon Salembier
Brian Selogie
Ian Sinderhoff
Sarah Stein
Laurie Stern
Andrew Stone
Peter Vetere
VERMONT JOURNAL OF ENVIRONMENTAL LAW
VERMONT LAW SCHOOL
Volume 11, Issue 2
Winter 2009
FACULTY ADVISORS
Jason J. Czarnezki
Professor of Law
Tseming Yang
Professor of Law
Christine Ryan
Environmental Research Librarian
ADMINISTRATION
Geoffrey B. Shields
President, Dean, and Professor of Law
Gil Kujovich
Vice Dean for Academic Affairs and Professor of Law
Lorraine Atwood
Vice President of Finance and Administration
Kathleen Hartman
Associate Dean for Enrollment Management
Shirley Jefferson
Associate Dean for Student Affairs and Diversity
Dorothy Behlen Heinrichs
Vice President for Institutional Advancement
VERMONT JOURNAL OF ENVIRONMENTAL LAW
VERMONT LAW SCHOOL
Volume 11, Issue 2
Susan Apel
Tracy L. Bach
Betsy Baker
Alexander W. Banks
Laurie Beyranevand
Teresa Clemmer
Liz Ryan Cole
Jason J. Czarnezki
Johanna Dennis
Michael Dworkin
Stephen Dycus
John Echeverria
Arthur C. Edersheim
Peg Elmer
Stephanie Farrior
Paul S. Ferber
David B. Firestone
Jackie A. Gardina
Oliver R. Goodenough
John M. Greabe
Winter 2009
FACULTY
Cheryl Hanna
Hillary M. Hoffmann
Gregory Johnson
Martha L. Judy
Donald Kreis
Gil Kujovich
Siu Tip Lam
Mark Latham
Reed E. Loder
Michelle Martinez
Campbell
James May
Michael McCann
David Mears
Philip Meyer
Marc Mihaly
Janet E. Milne
Laura Murphy
Sean Nolon
Cathryn C. Nunlist
Patrick Parenteau
Robert Gagnon
VISITING FACULTY
Jack Tuholske
Bonnie Barnes
Robin Barone
Emily Bergquist
Bud Carrey
John Evers
Randy Foose
ADJUNCT FACULTY
Clara Gimenez
Kevin Griffin
Shirley Jefferson
Matthew Levine
Eric Lopez
Larry Novins
Craig M. Pease
Brian Porto
Robert Rachlin
Anthony Renzo
Giuliana Robertson
Hilary Robinson
Geoffrey B. Shields
Linda Smiddy
Gus Speth
Pamela J. Stephens
Peter Teachout
Pamela Vesilind
Joan Vogel
Jeff White
Stephanie Willbanks
L. Kinvin Wroth
Tseming Yang
Carl A. Yirka
Maryann Zavez
William Rodgers
Jason Okai
William Russell
Anna Saxman
Stefanie Sidortsova
Judy Stern
SUPERFUND AT 30
Martha L. Judy & Katherine N. Probst∗
TABLE OF CONTENTS
Introduction ............................................................................................... 192
I. Cleaning Up Contaminated Sites .......................................................... 200
A. Program Accomplishments ............................................................. 205
B. NPL Listings/Deletions ................................................................... 206
C. Construction Complete .................................................................... 209
D. Government Performance Results Act Measures ............................ 210
E. Removal Actions.............................................................................. 212
II. Paying for Cleanup: Taxes and Liability ............................................. 213
A. Superfund Trust Fund ...................................................................... 216
B. Liability as a Funding Mechanism................................................... 221
C. Spreading Costs Among Private Parties .......................................... 230
III. Issues on the Horizon ......................................................................... 241
A. Can the Quality and Pace of Cleanups Be Improved? ..................... 242
B. What Steps Need to Be Taken to Ensure a Robust and Reliable Longterm Stewardship—or Post-cleanup—Program? ........................... 243
C. Will Private Parties Continue To Contribute Work and Money to
NPL Cleanups, and Will They Continue To Clean Up Sites on Their
Own?.............................................................................................. 244
D. What Are the Funding Needs for the Superfund Program over the
Next Five to Ten Years? ................................................................ 246
∗ Martha Judy’s research and teaching has focused on Superfund since she joined Vermont
Law School’s faculty in 1991. She has written numerous articles and contributed to briefs related to
Superfund liability, cost recovery, and contribution. She represented Vermont Law School as the
research director for the National Commission on Superfund, a group of environmentalists, corporate
CEOs, citizens, and government leaders seeking consensus on ways to improve the cleanup of
Superfund sites. Professor Judy earned her J.D. from Yale Law School in 1989. Katherine Probst is a
Senior Fellow at Resources for the Future, a think tank that works on environmental issues in
Washington, DC. She has authored numerous studies of the Superfund program, and she was the lead
author of Superfund’s Future: What Will It Cost?, a report requested by Congress. Ms. Probst received
her M.A. in city and regional planning from Harvard University in 1980.
192
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
E. Improving Superfund Requires Increased Program Evaluation and
Greater Transparency .................................................................... 247
INTRODUCTION
Congress enacted the Comprehensive Environmental Response,
Compensation, and Liability Act (CERCLA, or “Superfund” as it is more
commonly known)1 in the waning hours of the 96th Congress to address
contamination of land and water by hazardous substances dumped, spilled,
or otherwise released into the environment.2 Prior to CERCLA such
contamination was addressed, if at all, by common law causes of action
such as nuisance,3 trespass,4 and strict liability for ultra-hazardous
activities.5 Congress drew on this history to authorize the cleanup and the
recovery of costs associated with responding to releases of hazardous
substances into the environment.6 A lame-duck President, Jimmy Carter,
signed the bill into law during the final hours of the session on December
11, 1980.7
Two sites regularly featured on the television news and in news
magazines in the late 1970s and early 1980s set the stage for passage of
Superfund and provide a context for our understanding of the goals of the
statute.8 One site, “Valley of the Drums,” imprinted on the screen and in
the minds of the American public colorful images of erupting, smoking,
1. Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA) of
1980, Pub. L. No. 96-510, 94 Stat. 2767 (codified as amended at 42 U.S.C. §§ 9601–9675 (2006)).
2. The contamination at Love Canal, N.Y. is frequently cited as the site that spurred Congress
to act. See, e.g., Metro. Water Reclamation Dist. v. N. Am. Galvanizing & Coatings, Inc., 473 F.3d 824,
826–27 (7th Cir. 2007). For a discussion of the history of that site, see United States v. Hooker Chems.
& Plastics Corp., 850 F. Supp. 993 (W.D. N.Y. 1994) (providing an in-depth examination of the
contamination at Love Canal Channel, N.Y.).
3. E.g., Boomer v. Atl. Cement Co., 257 N.E.2d 870 (N.Y. 1970).
4. E.g., Martin v. Reynolds Metals Co., 342 P.2d 790 (Or. 1959).
5. See Atl. Research Corp. v. United States, 459 F.3d 827, 830 (8th Cir. 2006) (“CERCLA
effectively transformed centuries of real property and tort liability law by making those who
contaminate a site strictly liable for the costs of subsequent cleanup by others,”); Ronald G. Aronovsky,
Federalism and CERCLA: Rethinking the Role of Federal Law in Private Cleanup Cost Disputes,
33 ECOLOGY L.Q. 1, 9–10 (2006) (stating that strict liability for ultra-hazardous activity was a common
law action for private claims before CERCLA).
6. ROBERT V. PERCIVAL ET AL., ENVIRONMENTAL REGULATION: LAW, SCIENCE, AND POLICY
366–67 (5th ed. 2006).
7. The Daily Diary of President Jimmy Carter, December 11, 1980,
www.jimmycarterlibrary.gov/documents/diary/1980/d121180t.pdf.
8. James Bruggers, Valley of the Drums 30 Years Later, COURIER-JOURNAL (Louisville, KY),
Dec. 14, 2008, at A1, available at 2008 WLNR 26348599.
2009]
Superfund at 30
193
seeping, and corroding drums. News of the other site, “Love Canal,”
featured a brief hostage-taking of government officials by a charismatic
housewife turned activist named Lois Gibbs. Gibbs and others in her
community were concerned for their children and their suburban homes
after government officials issued test results suggesting “damaged
chromosomes” in some of her neighbors. The industrial waste site that
surrounded the community and on which the school was built became the
villain in the drama that ensued.9 Prior to the passage of CERCLA, state
and federal authorities struggled to respond to these events but found that
many of the environmental statutes passed earlier in the 1970s failed to
authorize remedial action that involved prior contamination of land or nonnavigable waters.
The goals of the Superfund law were to identify contaminated sites,
assure cleanup by either the federal government or private parties, make
those connected to the contamination pay for cleanup, and enlist private
interests in the identification and prevention of further hazards. While these
goals are not explicitly stated in the statute, courts infer CERCLA’s goals
from the legislative history and text of the statute. “The Comprehensive
Environmental Response, Compensation, and Liability Act (CERCLA) is
designed to promote the cleanup of hazardous waste sites and to ensure that
cleanup costs are borne by those responsible for the contamination.”10
“CERCLA’s dual goals are to encourage quick response and to place the
cost of that response on those responsible for the hazardous condition;”11
“CERCLA was enacted both to provide rapid responses to the nationwide
threats posed by the 30[,000]–50,000 improperly managed hazardous waste
sites in this country as well as to induce voluntary responses to those
sites.”12 Others involved in the early years, such as Philip Cummings, the
chief counsel of the Senate Environment Committee at the time of
CERCLA’s passage, interpreted the goals even more broadly:
CERCLA is not primarily an abandoned dump cleanup
program, although that is included in its purposes . . . . The
main purpose of CERCLA is to make spills or dumping of
9. See MARTIN LINSKY, CTR. FOR PRESS, POLITICS AND PUB. POLICY AT THE INST. OF
POLITICS, JOHN F. KENNEDY SCH. OF GOV’T, HARVARD UNIV., SHRINKING THE POLICY PROCESS: THE
PRESS AND THE 1980 LOVE CANAL RELOCATION (1985) (chronicling the events of Love Canal and Lois
Gibbs’s involvement).
10. Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870, 1872 (2009).
11. Control Data Corp. v. S.C.S.C. Corp., 53 F.3d 930, 936 (8th Cir. 1995).
12. United States v. Chem-Dyne Corp., 572 F. Supp. 802, 805 (S.D. Ohio 1983) (citing
5 U.S.C.C.A.N. 6119, 6119–6120 (1980)).
194
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
hazardous substances less likely through liability, enlisting
business and commercial instincts for the bottom line in
place of traditional regulation. It was a conscious intention
of the law’s authors to draw lenders and insurers into this
new army of quasi-regulators, along with corporate risk
managers and boards of directors.13
These goals are evident in CERCLA’s provisions giving the President
sweeping powers to respond to actual and threatened environmental
contamination of ambient air, surface and ground water, and land.
CERCLA empowered the President to clean up or otherwise respond to
releases of hazardous substances and releases of pollutants and
contaminants when these latter releases cause an imminent and substantial
endangerment.14 The President delegated most of his authority to the U.S.
Environmental Protection Agency (EPA).15
One of EPA’s first tasks was to amend the National Contingency Plan to
effectuate the statute’s purposes and to provide priorities, procedures, and
protocols for site identification, spill reporting, and emergency and longterm cleanup.16 The National Oil and Hazardous Substances Pollution
Contingency Plan (commonly referred to as the “National Contingency
Plan” or the “NCP”) was amended again in 1994 and serves to guide EPA
(and in some circumstances private party) actions related to CERCLA.17
EPA may clean up sites itself,18 and, if necessary, compel responsible parties
to undertake and to pay for site cleanups.19 The majority of cleanups are the
result of negotiated agreements whereby the parties reimburse EPA for its
costs (“cost recovery”), pay for and conduct site cleanups themselves
(“work agreements”), or both.20
13. PERCIVAL ET AL., supra note 6, at 367–68 (quoting Philip Cummings, Completing the
Circle, ENVTL. F. 11 (Nov–Dec. 1990)).
14. 42 U.S.C. § 9604(a)(1) (2006).
15. Exec. Order No. 12,580, 52 Fed. Reg. 2923 (Jan. 23, 1987).
16. See 42 U.S.C. § 9605(a)–(b) (2006) (requiring the National Contingency Plan to be revised
and republished to “specify procedures, techniques, materials, equipment, and methods to be employed
in identifying, removing, or remedying releases of hazardous substances . . . .”).
17. 59 Fed. Reg. 47,384 (Sept. 15, 1994) (codified at 40 C.F.R. §§ 300.1–300.3 (2008)).
18. See § 9604(a)(1) (stating that the President is authorized to “remove or arrange for removal
of, and provide for remedial action related to such hazardous substance, pollutant, or contaminant”).
19. See 42 U.S.C. § 9606(a) (2006) (“[T]he President may . . . issu[e] such orders as may be
necessary to protect public health and welfare and the environment.”).
20. U.S. GOV’T ACCOUNTABILITY OFFICE, GAO-09-656 SUPERFUND: LITIGATION HAS
DECREASED AND EPA NEEDS BETTER INFORMATION ON SITE CLEANUP AND COST ISSUES TO ESTIMATE
FUTURE PROGRAM FUNDING REQUIREMENTS, GAO-09-656, at 27 tbl.5 (2009) [hereinafter GAO-09656].
2009]
Superfund at 30
195
CERCLA also created a dedicated trust fund, financed primarily by
excise taxes on petroleum and chemical feedstocks, to enable the
government to pay for cleanups when the parties responsible either could
not be found, or were recalcitrant.21 While the title of the law includes the
word “compensation,” CERCLA does not, in fact, require compensation for
health damages.
CERCLA’s expansive liability scheme both preserves trust fund
resources (any site study or cleanup paid for and conducted by potentially
responsible parties (PRPs) is money saved) and also restores monies to the
trust fund through cost recovery actions and revenues from penalties.22 The
liability scheme provides a strong incentive for PRPs to negotiate
settlements with EPA,23 and it also provides incentives for pollution
prevention and voluntary cleanup of sites not listed on the National
Priorities List (NPL) or subject to federal enforcement.24 An August 2009
Government Accountability Office (GAO) report valued these contributions
to the trust fund and cleanup at near thirty billion dollars from all EPA
enforcement activities, including the recovery of past and future costs,
private party work commitments, and penalties.25
Superfund liability is retroactive,26 strict, and often joint and several.
Liability is retroactive because it applies to contamination that occurred
before CERCLA was enacted in 1980. Liability is strict because a
responsible party is liable even if it was not negligent. As in other forms of
tort law when the harm is not divisible, liability is joint and several; that is,
the government may hold one or more parties liable for the full costs of
cleanup, even if there are other parties at the site.27
In 1980, Congress gave the President considerable leverage to make
sure sites are cleaned up and costs are paid by private parties. In addition to
the broad liability provisions and presidential authority to compel private
21. TAX POLICY CTR., URBAN INST. AND BROOKINGS INST., REINSTATE SUPERFUND TAXES,
http://www.taxpolicycenter.org/taxtopics/2010_budget_superfund.cfm.
22. GAO-09-656, supra note 20, at 79.
23. Id. at 27. Nearly eighty percent of all EPA enforcement actions end in settlement. Id. at
27, tbl.5. After 1994 on average over half of the cases the government files against private parties have
pre-negotiated consent decrees that greatly decrease the time in litigation. Id. at 40 fig.6. Only twenty
percent of EPA enforcement actions at NPL sites involved judgments or unilateral administrative orders.
Id. at 25 tbl.4. Generally, though not always, unilateral administrative orders authorized by CERCLA
section 106 are reserved for recalcitrant parties. Id. at 29.
24. Id. at 10 n.7.
25. Id. at 31 tbl.7.
26. United States v. Olin Corp., 107 F.3d 1506, 1511–12 (11th Cir. 1997); United States v.
Mottolo, 695 F. Supp. 615, 622 (D.N.H. 1988).
27. Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870, 1881 (2009).
196
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
party cleanup, Congress provided only three narrow defenses to
CERCLA.28 In the early years, the government’s powers of persuasion
were all the clearer because the ease of proving liability required evidence
as little as a deed showing who owned the contaminated property.29 Over
time, amendments to CERCLA,30 judicial decisions,31 and, in some cases,
administrative reforms32 have softened CERCLA’s hard edges by providing
more clarity about the innocent landowner defenses,33 adding limited
exemptions,34 creating more opportunity to spread costs among parties,35
28. Congress set CERCLA’s cost recovery and liability provisions apart from other statutes and
defenses when it began section 107 as follows: “Notwithstanding any other provision or rule of law, and
subject only to the defenses set forth in subsection (b) of this section . . . .” 42 U.S.C. § 9607(a) (2006).
Section 107(b) protected parties from releases caused solely by acts of God or acts of war (onsite), and
protected innocent landowners that looked for contamination and found none but somehow ended up
owning contaminated property through no fault of their own. § 9607(b). This last defense encompasses
protection from actions by unaffiliated third parties as long as the landowner took steps to prevent such
contamination from those who inherited contaminated property, and from governments that acquire
property involuntarily or through eminent domain authorities. Id. Few of these defenses were
particularly effective to shield potentially liable parties, because even the innocent landowner defenses
were narrowly written to require an extensive (and ambiguous) pre-purchase investigation into prior site
contamination that almost no one could meet and also have hazardous substances on the property.
29. § 9607(a)(1).
30. Superfund Amendments and Reauthorization Act (SARA) of 1986, Pub. L. No. 99-499,
100 Stat. 1613 (codified at 42 U.S.C. § 9601 (2006)); Asset Conservation, Lender Liability, and Deposit
Insurance Protection Act of 1996, Pub. L. No. 104-208, § 107, 110 Stat. 3009–462 (codified at 42
U.S.C. §§ 9601, 9607(n) (2006)); Superfund Recycling Equity Act of 1999, Pub. L. No. 106-113, 113
Stat. 1536 (codified at 42 U.S.C. § 9627 (2006)); Small Business Liability Relief and Brownfields
Revitalization Act of 2002, Pub .L. No. 107-118, 115 Stat. 2356 (2002) (codified as amended in
scattered sections of 42 U.S.C. §§ 9601, 9604, 9605, 9607, 9622(g)) (amending CERCLA to add
additional circumstances wherein parties may avoid liability if certain conditions are met (e.g., the de
micromis exemption, section 107(o), and contiguous landowner exemption, section 107(q)).
31. See, e.g., United States v. Best Foods, 524 U.S. 51, 68–71 (1998) (holding that a parent
corporation cannot be held liable as an operator under CERCLA for the actions of its subsidiary without
evidence of actual control of the facility).
32. See GAO-09-656, supra note 20, at 11–12 (summarizing the sixty-two administrative
reforms, many of which were enacted in the Small Business Liability Relief and Brownfields
Revitalization Act).
33. Superfund Amendments and Reauthorization Act; Small Business Liability Relief and
Brownfields Revitalization Act (amending CERCLA to add additional circumstances wherein parties
may avoid liability if certain conditions are met, e.g., contiguous landowner exemption, section
9607(q)); 40 C.F.R. § 312 (2008) (establishing regulations, required by the Small Business Liability and
Relief and Brownfields Revitalization Act section 223, to clarify all appropriate inquiry sufficient to
protect owners from liability).
34. Asset Conservation, Lender Liability, and Deposit Insurance Protection Act of 1996, Pub.
L. No. 104-208, § 107, 110 Stat. 3009–462 (codified at 42 U.S.C. §§ 9601, 9607(n) (2006)); Superfund
Recycling Equity Act of 1999, Pub. L. No. 106-113, 113 Stat. 1536 (codified as 42 U.S.C. § 9627
(2006)); Small Business Liability Relief and Brownfields Revitalization Act § 102(a) (amending
CERCLA to add additional circumstances wherein parties may avoid liability if certain conditions are
met (e.g., de micromis exemption, § 9607(o); municipal solid waste exemption, § 9607(p)).
2009]
Superfund at 30
197
voiding CERCLA’s provisions creating strict liability for states,36 limiting
liability for parent corporations and bankrupt entities,37 and clarifying the
extent of insurance coverage.38
While private parties agonized over the stringent liability provisions,
the focal point of the Superfund program for the general public, however,
was and is the NPL. The NPL is commonly referred to as the list of the
highest priority sites in the country. What makes NPL sites unique is that
these are the only sites where trust fund monies can be used for long-term
cleanups called “remedial actions.”39 In the original legislation, Congress
required EPA to “list” 400 NPL sites to get the program up and running.40
At the end of FY 2009, EPA had listed 1607 final and deleted sites on the
NPL41 of over 47,000 sites that EPA evaluated for potential listing.42
While the NPL is a major focus of the Superfund program, EPA can
pursue enforcement actions at any site where hazardous substances are
releasing or threatening release into the environment, not just NPL sites. It
can also implement emergency cleanup or other, usually shorter-term,
cleanups called “removal actions” at any site in the country.43 Removal
actions can take many forms: they can be used to provide an alternate
source of drinking water where water has been contaminated, to remove
35. Superfund Amendments and Reauthorization Act of 1986, Pub. L. No. 99-499, § 113, 100
Stat. 1647 (codified as 42 U.S.C. § 9613 (2006)); United States v. Atl. Research Corp., 551 U.S. 128
(2007).
36. See Seminole Tribe of Fla. v. Florida, 517 U.S. 44, 62 (1996) (overruling Pennsylvania v.
Union Gas Co., 491 U.S. 1, 17 (1989)) (stating that, in situations where states are unable to legislate,
neither Congress nor anyone else may obtain money damages from the states).
37. See United States v. Best Foods, 524 U.S. 51, 68 (1998) (The Court held that “a
participation-and-control test looking to the parent’s supervision over the subsidiary, especially one that
assumes that dual officers always act on behalf of the parent, cannot be used to identify operation of a
facility resulting in direct parental liability.”); In re Chateaugay Corp. v. LTV Corp., 944 F.2d 997, 1004,
1008 (2d Cir. 1991) (ruling that EPA’s future response costs are dischargeable claims in bankruptcy if
the release or threatened release of hazardous substances occurred pre-petition, as are injunctive
remedies, to the extent that they impose obligations distinct from any obligations to stop or ameliorate
ongoing pollution).
38. See, e.g., Morton Int’l v. Gen. Accident Ins. Co., 629 A.2d 831, 878–79 (N.J. 1993)
(limiting insurer liability).
39. 40 C.F.R. § 300.425 (2008).
40. Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA) of
1980, Pub. L. No. 96-510, § 105, 94 Stat. 2767.
41. U.S. ENVTL. PROT. AGENCY, NAT’L PRIORITIES LIST: NPL SITE TOTAL BY STATUS AND
MILESTONE (Oct. 27, 2009), www.epa.gov/superfund/sites/query/queryhtm/npltotal.htm.
42. GAO-09-656, supra note 20, at 13 n.11.
43. 42 U.S.C. § 9604(a)(1) (2006); United States v. W.R. Grace & Co., 429 F.3d 1224, 1227–
28, 1237 (9th Cir. 2005) (observing that removal actions are usually “time-sensitive” but then
concluding that EPA’s several year cleanup of a Libby, Montana asbestos mine and processing site for
asbestos-containing vermiculite qualified as a removal action).
198
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
barrels of chemicals likely to explode at a site, or to relocate families at
imminent risk from the contamination.44 Removal actions also may take the
form of long-term action (“non-time critical”) in anticipation of settlement
with private parties, listing on the NPL, or in combination with other
authorities to address contamination,45 such as the Resources Conservation
and Recovery Act’s (RCRA) Corrective Action program..46 Funding for
removal actions implemented by the government is generally restricted to
$2,000,000 and 12 months, unless the President has determined that the
action is consistent with a remedial action.47
Additionally, private parties can agree to conduct an EPA-supervised
cleanup of a site that might otherwise be listed on the NPL. Federal
funding of the cleanup will not be necessary because private parties are
cooperating with the cleanup. Under these circumstances, EPA may decide
against listing the site on the NPL since EPA authority (other than funding)
extends equally to listed and non-listed sites.48 The agency often refers to
such federally supervised private party response actions on sites that
otherwise would be listed on the NPL as “Superfund Alternative” sites.49
So far, sites officially designated as Superfund Alternative sites are
relatively few in number. Only twenty-two sites were so designated as of
September 2007, with another forty more sites under consideration for
inclusion in the program.50 EPA may also use its authority under the RCRA
Corrective Action program for certain sites rather than listing them on the
NPL. Many sites do not receive much EPA attention;51 however, these sites
may also be cleaned up under Superfund authority by Indian tribes, state or
local governments, or, most frequently, private parties.52
44. § 9604(a)(1); 42 U.S.C. § 9601(23) (2006).
45. OFFICE OF SOLID WASTE & EMERGENCY RESPONSE, U.S. ENVTL. PROT. AGENCY,
CONDUCTING NON-TIME-CRITICAL REMOVAL ACTIONS UNDER CERCLA, PUB. NO. 9360.0-32FS
(1993).
46. 42 U.S.C. §§ 6924(a), (u), (v), 6925(c) (2006).
47. § 9604(c)(1)(A)–(C).
48. See § 9604(a)(1) (“Whenever . . . any hazardous substance is released or there is a
substantial threat of such a release into the environment, . . . the President is authorized to act, consistent
with the national contingency plan . . . .”); see also 42 U.S.C. § 9605(h) (2006) (“NPL Deferral”).
49. See generally Memorandum from Susan E. Bromm, Dir., Office of Site Remediation
Enforcement (OSRE) & Michael B. Cook, Dir., Office of Superfund Remediation & Technology
Innovation (OSRTI), to Superfund Nat’l Policy Managers, Regions I–X & Reg’l Counsel, Regions I–X,
Revised Response Selection and Settlement Approach for Superfund Alternative Sites (June 17, 2004),
http://www.epa.gov/compliance/resources/policies/cleanup/superfund/rev-sas-04.pdf.
50. GAO-09-656, supra note 20, at 52 n.56.
51. See id. at 39 tbl.8 (listing CERCLA cases filed by plaintiff: private parties accounting for
fifty percent; federal government for thirty-seven percent; and state government for thirteen percent).
52. 42 U.S.C. § 9607(a)(4) (2006).
2009]
Superfund at 30
199
When the law was first enacted, estimates varied widely about the
number of sites that would warrant federal attention, as well as about the
likely cost of the program. Some thought that cleanup of contaminated
sites would be done after the initial five-year authorization and funding;
others estimated that there were thousands of sites that would need to be
addressed, with a price tag of fifty billion dollars.53
Now, almost thirty years later, Congress has appropriated a total of
$33.4 billion to the EPA for the Superfund program.54 Although a recent
GAO report estimated that PRP work at NPL sites was worth $22.5 billion
through FY 2007,55 just how much private parties have spent is unknown, as
they are not required to report actual spending, even for actions at NPL
sites.
We know now that addressing the risk at Superfund sites is a lengthy,
expensive, and complicated endeavor. Many, if not most, sites will require
some kind of monitoring and maintenance for decades. Thus, while some
would like to see the Superfund program go away, it is likely to be with us
for many more years. Therefore, it is worth considering what we knew
about the scope of the problem when the program was first begun, how the
major elements of the program have evolved, and what issues are on the
horizon as the Superfund program enters its fourth decade.
In this article we examine two major areas of the Superfund program:
(1) the cleanup program and its accomplishments; and (2) cleanup funding,
including both the trust fund and liability provisions as mechanisms to pay
for site cleanup. For each of these topics we first provide some background
on what the original CERCLA legislation and early program guidance
called for, and then describe how litigation, administrative changes, and/or
statutory amendments shaped the cleanup of contaminated sites for the next
thirty years. Where information is available, we provide data on relevant
program accomplishments to date. We close with a discussion of a handful
of issues we believe are critical to improving the Superfund program in the
future.
53. David F. Salisbury, Superfund Set to Start Cleaning Up Abandoned Hazardous Waste Sites,
CHRISTIAN SCI. MONITOR, Dec. 8, 1980, at 7.
54. E-mail from Alan Youkeles, Associate Branch Chief, Budget, Planning and Evaluation
Branch, Office of Superfund Remediation & Technology Innovation, Office of Solid Waste and
Emergency Response, U.S. Environmental Protection Agency, to authors (Aug. 24, 2009, 2:00 p.m.
EST) (on file with authors).
55. GAO-09-656, supra note 20, at 30.
200
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
I. CLEANING UP CONTAMINATED SITES
Most Superfund stakeholders agree that the cleanup of sites
contaminated with hazardous substances is the fundamental objective of the
Superfund statute. That said, just what is meant by “cleanup” has been the
subject of much debate and disagreement ever since the program’s
inception.56
When Congress enacted CERCLA in 1980, the statute provided only
vague guidance regarding the objectives that cleanup was to achieve.
Section 104(c)(4) provided that:
[T]he President shall select appropriate remedial actions
determined to be necessary to carry out this section which
are to the extent practicable in accordance with the national
contingency plan and which provide for that cost-effective
response which provides a balance between the need for
protection of public health and welfare and the
environment at the facility under consideration, and the
availability of amounts from the Fund established under
Title II of this Act to respond to other sites which present
or may present a threat to public health or welfare or the
environment, taking into consideration the need for
immediate action.57
The lack of clarity of this section was one of the reasons why Congress
enacted major substantive amendments to CERCLA with the 1986
Superfund Amendments and Reauthorization Act (SARA).58 Even after
these amendments, the Superfund statute does not require national sitespecific cleanup standards that apply to all sites. Instead, SARA established
certain criteria for EPA to consider in selecting a remedy for each site. The
statute gives preference to remedial actions in which treatment
56. See, e.g., ENVTL. DEF. FUND ET AL., RIGHT TRAIN WRONG TRACK: FAILED LEADERSHIP IN
SUPERFUND CLEANUP PROGRAM 1 (1988) [hereinafter RIGHT TRAIN] (comparing initial cleanup
levels with cleanup levels eight years after enactment of CERCLA and finding them “virtually
indistinguishable”).
57. Comprehensive Environmental Response, Compensation, and Liability Act of 1980, Pub.
L. No. 96-510, § 104(c)(4).
58. Superfund Amendments and Reauthorization Act (SARA) of 1986, Pub. L. No. 99-499,
100 Stat. 1613, § 121(a) (codified at 42 U.S.C. § 9621(a) (2006)).
THE
2009]
Superfund at 30
201
“permanently and significantly reduces the volume, toxicity or mobility of
the hazardous substances, pollutants, and contaminants.”59
EPA currently establishes cleanup standards at each site based on sitespecific risk assessments and requirements under other statutes, commonly
referred to as “ARARs” (Applicable or Relevant and Appropriate
Requirements). Other criteria used in evaluating a remedy include: overall
protection of human health and the environment, compliance with ARARs,
short-term effectiveness, implementability, cost, and state and local
acceptance.60 Because these criteria often conflict, EPA developed a
complicated method for applying them, which includes categorizing the
nine criteria as “threshold,” “primary balancing,” or “modifying.”61 EPA
has considerable flexibility in its selection of a remedy as it attempts to
reconcile the criteria with the interests of the affected parties. Based on
these somewhat loose criteria, EPA selects a remedy at each site. In sum,
section 121 of CERCLA provides some guidance to EPA, but still leaves
much room for discretion—and controversy.
Many Superfund cleanup plans in the early years of the program,
however, failed to state these cleanup goals clearly or, in some cases, failed
to establish cleanup standards.62 For many years, the remedy selection
process was criticized for:
* Excessive Complexity — A cumbersome, repetitive, and slow
process created by the complexity of evaluating ARARs, risk assessments,
and alternative remedies;63 incompatible and inadequate management of
information important to the remedy selection process.64
* Unclear Standards — Cleanup standards and goals that are
determined (if at all) after an in-depth evaluation of all alternative remedies
instead of setting the goal first and then studying only the remedies that
meet that goal;65 determination of cleanup remedies based on the available
technologies, not the cleanup standards, since some cleanup standards
59.
60.
61.
62.
42 U.S.C. § 9621(b)(1) (2006).
40 C.F.R. § 300.430(e)(9)(A)–(I) (2008).
Id. § 300.430(f)(1)(i)(A)–(C).
U.S. GEN. ACCOUNTING OFFICE, GAO/RCED-92-138 SUPERFUND: PROBLEMS WITH THE
COMPLETENESS AND CONSISTENCY OF SITE CLEANUP PLANS 30 (1992).
63. E. Donald Elliott, Superfund: EPA Success, National Debacle?, 6 NAT. RESOURCES &
ENV’T 11, 12 (1991–1992).
64. U.S. GEN. ACCOUNTING OFFICE, GAO/PEMD-90-3 HAZARDOUS WASTE: EPA’S
GENERATION AND MANAGEMENT DATA NEED FURTHER IMPROVEMENT 2 (Feb. 1990).
65. CLEAN SITES, IMPROVING REMEDY SELECTION: AN EXPLICIT AND INTERACTIVE PROCESS
FOR THE SUPERFUND PROGRAM, at iv (1990).
202
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
cannot be achieved with present technologies;66 failure to define clear
priorities, such as how current and future risks are to be balanced.67
* Inconsistent Remedy Selection — Inconsistent application of the
statutory criteria, resulting in very different remedies for similar sites;68
giving too much69 or too little70 emphasis to cost in the selection of a
remedy; giving too little emphasis to baseline risk assessments of sites
where humans are not currently exposed to contamination;71 inadequate
consideration of the risks and damage to human health and the environment
posed by the implementation of the selected remedy;72 infrequent use of
treatability studies (laboratory or small-scale tests to determine the
effectiveness of technologies on the specific waste at a site) in evaluating
remedies and resolving technical disputes among PRPs, EPA, states, and
community groups;73 failure to consider current and future land use at the
site when establishing a cleanup standard or selecting a remedy.74
66. See Dr. Joel S. Hirschhorn, Definition and Analysis of Stubborn Superfund Problems, in
WORKING PAPERS ON SUPERFUND REFORM: PROBLEM DEFINITION AND POLITICAL MAPPING 55, 69
(1992) (recognizing that cleanup requirements originally set for a site may not determine what is an
acceptable cleanup because the cleanup may be limited by the performance of the technology chosen).
67. Carolyn B. Doty & Curtis C. Travis, The Superfund Remedial Action Decision Process: A
Review of Fifty Records of Decision, 39 J. AIR & WASTE MGMT. ASS’N 1535, 1537–38 (1989).
68. Id. at 1538.
69. See, e.g., FRANK R. LAUTENBERG & DAVE DURENBERGER, SENATE SUBCOMM. ON
SUPERFUND, OCEAN, & WATER PROT., LAUTENBERG-DURENBERGER REPORT ON SUPERFUND
IMPLEMENTATION: CLEANING UP THE NATION’S CLEANUP PROGRAM 63–64 (1989); Doty & Travis,
supra note 67, at 1538 (studying fifty RODs and finding “cost to be the most significant factor in the
selection of remedial alternatives”).
70. John Butler, III, Superfund: Super Costs, in RETHINKING SUPERFUND: IT COSTS TOO MUCH,
IT’S UNFAIR, IT MUST BE FIXED 67, 69 (1991).
71. Doty & Travis, supra note 67, at 1537 (observing that in this study of thirty-six sites for
which a quantitative risk assessment was conducted, all but seven sites had risk levels prior to cleanup
that were within the risk range required after cleanup yet remedial action often occurred, suggesting that
baseline risk assessments, where no current human exposure exists, does not play a significant role in
deciding whether to remediate a site).
72. 1 COMM. ON ENVTL. EPIDEMIOLOGY, NAT’L RESEARCH COUNCIL, ENVIRONMENTAL
EPIDEMIOLOGY: PUBLIC HEALTH AND HAZARDOUS WASTES 20 (1991).
73. CLEAN SITES, MAKING SUPERFUND WORK: RECOMMENDATIONS TO IMPROVE PROGRAM
IMPLEMENTATION 22 (1989).
74. Administration of the Federal Superfund Program: Hearings Before the Subcommittee on
Investigations and Oversight of the Committee on Pubic Works and Transportation, 100 Cong. 1 (1991)
(statement of Joel Robinson, Director, Environmental Science, Unical, Representing the American
Petroleum Institute) (SUDOC: Y4.P96/11:102-44 at 325); Daniel Koshland, Jr., Toxic Chemicals and
Toxic Laws, 253 SCI. 949, 949 (1991). But see, Donald A. Brown, What Is Wrong with the National
Contingency Plan?, 20 ENVTL. L. REP. NEWS & ANALYSIS 10371, 10373 (1990) (citing 40 C.F.R.
§ 300.430(a)(1)(iii)(B)–(D) (1990)) (arguing that the new National Contingency Plan allows EPA to use
land use controls and other institutional controls for groundwater cleanup when other measures are not
practicable).
2009]
Superfund at 30
203
* Questionable Results — Focus on the issue of how to choose
cleanup remedies for sites to the detriment of practical consideration of
whether the remedies chosen will actually work;75 inadequate assessment
and protection of ecological concerns;76 inconsistent remedy selection and
completeness depending on whether the PRP or the EPA conducts the
remedial investigation and feasibility studies;77 slower cleanup and more
frequent selection of lower quality remedies at sites in communities of
color;78 failure to select permanent remedies;79 failure to consider
“treatment trains,” consisting of a combination of technologies, which
together can clean up the site;80 over reliance by EPA on incompetent and
expensive contractors.81
* Inadequate Public Involvement — Lack of timely
communication about the selection process and opportunity for public input
prior to selection of the remedy, essentially closing the public out of the
reasoning for and participation in the most important decisions.82
In response to the many external studies criticizing these various
aspects of the remedy selection process, EPA implemented many reforms to
75. U.S. GEN. ACCOUNTING OFFICE, GAO/T-RCED-92-15 SUPERFUND: ISSUES THAT NEED
BE ADDRESSED BEFORE THE PROGRAM’S NEXT REAUTHORIZATION 10 (1991) (discussing problems
associated with clay caps for containment, and the problems associated with pump and treat for
groundwater); Doty & Travis, supra note 67, at 1542 (criticizing the remedy selection process for a lack
of explanation or analysis of the effectiveness of selected remedies, including failure to assess the
degree of risk reduction provided by remedial alternatives in all but twelve percent of fifty RODs
studied).
76. Doty & Travis, supra note 67, at 1539; RIGHT TRAIN, supra note 56, at 5; ENVTL. DEFENSE
FUND ET AL., TRACKING SUPERFUND: WHERE THE PROGRAM STANDS 4–5 (1990).
77. U.S. GEN. ACCOUNTING OFFICE, GAO/T-RCED-92-69 SUPERFUND: PROBLEMS WITH THE
COMPLETENESS AND CONSISTENCY OF SITE CLEANUP PLANS 12 (1992).
78. Marianne Lavelle & Marcia Coyle, Unequal Protection: The Racial Divide in
Environmental Law, THE NAT’L L.J., Sept. 21, 1992, at S2.
79. Doty & Travis, supra note 67, at 1541; RIGHT TRAIN, supra note 56, at 4; ENVTL. DEFENSE
FUND ET AL., supra note 76, at 3–4.
80. RIGHT TRAIN, supra note 56, at 4; ENVTL. DEFENSE FUND ET AL., supra note 76, at 3.
81. U.S. GEN. ACCOUNTING OFFICE, GAO/T-RCED-92-37 FEDERALLY SPONSORED
CONTRACTS: UNALLOWABLE AND QUESTIONABLE INDIRECT COSTS CLAIMED BY CH2M HILL (1992);
U.S. GEN. ACCOUNTING OFFICE, GAO/RCED-92-45 SUPERFUND: EPA HAS NOT CORRECTED LONGSTANDING CONTRACT MANAGEMENT PROBLEMS (1991); U.S. GEN. ACCOUNTING OFFICE, GAO/TRCED-91-5 EPA’S CONTRACT MANAGEMENT: AUDIT BACKLOGS AND AUDIT FOLLOW-UP PROBLEMS
UNDERMINE EPA’S CONTRACT MANAGEMENT (1990); U.S. GEN. ACCOUNTING OFFICE, GAO/RCED89-57 SUPERFUND CONTRACTS: EPA’S PROCEDURES FOR PREVENTING CONFLICTS OF INTEREST NEED
STRENGTHENING (1989).
82. RIGHT TRAIN, supra note 56, at 4; ENVTL. DEFENSE FUND ET AL., supra note 76, at 4;
Hirschhorn, supra note 66, at 72; CLEAN SITES, MAKING SUPERFUND WORK, supra note 73, at 23.
TO
204
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
the Superfund program.83 Controversy about the pace and quality of
cleanup persists.84 In the past eight years, however, there has been much
less congressional oversight of Superfund cleanups than in the earlier years
of the program and a dearth of independent evaluations of this aspect of the
program. As a result, it is difficult to ascertain the quality of recent remedy
decisions or of the remedy selection process. Notwithstanding these data
gaps, the most controversial aspect of Superfund remedies continues to be
the basic question: “How clean is clean?” That is, is the goal of site
cleanups to: (1) remove all contamination at a site (assuming this is
technically possible), or is it to (2) reduce risk to an acceptable level, which
can often be accomplished by limiting exposure to contaminants left onsite?
While there are many nuances to this argument, this has been a major
issue dividing critics of the Superfund program. Some believe that
cleanups are not stringent enough,85 whereas others say that cleanups are
too stringent and not cost-effective.86 Some have argued that the law, by
requiring permanent remedies, requires that contaminated sites be restored
to “pre-contamination” conditions; others believe this to be impractical and
a waste of money, and argue that risks can be greatly reduced by containing
exposure, that is, ensuring that people do not come into contact with any
contaminants left on site (or in ground or other waters).87
While these disagreements regarding the proper cleanup standards and
remedy selection criteria were among the most divisive issues during the
early and middle years of the program, in the last decade, focus on this
issue has waned at the national level (although this problem still arises at
individual sites). This is due to a number of factors. Under the George W.
Bush administration there was simply less attention paid to Superfund in
general, both in the White House and Congress. In addition, the focus of
national environmental organizations shifted to climate and other issues,
and few had senior staff following the Superfund program. And, the reality
83. U.S.
Envtl.
Prot.
Agency,
Superfund
Reforms,
http://www.epa.gov/superfund/programs/reforms (last visited Oct. 27, 2009).
84. See RENA STEINZOR & MARGARET CLUNE, THE TOLL OF SUPERFUND NEGLECT: TOXIC
WASTE DUMPS AND COMMUNITIES AT RISK (2007) (documenting cleanup and funding failures at fifty
toxic waste sites in ten of the most populated states.)
85. Id.
86. LISA GREY, CRS REPORT TO CONGRESS: CLEANUP STANDARDS RECONSIDERED (1995).
87. See, e.g., JAMES V. DE LONG, CATO INST., PRIVATIZING SUPERFUND: HOW TO CLEAN UP
HAZARDOUS WASTE (1995); JAMES T. HAMILTON & W. KIP VISCUSI, CALCULATING RISKS?: THE
SPATIAL AND POLITICAL DIMENSIONS OF HAZARDOUS WASTE POLICY 164 (Nancy L. Rose & Richard
Schmalensee eds., 1999).
2009]
Superfund at 30
205
of the challenges of site cleanup, both in terms of cost and technical
feasibility, came into play as well. It became clear that for some sites there
simply were not effective technologies to remove contamination, and that
for other sites, notably large mining sites and contaminated waterways,
cleanup was so expensive that complete cleanup may not be achievable or
is decades away.
Also, the types of sites EPA placed on the NPL began to change, and
the sites included on the NPL have become more complex and expensive.
According to a U.S. GAO report released in July 2009, for sites expected to
cost fifty million dollars or more to complete (“mega sites”), the median
duration is 14.8 years to reach the “construction complete” phase.88 For
non-mega sites, the average duration of cleanup activities is 10.1 years.89
As the program matured, it became clear that contamination would be
left at many sites. On-site cleanup techniques include: containing the
contamination on sites with soil and textile caps, impermeable landfill
liners, underground slurry walls, fences, bioremediation and ground water
pump-and-treat, and monitoring of groundwater and other contamination.90
These on-site remedies often required institutional controls to keep people
from coming into contact with contamination both now and in the future.
Such controls include: permit restrictions, zoning changes, public notices,
advisories and other warnings, and easements and deed restrictions.91
A. Program Accomplishments
There are many ways to examine the progress of the Superfund
program. For sites on the NPL, one can look at how many sites have been
added to the NPL each year, how many sites have been deleted from the
NPL, the number of sites where the remedy is actually constructed, or the
number of sites where Government Performance and Results Act (GPRA)
goals have been achieved. Other milestones are how many site cleanups
have been completed, the number and value of settlements reached with
responsible parties to compensate EPA for cleanup activities or to agree to
undertake various site-specific activities, and the number of unilateral
administrative orders issued to force private party cleanup.
88. GAO-09-656, supra note 20, at 70 tbl.15.
89. Id.
90. OFFICE OF SITE REMEDIATION AND ENFORCEMENT, U.S. ENVTL. PROT. AGENCY,
REVITALIZING CONTAMINATED SITES: ADDRESSING LIABILITY CONCERNS 38 (2008).
91. Id. at 39.
206
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
In addition, success can be characterized by the number of sites that
need never be listed because removal actions alleviated the hazard or
because private parties chose to clean up contamination using the liability
provisions of the statute to spread the costs to other responsible parties.
Other parties chose to settle with EPA and then pursue their own cleanup
without listing the site on the NPL. Some of these latter sites became
“Superfund Alternative” sites for which information is available, but for
many of these sites where EPA took no action or for sites that would not
otherwise qualify as NPL sites, information is limited.92
It is beyond the scope of this article to report on all the various areas of
program performance for the Superfund program, but it is important to give
the reader a sense of some of the key elements of the Superfund program
and how these accomplishments have changed over time.
B. NPL Listings/Deletions
One of the main indicators used to assess the scope and pace of the
Superfund program is the number of sites added to the NPL each year. It is
important to note that NPL listing is more of an art than a science, and the
EPA Administrator has tremendous discretion in deciding which, and how
many, sites should be listed. Typically, sites are brought to EPA’s attention
by state government representatives or by citizens living or working near a
site. While there is a formal system for evaluating whether a site is eligible
for the NPL, called the Hazard Ranking System,93 the final decision is the
EPA Administrator’s.94 The only legal requirement related to listing is that
a site may not be made a final NPL site (sites are first “proposed” to the
NPL, then most, but not all, sites are ultimately named as “final” NPL sites)
without the concurrence of the governor of the state where the site is
located.95 Thus, while the number of sites listed each year is telling in
terms of the agenda that the EPA is giving to itself, it does not, in fact,
92. Although a measure of private party interest in Superfund can be found by the number of
suits filed by private parties, the problem is how to distinguish contribution actions for costs incurred at
NPL sites from actions for recovery of costs for voluntary cleanup. See, e.g., GAO-09-656, supra note
20, at 39 tbl.8 (breaking out of the number of cases filed by private parties as compared to state and
federal government filings).
93. 40 C.F.R. § 300 app. A (2008).
94. For a description of the listing process, see Katherine N. Probst et al., Superfund’s Future:
What Will It Cost? A Report To Congress. KATHERINE N. PROBST ET AL., SUPERFUND’S FUTURE: WHAT
WILL IT COST? A REPORT TO CONGRESS 77–81 (2001).
95. Omnibus Consolidated Rescissions and Appropriations Act of 1996, Pub. L. No. 104-134,
110 Stat. 1321 (1996).
2009]
Superfund at 30
207
provide useful information to the public about the number of sites across
the country that need federal attention.
In the original statute, Congress required EPA to place 400 sites on the
NPL,96 which occurred in 1983. At the end of FY 2009, there were 1607
final and deleted NPL sites.97 NPL sites include not only sites where
private parties and state and local government organizations are liable, but
also sites owned or operated by the federal government, called “federal
facilities.” The overwhelming majority of the sites on the NPL are owned
and operated by private entities; however, eleven percent, or 173 of the
final and deleted NPL sites, are owned and operated by federal agencies.98
The cost of site studies and cleanups for federal facility sites are paid for
out of the budget of the federal agency responsible. The Department of
Energy has some of the most costly and complex NPL sites, and the
Department of Defense also has many contaminated sites on the NPL. As
shown in Figure 1 below, the majority of sites were listed before FY 1991.
Since that year, annual listings have numbered fewer than forty-five sites
per year. In the last eight years, twenty or fewer sites have been added each
year as final NPL sites.
96. Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA) of
1980, Pub. L. No. 96-510, § 105, 94 Stat. 2767.
97. U.S. Envtl. Prot. Agency, NPL Site Totals by Status and Milestone,
www.epa.gov/superfund/sites/query/queryhtm/npltotal.htm (last visited Oct. 27, 2009).
98. Some “federal facility” sites are among the most costly in the nation. However, we do not
address the many specific issues related to these sites in this article. See generally CONG. BUDGET
OFFICE, CLEANING UP THE DEPARTMENT OF ENERGY’S NUCLEAR WEAPONS COMPLEX (2004); U.S.
OFFICE OF TECH. ASSESSMENT, COMPLEX CLEANUP: THE ENVIRONMENTAL LEGACY OF CLEANING UP
NUCLEAR WEAPONS PRODUCTION (1991); U. S. CONG. BUDGET OFFICE, CBO PAPERS: CLEANING UP
DEFENSE INSTALLATIONS: ISSUES AND OPTIONS (1995).
208
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
Figure 1: Final NPL Listings, FY 1983–FY 200999
In the early years of the program, EPA developed the notion of
“deleting” a site from the NPL as the key measure of progress. According
to EPA policy, a site would be formally deleted from the NPL when EPA
has determined, in consultation with the state, that no further response
action is required to protect human health and the environment. Almost
thirty years later, only twenty-one percent of NPL sites have met that
goal.100 As of the end of FY 2009, 338 sites had been deleted, out of a total
of 1607 final NPL sites.101 Some of these sites were sites listed early in the
program and required little if any cleanup activity. Interestingly, as of the
end of FY 2007, according to GAO, eighty of the sites listed in 1983 were
still not construction complete.102
99. E-mail from Alan Youkeles, Associate Branch Chief, Budget, Planning and Evaluation
Branch, Office of Superfund Remediation & Technology Innovation, Office of Solid Waste and
Emergency Response, U.S. Environmental Protection Agency, to authors (Aug. 21, 2009, 10:45 a.m.
EST) (on file with authors); U.S. Envtl. Prot. Agency, Nat’l Priorities List: Number of NPL Site Actions
and Milestones by Fiscal Year, www.epa.gov/superfund/sites/query/queryhtm/nplfy.htm accessed (last
visited Dec. 16, 2009).
100. E-mail from Alan Youkeles, Associate Branch Chief, Budget, Planning and Evaluation
Branch, Office of Superfund Remediation & Technology Innovation, Office of Solid Waste and
Emergency Response, U.S. Environmental Protection Agency, to authors (Oct. 27, 2009, 1:25 p.m. EST)
(on file with authors).
101. Id.
102. GAO-09-656, supra note 20, at 68 fig.11.
2009]
Superfund at 30
209
C. Construction Complete
When it became clear to EPA that many sites would either never be
deleted from the NPL, or that attaining that status would take decades, EPA
developed a new measure for documenting progress: “construction
complete.” This new metric was established in 1990 and subsequently
clarified in 1993.103 A site is categorized as “construction complete” when
construction of the remedy is complete, that is, all the engineering work at
the site has been implemented.104 A site that is construction complete but
has not been deleted has not achieved the cleanup goals set out by EPA in
the selected remedy—in other words, more action at the site will be
required to achieve or maintain cleanup goals for the site.
Many still consider “construction complete” to be the best measure of
NPL site progress, as it indicates when all physical implementation of the
remedy is complete. As shown in Figure 2 below, the number of sites
categorized as construction complete hit a high of eighty-eight in FY 1992,
and then dropped to the sixties for the next few years until again reaching a
high of eighty-eight in FY 1997. In FY 2001, the number of sites that
attained this status dropped markedly, to forty-seven, and stayed in the
forties for the next few years. In FY 2009, the number of construction
complete sites reached an all time low—twenty sites—since the measure
was introduced.105
103. National Oil and Hazardous Substances Pollution Contingency Plan, 55 Fed. Reg. 8666,
8699 (Mar. 8, 1990) (to be codified at 40 C.F.R. pt. 300); Notification of Policy Change: Federal
Register Notice, 58 Fed. Reg. 12142, 12142 (Mar. 2, 1993) (to be codified at 40 C.F.R. pt. 300).
104. U.S.
Envtl.
Prot.
Agency,
Superfund:
Construction
Completions,
http://epa.gov/superfund/cleanup/ccl.htm (last visited June 3, 2009).
105. When EPA introduced the “construction complete” measure in the early 1990s, it then
“backdated” some sites that had reached that status before the measure was introduced. E-mail from
Alan Youkeles, Associate Branch Chief, Budget, Planning and Evaluation Branch, Office of Superfund
Remediation & Technology Innovation, Office of Solid Waste and Emergency Response, U.S.
Environmental Protection Agency, to authors (Aug. 20, 2009, 3:26 p.m. EST) (on file with authors).
210
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
Figure 2: Construction Complete Sites, FY 1983–FY 2009106
D. Government Performance Results Act Measures
As the pace of sites reaching “construction complete” has diminished in
recent years, EPA has again developed new performance measures to try to
provide a more attainable measure of progress for NPL cleanups. Most
NPL sites will require some kind of long-term monitoring and maintenance,
and many will never be deleted from the NPL. Thus, in the past decade,
increasing attention has been paid to the “post-cleanup” period, that is, what
is needed to ensure protection at a site after the remedy is fully
implemented. A critical element of post-cleanup activities is ensuring that
any institutional controls needed to prevent exposure are in fact being
implemented and enforced. Under the Government Performance Results
Act (GPRA),107 each federal agency is required to develop performance
106. E-mail from Alan Youkeles, Associate Branch Chief, Budget, Planning and Evaluation
Branch, Office of Superfund Remediation & Technology Innovation, Office of Solid Waste and
Emergency Response, U.S. Environmental Protection Agency, to authors (Aug. 21, 2009, 10:44 a.m.
EST) (on file with authors); U.S. Envtl. Prot. Agency, Nat’l Priorities List: NPL Site Total by Status and
Milestone, www.epa.gov/superfund/sites/query/queryhtm/npltotal.htm (last visited Oct. 27, 2009).
107. 5 U.S.C. § 306 (2006).
2009]
Superfund at 30
211
measures. The Superfund program established three new measures, in
addition to construction complete, for its GPRA goals.108 The EPA
definitions for each of these measures are:
Sitewide Ready for Anticipated Use: This measure tracks sites on
the NPL where: 1) construction of the remedy is completed, 2) all cleanup
goals have been achieved to reduce unacceptable risk that could affect
current and reasonably anticipated future land uses of the sites, and 3) all
institutional controls at the site have been implemented.
Human Exposure Under Control: This measure tracks sites on
the NPL where all identified unacceptable human exposures from site
contamination for current land and/or ground water use conditions have
been controlled.
Groundwater Migration Under Control: This measure tracks
sites on the NPL where either: 1) contamination is below protective, riskbased levels or, if not, 2) where the migration of contaminated ground water
is stabilized, there is no unacceptable discharge to surface water, and
monitoring will be conducted to confirm that affected ground water remains
in the original area of contamination.109
What all of these GPRA measures have in common is that they are
measures of interim progress: they relate to progress cleaning up a site that
is less than achieving the cleanup goals originally determined for the site.
Another element of these goals is that they relate directly to another
contentious cleanup issue—the question of whether cleanups should be
targeted to current or likely future land use or whether cleanups should be
aimed at enabling quite different possible future land uses. For example,
should an industrial site be cleaned up to a standard that would allow
industrial use or should it be cleaned up to a higher standard that would
allow residential use, with residents growing tomatoes in their yards and
children playing in the soil?
As of the end of FY 2008, EPA had “controlled all identified
unacceptable human exposure” at 1306 sites and “controlled the migration
of contaminated ground water through engineered remedies or natural
108. A fourth measure, “final assessment decision,” does not relate directly to cleanup progress,
and is therefore not included here.
109. U.S. Envtl. Prot. Agency, Accomplishments and Performance Measures,
www.epa.gov/superfund/accomplishments.htm (last visited Oct. 27, 2009).
212
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
processes” at 997 sites.110 Needless to say, the wording of these measures
raises many red flags. It is unclear what constitutes “unacceptable” human
exposure as well as how migration of contaminated groundwater can be
“controlled through natural processes.” Clearly, there is still an opportunity
for increased transparency in the measures EPA uses to document progress
for the Superfund program.
The third measure, sitewide ready for anticipated use, reflects the
agency’s increasing focus over the past decade on reuse of contaminated
properties. Importantly, a site cannot be considered ready for reuse unless
all institutional controls have been implemented. As of the end of FY 2008,
343 sites had met this measure.111
E. Removal Actions
The removal program is the understudy to the remedial program’s
starring role in Superfund. The removal program provides emergency aid
and initial cleanup at thousands of sites across the country. Removal
actions usually involve short-term actions to limit the immediate hazard at
sites.112 At some sites, however, such as EPA’s several-year cleanup of a
Libby, Montana asbestos mine and processing site for asbestos-containing
vermiculite, EPA implemented major cleanup actions under removal
authorities.113 Removal actions are tailored to the needs of the site and may
include a wide range of actions, such as providing alternative sources of
drinking water where water has been contaminated, removing barrels of
substances likely to explode at a site, or relocating families at imminent risk
from the contamination.114 The statute restricts funding for removal actions
funded by the government to $2,000,000 and twelve months, unless the
EPA has determined that continued action beyond these limits is necessary
to address an emergency, or other immediate risks to public health, welfare,
or the environment, or the action is consistent with the remedial action to be
taken.115
110. U.S. Envtl. Prot. Agency, Superfund National Accomplishments Summary Fiscal Year
2008, http://www.epa.gov/superfund/accomp/numbers08.htm (last visited Oct. 27, 2009).
111. Id.
112. 42 U.S.C. § 9604(a)(1) (2006).
113. United States v. W.R. Grace & Co., 429 F.3d 1224, 1227–28, 1237 (9th Cir. 2005)
(observing that removal actions are usually “time-sensitive” but then concluding that EPA’s several year
cleanup of a Libby, Montana asbestos mine and processing site for asbestos-containing vermiculite
qualified as a removal action).
114. § 9604(a)(1); 42 U.S.C. § 9601(23) (2006).
115. § 9604(c)(1)(A)–(C).
2009]
Superfund at 30
213
As of the end of FY 2007, EPA or private parties had started over
10,000 removal actions, including actions at federal facilities.116 EPA may
implement removal actions at any site in the country. CERCLA’s removal
program addresses hazardous pollution wherever and whenever it occurs,
without concern for whether the site is listed on the NPL.117 Seventy-two
percent of the removal actions through FY 2007 are at sites not listed on the
NPL.118 The majority of removal actions are paid for by EPA, not
responsible parties.119 This is almost certainly the result of the need to
move quickly to implement these types of activities.
II. PAYING FOR CLEANUP: TAXES AND LIABILITY
Congress provided two mechanisms for funding the cleanup of
contaminated sites: 1) a dedicated trust fund, and 2) a liability scheme that
encourages settlement with government authorities, compensates
government action, rewards voluntary cleanup, and, in some cases, forces
cleanup. The trust fund—officially titled the “Hazardous Substance
Response Trust Fund” and the source of the statute’s “Superfund”
moniker—provides resources for removal actions, site studies and cleanups,
and enforcement actions authorized under the liability scheme. Only sites
listed on the NPL may receive funding for remedial actions. Program
expenses and operating costs are also supported by the trust fund. By
giving EPA the resources to pay for cleanups directly, the government can
quickly clean up emergency spills and threatened releases, address orphan
or abandoned sites where no responsible parties can be found, and
accelerate site cleanup activities at those sites where there are responsible
parties but these parties fail to conduct sites studies or cleanups
themselves.120 Those activities paid for by the government are called “fundlead” activities.
116. GAO-09-656, supra note 20, at 52–53 & n.59.
117. § 9604(a)(4) (“[T]he President may respond to any release or threat of release if in the
President’s discretion, it constitutes a public health or environmental emergency and no other person
with the authority and capability to respond to the emergency will do so in a timely manner.”). This
authority extends even to actions otherwise excluded from federal response by § 9604(a)(3).
118. GAO-09-656, supra note 20, at 53.
119. See id. at 29 tbl.6, 52–53 (showing 1,838 removal actions taken by PRPs, a small
percentage of the over 10,000 removal actions taken over the life of the program).
120. 42 U.S.C. § 9611 (2006).
214
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
While the trust fund earned the statute its nickname, the liability
provisions earned the statute its reputation as a “polluter pays” statute.121
Superfund liability extends to owners and operators of a site at the time of
disposal,122 those who arrange for disposal (“generators”),123 and certain
transporters (those who selected the site for disposal).124 However,
Superfund liability, at least as originally enacted and continuing today in
some circumstances, extends beyond polluters to include those who benefit
from cleaned sites, such as current owners and operators,125 some of whom
may have had little to do with contaminating the site. Parties subject to
Superfund liability have come to be known as “potentially responsible
parties” (PRPs) rather than “polluters.” When PRPs agree through a
negotiated settlement to conduct work at the site rather than face a court
ruling under the statute’s liability provisions, or when EPA uses the
provisions provided to force parties to clean up, the effort is said to be
“PRP-lead.”
The remediation of a site may at various times be fund-lead and then
become PRP-lead, or vice-versa, as agreements are reached or orders are
issued. Typically, the majority of actions in the early stage of the cleanup
process—site studies and remedial designs—are paid for by EPA, with the
majority of the latter (and more expensive) actions—the actual remedies—
paid for by PRPs, as shown in Table 1 below.
121. OFFICE OF SITE REMEDIATION ENFORCEMENT, U.S. ENVTL. PROT. AGENCY., REVITALIZING
CONTAMINATED SITES: ADDRESSING LIABILITY CONCERNS: THE REVITALIZATION HANDBOOK 6 (2008).
122. 42 U.S.C. § 9607(a)(2) (2006).
123. § 9607(a)(3).
124. § 9607(a)(4).
125. § 9607(a)(1).
2009]
Superfund at 30
Remedial
Investigation/Feasibility
Study
Period
(FY)
1980–1986
215
Remedial Design
Remedial Action
Fund-Lead
PRP-Lead
Fund-Lead
PRP-Lead
Fund-Lead
PRP-Lead
76%
24%
63%
37%
67%
33%
1987–1990
52%
48%
49%
51%
54%
46%
1991–1999
54%
46%
28%
72%
27%
73%
2000–2008
66%
34%
51%
49%
36%
64%
Table 1: Comparison of Leads for Remedial Pipeline Actions by Time
Period (Percentage)126
Both the trust fund and the liability provisions have in some ways been
wildly successful—the Superfund taxes raised much-needed revenues for a
new federal program and the liability provisions have meant that the
majority of cleanups under the law have been conducted by companies
connected to contaminated sites. EPA is able to persuade parties to
undertake work at sites directly, and to reimburse EPA for work they have
already done.127 Both aspects of the legislation have, however, been quite
controversial. Many consider the liability provisions “draconian,”128 and
the taxes unfair.129 Others feel that both provisions enhance the preventive
aspects of the statute by providing an economic incentive to clean up and a
deterrent to making more waste or using more hazardous substances in
production.130 We explore the evolution of these two important funding
mechanisms below, as well as the accomplishments of each tool for funding
cleanups over time.
126. E-mail from Alan Youkeles, Associate Branch Chief, Budget, Planning and Evaluation
Branch, Office of Superfund Remediation & Technology Innovation, Office of Solid Waste and
Emergency Response, U.S. Environmental Protection Agency, to authors (Sept. 8, 2009, 5:15 p.m. EST)
(on file with authors); KATHERINE N. PROBST ET AL., supra note 94, at 43 tbl.3-2.
127. See GAO-09-656, supra note 20, at 26 fig.4 (showing that only ten percent of EPA
enforcement actions do not result in either reimbursement of EPA costs, site work, or both).
128. 141 Cong. Rec. E31 (1995) (statement of Hon. Fred Upton).
129. BNA Environmental Compliance Bulletin, 2010 Budget Plan Assumes Superfund Tax Will
Be Reinstated, http://subscript.bna.com/pic2/ecln.nsf/id/BNAP-7SAK5P?OpenDocument (last visited
Nov. 1, 2009).
130. GAO-09-656, supra note 20, at 10 n.7.
216
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
A. Superfund Trust Fund
For the first five years of the program, appropriations for the Superfund
program totaled $1.6 billion.131 The majority of these funds came from
excise taxes on petroleum and chemical feedstocks, with the remaining
funding from general revenues. The balance in the trust fund was
replenished over the years from penalties and cost recovery actions, as well
as from interest on the fund balance,132 as shown in Figure 3 with data from
a recent GAO report below.
Figure 3: Major Sources of Revenue for the Superfund Trust Fund,
FY 1981–FY 2007133
At the time of CERCLA’s enactment, public attention was on the trust
fund and the prospect of federal monies to speed cleanup.134 The focus on
131. Kate Probst, Reinstating the Superfund Taxes: Good or Bad Policy?, RESOURCES FOR THE
FUTURE, Aug. 24, 2009, http://www.rff.org/Publications/WPC/Pages/Reinstating-the-SuperfundTaxes.aspx.
132. U.S. GOV’T ACCOUNTABILITY OFFICE, SUPERFUND: FUNDING AND REPORTED COSTS OF
ENFORCEMENT AND ADMINISTRATION ACTIVITIES 7 fig.2 (2008) [HEREINAFTER U.S. GOV’T
ACCOUNTABILITY OFFICE, SUPERFUND: FUNDING].
133. U.S. GOV’T ACCOUNTABILITY OFFICE, SUPERFUND: FUNDING AND REPORTED COSTS OF
ENFORCEMENT AND ADMINISTRATIVE ACTIVITIES, GAO-08-841R (2008), (data obtained from the
“accessible text” version available at http://www.gao.gov/htext/d08841r.html).
2009]
Superfund at 30
217
the use of trust fund resources continued in the form of a scandal that
engulfed the program almost as soon as it was established. EPA
Administrator Anne Gorsuch Burford and her Assistant Administrator for
Solid Waste and Emergency Response, Rita M. Lavelle, became lightning
rods for accusations about favoritism, political corruption, and
mismanagement of Superfund resources.135 Questions arose about Lavelle’s
truthfulness in responding to congressional inquiries about her use of
Superfund monies, her alleged harassment of critics within the agency,136
her willingness to leak government settlement strategies to private
parties,137 and her “sweetheart deals”138 with private parties that fully
released them from significant portions of their cleanup obligations.139
Testimony before Congress revealed that Lavelle had issued orders for
staff “to speed up a survey of 15,000 hazardous waste sites around the
country ‘to verify her view that most of them were not a problem.’”140 The
scandal was serious enough to warrant Lavelle’s dismissal,141 indictment in
1983, and later imprisonment. For her part, Lavelle argued that she was
following orders.142 Indeed, Administrator Burford (then Gorsuch) had
doubted the need for the trust fund much beyond the statute’s then five-year
timeframe.143 Administrator Burford’s departure followed soon after
134. See Jonathan Harsch, Waukegan Harbor PCBs: Job For “Superfund,” CHRISTIAN SCI.
MONITOR, Dec. 3, 1980, http://www.csmonitor.com/1980/1203/120354.html (stating that, if passed by
Congress, the Superfund bill would provide 1.6 billion dollars to the federal fund for cleanup).
135. Philip Shabecoff, Forecast for E.P.A. Was Stormy from the Start, N.Y. TIMES, Feb. 20,
1983, § 4, at 2.
136. The Superfund Turned Upside Down, N.Y. TIMES, Dec. 28, 1982, at A22.
137. Shabecoff, Forecast, supra note 135; E.P.A. Administrator Said to Ask Inquiry on Miss
Lavelle’s Actions, N.Y. TIMES, Feb. 21, 1983, at A1, available at 1983 WLNR 484436 (stating that
Lavelle had met, without informing the government, with officials of two companies with which the
government was negotiating a hazardous waste cleanup in order to inform them that the E.P.A. would
consider a new settlement offer).
138. Shabecoff, Forecast, supra note 135.
139. See N.R. Kleinfield, $7 Million Settlement for Cleaning Up Hazardous Waste Dump Draws
Fire, N.Y. TIMES, Mar. 3, 1983, at B12, available at 1983 WLNR 463387 (referencing political
manipulation and favoritism in the cleanup of a contaminated site).
140. Philip Shabecoff, Toxic Cleanup Delay Laid to 2 Ex-Aides, N.Y. TIMES, Apr. 9, 1983, § 1,
at 7, available at 1983 WLNR 477278.
141. Howard Kurtz, Negotiation Approach Was Dictated By Burford, Lavelle Tells House Unit,
WASH. POST, Feb. 25, 1983, at A2.
142. Id.
143. Philip Shabecoff, 418 Toxic Dumps Listed in Cleanup, N.Y. TIMES, Dec. 21, 1982, at A1
(“Mrs. Gorsuch said in response to a question that she did not believe the fund would have to be
reauthorized after 1985 . . . .”).
218
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
Lavelle’s. One headline called the mess a “Made-for-Washington Epic
Starring EPA and Congress.”144
Under Administrator Burford’s leadership, trust fund dollars
accumulated because cleanup and enforcement suits were slow to start.145
During confirmation hearings for Burford’s successor in May 1983, EPA
Administrator-designee William Ruckelshaus committed to properly
funding the program and re-orienting Superfund to clean up first and
negotiate second.146
When Congress reauthorized the program in 1986, it increased annual
appropriations to $1.6 billion, thus quintupling the size of the program.147
Congress made some minor changes to the existing taxes and added a third
tax to generate additional revenues for the trust fund, the corporate
environmental income tax. This new tax was a broad-based tax, based on
every corporation’s modified alternative minimum taxable (AMT) income.
Many different types of companies paid the AMT, including not only the
same chemical and petroleum companies subject to the excise taxes, but
also companies in all major industrial sectors, including mining, insurance,
metals, and transportation, to name just a few.
Authorization for the three Superfund taxes expired at the end of 1995.
While the Clinton Administration called for reinstatement of the Superfund
taxes in its annual budget requests to Congress in subsequent years, little
effort was actually made to try to persuade Congress to enact them. Once
President George H. W. Bush was elected, the President’s budget no longer
requested that the taxes be reinstated, although in some years, the
President’s budget did request a funding increase for cleanup activities.
Over the years, various Members of Congress have introduced legislation to
reinstate one or more of the Superfund taxes, never with any success.
Since the lapse of the tax, the balance on the Superfund Trust Fund
reached its all time high in 1997 but thereafter declined steadily, arriving at
a zero balance in 2004, as it did in 1986 prior to passage of the first round
of amendments to CERCLA. The balance has rebounded above zero since
then—due to revenues from cost recovery, penalties, and interest on the
fund balance—but the amount of money in the trust fund remains quite
144. Peter Grier, Made-for-Washington Epic Starring EPA and Congress, CHRISTIAN SCI.
MONITOR, Feb. 11, 1983, http://www.csmonitor.com/1983/0211/021180.html.
145. Cass Peterson, Ruckelshaus Promises to Speed Dump Cleanup, WASH. POST, May 6, 1983,
at A3.
146. Id.
147. Probst, Reinstating the Superfund Taxes, supra note 131.
2009]
Superfund at 30
219
low.148 General fund appropriations, settlements, and judgments from the
cost recovery provisions of the statute slow the drain on the fund as is
demonstrated by the 1997 high after expiration of taxing authorities. In
fact, annual appropriations, which reached a high of $1.6 billion in
FY 1991, decreased to approximately $1.5 billion in FY 1994, where they
stayed through FY 1999.149 In FY 2000, however, annual appropriations
began a decline, and funding has leveled off at approximately $1.3 billion
annually, as shown in Figure 4 below.150
Figure 4: Superfund Appropriations, FY 1981–FY 2009151
As the balance in the trust fund decreased, so too did annual
appropriations from Congress. With the trust fund balance decreasing,
Congress increased funding from general revenues. According to a 2008
report from the Congressional Research Service, annual appropriations
from general revenues, which were $250 million a year from FY 1993
through FY 1998, were increased to $634 and $700 million in FY 2000–
148. U.S. GOV’T ACCOUNTABILITY OFFICE, SUPERFUND: FUNDING, supra note 132.
149. Critical Issues Facing the Superfund Program: S. Env’t & Pub. Works Comm. Subcomm.
On Superfund, 109th Cong. 4 (2006) (statement of Katherine N. Probst, Senior Fellow, Resources for
the Future).
150. Some of this decrease is due to funding for certain accounts, such as the Brownfields
program, now receiving appropriations separate from Superfund.
151. E-mail from Alan Youkeles, Associate Branch Chief, Budget, Planning and Evaluation
Branch, Office of Superfund Remediation & Technology Innovation, Office of Solid Waste and
Emergency Response, U.S. Environmental Protection Agency, to authors (Aug. 24, 2009, 2:00 p.m.
EST) (on file with authors).
220
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
FY 2002.152 This amount increased to $1.1 billion in FY 2004, and by
FY 2005, 100% of the annual Superfund appropriation came from general
revenues.153 According to data compiled by the U.S. GAO, for FY 1981–
FY 1995 (when the taxes expired) the dedicated Superfund taxes accounted
for the overwhelming majority (68%) of all trust fund revenues, as shown
in Table 2 below. During this period general revenues contributed 17% of
revenues, and the remainder of the funding came from interest on the trust
fund balance (9%), and fines, penalties, and cost recoveries (6%).154 For
FY 1996–FY 2007, however, with the taxing authority expired, dedicated
Superfund taxes accounted for only 7% of all trust fund revenues, a
dramatic decrease.155 During this period, general revenues accounted for
the majority (59%) of funds.156 Again, the remaining funds came from
interest on the trust fund balance (16%), and fines, penalties, and cost
recoveries (18%).157
Fiscal Year
General
Revenues
1981–1995
17%
Dedicated
Superfund
Taxes
68%
9%
6%
1996–2007
59%
7%
16%
18%
Total
33%
45%
12%
11%
Interest on
Trust Fund
Fines, Penalties,
and Recoveries
Table 2: Major Sources of Revenue for the Superfund Trust Fund,
FY 1981–FY 2007 (Percentage)158
At the same time, by the late 1990s, it was clear that EPA was
experiencing a shortfall in funds needed for cleanup. Lacking faith in
EPA’s estimates of its own funding needs, Congress asked Resources for the
Future (RFF) to estimate the future costs of the program to the EPA from
FY 2000–FY 2009.159 RFF identified a shortfall in the funds needed to
clean up sites on the National Priorities List. In contrast to “business as
usual” annual appropriations of approximately $1.3 billion each year, RFF
estimated a “base case” funding need of $15.1 billion over the ten years
152. CONG. RESEARCH SERV. REPORT FOR CONG., SUPERFUND TAXES OR GENERAL REVENUES:
FUTURE FUNDING ISSUES FOR THE SUPERFUND PROGRAM 7 (2008).
153. Id.
154. U.S. GOV’T ACCOUNTABILITY OFFICE, SUPERFUND: FUNDING, supra note 132, at 8 tbl.1.
155. Id.
156. Id. at 3.
157. Id. at 3, 8 tbl.1.
158. Id.
159. CONG. RESEARCH SERV. REPORT FOR CONG., supra note 152.
2009]
Superfund at 30
221
from FY 2000 through FY 2009—amounting to a shortfall of just over $2
billion in nominal dollars for the decade.160 With appropriations of around
$1.3 billion for each year since then, subsequent reports by the
Congressional Research Service, EPA’s Office of the Inspector General, and
others have documented funding shortfalls at specific sites in the years
since then, with remedial activities ready to go but put on hold for lack of
funding.161 The fact that the Superfund program was able to quickly
identify projects for $600 million in cleanup funding received under the
American Recovery and Reinvestment Act of 2009 attests to this funding
shortfall.162
B. Liability as a Funding Mechanism
1. Background
The legacy of the first few years of scandal and bureaucratic infighting
tainted the reputation of both the cleanup aspects of the program and its
enforcement and liability components.163 The mantra for the Superfund
program under Administrator William Ruckelshaus’s second term was
“cleanup first,” seeking to show a new, more aggressive Superfund
program.164 When this approach was criticized by many because the
government—not responsible parties—was paying for most of the cleanups,
Ruckelshaus’s successor, EPA Administrator William Reilly, revised the
Agency’s approach and created a new motto, “enforcement first.”165 The
Enforcement First Initiative used CERCLA’s liability provisions166 to
leverage EPA’s fund-lead cleanup efforts. EPA has filed numerous
administrative—and through the Department of Justice (DOJ) also
160. PROBST ET AL., supra note 94, at xxii-xxiii. The authors also estimate a “low” and “high”
funding scenario as well.
161. See generally CONG. RESEARCH SERV. REPORT FOR CONG., supra note 152; Letter from
Nikki L. Tinsley, EPA Inspector General, to Congressman John D. Dingell, Comm. on Energy &
Commerce (July 24, 2002) (detailing funding shortfalls and limitations), available at
http://www.epa.gov/oig/reports/2002/boxer.pdf.
162. American Recovery and Reinvestment Act of 2009, Pub. L. No. 111-5, 123 Stat. 115, 168;
see U.S. Envtl. Prot. Agency, Superfund Sites Receiving American Recovery and Reinvestment Act
Funding, http://www.epa.gov/superfund/eparecovery/sites.html (last visited Oct. 27, 2009) (listing sites
by region).
163. See supra notes 134–44 and accompanying text (discussing scandalous beginnings).
164. Peterson, supra note 145.
165. WILLIAM K. REILLY, ADM’R, U.S. ENVTL. PROT. AGENCY, A MANAGEMENT REVIEW OF
THE SUPERFUND PROGRAM 2–13 (1989).
166. 42 U.S.C. § 9607(a)(4)(A) (2006).
222
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
judicial—claims to recover its Superfund costs.167 Parties also entered
agreements to conduct site studies and cleanup as a consequence of these
actions.168
Congress supported the “enforcement first” initiative when, in 1986, it
pressed for even faster cleanup by barring private party challenges to
response actions until after the cleanup is complete.169 Congress had not
forgotten, however, that private party agreements to do work and to pay
limited sums to clean up—“sweetheart” deals, as they were called170—were
part of the scandal that unseated the first EPA Administrator to implement
Superfund, Administrator Anne Gorsuch Burford.171 Congress sought to
alleviate concerns about such deals by giving EPA clear guidance on their
settlement authorities, including specific direction about clauses that may
be included with any settlement agreements and specific procedures for
settlement.172 Final approval of settlements in excess of $500,000 requires
approval from the Attorney General173 and lodging of the consent decree in
federal district court.174 Congress’s support for settlement went so far as to
expressly encourage settlement “whenever practicable and in the public’s
interest.”175
After the 1986 amendments to CERCLA, EPA also sought agreements
with parties for them to take over cleanup at sites, i.e., to transfer the site
from a fund- to a PRP-lead cleanup. If such agreements were not
forthcoming, EPA threatened to force parties to clean up and in fact resorted
167. GAO-09-656, supra note 20 (listing between FY 1979 and FY 2007, 1,695 cost recovery
related enforcement actions, of which ninety-seven percent were “consensual,” i.e., resolved through
administrative and judicial settlements, and 1,005 combined cost recovery and PRP work agreements of
which ninety-eight percent were listed as consensual)
168. Id. (listing 1,473 site-related work enforcement actions of which fifty-four percent were
consensual).
169. Superfund Amendments and Reauthorization Act (SARA) of 1986, Pub. L. No. 99-499,
100 Stat. 1613 (codified at 42 U.S.C. § 9613(h) (2006)).
170. Shabecoff, Forecast, supra note 135.
171. Grier, supra note 144.
172. See 42 U.S.C. § 9622(b) (authorizing “mixed funding,” i.e., fund payment for PRP work
under certain circumstances); see also § 9622(f) (authorizing “covenants not to sue”); § 9622(h)(4)
(authorizing contribution protection); § 9622(a) (granting authority to enter into settlement agreements);
§ 9622(c) (providing for the scope of the agreement’s effect on agency settlement authority); § 9622(d)
(detailing the use of consent decrees and the role of DOJ and the federal courts in their review and
approval); § 9622(e) (authorizing the President to facilitate settlement agreements); § 9622(i) (requiring
a public comment period for cost recovery and de minimis settlements); § 9622(j) (encouraging
participation of natural resources trustees in settlement negotiations and their approval of settlement
provisions related to natural resources damages).
173. § 9622(g)(4), (h).
174. 28 C.F.R. § 50.7 (2008).
175. § 9622(a).
2009]
Superfund at 30
223
to issuing section 106 orders mandating cleanup in some situations.176 The
enforcement first theme has continued throughout Superfund’s
maturation.177 Ironically, given the history of suspicion and scandal that
overshadowed the negotiated agreements with private parties during the
Burford-Lavelle years,178 the overwhelming majority of agreements
resulting from EPA’s enforcement efforts are voluntary. As a recent GAO
report documented, “[f]rom fiscal years 1979 through 2007, EPA completed
4,642 enforcement actions at NPL sites, of which 3,682, or 80%, were
consensual.179 Moreover, EPA resolved negotiations with responsible
parties through administrative—rather than judicial—actions more than
60% of the time.”180
Many of these consensual agreements are documented by “consent
decrees” which are “lodged” in federal court for final approval.181 Consent
decrees may document a settlement after years of litigation or they may
memorialize agreements reached without any litigation. In fact, many
Superfund consent decrees are negotiated before DOJ files an action in
federal court. Pre-negotiated consent decrees make up the majority of
CERCLA-related federal court cases filed by the United States. These
agreements have increased to seventy-seven percent of the cases related to
Superfund filed by the federal government in FY 2007 from fifty-one
percent of these types of federal government cases filed in FY 1994.182
Figure 5 below shows the extent to which pre-negotiated consent decrees
make up the docket of cases filed by the federal government and how this
differs from the types of claims filed by other parties.183 Cases with prenegotiated settlements are faster than other types of CERCLA cases, and,
when counted along with the administrative version of these agreements,
176. GAO-09-656, supra note 20, at 27–28 (“Only 16 of EPA’s 901 unilateral administrative
orders issued at NPL sites—less than 2 percent—included requirements related to recovering agency
costs.”).
177. See, e.g., Memorandum from John Peter Suarez, Assistant Adm’r, Office of Enforcement
and Compliance Assurance & Marianne Lamont Horinko, Assistant Adm’r, Office of Solid Waste and
Emergency Response, to Regional Administrators, Enforcement First for Remedial Action and
Superfund Sites (Sept. 20, 2002) (stating that the EPA has a longstanding “enforcement first” policy
because such policy promotes the “polluter pays” principle and helps conserve the resources of the
Hazardous Substance Trust Fund); OFFICE OF SOLID WASTE & EMERGENCY RESPONSE, U.S. ENVTL.
PROT. AGENCY, DIRECTIVE NO. 9208.2 (2006) (making it explicit that the “enforcement first” policy
“applies to any actions needed to ensure the implementation and effectiveness of institutional controls”).
178. Shabecoff, Forecast, supra note 135.
179. GAO-09-656, supra note 20, at 24.
180. Id.
181. 28 C.F.R. § 50.7 (2008); 42 U.S.C. § 9622(d), (g)(4), (h) (2006).
182. GAO-09-656, supra note 20, at 39.
183. Id. at 111 fig.18 (data found at http://www.gao.gov/htext/d09656.html).
224
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
“administrative orders on consent”184 reflect the pervasive nature of
negotiation as a means to funding and cleaning up Superfund sites.185
Figure 5: Number of CERCLA Cases, by Outcome and Type of Plaintiff,
FY 1994–FY 2007186
The discussion that follows describes the relevant liability provisions of
the statute in the context of how liability affects funding and cleanup,
including liability’s role in compensating EPA for past and future costs,
enlisting private parties to conduct government-supervised cleanup, and,
finally, shifting costs among private parties and rewarding private party
voluntary cleanup outside the purview of the NPL.
184. Id. at 22 tbl.3.
185. Id. at 39.
186. Id. at 111 fig.18 (data found at http://www.gao.gov/htext/d09656.html). Cases that include
more than one type of outcome are counted in more than one category. Therefore, this figure over
counts the total number of cases. Cases can result in no outcome in the GAO analysis for a number of
reasons: the parties may not have reached any final outcomes in a case that was still open as of the end
of FY 2007; the parties may have reached only outcomes not relating to liability, such as gaining access
to a site; or, a case may have ended by being consolidated with another case before any outcomes were
reached.
2009]
Superfund at 30
225
2. Compensating the Trust Fund for Past and Future Costs at NPL Sites
As described above, when EPA pays for costs for a site study or cleanup
(removal and/or remedial actions) at a site, the activity is said to be “fundlead.” There are, however, a number of ways that EPA may be reimbursed
for costs they incur at NPL sites. First, states must pay for or reimburse
EPA for ten percent of fund-lead costs incurred, and, second, when the state
or a political subdivision of a state has a significant role in the pollution of
the site, states must pay fifty percent or even more of the costs incurred by
the fund.187 States must also agree to provide all future operations and
maintenance of the removal action or remedy for a fund-lead cleanup.188
Understandably, states no longer viewed the Superfund Trust Fund as “free
money” once their obligations for federally funded cleanup became clear.
The result is greater state support for EPA’s transitioning from fund-lead to
PRP-lead cleanup where possible. Part of this transition may include
private party compensation of past costs or private party commitments to
pay future costs.
The cost recovery provisions of the statute authorize the EPA, State, or
a tribe to recover all cleanup costs that are not inconsistent with the
National Contingency Plan.189 The standard of liability under section 107 is
the same as that provided for under section 311 of the Clean Water Act.190
A variety of courts have held that standard to be retroactive,191 strict, and,
pursuant to common law traditions, joint and several where the harm is
incapable of divisibility.192 With the exception of certain defenses and
exemptions discussed earlier,193 EPA may recover these costs from those
parties that own or operate the site currently or those that owned or
operated the site at the time of disposal of the hazardous substances, as well
as from generators of hazardous substances (those who arranged for
disposal), and those who transport the substances to a site selected by the
transporter. EPA may use its authority anywhere, not just at sites listed on
187. 42 U.S.C. § 9604(c)(3)(C) (2006).
188. § 9604(c)(3)(A).
189. 42 U.S.C. §§ 9605, 9607(a)(4)(B) (2006) (referring to 40 C.F.R. §§ 300.1–300.3 (2008)).
190. Id. § 9601(32).
191. United States v. Olin Corp., 107 F.3d 1506, 1511–12 (11th Cir. 1997); United States v.
Mottolo, 695 F. Supp. 615, 622 (D.N.H. 1988).
192. Burlington N. & Santa Fe Ry. v. United States, 129 S. Ct. 1870, 1881 (2009); see, e.g.,
United States v. Chem-Dyne Corp., 572 F. Supp. 802, 805 (S.D. Ohio 1983) (citing 5 U.S.C.C.A.N.
6119, 6119–6120 (1980)).
193. See supra notes 30–38 and accompanying text (discussing changes in CERCLA through
amendments, judicial decisions, and administrative reform).
226
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
the NPL, where hazardous substances release or threaten release into the
environment and for which the agency had incurred appropriate costs
toward cleanup.194
While these liability provisions might suggest that EPA should recover
almost 100% of their past costs, PRP reimbursement of EPA’s costs for NPL
sites has not attained that goal. Although EPA data is incomplete and
significant cost recovery cases are still pending, EPA provided GAO with
estimates that it has recovered approximately thirty-six percent of its sitespecific costs, from government work such as removal actions, site studies,
and remedial actions, since Superfund’s inception.195 The total value of
EPA’s cost recovery efforts, however, has climbed over the years. From
1979 through March 31, 1992, PRPs reimbursed EPA for $415 million of
Superfund appropriations spent on government-funded cleanup-related
activities at NPL sites.196 Now, however, total compensation for EPA’s costs
is counted in the billions of dollars. By the end of FY 2007, EPA had
recovered past costs of $5.1 billion and has received future costs of $2.2
billion to total $7.3 billion in cost recovery since the inception of the
program.197 From FY 1999–FY 2007, both past and future cost recovery at
NPL sites returned $302 million annually, ranging from a high of $568.5
million in FY 1999 to a low of $161.3 million in FY 2006.198 During this
same period, EPA’s total enforcement expenditures have averaged about
$199.2 million each year.199
The frequency of agreements to reimburse EPA for work done or yet to
be done at NPL sites has also increased. EPA obtained some form of
monetary compensation for activities at 1141 out of 1160 NPL sites for
which EPA provided enforcement data.200 One reason for the large number
of compensation agreements is administrative reform,201 and later statutory
reform, that resulted in EPA authority to offer “cash out” settlements for de
194. 42 U.S.C. § 9607(a) (2006).
195. GAO-09-656, supra note 20, at 30.
196. U.S. GEN. ACCOUNTING OFFICE, SUPERFUND: CURRENT PROGRESS AND ISSUES NEEDING
FURTHER ATTENTION, STATEMENT OF PETER F. GUERRERO BEFORE THE SUBCOMMITTEE ON OVERSIGHT,
COMMITTEE ON WAYS AND MEANS, HOUSE OF REPRESENTATIVES 4 (June 11, 1992). This number does
not include the costs EPA incurred in trying to recover these costs.
197. U.S. GOV’T ACCOUNTABILITY OFFICE, SUPERFUND: FUNDING, supra note 132, at 15 tbl.2.
198. Id. at 16.
199. Id.
200. Id. at 17. EPA took enforcement actions at 243 NPL sites for which compensation was not
received, but other results were achieved such as site access or other action to allow EPA or private party
cleanup to continue. Id.
201. GAO-09-656, supra note 20, at 45.
2009]
Superfund at 30
227
minimis parties.202 Through FY 2007, at least 438 of 1695 cost recovery
agreements were with these de minimis parties, and all of these de minimis
agreements were voluntary.203 Of the remaining 1257 agreements, only
forty-eight were the result of parties involuntarily reimbursing EPA’s
costs.204 GAO attributes EPA’s success in reaching cooperative settlements
in part to the strength of EPA’s case against a party if they had chosen not to
settle, and greater EPA willingness to cover a portion of the shares
attributable to bankrupt parties (“orphan shares”) and enter into ability-topay settlements and other administrative and statutory reforms.205
3. Liability Allows EPA to Obtain Work Agreements from PRPs
Settlements where PRPs voluntarily agree to undertake sites studies and
cleanups are significant sources of indirect income for the program.206
Work done by the PRPs is work that is not paid for by the fund and, thus,
preserves trust fund monies. Parties agree to undertake work at sites for
many of the same reasons they agree to compensate EPA for its prior costs,
with the additional reason that if a party implements site studies or remedies
themselves, it has greater control over both the site activities and costs.207
Private party response actions at NPL sites can take the form of removal
actions, site study, remedial action, and other site work.208
While the majority of EPA settlements with PRPs relate to payment of
past or future costs of response, the greater value is achieved from PRP
commitments to conduct work at NPL-listed sites.
Out of 4642
enforcement actions from FY 1979 through FY 2007, 799, or seventeen
percent, involved voluntary agreements to conduct response actions
(cleanup related work), as contrasted with 1647 voluntary agreements
(thirty-five percent) to compensate EPA for their costs.209 Twenty-one
percent, or 989 voluntary agreements, cover both site work and
202. Id. at 31 n.33; 42 U.S.C. § 9622(g) (2006).
203. GAO-09-656, supra note 20, at 27.
204. See id. at 27 tbl.5 (charting 48 nonconsensual cost recovery actions).
205. Id. at 45.
206. “Voluntarily” means that the parties did not contest the agreement, and perhaps even
sought the settlement with the EPA. The U.S. GAO characterizes these agreements as “consensual” as
opposed to those that parties contest, which GAO characterizes as “nonconsensual.” Id. at 22.
207. EPA acknowledges PRP influence on remedy selection indirectly by its reluctance to enter
into site studies with recalcitrant PRPs “because their work can influence the selection of an appropriate
remedial action.” Id. at 28.
208. See id. at 29 tbl.6 (providing this list of response actions in the context of EPA enforcement
actions at NPL sites organized by types of work sought).
209. Id. at 27 tbl.6.
228
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
compensation.210 Figure 6 below compares the number of EPA enforcement
actions that result in agreements to reimburse EPA for their response costs
(“cost recovery”), agreements to provide site-related work, and agreements
to provide both cost recovery and site work. It also demonstrates the
significant proportion of these that are consensual or voluntary in nature.
Figure 6: Outcome of EPA Enforcement Actions at NPL Sites, FY 1979–
FY 2007211
4. Liability Allows EPA to Force Recalcitrant Parties to Conduct Response
Actions
Another element of the Enforcement First initiative is federally
compelled cleanup. Congress empowered the President (usually acting
through the EPA, which is the President’s designee) to seek injunctive relief
in federal court or with the administrative authority to force responsible
parties to clean up when contamination presents an imminent and
substantial endangerment to human health or the environment.212 This
provision is bolstered by federal authority to assess treble damages against
those who fail to comply with a “106 order,” or “Unilateral Administrative
Order” (“UAO”) as these orders are often called.213 This can amount to
210.
211.
212.
213.
Id.
Id. at 27 tbl.5 (analyzing GAO data).
42 U.S.C. § 9606(a) (2006).
Id. § 9607(c)(3).
2009]
Superfund at 30
229
millions of dollars, since many response actions may exceed twenty million
dollars with the addition of another sixty million dollars or more if treble
damages are assessed.
Between the strict, joint and several liability provided by the cost
recovery provisions granted to state, federal, and tribal governments, and
the federal authority to force cleanup with the threat of huge damage
awards, Congress provided EPA with tremendous leverage. EPA has issued
901 unilateral administrative orders, meaning that cleanup actions
conducted by unwilling parties214 made up nineteen percent of the
enforcement actions taken at NPL sites through FY 2007.215 Most of the
time the treble damages are not invoked since most parties will comply with
these orders. GAO research indicates, however, that compelled action may
not be as effective as one might think since parties are more likely to balk at
cleanup requirements under these conditions.216 Of course, this data may
also reflect a propensity among parties unwilling to settle voluntarily to
contest EPA orders related to Superfund.
Together both voluntary and federally-mandated private party sitespecific study and cleanup actions make up a significant portion of the work
to address contamination at NPL sites. Figure 7 below details the nature of
the work done as a result of voluntary and mandated private party cleanup
actions.
214. GAO-09-656, supra note 20, at 25 tbl.4. GAO characterizes these actions as
“nonconsensual” while recognizing, as this article does here, that some parties in fact prefer to receive
and comply with a unilateral administrative order or CERCLA section 106 order. Id. at 24.
215. Id. at 22 n.4.
216. See id. at 28 (“When parties deny access or information, it may be less likely that EPA can
resolve issues through consensual agreements.”).
230
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
Figure 7: Outcome of EPA Enforcement Actions at NPL Sites, FY 1979–
FY 2007, by Type of Site Work Sought217
5. Liability Provides Resources to the Superfund Program
Work done by private parties is work that EPA does not have to pay for
out of its own budget. In reality, all private party work commitments
constitute “additional” funding for the Superfund program, above and
beyond annual appropriations from Congress, and are a significant source
of cleanup funding. The U.S. GAO estimates the total value of PRP work
from FY 1979–FY 2007 at $22.5 billion, seventy-five percent of the value
of all EPA’s enforcement actions.218
C. Spreading Costs Among Private Parties
When private parties reimburse EPA’s costs or agree to remediate a
NPL site, one of the factors they consider before entering into such an
agreement is whether or not they will be able to spread some of their costs
to other PRPs.219 When parties reimburse EPA for its costs, they may file
217. Id. at 29 tbl.6 (analyzing GAO data).
218. Id. at 30.
219. Id. at 33.
2009]
Superfund at 30
231
claims for contribution.220 Contribution claims are common law or
statutorily provided claims to reallocate costs among parties that have paid
more than their fair share as a consequence of joint and several liability.221
When parties conduct work at the site, they may be able to also file claims
for cost recovery from other liable parties.222 Cost recovery claims may
proceed without a prior finding of joint liability.223 But the distinction
between cost recovery and contribution was not always so clear.
Initially, the statute did not include an explicit right to contribution. It
did, however, provide language in section 107(a)(4)(B) that some courts
studied for authority to resolve claims for contribution224 and that other
courts considered as providing an independent cause of action for cost
recovery.225 SARA resolved the contribution question when Congress
added provisions in section 113 to explicitly authorize contribution claims
between potentially liable parties.226
Over the years, the cost recovery and contribution provisions of the
statute have been invoked to compensate parties engaged in cleanup-related
activities at sites not listed on the NPL and perhaps not even subject to state
or federal oversight, which we characterize as “voluntary cleanup.”227
While the line between the cost recovery provisions, authorized in section
220. 42 U.S.C. § 9613(f)(1), (f)(3)(b).
221. United States v. Atl. Research Corp., 551 U.S. 128, 138–39 (2007).
222. 42 U.S.C. § 9607(a)(4)(B) (2006); Atl. Research Corp., 551 U.S. at 138; United States v.
Kramer, Civil Action Nos. 89-4340 (JBS), 89-4380 (JBS), 2009 WL 2339341, at *8 (D.N.J. July 27,
2009) (granting a settling party’s motion to file a claim under 42 U.S.C. § 9607(a)(4)(B) against a nonsettling party).
223. Atl. Research Corp., 551 U.S. at 138 n.5.
224. See United States v. New Castle County, 642 F. Supp. 1258, 1262 (D. Del. 1986) (noting
uncertainty about whether CERCLA originally provided contribution rights and finding a right to
contribution under federal common law); Colorado v. Asarco, Inc., 608 F. Supp. 1484, 1486–93 (D.
Colo. 1985) (noting federal common law provides a right of contribution); Wehner v. Syntex
Agribusiness, Inc., 616 F. Supp. 27, 31 (E.D. Mo. 1985) (finding that a right of contribution implied in
section 107(e)(2)); United States v. Westinghouse Elec. Corp., No. IP 83-9-C, 1983 WL 160587, at *3
(S.D. Ind. June 29, 1983) (finding no right of contribution).
225. See City of New York v. Exxon Corp., 633 F. Supp. 609, 615–18 (S.D. N.Y. 1986) (holding
that the City does not need prior approval from federal or state authorities before bringing a cause of
action under section 107(a)(4)(B); Wickland Oil Terminals v. Asarco, Inc., 792 F.2d 887, 890–92 (9th
Cir. 1986) (holding section 107(a)(2)(B) to create a cause of action to recover private response costs
regardless of whether the government had instituted a cleanup action); Walls v. Waste Res. Corp., 761
F.2d 311, 317–18 (6th Cir. 1985) (same); City of Phila. v. Stepan Chem. Co., 544 F. Supp. 1135, 1140–
43 (E.D. Pa. 1982) (same).
226. Superfund Amendments and Reauthorization Act (SARA) of 1986, Pub. L. No. 99-499,
100 Stat. 1613 (codified at 42 U.S.C. § 9613(f)(1), (f)(3)(b) (2006)).
227. We characterized these non-NPL cleanups as “voluntary” although we recognize that
parties may be conducting these removal actions or other remedial work at the behest of state
Brownfields programs or to limit future liability.
232
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
107, and the contribution provisions, authorized in section 113, was not
clearly delineated for many years thereafter, many jurisdictions ultimately
decided that claims between private parties for reimbursement of voluntary
cleanup actions were claims for contribution—not claims for cost
recovery—and only allowed parties to pursue their claims under section
113.228
And pursue recovery under section 113 they did. Private party claims
make up the majority of CERCLA cases filed between FY 1994 and
FY 2007.229 Figure 8 below shows the percentage of CERCLA cases filed
by type of plaintiff for FY 1994–FY 2007.
Figure 8: Percentage of CERCLA Cases Filed by Type of Plaintiff,
FY 1994–FY 2007230
228. See Atl. Research Corp. v. United States, 459 F.3d 827, 832–33 (8th Cir. 2006) (“In the preAviall analysis, § 113 was presumed to be available to all liable parties, including those which had not
faced a CERCLA action.”). See Akzo Coatings, 30 F.3d at 763 n.4 (liable party’s § 113 claim for costs
voluntarily incurred held barred by settlement); Pinal Creek, 118 F.3d at 1306 (liable party’s claim for
costs voluntarily incurred governed by both §§ 107 and 113). Accordingly, most courts concluded liable
parties could not use § 107. See Pnuemo Abex, 142 F.3d at 776 . . . .”).
229. GAO-09-656, supra note 20, at 39 tbl.8.
230. Id. Because of the blurred line between contribution claims for reimbursement of costs
paid and voluntary cleanup costs, it is difficult to distinguish contribution claims from voluntary cleanup
claims for the purposes of this chart.
2009]
Superfund at 30
233
This parade of claims and suits to reallocate private party costs paid to
EPA or to share costs from a voluntary cleanup has slowed in recent years.
GAO estimates that private parties filed 142 cases in FY 1994 as contrasted
with forty-four cases filed in FY 2007, a decline of sixty-nine percent.231
Claims filed against private parties by state and federal governments have
remained relatively stable over those same years.232
The decline in private party suits may be an anomaly due to an
unsettling decision that moved through the ranks of the court system to the
U.S. Supreme Court between FY 2000 and the beginning of FY 2004. On
December 13, 2004, the U.S. Supreme Court ruled in Cooper Industries v.
Aviall Services Inc.233 that contribution claims brought under section
113(f)(1) may only proceed during or following a civil action under section
106 or section 107(a).234 The decision was contrary to the precedent in
nearly every circuit to consider the issue, most of which had directed
private parties, even those who had incurred costs voluntarily, to recover
their costs only under section 113(f)(1)’s contribution provisions rather than
under section 107(a)’s cost recovery provisions.235 Pending contribution
231. Id. at 37.
232. Id.
233. Cooper Indus. Inc. v. Aviall Servs. Inc., 543 U.S. 157 (2004).
234. Id. at 166.
235. E.g., Dico, Inc. v. Amoco Oil Co., 340 F.3d 525, 527 (8th Cir. 2003) (dismissing the
plaintiff’s direct cost recovery action); Bedford Affiliates v. Sills, 156 F.3d 416, 423–24 (2d Cir. 1998)
(limiting recovery by PRPs to section 113(f) contribution claims); Centerior Serv. Co. v. Acme Scrap
Iron & Metal Corp., 153 F.3d 344, 350 (6th Cir. 1998) (“Claims by PRPs, however, seeking costs from
other PRPs are necessarily actions for contribution, and are therefore governed by the mechanisms set
forth in section 113(f).”); Pneumo Abex Corp. v. High Point Thomasville & Denton R.R. Co., 142 F.3d
769, 776 (4th Cir. 1998) (“As the case before the court involves entirely potentially responsible parties,
such parties must seek contribution under [section 113].”); Pinal Creek Group v. Newmont Mining
Corp., 118 F.3d 1298, 1301 (9th Cir. 1997) ( “Because all PRPs are liable under the statute, a claim by
one PRP against another PRP necessarily is for contribution.”); New Castle County v. Halliburton NUS
Corp., 111 F.3d 1116, 1122 (3d Cir. 1997) (“Thus, section 113 does not in itself create any new
liabilities; rather, it confirms the right of a [PRP] under section 107 to obtain contribution from other
[PRPs].”); Redwing Carriers, Inc. v. Saraland Apartments, 94 F.3d 1489, 1496 & n.7 (11th Cir. 1996)
(distinguishing between section 113(f) contribution claims and section 107(a) cost recovery actions);
United States v. Colo. & E. R.R. Co., 50 F.3d 1530, 1539 (10th Cir. 1995) (Holding that “claims
between PRPs to apportion costs between themselves are contribution claims pursuant to § 113
regardless of how they are pled . . . .”); United Techs. Corp. v. Browning-Ferris Indus., 33 F.3d 96, 103
(1st Cir. 1994) (prohibiting recovery under section 107 but allowing recovery by and between PRPs for
contribution under section 113.). Cf. Akzo Coatings, Inc. v. Aigner Corp., 30 F.3d 761, 764 (7th Cir.
1994) (characterizing a section 113(f) claim as a “quintessential claim for contribution,” but leaving
open the possibility of section 107 recovery under certain circumstances).
234
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
claims under section 113(f)(1) brought without the necessary concurrent or
prerequisite civil action were dismissed.236
Justice Thomas’s Aviall decision, though clear and direct in its
reasoning, confounded private parties seeking recovery of costs for various
forms of voluntary—or at least not involuntary—cleanup, such as those
involved in brownfields redevelopment, transfer and sale of contaminated
properties, and state directed response actions.237 Aviall provided one more
excuse for the slow cleanup of facilities contaminated at least in part, but
frequently most significantly, by federal government actions because the
U.S. government controlled who could file contribution actions by its
choice of parties to enforce against in the first instance.238 By not filing a
section 106 or section 107 action, it seemed that no private party could
initiate a recovery action against the U.S. government.
Atlantic Research Corp. faced this dilemma when it sought recovery
from the United States under theories of cost recovery and implicit
contribution for costs it had paid in cleaning up contamination, for which
the federal government would have been liable prior to the Aviall decision,
but for which the United States had not filed a civil action for cost recovery
or forced cleanup. When the issue reached the U.S. Supreme Court three
years after the Aviall decision, Justice Thomas’s unanimous decision
resolved the problem the Court had created and developed clearer
distinctions between private party claims for contribution and cost recovery.
In United States v. Atlantic Research Corp., the Court restored the ability of
236. See, e.g., Major v. Astrazeneca, Inc., Nos. 5:01-CV-618, 5:00-CV-1736, 2006 U.S. Dist.
LEXIS 65225 (D.N.Y. Sept. 13, 2006) (“[A] party may only seek contribution pursuant to § 113(f)
during or following a civil action brought pursuant to [section 106] or [section 107(a)].”); Kaladish v.
Uniroyal Holding, Inc., No. 3:00CV854(CFD), 2005 U.S. Dist. LEXIS 17272 (D. Conn. Aug. 9, 2005)
(prohibiting a section 113(f) contribution claim unless that party has been sued under section 106 or
section 107); CadleRock Props. Joint Venture, L.P. v. Schilberg, No. 3:01CV896(MRK), 2005 U.S. Dist.
LEXIS 14701 (D. Conn. July 18, 2005) (dismissing contribution action for lack of an underlying federal
legal claim); Champion Labs., Inc. v. Metex Corp., Civ. No. 02-5284(WHW), 2005 U.S. Dist. LEXIS
37068 (D.N.J. July 8, 2005) (allowing a party to seek a section 113(f) contribution claim during or after
a civil action under section 106 or section 107).
237. See Richard O. Faulk & Cynthia J. Bishop, There and Back Again: The Progression and
Regression of Contribution Actions Under CERCLA, 18 TUL. ENVTL. L.J. 323, 337 (2005) (concluding
that following the Aviall decision there is confusion as to whether responsible parties have a recovery
claim under CERCLA if they clean up the contaminated site); Karl S. Bourdeau & W. Parker Moore,
Options for Potentially Responsible Parties in the Wake of the Aviall Decision, 38 ENVTL. REP. 440, 461
(2007) (recognizing that following the Aviall decision, both the Second and Eighth Circuit Courts have
limited private party recovery claims in certain circumstances).
238. See City of Moses Lake v. United States, 472 F. Supp. 2d 1220, 1224 n.4 (E.D. Wash.
2007) (leaving private PRPs without any recourse against potentially liable agencies unless sued by
EPA).
2009]
Superfund at 30
235
PRPs to pursue cost recovery actions under section 107 of CERCLA.239
PRPs once again were able to recover their costs of cleanup under theories
of contribution or cost recovery, or perhaps even both, depending on the
procedural circumstances surrounding the cleanup.240 Nonetheless, Aviall’s
dark cloud of confusion hung over private parties from FY 2004–FY 2007.
Private parties were likely hesitant to file either cases for contribution or
cost recovery for many of the later years that the GAO studied.241
1. Superfund Liability: Powerful and Complex
In addition to serving as a tool to generate revenue for cleanup, there
are many other aspects of Superfund liability that have garnered attention
over the years. While we cannot address them all, in the section below we
provide a brief summary of two concerns most commonly heard during this
period. Many of these issues relate to criticisms raised by the insurance
industry and many corporations throughout the late-1980s and early-1990s
in anticipation of the Superfund reauthorization debates in the mid-1990s.
These companies lobbied to end the liability regime. The conundrum was
how to fund cleanups if liability were eliminated.
There were many efforts to find a workable proposal. Included among
these efforts were those of two groups that each brought together a panoply
of stakeholders. Vermont Law School and The Keystone Center organized
the first group, the National Commission on Superfund,242 which brought
together CEO’s from key industry sectors—including insurance, chemical,
petroleum, and waste management—with representatives from state and
local governments, citizen groups, tribes, communities of color, and
national environmental organizations. The National Commission on
Superfund met from 1992 to 1993 and issued a consensus report in 1994.243
Later, the EPA Administrator commissioned an advisory panel of
experts to advise her on Superfund reforms under the auspices of the
Superfund Evaluation Subcommittee of the National Advisory Council for
Environmental Policy and Technology (NACEPT).
The Superfund
Evaluation Subcommittee of NACEPT met from 1993 to 1994 and issued a
series of policy recommendations (though no formal final report) in 1994.
239. United States v. Atl. Research Corp., 551 U.S. 128, 139 (2007).
240. Id. at 140.
241. See GAO-09-656, supra note 20, at 48–50 (discussing unresolved issues that recent cases
may or may not have settled).
242. KEYSTONE CTR. & ENVTL. LAW CTR. OF VT. LAW SCH., FINAL CONSENSUS REPORT OF THE
NATIONAL COMMISSION ON SUPERFUND 1 (1994).
243. Id.
236
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
These two efforts, in addition to myriad other reform initiatives, culminated
in an Administration Bill, HR 3800, the Superfund Reform Act of 1994,
which was introduced by the Clinton Administration in June.244 While a lot
of effort went into trying to find a solution that all could live with, in the
end there was no reform of the liability regime and no reauthorization of
Superfund in 1994.
The two criticisms raised most frequently during this period of
uncertainty in the early to middle years of the Superfund program were
complaints about transaction costs and fairness. Some aspects of these
issues resolved themselves over time and others were addressed, at least in
part, through subsequent amendments to the law.
a. Transaction Costs
Among the most frequent criticisms of the liability system was that too
much money was spent on lawyers and consultants and too little on
cleanup. Now the transaction cost issue is rarely mentioned, and for good
reason: many of the questions about the meaning of certain provisions of
the law have been answered by federal courts all the way up to the U.S.
Supreme Court and cleanup studies and the construction or remedies are
farther along. In the early years of the program, relatively little information
was available on the exact nature and extent of the transaction costs
involved in a Superfund cleanup. Little was known about the meaning of
the statute and little progress had been made toward cleanup.
A study by the RAND Institute for Civil Justice in 1992 attempted to
assess the transaction costs for five very large industrial firms (Fortune 100
companies), representing a mix of petroleum, chemical, and manufacturing
firms, and four major insurance companies, representing approximately
fifteen percent of the 1980 market for Comprehensive General Liability
Policies.245 RAND defined transaction costs as those costs not directly
related to cleanup, such as allocating costs among PRPs and litigation
issues. In some cases it was difficult to determine whether, for example,
some engineering studies of the waste at a site are to assist in cleanup or to
identify PRPs and allocate costs.246
Researchers found that for the years 1984–89, on average, transaction
costs accounted for twenty-one percent of the industrial firms’ total
244. Superfund Reform Act of 1994, H.R. 3800, 103rd Cong. (1994).
245. JAN PAUL ACTON & LLOYD S. DIXON, SUPERFUND AND TRANSACTION COSTS: THE
EXPERIENCES OF INSURERS AND VERY LARGE INDUSTRIAL FIRMS 15, 33 (1992).
246. Id. at 36.
2009]
Superfund at 30
237
hazardous waste expenditures.247 Transaction costs decline significantly as
a percentage of total costs once cleanup begins. RAND found that
transaction costs at sites in the cleanup-ongoing stage averaged twenty-five
percent, twenty percentage points lower than transaction cost share for sites
in the site-study stage. Transaction costs for sites in the constructioncomplete stage averaged sixteen percent of total costs, thirty-three
percentage points lower than the transaction cost share for sites in the sitestudy stage.248 This explains why the transaction cost issue is rarely
mentioned now that more sites are farther along in the construction of the
remedy.
In contrast, RAND found that insurance company transaction costs
averaged eighty-eight percent of total expenditures for the years 1986–89.249
The large discrepancy in this study between the insurance companies’
(eighty-eight percent) and PRPs’ (twenty-one percent) average transaction
cost shares can partly be explained by the fact that insurers were still
litigating whether or not they could be held liable for Superfund costs under
Comprehensive General Liability (CGL) policies. The issue for insurance
companies is not per se liability under Superfund since, in most cases,
insurance companies do not fall into any of the categories of PRPs (owner,
operator, generator, etc.). The issue for insurance companies is how these
CGL policies are interpreted in light of Superfund liability imposed on the
policy holders.
Insurance law is determined by each state, so insurance cases are unlike
cost recovery claims against PRPs, which are resolved in federal court.
PRPs learned comparatively quickly, though many precedents were not
fully resolved until the late-1990s or later, that Superfund liability could
apply to them. Insurance companies, however, were disputing whether
CGL policies cover costs incurred as a result of Superfund liabilities. Most
of these disputes related to old CGL policies. Policies now include
“pollution exclusion” language. RAND found coverage disputes accounted
for about forty percent of the insurers’ total hazardous waste cleanup
expenditures.250 Similarly, about forty percent of insurance company
transaction costs in the study fell into the category “policy holder
247. Id. at 39. The study included state and voluntary cleanups in addition to those under
Superfund.
248. Id. at 51, 59.
249. Id. at 24.
250. ACTON & DIXON, supra note 245, at 24 (showing coverage disputes accounted for fortytwo percent).
238
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
defense.”251 Many insurance policies obligate the insurer to defend the
policy holder in litigation.
An important reflection on insurance company transaction costs is that,
given their duty to defend private parties from suits requiring private parties
to clean up or to reimburse cleanup costs, high transaction costs may mean
that the insurance company succeeded in defending the private party from
having to pay any cleanup costs. This means that an insurance company
with 100% transaction costs paid no cleanup costs because all its costs
related to litigation that successfully defended its policy holder from
liability.
Still other researchers pointed out that other ways to finance and
cleanup hazardous sites, such as taxes and public works programs, would
have transaction costs as well,252 and that there would certainly be
significant transaction costs involved in the transition between the current
program and the new programs.253
b. Fairness
Many critics objected to the inequity of Superfund’s liability
provisions. There are a variety of ways Superfund is said to be “unfair.”
One way is through the broad liability provisions. PRPs complain that
strict and retroactive liability is unfair because they may be held liable for
the costs to clean up waste that was disposed of in what was thought, at the
time, to be a responsible manner.
Joint and several liability is also raised in discussions about fairness of
the program. The PRPs that are solvent or especially profitable point to the
unfairness of being tagged for a greater share of the cleanup than the share
allocated to PRPs that are either insolvent or so poor that the legal costs of
pursuing them exceeds the value of their likely contribution to the costs of
cleanup. In other instances, EPA might have identified several solvent
parties, but for a variety of reasons, including reducing their own
transaction costs, EPA might pursue only the “deepest pocket.” Efforts to
address some of the inequities created by strict, joint and several liability
251. Id.
252. Henry L. Diamond, A Lawyer's View of Superfund Problems: Promote Hazardous Site
Cleanup by Increasing Voluntary Private Participation, in WORKING PAPERS ON SUPERFUND REFORM:
PROBLEM DEFINITION AND POLITICAL MAPPING 15 (1992).
253. See generally KATHERINE N. PROBST & PAUL R. PORTNEY, RESOURCES FOR THE FUTURE,
ASSIGNING LIABILITY FOR SUPERFUND CLEANUPS: AN ANALYSIS OF POLICY OPTIONS 14 (1992) (stating
that there would be new transaction costs associated with a transition from the existing liability scheme
to a new liability and funding approach for Superfund).
2009]
Superfund at 30
239
initially met with mixed success. As discussed above, in 1986 Congress
amended Superfund to authorize parties to seek contribution for their
cleanup costs.254 Any settling party at a site may sue any other non-settling
party to recover cleanup costs. This provision was intended to increase the
fairness of the program and to create additional incentives for parties to
settle with EPA. The only way for a PRP to shield itself effectively from
such third party contribution actions is to settle with EPA and as a part of
the settlement obtain contribution protection.255
As Superfund has matured the role of contribution protection has
remained important to balancing equities and to encouraging private party
settlements. Due the last three U.S. Supreme Court opinions on
CERCLA,256 contribution protection has also become more complex.
Contribution protection has always been limited only to “matters addressed
in the settlement.”257 Furthermore, the Court’s decision authorizing cost
recovery actions under section 107 for private parties in Atlantic Research
Corp. called into question contribution protection provided by settlements
that did not anticipate private party cost recovery actions.258
On the other hand, contribution protection may have new benefits for
settling parties after Atlantic Research. The issue has arisen in the context
of a non-settling PRP being sued by a party that settled with EPA and thus
has contribution protection.259 Because the standard of liability in cost
recovery claims is joint and several liability, non-settling parties that are
unable to meet the burden of showing that the harm is divisible risk being
held responsible for all the settling PRP’s costs.260 This is because the
settling party’s contribution protection prevents the non-settling party from
successfully making a contribution counter-claim.261
254. Superfund Amendments and Reauthorization Act (SARA) of 1986, Pub. L. No. 99-499,
100 Stat. 1613 (codified at 42 U.S.C. § 9613(f)(1), (f)(3)(B) (2006)).
255. See 42 U.S.C. §§ 9613(f)(2), 9622(h)(4) (2006) (Both sections state that “[a] person who
has resolved its liability to the United States or a State in an administrative or judicially approved
settlement shall not be liable for claims for contribution regarding matters addressed in the settlement.”).
256. Cooper Indus. Inc. v. Aviall Servs. Inc., 543 U.S. 157 (2004); United States v. Atl. Research
Corp., 551 U.S. 128 (2007); Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870 (2009).
257. § 9613(f)(2).
258. Atl. Research Corp., 551 U.S. at 139–40. See, e.g., United States v. Kramer, Civil Action
Nos. 89-4340 (JBS), 89-4380 (JBS), 2009 WL 2339341, at *8 (D.N.J. July 27, 2009) (granting a settling
party’s motion to file a claim under 42 U.S.C. § 9607(a)(4)(B) against a non-settling party who will
have no counter-claim for contribution because the settling party is shielded from such claims,
potentially resulting in the non-settling party being held responsible for all of the settling party’s costs).
259. Id.
260. Burlington N., 129 S. Ct. at 1881.
261. See, e.g., Kramer, 2009 WL 2339341, at *8.
240
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
The most recent in the trilogy of U.S. Supreme Court cases to address
CERCLA liability and its related fairness issues is the May 2009 decision
Burlington Northern & Santa Fe Railway Co. v. United States.262 Justice
Stevens’s decision went to the heart of many PRPs’ fairness concerns—
joint and several liability. The Court confirmed that joint and several
liability is applicable to sites where the harm is not divisible or otherwise
capable of a reasonable basis for division and applied that requirement to
the facts presented in a less rigorous manner than had been the case in the
past.263 Whether this decision will increase uncertainty or alleviate fairness
issues concerning the role of joint and several liability remains to be seen as
the circuit courts resolve future claims concerning joint and several liability
under CERCLA.264
Others to complain about the liability regime included local
governments, lenders, small businesses, real estate developers, property
insurance and re-insurance companies, and neighboring landowners.
Ultimately many, though not all of these groups, received some relief.
Exemptions and defenses from liability were added in amendments to the
statute, including protections for lenders that foreclose on contaminated
properties265 and some generators and transporters that recycle certain types
of materials.266 Most significant were the exemptions provided in the Small
Business Liability Relief and Brownfields Revitalization Act of 2002.267
These exemptions provided additional protections to landowners, including
those that meet the requirements of “bona fide prospective purchasers,”268
262. Burlington N., 129 S. Ct. at 1881.
263. Id. at 1882–84 (affirming a lower court ruling that had voided joint and several liability
based on an analysis better suited to contribution allocation than divisibility of harm.) Contrary to
precedent, the district court acted on its own initiative to apportion the harm at the site by calculating
how long the property was leased and how much of the contaminated property was owned by the
railroads, with some consideration of the limited number of hazardous substances, the monitoring nature
of the remedy imposed on the railroad’s portion of the property, and an allowance for “calculation
errors.” Id.
264. For one example of how courts are interpreting these Supreme Court decisions, see
Kramer, 2009 WL 2339341, at *8.
265. Asset Conservation, Lender Liability, and Deposit Insurance Protection Act of 1996,
§ 2502(a) (amending 42 U.S.C. § 9607 by adding § 107(n)).
266. Superfund Recycling Equity Act of 1999, Pub. L. No. 106-113, 113 Stat. 1501A (codified
at 42 U.S.C. § 9627 (2000)).
267. Small Business Liability Relief and Brownfields Revitalization Act of 2002, Pub .L. No.
107-118, 115 Stat. 2356 (2002) (codified as amended in scattered sections of 42 U.S.C. §§ 9601, 9604,
9605, 9607, 9622(g)) (amending CERCLA to add additional circumstances wherein parties may avoid
liability if certain conditions are met.).
268. 42 U.S.C. § 9607(r) (2006).
2009]
Superfund at 30
241
“contiguous property owners,”269 and residential landowners.270
Municipalities,271 certain 501(c)(3) certified organizations,272 home owners
and renters,273 some small businesses,274 and very small quantity
generators,275 like the kennels and pizza parlors that were once the subject
of perhaps over-expansive contribution claims,276 garnered additional
protections from liability associated with sites listed on the NPL.
III. ISSUES ON THE HORIZON
CERCLA, and the Superfund Program it established, was designed to
identify contaminated sites, assure cleanup by either the government or
private parties, make those connected to the contamination pay for cleanup,
and enlist private interests in the identification and prevention of further
hazards. In many ways CERCLA has been wildly successful. The EPA has
investigated over 47,000 sites suspected of releasing hazardous substances
into the environment.277 Many of these sites have been addressed by
removal actions, state authorities, or by private parties.278 Just over 1600
sites have been placed on the NPL, and cleanups have been fully
implemented at more than two-thirds of these sites.279 Private parties have
contributed dollars or work to clean up at an estimated value of nearly thirty
billion dollars over the history of the program,280 averaging twenty-six
million dollars per site.281 In addition, Superfund liability has changed the
way many companies handle hazardous substances. Many private parties
changed their business practices to prevent or limit future liability once it
became clear that liability was real—and expensive—both in terms of
dollars and, for many companies, in terms of public relations.
269. § 9607(q).
270. 42 U.S.C. § 9601(40)(B)(iii) (2006).
271. § 9607(p).
272. § 9607(p)(1)(C).
273. § 9607(p)(1)(A).
274. § 9607(p)(1)(B).
275. § 9607(o).
276. Robert Tomsho, Pollution Ploy: Big Corporations Hit by Superfund Cases Find Way to
Share Bill—They Sue Small Businesses, Others That Put Garbage into the Same Landfills—An Effort to
Change the Law?, WALL ST. J., April 2, 1991, at A1.
277. GAO-09-656, supra note 20, at 13 n.11.
278. Id.
279. U.S. Envtl. Prot. Agency, National Priorities List: NPL Site Total by Status and Milestone,
www.epa.gov/superfund/sites/query/queryhtm/npltotal.htm (last visited Oct. 27, 2009).
280. U.S. GOV’T ACCOUNTABILITY OFFICE, SUPERFUND: FUNDING, supra note 132, at 15 tbl.2.
281. Id. at 17.
242
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
The fact that so much has been accomplished does not mean, however,
that Superfund’s work is done. Unfortunately, there are still more sites
across the country needing cleanup, and some of these sites are technically
complex and extremely expensive to address. And questions still remain
about the degree to which government and private parties achieve quality
cleanups in a reasonable amount of time and whether those sites likely to
have long-term contamination are properly monitored. Other questions
relate to funding and liability. Below we identify four questions that we
believe must be addressed now to ensure that, over the next ten years, the
Superfund program achieves steady progress. We end with a plea for
increased transparency and an investment in independent program
evaluation in order to ensure increased credibility for the Superfund
program in the years to come.
A. Can the Quality and Pace of Cleanups Be Improved?
Superfund’s ability to advance the goals envisioned by its creators
would be much improved by better information about what works and what
does not in achieving the fundamental goal of the statute: cleanup.
Questions about the pace and quality of cleanup still hound the Superfund
program. While it is clear that the challenges of remediating contaminated
sites preclude “speedy” cleanup, it can take a decade or longer to address
some NPL sites. EPA needs to investigate the pace of cleanup for two
important reasons: 1) to evaluate whether there are program reforms that
could be implemented that would accelerate high quality cleanups, and 2) to
be able to communicate to Congress and the public why, in some cases,
cleanup actually cannot be accelerated (or not by much). We recommend
an independent study to addresses these issues.
Concern about the quality of Superfund cleanup is not helped by a
dearth of information about what cleanups are accomplishing. Through the
1990s, EPA issued annual reports summarizing the remedies selected at
NPL sites. During this period there were also many external studies
evaluating cleanup remedies. In the past decade, however, there have been
few evaluations of Superfund cleanups from either EPA or external experts.
Yet, assessing what is—and is not—being accomplished in terms of setting
and achieving cleanup goals is crucial to program success. EPA should
synthesize information on the selection of remedies and cleanup goals
across all NPL sites and make this information publicly available, along
with clear information about whether these goals are being achieved.
This is even more of an issue for those voluntary cleanups conducted
by private parties without oversight from EPA or states. These cleanups
2009]
Superfund at 30
243
take place outside of the federal program. Parties use the liability
provisions of the law to help them recover costs for their own efforts to
clean up or they clean up sites to avoid facing possible enforcement actions.
It is unclear to what extent these privately addressed sites really are “clean”
or are properly monitored to assure no further exposure to the hazards that
remain after a removal or remedial action. Most of the information that is
publicly available about private party practices outside federal and state
requirements stems from private efforts in court to spread the cost of
cleanup to other parties, which until recently usually involved relatively
little judicial review of the actual cleanup. Moreover, to the degree parties
resolve cleanup and cost allocation issues outside of court or government
supervision, no information is available.
B. What Steps Need to Be Taken to Ensure a Robust and Reliable
Long-term Stewardship—or Post-cleanup—Program?
It is now generally accepted that at many NPL sites, even after cleanup
is completed, the site will not be appropriate for unrestricted use and that
some kind of “long-term stewardship”282 program will be required for years,
if not decades, to come. Long-term stewardship typically involves: legal
controls over the use of a site (and water and groundwater resources) when
contamination remains on site at levels that preclude unrestricted use; site
monitoring and maintenance; and mechanisms for ensuring that site
restrictions are enforced and that any site risks are communicated to the
public.
While the topic of long-term stewardship has garnered increasing
attention over the past decade, with reports sponsored by the Department of
Energy as well as guidance documents from EPA, there are, in fact, no
regulatory requirements governing long-term stewardship. As a result, each
site, whether on the NPL or not, is faced with the somewhat daunting task
of creating its own stewardship program and requirements. EPA needs to
fill this regulatory “hole” and develop regulatory requirements and, equally
important, program support for stewardship activities at contaminated sites
across the nation.
282. See, e.g., KATHERINE N. PROBST AND MICHAEL H MCGOVERN, LONG-TERM STEWARDSHIP
NUCLEAR WEAPONS COMPLEX: THE CHALLENGE AHEAD, at viii-ix (1998) (stating that some
contaminants at DOE sites will remain a concern for hundreds of years, and outlining the key elements
of a long-term stewardship program for contaminated sites).
AND THE
244
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
C. Will Private Parties Continue To Contribute Work and Money to
NPL Cleanups, and Will They Continue To Clean Up Sites on Their
Own?
In the last few years, the U.S. Supreme Court has shaken the
foundations of CERCLA liability, with the possible result that private
parties may no longer so willingly settle with government authorities or
conduct voluntary cleanup. During the last five years, the U.S. Supreme
Court issued three CERCLA opinions, the last of which, Burlington
Northern, was decided on May 4, 2009 and may be the most troubling.
The first decision, Aviall,283 limited private party contribution claims to
work done or paid for as a consequence of a civil action or certain
settlement agreements. In effect, the Court concluded that claims for
contribution may only be used when a civil action or related settlement has
made the claiming party liable for more than its fair share of cleanup costs.
This decision demolished the foundation of many private party contribution
claims. All such private party claims that related to costs for voluntary,
state and federally supervised cleanup without the prerequisite civil action
were dismissed because so many circuit courts had restricted all private
party efforts to allocate costs, even voluntary cleanup costs, to the
contribution provisions of the statute that the Court curtailed in Aviall.
Cooperating with government authorities or cleaning up on one’s own
suddenly was not a promising way to spread cleanup costs to noncooperating parties.
The cracks in the foundation were patched a few years later when the
U.S. Supreme Court decided Atlantic Research Corp.284 In this case, Justice
Thomas reminded CERCLA advocates and judges that authority for
recovering costs for voluntary cleanup resides in the liability provisions
originally written into the statute, rather than the contribution provisions
drafted later and on which most courts had relied prior to Aviall. While this
decision seemed to repair the liability regime, it also has called into
question the value of the contribution-protection provisions provided in
settlement agreements with EPA. Private party claims may—under some
circumstances created by private party cost recovery after Atlantic
Research—pierce these shields from private party suits. On the other hand,
contribution protection provisions in settlement agreements may prevent
contribution counter-claims when parties that have settled with the U.S.
283. Cooper Indus. Inc. v. Aviall Servs. Inc., 543 U.S. 157 (2004).
284. United States v. Atl. Research Corp., 551 U.S. 128 (2007).
2009]
Superfund at 30
245
bring subsequent cost recovery claims.285 Depending on the role of
contribution protection clauses in the face of private party cost recovery
claims, private parties may have one less reason or one more reason to
settle with EPA.
The last decision, however, is the most puzzling one of the triad and the
most significant reason why we must ponder whether liability will continue
to spur voluntary cleanup and government settlement agreements for
cleanup work and compensation. The Burlington Northern286 decision may
be a hairline crack in CERCLA’s liability regime or it may be an earthquake
destroying much that has made the statute successful at spurring private
party action. The Court found that joint and several liability is part of the
common law of CERCLA, and, as such, its application may be avoided
when the harm is divisible. So far, so good, since this principle was well
established in CERCLA. The practical application of the principle means
that few private parties avoid joint and several liability, since more often
than not contamination crosses borders, co-mingles with other hazardous
substances, and generally results in cleanup response that cannot be
distinguished according to a PRPs role at the site. This means that EPA
could recover all its costs from one or more PRPs in these cases.287
In Burlington Northern, however, the Supreme Court affirmed a district
court ruling that voided joint and several liability based on an analysis
better suited to contribution allocation than divisibility of harm.288 Contrary
to precedent, the district court acted on its own initiative to apportion the
harm at the site by calculating how long the property was leased and how
much of the contaminated property was owned by the railroads, with some
consideration of: the limited number of hazardous substances, the
monitoring nature of the remedy imposed on the railroad’s portion of the
property, and an allowance for “calculation errors.”289
285. See, e.g., United States v. Kramer, Civil Action Nos. 89-4340 (JBS), 89-4380 (JBS), 2009
WL 2339341, at *8 (D.N.J. July 27, 2009) (granting a settling party’s motion to file a claim under 42
U.S.C. § 9607(a)(4)(B) against a non-settling party whom will have no counter-claim for contribution
because the settling party is shielded from such claims, potentially resulting in the non-settling party
being held responsible for all of the settling party’s costs).
286. Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870 (2009).
287. Fairness among private parties was introduced in concurrent or subsequent claims for
contribution that allowed district courts to allocate the costs—paid to clean up the site or to reimburse
EPA under joint and several liability—among the PRPs after EPA or the private party initiating cleanup
had been paid in full. Without the application of joint and several liability, contribution claims are
unlikely to succeed.
288. Burlington N., 129 S. Ct. at 1884.
289. Id. at 1882.
246
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
How disruptive will these decisions be to the frequency and pace of
voluntary or other private party cleanup in the future? Specifically, will
EPA be able to continue to negotiate cleanup and reimbursement
settlements after Atlantic Research Corp. and Burlington Northern? Will
Atlantic Research Corp. increase voluntary cleanup because of the possible
option of suing parties that had settled with EPA in the past? Will
Burlington Northern result in more litigation with EPA because of
divisibility arguments, and less cleanup because of reluctance to engage in
voluntary cleanup for lack of contribution claims after an apportionment?
Based on the Court’s interpretation of the law alone, nothing should change;
but based on the Court’s affirmation of Burlington Northern’s unusual
divisibility of harm analysis, the foundation of CERCLA liability may be
shattered.
The answers to these questions will affect the continued progress of
cleanup and are one factor to consider in how much money the Superfund
program will need for the next five to ten years. DOJ and EPA will likely
need more money for enforcement until the ramifications of these decisions
are known. And, if the result of these cases is that PRPs are much less
willing to agree to undertake site studies and cleanups directly, the federal
Superfund program could need a lot more money than is currently being
appropriated.
D. What Are the Funding Needs for the Superfund Program over the
Next Five to Ten Years?
A well-funded Superfund program is critical not only to ensuring that
EPA has the funds it needs to evaluate potential sites, move contaminated
sites through the cleanup process, and address contamination where private
parties cannot be found or are recalcitrant, but it is also a critical element of
ensuring that the clout of the government will encourage private parties to
come forward. For the threat of EPA action to be credible, the agency must
have the funds to remediate sites when responsible parties fail to do so.
But what are the funding needs for the Superfund program over the
next five to ten years? The last comprehensive analysis of the future
funding needs for the Superfund program was the report to Congress
published by Resources for the Future in July 2001. Since that time, while
the question of future funding needs has been raised repeatedly and there
have been myriad reports documenting funding shortfalls for cleanup
activities, there has been no comprehensive, credible estimate of the
program’s future funding needs, or of the number of sites warranting listing
on the NPL in the future. This makes it impossible to ascertain if current
2009]
Superfund at 30
247
annual appropriations of approximately $1.3 billion are too little, too much,
or just the right amount of funding, or whether reinstatement of the
Superfund taxes that expired at the end of 1995 is needed. This lack of
information needs to be remedied.
Ideally, the EPA Administrator should initiate such a process herself,
but, if not, then Congress should again request an estimate of future funding
needs. To ensure the credibility of the analysis, the study could either be
done by an independent government organization, such as the
Congressional Budget Office, or by an independent non-profit research or
academic organization. What is critical is that the analysis be independent
and credible, and that the methodology used to derive cost estimates be
clearly documented for all to see.
E. Improving Superfund Requires Increased Program Evaluation
and Greater Transparency
The Superfund program needs to increase the transparency of its
operations, accomplishments, and planned activities—both in regards to the
cleanup program and to enforcement activities. The agency also needs, as
suggested above, to invest in evaluations and studies that will provide the
agency the information it needs to improve its effectiveness. A betterinformed debate about what works in the Superfund program and what does
not work is in the public interest. Without a sound analytical basis for
determining program reforms, the result is policy reform by anecdote,
something that should be avoided.
The EPA should develop a proactive agenda for evaluating key aspects
of the cleanup and enforcement programs, and it should reinstate the annual
reports to Congress, required under section 301(c) of CERCLA, that
included concise but comprehensive information about annual program
accomplishments as well as issues on the horizon. This kind of effort is
needed to ensure not only that the program has the information it needs to
guide future policy decisions, but it will also serve to improve the
credibility of the program.
Chances are the Superfund program will be with us for many years to
come. To make sure we spend scarce dollars—whether paid for by the
taxpayer or by private companies—wisely and effectively, we need to
investigate ways to improve implementation of what is an important
environmental program.
THE SUPERFUND COST ALLOCATION LIABILITY
CONFLICTS AMONG THE FEDERAL COURTS
Steven Ferrey∗
TABLE OF CONTENTS
Introduction ............................................................................................... 250
I. The First Step: Ocean’s Eleven—How Eleven Federal Circuit Court
Uniformly Misread the “Plain Meaning” of CERCLA for a Decade 253
A. Wrong at Every Turn ....................................................................... 253
B. How Ocean’s Eleven Disrupted the Superfund ............................... 259
II. The Second Step: The 2004 Aviall Supreme Court Decision ............. 264
III. The Third Step: The 2007 Atlantic Research Detonation .................. 267
A. U-Turns in the Second and Eighth Circuits ..................................... 267
B. The New Path of Atlantic Research................................................. 271
IV. How Have Various Courts Responded to the Supreme Court?.......... 276
A. Responding to the Aviall Section 113 Settlement Blockade ........... 277
B. Summary .......................................................................................... 291
V. Responding to the 2007 Unanimous Atlantic Research Reversal of
Section 107 Precedent ....................................................................... 292
A. First Circuit...................................................................................... 292
B. Second Circuit ................................................................................. 293
C. Third Circuit .................................................................................... 296
D. Fourth Circuit .................................................................................. 298
E. Fifth Circuit...................................................................................... 298
F. Sixth Circuit ..................................................................................... 299
G. Seventh Circuit ................................................................................ 300
H. Eighth Circuit .................................................................................. 301
I. Ninth Circuit ..................................................................................... 301
J. Tenth Circuit ..................................................................................... 303
∗ Professor of Law, Suffolk University Law School; Visiting Professor of Law, Harvard Law
School, 2003. Professor Ferrey advises the United Nations and World Bank Sustainable Development
Vice Presidency on international energy and environmental issues around the world. He is the author of
6 books and more than 80 environmental articles. He holds a B.A. in economics, an M.A. in
environmental planning, a J.D., and upon graduation from law school was a postdoctoral Fulbright
Fellow at University College, London.
250
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
K. Eleventh Circuit ............................................................................... 303
L. D.C. Circuit ...................................................................................... 303
Conclusion................................................................................................. 304
INTRODUCTION
Seldom in judicial history does the Supreme Court decide a case where
it can, let alone will, reverse the standing precedent articulated by
essentially every federal judicial circuit court across the nation. Seldom is
there such a profound difference of opinion. Step one: Eleven federal
circuits, one after the other in a compressed period, barred the use of
section 107 of the Comprehensive Environmental Response, Compensation,
and Liability Act (CERCLA),1 for cost recovery by most plaintiffs,
including: the Seventh2 (July 1994), the First3 (August 1994), the Tenth4
(March 1995), the Eleventh5 (September 1996), the Third6 (May 1997), the
Ninth7 (July 1997), the Fifth8 (July 1997), the Fourth9 (April 1998), the
Sixth10 (August 1998), the Second11 (September 1998), and the Eighth
(August 2003).12 This judicial cascade occurred at a time when the
Superfund hazardous substance cleanup effort was starved by budgetary
depravation. The Environmental Protection Agency (EPA) had to delay
beginning remediation activities at thirty-four separate priority sites in
fiscal year 2004 because of funding shortfalls.13
Step two: The practical result of the 2004 Supreme Court decision in
Cooper Industries v. Aviall Services, Inc.,14 coupled with the decisions of
the eleven circuit courts regarding section 107 in step one,15 coalesced to
1. Comprehensive Environmental Response, Compensation, and Liability Act, 42 U.S.C.
§§ 9601–9675 (2006).
2. Akzo Coatings, Inc. v. Aigner Corp., 30 F.3d 761 (7th Cir. 1994).
3. United Techs. Corp. v. Browning-Ferris Indus., 33 F.3d 96 (1st Cir. 1994).
4. United States. v. Colo. & E. R.R. Co., 50 F.3d 1530 (10th Cir. 1995).
5. Redwing Carriers, Inc. v. Saraland Apartments, 94 F.3d 1489 (11th Cir. 1996).
6. New Castle County v. Halliburton NUS Corp., 111 F.3d 1116 (3d Cir. 1997).
7. Pinal Creek Group v. Newmont Mining Corp., 118 F.3d 1298 (9th Cir. 1997).
8. OHM Remediation Serv. v. Evans Cooperage Co., Inc., 116 F.3d 1574 (5th Cir. 1997).
9. Pneumo Abex Corp. v. High Point, Thomasville & Denton R.R. Co., 142 F.3d 769 (4th Cir.
1998).
10. Centerior Serv. Co. v. Acme Scrap Iron & Metal Corp., 153 F.3d 344 (6th Cir. 1998).
11. Bedford Affiliates v. Sills, 156 F. 3d 416, 432 (2d Cir. 1998).
12. Dico, Inc. v. Amoco Oil Co., 340 F.3d 525, 531 (8th Cir. 2003).
13. Linda Roeder, Insufficient Funds for Cleanup Operations, Supreme Court Decision Lead
EPA Concerns, 36 ENV’T REP. (BNA) S-14, S-15 (2005).
14. Cooper Indus. Inc. v. Aviall Servs. Inc., 543 U.S. 157 (2004).
15. See infra Section II (discussing and analyzing the circuit court opinions).
2009] Cost Allocation Liability Conflicts Among the Federal Courts
251
greatly discourage voluntary remediation activities at many of the 450,000
contaminated sites in the United States.16 The nation’s hazardous liability
scheme had descended into total chaos because of the federal circuit court
decisions on section 107 cost recovery being unavailable and the 2004
Supreme Court Aviall decision walling off section 113 contribution to
remediation.
Step three: In 2007, in Atlantic Research Corp. v. United States,17 the
Supreme Court reversed the prior wall of consistent precedent from these
eleven unanimous federal circuit courts.18 It declared every federal circuit
court that had decided these cost recovery cases to be in total error in
statutory interpretation. Even more remarkably, the often divided Supreme
Court did so in a rare unanimous opinion in the Roberts Court.
Step four: The Atlantic Research decision nonetheless resulted in
uncertainty in the lower courts. Several circuits have not embraced or
implemented the Supreme Court decisions nor reversed their own contrary
precedent. Now, more than two years since Atlantic Research, there are
still critical gaps in the lower federal courts regarding CERCLA cost
allocation liability and recovery.19 These gaps are apparent and the
implications are key, with an almost half of a million estimated
contaminated sites in the United States that will cost hundreds of billions of
dollars to remediate.20
This stepped evolution has been a long and often conflicted process. I
published an article in 1994 that suggested that to resolve lower court
judicial splits on use of section 113 and section 107, the correct
interpretation of CERCLA by appellate courts was that section 107 was
16. U.S. GEN. ACCOUNTING OFFICE, “COMMUNITY DEVELOPMENT: LOCAL GROWTH ISSUESFEDERAL
OPPORTUNITIES
AND
CHALLENGES
118
(2000),
available
at
http://www.gao.gov/archive/2000/rc00178.pdf (“[N]ationwide, 450,000 brownfields await cleanup.”).
17. Atl. Research Corp. v. United States, 459 F.3d 827, 834–35 (8th Cir. 2006), aff’d, 551 U.S.
128 (2007).
18. See infra Section II.
19. See, e.g., Consol. Edison Co. of N.Y. v. UGI Util. Inc., 423 F.3d 90, 95 (2d Cir. 2005)
(finding that a contribution right only exists when liability for a CERCLA claim is resolved); City of
Waukesha v. Viacom, Int’l, Inc., 404 F. Supp. 2d 1112, 1115 (E.D. Wisc. 2005) (finding no contribution
right exists when only state environmental claims are resolved); Differential Dev.-1994 Ltd. v. Harkrider
Distrib. Co., 470 F. Supp. 2d 727, 729–30 (S.D. Tex. 2007) (discussing issues left unresolved by the
U.S. Supreme Court decision).
20. See EPA About Brownfields (2008), www.epa.gov/swerosps/bf/about.htm (estimating that
more than 450,000 contaminated sites require cleanup); EPA, CLEANING UP THE NATION’S WASTE
SITES: MARKETS AND TECHNOLOGY TRENDS 1–4 (2004), available at www.cluin.org/download/market/2004market.pdf (estimating the cost to clean up the remaining sites at $209
billion).
252
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
available to all private parties.21 Thereafter, as these disputes progressed to
appeal, the federal circuit courts commenced a 1994–2003 cascade of
unvarying opinions all holding contrary to the position I articulated in the
article, and often reversing their lower federal trial courts. This result
discouraged private party voluntary remediation at multi-party
contaminated sites precisely during the period when public Superfund
resources were dramatically scaled back by Congress. Simultaneously,
public resources were truncated by Congress and private incentives to
undertake voluntary remediation were obliterated by the federal circuit
court decisions.
In 2007, the Supreme Court was finally able to identify a conflict (set
up by circuit court equivocation after its former 2004 CERCLA opinion in
Aviall), grant certiorari, and, via a unanimous decision, reverse this decade
of circuit court decisions blocking section 107. It is this past fifteen years
of radical stepped evolution of Superfund hazardous substance cost
responsibility allocation that is analyzed here. It is a retrospective look at
the seeds of judicial conflict, as well as a look forward at how the courts are
internalizing, or not reflecting, the recent Supreme Court determinations.
This article analyzes each of the following steps.
Section I of this article analyzes step one, the action of all the federal
circuits between 1994–2003 walling off private hazardous waste cost
recovery actions pursuant to the long-standing and never amended section
107 of CERCLA, causing the entire national cost-recovery mechanism to
descend into chaos.
Section II examines step two, the pivotal impact of the 2004 Supreme
Court decision in Aviall, greatly limiting CERCLA section 113 contribution
liability allocation and complicating the circuit court precedential chaos.
Section III examines step three, the 2007 unanimous Supreme Court
opinion in Atlantic Research reopening section 107 based on “plain
meaning” statutory interpretation that none of the circuit courts had gotten
right.
Sections IV and V carefully chart how the lower courts have responded
to these new commands: specifically, whether they have implemented the
Supreme Court jurisprudence on CERCLA since 2004. The checkerboard
of results is surprising.
21. See Steven P. Ferrey, Allocation & Uncertainty in the Age of Superfund: A Critique of the
Redistribution of CERCLA Liability, 3 N.Y.U. ENVTL. L. J. 36 (1994).
2009] Cost Allocation Liability Conflicts Among the Federal Courts
253
I. THE FIRST STEP: OCEAN’S ELEVEN—HOW ELEVEN FEDERAL
CIRCUIT COURT UNIFORMLY MISREAD THE “PLAIN MEANING” OF
CERCLA FOR A DECADE
A. Wrong at Every Turn
For a decade, from 1994 to 2004, there were uncontradicted, parallel
decisions from every one of the eleven federal circuits considering section
107 Superfund private party liability allocation. One after the other, federal
circuits, often overruling their lower courts, cascaded down the same
chute—negating all rights of private potentially responsible parties (PRPs)
to utilize section 107 of Superfund to recover or share their remediation
response costs.
1. Roadblock: No Private PRP Circuit Court Access to Section 107
Since suggested by a federal district court in the Kramer opinion, the
section 107 private cost recovery route has been the preferred path of
private party plaintiffs for cost reallocation.22 Prior to 1994, none of the
circuit courts had directly addressed the issue of whether a PRP had
standing under section 107 to recover cleanup costs, and the Supreme Court
had only touched upon the question as a background issue in dicta.23
District courts split on whether a PRP could elect between a section 107 and
a section 113 claim, or was restricted to section 113.24
On its face, section 107(a)(4)(B) is available to “any . . . person” other
than the sovereigns who are listed and otherwise enabled in section
107(a)(4)(A).25 In 1994, there began a decade of cascading federal circuit
opinions. Over the course of four years, ten circuits confronted the question
22. United States v. Kramer, 757 F. Supp. 397 (D.N.J. 1991).
23. See Key Tronic Corp. v. United States, 511 U.S. 809, 816 (1994) (noting that the Court did
not answer the question of whether only “innocent” parties had standing under section 107 cost-recovery
claims; rather, the Court merely held that section 107 did not provide for the award of attorney’s fees).
24. For district courts that had allowed PRPs to raise section 107 claims, see, e.g., Companies
for Fair Allocation v. Axil Corp., 853 F. Supp. 575, 579 (D. Conn. 1994), finding that PRPs were
allowed to raise section 107 claims. See also United States v. SCA Serv. of Ind., Inc., 849 F. Supp.
1264, 1282 (N.D. Ind. 1994) (permitting a PRP to pursue a section 107 claim); Transp. Leasing Co. v.
California (CalTrans), 861 F. Supp. 931, 938 (C.D. Cal. 1993) (finding that PRPs were allowed to bring
claims under section 107). For district courts that have held that PRP may not use section 107 to
recover response costs, see, e.g., SC Holdings, Inc. v. A.A.A. Realty Co., 935 F. Supp. 1354, 1362–65
(D.N.J. 1996) (finding that CERCLA limits PRPs to contribution claims under section 113) and
Kaurfman v. Unisys Corp., 868 F. Supp. 1212, 1215 (N.D. Cal. 1994) (finding that PRPs are confined to
section 113).
25. 42 U.S.C. § 9607(a) (2006).
254
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
of whether section 107 of CERCLA can be utilized by PRPs to reallocate
their cost of voluntary cleanup at a hazardous waste site. Each of the
circuits, many reversing their district courts, blocked the path dictated by
section 107’s unambiguous language.
In a compressed period of four years, ten of the twelve circuits came to
decide or opine on the availability of section 107 cost recovery in the
following sequence: the Seventh26 (July 1994), the First27 (August 1994),
the Tenth28 (March 1995), the Eleventh29 (September 1996), the Third30
(May 1997), the Ninth31 (July 1997), the Fifth32 (July 1997), the Fourth33
(April 1998), the Sixth34 (August 1998), and the Second35 (September
1998). The Eighth Circuit followed in 2003.36 All of these circuits closed
off section 107 to private party plaintiffs. While the D.C. Circuit court had
never heard such a section 107 case, all of the remaining eleven federal
circuits were unanimous in their holding on this essential interpretation.
Despite similar fact patterns, each circuit court took its own approach to
disposing of the arguments put forward in favor of plaintiff PRP standing
under section 107. Some courts attempted to interpret the language in
section 107 and section 113 by looking at legislative and legal history and
engaging in a textual analysis of the provisions. Other circuit courts simply
ignored the express operative “any other person” language in section
107(a)(4)(B), and instead construed only section 113 as a backhanded way
to limit section 107. Some circuit courts hold that only their often strained
reading of CERCLA does not frustrate the goals of the statute.
Several circuits wrestled with the nature of section 113 contribution. In
an attempt to decipher the parameters of the claims under section 107 and
section 113, six37 of the eleven circuit courts discuss the Superfund
26. Akzo Coatings, Inc. v. Aigner Corp., 30 F.3d 761 (7th Cir. 1994).
27. United Techs. Corp. v. Browning-Ferris Indus., 33 F.3d 96 (1st Cir. 1994).
28. United States v. Colorado & E. R.R. Co., 50 F.3d 1530 (10th Cir. 1995).
29. Redwing Carriers, Inc. v. Saraland Apartments, 94 F.3d 1489 (11th Cir. 1996).
30. New Castle County v. Halliburton NUS Corp., 111 F.3d 1116 (3d Cir. 1997).
31. Pinal Creek Group v. Newmont Mining Corp., 118 F.3d 1298 (9th Cir. 1997). The court in
Adhesives Research v. American Inks & Coatings, 931 F. Supp. 1231, 1243–46 (M.D. Pa. 1996), adopts
the Pinal Creek court’s discussion of CERCLA policy in its entirety.
32. OHM Remediation Serv. v. Evans Cooperage Co., Inc., 116 F.3d 1574 (5th Cir. 1997).
33. Pneumo Abex Corp. v. High Point, Thomasville & Denton R.R. Co., 142 F.3d 769 (4th Cir.
1998).
34. Centerior Serv. Company v. Acme Scrap Iron & Metal Corp., 153 F.3d 344, 350 (6th Cir.
1998).
35. See Bedford Affiliates v. Sills, 156 F.3d 416, 432 (2d Cir. 1998) (holding that Bedford
could not pursue a section 107(a) claim).
36. Dico, Inc. v. Amoco Oil Co., 340 F.3d 525 (8th Cir. 2003).
37. The six circuit courts are the following: First, Second, Third, Sixth, Ninth, and Tenth.
2009] Cost Allocation Liability Conflicts Among the Federal Courts
255
Amendments and Reauthorization Act (SARA)38 and how this 1986
amendment codified the common law right to contribution. The Second,
Third, and Tenth Circuits most exhaustively develop the arguments for why
SARA, which did not alter the pre-existing section 107, by silent
implication precludes a PRP from using section 107.39 In distinguishing
“cost recovery actions” under section 107 from “contribution actions”
under section 113, five40 of the eleven circuit courts determined that an
action between PRPs for apportionment of cleanup costs is always an action
for contribution, notwithstanding section 107's contrary language. The
Sixth, Seventh, and Tenth Circuits reached back to either Black’s Law
Dictionary,41 the Restatement (Second) of Torts42 or American
Jurisprudence43 in defining the term “contribution” in the legal context.44
The First Circuit countered the clear inclusiveness of the express “any
other person” language of section 107 by stating that courts must strive to
give effect to each subsection in a statute, such as section 113, “indeed, to
give effect to each word and phrase.”45 To give effect and force to section
113, these courts then ignored the language of section 107. Many of these
circuits strain to avoid confronting or construing the plain language of
38. Superfund Amendments and Reauthorization Act of 1986 (SARA), Pub. L. No. 99–499, §§
101–405, 100 Stat. 1613 (1986).
39. The three courts reason that SARA, as well as the pre-SARA case law (recognizing an
implicit right of contribution), establish the principle that PRPs should not be exposed to joint and
several liability in actions by other PRPs seeking to recover cleanup costs. See Bedford Affiliates, 156
F.3d at 435 (holding that a PRP can never recover 100% of response costs since it is a joint tortfeasor);
New Castle County v. Halliburton NUS Corp., 111 F.3d 1116, 1122–24 (3d Cir. 1997) (holding that
actions for contribution must be made under section 113); United States v. Colo. & E. R.R. Co., 50 F.3d
1530, 1535 (10th Cir. 1995) (citing legislative history stating that a principle goal in creating section 113
was to “clarif[y] and confirm the right of a person held jointly and severally liable under CERCLA to
seek contribution from other potentially liable parties, when the person believes that it has assumed a
share of the cleanup or cost that may be greater than its equitable share under the circumstances”) (citing
S. Rep. No. 99–11, at 44 (1985)).
40. The five circuit courts are the following: First, Third, Sixth, Seventh, and Tenth.
41. Defining contribution as the “right of one who has discharged a common liability to
recover of another also liable, the aliquot portion which he ought to pay or bear.” BLACK’S LAW
DICTIONARY 328 (6th ed. 1990).
42. RESTATEMENT (SECOND) OF TORTS § 866(A) (1977).
43. 18 AM. JUR. 2D Contribution § 9 (2004).
44. See Centerior Serv. Co. v. Acme Scrap Iron & Metal Corp., 153 F.3d 344, 350 (6th Cir.
1998) (applying RESTATEMENT (SECOND) OF TORTS and BLACK’S LAW DICTIONARY); Akzo Coatings,
Inc. v. Aigner Corp., 30 F.3d 761, 764 (7th Cir. 1994) (applying Restatement (Second) of Torts); United
States v. Colorado & E. R.R. Co., 50 F.3d at 153 (applying Am Jur.)
45. A broad reading of the statute is unacceptable because allowing PRPs to have standing
under section 107 would eviscerate section 113(g)(3), and PRPs would readily abandon a section 113
claim for a section 107 claim due to the significant procedural advantages. Consequently, section
113(g)(3) would become a nullity and section 113 would eventually be swallowed by section 107.
United Techs. Corp. v. Browning-Ferris Indus., 33 F.3d 96, 101 (1st Cir. 1994).
256
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
section 107's empowerment of “any other person” as a plaintiff entitled to
cost recovery.
The Ninth46 and Sixth Circuits47 ignored the “any other person”
language in section 107 as moot because they surmised that section 107 and
section 113 work in conjunction in contribution claims. Thus, a
contribution claim brought pursuant to section 107 and its “any other
person” language is transformed and limited by the mechanisms of section
113. This leaves only several liability to be in play. The First and Sixth
Circuits left open the question of which statute of limitations provision
applies if a PRP initiates a cleanup with government prodding.48 The
statute of limitation periods under section 107 and section 113 are different
by a factor of 100%.
All of these opinions would be shattered by the Supreme Court a
decade later, although, as discussed below, not all of the circuits have taken
obvious action to conform their precedents in the two years since.
2. Contradicting the Trial Courts
Of particular note is that many of these federal circuits had to overrule
their trial courts to arrive at these opinions. A series of more than a dozen
primarily trial court decisions found no legislative barrier to section 107
action by “any other person” private-plaintiff parties.49 These decisions,
46. Pinal Creek Group v. Newmount Mining Group., 118 F.3d 1298, 1301–02 (9th Cir. 1997).
The court also points to the legislative history behind section 113 to support the contention that the two
provisions work together. Id. at 1301.
47. See Centerior Serv. Co., 153 F.3d at 350 (The court agreed with the government’s
contention that “CERCLA does not provide two separate and distinct causes of action, but that the two
sections, § 107(a) and §113(f) work in conjunction.”). The Sixth Circuit added:
[P]arties seeking contribution under § 113(f) must look to § 107 to establish the
basis and elements of the liability of the defendants . . . . While a party seeking
contribution under § 113(f) may not recover under joint and several liability, it is
clear that under a plain reading of the statute, the party is seeking to recover its
“necessary costs of response” as referred to in § 107(a).
Id.
48. Id. at 355; United Techs. Corp., 33 F.3d at 99 n.8.
49. Amcast Indus. Corp. v. Detrex Corp., 2 F.3d 746, 748 (7th Cir. 1993); Amcast Indus. Corp.
v. Detrex Corp., 2 F. Supp. 545, 553 (N.D. Ind. 1992); United States v. Kramer, 757 F. Supp. 397, 414
(D.N.J. 1991); Barton Solvents, Inc. v. Sw. Petro-chem, Inc., No. 91-2382-GTV, 1993 WL 382047, at
*3 (D. Kan. Sept. 14, 1993); Charter Twp. of Oshtemo v. Am. Cyanamid Co., No. 1:92:CV:843, 1993
WL 561814 (W.D. Mich. Aug. 19, 1993); Companies for Fair Allocation v. Axil Corp., 853 F. Supp.
575, 579 (D. Conn. 1994); Town of Walkill v. Tesa Tape, Inc., 891 F. Supp. 955, 958–59 (S.D. N.Y.
1995); Allied Corp. v. Acme Solvents Reclaiming, Inc., 691 F. Supp. 1100, 1119 (N.D. Ill. 1988);
United States v. Atlas Minerals & Chemicals, Inc., No.91-5118, 1995 U.S. Dist. LEXIS 13097, at *216
(E.D. Pa. Aug. 22, 1995); Bethlehem Iron Works, Inc. v. Lewis Indus. Inc., 891 F. Supp. 221, 225 (E.D.
Pa. 1995); Chesapeake & Potomac Tel. Co. of Virginia v. Peck Iron & Metal Co., 814 F. Supp. 1269,
1277–78 (E.D. Va. 1992); United States v. SCA Serv. of Ind., 849 F. Supp. 1264, 1281 (N.D. Ind.
1994); Gen. Elec. Co. v. Litton Indus. Automation Sys. Inc., 920 F.2d 1415, 1418 (8th Cir. 1990);
2009] Cost Allocation Liability Conflicts Among the Federal Courts
257
beginning with the Sand Springs decision in 1987,50 were elevated to a
detailed, eloquent analysis in the Kramer opinion in 199151 and proliferated
through the mid-1990s. The wholesale contradiction of many of these trial
courts is notable.
These district courts generally based their conclusions on a plain
meaning review of CERCLA. One of these trial courts had the case that
would eventually get to the Supreme Court in 2007. In Adhesives Research
v. American Inks & Coatings, the court began its inquiry by dissecting the
plain language of the statute,52 which invests standing under section 107 to
“any other person” who incurs response costs. The court found each of
these terms to be unambiguous.53 The court stated that if a statutory term
was clear in meaning, a court should not alter the meaning of the term
solely because its makes the statute broad in scope.54 The term “any,”
although broad in scope, is clear and unambiguous; applying the plain
meaning standard of review ends judicial review of the statute.55 The
Adhesives Court held that the “any other person” language of section
107(a)(4)(B) confers standing on persons who incur response costs
regardless of their own potential liability. Following this logic to its end,
the court held that a plaintiff PRP has standing to bring a cost recovery
action under CERCLA section 107.
A number of district courts also agreed. The district court in Laidlaw
Waste Systems v. Mallinckrodt56 spurned its own circuit precedent and
likewise found that the plain language of section 107 and section 113 does
not deny plaintiff PRPs bringing claims pursuant to section 107.57 The
district court in Pinal Creek Group v. Newmont Mining Corp.58 explained
that section 107 confers standing upon any party who has incurred response
Transp. Leasing Co. v. California, 861 F. Supp. 931, 937–38 (C.D. Cal. 1993); Kelley v. Thomas
Solvent Co., 790 F. Supp. 710, 717 (W.D. Mich. 1990); Barmet Aluminum Corp. v. Doug Brantley &
Sons, Inc., 914 F. Supp. 159, 164 (W. D. Ky. 1995); Sand Springs Home v. Interplastic Corp., 670 F. S.
pp. 913, 916 (N.D. Okla. 1987).
50. See Sand Springs Home, 670 F. Supp. at 916 (imposing no bar to a section 107 cause of
action for potentially responsible parties).
51. See Kramer, 757 F. Supp. at 416–17 (holding that any temporary windfall to the private
plaintiff employing section 107 for cost recovery was justified by the incentives for voluntary private
clean up to foster the purpose of the statute).
52. Adhesives Research Inc. v. Am. Inks & Coatings Corp., 931 F. Supp. 1231, 1238 (M.D. Pa.
1996).
53. Id. at 1239.
54. Id.
55. Id. at 1238.
56. Id. at 1239; Laidlaw Waste Sys., Inc. v. Mallinckrodt, Inc., 925 F. Supp. 624 (E.D. Mo.
1996).
57. Laidlaw Waste Sys., Inc., 925 F. Supp. at 630.
58. Pinal Creek Group v. Newmont Min. Corp., 926 F. Supp. 1400 (D.Ariz. 1996).
258
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
costs,59 as the plain meaning of the statute does not provide a modifier that
should be applied to “any other person.”60
The Supreme Court in 2007 would unanimously confirm these district
court opinions regarding section 107, even though every federal appellate
court had reversed these district court opinions in the interim, creating
chaos in hazardous substance cost recovery.
3. Equitable Factors Supersede Plain Meaning
The desire of the federal appellate courts to broadly apply equitable
factors seems to have influenced these courts’ interpretation of section 107.
Many of the circuits (with the exception of the First, Fifth, and Twelfth
Circuits) held that there are no equitable defenses available against a
section 107 claim. In Veliscol Chemical Corp. v. Enenco, Inc., the court
disagreed with the district court’s allowance of the doctrine of laches as a
defense under section 107 because the only statutorily available defenses
under section 107 are those enumerated thereunder.61 In General Electric
Co. v. Litton Industrial Automation Systems, Inc., the appellants argued
unsuccessfully that the court should allow a section 107 “unclean hands”
defense.62 The court in State of California v. Neville Chemical Company,
held that the “three defenses to CERCLA liability expressly listed in
§ 107(b) are the only defenses available, and traditional equitable defenses
are not.”63 To have equitable discretion at its disposal, the appellate courts
gravitated to section 113 claims and did not sanction section 107 claims.
Some circuits required the PRPs to bring a claim for contribution under
section 113, but then due to equitable considerations applicable under
59. Id. at 1405.
60. Id.
61. Veliscol Chem. Co. v. Eneco, Inc., 9 F.3d 524, 530 (6th Cir. 1993); Town of Munster v.
Sherwin-Williams Co., 27 F. 3d 1268, 1270 (7th Cir. 1994).
62. Gen. Elec. Co. v. Litton Indus. Automation Sys., 920 F.2d 1414, 1418 (8th Cir. 1990). The
court held that “unclean hands” is not a defense to private party’s action to recover CERCLA response
costs under section 107. Specifically, the court stated that “CERCLA is a strict liability statute, with
only a limited number of statutorily-defined defenses available [under section 107(b)] . . . CERCLA
does not provide for an ‘unclean hands’ defense; the liability imposed by § 107(a) is subject only to the
defenses [in section 107(b)] . . . the purpose of allowing a private party to recover its response costs is to
encourage timely clean-up of hazardous waste sites. This purpose would be frustrated if a plaintiff’s
motives were subject to question.” Id.
63. See California v. Neville Chem. Co., 358 F.3d 661, 672 (9th Cir. 2004) (holding that
equitable considerations are not premised under section 107; however, they are considered under section
113); see also Blasland, Bouck & Lee, Inc. v. N. Miami, 283 F.3d 1286, 1304 (11th Cir. 2002) (holding
that CERCLA bars equitable defenses); Veliscol Chem. Co. v. Enenco, Inc., 9 F.3d 524, 530 (6th Cir.
1993) (discussing the reluctance of the court to permit non-enumerated equitable defenses); Town of
Munster, 27 F.3d at 1270. In resolving contribution claims, the court may allocate response costs among
liable parties using such equitable factors as the court determines are appropriate. Id.
2009] Cost Allocation Liability Conflicts Among the Federal Courts
259
section 113, allowed for 100% cost recovery rather than partial cost
contribution.64 This result is exactly what some courts feared would happen
if section 107 were enforced and the entire PRP liability was shifted from
plaintiff PRPs to defendant PRPs.
According to many of the federal circuit courts, the general rule under
CERCLA was that PRPs may bring only a claim for contribution under
section 113 and cannot recover under section 107 unless they are asserting
one of section 107(b)’s defenses.65 There is no statutory provision or
supportive legislative history to this effect. To get there, the circuit courts
ignored plain statutory language, their own decisions in analogous
CERCLA cases, and several canons of statutory construction.
In United Technologies Corporation v. Browning-Ferris Industries, the
First Circuit made mention in dicta of the possibility of allowing a PRP who
does the cleanup without prodding by the government to recover under
section 107. The court did not decide the issue in this case because the
parties began cleanup after governmental prodding. The 2007 Supreme
Court decision in Atlantic Research dismissed the interpretation of these
circuit courts and found that this flexibility is indeed available.
B. How Ocean’s Eleven Disrupted the Superfund
Here is the cost allocation and recovery scheme that the decisions of the
eleven federal circuits disturbed by disposing of section 107 into the ocean.
The preferred EPA enforcement approach to hazardous substance releases
under CERCLA is cleanup by private parties, either voluntarily or pursuant
to enforcement orders issued by the EPA.66 The EPA compiles lists of
priority sites and related PRPs to promote and maximize the number of
privately funded cleanups, marshalling the Superfund to finance those
priority cleanups for which no or inadequate private response activity
transpires.67
64. Dent v. Beazer Materials & Serv. v. Braswell Shipyards, Inc., 156 F. 3d 523, 530–31 (4th
Cir. 1998); PMC, Inc. v. Sherwin-Williams Co., 151 F.3d 610, 616 (7th Cir. 1998); W. Prop. Serv.
Corp. v. Shell Oil Co., 358 F.3d 678, 689–90 (9th Cir. 2004); Morrison Enter v. McShares, Inc., 302
F.3d 1127, 1135 (10th Cir. 2002); Bedford Affiliates v. Sills, 156 F.3d 416, 424 (2d Cir. 1998) (holding
that a PRP can never recover 100% of costs under section 107(a)).
65 42. U.S.C. § 9607(b) (2006).
66. Broward Gardens Tenants Ass’n v. U.S. E.P.A., 311 F.3d 1066 (11th Cir. 2002).
67. Cf. Fireman’s Fund Ins. Co. v. City of Lodi, Cal., 271 F.3d 911, 924–25 (9th Cir. 2001)
(noting that CERCLA requires the EPA to compile a national priorities list and that the EPA may use
funds form the Superfund to finance the cleanup of priority sites).
260
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
The allocation of liability under CERCLA usually involves two
stages.68 The first stage is the macro-level shift of remediation expenses
from the plaintiff(s) to the defendant(s). This stage involves determining
whether liability should be imposed on the entire group of liable PRPs
jointly and severally or should be divided severally or equitably. In the
Superfund context, this dimension triggers issues of joint and several
responsibility versus several contribution, allocation among categories of
responsible parties, and plaintiff’s choice of the cause of action and
designation of named defendants.
The second stage entails a micro-level reallocation of response costs
among the universe of liable defendant parties. This allocation may be
accomplished as part of the original imposition of joint and several liability,
or subsequently in reallocation of shares among PRPs.69 This second
dimension of allocation can divide those who choose to settle their liability
with the government from nonsettlers, juxtaposing the interests of settling
and nonsettling responsible parties.
Either to compel PRPs to clean up contamination or to recover its own
response costs, the government could bring suit under section 107 to shift
the liability for government-incurred cleanup costs on to the defendant
PRPs. This shifting or imposition of liability constitutes the first macrolevel stage of allocation. The defendant PRPs then must equitably allocate
this liability among themselves in the second micro-level stage of
allocation. PRPs who settle with the government should be able to recover
from other co-responsible parties those expended response costs under
section 107(a) or contribution to cleanup expenses under section 113(f) of
CERCLA.70 Under section 107(a), liability of PRPs in cost recovery
actions against PRPs is strict.71 Section 107 often shifts joint and several
liability to the defendants, unless a defendant can affirmatively demonstrate
that the harm is divisible.72 Joint and several liability, however, generally
68. However, some courts have merged these two stages together. See infra. Also, when harm
is divisible, there is no need for apportionment of harm that occurs in the second stage. See infra.
69. See 42 U.S.C. § 9613(f)(1) (2006) (“Any person may seek contribution from any other
person who is liable or potentially liable under section 9607(a) of this title, during or following any civil
action . . . under section 9607(a) of this title.”).
70. See id. § 9607(a) (establishing avenues for cost recovery); § 9613(f) (establishing
contribution requirements).
71. Centerior Serv. Co. v. Acme Scrap Iron & Metal Corp., 153 F.3d 344, 348 (6th Cir. 1998).
72. See, e.g., Acushnet Co. v. Mohasco Corp., 191 F.3d 69, 75 (1st Cir. 1999) (“[W]e embrace
the Restatement (Second) of Torts approach in construing [CERCLA], stating that a defendant may
avoid joint and several liability if the defendant demonstrates that the harm is divisible.”); Centerior
Serv. Co., 153 F.3d at 348 (noting that a plaintiff to a section 107 cost recovery action need only show
that “each defendant is a ‘liable’ party” and that damages will only be “apportioned according to fault”
if the defendant can “affirmatively demonstrate that the harm is divisible”).
2009] Cost Allocation Liability Conflicts Among the Federal Courts
261
has been the norm because of the difficulty imposed on a PRP to
affirmatively demonstrate the divisibility of the harm.73
There are situations where the government is willing to make
To the extent that
advantageous settlements with some parties.74
unrecovered costs remain, early settlement leaves both the government and
the settling parties free to initiate section 113 contribution or section 107
response costs actions against non-settling parties. There is a potential
advantage to the government because it gets an immediate settlement,
whether in the form of a cash settlement or a commitment to perform
response actions, or both. Whatever discount the government affords
settlers in litigation, thereby leaving it with unreimbursed response costs,
can be recouped pursuant to section 107 jointly and severally from any
single or multiple nonsettler(s). Where the government knows that not all
the parties will settle and that viable non-settling parties will remain, this
strategy may have advantages and few risks to the government.75
A settling PRP can also strategically choose which non-settling
defendants to sue if it is allowed access to CERCLA section 107. It is much
easier for a plaintiff to prove damages against a fewer number of
defendants; if section 107 is employed, only a few defendants need to be
named to shift liability to those defendants. This is much easier than
bearing the burden of proof severally for the contribution share against
every PRP.
Recovery under section 113 is more complicated. Since liability under
section 113(f) is not joint and several, but merely several, plaintiffs bear the
burden of proving the proportionate share of liability for each and every
73. Centerior Serv. Co., 153 F.3d at 348.
74. EPA Region I and the Commonwealth of Massachusetts, in settling the United States v.
Cannons Engineering Corp., 720 F. Supp. 1027, 1041 (D. Mass. 1989), aff’d, 899 F.2d 79 (1st Cir.
1990) case, involving four sites, employed this strategy. Most of the major non-de minimis parties
received an initial settlement at a discount of approximately 75% of their proportionate shares based on
waste-in volume. Id. Subsequently, individual governments sued several dozen additional nonsettling
parties, requiring them to make the governments whole. Id. at 1033. Small volume parties received de
minimis settlements crafted in several stages. Id. Also note that non-settling PRPs may not bring
contribution actions against settling de minimis parties. 42 U.S.C. § 9613(f)(2) (2006); Id. § 9622;
Avnet, Inc. v. Allied-Signal, Inc., 825 F. Supp. 1132, 1139 (D.R.I. 1992).
75. Of course, after the initial settlement the government may seek either to litigate against
nonsettlers or use the initial settlement as leverage to compel a subsequent settlement with the
nonsettling parties. While reducing the number of parties by an initial settlement, the government may
or may not have resolved all liability or maximized recovery. This partial settlement mode has the
advantage for the government of prompting a quick settlement with some liable parties, while leaving
additional nonsettling parties against whom unreimbursed present or future response costs can be
incurred. In certain applications, this is an ideal solution in that it operates much like an insurance
policy, while expediting some initial settlement.
262
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
defendant severally.76 Section 113 allows an equitable reallocation of total
costs incurred severally among PRPs. Pursuant to this avenue, the liability
is divided among the PRPs according to their proven equitable
proportionate shares.77 This follows the general common law rule of
contribution that all joint tortfeasors must contribute equally to satisfy a
collective burden.78 A primary difference between section 107 and section
113 is that district courts are afforded great discretion in allocating liability
on an equitable basis under section 113(f)(1).79 The difference between
several liability and joint and several liability is profound.
Much of the circuit court imbroglio over section 107, and resulting
chaos after the 2004 Supreme Court decision in Aviall, was justified by
circuit courts that claimed that only by finding an implied congressional
intent for use of “equitable factors” under section 113 did distinctions make
sense. However, there is no express statutory prohibition against equitable
considerations applied to claims adjudicated under section 107.80
Not only do these two routes of section 107 and section 113 yield
potentially different reallocations, but they yield distinct outcomes
depending on whether or not the plaintiff previously has settled its liability
with the EPA. To encourage settlement and reduce litigation costs,
Congress provided contribution protection to all settling PRPs pursuant to
section 113(f) under section 107 or section 113.81 A PRP which has settled
with the government in a judicially or administratively approved settlement
is protected from additional liability from both other private PRPs and the
government for matters which are covered in the settlement.82 Contribution
76. See also § 9613(f) (dealing with contribution claims).
77. Id. (stating that courts should determine the equitable share on a case-by-case basis).
78. McDonald v. Magruder, 28 U.S. (1 Pet.) 470, 477 (1830); Adamson v. McKeon, 225 N.W.
414, 416 (1929); Easterly v. Barber, 66 N.Y. 433, 440 (1876), available at 1876 WL 12250.
Contribution is inherently equitable, and courts are free to fashion equitable remedies. Yates v.
Donaldson, 5 Md. 389, 394–95 (1854).
79. United States v. R.W. Meyer, Inc., 932 F.2d 568, 571–73 (6th Cir. 1991). Note that this
case is a subsequent opinion in United States v. Northernaire Plating Co., 670 F. Supp. 742, 748 (W.D.
Mich. 1987) (holding defendants jointly and severally liable under section 107(a)), aff’d sub nom.,
United States v. R.W. Meyer Inc., 889 F.2d 1497 (6th Cir. 1989).
80. See United States v. Hardage, 116 F.R.D. 460, 465 (W.D. Okla. 1987) (allowing equitable
factors raised as a defense to bar government’s section 107 claims). However, this opinion is not
followed by many other courts.
81. § 9613(f).
82. § 9613(f)(2). See generally STEVEN FERREY, ENVIRONMENTAL LAW: EXAMPLES AND
EXPLANATIONS (4th ed. 2007) (Typically settlements of CERCLA liability by private parties involve
both the EPA and state government. In the author’s experience, the interests of the state and federal
government can be quite distinct. The federal government incurs 90% of government capital response
costs, while the state government typically incurs the remaining 10% plus ongoing obligations for
operations and maintenance. Therefore, a state government may be particularly attuned to long-term
risks and costs associated with operating site O&M systems).
2009] Cost Allocation Liability Conflicts Among the Federal Courts
263
protection is effective as soon as a settlement is signed and it is not
dependent on the fulfillment of any duties undertaken by the settler.83 A
settlement with the government should confer absolute protection against
counterclaims by non-settling defendants.84 In order to ensure this
protection, however, the settling PRP must settle via a consent decree, as
opposed to responding to a unilateral administrative order from the EPA
issued pursuant to section 106.85 This distinction is critical. Whether a
state government settlement can trigger this federal contribution protection
emerged as a critical issue after the Supreme Court’s Aviall opinion, as
discussed below.
A PRP or group of PRPs can settle for the entire response cost, their
proportionate share, or even less than their proportionate share. If the
federal appellate courts had correctly interpreted the statute as ultimately
determined by the Supreme Court, the settling PRP can then pursue nonsettling PRPs for the response costs incurred under the settlement via either
of two alternative allocation avenues. First, pursuant to section 107, the
settling party typically seeks restitution of response costs under a theory of
joint and several liability.86 In the alternative, pursuant to section 113, the
settling party seeks contribution of liability under a several liability theory.87
The ultimate final allocation among all liable parties is only clear after a
subsequent cost recovery or contribution action lodged by the settling PRP
against other non-settling PRPs.88
There are significant differences between section 107 and section 113.
Key advantages of section 107 are the application of joint and several
liability, a doubly long statute of limitations period in which to initiate suit,
83. See Dravo Corp. v. Zuber, 13 F.3d 1222, 1225–26 (8th Cir. 1994) (holding that, pursuant to
section 122(a), de minimis settlers receive automatic and instantaneous contribution protection subject to
a condition subsequent to fulfill duties).
84. § 9613(f).
85. 42 U.S.C. § 9606 (2006). While there are many differences between the EPA’s model
consent decree and the EPA’s model unilateral administrative order, a critical distinction is in the
provision of contribution protection to settlers. The model consent decree utilized by the EPA contains
an optional paragraph pertaining to the contribution protection contained in section 113(f)(2). See
Dravo Corp., 13 F.3d at 1227–28 (holding that contribution protection applies to administrative
settlement as well as to consent decrees). The Dravo court held that it lacked jurisdiction to secondguess the EPA on administrative settlements. While nonparticipating settlers can object during the
standard thirty-day public comment period, the court found that there was no other recourse for a third
party to challenge a settlement, even where that party would be prevented from seeking contribution
against the settling party. Effectively, this removes judicial review of administrative de minimis
settlements. Id.
86. 42 U.S.C. § 9607(a)(4)(B) (2006). See E.I. Dupont de Numours v. United States, 460 F.3d
515, 522 (3d Cir. 2006) (Section “107 imposes strict, joint, and several liability on all PRPs.”).
87. § 9613(f)(1).
88. § 9613(f).
264
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
the necessity only to name and prosecute a few and not all of the liable
parties, and traditionally the unavailability of equitable defenses to
defendants beyond the statutorily prescribed defenses. Section 107 is less
likely to result in the plaintiff absorbing “orphan shares” of unfunded PRP
liability.89
If the plaintiff chooses to use a section 113 claim as a “sword,” the
settler can theoretically recover severally from nonsettlers for their
equitable shares of the incurred remediation costs. However, the plaintiff’s
burden to demonstrate several liability of each and every individual and
potential defendant is formidable90 and diminishes the probability of a full
recovery. Comparatively, the settler can settle for any amount—more or
less than the settler’s proportionate share—and then initiate litigation
against some non-settling PRPs91 under section 107 to shift the settler’s
direct response costs under joint and several liability principles.
Under this relatively detailed statutory scheme, the eleven federal
circuit courts walled off access to section 107 by PRPs acting as plaintiffs.
This fundamentally killed a lone PRP’s incentive to voluntarily undertake
remediation expenditures at multi-party sites, as it could get stuck with the
responsibilities of non-responding PRPs. It would turn out that all of these
circuit courts were wrong at every turn.
II. THE SECOND STEP: THE 2004 AVIALL SUPREME COURT DECISION
In its decision in Cooper Industries, Inc. v. Aviall Services, Inc.,92 the
Supreme Court went around the barrier created by the federal circuits
between 1993 and 2004.93 This chaos-causing decision allowed for billions
of dollars more in hazardous waste liability allocations than previously
allowed by most of the circuit courts. The Supreme Court opinion in Aviall
prohibits a private party from initiating a claim under section 113(f)(1)
against other PRPs for contribution to hazardous waste cleanup expenses,
unless and until that plaintiff party itself has been sued first for response
89. See FERREY, supra note 82, at 401–11 (providing a detailed example of orphan-share
allocation under section 107 and section 113).
90. In a section 113 action, a plaintiff must shoulder the burden of proving the liability of every
PRP defendant and successfully rebut all defenses raised by the defendants. Additionally, discovery
must be conducted against each defendant. For each defendant who is not brought into the litigation or
is not successfully prosecuted, a piece of the amount necessary for the plaintiff’s full recovery is lost.
The parts may not equal the whole under section 113.
91. Compared to a section 113 action, only a few defendants need be named, and only one
needs to be successfully prosecuted to shift liability jointly and severally.
92. Cooper Indus. Inc. v. Aviall Servs. Inc., 543 U.S. 157 (2004).
93. See supra Section II.
2009] Cost Allocation Liability Conflicts Among the Federal Courts
265
costs by (or settled with) the government under section 107(a) or section
106.
With the eleven circuits previously having blocked use of section 107,
this decision of the Supreme Court blocked the other avenue of section 113
in many instances. The federal government only sues PRPs in those few
situations resulting from very high-profile waste sites. The existing system
relies substantially on voluntary private action to remediate hazardous
waste contamination and thereafter private judicial proceedings to
reallocate the cleanup cost from the volunteering party to others PRPs who
are also liable for or contributed to the contamination.94 With the costreallocation mechanism of both section 107 and section 113 judicially
disabled, the incentives for voluntary clean up of hazardous waste sites
disappear because cost recovery for the volunteering party becomes
difficult or impossible.
The facts in Aviall are straightforward.95 The federal district court
found for Cooper, barring Aviall’s section 113 claim on the basis that it had
not been brought during or after a government action against Aviall nor
subject to an approved settlement.96 The Fifth Circuit initially upheld the
district court’s ruling,97 but then reversed it in an en banc hearing.98 The en
banc decision relied on the purpose of CERCLA “to promote prompt and
effective cleanup.”99 The Fifth Circuit en banc found that the statute’s
savings clause was not limited to state claims and that a private section 113
action by Aviall was not dependent upon a prior or pending action.
On certiorari, the odds on paper were stacked: twenty-three states
joined Aviall as amici, as did numerous corporations and others, to argue in
support of the final Fifth Circuit en banc decision allowing unfettered use of
section 113. The United States Government filed the sole amici brief
94. See FERREY, supra note 82, at 407–09 (discussing cost reallocation among private parties).
95. Aviall purchased four aircraft manufacturing and maintenance facilities from Cooper
Industries. Aviall, 543 U.S. at 163–64. Hazardous substance contamination at the sites was created by
both companies’ operations before and after the sale. Id. Aviall, the then-current owner, remediated the
site under direction of the Texas environmental agency at a cost of almost $5 million, as a prerequisite
to its sale of the property to a third party, whereupon it sought section 113 contribution and/or section
107 CERCLA response cost recovery from Cooper, the former owner. Id. at 164. These two claims
were later amended to consolidate and launch just a single section 113 claim, as per the controlling
precedent in the circuit. Id. Aviall performed the remediation voluntarily, having never been sued by
any environmental enforcement agency, nor had Aviall ever entered a formal judicially or
administratively approved settlement with any environmental agency. Id.
96. Id. at 164–65.
97. Aviall Servs. Inc. v. Cooper Indus. Inc., 263 F.3d 134 (5th Cir. 2001).
98. Aviall Servs. Inc. v. Cooper Indus. Inc., 312 F.3d 677 (5th Cir. 2002).
99. Id. at 681.
266
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
supporting Cooper Industries.100 The numbers of parties here did not
presage the outcome. The Supreme Court reversed the Circuit in a forceful
opinion requiring as a prerequisite to any section 113(f) CERCLA
contribution action that the private party plaintiff has either (1) suffered
prosecution for liability from the federal government or, (2) entered a
judicially or administratively approved settlement of such dispute with the
government.101 The Court maintained the distinction between section 107
“cost recovery” and section 113 “contribution actions”: “[a]fter SARA,
CERCLA provided for a right to cost recovery in certain circumstances, §
107 (a), and separate rights to contribution in other circumstances, § 113
(f)(1), § 113(f)(3)(B).”102
It required no leaps of judicial logic for the Supreme Court to reach this
outcome. It followed the plain meaning of the exact language of section
113 of the statute, finding that the authorization to initiate a contribution
action after or during such other litigation or settlement is the only means,
not an illustration of one of a host of means, to entitle one to bring
contribution claims against other potentially liable parties under the statute.
Otherwise, a more permissive interpretation of the section would “render . .
. entirely superfluous” the conditional “during or following” language of
the Act.103 This interpretation, in the view of the Court majority, gave every
word of the statute meaning—a plain meaning interpretation following the
canons of statutory construction.
Of particular note, while the Supreme Court decision in Aviall was by a
seven to two plurality, the two dissenters did not seem to contest this core
holding. Rather they sought to go further to address in the opinion the even
more pressing issue of whether there was a private right to cost recovery
along the alternative road of section 107(a), notwithstanding the prior
opinions of eleven circuit courts. However, because Aviall had been forced
in the district court to drop its alternative section 107 claim after the
initiation of the litigation, thereby consolidating all claims under the less
conducive section 113 recognized by the Fifth Circuit, the issue was not
addressed in the circuit court opinion on review, and therefore formally was
100. Brief for United States as Amicus Curiae Supporting Petitioner, Cooper Indus. Inc., v.
Aviall Servs. Inc., 543 U.S. 157 (2004) (No. 02–1192).
101. Aviall, 543 U.S. at 166.
102. Id. at 163.
103. Id. at 166. Justice Thomas, in his opinion for the majority, rejected the notion that “may”
should be read permissively so the “during or following” statutory language was one of several
mechanisms to utilize section 113. Id. Rather, the opinion held that only during or after one of the
statutorily specified requisites could a party as plaintiff initiate a section 113 civil contribution action.
Id. Therefore, the “may” language was read as exclusive rather than inclusive of the only statutorily
authorized means to utilize the section 113 path for private cost contribution to waste remediation. Id.
2009] Cost Allocation Liability Conflicts Among the Federal Courts
267
not before the Court on certiorari.104 The Supreme Court majority in Aviall
would not adjudge the section 107 rights of private party litigants sua
sponte.105 In dissent, Justice Ginsburg, joined by Justice Stevens, would
have ruled on the section 107 issue notwithstanding it not being briefed to
the Court, potentially overturning the impenetrable section 107 roadblocks
erected by the circuits.106
This 2004 Supreme Court decision cut off section 113 as an effective
cost reallocation route. Four of the federal circuits which had previously
blocked any use of the alternative section 107 cost recovery route by PRPs
debated then whether section 107 had been revitalized in the shadows of the
Supreme Court’s Aviall decision, or whether the impenetrable wall of the
eleven circuits remained un-breached. 107 Three of these four courts decided
to reopen access to section 107, which they had previously blocked, with
one stating that if the Supreme Court did not compel reopening section 107,
its prior precedent stood un-impeached. This provided the Supreme Court a
second chance to reach the decision it could not stretch to reach in 2004.
III. THE THIRD STEP: THE 2007 ATLANTIC RESEARCH DETONATION
A. U-Turns in the Second and Eighth Circuits
Immediately after the 2004 Supreme Court opinion in Aviall, the
Second Circuit did an immediate U-turn. Despite its 1998 Bedford decision
denying all PRPs access to section 107 in order not to render section 113 “a
nullity,” the Second Circuit had second thoughts in Consolidated Edison
104. Id. at 170.
105. Id.
106. Id. at 174 (Ginsburg, J., dissenting). Justice Ginsburg’s dissent notes that in Key Tronic,
the Supreme Court in dicta stated that section 107 “unquestionably provides a cause of action for
[PRPs].” Id. at 172. Justices Scalia and Thomas, dissenting in Key Tronic, but in the majority in Aviall,
also favored a private right of action under section 107. Justice Ginsberg notes that in KeyTronic “no
Justice expressed the slightest doubt that section 107 indeed enables a PRP to sue other covered persons
for reimbursement . . . of cleanup costs.” Id. at 172.
107. Three circuits reinvigorated section 107 after the Aviall decision. See Consol. Edison Co.
of New York v. UGI Util. Inc., 423 F.3d 90, 97 (2d Cir. 2005) (finding that the cost of recovery may be
pursued under section 107(a) because costs to clean up sites are “costs of response” under the statute);
see also Metro. Water Reclamation Dist. of Greater Chi. v. N. Am. Galvanizing & Coatings, Inc., 473
F.3d 824, 830–32 (7th Cir. 2007); Atl. Research Corp. v. United States, 459 F.3d 827, 834–35 (8th Cir.
2006), aff’d, 551 U.S. 128 (2007). The Third Circuit held to its prior precedent. E.I. du Pont de
Nemours & Co. v. United States, 460 F.3d 515, 531 (3d Cir. 2006).
268
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
Co. v. UGI Utilities, Inc.108 There, the Second Circuit noticed the “plain
language” of section 107 and, fearing that it would “impermissibly
discourage voluntary clean-up”109 and that all voluntary cleanups would
grind to a halt, stated that “[t]his would undercut one of CERCLA’s main
goals, ‘encouraging private parties to assume the financial responsibility of
cleanup by allowing them to seek recovery from others.’”110
To make this U-turn, the Second Circuit pivoted off the 1994 CERCLA
decision of the U.S. Supreme Court in Key Tronic Corp.111 Regarding its
prior contrary precedent in Bedford, the circuit court stated that “[t]his
holding impels us to conclude that [Bedford] no longer makes sense.”112
However, rather than formally overrule Bedford, the Second Circuit tried to
distinguish it by noting that Consolidated Edison had not been sued or
found partially liable, unlike in Bedford. However, here the logic stops.
Without settling one’s liability in an administratively or judicially approved
settlement with the federal government there is no contribution protection
under section 113(f)(2) and this minimizes incentives for voluntary
remediation. Even with section 107 available, without contribution
protection against cross-claims there is scant incentive for voluntary
remediation when more litigation will result in claims against the settler.
The Consolidated Edison Court viewed section 107(a) and section
113(f) as operating in tandem because “[e]ach of those sections . . .
embodies a mechanism for cost recovery available to persons in different
procedural circumstances.”113 The court’s holding implied that a section
113(f)(1) claim for contribution will be permitted for PRPs who are subject
108. Consol. Edison Co. 423 F.3d at 96–97. The plaintiff entered into a Voluntary Cleanup
Agreement with the state of New York. The plaintiff alleged that under that agreement it had resolved
its liability to the state, but the court held that the resolution of liability must pertain to liability of claims
under CERCLA. The court noted that while the plaintiff may have resolved its liability to the State of
New York for claims arising under the state’s environmental laws, the agreement contained a
“Reservation of Rights” whereby the state reserved its right to bring CERCLA claims against the
plaintiff. Because the state reserved a right to bring future CERCLA claims against the plaintiff, the
court held that the plaintiff had not resolved its CERCLA liability to the state and therefore it could not
bring a contribution claim under section 113(f)(3). Id.
109. Id. at 100.
110. Id.
111. Key Tronic Corp. v. United States, 511 U.S. 809 (1994).
112. Consol. Edison Co., 423 F.3d at 99.
113. Id. The court found that “section 107(a) permits a party that has not been sued or made to
participate in an administrative proceeding, but that, if sued, would be held liable under section 107(a),
to recover necessary response costs incurred voluntarily, not under a court or administrative order or
judgment.” Id. at 100. The court authorized PRPs to bring claims for contribution under section 107(a)
or section 113(f), depending on their “procedural circumstances,” but precludes those PRPs who are
subject to an administrative proceeding from bringing a claim for contribution under section 107(a). Id.
at 99.
2009] Cost Allocation Liability Conflicts Among the Federal Courts
269
to an administrative proceeding, even if they were not subject to a judicial
proceeding. The Supreme Court stated in Aviall “that [one] must, if
possible, construe a statute to give every word some operative effect.”114
Section 107(a) does not exclude liable parties from raising a section 107
claim, and thus section 107(a) is available to “any other person” who has
incurred “any other necessary costs of response.”115
In a similar U-turn, the Eighth Circuit provided an interesting
mechanism for the circuits to side-step their prior precedent blocking the
use of section 107. After Aviall, it decided that a PRP plaintiff may avail
itself of a section 107 cost recovery action.116 To overcome its prior
contrary decision in Dico, above, the Circuit did not attempt to reverse this
precedent, but rather indicated that a different three-judge panel can depart
from a prior panel’s decision where its rationale has been undermined.117
The Circuit found that once free of this prior decision, the Aviall decision
requires that section 107 and section 113 be regarded as separate and
distinct avenues both accessible to PRP plaintiffs. In section 107, “any
other person” includes any parties other than governments or Indian Tribes
which are previously expressly included.118
The Eighth Circuit, in this 2006 reversal of past decisions, found no
congressional intent to have silently repealed just the “any other person”
provision of the preexisting section 107 when it enacted section 113(f) in
1986.119 The court proceeded to limit the amount of response costs owed by
a PRP who has neither been sued nor settled its liability to only its fair share
costs under the “any other necessary costs of response” provision of section
107.120 This proportionate concept removes one of the few logical
criticisms of section 107’s shifting of private costs: that it could allow a
temporary windfall.121 The Circuit goes even further to read into section
107(c) an implied right to contribution so as not to penalize parties who
114. Cooper Indus. Inc. v. Aviall Servs. Inc., 543 U.S. 157, 167 (2004).
115. Consol. Edison Co., 423 F.2d at 94. The Second Circuit in Consolidated Edison did not
seem to recognize that a PRP counterclaim in response to a private plaintiff section 107 action could be
barred by section 113 counterclaim protection if the plaintiff had previously entered an approval
settlement with the EPA. Id. at 100 n.9. See discussion infra Section V.B.
116. Atl. Research Corp. v. United States, 459 F.3d 827 (8th Cir. 2006).
117. Id.
118. Id. at 835. See 42 U.S.C. § 9607(a)(4)(A)–(B) (2006) (PRPs shall be liable for “all costs
. . . incurred by the United States Government or a State or an Indian tribe” and “any other necessary
costs . . . incurred by any other person”).
119. Atl. Research Corp., 459 F.3d at 836.
120. Id. at 835.
121. See United States v. Kramer, 757 F. Supp. 397, 416–17 (D.N.J. 1991) (acknowledging that
allowing a section 107 can create a temporary windfall).
270
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
voluntarily remediate sites.122 Therefore, this opens up to all PRPs—even
those who have not settled with the government, and thus have no recourse
to section 113 action after the Supreme Court’s decision in Aviall—to utilize
section 107 for cost recovery of their direct costs.
Notwithstanding these two quick U-turns in these even-numbered
circuits, the reaction was not uniform. The Third Circuit remained fixed: if
the Supreme Court in Aviall did not overrule the circuits, then the circuits
should continue to deny PRP access to section 107.123 Of course, while a
minority of Justices in Aviall wished to address whether the circuit blockage
of section 107 was incorrect, the Court could not consider this because the
section 107 issue was neither before the Court nor briefed in Aviall.
On remand, the district court in the Fifth Circuit reexamining the
controversy similarly refused to recognize the availability of the section 107
path.124 Over the years, different panels of the Fifth Circuit had suggested,
but never squarely held, that a private PRP could utilize section 107 for cost
recovery.125 Holding that the Fifth Circuit had not squarely addressed the
issue, the district court took license to interpret “any other person” in
section 107(a) to apply only to innocent parties and not PRPs. The district
court’s ultimate defense essentially was that the Supreme Court in its Aviall
decision did not command that the circuit precedent on section 107 must
yield.126 Their seeming rationale was that allowing access to Section 107
would negate the contribution protection of section 113(f)(2).127 The court
concluded that section 107(a) also cannot be used for seeking contribution,
so that parties prior to a judicially approved settlement with the federal
government can not recover costs under any federal scheme, as both section
107 and section 113 are precluded.128
Some other courts followed both the Second and Eighth Circuit access
to section 107,129 while other courts followed the Third Circuit and Fifth
122. Atl. Research Corp., 459 F.3d at 836.
123. E.I. du Pont de Nemours & Co. v. United States, 460 F.3d 515, 544–45 (3d Cir. 2006).
124. Aviall Servs., Inc. v. Cooper Industries, LLC, No. 3:97–CV–1926–D, 2006 WL 2263305,
at *10 (N.D. Tex. Aug. 8, 2006). On remand, the Firth Circuit, en banc, ordered the district court to
permit Aviall to amend its complaint to reassert its original section 107 claim.
125. See Amoco Oil Co. v. Borden, Inc., 889 F.2d 664, 672 (5th Cir. 1989) (stating that a
private party may recover only those response costs consistent with the national contingency plan);
Tanglewood E. Homeowners v. Charles-Thomas, Inc., 849 F.2d 1568, 1575 (5th Cir. 1988) (holding that
government involvement is not a prerequisite for a PRP cost-recovery claim under section 107).
126. See Tanglewood E. Homeowners, 849 F.2d at 1568; Dico, Inc. v. Amoco Oil Co., 340 F. 3d
525, 529 (8th Cir. 2003).
127. See Tanglewood E. Homeowners, 849 F.2d at 1568.
128. See id.
129. See City of Bangor v. Citizens Commc’ns Co., 437 F. Supp. 2d 180, 222 (D. Me. 2006)
(declining to interpret Aviall as stripping PRPs of their rights to section 107 claims); see also Raytheon
2009] Cost Allocation Liability Conflicts Among the Federal Courts
271
Circuit denials of access to section 107 cost recovery.130 The result was
total chaos.
B. The New Path of Atlantic Research
1. A New Direction
The even-numbered circuits would eventually be affirmed by the
Supreme Court in 2007. In United States v. Atlantic Research Corp.,
decided in mid-2007, the Supreme Court affirmed a 2006 post-Aviall
opinion of the Eighth Circuit131 that opened up private PRP access to
section 107 cost recovery. In Atlantic Research, the United States had
Aircraft Co. v. United States, 435 F. Supp. 2d 1136 (D. Kan. 2006) (holding that a PRP has an implied
right to contribution under section 107); McDonald v Sun Oil Co., 423 F. Supp. 2d 1114, 1133 (D. Or.
2006) (following Ninth Circuit precedent that section 107 “continues to exist as a viable cause of action”
for PRPs); Aggio v. Aggio, No. C 04–4357 PJH, 2005 WL 2277037, at *5, 6 (N.D. Cal. Sept. 19, 2005)
(holding that in the Ninth Circuit PRPs continue to have rights to claims under section 107); Ferguson v.
Arcata Redwood Co., No. C 03–05632 SI, 2005 WL 1869445, at *6 (N.D. Cal. Aug. 5, 2005)
(recognizing that the Ninth Circuit allows PRPs to bring contribution claims under section 107);
Viacom, Inc. v. United States, 404 F. Supp. 2d 3, 7 (D.D.C. 2005) (noting the change in the circuit courts
and allowing PRPs to bring a section 107(a) action); Kotrous v. Goss-Jewett Co., No. Civ. S02–1520
FCD JFM, 2005 WL 1417152, at *3 (E.D. Cal. 2005) (acknowledging controlling Ninth Circuit
precedent and allowing PRPs to file contribution claims under section 107(a)); Metro. Water
Reclamation Dist. of Greater Chi. v. Lake River Corp., 365 F. Supp. 2d 913, 917–18 (N.D. Ill. 2005)
(allowing PRPs to sue under section 107(a)); Vine Street L.L.C. v. Keeling, 362 F. Supp. 2d 754, 764
(E.D. Tex. 2005) (allowing Vine Street to state a claim for cost recovery under section 107(a)).
130. See Spectrum Int’l Holding, Inc. v. Universal Coops., Inc. No. 04–99, 2006 WL 2033377,
at *5 (D. Minn. July 17, 2006) (noting the Eighth Circuit precedent and holding that a PRP may not
bring a section 107(a) claim); see also Adobe Lumber, Inc. v. Hellman, 415 F. Supp. 2d 1070, 1076
(E.D. Cal. 2006) (“[A] PRP does not have a cost recovery action, and instead has only a contribution
action . . . .”) (emphasis in original); R.E. Goodson Constr. Co. v. Int’l Paper Co., No. C/A 4:02-4184RBH, 2005 WL 2614927, at *29 (D.S.C. Oct. 13, 2005) (dismissing plaintiff’s section 113 claim for
lack of civil action under section 107(a) against the plaintiffs); Montville Twp. v. Woodmont Builders,
No. Civ.A. 03-2680DRD, 2005 WL 2000204, at *13 (D.N.J. Aug. 17, 2005) (precluding a PRP from
obtaining recovery costs from another PRP under section 107(a)); City of Rialto v. U.S. Dep’t of Def.,
No. EDCV 04-00079-VAP (SSx), 2005 U.S. Dist. LEXIS 26941 (C.D. Cal. Aug. 15, 2005) (finding that
PRPs may only pursue contribution claims under the combined effect of section 107(a) and section
113(f)); Boarhead Farm Agreement Group v. Advanced Envtl. Tech. Corp., 381 F. Supp. 2d 427, 435
(E.D. Pa. 2005) (finding that the law in the circuit states that section 113 is the only avenue for
contribution claims by PRPs); Blue Tee Corp. v. ASARCO, Inc., No. 03-5011-CV-SW-F-JG, 2005 WL
1532955, at *6 (W.D. Mo. June 27, 2005) (noting that PRPs cannot seek contribution against other PRPs
under section 107); Atl. Research Corp. v United States, No. 02-CV-1199, 2005 U.S. Dist. LEXIS 20484
(W.D. Ark. June 1, 2005) (barring section 107 cost recovery); City of Waukesha v. Viacom Int’l Inc.,
362 F. Supp. 2d 1025, 1027–28 (E.D. Wis. 2005) (stating that to allow a section 107 claim would be
“futile”); Mercury Mall Assocs. v. Nick’s Mkt., Inc., 368 F. Supp. 2d 513, 520 (E.D. Va. 2005) (denying
a section 107 recovery claim).
131. Atl. Research Corp. v. United States, 459 F.3d 827 (8th Cir. 2006), aff’d, 551 U.S. 128
(2007).
272
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
argued that “any other person” in section 107(b) referred to parties other
than the four groups of liable PRPs identified in section 107(a)(1)–(4).132
The Court dismissed this argument as making “little textual sense.”133 The
Court relied on a plain language interpretation in holding against the
government’s strained interpretation of section 107 of Superfund. The
Court acknowledged that persons in different procedural positions have
access to either section 107 or the recently limited section 113,134 as the
situation merits. In its articulation of “plain meaning,” the Court noted that
“[t]he provisions are adjacent and have remarkably similar structures. Each
concerns certain costs that have been incurred by certain entities and that
bear a specified relationship to the national contingency plan.”135
The Supreme Court opined on issues that every circuit decision
between 1993 and 2003 obscured or missed: section 113 is available for
equitable apportionment of costs (including those not directly incurred)
among jointly liable parties. Section 107, by contrast and pursuant to plain
meaning, is freely available for any person or party to utilize to recover its
own actually expended costs of response to hazardous substance
remediation. Sections 107 and 113 are complementary avenues, not the
excluded avenues determined by eleven circuits.
Despite the section 113(f)(2) contribution protection afforded settling
parties, the Court assumed, without directly ruling, that plaintiffs utilizing
section 107 would not be immune from counterclaims pursuant to section
113, which would cause a court to equitably apportion de novo the total
cost burden among co-liable litigants. In Atlantic Research, the Supreme
Court observed that an equitable allocation of response costs could be
achieved by bringing “a Section 113(f) counterclaim.”136 The Court noted
that the section 113(f)(2) “settlement bar does not by its terms protect
against cost-recovery liability under Section 107(a).”137 The Supreme Court
notes that “a defendant PRP in such a Section 107(a) suit could blunt any
inequitable distribution of costs by filing a Section 113(f) counterclaim.”138
There was less protection against the very section 107 cost-recovery claims
that the Court had liberated from the blockage of eleven circuit courts. The
Court did not seem to bar any section 113(f) counterclaims by defendants,
and noted that they would be protected from reimbursement claims by their
132. United States v. Atl. Research Corp., 551 U.S. 128, 134 (2007).
133. Id. at 136.
134. See Cooper Indus. Inc. v. Aviall Servs. Inc., 543 U.S. 157 (2004) (limiting access to section
113 where suit and settlement had not preceded action).
135. Atl. Research Corp., 551 U.S. at 135.
136. Id. at 140.
137. Id.
138. Id. at 140.
2009] Cost Allocation Liability Conflicts Among the Federal Courts
273
own prior settlement with the government, if not from section 107 cost
recovery.139 Settling parties are now opened up to section 107 claims,
which are distinguished from contribution claims and not limited by any
contribution protection under section 113(f)(2).
This dicta by the Supreme Court resolves the intriguing question
originally posed in Kramer, which is whether a section 107 plaintiff could
recover a windfall from shifting an inordinate share of its own equitable
share of costs to defendants. In the view of the Supreme Court, a plaintiff
could not recover a windfall. Section 107 could only be used by a party
who actually incurred response costs itself, rather than by a party who
reimbursed the costs incurred by another. “The remedies available in
Sections 107(a) and 113(f) complement each other by providing causes of
action ‘to persons in different procedural circumstances.’”140
“The choice of remedies simply does not exist” for a party.141 The
Court noted that voluntarily incurred costs can only be allocated by
recourse to section 107.142 However, the line between what is and is not
voluntary is not precisely defined. Administrative orders from, and consent
decrees with, the EPA, which are a common means of resolving alleged
responsibility at Superfund sites, are not specified as to which such group
they fall into but are left to the lower courts to determine. Prior to 2000,
those federal district courts disagreed with each other, and even with their
circuit courts, on section 107 and will do so again. The Supreme Court
clarified that “a PRP that pays money to satisfy a settlement agreement or a
court judgment may pursue § 113(f) contribution. But by reimbursing
response costs paid by other parties, the PRP has not incurred its own costs
of response and therefore cannot recover under § 107(a).”143
Moreover, the contribution protection of section 113(f) was held to only
protect against contribution actions under section 113, not section 107(a)
cost recovery actions.144
2. A New Canon Is Fired
Building on the rationale of the 2007 global warming decision of the
Supreme Court,145 the Atlantic Research decision establishes the
interpretive rule of “plain meaning” construction of federal statutes enacted
139.
140.
141.
142.
143.
144.
145.
See id. at 139–40 (discussing various section 113(f) counterclaims).
Id. at 139.
Id. at 140.
Id. at 140 n.6.
Id. at 139.
Id. at 140.
Massachusetts v. EPA, 549 U.S. 497 (2007).
274
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
by the legislature.146 This canon asserts that the actual words of the statute
are “the most important evidence of its meaning,”147 “the final expression of
the meaning intended,”148 and “the most authoritative interpretive
criterion.”149 Under the plain meaning rule, the language used in the statute
can be objectively determined without recourse to, for example, legislative
This canon restricts statutory interpretation to those
history.150
circumstances unambiguously addressed in the legislative process, as
evidenced by the specific terms of the law.151
This articulation was building in recent Supreme Court decisions. In
2005, the Supreme Court announced that it should interpret statutes by a
reading “that makes sense of each phrase” and “the one favored by our
canons of interpretation.”152 The Supreme Court also has held that it is the
duty of the courts to regard separate statutes—or in this case, by analogy,
separate provisions of the same CERCLA statute—as each fully effective:
“Judges ‘are not at liberty to pick and choose among congressional
enactments, and when two [or more] statutes are capable of co-existence, it
is the duty of the courts, absent a clearly expressed congressional intention
to the contrary, to regard each as effective.’”153
In 2007, in rendering its Massachusetts v. EPA decision on carbon
dioxide regulation, the Supreme Court again resorted to plain meaning
interpretation of the Clean Air Act.154
The eleven federal circuits, in blocking access of PRPs to section 107
between 1994 and 2003, did not apply the canons of statutory construction
as did the Supreme Court in Atlantic Research. Some of the circuit courts,
which held that section 107 was not available to private parties, inferred a
congressional purpose in the 1986 section 113(f) SARA amendments to
have impliedly corrected and preempted the use of the previously enacted
146. United States v. Atl. Research Corp., 551 U.S. 128 (2007).
147. See Richard A. Posner, Statutory Interpretation– in the Classroom and in the Courtroom,
50 U. CHI. L. REV. 800, 808 (1983) (refuting conception that most judges actually begin statutory
interpretation by looking at the language of the act).
148. Blake A. Watson, Liberal Construction of CERCLA Under the Remedial Purpose Canon:
Have the Lower Courts Taken a Good Thing Too Far?, 20 HARV. ENVTL. L. REV. 199, 212–13 (1996)
(quoting United States v. Missouri Pac. R.R., 278 U.S. 269, 278 (1929)).
149. Id. at 243 (quoting William N. Eskridge, Jr. & Philip P. Frickey, Statutory Interpretation As
Practical Reasoning, 42 STAN. L. REV. 321, 354 (1990)).
150. Id. at 219–20. Adherents to the Plain Meaning Rule are also known as textualists, and the
most well-known modern textualist is the U.S. Supreme Court Justice Antonin Scalia. See id.
151. Id.
152. Bates v. Dow Agrosciences LLC, 544 U.S. 431, 449 (2005).
153. County of Yakima v. Confederated Tribes & Bands of the Yakima Indian Nation, 502 U.S.
251, 265–66 (1992) (quoting Morton v. Mancari, 417 U.S. 535, 551 (1974)).
154. Massachusetts v. EPA, 549 U.S. 497, 528–29, 532 (2007).
2009] Cost Allocation Liability Conflicts Among the Federal Courts
275
section 107. However, there was nothing to “correct.” The first court
decision granting broad section 107 rights to private PRPs did not even
occur until 1987, after the Congress added section 113 with the SARA
amendments in 1986. Section 113(f) was added before any critical section
107 jurisprudence was decided.
Therefore, the rationale adopted by some circuits, that section 113 was
added to correct and limit the rampant application of section 107 by private
party plaintiffs to recover costs, does not comport with a true time line. The
SARA amendments’ legislative history is clear that SARA’s creation of the
new section 113 was only meant to accomplish the codification of the
implied right to contribution found in court decisions, not to impliedly
preempt any other avenues for cost recovery contained elsewhere in
CERCLA.155
In United Technologies Corp. v. Browning-Ferris Industries, Inc., the
First Circuit looked to the traditional section 113 meaning of the term
“contribution” but never interpreted section 107. The Second Circuit first
looked at this issue in Bedford Affiliates v. Sills, and similarly held that
section 107 was not available for a potentially responsible party because it
would render section 113 a nullity.156 Similarly, the Third Circuit also held
that a PRP may not bring a claim for recovery of costs against another
PRP,157 assiduously avoiding construing the meaning of section 107’s
language.158 The Fourth Circuit Court of Appeals used similar reasoning to
hold that a PRP must sue another PRP under section 113.159
In confronting the issue of standing to sue under section 107, the Sixth
Circuit160 construed only section 113’s language, not the plain language of
section 107.161 The Seventh Circuit,162 while leaving section 107 available
155. H.R. REP. NO. 99–253(I), at 79 (1985), reprinted in 1986 U.S.C.C.A.N. 2835, 2861 (House
Energy and Commerce Committee); S. REP. No. 99–11, at 44 (1985).
156. Bedford Affiliates v. Sills, 156 F.3d 416, 423–24 (2d Cir. 1998).
157. New Castle County. v. Halliburton NUS Corp., 111 F.3d 1116, 1119 (3d Cir. 1997).
158. See id. at 1123 (allowing a PRP to recover cleanup costs against another PRP under section
107 would strip section 113 of “any meaningful application”). The Third Circuit Court of Appeals cited
the arguments of other circuits in holding that PRPs may not bring section 107 cost recovery actions. Id.
at 1122–23 (citing First, Seventh, and Tenth Circuit Court opinions).
159. See Pneumo Abex Corp. v. High Point, Thomasville, & Denton R.R. Co., 142 F.3d 769,
776 (4th Cir. 1998) (explaining that a PRP must bring a claim against another PRP under § 9613); see
also Axel Johnson, Inc. v. Carroll-Carolina Oil Co., Inc., 191 F.3d 409, 415 (4th Cir. 1999) (citations
omitted). By the time of the Axel Johnson decision, the Sixth and Ninth Circuits had also ruled on this
issue.
160. Centerior Serv. Co. v. Acme Scrap Iron & Metal Corp., 153 F.3d 344, 349–50 (6th Cir.
1998).
161. See id. (noting that actions by PRPs are governed by section 113(f), not by section 107(a)).
In interpreting the “any other person” language of section 107, the court held that “any person may seek
276
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
for “innocent” parties, made no straightforward interpretation of section
107’s language. 163 The Ninth Circuit Court of Appeals, in ruling that a PRP
may not jointly and severally recover its cleanup costs from another liable
party,164 again focused only on the ordinary meaning of the term
“contribution” in section 113.165 The Tenth Circuit166 proceeded without
fully explaining its logic that “§ 113(f) would be rendered meaningless” if
section 107 was available to PRPs.167 The Eleventh Circuit simply held that
PRPs may not assert claims for recovery of costs under section 107.168
IV. HOW HAVE VARIOUS COURTS RESPONDED TO THE SUPREME
COURT?
Each Supreme Court decision on cost allocation in 2004 and 2007
reversed the holdings of every circuit court that had rendered a decision,
save one circuit, the D.C. Court of Appeals, which had not been asked to
render a decision. Have the lower courts negated their prior precedent and
conformed their rulings? Now more than two years past the last of these
two Supreme Court decisions, the checkerboard is very uneven. In this
section, I examine how the lower courts have or have not responded to
these Supreme Court rulings.
Below, I first examine the scurry to circumvent the 2004 decision in
Aviall. I examine whether this flurry of activity has or has not been
successful in refracting section 113 liability for contribution. In Section V, I
then examine the response of the district and circuit courts to the 2007
wholesale reversal of all federal circuit court CERCLA section 107
to recover costs under § 107(a), but . . . it is the nature of the action which determines whether the action
will be governed exclusively by § 107(a) or by § 113(f) as well.” Id. at 353.
162. Azko Coatings, Inc. v. Aigner Corp., 30 F.3d 761, 764–65 (7th Cir. 1997).
163. Id. at 764–65. This was one of the first of many times where a circuit court of appeals has
called these types of claims “a quintessential claim for contribution.” Id. at 764. However, the court
suggested that a landowner required to clean up a release of hazardous substances deposited on its land
by entirely unrelated third parties might be able to pursue a section 107 cost recovery action. See id.
Subsequent Seventh Circuit decisions reiterated this “innocent landowner” exception. See, e.g., AM
Int’l, Inc. v. Datacard Corp., 106 F.3d 1342, 1346–47 (7th Cir. 1997); Rumpke of Ind., Inc. v. Cummins
Engine Co., Inc., 107 F.3d 1235, 1239 (7th Cir. 1997) (discussing innocent landowner exception);
NutraSweet Co. v. X-L Eng’g Co., 227 F.3d 776, 784 (7th Cir. 2000) (describing the requirements for
the innocent landowner exception).
164. Pinal Creek Group v. Newmont Mining Group, 118 F.3d 1298, 1299–1300 (9th Cir. 1997).
165. Id. at 1301.
166. United States v. Colo. & E. R.R. Co., 50 F.3d 1530, 1535 (10th Cir. 1995).
167. Id. at 1536.
168. See Redwing Carriers, Inc. v. Saraland Apartments, 94 F.3d 1489, 1512–13 (11th Cir.
1995).
2009] Cost Allocation Liability Conflicts Among the Federal Courts
277
precedent in the Atlantic Research opinion. Here again, the response has
been less than uniform. In both of these implementing responses by the
courts, is writ part of the future of hazardous substance liability.
A. Responding to the Aviall Section 113 Settlement Blockade
In the words of Aviall's counsel more than a year after the 2004
decision, “the controversy continues, sites remain un-remediated, parties
who act responsibly remain uncompensated, and flagrant polluters still
scoff at potential liabilities.”169 Post-Aviall, there were two types of newlyminted initiatives by private party plaintiffs to try to end-run the roadblocks
for cost contribution and recovery created by the combination of this
decisive Supreme Court decision limiting access to section 113 along with
the prior uniform circuit court opinions cutting off access to section 107.
These initiatives dealt with the past and with the future.
First, the past: there remained a limbo of twenty-five years of prior past
settlements. To get around the newly articulated Aviall Supreme Court
requirement of a prior settlement to utilize section 113, previous settling
parties claimed that even though not recited in past responses to orders or
threats from state environmental agencies, the real hidden intent of the
parties was to resolve federal CERCLA liability to qualify to utilize section
113. Some states also supported this revisionist initiative to empower such
settlers to avail themselves of section 113 federal contribution rights. The
cases170 made it not a straightforward proposition.
Second, the future: in new CERCLA settlements, parties were insisting
that state agencies now include language that the agency otherwise would
not include—to recite that its purpose was to resolve federal CERCLA
liability as well as state liability. Some parties were proactively working
with state environmental authorities to have the state file a friendly suit,
immediately resolved by a consensual settlement, which was then judicially
169. Richard O. Faulk et al., Cost Recovery Under CERCLA Section 107 After Cooper v. Aviall,
37 ENV’T REP. (BNA) 640, 641 (2006).
170. See Vine St. LLC v. Keeling, 362 F. Supp. 2d 754, 761 (E.D. Tex. 2005) (“[A] claim for
contribution under § 113(f)(1) can only be brought by a potentially responsible party ‘during or
following any civil action’ under § 106 or § 107(a) of CERCLA.”); see also City of Waukesha v.
Viacom Int’l Inc., 362 F. Supp. 2d 1025, 1027 (E.D. Wis. 2005); W.R. Grace & Co.-Conn. v. Zotos Int’l,
Inc., No. 98-VC-838S(F), 2005 WL 1076117, at *7 (W.D. N.Y. May 3, 2005); Esso Standard Oil Co. v.
Rodriquez Perez, No. Civ. 01–2012(SEC)(JA), 2005 WL 643484 (D.P.R. Mar. 21, 2005); Pharmacia
Corp. & Solutia Inc. v. Clayton Chem. Acquisition LLC., 382 F. Supp. 2d 1079 (S.D. Ill. 2005); Johnson
v. City of San Diego, 2005 Cal. App. Unpub. LEXIS 1979, *26–27 (Cal. Ct App. Mar. 4, 2005);
Elementis Chem., Inc. v. T.H. Agric. & Nutrition, L.L.C., 373 F. Supp. 2d 257, 266 (S.D. N.Y. 2005);
AMW Materials Testing, Inc. v. Town of Babylon, 348 F. Supp. 2d 4, 11 (E.D. N.Y. 2004).
278
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
approved under stipulation of the parties in order to allow access to section
113.
A few states tried to chart an alternative path under state statute.171 A
number of courts post-Aviall moved immediately to dismiss contribution
suits under section 113 by plaintiffs who had not reached a judicially or
administratively approved settlement with, nor been sued by, the
government.172 Trial courts in the Fifth Circuit173 and the Ninth Circuit174
took a pragmatic response and allowed section 107 claims to be amended to
comport with allowable pending litigation post-Aviall.
1. Reconfiguring Past pre-Aviall Settlements Retroactively
Some states were willing to collude with PRPs post hoc in an effort to
create retrospective federal rights for some PRPs who previously settled
only with the state. This involved state agencies going back months or
171. See, e.g., R.R. St. & Co., Inc. v. Pilgrim Enter., Inc., 166 S.W.3d 232, 238, 240 (Tex. 2005)
(noting that private cost-recovery actions under Texas’s state counterpart to CERCLA requires different
elements to maintain a cause of action).
172. W.R. Grace & Co.-Conn., 2005 WL 1076117, at *2; see also Vine Street, 362 F. Supp. 2d at
761 (dismissing Vine Street’s contribution claim under section 113(f)(1) because “113(f)(1) only allows
claims ‘by and between jointly and severally liable parties for an appropriate division of the payment
one of them has been compelled to make’”); Mercury Mall Assocs. v. Nick’s Mkt., Inc., 368 F. Supp.
513, 518 (E.D. Va. 2005); Waukesha, 362 F. Supp. 2d at 1026–27 (dismissing a section 113 claim in
light of the Supreme Court’s holding in Aviall); R.E. Goodson Constr. Co. v. Int’l Paper Co., No. C/A
4:02-4184-RBH, 2005 W.L. 2614927, at *21 (D.S.C. Oct. 13, 2005) (dismissing plaintiff’s section
113(f) claims for lack of a civil suit being brought prior to the section 113 claims); Metro. Water
Reclamation Dist. of Greater Chi. v. Lake River Corp., 365 F Supp. 2d 913, 917–18 (N.D. Ill. 2005)
(ruling that parties who voluntarily undertake cleanup efforts may not sue for contribution); E.I. DuPont
de Nemours & Co. v. United States, 460 F.3d 515, 543–544 (3d Cir. 2006) (noting that the Supreme
Court’s decision in Aviall precluded a PRP that voluntarily cleaned up a site from recovery under section
113).
173. See Vine St., 362 F. Supp. 2d at 763 (permitting access to section 107 as a default
alternative where section 113 is not available after Aviall). Note that Cooper v. Aviall came before the
Supreme Court from the Fifth Circuit returned there on remand. Cooper Indus. Inc. v. Aviall Servs. Inc.,
543 U.S. 157, 165, 171 (2004). On remand to the Fifth Circuit, the en banc remanded the case to the
district court and ordered the district court to permit Aviall to amend its complaint to reassert its original
section 107 claim for cost recovery. Aviall Serv. Inc. v. Cooper Indus., 572 F. Supp. 2d. 676, 683–84
(N.D. Tex. 2008). However, on remand the district court ultimately refused to recognize the availability
of the section 107 path. Aviall Servs. Inc. v. Cooper Indus., LLC, No. 3:97–CV–1926–D, 2006 U.S.
Dist. LEXIS 55040, at *10 (N.D. Tex. Aug. 8, 2006). The district court concluded that Aviall could not
use section 107(a) for either a cost recovery action or a contribution claim. Id. Thus, parties prior to a
judicially approved settlement with the federal government can not recover costs under any federal
scheme, as both section 107 and section 113 are walled off. Over the years, different panels of the Fifth
Circuit had suggested, but never squarely held, that a private PRP could utilize section 107 for cost
recovery.
174. See Kotrous v. Goss-Jewett Co. of N. Cal., No. Civ. S021520 FCD JFM, 2005 WL
1417152, at *3 (E.D. Cal. June 16, 2005) (rationalizing that section 107 is the original source of
contribution).
2009] Cost Allocation Liability Conflicts Among the Federal Courts
279
years after a settlement was final and unappealable to rewrite settlement
documents to fit the newly articulated contours of Aviall. These settlements
were reworked cooperatively by the state and the PRP to recite post hoc
purported federal elements of settlement that were not initially included
when the settlement was actually entered. What is the dispositive value of a
settlement that is cooperatively redrafted after the fact by the supposedly
legally adverse state enforcement agency and target PRPs?
The desire to paper-over prior state settlements that did not even
purport by their original written terms to address, let alone resolve, federal
CERCLA liability was vigorous post-Aviall. In essence, states were asked
to mask the true and originally cited facts for the settlement so as to provide
additional protection for certain parties.
Clearly, such post-hoc
modifications are not part of the original bargain, not supported by either
the original and certainly not new consideration, and could misstate the
original purpose and authority of the state in making the settlement.
Case law began to emerge post-Aviall on what type of prior state
settlements qualify as an administratively approved settlement for purposes
of the finality of CERCLA section 113(f). In Wisconsin, whether a plaintiff
could retroactively rewrite its settlement deficiencies post-Aviall was tested
where a city sued the successor to a corporation that had disposed of
hazardous substances at a city landfill.175 Even though the city had not been
sued by either the EPA or the state environmental authority, before the
Aviall decision, the court interpreted section 113(f)(1) to permit the city to
bring a section 113(f)(3) action for contribution.176 After the Aviall
decision, the defendants filed a motion to dismiss the prior ruling in
conformance with Aviall.177 Having initially allowed the section 113
contribution claim, the court after Aviall dismissed the action.178 It appears
that the city’s effort to disguise this deficiency by trying to create the
“settlement” after the fact, rather than resolving the issue, demonstrated that
175. Waukesha, 362 F. Supp. 2d at 1027–28.
176. Id. at 1026.
177. Id. at 1027. The city attempted to disguise the fact that it had no settlement with any
environmental enforcement agency. The city tried to amend its complaint by adding a section
113(f)(3)(B) claim. Id. at 1027. The city claimed that its clean-up cost-sharing pilot contract with the
state was the legal equivalent of a settlement, and for good measure submitted to the state environmental
agency a settlement agreement purporting to settle all possible CERCLA claims and state liability. Id.
The court stated that a PRP must resolve its liability to the state before bringing a section 113(f)(3)
contribution action and “[i]f the unsigned administrative settlement agreement demonstrates anything, it
demonstrates that the City has not yet resolved its CERCLA liability to the State.” Id.
178. Id. at 1026–27.
280
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
the city knew that its original arrangement did not legally resolve its
liability to the state.179
In an Illinois matter, Pharmacia Corp. v. Clayton Chemical Acquisition,
L.L.C., nineteen PRPs entered into an administrative order on consent
(AOC) with the EPA pursuant to section 106 of CERCLA to perform the
remedial investigation and feasibility study (RI/FS) at a Superfund site.180
The plaintiff PRP brought suit for contribution against a group of unsettled
PRPs who were not party to any of the EPA orders.181 The court held that
the AOC was not an “administrative settlement” contemplated by section
113 of CERCLA for purposes of a subsequent contribution action.182 The
court focused on the fact that the AOC entered by the parties did not qualify
as a civil action183 and contained standard language stating that the parties
did not admit to liability, which belied any argument that it was a
settlement.184 Therefore, the responding PRPs could not initiate a
179. Id. at 1027. The court held that the cost-sharing agreement between the city and the state
was not the requisite settlement that resolved the plaintiff’s liability, since the statute authorizing the
agreement expressly provided that it did not affect any statutory or common law liability and because
the city submitted a separate administrative settlement that attempted explicitly to resolve the city’s
liability to the state. Id.
180. Pharmacia Corp. v. Clayton Chem. Acquisition, L.L.C., 382 F. Supp. 2d 1079, 1081, 1085
(N.D. Ill. 2005). The EPA also issued an administrative order unilaterally after the administrative order
on consent. Id. at 1081.
181. Id.
182. Id. at 1085. The plaintiffs asserted their contribution claim under section 113(f)(1) on the
basis that its facts were distinguishable since, unlike Aviall, plaintiffs incurred clean-up costs by
responding to two separate orders issued by the EPA: in an Administrative Order on Consent and a
Unilateral Administrative Order pursuant to section 106. Id. at 1085–86. The court noted that CERCLA
section 122 authorized the EPA to enter into administrative settlements, but the administrative order on
consent was issued pursuant to section 106 rather than section 122(d)(3). Id. It was consistently
captioned as an “order” rather than a “settlement.” Id. The plaintiff was attempting to recoup through a
contribution action some of the $3 million that it had expended after entering the administrative order on
consent. The court noted that the caption of the AOC provided that it was issued pursuant to section
106, but that the provisions of section 106 did not provide for settlements, and the term “settlement”
does not appear in section 106. Moreover, the court found that the penalties for violating the AOC were
those imposed pursuant to section 106 and that if the AOC was intended as a settlement, then the
penalties provided in the document would have been consistent with the penalties provided by section
122(1). Id.
183. The court looked to the Federal Rules of Civil Procedure and found that under Rule 2(a)
“‘civil action’ refers to the ‘entire civil proceeding, including all component “claims” and “cases” within
that proceeding.’” Id. at 1087. Under Rule 3, “‘[a] civil action is commenced by filing a complaint with
the court.’” Nowhere within the Rules is an administrative order even discussed. Id.
184. Id. The court looked at the term “civil action” in Black’s Law Dictionary to find the term
defined as a “non-criminal litigation,” whereas “administrative order” is defined as “[a]n order issued by
a government agency after an adjudicatory hearing” as well as “[a]n agency regulation that interprets or
applies a statutory provision.” Id.
2009] Cost Allocation Liability Conflicts Among the Federal Courts
281
contribution action after they began cleanup in compliance with an
administrative order on consent.185
The plaintiffs in Boarhead Farm Agreement Group v. Advanced
Environmental Technology Corp.186 also sought to distinguish their situation
from Aviall. In Boarhead, the plaintiffs had entered into a private
cooperative agreement to share in the costs of cleaning up two areas of the
Boarhead Farm Superfund Site.187 The court stated that to “stretch the
holding of Cooper Industries in such a way . . . would torture the plain
meaning of the statute and discourage PRPs not sued from cooperating and
settling with PRPs who were sued . . . .”188 There was no civil action in
Boarhead.189
A court in New York limited the ability of a settlement with the state to
be stretched after the fact to resolve CERCLA liability so as to enable a
private party section 113 contribution action.190 Where a party enters an
AOC with the state that does not contain any reference to CERCLA and
does not purport to release federal CERCLA liability of the settler, that
settlement relieves only the settler’s state liability and does not enable a
federal contribution action under CERCLA’s section 113.191 In this case, the
settling party amended its complaint after the decision in Aviall to add a
section 113(f)(3) contribution claim.192 The plaintiff argued that it had
entered into “two administratively approved settlements” with the state
Department of Environmental Conservation that by their terms resolved its
liability to the state, and therefore it could bring a claim for contribution
under section 113(f)(3).193
185. Id. The court applied the Aviall reasoning, finding that if Congress intended to allow a
contribution action at any time it would not have created two separate avenues for a PRP to seek
contribution nor would it have specified separately the conditions of a civil action in section 113(f)(1) as
well as an administrative or judicially approved settlement in section 113(f)(3). Id. at 1087–88.
186. Boarhead Farm Agreement v. Advanced Envtl. Tech. Corp., 381 F. Supp. 2d 427, 436 (E.D.
Pa. 2005).
187. Id. at 429–30. All but one of the members in the group had entered into one or more
settlement agreements with the EPA, which were thereafter entered in the district court as consent
decrees. Id. at 436.
188. Id. at 436.
189. Id. Even though the EPA settlement in Boarhead was entered in the district court as
consent decrees, the actions in Boarhead were not section 106 or section 107 civil actions, but rather
were governed by section 122 of CERCLA and therefore do not fall within the scope of section
113(f)(1).
190. W.R. Grace & Co.-Conn. v. Zotos Int’l, Inc., No. 98-CV-838S(f), 2005 WL 1076117, at *7
(W.D. N.Y. May 3, 2005). The AOCs were entered with the state environmental agency during the
1980s. The AOC resolved only state liability, but it did not release CERCLA liability, and it did not
indicate any EPA concurrence with the settlement. Id.
191. Id.
192. Id. at *3.
193. Id.
282
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
The court disagreed and held that for a settlement to be valid under
CERCLA section 122 it “must be judicially approved, i.e., entered as a
consent decree in the appropriate United States district court.”194 Section
122 of CERCLA grants to the EPA the authority to enter into settlement
agreements with a PRP and states that the settlement “must be ‘entered in
the appropriate United States district court as a consent decree.’”195 Section
104 of CERCLA provides that the EPA may “‘enter into a contract or
cooperative agreement with the State’” whereby a state may exercise
CERCLA authority, including the EPA’s authority to enter into settlement
agreements.196
A Texas court held that state consent agreements do not constitute
settlements resolving CERCLA liability.197 It is questionable whether a
state settlement that is finalized without some form of public comment and
agency response, as is required by CERCLA regulations,198 could qualify
under section 113 of CERCLA as an administratively or judicially approved
settlement. In Ferguson v. Acata Redwood Co., the court held that the
letters exchanged between the plaintiff and state and federal authorities did
not qualify as a settlement agreement under section 113(f)(3) because the
words “settlement” and “CERCLA” were nowhere contained in the
letters.199 Additionally, the court noted that the agency documents failed to
show that the state was acting pursuant to authority granted by the EPA.200
State-law bases for settlement do not satisfy the requisite for section
107 cost recovery. In Asarco v. Union Pacific Railroad Co., the court found
that a memorandum agreement settling a dispute that had not evolved to
litigation did not constitute a settlement satisfying the post-Aviall section
113 requirement.201 And in Cadlerock Prov. Joint Venture v. Schilberg, the
194. Id. at *6.
195. Id. at *4 (citing 42 U.S.C. § 9622(d)(1)(A) (2000)).
196. Id. (citing 42 U.S.C. § 9604(d)(1)(A) (2000)). In looking at the terms of the 1988 Consent
Order issued by the state, the court found that it “does not state that the DEC was exercising any
authority under CERCLA, does not indicate that the EPA concurred with the remedy selected and does
not provide a release as to any CERCLA claims.” Moreover, nowhere within the order was the term
“CERCLA” used. Thus, the 1988 Consent Order only resolved Grace’s liability to the State of New
York, and it could not bring a claim for contribution under section 113(f)(3). Id. at *7.
197. See Vine Street LLC v. Keeling, 362 F. Supp. 2d 754, 761 (Mar. 24, 2005). In this case a
settlement with Texas’s environmental agency did not constitute an administratively approved settlement
for purposes of absolving CERCLA liability. Id.
198. 42 U.S.C. § 9622(d)(2)(B) (2006).
199. Ferguson v. Arcata Redwood Co., No. C 03–05632 SI, 2005 WL 1869445, at *5 (N.D. Cal.
Aug. 5, 2005).
200. Id.
201. Asarco, Inc. v. Union Pac. R.R. Co., No. CV 04–2144–PHX–SRB, 2006 WL 173662, at
*16 (D. Ariz. Jan. 24, 2006) (finding that the agreement satisfied neither section 113 nor section 122 of
CERCLA, and rejected the argument that where states were delegated to oversee clean-up, the states
2009] Cost Allocation Liability Conflicts Among the Federal Courts
283
court held that a state civil order to remediate a site was not a federal order
under section 106 of CERCLA and could not qualify to enable a section 113
contribution cost-recovery path.202
It can be difficult to convince a court of the true legal tenor of the past
once one rewrites long-ago legal agreements post hoc.
2. Controlling the Future of Settlements
Crafting future settlements that qualify for section 113 cost contribution
can be challenging. First, the EPA exercises unilateral discretion as to
whom it names as defendants in any section 106 administrative
enforcement action under CERCLA and with whom it will enter into an
administratively approved settlement. No party is deemed indispensable to
an EPA enforcement action and the EPA does not allow interlocutory
challenge to a section 106 order until after compliance or remediation.203
Therefore, under the Court’s new Aviall interpretive strictures, a PRP may
not bring a section 113 action without an approved settlement or by
establishing defendant liability through litigation by the EPA under
CERCLA.
could fashion CERCLA clean-up outside the normal contours of section 111); see generally Pharmacia
Corp. v. Clayton Chem. Acquisition, L.L.C., 382 F. Supp. 2d 1079, 1085 (N.D. Ill. 2005); W.R. Grace &
Co.-Conn., 2005 WL 1076117.
202. Cadlerock Properties Joint Venture v. Schilberg, No. 3:01CV895, 2005 WL 1683494, at *6
(D. Conn. July 19, 2005).
203. Section 113(h) deprives the federal courts of jurisdiction over any action seeking to
challenge a removal or remedial action decision made under section 104 (response authorities), or
embodied in a section 106(a) order (abatement actions). Exceptions to this rule include: (1) in the
context of a cost recovery, natural resource damages, or contribution actions that have been instituted
pursuant to section 107(a) (liability provisions); (2) a section 106-based action to enforce an order or
collect a penalty for violation of an order; (3) a section 106(b)(2) action against the government for
reimbursement of voluntary clean-up expenditures; (4) a citizen suit brought under section 310 of
CERCLA; or (5) a section 106 citizen action to compel remedial action. 42 U.S.C. § 9613(h) (2000).
See United States v. Princeton Gamma-Tech., Inc., 31 F.3d 138, 151–153 (3d Cir. 1994) (Nygaard, J.,
concurring) (arguing that, under the doctrine of sovereign immunity, Congress can limit suits against the
EPA as it wishes). Such a position is consistent with those taken by the Seventh, Eight, and Eleventh
Circuits. Id. at 151. CERCLA section 106 provides authority for the EPA to act by administrative order
instead of seeking judicial relief. Such orders appear to be sufficiently “final” actions to support an
action brought by the recipient seeking judicial review. However, courts such as the one in Wagner Seed
Co. v. Daggett, 800 F.2d 310 (2d Cir. 1986), have ruled that such orders are not reviewable except in
defense of the EPA’s enforcement suits.
284
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
a. Federal Settlements
In August 2005, in response to the Aviall decision, the EPA revised its
model administrative order on consent.204 The new language stated that the
AOC is designed to resolve liability to the federal government and protect
the settlers’ right to file contribution claims against other parties. The
semantics are also altered: the title of future orders is changed from AOC to
“administrative settlement and order on consent,” and the word
“settlement” is substituted for the term “order” in the text of the
document.205 These changes were enacted solely to gain traction under the
new interpretations of CERCLA liability for cost allocation.
However, regardless of terminology, courts are still confronted with the
fact that these orders/settlements are implemented pursuant to the EPA’s
authority under section 106 of CERCLA, which authorizes unilateral EPA
orders and not settlements.
Section 122 of CERCLA authorizes
settlements.206 It remains to be seen how courts will respond after twentyfive years of the EPA calling its unilaterally-issued administrative orders
pursuant to section 106 “orders.” The EPA is now relabeling these same
documents “settlements,” without invoking the separate settlement
authority under CERCLA.
Responding to an agency Administrative Order on Consent (AOC) or a
section 106 CERCLA unilateral order from the EPA may also not qualify to
trigger section 113 contribution rights. There are critical legal distinctions
between settlements and the unilateral EPA orders that are issued by the
agency. If a private party is forced to take action to clean up a site pursuant
to an EPA administrative order, section 113 is not invoked to empower
subsequent cost contribution from other parties.
The Eastern District of Missouri reconsidered whether a consent decree
can satisfy a settlement claim for contribution under section 113(f)(2).207
Plaintiff Mallinckrodt filed a counterclaim against several companies
alleging their liability for the contamination at an industrial park.208
Mallinckrodt later engaged in a settlement with the other companies
purporting to drop the suit if they helped pay a portion of the cleanup
204. U.S. ENVTL. PROT. AGENCY & U.S. DEP’T OF JUSTICE, MEMORANDUM: INTERIM
REVISIONS TO CERCLA REMOVAL, RI/FS AND RD AOC MODELS TO CLARIFY CONTRIBUTION RIGHTS
PROTECTION
UNDER
SECTION
113(F)
(2005),
available
at
AND
http://www.epa.gov/compliance/resources/policies/cleanup/superfund/interim-rev-aoc-mod-mem.pdf.
205. Id.
206. 42 U.S.C. § 9622 (2000).
207. United States v. Mallinckrodt, Inc., No. 4:02CV01488, 2006 WL 3331220 (E.D. Mo. Nov.
15, 2006).
208. Id. at 2.
2009] Cost Allocation Liability Conflicts Among the Federal Courts
285
costs.209 The court granted contribution protection for private party
settlement agreements, holding that the statutory protection in section
113(f)(2) covers settlements between private parties.210 The court thought
that to disallow private party settlements would be against the public policy
justifications of CERCLA, which encourage settlements.211 However, this
still is not a settlement with the EPA.212
Case law continues to emerge as to what prior state settlements qualify
as an administratively approved settlement for purposes of section 113(f).213
A remedial design/remedial action consent decree with the United States, or
an AOC with the EPA for a remedial investigation/feasibility study, removal
action, or reimbursement of response costs could give rise to a right of
contribution pursuant to section 113(f)(3)(B).214 Plaintiffs attempted to
recover contribution for costs under section 113 by construing
administrative orders as settlements.
District courts have differed as to whether an AOC under section 122
between the plaintiffs and the EPA constitutes a settlement for purposes of
section 113(f)(3)(B). The two sections of the statute are distinct. In
Responsible Environmental Solutions Alliance v. Waste Management, Inc.,
the plaintiffs entered into an AOC, which did not use the word
“settlement.”215 Using the Oxford English Dictionary’s definition of
“settle,”216 the court concluded that the AOC was in fact a settlement
agreement since it referenced the parties agreeing and the document
referred to the AOC as an “agreement.”217 Additionally, through the
209. Id. at 1.
210. Id.
211. Id. at 2.
212. Id.
213. The Supreme Court left open the possibility that administrative orders would qualify as
“civil actions” for CERCLA contribution purposes. See Cooper Indus. Inc. v. Aviall Servs. Inc., 543
U.S. 157, 168 n.5. (2004) (“Neither has Aviall been subject to an administrative order under § 106; thus,
we need not decide whether such an order would qualify as a ‘civil action under section 9606 or under
section 9607(a)’ of CERCLA.”).
214. In order to clarify this issue, the EPA and the U.S. Department of Justice signed a
settlement agreement. U.S. ENVTL. PROT. AGENCY & U.S. DEP’T OF JUSTICE, supra note 204, at 3, 5, 8.
“The parties agree that this Settlement Agreement constitutes an administrative settlement for purposes
of section 113(f)(3)(B) of CERCLA, 42 U.S.C. § 9613(f)(3)(B), pursuant to which Respondents have, as
of the Effective Date, resolved their liability to the United States for the Work.” Id.
215. Responsible Envtl. Solutions Alliance v. Waste Mgmt., Inc., 493 F. Supp. 2d 1017, 1023
(S.D. Ohio 2007).
216. The Oxford English Dictionary defines “settle” as to fix by “mutual agreement.”
Responsible Envtl. Solutions Alliance, 493 F. Supp. 2d at 1023 (citing OXFORD ENGLISH DICTIONARY 86
(2d ed. 1989)).
217. Contra Pharmacia Corp. v. Clayton Chem. Acquisition, L.L.C., 382 F. Supp. 2d 1079,
1085 (S.D. Ill. 2005) The Pharmacia court held that the administrative order on consent was not an
“administrative settlement” contemplated by section 113 of CERCLA for purposes of a subsequent
contribution action. Id. The decision in Pharmacia significantly narrowed the ability of a PRP to bring
286
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
execution of the AOC, the plaintiffs had demonstrated that they had settled
their liability with the EPA.218
Comparatively, the majority approach is articulated in Emhart
Industries, Inc. v. New England Container Co., where the court held that the
“balance of decisional authority suggests that the type of administrative
order to which Emhart is subject is not a ‘civil action’ within the meaning
of § 113(f)(1).” Thus, Emhart was precluded from bringing a claim for
contribution under section 113(f)(1).219 In ITT Industries, Inc. v.
BorgWarner, Inc. (ITT Industries (I)), the district court denied that the AOC
at issue constituted a settlement because it was an interim agreement with
the EPA that did not resolve the plaintiff’s liability and therefore did not fall
within the description of the agreements in section 113(g)(3)(b).220 On
appeal, the Sixth Circuit found that the plaintiff’s administrative agreement
with the EPA did not resolve the plaintiff’s CERCLA liability and thus did
not confer contribution rights under section 113(f)(3)(B).221
The 2004 Supreme Court decision in Aviall did not address the issue of
whether private parties cleaning contaminated sites pursuant to a Unilateral
Administrative Order (UAO) have a right of contribution under section
113(f)(1). Case law has shown that a UAO can qualify as a civil action for
purposes of a PRP’s contribution claim under CERCLA section 113(f)(1).222
Using the reasoning of the Sixth Circuit’s former precedent in Centerior
a contribution action pursuant to section 113(f)(1) by barring a PRP who has incurred clean-up costs
pursuant to a section 106 administrative order from bringing a claim for contribution. Id.
218. Responsible Envtl. Solutions Alliance, 493 F. Supp. 2d at 1024. Cf. Asarco, Inc. v. Union
Pac. R.R. Co., No. CV 04–2144–SRB, 2006 WL 173662, at *16 (D. Ariz. Jan. 24, 2006). In Asarco, the
Nebraska Department of Environmental Quality expended nearly $30 million to decontaminate the site,
and sought contribution from defendant for its share of the cleanup costs, pursuant to section
113(f)(3)(B). Id. The court reasoned that the agreement did not constitute such a settlement, because it
failed to resolve the plaintiff's liability under either state law or CERCLA and therefore did not give rise
to a right of contribution under section 113(f)(3)(B). Id.
219. Emhart Indus., Inc. v. New England Container Co., Inc., 478 F. Supp. 2d 199, 203 (D.R.I.
2007). “[B]ecause there is no evidence to support the contention that the plain meaning of ‘civil action’
includes EPA-issued administrative orders, this court will follow the majority of courts in concluding
that 113(f)(1) is unavailable for parties who are merely subject to administrative orders, as opposed to
final consent decrees, judgments, or apportionments of liability. Count II will accordingly be
dismissed.” Id.
220. ITT Indus., Inc. v. BorgWarner, Inc., No. 1:05-CV-674, 2006 WL 2460793, at*13 (W.D.
Mich. Aug. 23, 2006) (stating that an AOC is not an administrative settlement for purposes of section
113(f)(3)(B)).
221. ITT Indus., Inc. v. BorgWarner, Inc., 506 F.3d 452, 460 (6th Cir. 2007); see Raytheon
Aircraft Co. v. United States, 435 F. Supp. 2d 1136, 1142–43 (D. Kan. 2006) (holding that a section 106
administrative order does not qualify as a “civil action” and therefore does not authorize suit under
section 113(f)(1)).
222. Carrier Corp. v. Piper, 460 F. Supp. 2d 827, 840 (W.D. Tenn. 2006). A current owner of a
facility from which there was a release of a hazardous substance brought action against former facility
owner and current operator to recover past and future response costs, damages, and other relief. Id.
2009] Cost Allocation Liability Conflicts Among the Federal Courts
287
Service Co. v. ACME Scrap Iron & Metal Corp., a Tennessee district court
found that the UAO the plaintiff was issued under CERCLA’s section 106
qualified as a “civil action” because it satisfied the requirement in section
113(f)(1).223 The common law of contribution, which required only “that
plaintiff act under some compulsion or legal obligation to an injured party,”
was codified in the savings clause in CERCLA section 113(f)(1). Thus,
contribution under section 113(f)(1) was held to apply in claims where a
potentially responsible party has been compelled to pay for response costs
for which others are also liable, and who seeks reimbursement for such
costs.
b. State Settlements
Shifting from federal to state enforcement, settlements with a state
agency will not necessarily trigger access to CERCLA section 113 to allow
a private contribution action by the settling private party against other
recalcitrant and non-cooperating liable parties. In fact, the great bulk of
hazardous substance site remediation is driven by the enforcement efforts of
state environmental agencies rather than the federal EPA. The states
conduct ninety percent of all enforcement actions and ninety-seven percent
of the inspections at regulated facilities, while the remainder are performed
by the federal EPA.224 The states operate under their own state statutory
authority.
CERCLA section 113(f) does not define or limit what is an
administratively approved settlement. Moreover, the Court’s language
contemplates recourse to a section 113 contribution action “after an
administratively or judicially approved settlement that resolves liability to
the United States or a State.”225 In contrast, it might appear that an
approved settlement with a state environmental agency is an equally valid
portal to enable a subsequent section 113 federal CERCLA contribution
action by a private party to recover costs of cleanup.
However, it is not so straightforward. First, many state environmental
agencies do not have memoranda of understanding with the federal EPA to
allow the state to make settlements discharging or resolving any federal
claims.226 Second, the state only has rights under CERCLA or state law to
223. See Centerior Serv. Co. v. Acme Scrap Iron & Metal Corp., 153 F.3d 344, 351–52 (6th Cir.
1998) (stating that an administrative order under section 106 satisfies the requirement that a plaintiff be
under a legal obligation to pay costs before bringing an action under section 107(a)).
224. Dean Scott, State Officials Urge Funding Shift to Restore Fiscal Year 2007 Grants, 37
ENV’T. REP. (BNA) 400, 401 (2006).
225. 42 U.S.C. § 9613 (2006).
226. Id. § 9622.
288
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
recover its own specifically incurred site response and cleanup costs, which
by definition do not include any federal incurred response costs or rights.
Therefore, the state settlement can not qualify as a settlement under
CERCLA.
For example, the state of New York proactively pronounced that all of
its past hazardous substance settlements satisfied the requirements of
section 113(f)(3) of CERCLA.227 A federal judge subsequently thought to
the contrary, ruling that settlements by consent with the New York state
environmental agency only resolved liability as against the state and not
potential federal CERCLA claims, and therefore that a settling party was
not entitled to utilize federal section 113 contribution actions under the
Aviall precedent.228 The court held that a right of contribution under state
law cannot flow from the federal CERCLA statute.229 Regardless of the
defendant’s liability, the court found that plaintiff was precluded from
shifting liability to that defendant under section 113.
Section 113(f)(3)(B) permits a contribution action to be brought by a
party who has “resolved its liability to the United States or a State for some
or all of a response action or for some or all of the costs of such action in an
administrative or judicially approved settlement.”230
Courts have
consistently stated that a PRP must resolve its liability to the state before
bringing a section 113(f)(3) contribution action. In BASF Catalysts L.L.C.
v. United States, the Massachusetts district court continued to interpret
section 113 strictly.231 The court held that a consent order entered by a
contractor with the EPA did not resolve its liability for purposes of
CERCLA’s contribution provision, especially since the consent order stated
that the parties agreed to resolve liability to the state for purposes of
227. Memorandum of Law of the State of New York As Amicus Curiae in Opposition to
Defendants’ Motion to Dismiss, Seneca Meadows, Inc. v. ECI Liquidating, Inc., 427 F.2d 279 (W.D.
N.Y. 2006). The New York Attorney General filed an amicus brief representing the New York State
Department of Environmental Conservation opposing defendant’s motion to dismiss, arguing that
CERCLA constitutes the very basis of New York’s settlement authority, and all settlements are meant to
satisfy section 113(f) requirements. Id. The Attorney General cited the opinion in Pfohl Bros. Landfill
Site v. Allied Waste Sys., 255 F. Supp. 2d 134 (W.D. N.Y. 2003), which, before the Aviall decision, held
that orders of consent in New York qualified as approved settlements under section 113(f)(3) of
CERCLA. Id. The Attorney General argued that the court should distinguish the Pharmacia decision
because that involved a section 106 EPA order rather than a settlement with a state environmental
agency. Id.
228. W.R. Grace & Co.-Conn. v. Zotos Int’l, Inc., No. 98-CV-838S(F), 2005 WL 1076117, at
*12 (W.D. N.Y May 3, 2005).
229. Id.
230. § 9613(f)(3)(B).
231. BASF Catalysts L.L.C. v. United States, 479 F. Supp. 2d 214, 220 (D. Mass. 2007); see
City of Waukesha v. Viacom Int’l, 362 F. Supp. 2d 1025, 1027 (E.D. Wis. 2005) (holding that the PRP
must resolve its liabilities before it may bring a section 113(f)(3) contribution action).
2009] Cost Allocation Liability Conflicts Among the Federal Courts
289
CERCLA.
A contractor with unresolved liability is not allowed
contribution.232
The Second Circuit found that state consent orders, including those of
state agencies, qualify as contribution-conferring agreements for purposes
of CERCLA section 113(f)(3)(B).233 A consent order between an owner of a
landfill site and the New York State Department of Environmental
Conservation (DEC) amounted to an “administrative settlement,” and, thus
the owner could seek to recover from the PRP some of the costs that it
incurred in cleaning up contamination at the site. This is possible because
section 113(f)(3) allows a party to resolve its liability for part or all of a
response action or costs thereof through an administrative settlement. The
consent orders expressly stated that the parties agreed that the plaintiff had
resolved its liability to the state for purposes of CERCLA.234
Other district courts have found that a settlement with a state agency
that did not resolve CERCLA liability is not a basis for a CERCLA
contribution claim.235 Plaintiffs have to also absolve liability with the
EPA.236 A settlement with the state that contains unfulfilled conditions is
insufficient to trigger section 113(f)(3)(b).237
232. See, e.g., Waukesha, 362 F. Supp. 2d at 1027 (“If the unsigned administrative settlement
agreement demonstrates anything, it demonstrates that the City has not yet resolved its CERCLA
liability to the State.”); see also Vine St., L.L.C. v. Keeling, 362 F. Supp. 2d 754, 761 (E.D. Tex. 2005)
(holding that a settlement with the Texas environmental agency did not constitute an administratively
approved settlement for purposes of also having CERCLA liability).
233. Seneca Meadows, Inc., v. ECI Liquidating, Inc., 427 F. Supp. 2d 279 (W.D. N.Y. 2006);
see City Of Bangor v. Citizens Commc’ns Co., 532 F.3d 70 (1st Cir. 2008) (approving a consent decree
that allocated certain responsibilities among the City of Bangor, Maine and Citizens Communications
Company, and holding that non-settling defendants have standing to object to a partial settlement that
purports to strip it of a legal claim or cause of action); see also Appleton Papers Inc. v. George A.
Whiting Paper Co., 572 F. Supp. 2d 1034, 1042, 1043 n.8 (E.D. Wis. 2008) (noting that the clean-up
costs incurred pursuant to a consent decree were not incurred voluntarily, as payments made under
government duress are definitely not voluntary).
234. Seneca, 427 F. Supp. 2d at 286. The Seneca court referred to one commentator’s statement
that “until Congress directly addresses this issue [of when a private party can assert a contribution claim
under CERCLA in a purely private party cost recovery action] through an amendment to CERCLA, or
the Supreme Court resolves the issues that it left open in Cooper Industries . . . practitioners prosecuting
private party cost recovery actions would be wise to enhance the viability of contribution claims by
making sure that settlements with the EPA or a state clearly reference that CERCLA liability is being
resolved.” Id. at 290 n.9 (quoting Robert C. Goodman, CERCLA Contribution Actions After
Cooper/Aviall, 19 CAL. ENVTL. INSIDER 3, 4 (2005).
235. See, e.g., Niagara Mohawk Power Corp. v. Consol. Rail Corp., 436 F. Supp. 2d 398, 402
(N.D. N.Y. 2006) (“However, resolution of liability for state law claims does not meet the statutory
prerequisite. Rather, to bring a § 9613(f)(3)(B) claim, CERCLA liability must have been resolved.”)
(citation omitted); Asarco, Inc. v. Union Pac. R.R. Co., No. 04-2144–PHX–SRB, 2006 WL 173662, at
*7 (D.Ariz. Jan. 24, 2006) (“Just as to receive CERCLA contribution protection, one must comply with
CERCLA settlement procedures and resolve CERCLA liability, to initiate a CERCLA contribution
action, one must do the same.”).
236. W.R. Grace & Co.-Conn. v. Zotos Int’l, Inc., 559 F.3d 85, 90–91 (2d Cir. 2009) (holding
that the plaintiff could not seek contribution under section 133(f)(3)(B) because, although it had
290
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
After Aviall, plaintiffs in the Fifth Circuit claimed that by participating
in a Voluntary Cleanup Agreement (VCP) they had resolved some or all of
their liability to the State of Texas and the United States and had a claim for
contribution under section 113(f)(3)(B). The plaintiffs also asserted that
since the EPA could not enforce CERCLA claims against a party while that
party is participating in a voluntary state cleanup agreement, the VCP
resolved at least part of plaintiff’s liability to the federal government.238
The court held that a VCP was not a settlement and should not create a
contribution right under section 113.239 Quoting Consolidated Edison, the
court held that “the only liability that might some day be resolved under the
Voluntary Cleanup Agreement is liability for state law—not CERCLAclaims.”240 The VCP was an agreement for eventual resolution, but at the
time it did not resolve any claims.
Plaintiffs have successfully circumvented Aviall restrictions on the use
of section 113 by plaintiffs by asserting that state consent decrees entitle
them to section 113 contribution. In Booth Oil Site Administrative Group v.
Safety-Kleen Corp., the plaintiff sought contribution from other PRPs for
costs incurred pursuant to a consent order with the New York environmental
protection agency, since the plaintiff had resolved its proportionate share of
liability.241 Applying the analysis of Seneca Meadows, Inc. v. ECI
Liquidating, Inc.,242 the court concluded that consent decrees issued
resolved all liability with the state, it had not resolved all liability with the EPA); see Niagara Mohawk,
436 F. Supp. 2d at 402 (“A state has no CERCLA authority absent specific agreement with the federal
Environmental Protection Agency.”); see also Asarco, Inc., 2006 WL 173662, at *7 (“The question is
whether a settlement lacking EPA authorization can serve as a basis for a CERCLA contribution claim
under section 113(f)(3)(b). The court believes that it cannot.”) (emphasis in original).
237. The district court for the district of New Jersey held that outstanding conditions in the AOC
prevented Ford from recovering contribution. The AOC stated that its terms were subject to any
required public process. Despite language in the AOC that was intended to fully absolve Ford of
liability, the defendant presented evidence that the defendant opposed the consent order during the
public comment period and therefore the order lacked finality. See Asarco, Inc., 2006 WL 173662, at
*16 (holding that an agreement with a state that fails to resolve a plaintiff’s liabilities to that state does
not allow that plaintiff to maintain a section 113(f)(3)(B) action).
238. Differential Development-1994, Ltd. v. Harkrider Distributing Co., 470 F. Supp. 2d 727,
741 (S.D. Tex. 2007).
239. Id. at 743.
240. Id. at 740 (quoting Consol. Edison Co. of New York v. UGI Util. Inc., 423 F.3d 90, 96 (2d
Cir. 2005). The plaintiff entered VCP with the state agency and sought contribution under section 113(f)
(3)(B), asserting that its voluntary cleanup agreement with the state environmental agency was an
“administrative settlement” of CERCLA liability. Id. The court dismissed on the ground that the
agreement with the state agency only resolved state-law claims and did not resolve federal or state
CERCLA claims. Id.
241. Booth Oil Site Admin. Group v. Safety-Kleen Corp., 532 F. Supp. 2d 477, 502 (W.D. N.Y.
2007).
242. See Seneca Meadows, Inc., v. EIC Liquidating, Inc., 427 F. Supp. 2d 279 (W.D. N.Y. 2006)
(concerning a consent order between an owner of landfill site and the New York State Department of
2009] Cost Allocation Liability Conflicts Among the Federal Courts
291
pursuant to a New York state statute allowed the plaintiff to pursue a
contribution claim since the order purported to resolve the plaintiff's
CERCLA liability to the EPA.243
Sometimes, there is even less of a settlement: state agencies can
typically issue letters informing private PRPs that they are taking no further
action at a remediation site rather than enter a formal settlement when they
are not further pursuing an enforcement action against a responsible party at
a site. Such a letter does not rise to the level of an administratively or
judicially approved settlement post-Aviall.244 Such a letter of no further
action does not invoke the contribution protection for settling parties under
section 113(f), nor does it trigger the ability to utilize section 113 for a
private-party federal contribution action after Aviall. Some courts also
found that a failure to reference CERCLA prevents such a state settlement
from qualifying as the type of settlement that allows a federal section 113
contribution action post-Aviall.245
B. Summary
The jurisprudence is very mixed on the details of what does and does
not work legally after the Supreme Court set the record straight on section
113. Some courts have been concerned that the plaintiffs would not be able
to recover under any other theory, so in the interest of justice, they have
expressly stated they will reconsider section 113 cases, while others have
not countenanced voluntary cleanup agreements. Because the Aviall
decision did not define the specifications of eligible settlements, the
jurisprudence is split on whether AOCs constitute an eligible settlement for
purposes of allowing use of section 113 contribution actions. However, the
majority holding is that an AOC does not qualify, as it is not the result of a
Environmental Conservation that expressly stated that the parties agreed that plaintiff had resolved its
liability to the state for purposes of CERCLA).
243. Booth Oil, 532 F. Supp. 2d at 502 (“The Order on Consent states that it is issued pursuant
to the NYSDEC’s ‘authority under [New York's Environmental Conservation Law], Article 27, Title 13
and ECL 3–0301 . . . or pursuant to CERCLA, 42 U.S.C. § 9604.’”).
244. Some of the precedent interpreting what is an administratively approved settlement looks
to the requirements of section 122 of CERCLA, 42 U.S.C. § 9622 (2006), which sanctions settlements
of the agencies. See United States v. Cannons Eng’g Corp., 899 F.2d 79 (1st Cir. 1990) (discussing
agency discretion in settlements); United States v. Serafina, 781 F. Supp. 336, 338–39 (M.D. Pa. 1991)
(discussing the right to settle); Gen. Time Corp. v. Bulk Materials, Inc., 826 F. Supp. 471, 474–78 (M.D.
Ga. 1993) (reviewing the requirements for a judicially approved settlement).
245. See City of Waukesha v. Viacom Int’l Inc., 404 F. Supp. 2d 1112, 1115 (E.D. Wis. 2005)
(“[R]esolving liability with respect to non-CERCLA claims . . . does not create a CERCLA contribution
right under section 113(f)(3)(B).”); see also W.R. Grace & Co.-Conn. v. Zotos Int’l, Inc., No. 98-CV838S(F), 2005 WL 1076117, at *7 (W.D. N.Y. May 3, 2005) (stating that where a state may settle with a
PRP on its own authority, such a settlement may open the PRP to additional EPA liability).
292
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
“civil action.” Some courts have also required that liability is resolved with
a state before commencing a contribution action, although the precedents
are also split here on what qualifies and whether a state settlement can
substitute for resolving liability with federal agencies.
V. RESPONDING TO THE 2007 UNANIMOUS ATLANTIC RESEARCH
REVERSAL OF SECTION 107 PRECEDENT
When the Supreme Court reverses the dominant legal interpretations,
especially when those reversed decisions are embodied in the case
precedent of virtually every federal circuit in the country, one expects a
significant conforming reaction immediately thereafter across the courts. In
2004, the Court had also decided the Aviall decision, so there would soon
become two iterations of CERCLA liability reversals by the Supreme
Court. The section above analyzed what the agencies, parties, and courts
did in response to the 2004 Supreme Court decision in Aviall regarding
section 113 being closed off. Next, the U.S. Supreme Court opened a new
door of section 107 cost recovery in 2007 in its Atlantic Research decision.
I examine here the subsequent lower federal court decisions that chart
the new CERCLA liability landscape in this brave new world of hazardous
substance. As of late 2009, for such a hotly contested matter involving
thousands of sites nationwide, there is less movement than one might
expect more than two years after the Supreme Court told all of the Circuits
that their opinions were misinterpreting the plain meaning of one of the
major financial allocation statutes in the country. I examine each circuit’s
conformance, starting with the First Circuit.
A. First Circuit
The First Circuit is a jurisdiction where no movement is yet obvious.
The First Circuit’s 1994 precedent on section 107 liability and relief has not
been overruled. In United Technologies Corp. v. Browning Ferris
Industries, Inc., a PRP entered into a consent decree under which it agreed
to clean up a contaminated site in accordance with the remedial action
requested by the EPA.246 After spending thirteen million dollars to clean up
a toxic spill, the plaintiff sought recovery of monies from the defendants to
pay back the state and the EPA.247 The court determined that the UTC
action was for contribution under section 113, and that it should therefore
246. United Techs. Corp. v. Browning-Ferris Indus., Inc., 33 F.3d 96, 103 (1st Cir. 1994).
247. Id. at 97.
2009] Cost Allocation Liability Conflicts Among the Federal Courts
293
constrain the claim to section 113’s three-year contribution statute of
limitations and not the six-year statute of limitation pursuant to section
107.248
Some lower federal courts in the First Circuit have shown movement.
In light of the 2007 Atlantic Research decision, the federal district court of
Rhode Island found an implied right to contribution under section 107 even
though contribution technically is contained only in section 113.249 Emhart
Industries Inc. brought claims under both CERCLA section 113 and section
107(a), alleging that New England Container Co. (NECC) had stopped
complying with CERCLA section 106 unilateral administrative orders,
which mandated that both Emhart and NECC perform removal actions at a
site in Rhode Island.250 The district court held that the section 106
administrative orders were not a “civil action” under section 113(f)(1) of
CERCLA, and thus Emhart could sue NECC for contribution under that
section.251 The court also found that Emhart, as a non-“innocent” PRP,
could not use section 107(a) for a direct cost recovery action against
NECC.252 Note, though, that this decision occurred just months before the
decision in Atlantic Research put the section 107 issue in a new corrected
interpretative posture.
However, largely because Emhart would otherwise be precluded from
recovering its equitable share of costs from NECC, the court found that
section 107(a) granted PRPs an “implied right to contribution.”253
Therefore, it was consistent with the eventual resolution of Atlantic
Research, but sequentially before its time. Emhart was permitted to
proceed under CERCLA section 107 to recover a portion of its costs from
NECC. Thus, consistent with Atlantic Research, the court ruled that a PRP
may bring an action against another PRP under section 107(a) of CERCLA.
B. Second Circuit
The Second Circuit was in an opposite mode and extremely active: after
both of the 2004 and 2007 Supreme Court decisions, W.R. Grace & Co. v.
Zotos International, Inc. overruled the Second Circuit’s prior holding in
248. Id. at 103.
249. Emhart Indus., Inc. v. New England Container Co., Inc., 478 F. Supp. 2d 199, 204 (D.R.I.
2007).
250.
251.
252.
253.
Id. at 202.
Id. at 203.
Id. at 202.
Id. at 204.
294
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
light of, and consistent with, the Supreme Court’s opinions.254 Although
only in dicta, Zotos not only embraced Atlantic Research’s holding that a
private PRP may assert a claim under section 107, it also went further to
answer some questions left open by the 2007 Supreme Court decision.
Zotos held that a PRP who had voluntarily entered into an administrative
consent order prior to remediation can assert a claim under CERCLA
section 107(a).255 The plaintiff, Grace, voluntarily entered into an
administrative order on consent (AOC) with the New York state
environmental agency. According to the AOC, Grace would reimburse the
agency for some of the response costs already spent and would conduct
RI/FS and remediation activities. The court applied both Atlantic Research
and Aviall to hold that a private party that had not yet been sued by the state
or federal agency, but that had merely “voluntarily” entered into a
cooperative agreement with either of those entities and as a result incurred
response and remediation costs, could seek recovery of response and
remediation costs under section 107(a)(4).256
The Atlantic Research decision employs the term “voluntary,” but it
does not define the term. The Second Circuit tried to define section 107’s
“voluntary” cleanup by differentiating voluntary cleanup actions from
involuntary consent decrees. In Niagara Mohawk Power Corp. v.
Consolidated Rail Corp., the plaintiff entered into two consent decrees with
the state that did not give it the ability to bring a section 113 contribution
action.257 Since the plaintiff had entered into two consent orders, it could
not be deemed a “volunteer” who was entitled to seek recovery under
section 107. 258 The court held that nothing in Atlantic Research provides
authority to change the Circuit’s previous decisions, in which it ruled that a
party that has incurred expenditures under a consent order with a
government agency and was found partially liable under § 113(f)(1) could
not seek to recoup those expenditures under section 107(a).259
254. W.R. Grace & Co.-Conn. v. Zotos Int’l, Inc., 559 F.3d 85, 90 (2d Cir. 2009). The Second
Circuit’s former precedent, Bedford Affiliates v. Sills, 156 F.3d 416, 424 (2d Cir. 1998), limited
recoveries by PRPs to actions brought under section 113(f) and held that a PRP is limited to a section
113 claim for contribution. Bedford Affiliates points out that section 113(f)(1) “apportions liability
based on equitable considerations and has a three-year statute of limitations,” whereas “section 107(a)
has a six-year statute of limitations.” Id. Thus, the Second Circuit reasoned, if a PRP could choose,
section 113(f)(1) would become a nullity, as “[a] recovering liable party would readily abandon a §
113(f)(1) suit in favor of the substantially more generous provisions of § 107(a).” Id.
255. W.R. Grace & Co.-Conn.,559 F.3d at 93.
256. Id.
257. Niagara Mohawk Power Corp. v. Consol. Rail Corp., 436 F. Supp. 2d 398, 401 (N.D. N.Y.
2006).
258. Id. at 403.
259. Id.
2009] Cost Allocation Liability Conflicts Among the Federal Courts
295
During the 2004–2007 interim between the two Supreme Court
opinions, and thus before section 107 was reopened, the Second Circuit
held that section 107 was available for private parties to recoup cleanup
costs where cleanup had occurred “voluntarily, not under a court or
administrative order or judgment.”260 Additionally, the Court of Appeals
considered in this interim period whether liable parties have a cause of
action under section 107, holding that they might, pursuant to case-by-case
consideration on a fact-by-fact basis.261 According to the Second Circuit,
section 107(a) permits a party that has not been sued or made to participate
in an administrative proceeding, but would be held liable under section
107(a) if sued, to recover necessary response costs incurred voluntarily
rather than under a court or administrative order or judgment.262 Therefore,
this did not open the section 107 door as widely as the Supreme Court
would correctively later do in Atlantic Research.
After this second 2007 Supreme Court decision, the Second Circuit
found in 2008 that future CERCLA response costs are not covered under
section 107. In New York v. Next Millennium Realty, L.L.C., the defendant
based its section 107 claim on remediation costs it already incurred coupled
with the future costs anticipated under section 107(a). However, the
defendant did not demonstrate that it incurred these necessary costs of
response within the meaning of section 107(a).263 The court found that a
section 107(a) action is viable where a party itself incurred cleanup
response costs as opposed to reimbursing costs paid by other parties, which
are more appropriately covered under section 113(f).264
Contrary to this decision, at about the same time a different panel of the
Second Circuit held that section 107 is available not just for “volunteers”
but also for parties that have conducted a cleanup pursuant to a consent
decree entered with the government.265 The court held, however, that a
party conducting a cleanup pursuant to a consent decree must have
conducted at least some cleanup activities prior to entering into the consent
decree.
After both Supreme Court opinions, a district court within the Second
Circuit, in Champion Laboratories v. Metex Corp., amended its decision in
light of the Atlantic Research opinion and reinstated Champion’s section
260. Consol. Edison Co. of N.Y. v. UGI Util. Inc., 423 F.3d 90, 100 (2d Cir. 2005).
261. Id.
262. Id.
263. New York v. Next Millennium Realty, L.L.C., No. CV-03-5985(SJF)(MLO), 2008 WL
1958002, at *7 (E.D. N.Y. May 2, 2008).
264. Id. at *6.
265. Schaefer v. Town of Victor, 457 F.3d 188, 201–02 (2d Cir. 2006).
296
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
107 claim against Metex for contamination at the Champion site.266
Following both Supreme Court decisions in another district court case
within the Circuit, New York v. Solvent Chemical Co., the defendant
sustained expenses pursuant to a consent decree following suit under
section 107(a) but did not incur costs voluntarily or reimburse the costs of
another party.267 The court allowed the defendant to assert section 107
claims relating to response costs incurred following the consent decree.
The court ruled that Solvent Chemical Company had a cause of action for
cost recovery against the defendant under section 107, since it directly
incurred response costs in performing remedial activities: “regardless of
what section of CERCLA is involved, it will make every effort to fairly and
equitably apportion liability.”268
In a final district court opinion within the Circuit, a non-settling
defendant can be a PRP under CERCLA if it is a successor to a company
that disposed of waste at the landfill. In United States v. Kramer, an
extensively litigated case, settling defendants at the Helen Kramer landfill
in New Jersey were entitled to contribution for cleanup costs from a nonsettling defendant, Alumax, a successor entity to the company that had
arraigned for the disposal of waste materials at the landfill.269 The
successor was also responsible for the release of hazardous substances at
the facility.270
The Second Circuit, and federal district courts within the Circuit, have
implemented the recent Supreme Court CERCLA opinions and proceeded
further to flexibly interpret many of the open issues still pending after these
decisions.
C. Third Circuit
The Third Circuit has also responded to the Supreme Court holdings.
In E.I. du Pont de Nemours and Co. v. United States,271 the court vacated its
prior holding in light of the Supreme Court’s decision in Atlantic
266. Champion Labs. v. Metex Corp., Civ. No. 02-5284 (WHW), 2008 WL 1808309, at *4
(D.N.J. Apr. 21, 2008).
267. New York v. Solvent Chem. Co., No. 83-CV-1401C, 2008 WL 3211273, at *4 (W.D. N.Y.
Aug. 6, 2008).
268. Id. at *6.
269. United States v. Kramer, No. 89–4340(JBS), 2008 WL 5046846, at *6 (D.N.J. Nov. 19,
2008).
270. Id. at *7.
271. E.I. du Pont de Nemours & Co. v. United States, 508 F.3d 126, 135 (3d Cir. 2007).
2009] Cost Allocation Liability Conflicts Among the Federal Courts
297
Research.272 Du Pont admitted that it contaminated its industrial facilities
throughout the U.S. with hazardous substances but alleged that the U.S.
also contaminated parts of the sites. After du Pont voluntarily cleaned up a
site jointly polluted by both parties, du Pont filed suit under CERCLA
seeking an order requiring the government to reimburse it for a share of the
cleanup costs. The Third Circuit differentiated between “those who
voluntarily admitted their responsibility” and those who have “in fact been
held responsible (via adjudication or settlement with the EPA)” under
section 107(f).273 According to this decision, a PRP who communicates
with a regulatory agency concerning site cleanup does not make a PRP that
admits liability and accepts responsibility any less a volunteer under
CERCLA.
A federal district court within the Third Circuit has since addressed the
issues of “voluntary” action and joint and several liability left unresolved in
Atlantic Research.274 In Reichhold, Inc. v. U.S. Metals Refining Co., the
plaintiff brought suit to recover costs from other liable defendants for the
cleanup of hazardous waste on its copper refinery in New Jersey.275 In
determining joint and several liability, the court applied the Restatement
(Second) of Torts, section 433(A), wherein liability is joint and several
unless defendant can prove divisibility.276 Pursuant to the Restatement, the
New Jersey Court construed “voluntary” as having responded to no
litigation or order on the record, but it recognized the element of coercion
by concluding that any other interpretation of “voluntary” would leave such
plaintiffs with no remedy.
Updating the prior holding in another case in the Third Circuit, another
federal district court held that the decade-earlier decision in New Castle
County v. Halliburton NUS Corp.,277 which held that PRPs may not bring a
recovery action under section 107, no longer applied to state or federal
entities even when they are PRPs.278
272. See E.I. du Pont de Nemours & Co. v. United States, 460 F.3d 515, 544–45 (3d Cir. 2006)
(holding that section 113 provided the sole means for potentially responsible persons to obtain
contribution for clean-up costs). In their earlier opinion, the Third Circuit held that du Pont could not
pursue an action under CERCLA to recover from the United States a portion of its cleanup costs.
273. Id. at 546.
274. Reichhold, Inc. v. U.S. Metals Refining Co., Civ. No. 03–453 (DRD), 2008 WL 5046780,
at *6 (D.N.J. Nov. 20, 2008).
275. Id. at *1.
276. Id. at *7.
277. New Castle County v. Halliburton NUS Corp., 111 F.3d 1116, 1119 (3d Cir. 1997).
278. United States v. Simon Wrecking, Inc., 481 F. Supp. 2d 363 (E.D. Pa. 2007).
298
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
D. Fourth Circuit
The Fourth Circuit has not altered its precedent. Its early decision in
Pneunmo Abex Corp. v. High Point, Thomasville & Denton Railroad baring
use of section 107 remains undisturbed law.279 In 2008, a district court
recognized the recent holding of Atlantic Research in its decision in Ashley
II of Charleston, L.L.C. v. PCS Nitrogen, Inc.280 The affected site changed
owners many times and Ashley II, the present owner, filed a claim against
one of the previous owners claiming it was a PRP for the release of
hazardous substances. The court denied the defendant’s cost recovery
claim under section 107 because it was not a voluntary PRP and had only
reimbursed other parties’ remediation costs and had not incurred any of its
own cleanup costs.281
E. Fifth Circuit
Similarly, the Fifth Circuit precedent prohibiting the use of section 107
has not been overruled.282 In the interim, between the two 2004 and 2007
Supreme Court decisions, a federal district court held that a PRP could
bring CERCLA section 107 cost-recovery claims for voluntary cleanup
costs in Vine Street L.L.C. v. Keeling.283 The court held “that a potentially
responsible party that voluntarily works with a government agency to
remedy environmentally contaminated property should not have to wait to
be sued to recover cleanup costs since Section 113(f)(1) is not meant to be
the only way to recover cleanup costs.”284 The Vine Street Court reached
this decision by reasoning that a party not sued by an environmental
regulator was indeed “innocent” and should be entitled to the benefits of a
section 107 cost recovery action.285
The federal districts were not all in accord. The Southern District of
Texas mirrored the Texas Northern District’s post-2004 Supreme Court
decision on remand in Aviall v. Cooper Industries, 286 holding that section
107(a) does not give a PRP a statutory right to bring a cost recovery action
279. Pneumo Abex Corp. v. High Point, Thomasville & Denton R.R., 142 F.3d 769 (4th Cir.
1998).
280. Ashley II of Charleston, L.L.C. v. PCS Nitrogen, Inc., No. 2:05-2782-CWH, 2008 WL
2462862 (D.S.C. June 13, 2008).
281. Id. at *6.
282. OHM Remediation Serv. v. Evans Cooperage Co., 116 F.3d 1574, 1583 (5th Cir. 1997).
283. Vine Street L.L.C. v. Keeling, 362 F. Supp. 2d 754, 764 (E.D. Tex. 2005).
284. Id. at 763.
285. Id. at 762.
286. Aviall Servs. v. Cooper Indus., No. 3:97–CV–1926–D, 2006 WL 2263305, at *10 (N.D.
Tex. Aug. 8, 2006).
2009] Cost Allocation Liability Conflicts Among the Federal Courts
299
against other PRPs.287 Following Aviall, the Southern District of Texas
determined that the legislative intent behind section 113 was to have a
separate contribution remedy, subject to its own specific conditions and
limitations distinct from section 107(a).288 This was not immediately
expressly stricken, notwithstanding the 2007 Supreme Court decision in
Atlantic Research.
F. Sixth Circuit
Atlantic Research partially overruled the Sixth Circuit’s section 107
impediment precedent in Centerior Service Co. v. Acme Scrap Iron & Metal
Corp.289 After both Supreme Court decisions, there were more lower court
decisions in this Circuit. In Moraine Properties, L.L.C. v. Ethyl Corp.,
following ITT Industries, Inc. v. BorgWarner, Inc.,290 the court held that Ҥ
107(a) [was] the appropriate vehicle for a party that, as an assignee, ‘stands
in the shoes’ of the true plaintiff.”291
The Sixth Circuit struggled with whether a PRP who was required to
incur costs pursuant to a consent decree may bring a section 107(a) cost
recovery action.292 Prior to Atlantic Research, the Sixth Circuit held that
these PRPs were not entitled to bring a section 107(a) cost recovery action.
287. Differential Dev.-1994 Ltd. v. Harkrider Distrib. Co., 470 F. Supp. 2d 727, 753 (S.D. Tex.
2007).
288. Id.
289. Centerior Serv. Co. v. Acme Scrap Iron & Metal Corp., 153 F.3d 344, 356 (6th Cir. 1998),
abrogated by United States v. Atl. Research Corp., 551 U.S. 128, 144 (2007) (holding that section
107(a) provides a potentially responsible party with a cause of action to recover costs).
290. ITT Indus., Inc. v. BorgWarner, Inc., 506 F.3d 452, 454–55, 458 (6th Cir. 2007).
291. Moraine Properties, L.L.C. v. Ethyl Corp., No. 3:07-cv-229, 2008 WL 4758692, at *4 (S.D.
Ohio Oct. 27, 2008).
292. BorgWarner, 506 F.3d at 455. A successor corporation to a previous owner and operator at
a superfund site brought a CERCLA claim seeking recovery of response costs and contribution from
other PRPs after it incurred expenses in investigating and addressing hazardous conditions on two sites
regulated by the EPA. This PRP sustained expenses pursuant to a consent decree following a suit under
section 106 section 107(a). Therefore, in this case the PRP did not incur costs voluntarily and did not
reimburse the costs of another party. In BorgWarner, the Sixth Circuit affirmed a PRP’s right to pursue
cost recovery but rejected ITT’s right to bring a contribution action under section 113(f). Id. at 457–61.
BorgWarner concerned two different operable units of the North Bronson Industrial Area Superfund
Site, which the EPA placed on the National Priorities List after finding trichloroethylene contamination
at the site. The contamination stemmed from the manufacturing of fishing reels. As a corporate
successor to the fishing reel manufacturer, the EPA named ITT Industries a PRP, and ITT sought to
recover cleanup costs from others that had previously owned or operated the sites when hazardous
substances were deposited on the property. Id. at 455. ITT entered into an AOC and agreed to
undertake a remedial investigation and feasibility study (RI/FS) for one of the operable units, which cost
approximately $2 million. The Michigan Department of Environmental Quality (MDEQ) supervised
cleanup of the other operable unit.. ITT and several other parties entered into a consent decree with
MDEQ to perform the cleanup, incurring $1.6 million in costs pursuant. The Sixth Circuit reversed the
district court's dismissal of ITT’s section 107(a) cost recovery claim and remanded that action to the
district court for further consideration in light of the Atlantic Research decision. Id. at 458.
300
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
In BorgWarner, however, the Sixth Circuit reversed the district court’s
dismissal of the plaintiff's section 107(a) cost recovery claim and remanded
that action to the district court for further consideration in light of the
Atlantic Research decision.293 In 2009, the district court held that the party
was an involuntary plaintiff because it was required to incur cleanup costs
at one site pursuant to a consent decree and therefore could not bring a cost
recovery claim with respect to that site.294
G. Seventh Circuit
Atlantic Research distinguished the Seventh Circuit precedent in Akzo
Coatings, Inc. v. Aigner Corp.295 The Seventh Circuit relaxed its “innocent
parties” requirement by allowing section 107 to apply to PRPs who have
voluntarily initiated cleanup without having been subject to any kind of
EPA administrative order.296 In Metropolitan Water, the plaintiff-landowner
sought recovery of costs for voluntarily cleaning up hazardous waste
substances released by a tenant.297 The court, following Aviall, held that the
plaintiff could not bring a claim under section 113 because it had not been
sued by the government and, as a landowner, the plaintiff was not an
innocent third-party free to bring a claim under section 107.298 The court
held that a PRP could sue under section 107(a) where that PRP neither
settled any liability with the government, nor had been subject to a
CERCLA suit, and had also not been the subject of an EPA administrative
order under section 106.299
Despite the Atlantic Research Court’s use of the terms “voluntary” and
“involuntary” to distinguish between costs recoverable under section 107(a)
and those recoverable under section 113(f), at least one lower federal
district court adheres to the principle that section 107(a) is available to
293. Id. at 458.
294. ITT Indus., Inc. v. BorgWarner, Inc., 615 F. Supp. 2d 640, 646–48 (W.D. Mich. 2009). On
remand, the Western District of Michigan held that the plaintiff, who was required to pay clean-up costs
at one site pursuant to consent decree, could not bring a cost recovery claim for that site. Id. at 646.
295. Akzo Coatings, Inc. v. Aigner Corp., 30 F.3d 761, 770 (7th Cir. 1994). The court
unilaterally converted the section 107 claim against the plaintiff’s pleadings and upheld the district
court’s holding that the plaintiff was actually seeking contribution under section 113(f)(1), and not cost
recovery under section 107(a).
296. See Metro. Water Reclamation Dist. v. N. Am. Galvanizing & Coatings, 473 F.3d 824, 836
(7th Cir. 2007) (allowing a PRP to sue under section 107(a) where that plaintiff-PRP neither settled any
liability with the government, nor had been subject to a CERCLA suit, and had also not been the subject
of an EPA administrative order under section 106).
297. Id. at 825.
298. Id. at 829.
299. Id. at 836.
2009] Cost Allocation Liability Conflicts Among the Federal Courts
301
recover payments only in cases where section 113(f) is not.300 In Appleton
Papers Inc. v. George A. Whiting Paper Co., the court held that the cleanup
costs incurred pursuant to a consent decree were not incurred voluntarily;
the court noted that payments made under government duress are definitely
not voluntary.
H. Eighth Circuit
The Supreme Court’s Atlantic Research decision affirmed a 2006 postAviall opinion of the Eighth Circuit301 and opened up private PRP access to
section 107 cost recovery.302 A district court within the Eighth Circuit
changed its holding after this Supreme Court decision. On reconsideration,
the Eastern District of Missouri held that a state AOC can form the basis of
a contribution claim. In Westinghouse Electric Co. v. United States, the
Missouri court considered whether a state consent decree lacking an express
EPA delegation of authority to the state could be a judicially approved
CERCLA settlement triggering a section 113(f) claim.303 It held that a
section 107(a)(4)(a) response cost claim does not need a section 104
authorization from the state in order for the state to recover.304
I. Ninth Circuit
The Ninth Circuit officially overruled its section 107 liability holding
in Pinal Creek Group v. Newmount Mining Corp.305 in 2008 after Atlantic
Research.306 In Kotorous v. Gross-Jewett Co. of Northern California, a
plaintiff who had not been sued under section 106 or section 107 could file
a cost recovery action under section 107.307 The case also addressed an
300. Appleton Papers Inc. v. George A. Whiting Paper Co., 572 F. Supp. 2d 1034, 1042–43
(E.D. Wisc. 2008).
301. Atl. Research Corp. v. United States, 459 F.3d 827 (8th Cir. 2006), aff’d, 551 U.S. 128
(2007).
302. United States v. Atl. Research Corp., 551 U.S. 128, 134 (2007).
303. Westinghouse Elec. Co. v. United States, No. 4:03CV861(SNL), 2008 WL 2952759, at *3–
6 (E.D. Mo. July 29, 2008).
304. Id. at *7.
305. Pinal Creek Group v. Newmont Mining Corp., 118 F.3d 1298 (9th Cir. 1997). PRP actions
for directly incurring recovery costs are provided by section 107, which provides liability, and section
113, which provides relief. Metro. Water Reclamation Dist. v. N. Am. Galvanizing & Coatings, 473
F.3d 824 (7th Cir. 2007); see City of Rialto v. U.S. Dep’t of Defense, 274 F. App’x 515, 517 (9th Cir.
2008).
306. Kotrous v. Gross-Jewett Co. of N. Cal., 523 F.3d 924, 932–33 (9th Cir. 2008); see Adobe
Lumber, Inc. v. Hellman, No. CIV. 05–1510, 2009 WL 256553, at *2 (E.D. Cal. Feb. 3, 2009) (noting
the recent rulings by the Supreme Court and the Ninth Circuit).
307. Kotrous, 523 F.3d at 934.
302
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
issue not decided in Atlantic Research—whether a section 113 claim was an
appropriate counter-claim to a section 107 claim. The Ninth Circuit
determined that any defendant sued under a section 107 could counter with
a section 113 claim for contribution.308 This would seem contrary to the
precise nature of section 113(f)’s statutory contribution protection, but the
court plastically implemented it as a practical prudential element.
A federal district court within the Ninth Circuit recently held in Port of
Tacoma v. Todd Shipyards Corp. that Aviall and CERCLA definitively
establish that no section 113(f)(1) contribution action may be brought in the
absence of a prior CERCLA section 106 or section 107 action directly
against the would-be contribution plaintiff.309 The Port of Tacoma court
held that Atlantic Research did not alter the holding in Aviall with regard to
the viability of actions under section 113(f)(1).310 The court reiterated the
Aviall analysis that section 113(f)(1) requires would-be plaintiffs to have
been the subject of a section 106 or section 107 action in order to state a
claim for contribution.311
In BNSF Railway Co. v. California, the defendants counterclaimed
against the state agency for contribution under section 113 seeking to
recover attorney’s fees and to identify other parties that may have been
responsible for contaminating the agency’s property. The federal district
court found that the costs BNSF incurred did not advance the cleanup of the
site and were therefore litigation expenses, which are not recognized as
response costs recoverable under CERCLA.312
308. Id. at 933.
309. See generally Port of Tacoma v. Todd Shipyards Corp., No. C08-5132(BHS), 2009 WL
113852 (W.D. Wash. Jan. 14, 2009).
310. Id. at *3. The Port of Tacoma brought a section 113 contribution action against Todd
Shipyards Corp., seeking to recover costs the Port incurred in carrying out a consent decree it had
entered into with the United States to resolve a section 107 action filed against the Port. Port of Tacoma
v. Todd Shipyards Corp., No. C09-5132BHS, 2008 WL 4454136, at *2–3 (W.D. Wash. Sept. 30, 2008).
Todd Shipyards asserted a third party contribution claim against the government, and the government
moved to dismiss the claim because Todd had not been sued under section 106 or section 107. Port of
Tacoma, 2009 WL 113852, at *3.
311. Port of Tacoma, 2009 WL 113852, at *3. Despite previously concluding that a
contribution action was authorized under Atlantic Research if it “stemmed” from an action instituted
under section 106 or section 107, Judge Settle’s, reconsidered opinion, recognized that Atlantic
Research did not alter the holding in Cooper Industries with respect to the timing of actions under
section 113(f)(1). In order to proceed, such claims must be brought by PRPs “with common liability”
stemming from an action instituted under section 106 or section 107. A contribution plaintiff may not
rely on the mere possibility that it will be held liable under section 106 or section 107, but must actually
be subject to a claim brought under those sections. Id.
312. BNSF Railway Co. v. California, No. 2:08-CV-02225-JAM-JFM, 2009 U.S. Dist. Lexis
2802, at *7 (E.D. Cal. Jan. 6, 2009).
2009] Cost Allocation Liability Conflicts Among the Federal Courts
303
J. Tenth Circuit
The Tenth Circuit has not modified its earlier liability holdings. The
court’s precedent, United States v. Colorado & Eastern Railroad Co.,
remains standing law, holding that if PRPs can bring a claim under section
113(f), they could not bring a claim under section 107(a).313 In Raytheon
Aircraft Co. v. United States, the District Court of Kansas followed Atlantic
Research to conclude that a PRP can demand joint and several liability
under section 107.314 The court held that a non-settling defendant is entitled
to full credit for a plaintiff’s previous settlements, including amounts
received in settling insurance claims arising from liability for response
costs.315 The Tenth Circuit entitles non-settling plaintiffs to broader
flexibility, unless the settlement relates to divisible harm or contains
specific allocation language demonstrating the parties’ intent to venture
from the plain language of CERCLA.316
K. Eleventh Circuit
The Eleventh Circuit also has not made significant changes to its
precedent. The circuit precedent, Redwing Carriers, Inc. v. Saraland
Apartments, is still good law, allowing a PRP to demand joint and several
liability.317 In the interim between Aviall and Atlantic Research, the
Eleventh Circuit found that a PRP who entered into an agreement with the
government could sue another PRP under section 113 because the
agreement fulfills the civil action requirement.318
L. D.C. Circuit
Prior to the initial Aviall decision in 2004, the D.C. Circuit was the only
federal circuit court that had not decided a case on the private party plaintiff
313. United States v. Colo. & E. R.R. Co., 50 F.3d 1530, 1539 (10th Cir. 1995).
314. See Raytheon Aircraft Co. v. United States, 532 F. Supp. 2d 1306, 1309–13 (D. Kan. 2007)
(reasoning that permitting PRPs to pursue cost-recovery claims under section 107(a) would allow PRPs
to “eschew equitable apportionment under § 113(f) in favor of joint and several liability under §
107(a)”).
315. Id.
316. Friedland v. TIC-The Indus. Co., 566 F.3d 1203, 1210–11 (10th Cir. 2009) (describing cost
allocation and the need to clearly indicate any departure from the statutory language).
317. See Redwing Carriers, Inc. v. Saraland Apartments, 94 F.3d 1489, 1514 (11th Cir. 1996), as
recognized in Atl. Research Corp. v. United States, 459 F.3d 827 (8th Cir. 2006) (holding that if plaintiff
attempted to use section 107 to recover more than its fair share of reimbursement, a defendant would be
free to counterclaim for contribution under section 113(f)).
318. Atlanta Gas Light Co. v. UGI Util. Inc., 463 F.3d 1201, 1204 (11th Cir. 2006).
304
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
use of section 107 to recover response costs. Therefore, the court did not
join the cascade of every other federal circuit court, all of which ultimately
were reversed by a unanimous Supreme Court in 2007. It is unclear how
the D.C. Circuit would have ruled had it received a section 107 case to
adjudicate.
However, the D.C. Circuit did receive a section 107 case after Aviall
and before Atlantic Research. In Viacom, Inc. v. United States, a PRP sued
the federal government for recovery of response costs under section
107(a).319 The court held that a PRP who cannot bring a contribution claim
under section 113 may bring a claim to recover cleanup costs under section
107. The D.C. Circuit appears to have been able to anticipate the Supreme
Court decision in Atlantic Research.
CONCLUSION
Thus, the lasting precedential imprint of the 2007 Atlantic Research
decision is that it unknots the entire paralysis in the Superfund program
created by conflating overturned decisions of eleven circuit courts barring
access to section 107, with the Supreme Court’s 2004 Aviall decision
limiting access to section 113 contribution.
The Supreme Court
fundamentally altered the allocation of, and responsibility for, hazardous
substance liability and cost allocation in the United States. Settlement is
necessary before a private-plaintiff PRP can utilize section 113(f)(1) to
reallocate contribution, and the Court removed barriers to private-party use
of section 107 cost recovery. However, the Supreme Court did not decide
whether settlement provides a section 107 plaintiff with a particular level of
protection against a defendant’s counterclaim for equitable contribution
under section 113.
In Atlantic Research, the Court spoke generally about the lower court’s
discretion to apply “traditional rules of equity.”320 However, despite this
general statement in dicta, how the lower courts apply rules of equity will
vary from district to district. The mechanics and incentives for settlement
of Superfund responsibility have changed radically because of the
progression of federal court decisions. As a result of eleven federal circuit
courts blocking private-party PRP access to section 107 cost recovery
319. Viacom, Inc. v. United States, 404 F. Supp. 2d 3, 7 (D.D.C. 2005) (concluding that both a
“plain reading” of section 107 and the Supreme Court’s prior ruling in Key Tronic compelled the
conclusion that a right of cost recovery existed under section 107(a) for those PRPs not subject to
section 113(f) (citing Key Tronic Corp. v. United States, 511 U.S. 809, 818 (1994)).
320. United States v. Atl. Research Corp., 551 U.S. 128, 141 (2007).
2009] Cost Allocation Liability Conflicts Among the Federal Courts
305
between 1994 and 2003, the incentive to settle voluntarily at multi-party
sites was eliminated.321
The Supreme Court’s decision in Aviall further blocked the incentive to
settle because no mechanism for private-party cost contribution was
thereafter available to a PRP unless it had an eligible prior civil action or
settlement.322 Atlantic Research removed the circuit court blockade of cost
recovery under section 107 for PRPs but simultaneously clouded the shield
of section 113(f)(2) previously protecting those who settled with the
government against subsequent section 107 private cost-recovery claims.323
This progression has been a long and often conflicted process. My 1994
article’s324 legal conclusion, which convinced no federal circuit courts
during the 1994 to 2003 cascade of unvarying opinions discouraging
voluntary private party remediation of multi-party contaminated sites, was
endorsed when the Supreme Court unanimously reversed the decisions of
all circuits in Atlantic Research. Both a retrospective look at the legal crisis
created by the sum of the parts, as well as a look forward at how the courts
may internalize the recent Supreme Court determinations, reveal trends in
the administration of cost allocation law.
While the Supreme Court in its 2007 unanimous opinion reinvigorated
CERCLA cost recovery, the long-standing protections afforded by the
statute to settlers were left in limbo. This leaves uncertainties in the basic
mechanics of CERCLA cost allocation. Furthermore, over the two years
after these Supreme Court opinions, there is a checkerboard pattern of the
circuit courts conforming to this new precedent. While the Supreme Court
has provided a clear direction, it remains unclear whether and how the
lower courts will follow.
321. See supra Section II.
322. Cooper Indus. Inc. v. Aviall Servs. Inc., 543 U.S. 157, 167 (2004).
323. Atl. Research Corp., 551 U.S. at 139–41.
324. Steven S. Ferrey, Allocation & Uncertainty in the Age of Superfund: A Critique of the
Redistribution of CERCLA Liability, 3 N.Y.U. ENVTL. L. J. 36 (1994), cited in United States v. Kramer,
953 F. Supp. 592, 600 (D.N.J. 1997).
DIS-JOINTED? SEVERAL APPROACHES TO DIVISIBILITY
AFTER BURLINGTON NORTHERN
Steve C. Gold∗
TABLE OF CONTENTS
Introduction ............................................................................................... 307
I. Divisibility in CERCLA Jurisprudence ................................................ 311
A. Before Burlington Northern ............................................................ 311
B. Burlington Northern......................................................................... 317
II. Joint and Several Liability in the CERCLA Scheme ........................... 322
A. Implementing the Polluter Pays Principle ....................................... 323
B. Facilitating Response Action ........................................................... 326
III. Burlington Northern Provides Little Justification for Apportionment in
Other Factual Settings ....................................................................... 329
A. Separate Ownership Was a Critical Fact in Burlington Northern ... 329
B. Burlington Northern Did Not Relieve Arrangers Seeking
Apportionment of the Burden of Proving Their Shares................. 331
IV. New and Old Legal Theories of Indivisible Harm Continue to
Support Imposition of Joint and Several Liability After Burlington
Northern............................................................................................. 333
Conclusion................................................................................................. 369
INTRODUCTION
The Supreme Court’s sudden interest in the workings of the CERCLA
liability scheme seems somewhat surprising.1 Section 107, the crucial
∗ Assistant Professor of Law, Rutgers School of Law - Newark. J.D., Yale Law School; A.B.
cum laude in Biology, Harvard College. The author formerly was a Senior Attorney in the
Environmental Enforcement Section of the Environment and Natural Resources Division of the United
States Department of Justice. In that capacity, he litigated CERCLA cost recovery actions, but he was
not involved in any phase of the Burlington Northern case. The author gratefully acknowledges the
helpful comment, discussion, and information provided by Bernard Bell, Diane Connolly, Martha Judy,
Randall Stone, Michael Steinberg, Gregory Sukys, and Marc Zeppetello; the research contributed by
research assistant Caitlin Stephens and librarian Susan Lyons; and the enduring patience and support of
his wife, Jennifer Aley.
308
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
provision creating liability for cost recovery and for damages for injury to
natural resources,2 was, after all, revolutionary when it was enacted in 1980.
Very quickly thereafter, lower federal courts were peppered with cases that
required them to interpret and implement CERCLA’s liability provisions.
Petitions for certiorari followed in due course.3 Nonetheless, with few
exceptions,4 the Supreme Court was content—for nearly a quarter century
after CERCLA’s enactment, and nearly two decades after substantial
amendments to the statute—to let the lower federal courts sort out the
mechanics of CERCLA liability.
In a relatively short time, consensus emerged. Among other things, the
courts agreed that section 107 imposed joint and several liability for the
government’s response costs on each liable party, unless a defendant could
satisfy a very difficult burden of proving that it should be assigned only a
portion of the harm.5 But a liable person who paid more in response costs
than equity demanded—whether by undertaking response actions either
voluntarily or under government compulsion, or by reimbursing the
government’s response costs—had a right, initially read into the statute by
courts and then written into section 113(f) by Congress,6 to seek to recover
the excess in a contribution action against other responsible parties. A
liable party could not, however, invoke CERCLA’s cost recovery provision
in an attempt to hold other responsible parties jointly and severally liable.7
These rules became part of the landscape of CERCLA practice. Then, in
less than five years, the Supreme Court issued three seismic opinions in
CERCLA cases, each of which did or could work “an ‘extreme makeover’”8
1. See generally Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870 (2009);
Comprehensive Environmental Response, Compensation, and Liability Act, 42 U.S.C. §§ 9601–9675
(2006).
2. 42 U.S.C. § 9607 (2006).
3. The first cert petition in a CERCLA case was filed on August 1, 1985. See Opening Brief
for Petitioners at 8, Stringfellow v. Concerned Neighbors in Action, 489 U.S. 370 (1987) (No. 85-184).
4. Pennsylvania v. Union Gas Co., 491 U.S. 1, 19 (1989) (holding Congress intended to make
states liable under CERCLA and had constitutional authority to do so), overruled by Seminole Tribe of
Fla. v. Florida, 517 U.S. 44 (1996); Key Tronic Corp. v. United States, 511 U.S. 809, 819 (1994)
(clarifying the circumstances in which private CERCLA plaintiffs could recover their attorneys’ fees);
United States v. Bestfoods, 524 U.S. 51, 62 (1998) (announcing the rule of law that determines when a
corporation’s liability can be vicariously imposed on its parent corporation or other shareholders).
5. See infra Section I.A.
6. 42 U.S.C. § 9613(f) (2006).
7. E.g., Amoco Oil Co. v. Borden, Inc., 889 F.2d 664, 672 (5th Cir. 1989) (“When one liable
party sues another to recover its equitable share of the response costs, the action is one for contribution
. . . .”); Akzo Coatings, Inc. v. Aigner Corp., 30 F.3d 761, 764 (7th Cir. 1994); see also note 15, infra.
8. Michael W. Steinberg, Cloudy with a Chance of Contribution: Life After Cooper Industries,
2 SUPERFUND & NAT. RES. DAMAGES LITIG. COMMITTEE NEWSLETTER 2, 2 (Aug. 2005).
2009] Several Approaches to Divisibility After Burlington Northern
309
of an aspect of the CERCLA liability scheme that government, industry,
and the courts had come to take for granted.
The first case was Cooper Industries, Inc. v. Aviall Services, Inc.9
Cooper Industries upheld the judiciary’s belated discovery10 of the “plain
meaning” of CERCLA’s contribution provision: that when Congress
granted a right to parties to seek contribution either “during or following” a
civil enforcement action11 or once they had “resolved [their] liability to the
United States or a State” in a settlement,12 Congress meant those
circumstances to be the only ones in which a party could invoke the
statutory contribution remedy. Thus, a responsible party that conducted
response actions in the absence of a qualifying lawsuit or settlement could
not bring a CERCLA contribution claim.
Cooper Industries did not decide whether such a party could sue for
cost recovery under section 107, but lower court precedent seemed to
foreclose that option.13 Therefore, many parties seemed to have lost legal
recourse that they thought they had against other liable persons, and many
commentators concluded that as a result Cooper Industries had seriously
undermined incentives for voluntary cleanup of contaminated sites.14 After
two and a half years in which the lower courts attempted, to varying
degrees and in diverse ways, to restore those incentives, the Supreme Court
decided that section 107(a)(2) of CERCLA authorizes responsible parties to
bring cost recovery actions after all.15 Once again the Supreme Court held
9. Cooper Indus. Inc. v. Aviall Servs. Inc., 543 U.S. 157 (2004).
10. As the majority of the en banc Fifth Circuit put it: “In numerous published cases decided
after the enactment of SARA in 1986, this and other courts of appeals have ruled on CERCLA claims
for contribution where no [prior civil enforcement] action had been brought . . . . [O]ne must assume
that talented attorneys have had sufficient incentive and opportunity to explore statutory lacunae such as
those created by a cramped reading of § 113(f)(1). Yet all that existed before this case arose are isolated
dicta. The absence of direct precedent is like the dog that didn’t bark.” Aviall Servs. Inc. v. Cooper
Indus. Inc., 312 F.3d 677, 688–89 (5th Cir. 2002) (en banc) (footnotes omitted), rev’d, 543 U.S. 157, 171
(2004).
11. § 9613(f)(1).
12. § 9613(f)(2).
13. Steven Ferrey, Inverting the Law: Superfund Hazardous Substance Liability and Supreme
Court Reversal of All Federal Circuits, 33 WM. & MARY ENVTL. L. & POL’Y REV. 633, 663–64 (2009)
(citing cases).
14. E.g., Michael Mazzone et al., New Supreme Court Decision Discourages Voluntary
Clean-up Under CERCLA, 42 HOUSTON LAW. 52 (Jan./Feb. 2005); Jeannette Paul, Neither Innocent nor
Proven Guilty: The Aviall Services v. Cooper Industries Dilemma, 13 BUFF. ENVTL. L.J. 31, 56–57
(2005).
15. United States v. Atl. Research Corp., 551 U.S. 128, 139 (2007). Ironically, this was the
conclusion reached in what appears to be the first CERCLA decision published. City of Philadelphia v.
Stepan Chem. Co., 544 F. Supp. 1135 (E.D. Pa. 1982) (denying motion to dismiss). After the plaintiff in
City of Philadelphia cleaned up a hazardous substance release at a municipal dump, it sued several
defendants for cost recovery, alleging that the defendants had arranged for disposal of hazardous
310
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
that the lower courts had missed the meaning of CERCLA’s “plain terms.”16
And once again the Supreme Court left a major question, previously
thought settled, unanswered: “We assume without deciding[,]” the Court
wrote in a footnote, “that § 107(a) provides for joint and several liability.”17
In May, 2009 the Supreme Court answered that question in Burlington
Northern and Santa Fe Railway Co. v. United States.18 The Supreme Court
decided that for once the lower courts got it right: section 107(a) does
indeed provide for joint and several liability. But, as the lower courts also
had decided, there is an out: a liable defendant might limit its liability to
several liability for an apportioned share, if it were to provide proof
sufficient to support such apportionment. What it takes to satisfy that
burden was the subject of Burlington Northern.19 What Burlington
Northern should mean to the courts that must apply it, and to the future of
joint and several liability under CERCLA, is the subject of this article.
Interested parties’ first-blush response to the Supreme Court’s opinion
implies that lower courts will not find proper application of Burlington
Northern to be obvious or easy. The Department of Justice, emphasizing
the Court’s matter-of-fact agreement with the legal standard that lower
courts have long applied to CERCLA scope-of-liability disputes,20 contends
that the opinion illustrates that plus ça change, plus c’est la même chose.21
By contrast, the CERCLA defense bar claims that the opinion “reverses a
substances at the facility. The defendants argued that because the city, as owner of the dump, was also a
responsible party, the city could not bring a cost recovery suit under section 107. At the time, neither
case law nor statutory amendment had established any right of responsible parties to seek contribution.
The court denied the motion, keeping the city’s CERCLA claim alive. Id. at 1143. The court’s
reluctance to dismiss the city’s claim was influenced by the fact that the city owned and operated a
hazardous waste dump only unwillingly: the hazardous substances had been deposited illegally by waste
transporters who obtained access to the dump by bribing city employees. Id.
16. Atl. Research Corp., 551 U.S. at 141.
17. Id. at 140 n.7.
18. Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870 (2009).
19. Burlington Northern also addressed the issue of arranger liability under CERCLA. Id. at
1880. That holding is not addressed here.
20. Id. at 1880–81.
21. See, e.g., John C. Cruden, Acting Assistant Attorney Gen., Envtl. & Natural Res. Div., The
Supreme Court’s Decision in Burlington Northern & Santa Fe Railway Company v. United States (May
29, 2009), http://www.usdoj.gov/enrd/1306.htm (stating: (1) the “opinion clearly reaffirmed the basic
principles that we have advocated for years”; (2) with respect to apportionment, the decision “is
predominantly factual”; and (3) “[w]e think that [meeting burden of proof on apportionment] will
continue to be difficult to do in most cases”); United States’ Reply Memorandum in Support of Its
Renewed Motion for Summary Judgment on the Joint and Several Liability of Defendant Saul Senser
29, United States v. Atlas Lederer Co., No. 3:91cv309 (S.D. Ohio June 1, 2009) (arguing that because
Burlington Northern “did not change the governing legal principles,” courts should adhere to prior
rejection of apportionment based on evidence constituting “no more than a ‘best guess’” of each
responsible party’s contribution to the site).
2009] Several Approaches to Divisibility After Burlington Northern
311
long-standing presumption by the Environmental Protection Agency in
favor of joint and several liability in multiple-party [CERCLA] cases.”22
Some observers argue that Burlington Northern so severely and wrongly
diminished the proof required to avoid joint and several liability that
legislative action is required to correct the decision’s effects.23
I contend that the Burlington Northern decision should have relatively
limited impact on CERCLA litigation, if federal courts understand and
apply the Supreme Court’s opinion properly. Section I describes the
jurisprudence of joint and several liability under CERCLA that preceded
Burlington Northern, and it also describes the Burlington Northern decision
itself. Section II describes the statutory purposes courts must keep in mind
when applying Burlington Northern in CERCLA cases and, in particular,
the role of joint and several liability in fulfilling those purposes. In Section
III, I argue that Burlington Northern’s ruling on the sufficiency of evidence
of apportionment should not, and likely will not, be reflected in an upsurge
of successful proof of apportionment in commonly encountered CERCLA
fact patterns. In Section IV, I argue that a number of viable legal theories
should preserve joint and several liability under CERCLA even if
Burlington Northern has the effect of greatly diminishing the evidentiary
hurdles that had in the past largely stymied attempts at apportionment.
I. DIVISIBILITY IN CERCLA JURISPRUDENCE
A. Before Burlington Northern
From the outset, the government sought to hold defendants jointly and
severally liable as it grappled with cost recovery claims for which dozens of
22. Cost Recovery: Burlington Northern Decision Called New Path to Fairness Under
CERCLA 1, TOXICS L. RPTR. (BNA), May 21, 2009 (referring to comments by members of firm that
represented Burlington Northern); see, e.g., Allen A. Kacenjar et al., Burlington Northern and Santa Fe
Railway Co. v. United States, June 11, 2009, http://www.martindale.com/environmentallaw/article_Squire-Sanders-Dempsey-L.L.P._703524.htm (proposing that as a result of the Burlington
Northern decision, apportionment will be more available, more defendants will argue for apportionment,
and defendants will have increased leverage in settlement negotiations with government).
23. Mark R. Misiorowski & Joel D. Eagle, The Diminishing Role of Science in CERCLA After
Burlington Northern & Santa Fe, 40 ENV’T REP. (BNA) 1205, 1205 (May 22, 2009); see generally
Appleton Papers, Inc. v. George A. Whiting Paper Co., No. 08-C-16, 2009 U.S. Dist. LEXIS 111648, at
*8–9 (E.D. Wis. Nov. 18, 2009) (describing Burlington Northern as a “watershed” case that
“significantly eases the burden on defendants who seek to avoid joint and several liability”); Evansville
Greenway & Remediation Trust v. Southern Ind. Gas & Elec. Co., Inc., No. 3:07-cv-0066-DFH-WGH,
2009 U.S. Dist. LEXIS 95091, at *66 (S.D. Ind. Sept. 29, 2009) (noting “full import of Burlington
Northern is hotly debated”).
312
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
parties were potentially responsible under section 107.24 In the firstimpression opinion that has ever since been considered “seminal” on the
issue,25 Chief Judge Carl Rubin of the Southern District of Ohio confronted
defendants’ motion “for an early determination that they are not jointly and
severally liable for the clean-up costs”26 incurred in connection with the
Chem-Dyne Superfund site in Hamilton, Ohio.27
Judge Rubin found CERCLA’s “language ambiguous with regard to the
scope of liability.” He therefore consulted the legislative history and
concluded:
● Although earlier versions of the bill that became CERCLA included
language mandating joint and several liability that was deleted before final
passage, the deletion did not imply that Congress intended liability to be
only several.
● “Rather, the term was omitted in order to have the scope of liability
determined under common law principles, where a court performing a case
by case evaluation of the complex factual scenarios associated with
multiple-generator waste sites will assess the propriety of applying joint and
several liability on an individual basis.”
● The common law principles to be applied should be uniform and
federal, rather than borrowed from the law of the forum state.
Judge Rubin rejected the government’s invitation to fashion a common
law rule of mandatory joint and several liability. Instead, Judge Rubin
borrowed a framework from the Second Restatement of Torts:
24. The government sought to hold parties joint and severally liable for Love Canal, which
predated CERLCA and was the impetus for it. See United States v. Hooker Chem. & Plastics Corp., 607
F. Supp. 1052, 1053 n.2 (W.D. N.Y. 1985) (case involving Hooker dumps in Niagara Falls, including
Love Canal, filed in 1979). Hooker presented the question of the scope of liability of multiple
responsible parties. Id. at 1053. Although all of the millions of pounds of chemical waste at Love Canal
were put there by Hooker Chemical while Hooker Chemical owned the site, see United States v. Hooker
Chem. & Plastics Corp., 680 F. Supp. 546, 549, 556 (W.D. N.Y. 1988) (holding Hooker’s successor
jointly and severally liable), the City of Niagara Falls was also held liable because it acquired the
property from Hooker for nominal consideration and used it, among other things, as the location of a
school. United States v. Hooker Chem. & Plastics Corp., 965 F. Supp. 408, 413, 416 (W.D. N.Y. 1997)
(city liable under CERCLA, divisibility of harm to be determined in apportionment phase).
25. H.R. REP. NO. 253, pt. 1, at 74 (1985), reprinted in 1986 U.S.C.C.A.N. 2835, 2856 (“The
Committee fully subscribes to the reasoning of the court in the seminal case of United States v. ChemDyne Corp., 572 F. Supp. 802 (S.D. Ohio 1983), which established a uniform federal rule allowing for
joint and several liability in appropriate CERCLA cases.”); Burlington Northern, 129 S. Ct. at 1880
(“The seminal opinion on the subject of apportionment in CERCLA actions was written in 1983 . . . .”).
26. Chem-Dyne Corp., 572 F. Supp. at 804 (noting that “[a]t present, there is no case authority
specifically addressing” scope of liability under CERCLA). The United States alleged claims against
twenty-four defendants in Chem-Dyne. Id.
27. See 40 C.F.R. pt. 300, App. B (2008) (National Priorities List including and giving location
of Chem-Dyne site).
2009] Several Approaches to Divisibility After Burlington Northern
313
An examination of the common law reveals that when two
or more persons acting independently caused 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. Restatement (Second) of Torts,
§§ 433A, 881 (1976) . . . . But where two or more persons
cause a single and indivisible harm, each is subject to
liability for the entire harm. Restatement (Second) of
Torts, § 875 . . . . Furthermore, where the conduct of two
or more persons liable under § 9607 has combined to
violate the statute, and one or more of the defendants seeks
to limit his liability on the ground that the entire harm is
capable of apportionment, the burden of proof as to
apportionment is upon each defendant. Id. at § 433B . . . .28
Chem-Dyne thus presented a rare opportunity for a federal district judge
to announce a rule of law entirely in the abstract, a “judicial interpretation
of the nature and scope of liability under 42 U.S.C. § 9607” that was
“intended to assist the parties in expediting discovery and trial
preparation.”29
Judge Rubin predicted that the rule would prove
serviceable: “These rules clearly enumerate the analysis to be undertaken
when applying 42 U.S.C. § 9607 . . . .”30 But he never got to perform that
analysis in a concrete factual setting: the case settled less than two years
after Judge Rubin issued his opinion.31
Other district courts quickly adopted the Chem-Dyne analysis.32 The
courts of appeals unanimously followed suit, agreeing on the basic “federal
28. Chem-Dyne Corp., 572 F. Supp. at 810 (selected citations omitted).
29. Id. at 811.
30. Id. at 810–11.
31. See Proposed Consent Decrees in Action Under the Comprehensive Environmental
Response, Compensation and Liability Act and the Resource Conservation and Recovery Act to Require
Defendants to Reimburse the United States for Response Costs and to Complete Cleanup of the ChemDyne Hazardous Waste Site in Hamilton, OH, 50 Fed. Reg. 25,797 (June 21, 1985) (announcing and
soliciting public comment on proposed settlements related to Chem-Dyne site). In one very narrow
sense, Judge Rubin did apply the legal rule to the facts of the case: treating the defendants’ motion as a
motion for partial summary judgment that they were not jointly and severally liable; the court held that
in light of their burden of proof the defendants had not demonstrated the absence of genuine issues of
material fact. Chem-Dyne Corp., 572 F. Supp. at 810. It is clear, however, that because the defendants
argued that CERCLA did not provide for joint and several liability regardless of the facts, they had made
no attempt to build a record that might have satisfied their burden.
32. See e.g., United States v. Wade, 577 F. Supp. 1326, 1337 (E.D. Pa. 1983) (adopting ChemDyne rule and denying plaintiff’s summary judgment motion because facts not sufficiently developed);
United States v. Conservation Chem. Co., 589 F. Supp. 59, 63 (W.D. Mo. 1984) (adopting Chem-Dyne
rule and denying defendants’ summary judgment motion); United States v. S.C. Recycling & Disposal,
314
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
common law” principles that would govern the scope of liability in
CERCLA cases.33 Those principles, the courts agreed, derive primarily
from section 433A of the Second Restatement of Torts (hereinafter referred
to simply as “section 433A”):
§ 433 A. Apportionment of Harm to Causes
Inc., 653 F. Supp. 984, 994–95 (adopting Chem-Dyne legal analysis and granting plaintiff summary
judgment because undisputed facts showed that harm was indivisible); United States v. Argent Corp., 14
Envtl L. Rep. 20, 616 No. 83-0523 (D.N.M. May 4, 1984) (denying plaintiff’s summary judgment
motion because of disputed fact issues). A few early cases adopted the view that equitable factors,
rather than the causal principles of the Restatement, could provide a basis for apportionment in lieu of
joint and several liability. United States v. A&F Materials Co., 578 F. Supp. 1249, 1256 (S.D. Ill. 1984)
(rejecting “rigid” application of Restatement formula to allow court to apportion by applying equitable
“Gore factors”); United States v. Stringfellow, 14 Envtl L. Rep. 20,385 No. CV-83-2501-MML, (C.D.
Cal. Apr. 5, 1984) (suggesting federal courts may impose joint and several liability). The courts of
appeals decisively rejected that analysis. See United States v. Twp. of Brighton, 153 F.3d 307, 318–20
(6th Cir. 1998) (divisibility inquiry based on causal apportionment distinct from equitable inquiry
involving normative fault); United States v. Monsanto Co., 858 F.2d 160, 171 n.22 (4th Cir. 1988)
(equitable factors are relevant in subsequent actions for contribution but not to the question of joint and
several liability, which focuses on divisibility); United States v. Alcan Aluminum Corp. (Alcan-Butler),
964 F.2d 252, 270 n.29 (3d Cir. 1992) (“[T]he contribution proceeding is an equitable one in which a
court is permitted to allocate response costs based on factors it deems appropriate, whereas the court is
not vested with such discretion in the divisibility determination.”). Later, the Supreme Court did the
same. Burlington Northern, 129 S. Ct. at 1882 n.9.
33. United States v. Capital Tax Corp., 545 F.3d 525, 535–36 (7th Cir. 2008) (applying section
433A analysis, holding defendant failed to establish basis for apportionment); Chem-Nuclear Sys. v.
Bush, 292 F.3d 254, 259–61 (D.C. Cir. 2002) (applying Section 433A analysis, affirming holding that
responsible party failed to prove reasonable basis for apportioning harm and therefore could not obtain
partial reimbursement of response costs from Superfund); Carson Harbor Vill., Ltd. v. Unocal Corp.,
270 F.3d 863, 870 (9th Cir. 2001) (articulating Chem-Dyne rule without citing Chem-Dyne); United
States v. Hercules, Inc., 247 F.3d 706, 717 (8th Cir. 2001) (applying Chem-Dyne analysis, reversing
summary judgment that defendant was jointly and severally liable, because of disputed issues of fact);
United States v. Burlington N. R.R. Co., 200 F.3d 679, 697 (10th Cir. 1999) (following Chem-Dyne,
holding that district court that found harm was geographically divisible correctly apportioned associated
settlement credit); Redwing Carriers, Inc. v. Saraland Apartments, 94 F.3d 1489, 1513 (11th Cir. 1996)
(approving Chem-Dyne test while distinguishing apportionment of liability from equitable contribution);
In re Bell Petroleum Servs., 3 F.3d 889, 901–04 (5th Cir. 1993) (applying Restatement analysis similar
to Chem-Dyne, reversing judgment of joint and several liability entered after trial because where
different defendants operated same plant emitting same pollutant for successive periods, evidence of
duration and volume of operation provided reasonable basis for apportionment); United States v. Alcan
Aluminum Corp., 990 F.2d 711, 722 (2d Cir. 1993) (adopting Chem-Dyne approach, reversing plaintiff’s
summary judgment for consideration of disputed factual issues about whether defendant’s hazardous
substance contributed to environmental harm and if so whether contribution was divisible); AlcanButler, 964 F.2d at 268–69 (following Chem-Dyne, reversing plaintiff’s summary judgment for hearing
on disputed fact issues); United States v. R.W. Meyer, Inc., 889 F.2d 1497, 1507 (6th Cir. 1989)
(following Chem-Dyne, affirming summary judgment that defendant was jointly and severally liable,
because finding of indivisible harm not clearly erroneous); O’Neil v. Picillo, 883 F.2d 176, 178 (1st Cir.
1989) (adopting Chem-Dyne approach, affirming judgment of joint and several liability against
arrangers after trial); Monsanto Co., 858 F.2d at 171–72 (affirming judgment of joint and several
liability).
2009] Several Approaches to Divisibility After Burlington Northern
315
(1) Damages for harm are to be apportioned among two or
more causes where
(a) there are distinct harms, or
(b) there is a reasonable basis for determining the
contribution of each cause to a single harm.
(2) Damages for any other harm cannot be apportioned
among two or more causes.34
To apply those rules, courts struggled to figure out what terms like
“damages” and “harm” mean in the statutory context of CERCLA liability
for cost recovery or performance of response actions.35 The most frequent
and most challenging cases involved claims that “a single harm” could be
apportioned on a reasonable basis by “determining the contribution of each
cause.” Following the terminology of a Restatement comment, courts came
to refer to the issue as “divisibility.”36
Courts held, again following the Restatement, that liable defendants
bear the burden of proof on divisibility in order to avoid joint and several
liability.37 The courts found the divisibility inquiry to be “intensely
factual.”38 It required “a fairly complex factual determination.”39 And,
although “‘commingled’ waste is not synonymous with ‘indivisible’
harm,”40 for a typical Superfund site featuring “numerous, commingled
34. RESTATEMENT (SECOND) OF TORTS § 433A (1966). References in the text to the
“Restatement” refer to RESTATEMENT (SECOND) OF TORTS unless the context clearly indicates another
meaning.
35. See, e.g., O’Neil, 883 F.2d at 180 (expressing doubt as to whether “harm,” for divisibility
purposes, consists of response costs, of environmental contamination that actually occurred, or of
environmental contamination averted by response action); see also United States v. Burlington N. &
Santa Fe Ry. Co., 520 F.3d 918, 938–39 (9th Cir. 2008) (discussing possible interpretations of “harm” in
CERCLA cases).
36. See RESTATEMENT (SECOND) OF TORTS § 433A cmt. d (1966) (stating that certain kinds of
harm not clearly severable are still capable of division).
37. Chem-Dyne Corp., 572 F. Supp. at 810 (citing RESTATEMENT (SECOND) OF TORTS § 433B
(1966)).
38. Alcan-Butler, 964 F.2d at 269; see, e.g., Hercules, 247 F.3d at 717 (“[A]pproaches to
divisibility will vary tremendously depending on the facts and circumstances of each case.”).
39. Chem-Dyne Corp., 572 F. Supp. at 811.
40. Alcan-Butler, 964 F.2d at 270 n.29; accord United States v. Alcan Aluminum Corp., 990
F.2d 711, 722 (2d Cir. 1993) (“[C]ommingling is not synonymous with indivisible harm . . . .”);
Hercules, 247 F.3d at 718 (“‘[C]ommingling is not synonymous with indivisible harm.’”) (quoting
Alcan, 990 F.2d at 722); cf. O’Neil, 883 F.2d at 183 n.11 (“Because there was substantial commingling
of wastes [in contaminated soil], we think that any attempt to apportion the costs incurred by the state in
removing the contaminated soil would necessarily be arbitrary.”).
316
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
hazardous substances with synergistic effects and unknown toxicity,”41
proving divisibility was “a very difficult proposition.”42 As a result,
defendants “rarely succeed”43 in meeting their burden. The failure of proof
variously consisted of an inability to segregate the harm caused by a
defendant’s waste amid a toxic soup of potentially interacting hazardous
substances,44 or an inability to quantify a defendant’s volumetric or
geographic contribution as a proportion of the total waste,45 or both.
Until Burlington Northern, in only one significant CERCLA case—Bell
Petroleum—had a court clearly held that a defendant had established a
reasonable basis for apportioning a single harm The facility and the
hazardous substance release in Bell Petroleum, however, were quite
different from the prototypical CERCLA dump site. The contamination to
which the government responded consisted of a single hazardous substance
(chromium) that originated from industrial operations at a single plant,
although it had spread from that plant in groundwater. The plant had been
operated, in succession, by three different manufacturers. All of them
conducted essentially the same operations that resulted in chromium
reaching the aquifer. On these facts, the Fifth Circuit held as a matter of
law that the harm was at least theoretically capable of apportionment.46 As
a factual matter, it also held that the defendants had presented sufficient
41. In re Bell Petroleum Servs., 3 F.3d 889, 903 (5th Cir. 1993).
42. Control Data Corp. v. S.C.S.C. Corp., 53 F.3d 930, 934 n.4 (8th Cir. 1995); see, e.g., AlcanButler, 964 F.2d at 269 (“Alcan’s burden in attempting to prove the divisibility of harm . . . is
substantial, and the analysis will be factually complex. . . . .”). Although the Third Circuit, in AlcanButler, held that the district court had improperly granted summary judgment to the plaintiff, on remand
the district court found after a hearing that Alcan had failed to meet its burden, United States v. Alcan
Aluminum Corp., 892 F. Supp. 648, 648 (M.D. Pa. 1995), and the Third Circuit affirmed without
opinion. 96 F.3d 1434 (1996). The Second Circuit similarly affirmed Alcan’s joint and several liability
after remand. Alcan, 990 F.2d at 711.
43. Bell Petroleum, 3 F.3d at 901; accord United States v. Capital Tax Corp., 545 F.3d 525, 535
n.9 (7th Cir. 2008) (“[D]ivisibility is a ‘rare scenario.’”); O’Neil, 883 F.2d at 178–79 (“[P]ractical effect
of placing the burden on defendants has been that responsible parties rarely escape joint and several
liability . . . .”).
44. E.g., United States v. Alcan Aluminum Corp. (Alcan-Fulton), 315 F.3d 179, 187 (2d Cir.
2003); United States v. Alcan Aluminum Corp., 892 F. Supp. 648, 656–57 (M.D. Pa. 1995), aff’d mem.,
96 F.3d 1434 (3d Cir. 1996); Monsanto, 858 F.2d at 172–73. (“In this case, however, volume could not
establish the effective contribution of each waste generator to the harm at the Bluff Road site.”).
45. E.g., Capital Tax, 545 F.3d at 535–36; United States v. Vertac Chem. Corp., 453 F.3d 1031,
1047 (8th Cir. 2006); Alcan-Fulton, 315 F.3d at 186; Chem-Nuclear, 292 F.3d at 261 (“In short, while
CWM produces some circumstantial evidence to support its theory of geographic divisibility, it has not
managed the ‘very difficult proposition’ of proving its theory by a preponderance of the evidence.”);
O’Neil, 883 F.2d at 182.
46. Bell Petroleum, at 902–03 (“‘[T]he question whether the harm to the plaintiff is capable of
apportionment among two or more causes is a question of law.’”) (quoting RESTATEMENT (SECOND) OF
TORTS § 434 (1966)). The court distinguished cases involving chemical soups presenting possible
synergistic effects. Id. at 903.
2009] Several Approaches to Divisibility After Burlington Northern
317
documentary and testimonial evidence from which the trier of fact could
have determined the relative contribution of each defendant to the harm.47
The court of appeals remanded to the district court with instructions to
make that determination.48
B. Burlington Northern
The facility at issue in Burlington Northern was the Brown & Bryant
Superfund Site in Arvin, California. Brown & Bryant, Inc. (B&B) ran an
agricultural chemical distribution business there; as in Bell Petroleum, the
site was not a dump for waste from off-site manufacturers.49 B&B
originally conducted its operations on property it owned (the “B&B
Parcel”), but it later leased from two railroad companies (the “Railroads”) a
smaller adjacent parcel that the Railroads jointly owned (the “Railroad
Parcel”).50 B&B’s sloppy operation resulted in soil and groundwater
contamination with at least three hazardous substances: Dinoseb, Nemagon,
and D-D, the last two of which were sold to B&B by the Shell Oil
Company (“Shell”).51
The United States and the State of California eventually conducted
removal and remedial action in response to the release of hazardous
substances at the Arvin Site and sued for cost recovery. By the time the
cost recovery suit was tried, B&B had long since disappeared into
insolvency. The defendants at trial were the Railroads, allegedly liable as
current owners52 and as owners at the time of disposal,53 and Shell,
allegedly liable as an operator at the time of disposal54 and as an arranger.55
1. The District Court Decision
District Judge Oliver W. Wanger lived with the Brown & Bryant
Superfund site for a decade. He tried the governments’ cost recovery case
over a period of more than two months, and he issued his very long
47.
48.
49.
50.
51.
52.
53.
54.
55.
parties).
Id. at 904.
Id. at 889.
Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870, 1874 (2009).
Id.
Id. at 1875.
42 U.S.C. § 9607(a)(1) (2006).
§ 9607(a)(2).
See id.
See id. § 9607(a)(3) (persons who arranged for disposal or treatment as responsible
318
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
decision four years later.56 Judge Wanger concluded that the Railroads were
liable as current owners and owners at the time of disposal, and that Shell
was liable as an arranger but not as an operator.
The Railroads and Shell each had argued that it should not be held
liable at all. Neither defendant had pressed a theory of divisibility.57
Nevertheless, Judge Wanger felt that “the circumstances of the case”
required apportionment, so he sua sponte undertook a search of the record
for evidentiary support.58
There can be no question that concerns for equity drove Judge Wanger
to conduct this “heroic labor.”59 In the course of holding the Railroads
liable, Judge Wanger noted that CERCLA makes facility owners liable
regardless of whether they were “responsible for the actual disposal,”60 thus
ensuring “that at least some entity other than the taxpayers will be
potentially available, in the first instance, to redress the environmental
harm.”61 Tellingly, however, immediately after the holding the Railroads
liable, Judge Wanger added an additional and, in the context, apparently
gratuitous conclusion of law: “The volume of the hazardous substance
releasing activities on the B&B site is at least ten times greater than any
Railroad parcel releases.”62 The presence, in the court’s liability discussion,
of that curiously-worded conclusion of law—which began by referring to a
volume of activity and ended by referring to a volume of releases of
hazardous substances—illustrates Judge Wanger’s discomfort with the
“potentially unfair” consequences of CERCLA’s liability scheme.63 The
judge noted, however, that such “concerns are more appropriately addressed
in the harm (apportionment) stage.”64
56. United States v. Atchison, Topeka & Santa Fe Ry. Co., No. CV-F-92-6058, 2003 U.S. Dist.
LEXIS 23130, at *4 (E.D. Cal. July 14, 2003).
57. Id. at *236.
58. Id. at *237.
59. Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870, 1885 (2009)
(Ginsburg, J., dissenting).
60. Atchison, Topeka & Santa Fe Ry. Co., 2003 U.S. Dist. LEXIS 23130, at *141 (¶ 324a); see
id. at *142 (¶¶ 324b–c) (“Although owners may not actually engage in disposal of hazardous substances,
they are still deemed polluters for allowing ultrahazardous activity to occur on their property.”).
61. Id. at *143 (¶ 324d).
62. Id. at *146. Elsewhere, Judge Wanger made many factual findings that would support a
qualitative inference that the magnitude of releases on the B&B Parcel far exceeded that of the releases
on the Railroad Parcel. E.g., id. at *24–25, *33–35, *253–55.
63. Id. at *157 (¶ 352).
64. Id. at *159 (¶ 353); see id. at *237 (¶ 455) (parties, by declining to offer evidence on
apportionment, “left the court to independently perform the equitable apportionment analysis”).
2009] Several Approaches to Divisibility After Burlington Northern
319
In the apportionment stage, the court held that the case involved a
single “harm”—the plume of contaminated groundwater.65 The court
reasoned generally that this plume “predominately” consisted of chemicals
originally spilled on the B&B Parcel; the evidence showed that the volume
of spills resulting from B&B’s activities on the Railroad Parcel, although
“incalculable,” had been much smaller.66 To produce a quantitative
apportionment, Judge Wanger multiplied three fractions: the area of the
Railroad Parcel as a fraction of the total area of both parcels (19%), the
number of years during which B&B operated on the Railroad Parcel as a
fraction of the total number of years of B&B’s operation (0.45), and the
amount of site contamination attributed to the chemicals spilled on the
Railroad Parcel that could have contributed to the groundwater plume
(2/3rds).67 The product equaled approximately 6%, which the judge
increased by one-half, to 9%, to account for possible “calculation errors.”68
2. The Ninth Circuit Decision
The court of appeals reversed Judge Wanger’s apportionment of the
Railroads’ liability.69 In the course of determining the standard of review,
the Ninth Circuit held that the divisibility issue comprised two components:
first, whether “the particular harm at issue in the case is theoretically
capable of apportionment,” (i.e., divisible or indivisible) which is a question
of law;70 and second, if so, whether the defendant seeking apportionment
submitted sufficient evidence to establish a reasonable basis for it.71
The court of appeals agreed with the district court that the case
involved a single harm: “the contamination on the Arvin site.”72
65. Id. at *250–51 (¶ 472). The court included contaminated overlying soil in the single harm
because contamination could leach from the soil into the groundwater. Id.
66. Id. at *238 (¶ 458); see id. at *146, *254, *259 (referring to the spills in quantifiable
terms).
67. Id. at *260 (¶ 489). All three chemicals were found in soil on the Railroad Parcel, but the
court concluded that the “slight” D-D contamination could be “offset” based on other facts tending to
indicate that the D-D spilled there would not have required remediation. Id. at *259 (¶ 488). Of the
remaining two chemicals, the court wrote: “Estimates are that these two chemicals contributed to 2/3 of
overall Site contamination.” Id. at *259 (¶ 489).
68. Id. at *260 (¶ 489). Judge Wanger also apportioned Shell’s liability, assigning it a six
percent share (without any adjustment for calculation errors). Id. at *267 (¶ 498).
69. The Ninth Circuit’s original decision, United States v. Burlington N. & Santa Fe Ry. Co.,
479 F.3d 1113 (9th Cir. 2007), was amended after a petition for rehearing en banc. United States v.
Burlington N. & Santa Fe Ry. Co., 520 F.3d 918, 952 (9th Cir. 2008). The court denied the petition for
rehearing en banc over an eight-judge dissent. Id.
70. Burlington Northern, 520 F.3d at 942.
71. Id.
72. Id. at 943.
320
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
Accordingly, the court held that apportionment could be justified only if
defendants established, per section 433(A)(1)(b), a reasonable basis for
determining the contribution of each cause to that harm.73 The Ninth
Circuit found “no dispute” on the “purely legal question—whether the harm
is capable of apportionment.”74 The court of appeals also agreed with the
district court that proof of the portion of “contaminants found on the
[entire] Arvin parcel . . . attributable to the presence of toxic substances or
to activities on the Railroad parcel” would provide a reasonable basis for
apportionment.75 But it found clearly erroneous the district court’s finding
that such proof was in the record, noting that the Railroads presented no
evidence that the contribution of hazardous substances from their parcel
was in any way proportional to the parcel’s area, the number of years of
activity, or the number of chemicals involved. Therefore, the court of
appeals held, the district court’s calculations “bore insufficient logical
connection to the pertinent question.”76
3. The Supreme Court Opinion
The Supreme Court reversed, reinstating the district court’s
apportionment.77 The Supreme Court found no error in the Ninth Circuit’s
analysis of the legal standards for divisibility.78 The Court endorsed the
Chem-Dyne analysis and, apparently, the decisions of the courts of appeals
that “fully embraced” it.79 Quoting section 433A(1)(b), the Court stated
that “apportionment is proper when there is a reasonable basis for
determining the contribution of each cause to a single harm,” and agreed
that the party seeking apportionment bears the burden of proof on the
issue.80 “Neither the parties nor the lower courts dispute [these] principles,”
the Court noted.81
73. Id. at 939.
74. Id. at 942.
75. Id. at 946.
76. Id.
77. Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870, 1883 (2009).
78. In fact, the Supreme Court barely mentioned the Ninth Circuit’s opinion in this context.
See id. at 1880–81 (mentioning little of the Ninth Circuit’s opinion and instead discussing Chem-Dyne
and citing only decisions of other federal appellate courts).
79. Id. at 1881.
80. Id. (quoting RESTATEMENT (SECOND) OF TORTS § 433A(1)(b) (1963–64)). The opinion
does not even note the existence of the Third Restatement on apportionment of liability, which reflects
the widespread legislative modification of joint and several liability by comparative fault or comparative
responsibility statutes. See generally RESTATEMENT (THIRD) OF TORTS: APPORTIONMENT OF LIABILITY
§ 1 (2000). The Third Restatement describes regimes of pure joint and several liability, pure
apportionment, and several hybrids without expressing an ultimate preference for any. Id. § 17. It
suggests that its principles should inform interpretation of CERCLA and other liability-imposing
2009] Several Approaches to Divisibility After Burlington Northern
321
Both “the District Court and Court of Appeals agreed that the harm
created by the contamination . . . although singular, was theoretically
capable of apportionment,” the Court continued.82 It then reviewed the
district court’s apportionment and concluded that “facts contained in the
record reasonably supported the apportionment.”83 The Supreme Court
seemed heavily influenced by the qualitative evidence that made it
“abundantly clear that the primary pollution” at the site was deposited on
the B&B Parcel,84 and appeared to accept without question the district
court’s inference from that qualitative evidence that “spills . . . on the
Railroad parcel contributed to no more than 10% of the total site
contamination.”85 “With those background facts in mind,” the Supreme
Court was persuaded that it was reasonable for the district court to use the
relative area of the parcels and the relative number of years of operation as
the “starting point for its analysis.”86 Coincidentally, the relative area of the
parcels multiplied by the number of years of operation equals 9%—the
same percentage the district court ultimately apportioned to the Railroads,
and a close approximation of “no more than 10%.”87
The importance of this coincidence became apparent as the Supreme
Court considered the last factor in the district court’s apportionment
formula, 2/3rds, to represent the harm caused by the two chemicals spilled
in relatively large quantity on the Railroad Parcel. The district court’s
“conclusion that those two chemicals accounted for only two-thirds of the
contamination requiring remediation,” the Court understated, “finds less
support in the record.”88 But never mind—the district court’s 50% fudge
factor and its erroneous 1/3rd reduction, neither one of which had any
particular factual support in the record, happily offset one another.89 One
wonders what guidance the district court in the next case is supposed to
take from this ruling.90
statutes but also acknowledges that CERCLA and other statutory liability regimes may present “special
policy considerations” influencing the choice in the context of those statutes. Id. § 1 cmt. e.
81. Burlington Northern, 129 S. Ct. at 1881.
82. Id.
83. Id. at 1882. The Court acknowledged that the district court erred by invoking equity in the
apportionment analysis, but glossed over the error because the “actual apportionment decision was
properly rooted in evidence” of causal contribution. Id. at 1882 n.9.
84. Id. at 1883.
85. Id.
86. Id.
87. Id.
88. Id.
89. Id.
90. Is it always appropriate to increase a calculated apportionment to account for possible
error? If it is, is 50% necessarily the right amount by which to increase the calculated apportionment?
322
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
The Court might have reached the same conclusion more honestly. It
could have agreed with the Ninth Circuit’s correct observation that the
record was inadequate to support the district court’s assumption of
proportionality between contamination and surface area or number of
years.91 It could then have stated, however, that the qualitative evidence
relating to the amount of activity and spillage on the two parcels was
overwhelming—so overwhelming that it actually justified the inference that
10% was the highest plausible quantitative estimate. That ruling, too,
would have been open to questioning and criticism, but it would have
provided some useful guideposts to lower courts and would have indicated
clearly the fact-bound nature of the holding.
But that is not what the Supreme Court did. So now lower courts must
determine how broadly or narrowly to apply the Burlington Northern
decision. Section II explains that, in making that choice, courts should
consider the overall goals of CERCLA and the role that joint and several
liability plays in achieving them.
II. JOINT AND SEVERAL LIABILITY IN THE CERCLA SCHEME
Despite the imprimatur the Supreme Court gave to section 433A of the
Second Restatement,92 courts “follow the Restatement, however, only to the
extent that it is compatible with the provisions of CERCLA.”93 Under the
Restatement, for example, the decision of whether liability is apportioned or
is joint and several does not even arise unless the plaintiff has first proven
that each defendant’s conduct was a “substantial” factor in bringing about
the harm, but courts “decline to place this threshold burden on the
government in CERCLA actions” because to do so would be inconsistent
with the statute’s liability-creating provision.94 Judicial determination of
If it is, and 9% was the apportionment supported by the record, shouldn’t the Railroads’ share have been
increased to 13.5%? Is it never appropriate to include an adjustment for possible error?
91. See generally Misiorowski & Eagle, supra note 23 (criticizing the Supreme Court for
approving apportionment despite lack of expert scientific testimony). As Misiorowski & Eagle point
out, it is unlikely that an expert trying to testify in support of the district court’s apportionment formula
could have survived a Daubert challenge. It is a useful thought experiment to wonder how Burlington
Northern would have come out had the burden of proof been in the opposite direction. If a plaintiff
seeking cost recovery had to prove the Railroads’ contribution in order to hold the Railroads liable at all,
would the Supreme Court have affirmed a plaintiff’s verdict of nine percent liability on the same record?
92. Burlington N. v. Burlington N. and Santa Fe Ry. Co. v. United States, 520 F.3d 918, 944
(9th Cir. 2008) (stating section 433A is a “universal starting point” for divisibility analysis) (quoting
United States v. Hercules, Inc., 247 F.3d 706, 717 (8th Cir. 2001)).
93. Hercules, 247 F.3d at 717; accord United States v. Capital Tax Corp., 545 F.3d 525, 535 n.8
(7th Cir. 2008).
94. O’Neil v. Picillo, 883 F.2d 176, 179 (1st Cir. 1989).
2009] Several Approaches to Divisibility After Burlington Northern
323
the scope of CERCLA liability is not a pure exercise of common law
judging, but an exercise in interstitial statutory interpretation.95 Courts,
therefore, should consider how well their application of Restatement
provisions comports with the statute’s overall structure and purpose as well
as its literal provisions.96
Despite the superficial resemblance between CERCLA liability and tort
liability for pollution, the analogy between CERCLA and tort is far from
perfect.
In a traditional tort case involving multiple independent
tortfeasors, the “primary consequence of what form of joint and several or
several liability is imposed is the allocation of the risk of insolvency of one
or more responsible tortfeasors.”97 This is also a critical consequence of the
choice between joint or apportioned liability in CERCLA cases, of course,
as the Burlington Northern case thoroughly illustrates. But joint and
several liability is vital to the CERCLA scheme in important ways that do
not apply to tort actions. CERCLA “was designed to promote the ‘timely
cleanup of hazardous waste sites’ and to ensure that the costs of such
cleanup efforts were borne by those responsible for the contamination.”98
The routine imposition of joint and several liability, more than its mere
theoretical availability, has admirably served those purposes.
A. Implementing the Polluter Pays Principle
Most obviously, joint and several liability furthers the second of
CERCLA’s major goals set forth above. By making solvent liable parties
(rather than the responding government) bear the risk that other liable
parties are insolvent, joint and several liability places the financial burden
of CERCLA cleanup on those responsible for the contamination.99
One might object that “responsibility” for contamination extends only
so far as one’s several contribution to the problem, but that begs the
question. Someone must bear any “share” that would be assigned to an
insolvent liable party. The issue is whether Congress evinced any
95. United States v. Alcan Aluminum Corp., 964 F.2d 252, 268, 270 n.29 (3d Cir. 1992);
United States v. Chem-Dyne Corp., 572 F. Supp. 802, 808 (S.D. Ohio 1983).
96. See In re Bell Petroleum Servs., 3 F.3d 889, 902 (5th Cir. 1993) (“Restatement principles
must be adapted, where necessary, to implement congressional intent with respect to liability under the
unique statutory scheme of CERCLA.”) (emphasis in original).
97. RESTATEMENT (THIRD) OF TORTS: APPORTIONMENT OF LIABILITY § 10 cmt. a (2000).
98. Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870, 1874 (2009) (quoting
Consol. Edison Co. of N.Y. v. UGI Util., Inc., 423 F.3d 90, 94 (2d Cir. 2005)).
99. “The degree to which the United States will be able to protect its financial interest in the
trust fund is directly related to the scope of liability under CERCLA . . . .” Chem-Dyne Corp., 572 F.
Supp. at 808.
324
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
preference for allocating such share to the responding government, on the
one hand, or to viable responsible parties, on the other. The statute itself100
provides plenty of evidence of “the general congressional intent of placing
liability for toxic waste clean-up as nearly as possible on those responsible
for creating the hazard,”101 even though Congress also created the
Hazardous Substances Superfund as a backstop.
The most telling provisions are those that Congress enacted specifically
in response to the argument that joint and several liability unfairly makes
relatively minor contributors disproportionately liable for CERCLA
response actions. Congress did not mandate that the courts abandon or
limit the imposition of joint and several liability.102 Instead, endorsing the
Chem-Dyne approach, Congress chose to rely on contribution to ameliorate
the “harshness”103 of joint and several liability, even though, of course, a
contribution right would be futile in proportion to the extent of insolvency
among those allocated equitable shares.
When generators of municipal solid waste and of small amounts of
waste bearing hazardous substances continued to express concerns about
being subject to the CERCLA liability scheme, Congress again declined to
enact a scheme of several liability in proportion to causal contribution, or
even, for example, to specify that joint and several liability would apply
only to liable parties whose contributions exceeded some threshold of
significance. Instead, Congress chose to entirely exempt from liability
certain types of parties that presented compelling cases of unfairness in the
existing liability regime.104 In doing so, Congress again acknowledged the
general availability of joint and several liability under the statute,105 and
Congress again chose not to change it—preferring a different way to
“return a little bit of basic fairness to Superfund’s liability regime.”106 In
100. Statements of congressional intent in the legislative history that ratify the early joint and
several liability jurisprudence under CERCLA provide additional support. See supra note 31.
101. United States v. Wade, 577 F. Supp. 1326, 1339 (E.D. Pa. 1983).
102. The enacted provisions did again demonstrate, however, that “when Congress wanted to
draw distinctions based on concentration or quantity, it expressly provided as much.” B.F. Goodrich Co.
v. Murtha, 958 F.2d 1192, 1200 (2d Cir. 1992).
103. U.S. Bank Nat’l. Ass’n v. Envtl. Prot. Agency, 563 F.3d 199, 207 n.4 (6th Cir. 2009).
104. See 42 U.S.C. § 9607(o) (2006) (de micromis exemption); see also id. § 9607(p)
(municipal solid waste exemption). Congress enacted both provisions as part of the Small Business
Liability Relief and Brownfields Revitalization Act, Pub. L. No. 107-118 (2002).
105. H.R. REP. NO. 107-70, at 2 (2001) (CERCLA “has been interpreted by the courts to have
established responsibility for clean ups based upon a retroactive, strict, joint, and several liability
scheme”); 147 CONG. REC. H 10,903 (Dec. 19, 2001) (statement of Rep. Boehlert) (describing problems
of arranger for disposal of very small quantity of waste under “current strict, joint, and several liability
system . . . .”).
106. 147 CONG. REC. H 10,903 (Dec. 19, 2001) (statement of Rep. Boehlert).
2009] Several Approaches to Divisibility After Burlington Northern
325
fact, in response to the argument that exempting parties from liability would
reduce the government’s ability to recover costs to replenish the Superfund,
the pertinent House committee explained that “to the extent EPA could
recover the exempted PRP’s share of the costs from any other remaining
PRPs at a particular site, there would be no reduction in costs recovered.”107
Such recovery, of course, is possible only if liability is joint and several;
otherwise no liable party could be forced to cover the causally-apportioned
share of an exempted party.
Furthermore, the statute’s treatment of settlements, and in particular of
so-called “ability to pay” settlements, demonstrates congressional intent
that other responsible parties, and not the government, bear the shares of
insolvent parties. The 1986 amendments to CERCLA included provisions
designed to encourage and expedite settlements. One provision specified
that a settlement with one defendant “reduces the potential liability of the
others by the amount of the settlement,” rather than by the settling
defendant’s apportioned share of the liability.108 This places on non-settling
defendants the risk that the government will make an inadequate settlement.
In 2002, Congress further clarified that the government should reduce the
amount required in settlement for a party that demonstrated an inability or
limited ability to pay response costs.109 These provisions combine to assign
the responsibility of filling the gap between what a liable party can pay and
what it rightly ought to pay. Congress put that burden squarely on other
liable parties instead of the responding government—just as the imposition
of joint and several liability does.
Finally, if joint and several liability were to become rare under
CERCLA, or even if its availability in any given case were to become
subject to considerable doubt, government plaintiffs would have little
alternative but to sue as many potentially responsible parties as they
reasonably could in order to maximize the potential cost recovery. This
would drive up the government’s enforcement costs, both because it would
have to pursue more parties and because the cases would become inherently
107. H.R. REP. NO. 107-70, at 5 (2001).
108. 42 U.S.C. § 9613(f)(2) (2006); see id. § 9622(g)(5) (same, with respect to expedited “de
minimis” settlements with arrangers that contributed small amounts of hazardous substances to the
facility or with owners who did not dispose of hazardous substances or contribute to the release).
Contribution protection for settlers coupled with this “pro tanto” approach, which makes no sense if
parties can only be held severally liable for their apportioned share, strongly implies that Congress
understood that parties would be liable for shares that could be causally attributed to other parties, as in
joint and several liability.
109. § 9622(g)(7).
326
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
more complex. Those costs are recoverable under CERCLA,110 but they,
too, presumably would be recoverable only severally. The changed legal
regime would thus force the government to spend more money on
enforcement and to lose a portion of each dollar spent if some responsible
parties were insolvent. Not only would this exacerbate the erosion of the
polluter-pays principle, but it would also, if only by competition for
resources, likely cause a delay in government cleanup activities—thereby
frustrating CERCLA’s other primary goal.
B. Facilitating Response Action
“Timely cleanup” has been central not just to the legislative purpose,
but to the executive’s implementation of CERCLA. In 1989, EPA initiated
its “enforcement first” approach, which seeks to have responsible parties
undertake response action in preference to a government-conducted,
Superfund-financed cleanup followed by cost recovery litigation.111 The
near-certain, or even likely, prospect of joint and several liability for the
costs of responding to a release of hazardous substances is an obvious
incentive for one or a group of responsible parties to agree to undertake
response action,112 and under the evolved legal regime responsible parties
have very commonly entered such agreements.113 By contrast, a party that
confidently believed it faced only several liability would be much more
likely to prefer litigating against the government to attempt to establish an
x% share of liability, rather than undertaking to perform a response action
110. 42 U.S.C. § 9607(a)(4)(A) (2006); see id. § 9601(25) (“The terms ‘respond’ or ‘response’
means remove, removal, remedy, and remedial action; all such terms (including the terms ‘removal’ and
‘remedial action’) include enforcement activities related thereto.”).
111. EPA, SUPERFUND 20TH ANNIVERSARY REPORT, TEXT TIMELINE (2000),
http://www.epa.gov/superfund/20years/texttl.htm; EPA OFFICE OF WASTE PROGRAMS ENFORCEMENT,
SUPERFUND ENFORCEMENT STRATEGY AND IMPLEMENTATION PLAN 7, 15, 18–19 (Sept. 26, 1989),
http://www.epa.gov/compliance/resources/policies/cleanup/superfund/strategy-imp-rpt.pdf (describing
emphasis on private party response action implementation as general principle, in removal program, and
in remedial program).
112. See United States v. Atchison, Topeka & Santa Fe Ry. Co., No. CV-F-92-6058, 2003 U.S.
Dist. LEXIS 23130, at *229 (E.D. Cal. July 16, 2003) (“‘[I]mposing joint and several liability carries
out the legislative intent by insuring that responsible parties will fulfill their obligations to clean up the
hazardous waste facility.’”) (quoting United States v. Stringfellow, 661 F. Supp. 1053, 1060 (C.D. Cal.
1987); 131 Cong. Rec. 23,928, 23,953 (Sept. 17, 1985) (responses by EPA appended to statement of
Sen. Stafford, predicting that replacing joint and several liability with apportioned liability would delay
cleanup by making injunctive relief more difficult to obtain and reducing incentives for settlement).
113. A recent study identified nearly 1800 such agreements to perform response actions with an
estimated value of $22.5 billion, with respect only to the subset of response actions that involve sites on
the National Priorities List. See Martha L. Judy & Katherine N. Probst, Superfund at 30, 11 VT. J.
ENVTL. L. 191, 199 (2009).
2009] Several Approaches to Divisibility After Burlington Northern
327
and then trying to recover the difference between 100% and x% from a
multitude of other responsible parties.114
Even when responsible parties have not agreed to perform response
actions, the government nevertheless has very often secured private
performance by issuing a unilateral administrative order under section
106(a) of CERCLA.115 Failing to comply with such an order without
sufficient cause subjects the violator to stiff civil penalties116 and punitive
damages,117 which might elicit performance even from a party that believed
it was liable only for a portion of site response costs and that imposition of
joint and several liability was unlikely. But such a party presumably would,
upon completion of the action, seek reimbursement from the Superfund118
for the reasonable costs it incurred in excess of its apportioned liability.119
114. Agreements to perform response action provide a variety of benefits to settling defendants.
A performing party secures the right to select contractors (subject to government veto) and obtains a
degree of control over the work and costs. Moreover, the government sometimes offers a partial
compromise of past costs as an incentive for agreeing to the obligation to perform future response
actions (an offer made feasible by the government’s knowledge that it could pursue non-settlers jointly
and severally for any shortfall). All these benefits are worth proportionately less to a party that is
confident it faces only an apportioned share of liability. Compared to settling with the government and
suing many other responsible parties for contribution, defending against the government and seeking
apportionment would probably cost less, would be administratively easier, and would avoid both
litigation risk and insolvency risk against the other responsible parties. The analysis is no different for
attempted negotiations with a group of responsible parties, each of which believes it is only severally
liable.
115. 42 U.S.C. § 9606(a) (2006).
116. § 9606(b)(1) (“Any person who, without sufficient cause, willfully violates, or fails or
refuses to comply with, any order of the President under subsection (a) may . . . be . . . fined not more
than $25,000 for each day in which such violation occurs or such failure to comply continues.”).
117. 42 U.S.C. § 9607(c)(3) (2006) (“If any person who is liable for a release or threat of release
of a hazardous substance fails without sufficient cause to properly provide removal or remedial action
upon order of the President . . . such person may be liable to the United States for punitive damages in
an amount at least equal to, and not more than three times, the amount of any costs incurred by the Fund
as a result of such failure . . . .”).
118. § 9606(b)(2) (authorizes judicially reviewable petitions for reimbursement by “[a]ny
person who . . . complies with” a section 106(a) order). To obtain reimbursement, a petitioner must
“establish by a preponderance of the evidence that it is not liable for response costs.” § 9606(b)(2)(C).
119. The Railroads involved in Burlington Northern did just this. EPA ordered them to repair a
security fence, build a berm, and install monitoring wells on their parcel at the Brown & Bryant site. In
re Brown & Bryant, Inc. Site, Pet. for Reimbursement of Costs 7, No. 94-12 (EPA Oct. 5, 1992). Their
initial petition for reimbursement contended (as they contended in the district court) that they bore no
liability whatsoever for the site. Id. at 2, 13–20. After the district court’s judgment, the Railroads
moved in the administrative proceeding for a reimbursement award of 91% of their costs of compliance
with the 106(a) order, consistent with the district court’s finding that they bore a 9% share of liability. In
re Brown & Bryant, Inc. Site, Notice of Expiration of Stay and Request for Issuance of Final Order
Granting Reimbursement 4–5, 18, No. 94-12 (EPA Sept. 30, 2003). Based on review of the
Environmental Appeals Board docket as of this writing, the petition apparently remains pending. At
least one court decision strongly implies, however, that reimbursement would be available to a party in
the Railroads’ position. Chem-Nuclear, 292 F.3d at 259 (if it could meet its burden of proof on
328
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
If reimbursement claims based on apportionment of liability were to begin
to succeed routinely, EPA would face constant uncertainty as to the eventual
financial outcome of any section 106(a) order. That uncertainty could
inhibit EPA’s willingness to use its authority to issue orders, reducing the
utility of this critical statutory tool for achievement of the key statutory
objective of timely response actions.120
The statutory objective of achieving prompt response action—which
entails efficacious use of the statute’s provisions for responsible party
implementation under orders, injunctions, and settlements121—distinguishes
CERCLA from tort law, which is concerned only with where the financial
consequences of a completed loss will lie.122 This distinction should inform
courts’ efforts to apply common law principles in the CERCLA context.123
The common law’s willingness to apportion harm was predicated at least in
part on a concern for fairness: that “the one who did the least” not be “made
liable for the damages of others far exceeding” that.124 But CERCLA, as
courts have drily noted, “is not a legislative scheme which places a high
priority on fairness to” parties made liable;125 it imposes liability, for
instance, even if no evidence exists to satisfy the traditional common law
causation standard.126 Joint and several liability is a pretty big hammer to
apportionment, petitioner clearly liable for at least some costs could nonetheless “avoid joint-andseveral liability for the full” cost of response action performed under order; court held burden was not
met).
120. Even if EPA continued to issue section 106(a) orders, the need to plan amid the attendant
uncertainty, as well as to cover the eventual reimbursements, would surely inhibit EPA commitments of
resources to Superfund-lead sites, impeding achievement of the statute’s cleanup objective in another
way.
121. See 42 U.S.C. § 9604(a) (authorizing the President to allow a responsible party to take
response action including remedial investigations and feasibility studies); § 9606(a) (authorizing the
President to issue orders to protect public health, welfare, and the environment or to require the Attorney
General to secure injunctive relief); 42 U.S.C. § 9622(a) (2006) (authorizing the President to enter
settlements under which responsible parties will perform response actions).
122. In this respect, an administrative order or suit for injunction under CERCLA also is
different from a tort claim seeking abatement of a nuisance, because a nuisance defendant generally can
be enjoined only to cease its own nuisance-creating activity, while any party liable for a release of
hazardous substances can be ordered to take response action.
123. E.g., United States v. Chem-Dyne Corp., 572 F. Supp 802, 810 (S.D. Ohio 1983).
124. Chipman v. Palmer, 77 N.Y. 51, 53 (1879).
125. United States v. Rohm & Haas Co., 721 F. Supp. 666, 686 (D.N.J. 1989).
126. See, e.g., United States v. Alcan Aluminum Corp. (Alcan-Fulton), 315 F.3d 179, 184 (2d
Cir. 2003) (“[G]overnment is not required to show that a specific defendant's waste caused the
incurrence of cleanup costs in order for strict liability to attach to that defendant.”); United States v.
Alcan Aluminum Corp., 990 F.2d 711, 721 (2d Cir. 1993) (“CERCLA does away with a causation
requirement.”); United States v. Alcan Aluminum Corp. (Alcan-Butler), 964 F.2d at 265 (“Congress . . .
imposed liability upon a class of responsible persons without regard to whether the person specifically
caused or contributed to the release and the resultant response costs.”) (emphasis in orginal); United
States v. Hercules, 247 F.3d 706, 716 (8th Cir. 2001) (CERCLA “does not require the government to
2009] Several Approaches to Divisibility After Burlington Northern
329
use to achieve the goal of responsible-party-led cleanups,127 but that does
not mean courts should ignore the congressional objective.
The Supreme Court’s opinion in Burlington Northern did not consider
whether or how its ruling would affect the achievement of CERCLA’s
statutory goals. But it also did not instruct that the congressional purpose is
irrelevant. Those goals should suffuse the thinking of lower courts as they
try to figure out, in CERCLA cases with different fact patterns, whether
they should make the same kind of loose assumptions that the Supreme
Court thought sufficient in Burlington Northern.
III. BURLINGTON NORTHERN PROVIDES LITTLE JUSTIFICATION FOR
APPORTIONMENT IN OTHER FACTUAL SETTINGS
Burlington Northern and Bell Petroleum stand as the primary examples
of successful divisibility arguments by CERCLA responsible parties. Of
the two, Burlington Northern was the harder case to prove. There were
three hazardous substances, not one; and there were two parcels at which
different activities took place, not one plant at which successive owners did
the same thing. The business records available to the court in Bell
Petroleum, although incomplete and still requiring inference,128 were much
better evidence of each defendant’s contribution to the groundwater
contamination than Judge Wanger had in Burlington Northern. But even
though Burlington Northern was not as simple as Bell Petroleum, it also
was not like the many CERCLA cases that involve numerous commingled
chemicals from numerous sources. Should the fact that apportionment
succeeded in Burlington Northern lead to successful apportionment in the
more typical cases where, up to now, courts have found the evidence
insufficient? As this section explains, it should not.
A. Separate Ownership Was a Critical Fact in Burlington Northern
Both the district court and appellate court in Burlington Northern
emphasized that the two adjacent parcels on which B&B conducted its
prove as part of its prima facie case that the defendant caused any harm to the environment.”); United
States v. Burlington N. & Santa Fe Ry. Co., 520 F.3d 918, 933 (9th Cir. 2008) (parties can be liable
“without any finding . . . that they caused the contamination . . . .”).
127. See United States v. Conservation Chem. Co., 628 F. Supp. 391, 403 (W.D. Mo. 1985)
(“[A]ll other things being equal, this Court believes that public policy demands that preference be given
to the use of private funds for cleanup of hazardous waste sites.”).
128. In re Bell Petroleum Servs., 3 F.3d 889, 894 (5th Cir. 1993).
330
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
operations constituted a single CERCLA “facility.”129 Nevertheless, the fact
of separate ownership was critical to the district court’s apportionment,
which the Supreme Court eventually found reasonably supported by the
record.
To understand this requires looking at the first round of litigation over
the B&B Superfund site, which was filed by the Railroads. Having
complied with an EPA order requiring them to repair a security fence at the
site and to conduct investigatory activities on their parcel, the Railroads
sued B&B and related parties for the costs of those actions.130 The B&B
parties counterclaimed for contribution. The parties filed cross-motions for
summary judgment.
Judge Wanger rejected, as he later did in the governments’ case, the
Railroads’ arguments that they bore no CERCLA liability whatever for the
Arvin site.131 The ruling was not a total loss for the Railroads, however. In
hindsight, it clearly adumbrated the apportionment decision that was to
come after additional years of litigation and evidentiary development. The
decision sharply distinguished the two parcels. It reasoned that because
CERCLA liability attaches to the “owner and operator of . . . a
facility . . . from which there is a release . . . of a hazardous substance,”132
Both the Railroads and the Brown & Bryant parties . . .
incorrectly focus their liability analysis on the areas
affected by the contamination rather than upon the source
(“facility”) of the contamination. It is not material to the
issue of liability who owns the property requiring
remediation; what is essential to the issue of liability is who
129. United States v. Atchison, Topeka & Santa Fe Ry. Co., No. CV-F-92-6058, 2003 U.S. Dist.
LEXIS 23130, at *128 (¶ 304) (E.D. Cal. July 14, 2003); United States v. Burlington N. & Santa Fe Ry.
Co., 520 F.3d 918, 944 (9th Cir. 2008). The Supreme Court did not address the issue so clearly, but it
nonetheless repeatedly referred to “the Arvin facility.” E.g., Burlington Northern, 129 S. Ct. at 1883.
130. The Railroads described their action as one “for cost recovery and contribution.” See In re
Brown & Bryant, Inc. Site, Petitioners’ Motion to Stay Further Proceedings Pending Resolution of
Related CERCLA Actions in U.S. District Court and Request for Issuance of Final Order Granting
Reimbursement 2, No. 94-12 (EPA May 13, 1997). See generally Atchison, Topeka & Santa Fe Ry. Co.
v. Brown & Bryant, Inc., 42 ERC (BNA) 1605, 1995 U.S. Dist. LEXIS 20627, at *24–25 (E.D. Cal.
Nov. 15, 1995) (stating that railroads could seek joint and several liability against B&B parties for costs
of responding to releases from B&B Parcel). They alleged they had spent more than $3 million to
comply with the order. The district court ultimately found that the Railroads had incurred about $2.6
million in costs that were reasonable and thus recoverable by a private party under CERCLA. Id; see 42
U.S.C. § 9607(a)(2) (2006) (stating that parties are liable when they owned or operated a facility at the
time of hazardous material disposal).
131. Brown & Bryant, Inc., 45 ERC (BNA) 1605 at *17–18.
132. § 9607.
2009] Several Approaches to Divisibility After Burlington Northern
331
owned or operated the “facility” that caused the need for
remediation.133
Therefore, Judge Wanger held that “the Railroads are not liable for
response costs for releases from the Brown & Bryant property.”134 The
judge found “a genuine issue of material fact whether releases from the
Railroads’ property have resulted in response costs for groundwater
contamination” beneath B&B’s land, so he refused to rule, at that stage, that
the Railroads could not be jointly and severally liable for the cleanup of the
B&B property.135 Nevertheless, even before the governments put their
claims in issue, the trial court had conceived that the Railroads’ liability (1)
was fundamentally severable from B&B’s and (2) could be determined
based on the extent to which releases from the Railroad parcel contributed
to the groundwater contamination plume. This conception never altered. It
permeated the liability holding in the governments’ case136 as well as the
apportionment ruling, which amounted to an effort to figure out how much
of the contamination in the groundwater might have come from releases on
the Railroad parcel.
The district court’s approach cannot translate to paradigmatic multiple
arranger cases such as in Chem-Dyne, O’Neil, or the Alcan cases while
remaining consistent with those decisions. The equivalent analysis would
begin with the assertion that each arranger is liable only “for” the release of
its own hazardous substances, a contention the courts have firmly rejected.
Rather, CERCLA makes arrangers liable for the costs of responding to the
release from the facility—unless, of course, they can prove what their
volumetric share was and also that volumetric share is a reasonable basis
for apportionment.
B. Burlington Northern Did Not Relieve Arrangers Seeking
Apportionment of the Burden of Proving Their Shares
In Burlington Northern, the Supreme Court approved an apportionment
calculation based on strong assumptions of proportionality that were
133. Brown & Bryant, Inc., 45 ERC (BNA) 1605 at *12–13 (emphasis omitted) (footnote
omitted).
134. Id. at *19 (emphasis omitted); see id. at *21 (“[T]here is no basis for the Railroads’ liability
for response costs attributable to releases at Brown & Bryant’s property.”).
135. Id. at *18.
136. “The Railroads are strictly liable for the hazardous substances originating from the
‘Railroad parcel’ that in part caused the need for CERCLA response of the Site.” United States v.
Atchison, Topeka & Santa Fe Ry. Co., No. CV-F-92-6058, 2003 U.S. Dist. LEXIS 23130, at *159 (E.D.
Cal. July 14, 2003).
332
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
weakly supported by evidence.137 But the record in Burlington Northern
still allowed the district court to extract the two components that are critical
to allow apportionment of a fraction of liability: a numerator and a
denominator. The Restatement makes clear that before even a divisible
harm can be divided, there must be evidence of the appropriate dividend
and divisor.138
In many multiple-arranger cases, such evidence has been absent or
inadequate. Nothing in Burlington Northern undercuts the holding, for
example, of O’Neil, in which the court held that the arranger defendants
seeking apportionment had merely proven the minimum number of drums
they had sent to the site,139 or the similar holding of Chem-Nuclear.140
Burlington Northern would be analogous only if the Railroads had proven
that their parcel constituted at least 19% of the site but possibly more, or
that their lease had extended through at least 45% of the time B&B
operated but possibly longer. Similarly, the denominator was not in
question in Burlington Northern, but in many chemical dump cases the total
amount of hazardous substances disposed cannot reliably be estimated. In
such a case, it would be speculative to assign a percentage share even to an
arranger whose own contribution was fully proven.
Finally, even if a volumetric numerator and denominator could be
proven, they may not provide a reasonable basis for apportionment. In
“chemical soup” cases, apportionment has often failed also because variable
toxicity, variable mobility, and potential interactions among the different
chemicals in the mix have made it unreasonable to simply assume that
volume is a reasonable proxy for harm.141 Even in common law claims for
pollution damage, courts recognized that synergistic effects would warrant
joint and several liability.142 CERCLA defendants often cannot demonstrate
137. The district court (1) had very limited quantitative evidence of the actual amount of
chemicals spilled on the Railroad Parcel as compared to the B&B Parcel and instead relied on proxies of
unproven reliability, such as land area and number of years; and (2) assumed the proportionality of harm
to volume without having heard any testimony directed to that issue. See supra notes 67–69.
138. RESTATEMENT (SECOND) OF TORTS § 433A cmt. d (1966).
139. O’Neil v. Picillo, 883 F.2d 176, 182–83 (1st Cir. 1989).
140. Chem-Nuclear Systs., Inc. v. Bush, 292 F.3d 254, 260–61 (D.C. Cir. 2002) (stating that
proof that drums known to have come from the arranger defendant had been dumped in one part of the
site was insufficient to support the conclusion that none of the arranger’s drums had been dumped in
another portion of the site).
141. E.g., O’Neil, 883 F.2d at 182 (stating that appellants could be limited contributors only if
they documented “the whereabouts of their waste at all times after it left their facilit[y]”).
142. See Symmes v. Prairie Pebble Phosphate Co., 63 So. 1, 3 (Fla. 1913) (holding that
companies that independently dumped mine waste that destroyed plaintiff’s oyster beds were only
severally liable, in part because it did not “appear that the act of each one made the acts of the others
more injurious”).
2009] Several Approaches to Divisibility After Burlington Northern
333
the absence of interactions or variability that would vitiate the assumption
that harm is proportional to quantity.
The relatively simple situation in Burlington Northern, where the same
three chemicals were spilled on both parcels, did not present the possibility
of synergistic interactions or differential effects of different defendants’
hazardous substances.143 The Supreme Court’s decision, which after all
cited cases like Alcan and Monsanto that found the absence of such
evidence dispositive, cannot be read to alter those cases’ import.144
IV. NEW AND OLD LEGAL THEORIES OF INDIVISIBLE HARM CONTINUE
TO SUPPORT IMPOSITION OF JOINT AND SEVERAL LIABILITY AFTER
BURLINGTON NORTHERN
Section III explained why judicial application of the actual holding of
Burlington Northern—that the record in that case provided an adequate
basis for apportionment—should not extend to a broader range of factual
scenarios. But what if it does? What if Burlington Northern ends up
meaning that the burden of production and persuasion is no longer
“difficult,”145 so that in the future it will be satisfied commonly rather than
“rarely”?146 This section proceeds from that assumption and argues that
courts should still impose joint and several liability in a wide range of
circumstances.
As the Supreme Court noted, before evidence can support a reasonable
basis for apportioning a harm, the harm must be “theoretically capable of
apportionment.”147 The Supreme Court spent no analytical energy on this
requirement, merely observing that “both the District Court and Court of
Appeals agreed that the harm created by the contamination of the Arvin
site, although singular, was theoretically capable of apportionment.”148
Lower courts that rejected CERCLA defendants’ earlier attempts to justify
143. See Brief for Petitioners Burlington Northern and Santa Fe Ry. Co. and Union Pacific
Railroad Co. at 36, 38, Burlington N. and Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870 (2009) (No.
07-1601 & 07-1067) (contrasting Burlington Northern with case of “genuinely synergistic harms”).
144. See Burlington N. v. Burlington N. and Santa Fe Ry. Co. v. United States, 520 F.3d 918,
944 (9th Cir. 2008) (The Ninth Circuit referred to the “synergistic” use of two parcels, but it did not
suggest the existence of chemical synergies akin to those found in multiple arranger cases.).
145. United States v. Hercules, Inc., 247 F.3d 706, 716 (8th Cir. 2001) (“We have previously
observed that proving divisibility is a very difficult proposition.”) (internal quotation marks and citation
omitted).
146. O’Neil, 883 F.2d at 178 (“[R]esponsible parties rarely escape joint and several liability
. . . .”).
147. Burlington Northern, 129 S. Ct. at 1881.
148. Id.
334
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
apportionment generally relied, like the Ninth Circuit in Burlington
Northern, on the relatively easy finding of a failure of proof, without having
to grapple too closely with the preliminary question of when apportionment
is appropriate even in principle. CERCLA plaintiffs, who routinely won
with arguments that the defendants’ proof had failed to clear the bar, had
little reason to engage in esoteric legal argument that the defendants should
not be permitted even to attempt the vault.
Nevertheless, according to both the Second Restatement and the
Supreme Court, the apportionment inquiry is a two-step process. If
Burlington Northern is understood to imply that courts have misunderstood
the nature of the second step, that understanding warrants re-examination of
the first step. The requirement that a harm be “theoretically capable of
apportionment” must have some substantive content, or neither the Second
Restatement nor the Supreme Court would have bothered articulating it.
And it must mean something more than “the total amount of damages
can be divided into fractions,” which would make every harm
apportionable. “Not all harms are capable of apportionment,”149 the
Supreme Court acknowledged; such a thing exists as a “single, indivisible
harm.”150 This is true under section 433A of the Second Restatement in
general,151 and it must also be true of harms in CERCLA cases in particular.
For even if “Congress did not intend for joint and several liability to be
imposed without exception,”152 it is equally clear that Congress did not
intend for courts invariably to apportion liability. We know this, not only
for the reasons Judge Rubin expressed in Chem-Dyne, but also because of
the subsequent Congressional response in enacting the Superfund
Amendments and Reauthorization Act of 1986 (“SARA”).153 Not only did
Congress “fully subscribe[] to the reasoning” of Chem-Dyne itself,154 but it
also embraced the then-developed body of jurisprudence applying ChemDyne,155 which had already begun to evince joint and several liability as the
149. Id.
150. Id.
151. RESTATEMENT (SECOND) OF TORTS § 433A(2) (1966); see id. cmt. i.
152. O’Neil v. Picillo, 883 F.2d 176, 181 (1st Cir. 1989).
153. Superfund Amendments and Reauthorization Act of 1986, Pub. L. No. 99-499 (1986).
154. H.R. REP. NO. 253, pt. 1, at 74 (1985), reprinted in 1986 U.S.C.C.A.N. 2835, 2856; accord
131 Cong. Rec. 34,632 (Dec. 5, 1985) (statement of Rep. Dingell) (“The uniform Federal rule on joint
and several liability established in [Chem-Dyne] is correct and should be followed.”); see id. at 34,636
(statement of Rep. Eckart) (The bill “absolutely protects” CERCLA “liability scheme of strict, joint, and
several liability”; all committees involved “fully subscribe” to Chem-Dyne standard.); 132 Cong. Rec.
29,737 (Oct. 8, 1986) (statement of Rep. Glicksman) (conferees continued house committees’ approach
endorsing Chem-Dyne).
155. H.R. REP. NO. 253, pt. 1, at 74 (1985), reprinted in 1986 U.S.C.C.A.N. 2835, 2856
(“[C]ourts have made substantial progress” in applying traditional and evolving common law standards
2009] Several Approaches to Divisibility After Burlington Northern
335
ordinary, if not ordained, result in CERCLA cases.156 Moreover, when it
amended CERCLA, Congress codified a right of contribution among
responsible parties, which would be unnecessary if defendants were liable
to plaintiffs only severally.157
The challenge, then, is for courts simultaneously to abide by the
teaching of Burlington Northern, to apply the common law principles of the
Second Restatement, and to keep faith with the congressional intent of
CERCLA. Courts can do this, even if they believe the Supreme Court has
greatly reduced proof requirements for conventional CERCLA divisibility
arguments, by taking a fresh look at how the “traditional and evolving
principles of common law”158 fit with the overarching objectives of the
statute. To show how, I briefly review the role of joint and several liability
in the statutory structure and then describe several bases for imposition of
joint and several liability that survive even a greatly relaxed standard of
proof of divisibility under section 433A. These theories can ensure that
joint and several liability continues to play its appropriate role in sustaining
the objectives of CERCLA, without the need for legislative changes and
without disregarding either the Restatement or Burlington Northern.
A hypothetical will help illustrate several of these theories. Consider
the Manufacturers’ Surplus Landfill (“MSL”), a prototypical (but simpler
than is usual in the real world) dump site, contaminated with a halogenated
organic solvent hazardous substance (“HOS”) that the government first
investigates and then remediates by installing a cap to prevent infiltration
and a pump-and-treat system that will need to operate for thirty years to
manage migration. The following parties are liable:159
to scope of CERCLA liability.); 132 Cong. Rec. 29,716 (Oct. 8, 1986) (statement of Rep. Dingell)
(“Nothing in this legislation is intended to change the application of the uniform Federal rule of joint
and several liability enunciated in the Chem-Dyne case and followed by a number of other Federal
courts.”).
156. See United States v. S.C. Recycling and Disposal, Inc., 653 F. Supp. 984, 995 (C.D. S.C.
1984) (“[A]rbitrary or theoretical means of cost apportionment do not diminish the indivisibility of the
underlying harm”—refusing to apportion among arrangers by volume.) (emphasis in original); United
States v. Ottati & Goss, Inc., 630 F. Supp. 1361, 1396 (D.N.H. 1985) (ruling that although defendants
proved approximately how many of their drums were brought to site, “resulting proportionate harm to
surface and groundwater cannot be proportioned with any degree of accuracy as to any individual
defendant,” so “liability is joint, several and indivisible.”).
157. 42 U.S.C. § 9613(f) (2006); RESTATEMENT (THIRD) OF TORTS: APPORTIONMENT OF
LIABILITY § 23 (2000) (stating that contribution is applicable “when two or more persons are or may be
liable for the same harm and one of them discharges the liability of another”); see H.R. REP. NO. 253, pt.
1, at 79 (1985), reprinted in 1986 U.S.C.C.A.N. 2835, 2861 (stating that the contribution right among
liable parties is “concomitant” to their joint and several liability to government).
158. United States v. Chem-Dyne Corp., 572 F. Supp. 802, 807 (S.D. Ohio 1983).
159. In an actual case, liability might be disputed, but the assumption that all identified
defendants are liable facilitates a useful illustration. The same applies to the facts concerning volume of
336
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
Defendant O1 owned and operated the dump for the first twenty years
of its existence. Sixty-seven percent of the total volume of HOS was
dumped during this time. O1 is insolvent.
Defendant O2 is the current owner of the dump. After buying the dump
from O1 twenty years ago, O2 operated the dump for ten additional years,
during which 33% of the total volume of HOS was dumped, and then O2
stopped accepting waste there. O2 is insolvent.
Defendant T1 is a waste hauler that transported 60% of the total volume
of HOS to the dump. T1 is insolvent.
Defendant T2 is a waste hauler that transported 10% of the total volume
of HOS to the dump. T2 is insolvent.
Defendant A1 is a manufacturer that contracted with O1 and O2 (the
owner/operators) for disposal of HOS. A1 brought its HOS to the dump in
its own trucks and contributed 30% of the total volume of HOS. A1 is
viable.
Defendant A2 is a manufacturer that contracted with T1 to haul HOS
from its plant; T1 hauled the HOS to the dump. A2 contributed 20% of the
total volume of HOS. A2 is viable.
Defendant A3 is a manufacturer that contracted with T1 to haul HOS
from its plant; T1 hauled the HOS to the dump. A3 contributed 40% of the
total volume of HOS. A3 is insolvent.
Defendants A4 through A8 are machine shops that contracted with T2 to
haul HOS from their plants; T2 hauled the HOS to the dump. Each of these
five arrangers contributed 2% of the total volume of HOS. A4 has assets,
but not enough to satisfy the entire liability. A5, A6, A7, and A8 are
insolvent.
It is impossible to argue that the MSL facility presents “distinct harms.”
A defendant seeking to apportion liability for MSL would, like the
Railroads in Burlington Northern, try to establish that the single harm is
divisible, or reasonably capable of apportionment, under section
433A(1)(b). The theories described below all suggest reasons why the
assertion of divisibility might fail at the first step, the determination
whether the CERCLA harm is “theoretically capable of apportionment.” A
corollary is that none of these theories (even if it completely ruled out the
possibility of divisibility under section 433A(1)(b)) would lead to
invariable imposition of joint and several liability, because liability in cases
of distinct harms would still be several. Thus, these theories would not
contravene the dictum that “Congress did not intend for joint and several
each arranger’s and transporter’s hazardous substance, which in the real world would almost certainly be
much more in doubt.
2009] Several Approaches to Divisibility After Burlington Northern
337
liability to be imposed without exception.”160 Cases of distinct harm are not
hypothetical: they can arise in so-called “two-site” cases where harms are
geographically distinct, and perhaps in cases involving parties bearing
separate responsibility for contamination of separate environmental media.
With these preliminary matters addressed, the legal theories can be
presented. They begin with the one that requires the least rethinking of the
prevailing assumption of theoretical divisibility and ascend from there.
A. Of Soloists and Ensembles: Concert of Action
Section 433A may be the “universal starting point for divisibility of
harm analyses in CERCLA cases,”161 but it is not the only basis in the
Restatement for imposition of joint and several liability on multiple
tortfeasors. Section 433A provides a scope of liability rule that applies
“when two or more persons acting independently”162 cause harm.
The Restatement articulates another basis for imposition of joint and
several liability. Tortfeasors that act in concert are jointly and severally
liable for the harm their concerted action causes, even if their individual
acts cause only part of that harm.163 According to the Second Restatement,
concerted action leading to joint and several liability exists, subjecting a
party to liability “[f]or harm resulting to a third person from the tortious
conduct of another,”164 where that party:
(a) does a tortious act in concert with the other or pursuant
to a common design with him, or
(b) knows that the other's conduct constitutes a breach of
duty and gives substantial assistance or encouragement to
the other so to conduct himself, or
160. O’Neil v. Picillo, 883 F.2d 176, 181 (1st Cir. 1989); see Burlington N. & Santa Fe Ry. Co.
v. United States, 129 S. Ct. 1870, 1881 (2009) (CERCLA “did not mandate ‘joint and several’ liability in
every case.”) (citation omitted); United States v. Chem-Dyne Corp., 572 F. Supp. 802, 810 (S.D. Ohio
1983) (rejecting “blanket adoption” of joint and several liability).
161. Burlington Northern, 129 S. Ct. at 1881 (quoting United States v. Hercules, Inc., 247 F.3d
706, 717 (8th Cir. 2001)).
162. Chem-Dyne Corp., 572 F. Supp. at 810 (citing RESTATEMENT (SECOND) OF TORTS §§
433A, 881 (1966)).
163. “The original meaning of a ‘joint tort’ was that of vicarious liability for concerted action.”
WILLIAM L. PROSSER, LAW OF TORTS 291 (4th ed. 1971); see RESTATEMENT (THIRD) OF TORTS:
APPORTIONMENT OF LIABILITY § 15 cmt. a (2000) (“[J]oint and several liability for persons engaged in
concerted action applies regardless of the rule regarding joint and several or several liability for
independent negligent tortfeasors in the jurisdiction.”).
164. RESTATEMENT (SECOND) OF TORTS § 876 (1966).
338
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
(c) gives substantial assistance to the other in
accomplishing a tortious result and his own conduct,
separately considered, constitutes a breach of duty to the
third person.165
From Chem-Dyne on down, courts in CERCLA cases have simply
assumed the independence of action of multiple liable parties.166 Plaintiffs
have had little incentive to challenge the courts to do otherwise, in light of
the failure rate of attempts to prove a reasonable basis for apportionment.167
As a result, concert of action theories have not been discussed.168 But if
Burlington Northern has really changed the game on proof of a basis for
apportionment under section 433A, then both plaintiffs and courts must
examine whether that section’s underlying predicate of independent action
is truly satisfied.
165. Id. The Third Restatement similarly states that each tortfeasor acting in concert with others
is jointly and severally liable for the liability of all. RESTATEMENT (THIRD) OF TORTS: APPORTIONMENT
OF LIABILITY § 15 (2000). The Third Restatement takes no position on whether those acting in concert
are jointly and severally liable with others acting independently, leaving that to the jurisdiction’s
selection of one of five “tracks” in use after adoption of comparative responsibility. Id. cmt. a. That
section “does not address the rules regarding when concerted activity exists,” id. cmt. o, although the
reporters appear to endorse the Second Restatement’s formulation. Id. (quoting RESTATEMENT
(SECOND) OF TORTS § 876 (1966)).
166. When CERCLA defendants raised lack of concert of action as a basis for avoiding joint and
several liability, courts correctly held that independent actors can be jointly and severally liable under
CERCLA by application of the principles of section 433A. E.g., Kelley v. Thomas Solvent Co., 714 F.
Supp. 1439, 1449 (W.D. Mich. 1989) (Striking affirmative defense alleging “liability . . . is not joint
with other defendants because . . . there is no concert of action,” because “[c]oncert of action has no
relevance to the decision about whether to impose joint and several liability.”).
167. So long as a reasonable basis for apportionment remains hard to prove, it is easier for a
CERCLA plaintiff to achieve joint and several liability by relying on the defendants’ failure to satisfy
the burden that section 433B imposes than by trying to prove concert of action.
168. See generally Bulk Distrib. Ctrs., Inc. v. Monsanto Co., 589 F. Supp. 1437, 1443 (S.D.
Fla.1984) (“[I]f they worked in concert to produce a single indivisible release or threatened release, then
they may be held jointly and severally liable for the claimant’s response costs if the applicable state law
so provides.”). A few cases address whether concert of action could serve as the sole basis for imposing
liability in the first instance, e.g., Grine v. Coombs, No. 95-342 ERIE, 1997 U.S. Dist. LEXIS 19690,
*11–12 (W.D. Pa. 1997) (allegation that defendant arranged for disposal of hazardous substance by
conspiring with others sufficient to survive motion to dismiss), or the sufficiency of evidence of
concerted action, e.g., Gen. Elec. Co. v. AAMCO Transmissions, Inc., 962 F.2d 281, 288 (2d Cir. 1992)
(stating that where oil companies, who leased underground storage tanks to independent gas station
operators who used tanks to store contaminated waste oil subsequently disposed of at Superfund site,
did not participate in disposal or know of disposal practices, common law aider and abetter liability
under Second Restatement § 876(b) could not apply, so court did not decide “whether common law
doctrines can . . . supplement” CERCLA liability scheme); New Jersey Tpk. Auth. v. PPG Indus., Inc.,
16 F. Supp. 2d 460, 471 (D.N.J. 1998) (stating that the theory of “enterprise (or ‘concert-of-action’)
liability” did not apply where plaintiff could not show which of several defendants’ chromium ore waste
had been deposited at any of several facilities and did not contend or produce evidence showing there
was “any sort of common plan” to dispose of such waste at those facilities).
2009] Several Approaches to Divisibility After Burlington Northern
339
The Restatement gives the example of a group of people committing
intentional torts as the paradigmatic case of tortfeasors acting in concert,169
but concert of action theories have long been applied to negligence170 and,
more recently, to strict liability actions as well.171 A uniform law of the
scope of CERCLA liability based on “traditional and evolving principles of
common law,” applied in the context of the statutory goals, should apply
concert of action principles to impose joint and several liability on multiple
CERCLA defendants in appropriate factual circumstances.
How might this work? Assume arguendo that the facts of the MSL
hypothetical would support an apportionment under section 433A.172 In
that case, the very best outcome a plaintiff could hope for would be
apportionment of 52% of the liability to the three viable arrangers, based on
their share of the total volume of HOS dumped. The actual outcome would
probably yield a much lower percentage recovery to the plaintiff, for a court
truly apportioning liability on a purely causal basis (without considering
equities or the effect of certain defendants’ insolvency on the plaintiff)
would be hard-pressed to ignore the causal contribution of the defunct
owner-operators and transporters.
A concert of action theory, however, could and should restore a measure
of joint and several liability to this situation. At a minimum, A1 should be
liable for the liability apportioned to O1 and O2 as well as to itself; A2
should be liable for the liability apportioned to T1 as well as to itself; and
A4 should be liable for the liability apportioned to T2 as well as to itself.173
In each instance the arranger defendant “act[ed] in accordance with an
agreement to cooperate in a particular line of conduct or to accomplish a
particular result.”174 The line of conduct or particular result was the
disposal of hazardous substances belonging to the arranger, for that is what
169. See RESTATEMENT (SECOND) OF TORTS § 876 cmt. a, illus. 1 (1966).
170. E.g., Moore v. Foster, 180 So. 73, 74 (Miss. 1938) (applying joint liability to a negligence
claim); Oliver v. Miles, 110 So. 666, 668 (Miss. 1926) (applying joint liability to a negligence claim).
Both decisions are cited in RESTATEMENT OF (SECOND) TORTS § 876 Rptrs’ Note, cmt. b (1966).
171. E.g., Roney v. Gencorp, 431 F. Supp. 2d 622, 633–34 (S.D. W. Va. 2006) (applying West
Virginia law, denying motion to dismiss product liability claim alleging “aiding and abetting” concert of
action as described in subsection b of section 876 of the Second Restatement). In the Restatement, the
American Law Institute took “no position on whether the” concert of action rules apply to cases where
the conduct is neither intentionally tortious or negligent but “involves strict liability for the resulting
harm.” RESTATEMENT (SECOND) OF TORTS § 876 Caveat (1966).
172. The assumption is not valid.
173. See RESTATEMENT (THIRD) OF TORTS: APPORTIONMENT OF LIABILITY § 15 cmt. a (2000)
(Explaining that even where joint and severable liability has been replaced or modified by comparative
responsibility, “joint and several liability for persons engaged in concerted action applies regardless of
the rule . . . for independent negligent tortfeasors.”).
174. RESTATEMENT (SECOND) OF TORTS § 876 cmt. a (1966) (defining “acting in concert”).
340
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
creates the CERCLA liability. It should not matter that no party to any of
the transactions anticipated the eventual release of hazardous substances,
much less the eventual incurrence of response costs.175 Alternatively, each
of the three arrangers should be liable for additional apportioned shares as
described above under either or both of the “aiding and abetting” prongs of
the Second Restatement’s concert of action rule.176 In each case, the
arranger’s conduct gave “substantial assistance” to the liability-creating
conduct of the owner/operator or transporter, and vice versa.
On the same theory, the liability of the transporters, T1 and T2, includes
not only their individually-apportioned shares but also the liability
apportioned to the owner/operators, O1 and O2. Therefore, A2 and A4, who
are liable for the transporters’ apportioned liability as well as their own,
should also be jointly and severally liable for liability apportioned to the
owner/operators. This is only fair to A1, who contracted with the dump
owner directly and used A1’s own trucks to haul waste there. There is no
reason A1 should be liable for the potentially large causal share assigned to
the owner/operators while A2 and A4 should avoid that share because they
used an intermediary to participate in the same course of conduct (disposing
of hazardous substances in the dump).177
The viable arrangers would presumably object that holding them liable
for the shares of transporters and owner/operators is unfair and improper:
once the court has apportioned individual causal shares to A1, A2, and A4,
each of them should be liable for that causal share and no more, as section
433A envisions. Such an objection should not succeed. The theory behind
apportionment under section 433A is that each independent actor should be
liable only for the amount of harm it caused, if that amount can be
determined. The theory behind joint and several liability under section 876
is that an actor should be liable even for harm that it demonstrably did not
cause—even an entirely distinct harm—if the harm was caused by a person
acting in concert. Joint and several liability under a concerted action theory
trumps causal apportionment, not the other way around.178
175. In the drag racing cases, the “particular line of conduct” or “particular result” accomplished
in concert is the race, not the eventual collision between one of the racing cars and a third car. Yet both
participants in the race are jointly and severally liable for the resulting harm. RESTATEMENT (SECOND)
OF TORTS § 876 cmt. a, illus. 2 (1966).
176. Id. § 876(b), (c) (1966).
177. By contrast, making A2 and A4 (but not A1) liable for the apportioned liability of T1 and
T2 is not unfair. The causal apportionment presumably would have initially assigned a percentage of
liability to owner/operators versus transporters versus arrangers, and A1, as its own “transporter,” would
already have been assigned an aliquot of liability for its participation in that role.
178. See Reilly v. Anderson, 727 N.W.2d 102, 112 (Iowa 2006) (Concerted actors do not
commit independent tortious conduct and thus remain jointly and severally liable with one another.);
2009] Several Approaches to Divisibility After Burlington Northern
341
Consider, hypothetically, a site without transporters or liable
owner/operators: a remote farm, for example, where despite the farm
owner’s adequate precautions and due care,179 our hypothetical arrangers A1
through A8 manage to engage in midnight dumping using their own trucks.
Assuming, again, that their relative volumetric contributions provide a
reasonable basis for apportionment under section 433A, a plaintiff that
cleaned up the farm would be able to recover 52% of its costs from viable
liable parties—not nearly so nice as the 100% recovery that joint and
several liability would have provided, but still respectable. Adding
transporter or owner/operator participants to the enterprise of getting rid of
these arrangers’ hazardous substances should not allow the arrangers to
evade liability180 for costs that they would have borne in the absence of
other participants.181 It certainly should not serve to transfer those costs to a
plaintiff in the event that the other participants are judgment-proof at
collection time.
A more serious objection would be that various elements of concert of
action in traditional tort claims are missing from the hypothetical dump
case. The Restatement’s aiding and abetting provisions, for example,
require either knowledge that the other tortfeasor’s conduct is a breach of
duty or that the aider and abetter’s conduct is in its own right a breach of
duty to the victim. So, it would be argued, A4 could not be liable for the
share assigned to T2, and a fortiori the shares assigned to O1 and O2, unless
A4 knew that T2 was going to bring the waste to the dump, or unless A4
somehow acted wrongfully in consigning the waste to T2. Similarly, one
might argue that the types of arms-length commercial transactions between
arrangers, transporters, and owner/operators have hardly been considered
“substantial assistance or encouragement” in tort law, much less part of “a
common design.”
RESTATEMENT (THIRD) OF TORTS: APPORTIONMENT OF LIABILITY § 15 cmt. a (2000) (stating that
persons engaged in concerted action bear joint and several liability even where independent tortfeasors
would not).
179. The owner would thus qualify for CERCLA’s third-party defense. 42 U.S.C. § 9607(b)(3)
(2006).
180. “Evading liability” in this context means liability to the plaintiffs. The statutory
contribution remedy would remain available to the arranger defendants if a transporter or
owner/operator had assets from which to pay contribution.
181. Perverse incentives would arise, in a CERCLA regime in which apportionment were
routine, if a party’s apportioned liability were reduced by the presence in the case of other parties with a
different causal relationship to the same hazardous substances. A party arranging for disposal of waste,
for example, might prefer to contract with multiple transporters and disposal site operators, on the
theory that if something were to go wrong, the future CERCLA liability would be shared. The cause of
proper hazardous-substance management might be better served by encouraging responsible arrangers to
“take ownership” of their waste and minimize risk of releases.
342
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
These arguments might have much force if a plaintiff that incurred
response costs were to try to recover them as damages for a common law
tort. But, of course, such a plaintiff would sue under CERCLA, and courts
refer to common law only as a source of interpretation and interstitial
completion of the statutory scheme. Just as courts do not impose the
“substantial factor” burden on CERCLA plaintiffs because doing so would
be inconsistent with the statutory scheme, it would be inappropriate to
require a showing of a tortious act or breach of duty before finding concert
of action sufficient to hold the concerted actors jointly and severally liable
for the harm apportioned to all. The appropriate standard is the statutory
one. Did the defendant arrange for disposal or transport, or own or operate
a facility, in concert with another liable person? Did the defendant know
that the liable person, to whom defendant gave substantial assistance or
encouragement, was engaging in the conduct (or role) that gave rise to
CERCLA liability? Did the defendant, who gave substantial assistance or
encouragement to another liable party, itself engage in conduct (or role) that
gave rise to CERCLA liability?
Congress has, in this regard, provided courts with a strong hint of what
might constitute concerted action among responsible parties under
CERCLA. Congress addressed causation in section 107(b) of the statute,
the exclusive affirmative defenses to CERCLA liability.182 In a major
departure from the common law, Congress crafted CERCLA’s third-party
defense extremely narrowly.183 For a posited third-party cause to absolve an
otherwise responsible party of liability, the release of hazardous substances
must be “caused solely by”184 “an act or omission of a third party other than
. . . one whose act or omission occurs in connection with a contractual
relationship, existing directly or indirectly, with the defendant . . . .”185
182. § 9607(b). The defenses allow an otherwise responsible person to avoid liability by
proving that the release was “solely” caused by an act of God, an act of war, or an act or omission of a
third party. Id.
183. The statutory defense is so different from common law causation rules that courts routinely
struck from answers, as legally insufficient, alleged common law defenses such as “no proximate
cause,” “intervening cause,” etc. E.g., United States v. Kramer, 757 F. Supp. 397, 418 (D.N.J. 1991).
184. § 9607(b).
185. § 9607(b)(3). To establish the defense, a defendant also must demonstrate that it exercised
due care and took precautions. Id. Because parties liable under CERCLA are apt to have in some way
contributed to causing the release of hazardous substances (e.g., by arranging for disposal of a hazardous
substance at the facility) or to be in some contractual relationship (e.g., a contract for dumping or
hauling services or for the rental or purchase of real estate) with the putative third-party cause,
successful assertions of the third-party defense have been rare. The primary exception has been for socalled “innocent landowners” who acquire previously contaminated property, e.g., New York v. Lashins
Arcade Co., 91 F.3d 353, 360 (2d Cir. 1996) (finding that a recent purchaser of contaminated property
established third-party defense where prior owner’s act or omission was not “in connection with” the
purchase contract). Congress enacted special provisions for such landowners, 42 U.S.C. § 9601(35)(A)
2009] Several Approaches to Divisibility After Burlington Northern
343
This tells us just how far Congress thought it appropriate, for purposes
of a cost recovery action, to impute the causal relation between a
defendant’s status as a “covered person” and a release of hazardous
substances which causes the incurrence of response costs. A person
remains liable, not just for the acts of an employee or agent, but even of
independent contractors that cause the release—even if the contractor’s
relationship with the liable party exists only “indirectly.” It is true that
section 107(b)(3) directly addresses only the existence of liability, not the
scope of liability. But Congress’s willingness to extend liability through
multiple contractual links suggests strongly that a comparably broad
interpretation of the requirements for concerted action under CERCLA
would be consistent with the statutory goals.186 And even though concert of
action theories have not yet found application in CERCLA jurisprudence,
there is precedent that would support their use.187
O’Neil v. Q.L.C.R.I., Inc.188 involved land from which a faulty septic
system leaked raw sewage into a river. The state alleged that the landowner
was liable under the Federal Water Pollution Control Act189 (FWPCA) for
the illegal discharge of a pollutant.190 The state also sued a credit union that
held a mortgage on the property, even though the state could not allege that
the credit union itself discharged a pollutant. Instead, the state sought “to
use the common law concept of aiding and abetting to find [the credit
union] in violation of federal . . . statutory . . . law.”191 The state alleged
that the credit union aided and abetted the unlawful discharge by
participating in “straw conveyances” and by making the loans without
requiring the borrower to correct the sewage discharge, even though the
credit union knew about the discharge.192 The court held that the complaint
stated a claim for relief against the credit union:
(enacted by Pub. L. 99-499 (1986); amended by Pub. L. 107-118 (2002)). Congress later created other
exemptions from liability. See supra note 104 and accompanying text.
186. CERCLA is a remedial statute that is broadly construed to effectuate its purposes. E.g.,
United States v. Carolina Transformer Co., 978 F.2d 832, 838 (4th Cir. 1992); Colorado v. Idarado
Mining Co., 916 F.2d 1486, 1493 (10th Cir. 1990).
187. The very few court decisions that even refer to concerted action in connection with
CERCLA claims are described supra note 168. These opinions do not hold that concert of action
theories are unavailable under CERCLA, even though they also do not include a successful use of such a
theory.
188. O’Neil v. Q.L.C.R.I., Inc., 750 F. Supp. 551 (D.R.I. 1990).
189. Federal Water Pollution Control Act, 33 U.S.C. §§ 1251–1387 (2006) (more commonly
known as the Clean Water Act).
190. See Friends of the Sakonnet v. Dutra, 738 F. Supp. 623, 627 (D.R.I. 1990) (describing facts
relating to the discharge).
191. Q.L.C.R.I., Inc., 750 F. Supp. at 554 (citing RESTATEMENT (SECOND) OF TORTS § 876
(1966)).
192. Id.
344
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
[T]his Court must determine whether the common law
concept of aiding and abetting can be used to determine
liability under the FWPCA. The applicability of common
law doctrines in litigation under federal statutes depends on
whether these principles advance the goals of the particular
federal statute which plaintiffs allege has been violated
. . . . [A]iding and abetting is a concept that can be used to
interpret “in violation” under [33 U.S.C.] § 1365 [the
citizen suit provision of the Clean Water Act]. Nothing in
the FWPCA rules out the use of the aiding and abetting
doctrine . . . . Plaintiff has alleged sufficient involvement
and possible control on the part of [the credit union] to
present a viable claim of aiding and abetting.193
This analysis of why aiding and abetting may be used as a basis for
liability under FWPCA applies even more strongly to assessing whether the
common law concert of action doctrine (which includes aiding and
abetting) may be used to determine the scope of CERCLA liability. All
courts, including the Supreme Court, have recognized that Congress
intended for courts to refer to common law for the scope-of-liability
determination. Joint and several liability for persons acting in concert is
consistent with common law principles as well as with CERCLA, and will
further CERCLA’s well-understood objectives.
Applying concert of action to CERCLA as proposed herein would
sometimes impose joint and several liability even if the divisibility standard
of section 433A(1)(b) could otherwise be met. But it would not result
inevitably in the imposition of joint and several liability in all cases of
otherwise divisible harm.194
First, it is hard to see how, in many cases, a court could find concert of
action among multiple defendants in the same category of responsible
persons. In the dump hypothetical, the arrangers A1, A2, and A4 contracted
independently with O1/O2, T1, and T2, respectively, to dump or haul each
arranger’s waste. T1 and T2, similarly, each acted in concert with the
owner/operator parties and with its respective arranger customers, but not in
any obvious way with the other transporters.195
193. Id. at 555 (quoting Petro-Tech, Inc. v. Western Co. of N. Am., 824 F.2d 1349 (3d Cir. 1987)
(internal quotation marks and brackets omitted).
194. It would also not affect cases of distinct harms. See supra note 160 (discussing the nonuniversal application of joint and several liability).
195. Of course, in a given case, a plaintiff might develop facts tending to show that arrangers
(for example) acted in concert with other arrangers. Different arrangers might, for example, have
2009] Several Approaches to Divisibility After Burlington Northern
345
Second, in some factual scenarios even “vertical” concert of action
among arrangers, transporters, and owner/operators would lead to less than
complete joint and several liability. For example, consider a site at which,
first, an owner/operator created a waste-oil lagoon in which wastes from
various arrangers and transporters were mixed. Subsequently, a second
owner/operator who purchased the property rented storage tanks to various
industrial customers, some of the tanks’ substances leaked into the lagoon.
Assuming, again, that these parties first established a reasonable basis for
apportionment under section 433A, concert of action might hold the wasteoil arrangers and transporters jointly and severally liable for the first
owner’s share, and the tank-farm customers jointly and severally liable for
the second owner’s share. But it is hard to see how a concert of action
theory could allow the liability apportioned to the parties involved in the
waste-oil operation to be attributed to the parties involved in the tank-rental
operation, and vice versa.
Thus, concert of action theories would be legally appropriate and could
enable CERCLA plaintiffs to achieve greater cost recoveries in cases where
defendants met their burden of proof on apportionment under Burlington
Northern and some liability had been apportioned to insolvent parties.
However, concert of action theories would not necessarily result in a
defendant being jointly and severally liable for all response costs at a site.
In that respect, of course, being relegated to concert of action would be
substantially less desirable for plaintiffs than the prevailing legal regime
before Burlington Northern, in which defendants tried, but generally failed,
to avoid joint and several liability under section 433A. Allocation of the
burden of proof is another critical difference.
As all CERCLA
jurisprudence applying the Second Restatement has held, once a plaintiff
proves a defendant liable—a relatively light burden under CERCLA—the
liability is joint and several unless the defendant satisfies the burden of
justifying apportionment per section 433A.196 Concert of action, by
contrast, is an element of the case for a plaintiff who wishes to rely on it.197
Yet, of course, a CERCLA plaintiff need not allege concert of action in
order to state a claim that any individual owner/operator, arranger, or
transporter is liable under section 107(a) of CERCLA.
encouraged one another to use a particular transporter or disposal facility, or formed some type of joint
venture to achieve economies of scale in disposing of a particular waste generated by a particular
industry in a particular locality.
196. Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870, 1881 (2009);
RESTATEMENT (SECOND) OF TORTS § 433B (1966).
197. See, e.g., Roney v. Gencorp, 431 F. Supp. 2d 622, 634 (S.D. W. Va. 2006) (“While Plaintiff
may not be able to prove this ‘concert of action’ theory of liability . . . Plaintiff has alleged sufficient
facts to survive a motion to dismiss.”).
346
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
It would make sense, then, for CERCLA plaintiffs to continue to allege
joint and several liability, but to plead concerted action in the alternative in
the event that the court otherwise holds defendants’ liability to be several
only.
Defendants would retain the burden of proof on
divisibility/apportionment and would need to satisfy that burden to avoid
imposition of joint and several liability. If defendants satisfied the burden
on divisibility/apportionment, a plaintiff could attempt to prove concert of
action among one or more groups of defendants. As outlined above, in
many cases, establishing this proof would not be terribly difficult: it would
consist of the same evidence adduced to establish a defendant’s liability in
the first instance.198 If plaintiff succeeded, the court would hold each
member of each concerted action group jointly and severally liable for the
summed apportioned liability of the group.
Concert of action theories would not re-create a prior legal regime in
which CERCLA responsible parties’ liability was “typically joint and
several.”199 Concert of action theories would, however, blunt the
deleterious effects of a regime in which apportionment was too easy to
achieve, sometimes dramatically. Consider Burlington Northern itself, and
assume (counterfactually) that Shell remained liable as an arranger. A
strong case could have been made that Shell acted in concert with the nowpenniless Brown & Bryant. The success of that case would have made the
plaintiffs whole, notwithstanding the small share of liability apportioned to
the Railroads.
B. Of Two Lethal Wounds: Multiple Sufficient Causes
The premise of apportioned liability is that a defendant should be liable
only for harm that it has caused, and not for harm that others have caused.200
In Burlington Northern, for instance, both the district court and the
Supreme Court gagged on the notion that the Railroads should be liable for
the consequences of the activities that Brown & Bryant conducted on
Brown & Bryant’s own parcel. But what if each of the multiple defendants
caused all of the harm—or would have, in the absence of their codefendants?
198. For example, if an arranger brought hazardous substances to a dump in its own trucks,
admission of the dump tickets would demonstrate an express agreement with the owner/operator. If
there were no dump tickets, testimonial proof of the dumping would establish a tacit agreement, except
in a case of “midnight dumping.”
199. United States v. Hercules, Inc., 247 F.3d 706, at 715 (8th Cir. 2001).
200. See RESTATEMENT (SECOND) OF TORTS § 881 cmt. a (In cases of divisible harm, “each
[tortfeasor] is held liable only for the proportion of the total harm for which he is himself responsible.”).
2009] Several Approaches to Divisibility After Burlington Northern
347
This situation faced the only district court that, as of this writing, has
attempted to apply Burlington Northern’s apportionment ruling to another
CERCLA claim. The result in Reichhold, Inc. v. United States Metals
Refining Co.,201 issued by a court in an influential district that sees
considerable CERCLA litigation, is not encouraging for those who would
argue that the reach of Burlington Northern is closely cabined by its facts.
Reichhold, Inc. bought industrial property in Carteret, New Jersey from
United States Metals Refining Co. (USMRC). USMRC’s operations (as the
district court later found) had contaminated large swaths of the site with
significant quantities of metals. Reichhold later sold the property, but the
state environmental agency nevertheless required Reichhold to investigate
and clean up the mess. Reichhold recouped some of its response costs from
USMRC’s parent company in a settlement. The settlement agreement
included a reservation of rights or “re-opener” under which Reichhold
could sue USMRC for costs relating to any “Material New Environmental
Obligations” relating to the site. After the agency demanded several
additional response actions, Reichhold invoked the re-opener202 and sued
USMRC for cost recovery under CERCLA and the New Jersey Spill Act.
USMRC asserted various defenses and counterclaimed for contribution.
After a bench trial, Senior Judge Dickinson R. Debevoise found that
USMRC “had deposited vast amounts” of hazardous substances on the
property before the sale to Reichhold, and that Reichhold had neither
brought new contaminated fill to the property nor spread the pre-existing
contamination203 during its operations or cleanup. Thus, the court held
USMRC liable for the cost of most of the response actions subject to the reopener.204
The court found, however, that in one contaminated area, USMRC’s
operations were not the only source of contaminated fill. In that area, a
subsequent landowner or tenant205 placed two or three feet of fill on top of
201. Reichhold, Inc. v. United States Metals Refining Co., No. 03-453(DRD), 2009 U.S. Dist.
LEXIS 52471, at *131–33 (D.N.J. 2009).
202. Much of the district court’s opinion is devoted to determining whether each of several
response actions, for which Reichhold sought recovery, satisfied the conditions that the settlement
agreement’s elaborate definition of “Material New Environmental Obligation” required for invoking the
re-opener. Reichhold, 2009 U.S. Dist. LEXIS 52471, at *11–12.
203. Moving pre-existing material can be “disposal” under CERCLA. Kaiser Aluminum &
Chem. Corp. v. Catellus Dev. Corp., 976 F.2d 1338, 1342 (9th Cir. 1992).
204. The court held that a few of the additional response actions were not “Material New
Environmental Obligations” subject to the re-opener; the settlement had resolved USMRC’s liability to
Reichhold for those. Reichhold, 2009 U.S. Dist. LEXIS 52471, at *5.
205. Several entities owned or occupied the property after Reichhold. Id. at *132. The opinion
does not identify which one deposited the additional contaminated fill—presumably because that party
348
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
the contaminated slag that USMRC had previously deposited. Judge
Debevoise found that the new fill was contaminated as well, and that either
the pre-existing fill placed by USMRC or the new fill placed by the
subsequent owner or tenant would have required the additional response
action—a cap over the contaminated area—for which Reichhold sought
cost recovery from USMRC.
On those facts, the district court held that “the metals contamination [of
the area in question] was a distinct or single harm that USMRC and a third
party caused.” Burlington Northern, the court stated, “suggests that this
situation might be addressed by apportionment rather than equitable
principles.” Attempting to apply section 433A of the Second Restatement
and the teaching of Burlington Northern, the court held:
There is a reasonable basis for division according to the
contribution of each. The measurement is not the exact
amount of metals contamination for which each was
responsible; USMRC was undoubtedly the source of most
of it. Rather, it is the circumstances that each was
responsible for a sufficient amount of metals contamination
that required the cap . . . . [B]ecause under CERCLA,
USMRC would be responsible for only half these costs [of
the cap], Reichhold will be able to collect from USMRC
only one-half of its past and future expenditures in
connection with the . . . [c]ap.206
Like Judge Wanger in Burlington Northern, Judge Debevoise
apparently arrived at an apportionment sua sponte.207 The effort to apply
brand-new Supreme Court precedent was commendable, but the holding
was unsound.
The trial court in Reichhold found, as a matter of fact, that either the
USMRC slag or the later-added fill was sufficient, alone, to have caused
remediation by capping. This type of over-determined causation posed a
conundrum for the common law, which responded by modifying the
traditional “but for” test for causation in fact so that “[i]f two forces are
actively operating, one because of the actor’s negligence, the other not
was not identified in the record, either because the litigants did not know its identity or because they
chose not to present it to the district court despite knowing it.
206. Id. at *131–32.
207. The non-jury trial in Reichhold ended on March 4, 2009. Id. at *3. The Supreme Court
issued Burlington Northern on May 4, 2009. Burlington N. & Santa Fe Ry. Co. v. United States, 129 S.
Ct. 1870 (2009). Judge Debevoise issued the Reichhold opinion on June 22, 2009. The opinion makes
no reference to any argument about apportionment by a litigant.
2009] Several Approaches to Divisibility After Burlington Northern
349
because of any misconduct on his part, and each of itself is sufficient to
bring about harm to another, the actor’s negligence may be found to be a
substantial factor in bringing it about.”208
The classic case involves two fires that combine to burn a plaintiff’s
property.209 If either fire alone would have done the harm, each “is
regarded as a factual cause of the harm”210—a “substantial factor” in
bringing the harm about, in the Second Restatement formulation.
To say that each causally sufficient fire (or tortious act, or arrangement
for disposal) is a cause of the harm under section 432(2) of the Second
Restatement does not necessarily imply that the person who set each fire
should be jointly and severally liable for the entire harm under section
433A. Nevertheless, joint and several liability among multiple sufficient
causes makes sense.
If the predicate of causal sufficiency is correct—either fire alone would
have burned the plaintiff’s property—then each fire’s contribution to the
harm is not proportional to its size before it merged with the other. It
cannot be said that each fire merely contributed some fraction of the heat
and flame that, only when combined with the other fire’s fractional
contribution of heat and flame, burned plaintiff’s property. What harm did
the person who set each fire cause? Each caused the burning of the
property, not the burning of a portion of the property. The fairness-based
argument for apportionment, that no actor should be charged with liability
for harm it did not cause, does not support apportionment among multiple
sufficient causes.211
This analysis tracks precisely the reasoning of the Minnesota Supreme
Court in the type specimen of multiple sufficient cause cases.212 The
208. RESTATEMENT (SECOND) OF TORTS § 432(2) (1966).
209. Id. at illus. 3 (“Two fires are negligently set by separate acts of the A and B Railway
Companies . . . . The normal spread of either fire would have been sufficient to burn [C’s property] . . . .
It may be found that the negligence of either the A or the B Company or of both is a substantial factor in
bringing about C’s harm.”).
210. RESTATEMENT (THIRD) OF TORTS: LIABILITY FOR PHYSICAL HARM § 27 (Proposed Final
Draft No. 1, 2005) (“If multiple acts exist, each of which alone would have been a [but for] factual cause
under § 26 of the physical harm at the same time, each act is regarded as a factual cause of the harm.”).
211. See RESTATEMENT (SECOND) OF TORTS § 881 (1966) (“If two or more persons, acting
independently, tortiously cause distinct harms or a 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.”); cf. RESTATEMENT (THIRD) OF TORTS: APPORTIONMENT OF LIABILITY
§ 10 cmt. a (2000). One justification for joint and several liability is that each defendant’s tortious
conduct is the legal cause “of the entirety of the plaintiff’s damages,” although if comparative
responsibility is involved, plaintiff’s conduct is a cause of all the harm as well. Id.
212. Anderson v. Minneapolis, St. Paul & Sault Ste. Marie Ry. Co., 179 N.W. 45, 46 (Minn.
1920). See RESTATEMENT (SECOND) OF TORTS § 432 Rptrs’ Note, 3 App. (1966) (“Illustration 3 is based
on Anderson” and other cases).
350
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
plaintiff alleged that sparks from a railroad’s engines caused one or more
fires that the railroad failed to extinguish, and that those fires eventually
flared up and spread to plaintiff’s property. The day they did, however,
several “great fires . . . swept through Northeast Minnesota,” including
plaintiff’s property.213 The defendant argued that it should not be liable if
the damage was caused by the combination of a fire it set and others “of no
responsible origin, but of such sufficient or superior force that they would
have produced the damage to plaintiff's property, regardless of the fire
pleaded.”214 The court responded that “[i]f a fire set by the engine of one
railroad company unites with a fire set by the engine of another company,
there is joint and several liability, even though either fire would have
destroyed plaintiff's property”; therefore it would not absolve a railroad of
liability if the second fire happened to be “of no responsible origin” instead
of being set by another railroad.215 The court affirmed the plaintiff’s
verdict. The court did not suggest that damages were reduced because of
the other “innocent” causes.
The Second Restatement treats multiple sufficient causes explicitly
only as an exception to “but for” causation,216 and does not discuss how this
exception coheres with the causal apportionment principles of section
433A. The comments to section 433A open with a seemingly broad
statement of applicability: “The rules stated in this Section apply whenever
two or more causes have combined to bring about harm to the plaintiff, and
each has been a substantial factor in producing the harm, as stated in
sections 431 and 433.”217 But nothing in section 433A or its accompanying
commentary suggests that a single harm can or must be apportioned among
multiple sufficient causes.
The section 432(2) description of multiple independent causes— “two
forces are actively operating . . . and each of itself is sufficient to bring
about harm to another”—is not literally within the ambit of “whenever two
or more causes have combined to bring about harm,” and the comment to
section 433A notably omits any reference to section 432(2).218 The Second
Restatement gives thirteen illustrations that deem apportionment
appropriate, and not one involves a clear case of multiple sufficient
213. Anderson, 179 N.W. at 46.
214. Id. at 49.
215. Id.
216. Compare RESTATEMENT (SECOND) OF TORTS § 432(2) (1966), with id. § 432(1).
217. Id. § 433A cmt. a (1966).
218. See generally RESTATEMENT (THIRD) OF TORTS: LIABILITY FOR PHYSICAL HARM § 27 cmt.
i (1999) (Proposed Final Draft No. 1, 2005) (distinguishing multiple causes from multiple sufficient
causes).
2009] Several Approaches to Divisibility After Burlington Northern
351
causes.219 The conspicuous absence of combining fire cases from these
illustrations suggests strongly that at common law and under the Second
Restatement, parties responsible for multiple sufficient causes of harm
faced joint and several liability for the entire resulting harm. Even some
common law pollution cases that otherwise held apportionment appropriate
suggested that parties should be held jointly and severally liable for the
entire harm.220
In the CERCLA context, statutory objectives reinforce the common
law’s reasoning. To shift costs from the government to responsible parties,
Congress made parties liable for response costs under CERCLA even
without proof that their involvement with a facility (as owner, operator,
transporter, or arranger) caused the environmental damage addressed by the
response action. It follows that Congress surely intended to shift costs to
any party that could be shown to have caused, all on its own, the need for
an entire cleanup. It would be inconsistent with CERCLA for such a party’s
liability to be reduced simply because another party did the same thing—
especially in a case where, because of the other party’s insolvency, the
result would be to shift costs back to the taxpayer.
219. Illustrations that relate to separate torts that cause separate injuries (including successive
episodes of water pollution that deprive a plaintiff of the use of the water at different times) are
obviously distinguishable from multiple sufficient causes of the same harm. RESTATEMENT (SECOND)
OF TORTS § 433A(1) (1966); see id. cmts. b, c, f, & illus. 1, 2, 8, 10, 11. In the animal-trespass
examples, it is clear that each tortfeasor’s trespassing animals did some of the damage, and that neither
tortfeasor’s animals acting alone would have caused all of the harm. Id. cmt. d, illus. 3. The flooding
examples constitute the clear case in which each cause literally creates only a fraction of the damage:
the extent of a flood is proportionate to the amount of water released. Id. § 433A cmts. d, e, f, & illus. 4,
6, 9. The mine-debris example is exactly analogous to flooding. Id. § 881 cmt. d, illus. 1. The private
nuisance example of smoke from a roundhouse posits explicitly that “reasonable operation of the
roundhouse . . . would have caused one-third of the smoke and interference,” so neither the reasonably
emitted smoke nor the tortious smoke was a sufficient cause of the entire harm. Id. § 433A cmt. e, illus.
7. Similarly, the example of smelter fumes specifies that one of the sources would by itself cause no
harm at all, much less all of the harm. Id. § 881 cmt. d, illus. 2. The private nuisance example of
concurrent water pollution in which oil pollution, 70% discharged by A and 30% discharged by B,
“deprived [plaintiff] of the use of the water for his own industrial purposes” is the only illustration in
which it is even conceivable that each cause was a sufficient one. Id. § 433A cmt. d, illus. 5. But the
illustration does not say so, and the sufficiency of each cause is not logically compelled—the water
might have been usable for industrial purposes if polluted by only a fraction of the oil. The intellectual
underpinning of section 433A, and the context of the other illustrations, suggest powerfully that the
illustration conceives that each oil discharge separately would have caused actual harm in proportion to
the amount of oil discharged, rather than causing the same harm that resulted from the combined
discharges.
220. E.g., Mitchell Realty Co. v. City of W. Allis, 199 N.W. 390, 395 (Wis. 1924)
(distinguishing that case from situations in which each wrongful act would produce the identical
results); Pulaski Anthracite Coal Co. v. Gibboney Sand Bar Co., 66 S.E. 73, 74 (Va. 1909) (stating that
several liability is the rule among multiple independent tortfeasors, provided “neither being sufficient to
produce the entire loss . . . .”).
352
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
Multiple sufficient causes may exist commonly in CERCLA cases
where response action is triggered by environmental conditions that exceed
a certain threshold.221 In the MSL hypothetical, if twenty percent of the
total amount of HOS placed in the landfill would have made the
concentration of HOS in the groundwater high enough to require the
remedy, many of the liable parties should be considered multiple sufficient
“causes” in the context of the CERCLA liability scheme: O1 and O2, each
of whom owned and operated the landfill when more than twenty percent of
the HOS was disposed of there; T1, who transported more than twenty
percent of the HOS to the dump; and A1, A2, and A3, each of whom
arranged for disposal of at least twenty percent of the HOS dumped.222 The
amount of HOS for which each of these defendants is responsible was
sufficient to induce construction of the cap, installation of the pump-andtreat system and, with some further factual assumptions, the duration of
operation of the pump-and-treat system.223 A surface cap, as in Reichhold,
may be the purest example: if two companies arrange for disposal of the
same hazardous substance in a given area, and the amount dumped by each
causes the soil concentration to exceed the response action threshold, then
each has caused the need for construction of a cap over the area they jointly
contaminated.
How might this work in practice? CERCLA plaintiffs would continue
to prove the usual elements of CERCLA liability, without needing to allege
traditional causation and certainly without needing to allege that any
individual defendant’s behavior would have been sufficient to cause the
response action. A defendant that alleged divisibility would bear the burden
of establishing a reasonable basis for apportionment. Even under
Burlington Northern, establishing the basis for apportionment would
necessarily entail some type of proof of the defendant’s causal contribution.
In attempting to defeat the claim for apportionment, the plaintiff should be
entitled to produce evidence that the defendant’s causal contribution, even
221. But, of course, multiple sufficient causes will not always exist, so this theory, too, does not
produce joint and several liability in every case of a single harm under CERCLA.
222. This hypothetical is not unrealistic. The author was lead counsel for the United States in a
case involving a landfill with innumerable customers that may have arranged for disposal of hazardous
substances, but the evidence suggested that the overwhelming majority of the organic solvents that
“drove” the remedy decision came from just three or four companies. The issue was never litigated, but
it would have been completely plausible that the volume dumped by any one of these major contributors
would have required a remedy all by itself.
223. If the duration of operation of the pump-and-treat system to manage migration were a
function only of the presence under the dump of groundwater contaminated above the threshold and the
rate of flow of that groundwater to the landfill boundary, then each sufficient contributor to the abovethreshold contamination would have caused the need for the thirty years of operation.
2009] Several Approaches to Divisibility After Burlington Northern
353
if proven, would alone have been a sufficient cause of the harm. The
burden of proof on apportionment would remain with the defendant.224
Thus, in response to plaintiff’s production, the defendant would have to
establish by a preponderance of the evidence that its proven causal
contribution was insufficient to cause the harm.
Even though the burden of proof on apportionment would remain with
the defendant, this would still be a harder case than CERCLA plaintiffs
have been accustomed to presenting. In order to obtain and preserve
evidence that a particular responsible party’s basis for liability constituted a
sufficient cause, EPA or other responding agencies would likely need to
adjust the way in which response action decisions are documented.
Agencies would need to intensify investigation and discovery into
particular responsible parties’ roles (although that is a probable
consequence of Burlington Northern regardless of how the courts treat
multiple sufficient causes). In some cases, such as a cost recovery claim for
a site involving no dominant arrangers among many, the facts may not
implicate any party—or any solvent party—as a sufficient cause.225 In other
cases, the harm may be inherently proportional to causal contribution. Such
harm is analogous to a purely dose-dependent toxicological response,
instead of a threshold response.226 Thus, sometimes it might not be
possible, even theoretically, for agencies to assign causal sufficiency to any
party among several in a given category of responsible persons.227
224. RESTATEMENT (SECOND) OF TORTS § 433B (1966).
225. Often, owner/operator defendants could be found to have caused the entire harm, but they
are apt to be defunct, like Brown & Bryant. But mere owner/operator status would not always imply
joint and several liability on a multiple-sufficient-cause theory. In a case like Bell Petroleum, depending
on the facts, each of several owner/operator defendants might be found to have contributed too little to
have been a sufficient cause of the entire harm. In re Bell Petroleum Servs., 3 F.3d 889, 889 (5th Cir.
1993).
226. A drum removal, with no contamination of environmental media, may be the purest
example. If, for example, two arrangers brought drums of the same hazardous substance to a storage
facility, which the facility owner/operator abandoned after going bankrupt, it is hard to see how either
arranger could be found to have caused the entire removal action. But cf. United States v. Capital Tax
Corp., 545 F.3d 525, 536 (7th Cir. 2008) (stating that joint and several liability was warranted in drum
removal case because not all response actions were proportionate to number of drums and because “a
CERCLA owner may not move barrels of hazardous substances across property lines . . . in order to
reduce its liability . . . .”).
227. Defendants responding to an assertion of causal sufficiency would likely search for ways in
which the response action was “dose-dependent.” For example, in the MSL hypothetical, a defendant
might argue that duration or cost of operating the pump-and-treat system was a function of the total
volume of HOS in the soil and groundwater. In a soil cap scenario like Reichhold, a defendant might
argue that the combination of both companies’ dumped material increased the total contaminated area
and therefore proportionately increased the size and cost of the needed cap. Courts should scrutinize
such claims closely, even if they conclude that Burlington Northern established a relatively relaxed
proof standard for reasonable apportionment in general. If found to be a sufficient cause of the response
354
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
Despite these issues, proper judicial treatment of multiple sufficient
causes in the context of CERCLA liability should result in joint and several
liability, even in some cases where contributions could be estimated under a
broad view of Burlington Northern. Reichhold provides a perfect
illustration. For the capped area at issue in this case, the evidence had
given Judge Debevoise an idea of the relative contributions of USMRC and
the subsequent occupier’s metals contamination.228 But once the court
found as a fact that either one’s contribution would have required the cap
even in the absence of the other’s,229 the Second Restatement and CERCLA
dictated that the party responsible for either of these sufficient causes
should have been held jointly and severally liable for the resulting harm.
By apportioning230 half the harm to a third party that was not before the
court in any capacity, Judge Debevoise saved the plainly liable USMRC
more than $348,000231 and stuck the plaintiff with that expense.232 Future
action overall, a defendant should be jointly and severally liable even if it could argue for apportionment
based on loose approximations like those in Burlington Northern. An attempt to avoid that result by
nuanced dissection of the response action should be supported by correspondingly nuanced proof. So
the MSL defendant should be required to prove not just its contribution to the HOS problem, but that the
pump-and-treat operating costs are in fact proportional to that contribution; the soil cap defendant
should be required to prove not just its percentage share of the dumped contaminated fill, but how much
smaller the cap would have been in the absence of the other company’s dumping.
228. In discussing apportionment of the costs of the cap, the opinion stated that “USMRC was
undoubtedly the source of most of” the contamination of the relevant area. Reichhold, Inc. v. United
States Metals Refining Co., No. 03-453(DRD), 2009 U.S. Dist. LEXIS 52471, at *131–33 (D.N.J.
2009). Earlier in the opinion, the findings of fact gave an even starker qualitative comparison: “There
is evidence that a subsequent owner or tenant . . . introduced two or three feet of contaminated fill . . .
which was spread on top of an enormous amount of slag fill placed there by USMRC . . . .” Id. at *93.
Still more specifically, the evidence disclosed the area of the cap—125,000 square feet—and the volume
of slag fill dumped by USMRC—65,000 cubic yards. Id. at *10, *93–96. Had the judge deemed it
pertinent, he easily could have computed the greatest possible volume of later fill consistent with the
evidence (three feet deep times 125,000 square feet, a little less than 14,000 cubic yards), and concluded
that USMRC contributed about eighty-two percent of the total volume of contaminated fill in that area.
The calculation implies an assumption that the contamination levels were the same in the fill from each
source, but such an assumption hardly seems less of a stretch than the implicit assumptions the Supreme
Court found acceptable in Burlington Northern.
229. Reichhold, Inc. v. United States Metals Refining Co., No. 03-453(DRD), 2009 U.S. Dist.
LEXIS 52471, at *116 (D.N.J. 2009).
230. The simple apportionment approach of Reichhold seems at odds with the Third Circuit’s
view that proof of divisibility “will be factually complex as it will require an assessment of the relative
toxicity, migratory potential and synergistic capacity of the hazardous waste at issue.” United States v.
Alcan Aluminum Corp., 964 F.2d 252, 269 (3d Cir. 1992). Ironically, Judge Debevoise, sitting by
designation, was on the Alcan panel.
231. Reichhold, Inc., 2009 U.S. Dist. LEXIS 52471, at *132 (plaintiff awarded $348,212 in past
costs, plus one-half of future costs, for the cap).
232. One might be tempted not to weep much for Reichhold, which as a liable party itself might
have been required to contribute to the cleanup costs in any event. But nothing in the opinion’s
reasoning would preclude its application to a non-liable government plaintiff. It is disturbing, moreover,
that Reichhold, which agreed to construct the cap and then sought cost recovery, might have fared better
2009] Several Approaches to Divisibility After Burlington Northern
355
courts facing similar circumstances should apply Burlington Northern more
cautiously.233
C. Of the Weight of Lines and the Length of Masses:
Incommensurable Bases for Liability
The first Restatement of Torts placed the scope-of-liability rules in a
chapter that addressed a variety of problems involving multiple
tortfeasors.234
The Second Restatement recognized the theoretical
gravamen of these rules and nestled their successors comfortably amid
other causation principles.235 However, CERCLA imposes liability for
response actions without regard to whether the conduct of an owner,
operator, arranger, or transporter was a but-for cause of the response action
or even a substantial factor in bringing it about.236 Courts have recognized
had the state performed the response action and sued Reichhold for cost recovery, after which Reichhold
sued USMRC for equitable contribution. The equities tipped strongly against USMRC. USMRC placed
the lion’s share of the contamination into the capped area; by contrast, the court found “no evidence that
Reichhold has brought contaminating metals onto the Site” at all, even if its activities might have moved
some previously-existing contamination. Id. at *89.
233. It seemed as if the Third Circuit Court of Appeals would have an opportunity, as USMRC
appealed the judgment against it, Reichhold, Inc. v. United States Metals Refining Co., No. 03-453
(DRD), Notice of Appeal (D.N.J. filed July 10, 2009), and Reichhold cross-appealed. Reichhold, Inc. v.
United States Metals Refining Co., No. 03-453 (DRD), Notice of Cross-Appeal (D.N.J. filed July 17,
2009). Although Reichhold’s notice of cross-appeal did not refer to the apportionment ruling, USMRC
identified the apportionment as an issue on appeal in a preliminary filing with the Court of Appeals.
Reichhold, Inc. v. United States Metals Refining Co., No. 09-3027, Concise Summary of the Case (3d
Cir. filed July 30, 2009). But the case settled and the parties dismissed their appeals. Reichhold, Inc. v.
United States Metals Refining Co., No. 09-3027, Order (3d Cir. filed Oct. 2, 2009).
234. RESTATEMENT (FIRST) OF TORTS §§ 875, 879, 881 (1939).
235. RESTATEMENT (SECOND) OF TORTS § 433A Rptrs’ Note (1966) (stating that this section
“expands rules formerly stated in §§ 875, 879, and 881, which Sections now refer to this one”). But cf.
Gerald W. Boston, Apportionment of Harm in Tort Law: A Proposed Restatement, 21 U. DAYTON L.
REV. 269, 327 (1996) (arguing that causation principles alone cannot explain section 433A).
236. E.g., United States v. Capital Tax Corp., 545 F.3d 525, 530 (7th Cir. 2008) (CERCLA
imposes liability “when a party is found to have a statutorily defined ‘connection’ with the facility; that
connection makes the party responsible regardless of causation.”); United States v. Alcan Aluminum
Corp., 990 F.2d 711, 721 (2d Cir. 1993) (“What is not required is that the government show that a
specific defendant’s waste caused incurrence of clean-up costs.”); United States v. Alcan Aluminum
Corp., 964 F.2d 252, 266 (3d Cir. 1992) (“Decisions rejecting a causation requirement between the
defendant's waste and the release or the incurrence of response costs are well-reasoned, consistent with
the plain language of the statute and consistent with the legislative history of CERCLA”; plaintiff need
not prove that the arranger’s hazardous substances caused the release or the incurrence of response
costs.); Memphis Zane May Assocs. v. IBC Mfg. Co., 952 F. Supp. 541, 546 (W.D. Tenn. 1996) (“[T]ort
notions of causation do not apply in CERCLA, which utilizes a ‘status-based’ liability standard.”)
(quoting Peter M. Manus, Natural Resource Damages from Rachel Carson’s Perspective: A Rite of
Spring in American Environmentalism, 37 WM. & MARY L. REV. 381, 417 (1996)); cf. RESTATEMENT
(SECOND) OF TORTS § 432(1) (1966) (describing “but for” causation); see id. § 431 (causation requires
tortfeasor’s conduct to be “substantial factor in bringing about the harm”).
356
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
the square peg/round hole quality of apportioning liability based on
causation in the context of a statutory scheme in which liability does not
depend on causation.237
The most obvious incongruity results from Congress’s decision to make
current owners of facilities liable, even if they play no role in the operations
that result in a release and even if no disposal of hazardous substances
occurs during their ownership.238 In the MSL hypothetical, imagine an
additional defendant, O3, who recently bought the property from O2. O3
takes due care with respect to the HOS at the facility, takes precautions
against foreseeable acts of third parties that might make the situation worse,
cooperates with and does not impede the responding government, and
complies with required land use restrictions.239 O3 is nevertheless liable,
because O3 purchased the facility with full knowledge of the HOS
contamination.240 Yet during O3’s ownership the environmental problem at
the facility only improves. By the tempting volumetric measure of causal
apportionment, O3 should have a zero share. But a zero share for O3 would
contradict the plain language of CERCLA that makes current owners liable
based solely on their ownership status.241
237. Capital Tax, 545 F.3d at 535 (“Some courts have noted that the ‘fit’ between § 433A and
CERCLA is actually quite unclear . . . .”); see Alcan, 964 F.2d at 269 (arguing that allowing defendant to
avoid liability based on lack of causation is “consistent” with holding that plaintiff may establish
liability without proving causation); Alcan, 990 F.2d at 722 (divisibility analysis brings causation back
into case “through the backdoor, after being denied entry at the front door”).
238. See 42 U.S.C. § 9607(a)(1) (2006) (making “owner or operator of a vessel or a facility”
liable); see also Kerr-McGee Chem. Corp. v. Lefton Iron & Metal Co., 14 F.3d 321, 325 (7th Cir. 1994)
(current owner liable, although it did not cause disposal, because it could not satisfy “innocent
landowner” standard); United States v. Wedzeb Enters, Inc., 809 F. Supp. 646, 652 n.8 (S.D. Ind. 1992)
(Although section 107(a)(1) refers to “owner and operator,” liability also attaches to entity that either
owns or operates facility but does not do both.); cf. 42 U.S.C. § 9601(35) (2006) (providing so-called
“innocent landowner exception,” which narrows term “contractual relationship” in CERCLA’s thirdparty defense to allow successful assertion of defense by current owners who acquired facility after all
placement of hazardous substances there and who meet additional conditions); see id. § 9622(g)(1)(B)
(authorizing expedited settlements with “de minimis landowners” who did not conduct operations
involving or contribute to the release of hazardous substances, implying that absent settlement such
landowners would be liable).
239. See §§ 9601(35), 9607(b)(3) (listing requirements for innocent landowner exception).
240. § 9601(35)(A)(i).
241. This facial inconsistency with the statute distinguishes the case of a “non-contributing”
current owner from the “special exception” of a “backdoor” causation defense that the Alcan cases
recognized, which allows an arranger, at least in theory, to establish that the harm is reasonably capable
of apportionment and that its correct apportioned share should be zero. United States v. Alcan
Aluminum Corp., 990 F.2d 711, 722 (2d Cir. 1993); see United States v. Alcan Aluminum Corp., 964
F.2d 252, 269 (3d Cir. 1992) (similarly distinguishing use of causation in apportionment inquiry from
use of causation in liability inquiry). Of course, a court balancing the equities in a contribution case
might well decide that such an owner, despite being liable to the government, should contribute nothing
to other responsible parties.
2009] Several Approaches to Divisibility After Burlington Northern
357
The difficulty is not limited to current owners, but arises in any case
that involves a single harm to which liable parties have “contributed” in
incommensurable ways. If an arranger’s hazardous substance travels on a
transporter’s truck to an operator’s disposal business on an owner’s
property, by what metric can a court compare the extent to which each of
those liable parties contributed to the eventual environmental
contamination? If the answer is that each should bear an equal share,
because all contributed equally to causing the entire harm, then per capita
apportionment would prevail in every case. That outcome would
contravene the manifest will of Congress as well as the legal standard
approved in Burlington Northern. On that view, each liable party would
benefit because of the addition of more parties (such as hauling by a
transporter instead by the arranger or operator, or the non-identity of the
owner and the operator). If each liable party caused the entire harm
individually and equally with the others, it makes just as much sense for a
court to hold each party jointly and severally liable.
The MSL hypothetical extends the problem to a somewhat more
realistic, if still oversimplified, situation. It might be tempting to array the
volume of HOS that each responsible party “touched” in its particular role
and divide by the sum of all the individual shares to apportion, for example,
1/9th of the liability to A1 and 2/9th of the liability to T1. This calculation
assumes that cause is proportionate to volume and independent of activity:
arranging for disposal is equivalent to transporting which is equivalent to
operating the facility. A court might hold that assumption reasonable, but
only by its ipse dixit—volume can be measured, therefore I will apportion
by it. The assumption is essentially arbitrary. It becomes more arbitrary as
the actual facts become more complex—if, say, causal contribution varies
not only with each defendant’s class but with its behavior within that
class,242 or if some parties’ liability results purely from ownership status.243
Ultimately, a court attempting a quantitative causally based apportionment
242. For example, operators might conduct different activities involving different substances
with different frequencies of releases. Another imponderable, on the facts of the MSL hypothetical, is
whether A1 should be charged with a “double” share as both an arranger and its own transporter. A1
would presumably argue, with some force, that its causal contribution is what it is, and that A1 made its
contribution only once, not twice.
243. See generally United States v. Northernaire Plating Co., 670 F. Supp. 742, 748 (W.D. Mich.
1987) (Holding harm indivisible where defendants were owner, corporate operator, and individual
operator; although presence of hazardous substances was “directly attributable to the activities of”
operators, “plain language” of CERCLA making owners liable shows that “Congress clearly intended
that the landowner be considered to have ‘caused’ part of the harm. As such, the harm is indivisible, and
all of the defendants are jointly and severally liable.”), aff’d, United States v. R.W. Meyer, Inc., 889 F.2d
1497 (6th Cir. 1989).
358
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
among parties that contribute to a CERCLA harm in qualitatively different
ways must engage in something “like judging whether a particular line is
longer than a particular rock is heavy.”244
Not surprisingly, CERCLA case law does not support the idea that a
court can apportion a single harm among different classes of liable parties.
Chem-Dyne, the archetype, held that section 433A would govern “where a
court performing a case by case evaluation of the complex factual scenarios
associated with multiple-generator waste sites will assess the propriety of
applying joint and several liability on an individual basis.”245 The Alcan
cases, O’Neil, and many others similarly confronted courts with arrangers
attempting to apportion their liability against other arrangers.246 Another
typical fact pattern, and one that has provided the greatest success for
defendants seeking apportionment, has involved owner/operators of distinct
parcels or with distinct time periods of involvement.247 Serious claims of
divisibility in CERCLA cases have generally involved attempts to compare
the comparable.248
Burlington Northern might have been an exception. Brown & Bryant
and the Railroads were both liable as owners/operators, but the district court
244. Bendix Autolite Corp. v. Midwesco Enters., Inc. 486 U.S. 888, 897 (1988) (Scalia, J.,
concurring) (concurring in a case neither related to CERCLA nor involving apportionment). It is of
course true that in contribution actions, courts (or the parties to the contribution case, if they settle) must
perform allocations across responsible party classes all the time—although only after defendants have
some common liability to the plaintiffs. The contribution context can be distinguished. First, it invokes
equity, which allows consideration of numerous factors beyond causation and increases the amount of
discretion with which factors can be weighed. Cf. Burlington N. & Santa Fe Ry. Co. v. United States,
129 S. Ct. 1870, 1882 n.9 (2009) (equitable factors not considered in apportionment analysis). Second,
it does not invoke the principal statutory objectives that warrant imposition of joint and several liability
to plaintiffs in the first instance. For the same reasons, the apportionment analysis under CERCLA is
not the same as that undertaken by “any factfinder that has been called on to apportion fault under
comparative negligence statutes.” In re Bell Petroleum Servs., 3 F.3d 889, 903 n.16 (5th Cir. 1993).
245. United States v. Chem-Dyne Corp., 572 F. Supp. 802, 808 (S.D. Ohio 1983) (emphasis
added). “Generators,” as used in Chem-Dyne, refers to “arrangers” liable under § 9607(a)(3).
246. The arrangers framed the issue that way for the courts, in any event, regardless of whether
an owner/operator, such as Mr. Picillo, originally was a defendant in the case. See, e.g., O’Neil v.
Picillo, 883 F.2d 176, 178 (1st Cir. 1989).
247. E.g., Bell Petroleum, 3 F.3d at 904 (remanding for district court to perform apportionment
after successive operators that engaged in similar activity involving same chemical proved that
reasonable basis for apportionment existed); Dent v. Beazer Materials and Servs., Inc., 156 F.3d 523,
530 (4th Cir. 1998) (owner/operator of separate plant on separate parcel, that contributed different
hazardous substances, not jointly and severally liable with owner/operator of other plant); United States
v. Rohm & Haas Co., 2 F.3d 1265, 1280 (3d Cir. 1993) (defendant in case involving successive
owner/operators not entitled to apportionment because it failed to meet burden of proving a reasonable
basis “for determining the extent of its contribution to the harm”).
248. Cf. Boston, supra note 235, (proposing a general tort rule that would make joint and
several liability less available, yet in cases of causal apportionment requiring that “contributing causes
. . . are reasonably capable of measurement or comparison”).
2009] Several Approaches to Divisibility After Burlington Northern
359
also held Shell liable as an arranger and apportioned a six percent share to
Shell.249 Thus, the Supreme Court almost had an opportunity to provide
guidance on how, if at all, to apportion CERCLA liability among different
classes of liable persons. But the Court dodged that bullet when it held that
Shell was not liable as an arranger at all.250 Instead, Burlington Northern
actually demonstrates that a rule precluding apportionment of CERCLA
liability among responsible parties in disparate liability classes would not
necessarily preclude apportionment in every case involving a single harm.
D. Of Pollution As Paradigm and the Difference of Death: A Closer
Look at Section 433A
In Burlington Northern, the Railroads argued to the Supreme Court that
“the law has recognized pollution as the ‘paradigmatic’ apportionable
harm.”251 In CERCLA cases, many courts have seemed to acquiesce in this
view, at least in principle, even if they held that defendants had failed to
meet their burden of establishing a reasonable basis for a court to carry out
the “paradigmatic” apportionment.252
One passage in the Second
Restatement commentary on section 433A seems to have had particular
influence:
d. Divisible harm. There are other kinds of harm which,
while not so clearly marked out as severable into distinct
parts, are still capable of division upon a reasonable and
249. There is a fair argument that, even assuming the possibility of a reasonable basis of
apportionment among two owner-operators and an arranger, the district court performed the
apportionment improperly. The district court’s finding that operations on the B&B Parcel contributed at
least ten times as much contamination as operations on the Railroad Parcel, as well as the way in which
the district court computed the Railroads’ apportioned share of liability, suggests an implicit
apportionment of 91% to Brown & Bryant. United States v. Atchison, Topeka & Santa Fe Ry. Co., No.
CV-F-92-6058, 2003 U.S. Dist. LEXIS 23130, at *146, *260 (E.D. Cal. July 14, 2003). The
independent computation of a 6% share for Shell implies that the district court actually allocated 106%
of the liability to the three responsible parties. This illustrates the problem of incommensurable bases of
liability: there is no logical way to address the overlap between the arranger’s causal contribution and
the owners’/operators’ contribution. On the other hand, if one assumes that the district court really
meant to apportion 9% to the Railroads, 6% to Shell, and 85% to Brown & Bryant, then when the
Supreme Court held that Shell was not liable, Shell’s share should have been reallocated to keep the
total apportioned liability at 100%—the Railroads’ share would have increased to about 9.6%.
250. Burlington N. and Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870, 1880 (2009).
251. Brief for Petitioners the Burlington Northern and Santa Fe Railway Company and Union
Pacific Railroad Company at 26, Burlington N. and Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870
(2009) (No. 07-1601) .
252. As noted above, see supra note 82, in Burlington Northern the Supreme Court simply
noted that both the district court and the court of appeals had treated the harm at the Brown & Bryant
site as a single harm that was theoretically capable of apportionment.
360
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
rational basis, and of fair apportionment among the causes
responsible . . . .
....
Such apportionment is commonly made in cases of private
nuisance, where the pollution of a stream, or flooding, or
smoke or dust or noise, from different sources, has
interfered with the plaintiff’s use or enjoyment of his land.
Thus where two or more factories independently pollute a
stream, the interference with the plaintiff’s use of the water
may be treated as divisible in terms of degree, and may be
apportioned among the owners of the factories, on the basis
of evidence of the respective quantities of pollution
discharged into the stream.253
The Restatement provides an example, Illustration 5, in which oil
negligently discharged from two factories onto the surface of a stream
deprives a downstream riparian owner of the use of the water for industrial
purposes. “There is evidence” that seventy percent of the oil came from
one factory and thirty percent from the other; on that basis, each factory
owner is liable for the corresponding proportion of the plaintiff’s
damages.254
Courts have taken for granted the analogy between CERCLA liability
to the government and private nuisance liability to the Restatement’s
downstream landowner.255 A careful analysis of the treatment of pollution
cases in section 433A reveals that the analogy has been flawed from the
start.
1. Section 433A As a Whole
First, section 433A takes multiple views of pollution cases. The loss of
a stream’s use for industrial purposes by the combined effect of two oil
discharges is divisible if the basis for apportionment is proven, the
Restatement says in Illustration 5, but “[c]ontrast Illustrations 14 and 15”:
253. RESTATEMENT (SECOND) OF TORTS § 433A cmt. d (1966). This text is quoted or closely
paraphrased in, inter alia, United States v. Alcan Aluminum Corp., 964 F.2d 252, 269 (3d Cir. 1992); In
re Bell Petroleum Servs., 3 F.3d 889, 896 (5th Cir. 1993); United States v. Hercules, Inc., 247 F.3d 706,
718 (8th Cir. 2001).
254. RESTATEMENT (SECOND) OF TORTS § 433A cmt. d, illus. 5 (1966).
255. The assumed analogy has often made little difference, because evidence to provide a
reasonable basis for apportionment has usually been lacking.
2009] Several Approaches to Divisibility After Burlington Northern
361
14. A Company and B Company each negligently
discharge oil into a stream. The oil floats on the surface
and is ignited by a spark from an unknown source. The fire
spreads to C’s barn, and burns it down. C may recover a
judgment for the full amount of his damages against A
Company, or B Company, or both of them.
15. The same facts as Illustration 14, except that C’s cattle
drink the water of the stream, are poisoned by the oil and
die. The same result.256
How can courts reconcile Illustrations 5, 14, and 15? The Restatement
does not hinge the difference on the legal characterization of the claim in
Illustration 5 as sounding in “nuisance,” but rather on the nature of the
harm:
Certain kinds of harm, by their very nature, are normally
incapable of any logical, reasonable, or practical division.
Death is that kind of harm, since it is impossible, except
upon a purely arbitrary basis for the purpose of
accomplishing the result, to say that one man has caused
half of it and another the rest. The same is true of . . . the
destruction of a house by fire . . . .257
Thus, according to section 433A, those who contribute to the
“indivisible” burning of a barn or fatal poisoning of cows are liable, jointly
and severally, for all of the damage to which they contributed. The
unmistakable implication is that loss of use or enjoyment of land, by
contrast, is inherently capable of logical, reasonable, or practical division.
The coherence of this distinction can be criticized from both ends.258
However, the Supreme Court has directed courts to apply section 433A to
claims of divisibility in CERCLA cases. To do so, courts must consider a
novel question: is a CERCLA cost recovery claim more like a claim for lost
use of water, or more like a claim for a burned barn or poisoned cattle?
256. Id. § 433A cmt. I, illus 14, 15.
257. Id. § 433A cmt. i.
258. From one end, Gerald Boston argued that determining the causal contributions of, say,
cigarette smoking and asbestos exposure to death from lung cancer ought to be no different from
determining their contributions to illness from lung cancer. Boston, supra note 235, at 331. See
generally RESTATEMENT (THIRD) OF TORTS: APPORTIONMENT OF LIABILITY (2000). From the other end,
the loss of use of the water in Illustration 5 seems just as total as the loss of the burned barn or the dead
cattle; is it really possible to say, “except . . . for the purpose of accomplishing the result” that each
contribution of oil made some of the water unusable or made all of the water partly unusable?
362
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
Certainly, hazardous substance contamination, and/or the required
response action, can make property unmarketable and virtually unusable.
But then, burning down a house makes it pretty unusable too. So the lost
use analogy does not help much. The key, rather, is that many, if not all,
CERCLA cost recovery claims are for response actions that result from the
totality of an environmental situation. In the MSL hypothetical, the
remedial investigation and feasibility study would necessarily have been
conducted in response to the (as yet uncharacterized) contamination as a
whole; the cap and pump-and-treat systems are similarly comprehensive
responses to an overall problem. The facility from which hazardous
substances are released is the burning barn; each individual contribution of
HOS (whether by arrangement for disposal, transportation, or operation of
the landfill) is one of the individual fires that combined to create the
conflagration. The analogy may not be perfect, but it is no more imperfect
than the analogy of a typical CERCLA site to water rendered unusable by a
single substance in assumed proportion to the amount of the substance.
The Restatement illustrations include another apparent paradox that
further highlights this distinction. Illustration 15 treats cattle killed by
poisoned water as an indivisible harm.259 The apparent explanation is that
the cattle are dead, and death is an indivisible harm. But Illustration 3 holds
that if three dogs belonging to one owner and two dogs belonging to
another kill ten of plaintiff’s sheep, the harm is divisible in proportion to the
number of dogs of each owner, provided there “is evidence . . . that all of
the dogs are of the same general size and ferocity.”260 The sheep in the
latter illustration are no less dead than the cattle in the former. The best
explanation for the varying outcomes is that intuition tells us that the dogs
likely set upon the sheep individually or nearly so, and no sheep died by the
act of all of the dogs. By contrast, although it is theoretically possible that a
single steer drank only oil emitted from one factory, intuition tells us that
the cattle were poisoned by both defendants’ oil, just as a CERCLA
response action usually responds to environmental conditions created by all
of the responsible parties.
Neither the nature of the harm (e.g., death) nor the nature of the cause
(e.g., pollution) seems to explain satisfactorily all of these examples.
Looking beneath the text of the Restatement commentary to the court
decisions from which its illustrations are extracted confirms that these
distinctions are inadequate.
259. RESTATEMENT (SECOND) OF TORTS § 433A cmt. d, illus. 15 (1966).
260. Id. illus 3. The importance of the evidentiary qualification to CERCLA cases, where
defendants often cannot prove that their hazardous substances are of equal environmental ferocity, is
immediately obvious.
2009] Several Approaches to Divisibility After Burlington Northern
363
Illustration 14, where an oil slick contributed by two defendants catches
fire and the fire burns a barn, “is taken from Northup v. Eakes,”261 an
Oklahoma Supreme Court opinion.262 In Northup, numerous holders of oil
leases allowed crude oil to flow into a creek that ran past plaintiff’s land;
just as in the illustration, the oil ignited and the resulting fire spread to and
destroyed plaintiff’s barn. The court held that “the lessees holding separate
leases acted independent of each other, yet their several acts in permitting
the oil to flow into the stream combined to produce but a single injury. In
these circumstances each is responsible for the entire result . . . .”263
Illustration 15, where two defendants discharge oil that poisons
plaintiff’s cattle, also derives from an Oklahoma case, Tidal Oil Co. v.
Pease.264 Multiple defendants discharged petroleum well waste containing
oil and salt water onto their own properties, which eventually drained into
two creeks that crossed plaintiff’s land.265 Some defendants discharged into
only one of the creeks, but plaintiff produced evidence that his livestock
drank from both creeks.266 Some of the stock died from drinking the
contaminated water, but—contrary to the Restatement illustration—most of
the animals were merely “injured . . . in loss of flesh, appearance, class and
condition.”267 Again the court held that joint and several liability was
appropriate.268 The same court reached similar results in a case of pollution
causing crop damage269 and even in a case involving multiple causes of
flooding.270
In an industrializing America, it became commonplace for the effluvia
of numerous manufacturing or extractive firms to combine and to injure
their neighbor. The scope of each tortfeasor’s liability for those injuries
261. RESTATEMENT (SECOND) OF TORTS § 433A Rptrs’ Note, cmt. i (1966).
262. Northup v. Eakes, 178 P. 266 (Okla. 1918).
263. Id. at 268.
264. Tidal Oil Co. v. Pease, 5 P.2d 389 (Okla. 1931); see RESTATEMENT (SECOND) OF TORTS
§ 433A Rptrs’ Note, cmt. i (1966).
265. Id. at 390.
266. Id.
267. Id.
268. “While the defendants were acting independent of each other, if their acts combined to
produce the alleged injury to plaintiffs’ live stock . . . each so acting is responsible for the entire result,
even though the act of any one defendant might not have caused it.” Id. at 391; accord Kanola Corp. v.
Palmer, 30 P.2d 189, 190 (Okla. 1934). Interestingly, while Tidal Oil presents a case of indivisible
harm, it also illustrates the principle of distinct harms—the court held that plaintiff could not recover
damages for harm to certain cattle that were injured while plaintiff unlawfully pastured them on a
defendant’s land. Pease, 5 P.2d at 392.
269. Indian Terr. Illuminating Oil Co. v. Bell, 46 P.2d 481, 482 (Okla. 1935).
270. City of Skiatook v. Carroll, 21 P.2d 498, 499 (Okla. 1933). The application of joint and
several liability in City of Skiatook cannot be squared with Restatement (Second) of Torts section 433A
cmt. d, illus. 4.
364
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
was a pressing question for the common law courts. Downwind of the
smelters of Ducktown,271 in the Pennsylvania coal fields,272 and in many
other places,273 the courts protected industry from joint liability. In
Oklahoma, where the battle was between oil production and cattle
production, the ranchers won a principle of joint and several liability.
Louisiana, too, declined to apportion similar harm under its civil law.274
The pollution illustrations that accompany section 433A are incoherent
because the underlying decisions do not reflect the distinctions the
Restatement essays as explanations. The decisions, rather, presented two
different themes in response to the problem of multiple discharges
combining to harm receptors downstream—even in the early years of the
twentieth century. Courts adjudicating CERCLA claims that have been
directed to apply “traditional and evolving” common law standards should
consider which of these two views of divisibility of harm best fits both the
fact patterns and the statutory scheme of CERCLA.
2. Illustration 5 on Its Own
The foregoing discussion demonstrated that section 433A and
associated commentary, read as a whole, can support the view that most, if
not all, CERCLA harms are not divisible; courts need not feel chained to
Illustration 5. But courts have focused on Illustration 5, so it is worth
asking whether they properly have understood it. A look at the common
law source material for the illustration demonstrates that Illustration 5’s
treatment of private nuisance claims, even on its own terms, does not
support the view that CERCLA liability is analogously capable of
apportionment.
The Reporter’s Note to section 433A includes a long string citation of
cases that Illustration 5 is “based on.”275 Few of these decisions actually
involve a court conducting or reviewing an apportionment of liability, but
nearly all276 articulate in some way the proposition that, in the absence of
271. Swain v. Tenn. Copper Co., 78 S.W. 93 (Tenn. 1903).
272. Little Schuylkill Nav. R.R. & Coal Co. v. Richards’ Adm’r, 57 Pa. 142, 143 (1868).
273. See generally RESTATEMENT (SECOND) OF TORTS § 433A(1) Rptrs’ Note (1966) (citing to
cases in many states).
274. Williams v. Pelican Natural Gas Co., 175 So. 28, 34 (La. 1937) (imposing in solido liability
on three defendants where salt water discharged independently from each defendant’s oil well entered
plaintiff’s pond, killing fish and timber and making the water unfit for sale or swimming or watering
stock).
275. RESTATEMENT (SECOND) OF TORTS § 433A(1) Rptrs’ Note (1966).
276. Two decisions the Reporter’s Note cited as sources of Illustration 5 actually imposed joint
and several liability. Id. Phillips Petroleum Co. v. Hardee, 189 F.2d 205, 212 (5th Cir. 1951) (Applying
Louisiana law and stating that if “though acting separately, [defendants’] negligence combined to
2009] Several Approaches to Divisibility After Burlington Northern
365
concerted action, multiple defendants who contribute to pollution that
combines to cause a downstream private nuisance are only severally
liable.277 The question is why they so hold.
It turns out that within those dusty old reporters (or rarely viewed
corners of electronic databases) lurks some pretty hoary doctrine. The
cases are infused with procedural impediments to joinder of parties278 and,
most importantly, with the concept that joint liability could only attach
where the defendants committed a “joint tort.”279 The joint tort analysis
viewed the situation from the point of view of the tortfeasor rather than the
injured plaintiff. It focused on what act the tortfeasor did, not on the harm
its act caused. This issue dominates the cases cited in support of Illustration
5, and to the courts deciding those cases it was fundamental.
In the earliest of those cases, the defendants and other parties operated
coal mines and deposited “coal-dirt” into streams.280 All the coal-dirt filled
a pond behind plaintiff’s dam.281 The trial court instructed the jury that if
the defendants threw coal-dirt into the river with the knowledge that other
miners were doing so at the same time, the defendants were liable for “the
produce the pollution damage, plaintiffs may recover for the whole damage against one or all of those
contributing.”); City of Valparaiso v. Moffitt, 39 N.E. 909, 911 (Ind. App. 1895) (distinguishing public
nuisance from “general rule” that joint liability requires concert of action).
277. E.g., Snavely v. City of Goldendale, 117 P.2d 221, 224 (Wash. 1941) (Allowing, contrary to
prior authority, joinder of multiple independent tortfeasors in such a case, because “the several liability
of such tort-feasors . . . may be more accurately and justly determined in an action in which they are all
defendants.”); Mitchell Realty Co. v. City of West Allis, 199 N.W. 390, 396 (Wis. 1924) (remanding for
apportionment in case of nuisance caused by several sewage dischargers); Thomas v. Ohio Coal Co, 199
Ill. App. 50, 56 (1916) (holding that plaintiff adduced sufficient evidence to support jury verdict
estimating contribution of individual defendant to oil and brine pollution discharged from various
parties’ wells).
278. E.g., Watson v. Pyramid Oil Co., 248 S.W. 227, 228 (Ky. App. 1923) (plaintiff declined to
elect against which defendant he would proceed); Standard Phosphate Co. v. Lunn, 63 So. 429, 432 (Fla.
1913) (“For separate and distinct wrongs in no wise connected by the ligament of a common purpose . . .
wrongdoers are liable only in separate actions and not jointly in the same action.”).
279.
“Joint tort” has historically been used to describe the situation in which there was
a connection among defendants that justified holding each of them liable for the
acts of others, such as concerted action or the vicarious liability of an employer
for an employee’s negligence. The term “joint tort” has also been used to
describe the procedural issue of whether multiple defendants may be joined in the
same action, especially under the restrictive joinder requirements imposed by
common-law pleading.
RESTATEMENT (THIRD) OF TORTS: APPORTIONMENT OF LIABILITY § 10 Rptrs’ Note, cmt. b (2000)
(citations omitted).
280. Little Schuylkill Nav. R. & Coal Co. v. Richards’s Adm’r, 57 Pa. 142, 144 (1868).
281. Id.
366
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
combined results of all the . . . deposits . . . .”282 The Pennsylvania Supreme
Court reversed:
[T]he fallacy lies in the assumption that the deposit of the
dirt by the stream in the basin is the foundation of liability.
It is the immediate cause of the injury, but the ground of
action is the negligent act above. The right of action arises
upon the act of throwing the dirt into the stream—this is
the tort, while the deposit below is only a consequence.
The liability, therefore, began above with the defendant’s
act upon his own land, and this act was wholly separate,
and independent of all concert with others. His tort was
several when it was committed, and it is difficult to see
how it afterwards became joint, because its consequences
united with other consequences.
The union of
consequences did not increase his injury.283
This logic, and even language, reverberated for decades.284 The
physical separation between the wrongdoers’ individual tortious acts and
those acts’ combined harmful consequences provides a persistent
explanation for why private nuisance developed as the “one area”285 in
which the common law diverged from “the doctrine applicable to the
ordinary negligence cases.”286
282. Id. at 146.
283. Id.
284. E.g., Johnson v. City of Fairmont, 247 N.W. 572, 573 (Minn. 1933) (“[A]cts of
independent tort-feasors, each of which cause some damage, may not be combined to create a joint
liability at law for damages.”); Standard Phosphate Co. v. Lunn, 63 So. 429, 432 (Fla. 1913) (“Torts that
are several, separate, and independent acts when committed do not become joint by the subsequent
union or intermingling of their consequences . . . .”); Chipman v. Palmer, 77 N.Y. (32 Sickels) 51, 53
(1879) (“[R]ight of action arises from the discharge into the stream, and the nuisance is only a
consequence of the act . . . [which], being several when it was committed, cannot be made joint because
of the consequences which followed in connection with others . . . .”).
285. Brief for Petitioners the Burlington Northern and Santa Fe Railway Company and Union
Pacific Railroad Company at 27, Burlington N. and Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870
(2009) (No. 07-1601).
286. Mitchell Realty Co. v. City of West Allis, 199 N.W. 390, 394 (Wis. 1924); see Edmonds v.
Compagnie Generale Transatlantique, 443 U.S. 256, 260 (1979) (“[C]ommon law . . . allows an injured
party to sue a tortfeasor for the full amount of damages for an indivisible injury that the tortfeasor’s
negligence was a substantial factor in causing, even if the concurrent negligence of others contributed to
the incident.”); see generally Phillips Petroleum Co. v. Hardee, 189 F.2d 205, 212 (5th Cir. 1951)
(applying joint and several liability to claim for damage to crops and land resulting from pollution
despite recognizing nuisance as “exception” to rule that each of two persons whose independent tortious
acts comprise substantial factor in causing harm is liable for entire harm) (citing RESTATEMENT (FIRST)
OF TORTS § 879 (1939)). This does not mean to imply that causal considerations played no role in the
decisions approving apportionment in pollution cases. Some courts openly worried about the possibility
2009] Several Approaches to Divisibility After Burlington Northern
367
CERCLA liability is different. It derives only from the status of the
responsible party in relation to the facility that released hazardous
substances. Arranger liability does not arise when the hazardous substance
leaves the arranger’s property; transporter liability does not arise when the
transporter puts the hazardous substance on its truck. The liability exists
regardless of whether the liable party’s hazardous substances exceeded
some threshold quantity that would have occasioned the response action.287
By tying the liability-creating conduct to the facility at which the release
and response occur, CERCLA unites conceptually the liable parties’
separate “torts” and unites geographically the “tort” and the “consequence”
in a way that the private nuisance claims cited in the Restatement do not.288
Finally, the private nuisance cases are inapposite because they are, well,
private. City of Valparaiso v. Moffitt,289 a nineteenth-century case cited in
the Restatement in support of Illustration 5, held that “all persons who
created or continued” a public nuisance “are jointly and severally liable for
all the damages resulting therefrom, although they are not joint
tortfeasors.”290 If a government CERCLA claim is like a nuisance claim at
all, it is much more like a claim for public nuisance than private nuisance.
It is based not on lost use and enjoyment of the facility that is the subject of
the cleanup, but on the need to protect human health, welfare, and the
of disproportionate liability, particularly for actors whose own conduct alone would have caused no
harm, such as a party that discharges sewage in amounts too small to constitute a nuisance, but which
creates a nuisance when combined with others’ similarly small contributions. E.g., Martinowsky v. City
of Hannibal, 35 Mo. App. 70, 78 (1889) (holding that liability must be apportioned to avoid having a
person who dumps “some putrid matter” become liable for “a pestilence that depopulates a city” only
because others did the same thing). But similar, if less extreme, concerns can also arise in personal
injury cases, and causation doctrines short of a general rule of several liability can address them.
287. See United States v. Alcan Aluminum Corp., 964 F.2d 252, 260 (3d Cir. 1992) (“CERCLA
liability does not depend on the existence of a threshold quantity of a hazardous substance.”) (citation
omitted).
288. CERCLA is different from the common law antecedents of Illustration 5 in at least one
other significant way. A number of the common law opinions, especially ones that worried about
disproportionate liability, also noted the common law rule precluding joint tortfeasors from obtaining
contribution from fellow tortfeasors. E.g., Chipman, 77 N.Y. at 53–54 (noting that a defendant held
jointly and severally liable would be “without any means to enforce contribution or to adjust the amount
among the different parties”); Watson v. Pyramid Oil Co., 248 S.W. 227, 228 (Ky. App. 1923)
(identifying that such a defendant would have no remedy against the fellow tortfeasors).
289. City of Valparaiso v. Moffitt, 39 N.E. 909, 909 (Ind. App. 1895).
290. Id. at 911. In that case, the defendant city, which had built sewers that polluted a stream,
alleged an affirmative defense that an operator of a private gas works, with whom plaintiff had settled
separately, had caused part of the pollution; the court affirmed a plaintiff’s verdict. Id. at 910. The
Restatement cites the opinion as support for Illustration 5, presumably because of dicta stating a general
rule that “[i]f several distinct acts of several persons have contributed to a single injury, but without
concert of action or common intent, there is generally no joint liability.” Id. at 911.
368
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
environment, a governmental function that the statute specifically
authorized.291
Judge Rubin apparently understood this: the Chem-Dyne opinion cites
City of Valparaiso (and not any of the private nuisance cases).292 The idea
that the common law’s treatment of public nuisance is a relevant and
appropriate source of guidance for courts defining the contours of joint and
several liability under CERCLA has simply vanished from the
jurisprudence.293 It is time to bring it back from the phantom zone.
Burlington Northern, after all, instructs courts to consult the common law
for guidance and endorses the Chem-Dyne analysis.
In sum, neither Burlington Northern nor section 433A binds federal
courts to assume willy-nilly that CERCLA harms, simply because they
consist of pollution, are theoretically capable of apportionment. Courts
should engage that question anew. If they do, they will find in Illustrations
14 and 15, as well as in the differences between CERCLA and private
nuisance, strong grounds to reason that single CERCLA harms most closely
resemble harms that are “indivisible” under section 433A.
Would this mean that all parties liable under CERCLA will be jointly
and severally liable all the time? No. First of all, as described above,294
even if courts were to determine that there is never “a reasonable basis for
determining the contribution of each cause to a single [CERCLA] harm,”295
apportionment would still be available in CERCLA cases that present
“distinct harms.”296 Congress wanted the courts to impose joint and several
liability in “appropriate” cases; courts would not violate that intent by
deciding that it is always “appropriate” in one of two Restatement
categories. Second, the re-analysis of section 433A suggested here would
not necessarily foreclose holding a single harm divisible in the case where
291. An analogy to public nuisance may be pertinent to a court’s consideration of common law
sources to determine the scope of CERCLA liability, but this does not mean courts are free to import
into CERCLA’s comprehensive statutory liability scheme the substantive common law requirements of
public nuisance, e.g., RESTATEMENT (SECOND) OF TORTS §§ 821B, 821C (1966).
292. United States v. Chem-Dyne Corp., 572 F. Supp. 802, 810 (S.D. Ohio 1983).
293. Only one other CERCLA decision that addresses the appropriateness of joint and several
liability has cited City of Valparaiso: United States v. Ottati & Goss, Inc., 630 F. Supp. 1361, 1395
(D.N.H. 1985) (quoting Chem-Dyne Corp., 527 F. Supp. at 810). There the court refused to apportion
liability even though there was evidence of the number of drums that various arrangers of hazardous
substances had dumped at the site, because the subsequent mixing of the drums’ contents had created an
indivisible harm that could not be apportioned “with any degree of accuracy.” Id. at 1396.
294. See supra note 42 and accompanying text.
295. RESTATEMENT (SECOND) OF TORTS § 433A(1)(b) (1966).
296. Id. § 433A(1)(a).
2009] Several Approaches to Divisibility After Burlington Northern
369
the degree of harm displayed true proportionality or dose-dependence. The
Bell Petroleum holding, for example, might well survive.297
CONCLUSION
In Burlington Northern, the Railroads joined a select group of
CERCLA defendants that have succeeded in being liable to the government
for cost recovery under CERCLA without being jointly and severally liable.
And they got the highest court in the land to admit them to the group. This
was no mean feat.
How much their accomplishment will redound to the benefit of liable
parties in other CERCLA cases remains to be seen. The Burlington
Northern opinion might become a landscape-changing earthquake, or it
might—despite its august source—be a minor tremor with relatively few
aftershocks, like Bell Petroleum turned out to be. The outcome could
dramatically reduce CERCLA’s ability to continue meeting its goals. This
article has argued that courts can, and should, avoid that result without
inconstancy to the Burlington Northern opinion itself.
Of course, there is a third, and unsettling, possibility. The Supreme
Court might not be done remaking the liability scheme it let alone for
twenty-plus years. Time will tell.
297. See supra notes 46–47 and accompanying text.
RESTATEMENT FOR ARRANGER LIABILITY UNDER
CERCLA: IMPLICATIONS OF BURLINGTON NORTHERN
FOR SUPERFUND JURISPRUDENCE
Alfred R. Light∗
TABLE OF CONTENTS
Introduction ............................................................................................... 372
I. Restatement of the Law: CERCLA Arranger or “Generator” Liability 373
II. Comments and Illustrations ................................................................. 373
A. Scope ............................................................................................... 373
B. History ............................................................................................. 374
C. “Person” ........................................................................................... 375
D. “By Contract, Agreement, or Otherwise Arranged For” ................. 375
E. “Disposal or Treatment” .................................................................. 376
F. “Of Hazardous Substances Owned or Possessed by Such Person”.. 378
G. “By Any Other Party or Entity” ...................................................... 379
H. “At Any Facility or Incineration Vessel Owned or Operated by
Another Party or Entity and Containing Such Hazardous
Substances”.................................................................................... 379
I. “From Which There Is a Release, or a Threatened Release Which
Causes the Incurrence of Response Costs, of a Hazardous
Substance” ..................................................................................... 380
J. Defenses ............................................................................................ 384
III. Reporter’s Notes ................................................................................. 385
A. Scope ............................................................................................... 385
B. History ............................................................................................. 386
C. “Person” ........................................................................................... 386
D. “By Contract, Agreement, or Otherwise Arranged For” ................. 387
E. “Disposal or Treatment” .................................................................. 388
F. “Of Hazardous Substances Owned or Possessed by Such Person”.. 389
G. “By Any Other Party or Entity” ...................................................... 390
∗ Professor of Law, St. Thomas University School of Law, Miami Gardens, Florida. Ph.D.
University of North Carolina at Chapel Hill, 1976; J.D. Harvard, 1981; B.A. The Johns Hopkins
University, 1971.
372
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
H. “At Any Facility or Incineration Vessel Owned or Operated by
Another Party or Entity and Containing Such Hazardous
Substances”.................................................................................... 390
I. “From Which There Is a Release, or a Threatened Release Which
Causes the Incurrence of Responses Costs, of a Hazardous
Substance” ..................................................................................... 391
INTRODUCTION
This past May, the United States Supreme Court for the first time
addressed two issues that Congress left open in the Comprehensive
Environmental Response, Compensation, and Liability Act of 1980
(“CERCLA” or “Superfund”).1 These issues are: (1) the scope of arranger
or “generator” liability under the language of CERCLA section 107(a)(3)2;
and (2) the circumstances in which a party under section 1073 may be held
jointly and severally liable. The Court rejected the position of the United
States Government on both issues. The Court found that the Government
had attempted to extend CERCLA generator liability “beyond the limits of
the statute itself.”4 It also found that despite the Government’s “refusal to
acknowledge the potential divisibility of the harm,”5 the District Court’s
rough formula limiting the generator liability of the defendants to nine
percent of the Government’s total response costs “was supported by the
evidence and comports with the apportionment principles”6 to which
members of Congress had made reference in 1980.7
Because the Court first addressed these two issues more than twentyeight years after CERCLA’s enactment and because of the fact that the
Court rejected the Government’s litigation position, which it had asserted
regularly in the lower courts, Burlington Northern has precedential
implications for hundreds of lower court opinions relating to arranger
liability and to the application of joint and several liability. Below, I survey
decisions prior to Burlington Northern on the first of these issues, arranger
liability, and assess their precedential effect in light of that decision. This
1. Comprehensive Environmental Response, Compensation, and Liability Act, 42 U.S.C.
§§ 9601–9675 (2006).
2. Id. § 9607(a)(3).
3. § 9607.
4. Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870, 1879 (2009).
5. Id. at 1881.
6. Id. at 1883.
7. See id. at 1881 (stating Congress’s desire that CERCLA follow “traditional and evolving
principle of common law”).
2009]
Restatement for Arranger Liability Under CERCLA
373
article will adopt the artifice of a hypothetical “Restatement” for this area of
Superfund jurisprudence. It will follow a structure consisting of three
subdivisions: (1) the black-letter law, reflecting an assessment of the current
applicable legal rules in summary form after Burlington Northern; (2)
Comments and Illustrations reflecting an elaboration and application of
these rules with reference to prior cases; and (3) Reporter’s Notes,
reflecting commentary about these rules, applications, and likely areas of
ambiguity or dispute that courts may need to visit or revisit in light of
Burlington Northern.
I. RESTATEMENT OF THE LAW: CERCLA ARRANGER OR
“GENERATOR” LIABILITY
Notwithstanding any other provision or rule of law, and subject only to
the defenses expressly set forth in CERCLA, any person who by contract,
agreement, or otherwise arranged for disposal or treatment, or arranged
with a transporter for transport for disposal or treatment, of hazardous
substances owned or possessed by such person, by any other party or entity,
at any facility or incineration vessel owned or operated by another party or
entity and containing such hazardous substances . . . from which there is a
release, or a threatened release which causes the incurrence of responses
costs, of a hazardous substance . . . shall be liable for: all costs of removal
or remedial action incurred by the United States Government or a State or
an Indian tribe not inconsistent with the national contingency plan; any
other necessary costs of response incurred by any other person consistent
with the national contingency plan; damages for injury to, destruction of, or
loss of natural resources, including the reasonable costs of assessing such
injury, destruction, or loss resulting from such release; and the costs of any
health assessment or health effects study carried out under subsection i [of
CERCLA section 104].8
II. COMMENTS AND ILLUSTRATIONS
A. Scope
This Restatement addresses issues of arranger or “generator” liability
arising under CERCLA section 107(a)(3), 42 U.S.C. § 9607(a)(3), as they
8. § 9607(a).
374
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
stand after the Supreme Court’s decision in Burlington Northern.9 It
explores how the decision likely impacts prior holdings of the lower courts
on many arranger liability issues not directly addressed in the Supreme
Court’s holding. The Restatement is limited to a discussion of arranger
liability and does not generally address principles regarding other classes of
potentially responsible parties (PRPs), except insofar as they relate to
arranger liability.10
Arranger liability under section 107(a)(3) significantly expanded
hazardous waste cleanup liability from the federal law that existed prior to
its enactment in 1980, especially considering its intended retroactive
application.11
For the sake of simplicity, except where the context indicates otherwise,
this Restatement uses the word “damages” to encompass response costs,
natural resource damages, and the costs of health effects studies recoverable
under section 107(a).12 Similarly, the word “harm” is interchangeable with
“release, and threatened release” within the CERCLA context.13
B. History
Generally, CERCLA is “not a model of legislative draftsmanship”14: the
courts have described it as “hastily-drawn,”15 “fragmented,”16 with a “welldeserved notoriety for vaguely-drafted revisions and an indefinite, if not
contradictory, legislative history.”17
More specifically, Congress
deliberately left many liability issues open, indicating, “[i]t is intended that
issues of liability not resolved by this act, if any, shall be governed by
9. Burlington Northern, 129 S. Ct. at 1870.
10. See infra Parts II.G–H. Note that the Burlington Northern decision of the Supreme Court
did not directly address the apportionment of CERCLA liability to generators in its holding since it
found Shell Oil Company, the “generator” appellant, not liable at all under section 107(a)(3).
Burlington Northern, 129 S. Ct. at 1880 (“Having concluded that Shell is not liable as an arranger, we
need not decide whether the Court of Appeals erred in reversing the District Court’s apportionment of
Shell’s liability for the cost of remediation.”).
11. See United States v. Olin Corp., 107 F.3d 1506 (11th Cir. 1997) (upholding
constitutionality of CERCLA’s retroactive effect). But see Alfred R. Light, “Taking” CERCLA
Seriously: The Constitution Really Does Limit Retroactive Liability, 13 Toxics L. Rep. (BNA) 238, 238–
44 (July 15, 1998).
12. See 42 U.S.C. § 9601(6), (16), (22), (23), (24) (2006) (defining “damages”, “natural
resources”, “remove or removal”, and “remedy or remedial action”).
13. See § 9601(22) (definition of “release”).
14. Exxon Corp. v. Hunt, 475 U.S. 355, 363 (1986).
15. Violet v. Picillo, 648 F. Supp. 1283, 1288 (D.R.I. 1986), aff’d, O’Neil v. Picillo 883 F.2d
176 (1st Cir. 1989).
16. United States v. Conservation Chem. Co., 619 F. Supp. 162, 204 (W.D. Mo. 1985).
17. Amoco Oil Co. v. Borden, Inc., 889 F.2d 664, 667 (5th Cir. 1989) (quoting United States v.
Mottolo, 605 F. Supp. 898, 902 (D.N.H. 1985)).
2009]
Restatement for Arranger Liability Under CERCLA
375
traditional and evolving principles of common law.”18 The language of
section 107(a)(3) is a modification of language in Senate Bill S.1480 as
reported from the Senate Environment and Public Works Committee in
1980.19
C. “Person”
The “person” potentially liable as an arranger may be “an individual,
firm, corporation, association, partnership, consortium, joint venture,
commercial entity, United States Government, State, municipality,
commission, political subdivision of a State, or any interstate body.”20
Parent corporations are not derivatively liable absent a traditional basis for
piercing the corporate veil.21 Individuals are not derivatively liable absent
personal involvement in disposal practices.22
D. “By Contract, Agreement, or Otherwise Arranged For”
To qualify as an arranger, a person must have entered into a transaction
“with the intention that at least a portion of the product [which is the subject
of the transaction] be disposed of” or treated.23 The word “arrange” implies
action directed to a specific purpose.24 “[K]nowledge alone is insufficient
to prove that an entity ‘planned for’ the disposal, particularly when the
18. 126 CONG. REC. 30,932 (1980) (statement of Sen. Randolph), reprinted in 1 A LEGISLATIVE
HISTORY OF THE COMPREHENSIVE ENVIRONMENTAL RESPONSE, COMPENSATION, AND LIABILITY ACT OF
1980, at 686 (1983) [hereinafter LEGIS. HIST.].
19. Id. at 485–86.
20. 42 U.S.C. § 9601(21) (2006).
21. United States v. Bestfoods, 524 U.S. 51, 62–64 (1998).
22. See Riverside Mkt. Dev. Corp. v. Int’l Building Prods., Inc., 931 F.2d 327 (5th Cir. 1991)
(finding corporate officer not liable because he did not personally participate in company’s disposal of
asbestos); Rockwell Int’l Corp. v. IU Int’l Corp., 702 F. Supp. 1384, 1390 (N.D. Ill. 1988) (“Mere ability
to exercise control as a result of the financial relationship of the parties is insufficient for liability to
attach. The entity must actually exercise control.”). But cf. United States v. New Castle County, 727 F.
Supp. 854, 868–69 (D. Del. 1989), aff’d on other grounds sub nom., New Castle County v. Halliburton
NUS Corp., 111 F.3d 1116 (3d Cir. 1997) (declining to find liable a state closely involved in planning
and regulation of disposal site). An “operator” of the polluted facility requiring cleanup, however, is
directly liable. Bestfoods, 524 U.S. at 65. Operate means to “manage, direct, or conduct operations
specifically related to pollution, that is, operations having to do with the leakage or disposal of
hazardous waste, or decisions about compliance with environmental regulations.” Id. at 66–67. The
Supreme Court’s holding in Bestfoods concerning the operator liability of a parent corporation “also
logically applies to cases involving arranger liability.” Raytheon Constructors, Inc. v. Asarco, Inc., 368
F.3d 1214, 1220 (10th Cir. 2003) (citing Carter-Jones Lumber Co. v. Dixie Distrib. Co., 166 F.3d 840,
846 (6th Cir. 1999)).
23. Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870, 1879–80 (2009).
24. Id. at 1879.
376
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
disposal occurs as a peripheral result of the legitimate sale of an unused,
useful product.”25
Illustrations
1. A enters “into a transaction for the sole purpose of discarding a
used and no longer useful hazardous substance.”26 A is liable as an
arranger.27
2. A, the original owner of Facility 1 and the generator of
contaminated mud there, sold the property to B. B subsequently arranged
for disposal of the waste mud at Facility 2, owned and operated by C. C
sued A, alleging that by selling Facility 1, A had “arranged for” disposing of
the site’s leftover contaminated waste mud. Even though in selling the
entire property to B, A also sold whatever waste existed on Facility 1, A did
not arrange for the disposal of the waste mud from Facility 1 to Facility 2.
A is not liable as an arranger.28
3. A sells used transformers containing the hazardous substance PCB
to B. B subsequently, unbeknownst to A, disposes of the transformers in a
way that leads to contamination. A is not liable as an arranger.29
4. A is aware that minor, accidental spills occurred during the transfer
of a product containing a hazardous substance from a common carrier to B’s
bulk storage tanks after the product had arrived at the facility and had come
under B’s stewardship. This knowledge does not support an inference that
A intended such spills to occur. A is not liable as an arranger for the
contamination that occurred at B’s facility.30
E. “Disposal or Treatment”
The statute requires that the subject of the arrangement triggering
liability be “disposal or treatment” of hazardous substances. The statute
broadly defines “disposal” as “the discharge, deposit, injection, dumping,
spilling, leaking, or placing of any solid waste or hazardous waste into or
on any land or water . . . .”31 However, to qualify as an arranger, a person
25. Id. at 1880.
26. Id. at 1878.
27. Id.
28. Jersey City Redevelopment Auth. v. PPG Indus., 655 F. Supp. 1257, 1260–61 (D.N.J.
1987), aff’d, 1988 U.S. Dist LEXIS 16848 (3d Cir. 1987).
29. Fla. Power & Light v. Allis Chalmers Corp., 893 F.2d 1313, 1319 (11th Cir. 1990).
30. Burlington Northern, 129 S. Ct. at 1880.
31. 42 U.S.C. § 9603(3) (2006) (RCRA definition adopted for CERCLA at 42 U.S.C.
§ 9601(29)).
2009]
Restatement for Arranger Liability Under CERCLA
377
must have entered into a transaction “with the intention that at least a
portion of the product [which is the subject of the transaction] be disposed
of . . . by one or more of the methods described” in this definition.32 Simply
knowing that “spilling” and “leaking” may occur as a collateral
consequence of the transaction is not sufficient to justify liability.33
Illustrations
1. A supplies creosote and other hazardous chemicals to B for wood
treatment. A had also been involved in the design, construction, and
continuation of B’s facility, where the wastes were generated and deposited.
A’s transaction was not for the purpose of disposing or treating the
substance at B’s site. A is not liable as an arranger to remedy B’s site.34
2. A shipped used bearings to a Foundry for processing into new
bearings. In exchange for the used bearings, A received credit for the
weight of the bearings against a purchase of new wheel bearings (after a
deduction for the weight of dirt and grease). The used wheel bearings
transported to the Foundry were dirty and broken when they arrived.
Moreover, the bearings were melted down in a process that produced dirt
and slag (both of which were dumped in a back lot with sand, which was
found to be contaminated). However, slag and dust would be produced
even if virgin materials were used to make new bearings. Removal of
contaminants was not the purpose of the transaction; rather, the intent of
32. Burlington Northern, 129 S. Ct. at 1880.
33. Id.; Fla. Power & Light, 893 F.2d at 1317–18, cited in Burlington Northern, 129 S. Ct. at
1878–79; see Freeman v. Glaxo Wellcome, Inc., 189 F.3d 160, 164 (2d Cir. 1999) (finding sale of
unused chemicals not a disposal arrangement); see also Dayton Indep. School Dist. v. U.S. Mineral
Prods. Co., 906 F.2d 1059, 1066 (5th Cir. 1990) (finding that “CERCLA does not provide a remedy for
asbestos removal” under disposal arrangement).
34. Edward Hines Lumber Co. v. Vulcan Materials Co., 861 F.2d 155, 157 (7th Cir. 1988)
(“The statute does not fix liability on slipshod architects, clumsy engineers, poor construction
contractors, or negligent suppliers of on-the-job training—and the fact that [the defendant] might have
been all four rolled into one does not change matters.”); see Kelley ex rel. Mich. Natural Res. Co. v.
ARCO Indus. Corp., 739 F. Supp. 354, 359 (W.D. Mich. 1990) (holding that a sale was not an
arrangement even where hazardous substance was not necessary to functioning of product); Prudential
Ins. Co. v. U.S. Gypsum Co., 711 F. Supp. 1244, 1253 (D.N.J. 1989) (“In the absence of the disposal of
or an arrangement for the disposal of a hazardous substance, liability under § 9607(a)(3) cannot
attach.”); United States v. Westinghouse Elec. Corp., 22 ENV’T REP. CAS. (BNA) 1230, 1233 (S.D. Ind.
1983) (finding no liability for third-party defendant who did not contract for disposal of waste). Contra
United States v. Aceto Agric. Chem. Corp., 872 F.2d 1373, 1379–80 (8th Cir. 1989) (holding a chemical
manufacture liable for cleanup for sending chemical waste to independent contractor to formulate
pesticides, even though manufacturer had no control or knowledge over the formulator’s waste-disposal
practices, because generation of waste by the formulator was inherent in the process for which the
manufacturers had contracted).
378
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
both parties was to create new wheel bearings. A is not liable as an
arranger.35
F. “Of Hazardous Substances Owned or Possessed by Such Person”
The statute requires “proof of ownership, or at least possession” of the
hazardous substance that is the subject of the disposal arrangement as a
required factor for liability.36 Constructive rather than actual possession of
the waste may be sufficient, however, and liability may be established by
demonstrating a defendant’s control over the waste.37
Illustrations
1. A arranged for the disposal of wastes generated by B, due to a fire
at B’s chemical plant, which are transported and disposed of by C, an
independent contractor paid by B. A may or may not be liable as an
arranger, depending upon whether A assumed the obligation to control or
the duty to dispose of the hazardous materials at issue.38
2. A and B entered into an agreement to build a manufacturing plant
that would convert TDA (which B would produce) into TDI (which A
would purchase and use for its own production of urethane foam). The
generation of toxic waste was a natural byproduct of this manufacturing
process. A approved the plant design specifications and capital expenditure
requests. The construction plans specifically provided that the hazardous
waste generated by the TDI Plant would be placed in drums and buried at
35. Pnemo Abex Corp. v. High Point, Thomasville & Denton R. Co., 142 F.3d 769, 775 (4th
Cir. 1998), cited in Burlington Northern, 129 S. Ct. at 1879.
36. Morton Int’l, Inc. v. A.E. Staley Mfg. Co., 343 F.3d 669, 678 (3d Cir. 2003).
37. See id. at 677 (holding inter alia that one of the most important factors of arranger liability
is control over the production process); see also Am. Cyanamid Co. v. Capuano, 381 F.3d 6, 24 (1st Cir.
2004) (finding that “a waste broker may be liable as an arranger if the broker controls the disposal of the
waste”); United States v. Shell Oil Co., 294 F.3d 1045, 1057 (9th Cir. 2002) (finding the United States
not liable as arranger, because it never owned any of the waste products, did not exercise any control
over waste disposal, and did not have an obligation to control the manner of waste disposal); Gen. Elec.
Co. v. Aamco Transmissions, Inc., 962 F.2d 281, 286 (2d Cir. 1992) (finding that obligation to control
wastes sufficient to impose liability); Transp. Leasing Co. v. California, 861 F. Supp. 931, 949 (C.D.
Cal. 1993) (finding that ownership or possession can be actual or constructive). But cf. United States v.
Bliss, 667 F. Supp. 1298, 1306 (E.D. Mo. 1987) (holding broker liable if it has authority to control the
place and manner of the disposal).
38. Bliss, 667 F. Supp. at 1306–07 (holding that a party need not have actual ownership or possession of
the waste to be liable). But see Shell Oil Co., 294 F.3d at 1056 (distinguishing 8th Circuit decisions in
Aceto and Continental Ins. Cos. v. Northern Pharmaceutical & Chemical Co., 842 F.2d 977 (8th Cir.),
cert. denied, 488 U.S. 821 (1988), where the liable parties were either the source of the pollution or
managed its disposal by the arranger because the obligation to exercise control, rather than the mere
ability or opportunity to exercise control, is necessary for arranger liability).
2009]
Restatement for Arranger Liability Under CERCLA
379
an offsite location. A joint committee of the two companies included plans
and preparations for waste disposal, research and recommendations on
waste disposal locations, and approved methods for reducing the volume of
waste sent offsite. Considered together, these facts demonstrate that the
parties arranged for disposal of hazardous substances. A’s control of the
hazardous substances suffices to establish constructive ownership and
possession of the hazardous substances disposed of.39
G. “By Any Other Party or Entity”
Arranger liability applies only where the arrangement is for disposal by
a different party or entity. Because a disposer is liable as an “operator,”
there is no need to resort to the arranger provision to establish liability over
such a person. If a party actually disposes of the hazardous substance itself,
it is liable as “[a] person who at the time of disposal . . . operated any
facility at which such hazardous substances were disposed of.”40
H. “At Any Facility or Incineration Vessel Owned or Operated by
Another Party or Entity and Containing Such Hazardous
Substances”
Arranger liability applies only where the arrangement is at a facility or
vessel owned or operated by another party or entity containing the
hazardous substances that were the subject of the arrangement. Where the
same person who arranged for disposal of hazardous substances also owns
or operates the facility where those substances are disposed, the person is
liable as an “owner or operator” and arranger liability is not needed. Unlike
arranger liability, “owner or operator” liability may exist whether or not the
defendant planned the disposal.41 A present “owner or operator” may be
liable regardless of his involvement (or lack thereof) in the disposal of the
hazardous substances.42 To establish liability under the arranger provision,
the plaintiff must show that the hazardous substances that were the subject
of the defendant’s arrangement for disposal are contained at the facility
39. GenCorp, Inc. v. Olin Corp., 390 F.3d 433, 449 (6th Cir. 2004).
40. 42 U.S.C. § 9607(a)(2) (2006).
41. E.g., Nurad, Inc. v. William E. Hooper & Sons, 966 F.2d 837, 840 (4th Cir. 1992) (holding
that CERCLA section 107(a)(2) imposes liability on the party that owns the facility at the time
hazardous waste leaks from an underground storage tank on the premises, reversing a decision holding
that proof of an affirmative participation in hazardous waste disposal was a prerequisite to liability).
42. § 9607(a)(1); e.g., United States v. Twp. of Brighton, 153 F.3d 307, 330 (6th Cir. 1998),
cited in Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870, 1882 (2009).
380
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
from which there is a release or threat of release.43 Absent direct proof
regarding the removal of defendant’s particular waste from a facility to
which it was sent, a showing that waste chemically similar to waste of the
defendant is sufficient.44 A showing that the defendant intended that his
waste go to the particular facility at issue is unnecessary.45
I. “From Which There Is a Release, or a Threatened Release Which
Causes the Incurrence of Response Costs, of a Hazardous
Substance”
In his prima facie case, the plaintiff need not show that a specific
defendant’s waste caused the incurrence of cleanup-costs.46 Plaintiff need
only prove that the defendant’s hazardous substances were deposited and
contained at the site from which there was a release and that the release
caused the incurrence of response costs.47 This is the only required “nexus”
in the prima facie case.48 Where the plaintiff seeks recovery of natural
resource damages rather than response costs, it must show that the
defendant’s substances are a “contributing factor” to the damages.49
Illustrations
1. B disposed of hazardous substances on its property near the facility
requiring cleanup. B’s disposal, however, was not connected to the
different contamination on plaintiff’s property and no investigation or
remediation had been undertaken in response to the wastes disposed of by
defendant. Plaintiff has not established the required “nexus” between
defendant’s disposal and the incurrence of response costs.50
43. § 9607(a)(3).
44. United States v. Hercules, Inc., 247 F.3d 706, 716 n.8 (8th Cir. 2001).
45. E.g., United States v. Conservation Chem. Co., 619 F. Supp. 162, 234 (W.D. Mo. 1985)
(“[T]he statute does not require that the generator select the site in order to be liable . . . . [T]he statute
imposes liability upon any person who ‘arranged for disposal or treatment . . . at any facility owned or
operated by another party or entity.’”).
46. Hercules, Inc., 247 F.3d at 716 n.8, cited in Burlington Northern, 129 S. Ct. at 1881;
Dedham Water Co. v. Cumberland Farms Dairy, Inc., 889 F.2d 1146, 1151–54 (1st Cir. 1989); Durham
Mfg. Co. v. Merriam Mfg. Co., 294 F. Supp.2d 251, 268 (D. Conn. 2003).
47. United States v. Alcan Aluminum Corp., 964 F.2d 252, 266 (3d Cir. 1992), cited in
Burlington Northern, 129 S. Ct. at 1881; United States v. Union Corp., 259 F. Supp.2d 356, 384 (E.D.
Pa. 2003).
48. Niagara Mohawk Power Corp. v. Consol. Rail Corp., 291 F. Supp.2d 105, 135–37 (N.D.
N.Y. 2003).
49. Coeur D’Alene Tribe v. Asarco, Inc., 280 F. Supp.2d 1094, 1124 (D. Idaho 2003); United
States v. Montrose Chem. Corp. of Cal., 33 Env’t Rep. Cas. (BNA) 1207 (C.D. Cal. 1991).
50. Niagara Mohawk Power Corp., 291 F. Supp.2d at 136.
2009]
Restatement for Arranger Liability Under CERCLA
381
2. Hazardous substances originating at site A, owned by plaintiff, and
site B, owned by defendant, commingled. Site A requires remediation. The
plaintiff, on summary judgment, (a) identifies a contaminant at site A, (b)
identifies the same contaminant (or perhaps a chemically similar)
contaminant at the defendant’s site B, and (c) provides evidence of a
plausible migration pathway by which the contaminant could have traveled
from the defendant’s facility to the plaintiff’s site. The plaintiff has met its
burden of production. To avoid summary judgment, defendant must proffer
evidence sufficient to create a genuine issue of material fact as to its ability
to disprove causation.
3. A operated a portion of the Water Street Site, the Wynantskill
Creek, into which A spilled a number of hazardous substances. B brought
an action seeking to recover costs in response to releases of such hazardous
substances at Area 2 of the Water Street site. B fails to prove that hazardous
substances migrated from the Wynantskill Creek onto Area 2, which B
remediates. No investigations or remedial actions by B relate to
Wynantskill Creek. B has not shown the required nexus between any
purported release on A’s property and remediation at Area 2. Thus, A is not
liable.51
Even where a person falls within the scope of arranger liability, he may
escape partial or total responsibility for response costs or damages if those
costs or damages can be divided by causation.52 For example, no liability
accrues where spills of substances that were the subject of the arrangement
do not require remediation.53 The Restatement (Third) of Torts explains
that “[d]amages can be divided by causation when” any person or group of
persons “to whom the factfinder assigns a percentage of responsibility [or
any tortious act of such a person] was a legal cause of less than the entire
damages.”54 Divisibility is entirely a technical, scientific inquiry having
nothing to do with culpability, cooperation, or any other conduct factors
that may bear on the allocation of costs in a CERCLA contribution action.55
Any basis upon which contributions of a defendant can be measured or
51. Id. at 135–36.
52. The topic of joint and several liability after Burlington Northern is the subject of another
article by this author, Alfred R. Light, Restatement for Joint and Several Liability Under CERCLA After
Burlington Northern, 39 ENVTL. L. REP. 11058 (2009).
53. Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870, 1882–83 (2009)
(discounting the railroad’s liability because chemical contamination at the railroad parcel did not require
remediation).
54. RESTATEMENT (THIRD) OF TORTS: APPORTIONMENT OF LIABILITY § 26(b)(1) (1999).
55. Gerald W. Boston, Apportionment of Harm in Tort Law: A Proposed Restatement, 21 U.
DAYTON L. REV. 267, 359 (1996).
382
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
compared to contributions of others may provide a basis for divisibility.56
For a person held liable as an arranger, the basis for comparison may be
volume57 or comparative toxicity.58 Some bases which a court may use to
divide damages by causation also may be equitable factors, which the same
court may use to allocate costs among jointly and severally liable persons
where damages are not divisible.59
4. A manufactured aluminum sheet and plate products, which
produced as a byproduct an emulsion consisting of ninety-five percent
deionized water and five percent mineral oil. Trace quantities of hazardous
substances were also contained in the emulsion. A arranged for the disposal
of used emulsion with B, who disposed of it through a borehole into the
site. The release of hazardous substances at the site caused the government
to incur response costs. After assessing the relative toxicity, migratory
potential, and synergistic capacity of substances in the release causing the
incurrence of response costs at the site, A showed that the emulsion did not
and could not, when mixed with other hazardous substances, contribute to
the release and the resultant response costs. Though nominally a liable
56. Id. at 360; see United States v. Hercules, Inc., 247 F.3d 706, 719 (8th Cir. 2001) (explaining
that it is possible to prove divisibility of single harms based on volumetric, chronological, and other
types of evidence or establishment of “non-contiguous areas of contamination”).
57. In re Bell Petroleum Servs. Inc., 3 F.3d 889, 901–02 (5th Cir. 1993), cited in Burlington
Northern, 129 S. Ct. at 1881; Kamb v. U.S. Coast Guard, 869 F. Supp. 793, 799 (N.D. Cal. 1994)
(apportioning liability based on the volume of lead each defendant contributed to the site and based on
the divisibility of the site into two discreet sections—one section used by defendants and one unused by
defendants).
58. Control Data Corp. v. S.C.S.C. Corp., 53 F.3d 930, 937–38 (8th Cir. 1995) (allocating by
toxicity in a contribution action).
59. Courts frequently make reference to the Gore factors, named for Rep. Al Gore in an
amendment that did not become part of CERCLA, as equitable allocation factors. These include some
factors potentially relevant to divisibility, such as the distinguishability of a defendant’s discharge from
other discharges, amount of hazardous waste involved, degree of toxicity of the hazardous waste
involved, and the degree of involvement by the parties in generation, transportation, treatment, storage,
or disposal of the wastes. Other Gore factors appear unrelated to causation, such as the degree of care
exercised with respect to the hazardous waste concerned and the degree of cooperation with Federal,
State, or local officials to prevent any harm to the public health or environment. United States v. Twp.
of Brighton, 153 F.3d 307, 318 n.16 (6th Cir. 1998) (“Some of the Gore factors (1, 2, and 3) are
compatible with causation analysis; others (5 and 6) reflect fairness concerns; at least one (4) does
both.”), cited in Burlington Northern, 129 S. Ct. at 1881; Allied Signal, Inc. v. Amcast Int’l Corp., 177
F. Supp. 713 (S.D. Ohio 2001) (listing and analyzing the Gore factors and describing their legislative
background); United States v. Cantrell, 92 F. Supp. 2d 704, 711 (S.D. Ohio 2000) (“Divisibility
determinations are to be based on legal considerations of causation, not of equitable considerations of
fairness. Considerations of fairness which are not relevant to divisibility determinations include the
degree of care taken by the PRPs and the degree of cooperation by the PRPs with the government to
prevent harm to the environment or public.”) (internal citation omitted).
2009]
Restatement for Arranger Liability Under CERCLA
383
party under CERCLA as an arranger, A has divided the damages by
causation and is not liable for any response costs.60
5. A owned the real property at the site, a twenty-four-block area north
of the Brazo Street facility, a chrome-plating shop operated successively
from 1971 through 1977 by A, B, and C. A owned the property from 1967
through 1981 and conducted chrome-plating activities there in 1971 and
1972. In 1972, B purchased the shop and leased the property from A. B
continued to conduct similar, but more extensive, chrome-plating activities
there until mid-1976. In August 1976, C purchased the assets from B,
leased the property from A, and conducted similar chrome-plating activities
there until late 1977. The release of hazardous substances at the site caused
the government to incur response costs and to sue A, B, and C for
reimbursement. Various witnesses testify regarding the rinsing and
wastewater disposal practices of each defendant and the amount of chromeplating activity conducted by each. C introduces expert testimony
regarding a volumetric approach to apportionment, calculating the total
amount of chromium that had been introduced into the environment by A,
B, and C, collectively and individually. A second expert estimated the
amount of chromium on the basis of electrical usage records. C has
established a reasonable basis for division of the damages by causation.61
“Whether damages are divisible is a question of fact.”62 “A party
alleging that damages are divisible [ordinarily] has the burden to prove they
are divisible.”63 “The magnitude of each divisible part is also a question of
fact. The burden to prove the magnitude of each part is on the party who
60. United States v. Alcan Aluminum Corp., 964 F.2d 252, 268–69, 271 (3d Cir. 1992), cited in
Burlington Northern, 129 S. Ct. at 1881. In this illustration—unlike in Alcan on remand—A proffered
sufficient evidence showing that the emulsion could not have contributed to the release and is therefore
able to avoid liability. Cf. United States v. Alcan Aluminum Corp., 892 F. Supp. 648, 657 (M.D. Pa
1995) (granting summary judgment in favor of the Government and full reimbursement of its response
costs upon Alcan’s failure to present evidence on apportionment of the harm).
61. In re Bell Petroleum Servs., Inc., 3 F.3d at 902–04, cited in Burlington Northern, 129 S. Ct.
at 1881. Professor Boston characterizes this decision and the two circuit decisions involving Alcan
Aluminum Corporation as “agreeing that divisibility of the harm could be established even if wastes
were commingled.” Boston, supra note 55, at 357.
62. RESTATEMENT (THIRD) OF TORTS: APPORTIONMENT OF LIABILITY § 26 cmt. h (1999); see
generally Price v. U.S. Navy, 39 F.3d 1011, 1018 (9th Cir. 1994) (affirming assignment of percentages
of liability to two defendants [95% to Navy, 1% to home builder] and declining to second guess
discretion of court in apportioning liability). But cf. In re: Dana Corp., 379 B.R. 449, 457–58 (S.D. N.Y.
2007) (divisibility analysis is “intensely factual” but the preliminary matter of whether harm is capable
of apportionment is a question of law).
63. RESTATEMENT (THIRD) OF TORTS: APPORTIONMENT OF LIABILITY § 26 cmt. h (1999); see
Raytheon Aircraft Co. v. United States, 532 F. Supp. 2d 1306, 1310 (D. Kan. 2007) (placing divisibility
burden on Government as defendant in private cost recovery action); In re: Dana Corp., 379 B.R. at
457.
384
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
Notwithstanding this ordinary burden, both the
seeks division.”64
Restatement (Second) and the Restatement (Third) of Torts have noted the
potential unfairness of this general rule because it can impose full liability
on a defendant who only caused part of the damages.65 The Third
Restatement recommends that the “more attractive solution is to place the
burden of proof on the party seeking to avoid responsibility for the entire
injury, along with relaxing the burden of production.”66 Under Burlington
Northern where an apportionment analysis is demanded by the
circumstances of a case, a court may apportion liability sua sponte, even if
not advanced by a defendant.67
J. Defenses
Affirmative defenses expressly set forth in CERCLA are found in
sections 107, 113, and 122.68 They are not addressed in this Restatement.69
64. RESTATEMENT (THIRD) OF TORTS: APPORTIONMENT OF LIABILITY § 26 cmt. h (1999).
65. Id.; RESTATEMENT (SECOND) OF TORTS § 433B (1964).
66. RESTATEMENT (THIRD) OF TORTS: APPORTIONMENT OF LIABILITY § 26 cmt. h (1999); see
United States v. Alcan Aluminum Corp., 990 F.2d 711, 722 (2d Cir. 1993) (“Alcan as the polluter bears
the ultimate burden of establishing a reasonable basis for apportioning liability.”); United States v. Alcan
Aluminum Corp., 2006 U.S. Dist. LEXIS 24309, *34–36 (N.D. N.Y. Apr. 26, 2006) (denying motion for
summary judgment in favor of jury trial on divisibility, indicating potential divisibility of Alcan wastes
“chronologically, physically, and geographically,” apportionment by volume, and an analysis of issues
of relative toxicity, migratory potential, and synergy); United States v. Agway, Inc., 193 F. Supp. 2d 545,
548 (N.D. N.Y. 2002) (“To defeat a motion for summary judgment on the issue of divisibility, a
defendant ‘need only show that there are genuine issues of material fact regarding a reasonable basis for
apportionment of liability.’” (quoting Alcan Aluminum Corp., 990 F.2d at 722)); United States v.
Chrysler Corp., 29 Chem. Waste Litig. Rep. 566, 576 (D. Del. Dec. 9, 1994) (“[A]n evidentiary hearing
should be conducted and summary judgment should not be granted when there is a genuine issue of
material fact as to whether there is a reasonable basis for apportioning liability.”).
67. In Burlington Northern, the parties “left the court to independently perform the equitable
apportionment analysis demanded by the circumstances of the case.” Burlington N. & Santa Fe Ry. Co.
v. United States, 129 S. Ct. 1870, 1881 (2009). The Supreme Court’s upholding implicitly found the
District Court’s sua sponte apportionment analysis permissible. Id. It did this despite the United
States’s argument that this mode of procedure deprived it of a fair opportunity to respond to the court’s
theories of apportionment and to rebut their factual underpinnings. Id. at 1886 (Ginsburg, J.,
dissenting).
68. 42 U.S.C. §§ 9607, 9613, 9622 (2006).
69. That only express defenses are recognized does not mean that certain other universally
applicable legal principles, sometimes denominated as affirmative defenses but not explicitly endorsed
in CERCLA’s language (e.g., collateral estoppel, accord and satisfaction, and res judicata), are
unavailable. See Town of Munster v. Sherwin-Williams Co., 27 F.3d 1268, 1272 (7th Cir. 1994)
(explaining that a CERCLA defendant should be able to raise res judicata as a matter of public policy);
Cal. Dep’t. of Toxic Substances Control v. Alco Pac., Inc., 217 F. Supp. 2d 1028, 1038 (C.D. Cal. 2002)
(noting that res judicata, collateral estoppel, and accord and satisfaction are available to CERCLA
defendants).
2009]
Restatement for Arranger Liability Under CERCLA
385
III. REPORTER’S NOTES
A. Scope
The parameters of the arranger or “generator” liability provision set
forth in section 107(a)(3) have been among the most fiercely litigated issues
under CERCLA. As the Supreme Court put it in Burlington Northern, the
primary reason for fierce litigation has been the government’s push for
CERCLA liability to “extend beyond the limits of the statute itself.”70
Without clear statutory language or legislative history supporting its
position, the government resorted to policy arguments. For example, they
argued that CERCLA intends to place responsibility for cleanup on “those
responsible for problems caused by the disposal of chemical poisons.”71
They also counseled for a “liberal judicial interpretation . . . consistent with
CERCLA’s ‘overwhelmingly remedial’ statutory scheme”72 and broad,
conclusory principles not to interpret the statute “in any way that apparently
frustrates the statute’s goals, in the absence of a specific congressional
intent otherwise.”73 A broad goal is simply to provide the government with
the “tools” it desires for a “prompt and effective response to the problems
of national magnitude resulting from hazardous waste disposal.”74 It
buttressed the approach by resorting to common law liability principles,
without reference to CERCLA’s statutory language.75 The government has
used this approach in other areas of CERCLA liability.76
Starting in the late 1990s, however, circuit courts began to put an end to
this policy approach, which amounts to the notion that the government
should win simply because it is the government. Burlington Northern,
which uses the traditional approach to statutory interpretation found in the
opinions of Justice Thomas in Atlantic Research77 and Cooper Industries,78
70. Burlington Northern, 129 S. Ct. at 1879.
71. United States v. Aceto Agric. Chem. Corp., 872 F.2d 1373, 1377 (8th Cir. 1989) (quoting
Dedham Water Co. v. Cumberland Farms Dairy, Inc., 805 F.2d 1074, 1081 (1st Cir. 1986)).
72. Aceto, 872 F.2d at 1380 (quoting United States v. Ne. Pharmaceutical & Chem. Co., 810
F.2d 726, 733 (8th Cir. 1986)).
73. Id. (quoting Dedham Water Co., 805 F.2d at 1081).
74. Id.
75. Id. at 1379 (dealing with the liability of the employer of an independent contractor) (citing
RESTATEMENT (SECOND) OF TORTS § 427A (1964)).
76. See Alfred R. Light, CERCLA’s Cost Recovery Statute of Limitations: Closing the Books or
Waiting for Godot? 16 SE. ENVTL. L.J. 248, 281–84 (2008) (arguing that the earliest courts interpreting
CERCLA improperly deferred to the government’s litigation positions in interpreting CERCLA’s statute
of limitations using “conclusory policy reasoning”).
77. United States v. Atl. Research Corp., 551 U.S. 128, 135–37 (2007) (interpreting CERCLA
section 107(a)(4)(B)).
386
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
indicates that the United States Supreme Court also squarely rejects this
jurisprudential approach.79
B. History
Despite the well-deserved notoriety of CERCLA’s legislative process in
1980, the arranger liability provision in section 107(a)(3) bears hallmarks of
careful legislative attention to the language delimiting the parameters of the
provision. In 1980, the Senate Committee on Environment and Public
Works narrowed CERCLA’s language, holding liable any “person who
caused or contributed to or is causing or contributing to . . . such release,
including . . . generators.”80 The description of “generator” in the bill tracks
the current language, except that it requires the hazardous substances be “at
facilities or sites owned or operated by such other party or entity,” seeming
to imply a requirement that the “generator” arrange for disposition of waste
at the other contracting party’s facility.81 If so, the final language negotiated
in the lame-duck session clarified or broadened the “generator” category to
include a broader class of disposal arrangements through deletion of the
word “such.” That Senators on this Committee (or committee staff)
focused on “such” is of some interest in light of the problematic judicial
construction of the word as it appears in the final language.82 The final
language, however, is narrower than the language in the bill that had passed
the House, which was very similar to the language in the original Senate
bill described above.83
C. “Person”
The Supreme Court’s 1998 decision in Bestfoods84 superseded a number
of lower court opinions that had held derivatively liable a person who
exercised pervasive control over a liable entity even in the absence of
78. See Cooper Indus., Inc. v. Aviall Servs., Inc., 543 U.S. 157, 165–66 (2004) (interpreting
CERCLA section 113(f)).
79. Light, supra note 76, at 284.
80. LEGIS. HIST., supra note 18, at 169.
81. Id. at 486 (emphasis added).
82. See infra notes 99–103 and accompanying text (discussing the literal interpretation of
“such” versus traditional canons of statutory construction).
83. H.R. REP. NO. 96-510, § 107(a)(4) (1980) (adding section 3071(a)(1)(D)(b), LEGIS. HIST.,
supra note 18, at 438) (imposing liability on “any person who caused or contributed to the release or
threatened release”).
84. United States v. Bestfoods, 524 U.S. 51 (1998).
2009]
Restatement for Arranger Liability Under CERCLA
387
specific evidence about control over waste-disposal practices.85 An opinion
holding a secured creditor liable on such a theory has been superseded by
statutory amendment.86
D. “By Contract, Agreement, or Otherwise Arranged For”
At the time Burlington Northern was decided, several circuits had
adopted standards similar to the Ninth Circuit’s, which the Supreme Court
disapproved. One district court categorized arranger case law prior to
Burlington Northern as dividing into three distinct approaches: (1) a strict
liability approach; (2) a specific intent approach; and (3) a “totality of the
circumstances” or case-by-case approach.87 Plainly, Burlington Northern
limits liability to situations where there is an intent to purposefully dispose
or treat, opting for the specific intent approach.88
85. E.g., United States v. Kayser-Roth Corp., 910 F.2d 24, 27–28 (1st Cir. 1990) (finding the
parent corporation was an “operator” because of its pervasive control over subsidiary empowering it to
control release or threat of release of hazardous substance); see United States v. Nicolet, Inc., 712 F.
Supp. 1193, 1202 (E.D. Pa. 1989) (“Where a subsidiary is or was at the relevant time a member of one
of the classes of persons potentially liable under CERCLA; and the parent had a substantial financial or
ownership interest in the subsidiary; and the parent corporation controls or at the relevant time
controlled the management and operations of the subsidiary, the parent’s separate corporate existence
may be disregarded.”); see also Idaho v. Bunker Hill Co., 635 F. Supp. 665, 671–72 (D. Idaho 1986)
(holding parent liable because familiar with subsidiary’s waste disposal practices, and having capacity to
control, placed limits on subsidiary’s pollution control expenditures, and other factors). The Eleventh
Circuit in United States v. Fleet Factors Corp., 901 F.2d 1550, 1557 (11th Cir. 1990) said in dictum that
a security creditor “may incur liability, without being an operator, by participating in the financial
management of a facility to a degree indicating a capacity to influence the corporation’s treatment of
hazardous wastes.”
86. 42 U.S.C. § 9601(20)(F)(i)(II) (2006) (stating that the term “participation in management”
“does not include merely having the capacity to influence, or the unexercised right to control, vessel or
facility operations”).
87. United States v. Gordon Stafford, Inc., 952 F. Supp. 337, 339–40 (N.D. W. Va. 1997); Fleet
Factors, 901 F.2d at 1550; Randy Boyer, Note, Morton Int’l, Inc. v. A.E. Staley Mfg. Co.: The Third
Circuit Establishes a Standard for Arranger Liability, 17 TULANE ENVTL. L.J. 201, 204–05 (2003).
88. E.g., Morton Int’l, Inc. v. A.E. Staley Mfg. Co., 343 F.3d 669, 678 (3d Cir. 2003) (rejecting
a strict liability approach to arranger liability); see S. Fla. Water Mgmt. Dist. v. Montalvo, 84 F.3d 407
(11th Cir. 1996) (recognizing that intent is one relevant factor to determine arranger liability); see also
United States v. Cello-Foil Prods., Inc., 100 F.3d 1227, 1232 (6th Cir. 1996) (“[I]n the absence of a
contract or agreement, a court must look to the totality of the circumstances, including any ‘affirmative
acts to dispose,’ to determine whether the Defendants intended . . . an arrangement for disposal. . . . .
[A] party can be responsible for ‘arranging for’ disposal, even when it has no control over the process
leading to the release of substances.”); Amcast Indus. Corp. v. Detrex Corp., 2 F.3d 746, 751 (7th Cir.
1993), cited in Burlington Northern, 129 S. Ct. at 1879 (explaining that the “words ‘arranged for’ imply
intentional action”); FMC Corp. v. U.S. Dep’t of Commerce, No. 92-1945, 1993 WL 489133 (3d Cir.
1993) (finding Congress’s intent when drafting CERCLA was not to impose strict liability on the
government for hazardous WWII facilities), rev’d, FMC Corp. v. U.S. Dep’t of Commerce, 29 F.3d 833
(3d Cir. 1994); Fla. Power & Light v. Allis Chalmers Corp., 893 F.2d 1319 (11th Cir. 1990) (granting
motion for summary judgment where party could not support the contention that manufacturers intended
388
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
E. “Disposal or Treatment”
Read in the context of the phrase, “arranged for disposal or treatment,”
the phrase “disposal or treatment” requires that there be an intentional
human act by the “person” potentially liable under section 107(a)(3).89
Similarly, the phrase “arranged with a transporter for transport for disposal
or treatment” does not reach a shipper who arranged for transport of a
useful product containing hazardous substances that are spilled en route,
since the arrangement for shipment of the product does not imply an intent
to “dispose” of the product through such accidental spillage.90 This does
not necessarily mean that “disposal” in other CERCLA contexts implies
intentional action. “In the context of the operator of a hazardous-waste
dump, ‘disposal’ includes accidental spillage.”91 Thus, Burlington Northern
does not necessarily carry implications for cases that have imposed liability
under section 107(a)(2) on a “person who at the time of disposal owned or
operated any facility at which such hazardous substances are disposed of.”92
Whether the terms “disposal” and “disposed of” in that paragraph imply
affirmative human conduct is beyond the scope of this arranger
Restatement.93
to dispose of the waste); United States v. Aceto Chem. Corp., 872 F.2d 1373 (8th Cir. 1989); New York
v. Gen. Elec. Co., 592 F. Supp. 291 (W.D. N.Y. 1984); see generally Benjamin J. Rodkin, Note,
Deciphering CERCLA’s Vocabulary: United States v. Burlington–“Reasonable” Division and
“Arranger” Liability, 20 VILL. ENVTL. L.J. 275 (2009) (evaluating a Ninth Circuit decision to impose
joint and several liability on defendants in Burlington Northern); Boyer, supra note 87, at 206
(discussing a Seventh Circuit decision establishing a specific intent approach).
89. Amcast Indus. Corp. v. Detrex Corp., 2 F.3d 746, 751 (7th Cir. 1993).
90. Id.
91. Id.
92. 42 U.S.C. § 9607(a)(2) (2006).
93. Several circuit courts have held that passive migration does not constitute a “disposal”
within the meaning of CERCLA. Bob’s Beverage, Inc. v. Acme, Inc., 264 F.3d 692, 697 (6th Cir. 2001);
United States v. 150 Acres of Land, 204 F.3d 698, 705 (6th Cir. 2000); ABB Indus. Sys., Inc. v. Prime
Tech, Inc., 120 F.3d 351, 357–58 (7th Cir. 1997); United States v. CDMG Realty Co., 96 F.3d 706, 713
(3d Cir. 1996); see Tanglewood E. Homeowners v. Charles Thomas, Inc., 849 F.2d 1568, 1573 (5th Cir.
1998) (finding “disposals” may occur when a party releases previously existing hazardous substances
into the environment through landfill excavations and fillings); Nurad, Inc. v. William E. Hooper &
Sons, Co., 966 F.2d. 837, 844–46 (4th Cir. 1992) (finding CERCLA’s definition of “disposal” includes
passive migration); United States v. 175 Inwood Assocs., 330 F. Supp. 2d 213, 225–26 (E.D. N.Y. 2004)
(finding defendants liable as owners of the site at the time drums of hazardous waste were releasing
chemicals into the ground); Ecodyne Corp. v. Shah, 718 F. Supp. 1454, 1457 (N.D. Cal. 1989) (allowing
civil action only against prior owners or operators who owned the land at the time the hazardous
substances were introduced); Michael S. Caplan, Escaping CERCLA Liability: The Interim Owner
Passive Migration Defense Gains Circuit Recognition, 28 Envtl. L. Rep. (Envtl. L. Inst.) 10,121 (1998)
(analyzing “CERCLA’s language and structure as it pertains to liability for passive migration of
hazardous substances”); Robert L. Bronston, Note. The Case Against Intermediate Owner Liability for
Passive Migration of Hazardous Waste, 93 MICH. L. REV. 609 (1994) (arguing that “Congress intended
2009]
Restatement for Arranger Liability Under CERCLA
389
F. “Of Hazardous Substances Owned or Possessed by Such Person”
Prior to Burlington Northern, the broad interpretation of “arranged for”
advocated by the Environmental Enforcement Section of the Department of
Justice (the “A Team”) presented serious challenges for the Environmental
Defense Section (the “B Team”) in its defense of federal agencies seeking
to avoid liability under section 107(a)(3). For example, Shell Oil argued in
a counterclaim against the United States that if a party has “substantial
control over a manufacturing process wherein a hazardous waste stream is
generated and disposed of, then that party assumes the obligation to control
the disposal of that waste stream.”94 The Ninth Circuit ruled in favor of the
United States, holding that it was not liable on the counterclaim because it
never actually “owned or possessed [the waste] and never had an authority
to control or duty to dispose of, the hazardous materials.”95 In effect, by
winning this case the Government undermined its earlier precedents as
plaintiff, holding waste brokers and other such intermediaries liable. The
Shell Oil precedent narrowed the circumstances in which such persons
might be said to have “owned or possessed” the substances disposed of.96
As pointed out in law review commentary in early 2009, the Ninth Circuit’s
distinguishing of its earlier opinion in Shell in its Burlington Northern
decision (ironically involving the same corporation on the “other side” of
the same issue) is unpersuasive because it diluted the level of control over
substances that trigger liability to mere awareness of spills.97 Because the
Burlington Northern decision focuses on the absence of the intent required
to show that Shell “arranged for” the disposal or treatment of hazardous
substances, the law regarding the separate element that the substances be
“owned or possessed” by that person remains confused. The matter is
likely to be litigated in cases where the government sues as arranger
persons who brokered arrangements between generators and disposers but
who never actually owned or possessed the generator’s substances. In
short, it is unclear whether the “A Team” or the “B Team” now holds the
upper hand.
disposal to have an active meaning and therefore that courts should not hold prior intermediate owners
liable for the passive migration of hazardous waste”).
94. United States v. Shell Oil Co., 294 F.3d 1045, 1055 (9th Cir. 2002).
95. Id. at 1058.
96. See, e.g., United States v. Bliss, 667 F. Supp. 1298, 1306–07 (E.D. Mo. 1987) (representing
the previous precedent that a party does not need to have actual ownership of the waste to be held liable,
just the ability to exercise control of the disposal, as shown in Part II.F.Illustration.1).
97. Rodkin, supra note 88, at 275, 302–04.
390
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
G. “By Any Other Party or Entity”
There is no reason to think that the requirement that only “off-site”
arrangers are liable would generate any controversy in light of the broad
scope of liability of past and present owners and operators under section
107(a)(1) and section 107(a)(2).98
H. “At Any Facility or Incineration Vessel Owned or Operated by
Another Party or Entity and Containing Such Hazardous
Substances”
In the context of arranger liability, the Court has emphasized the
“ordinary meaning” of statutory language when interpreting “arranged.”99
This has led to the consequent view that a defendant’s knowledge of spills
is only relevant if it is indicative of the ultimate issue of whether the
defendant intended spills to occur.100 This view hints that the Court would
approach the significance of the phrase “containing such hazardous
substances” differently from the position that the Government has argued,
often successfully, in the lower courts. Beginning with United States v.
South Carolina Recycling & Disposal, Inc.,101 some lower courts adopted
the Government’s view that Congress intended to omit any causation
requirement in CERCLA and read section 107(a)(3) only to require that the
plaintiff show that hazardous substances “alike, similar, or of a like kind”
are contained at the facility rather than to require a showing that the
defendant’s substances, which were the subject of his disposal arrangement,
are contained at the site.102 This construction, however, presents logical
difficulties. Read literally with this interpretation of “such,” any person
who arranges for disposal of a substance at a facility is liable for cleanup of
the facility, whether their waste ever went to the facility at all. The Fourth
Circuit solved this difficulty by inferring a requirement that the generator’s
waste must be “sent” to the site needing cleanup.103 But this inference
offends traditional canons of statutory construction because there is no
textual basis for this judicial gloss. Burlington Northern suggests that the
Supreme Court might well reach the same practical result as the Fourth
Circuit in Monsanto, but that it would use a different jurisprudential
98. 42 U.S.C. § 9607(a)(2) (2006).
99. Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870, 1879 (2009).
100. Id. at 1880.
101. United States v. S.C. Recycling & Disposal, Inc., 653 F. Supp. 984, 992 (D.S.C. 1984),
aff’d, United States v. Monsanto Co., 858 F.2d 160 (4th Cir.1988).
102. Monsanto, 858 F.2d at 169.
103. Id. at 169 n.15.
2009]
Restatement for Arranger Liability Under CERCLA
391
approach. The Court might rule that the plaintiff could meet its burden of
production on the issue of whether the defendant’s waste is contained at the
facility by showing (1) the arrangement for disposal at the facility and (2)
that waste similar to that of the defendant was at the site at the time of
remediation. In other words, absent direct proof that a defendant’s waste
had been removed from the site, an inference that the defendant’s waste is
“contained at the site” would arise from this showing. This would avoid the
strained construction of the word “such,” which the Government initially
wrote into the South Carolina Recycling district court opinion it drafted.104
I. “From Which There Is a Release, or a Threatened Release Which
Causes the Incurrence of Responses Costs, of a Hazardous
Substance”
As several circuit courts candidly have admitted, even if the plaintiff
does not have to prove a precise nexus between the defendant and the
release requiring cleanup, the potential that damages may be apportioned,
and liability limited under Restatement principles, means that a defendant
“may escape any liability for response costs if it either succeeds in proving
that [its hazardous substance] did not contribute to the release and the
cleanup that followed, or contributed at most only to a divisible portion of
the harm.”105 Though Burlington Northern does not address causation as a
separate element of proof under CERCLA, its approval of the district
court’s reduction of Railroad responsibility based in part on the absence of a
substantial contribution of a particular chemical, D-D, to the contamination
remediated implicitly acknowledges this method of escaping CERCLA
liability.106
Under the Restatement standard applicable under CERCLA, explained
in Burlington Northern, a single harm is divisible when it is possible to
discern the degree to which different parties contributed to the damage.107
104. See Alfred R. Light, The Importance of “Being Taken”: To Clarify and Confirm the
Litigative Reconstruction of CERCLA’s Text, 18 B.C. ENVTL. AFF. L. REV. 1, 31–32 (1990) (describing
the district court’s crafting of the word “such” in South Carolina Recycling).
105. United States v. Hercules, Inc., 247 F.3d 706, 716 (8th Cir. 2001), cited in Burlington N. &
Santa Fe Ry. Co. v. United States, 129 S. Ct. at 1881; United States v. Alcan Aluminum Corp., 990 F.2d
711, 722 (2d Cir. 1993), cited in Burlington Northern, 129 S. Ct. at 1881 (2009); see Textron, Inc. v.
Barber-Coleman Co., 903 F. Supp. 1570, 1583 (W.D.N.C. 1995) (“It seems the point is to shift the
burden of proving causation to defendants.”).
106. Burlington Northern, 129 S. Ct. at 1882–84.
107. See Burlington Northern, 129 S. Ct. at 1881 (“[T]he District Court ultimately concluded
that this was ‘a classic divisible in terms of degree case . . . .’”); see also Hercules, 247 F.3d at 718, cited
in Burlington Northern, 129 S. Ct. at 1881; United States v. Vertac Chem. Corp., 453 F.3d 1031, 1040
392
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
Divisibility may be provable even where wastes have become crosscontaminated and commingled, for “commingling is not synonymous with
indivisible harm.”108
Though the burden of proof on divisibility ordinarily lies with the
defendant, this is not always the case in light of the Supreme Court’s
affirmation of the district court’s independent apportionment in that case.
After Burlington Northern, the plaintiff, frequently the Government, more
often will have to come forward with its own evidence on the issue and will
not be able to avoid trial of the issue through summary judgment.109 Most
of the early CERCLA decisions regarding joint and several liability were
district court rulings on motions for summary judgment. For example, the
Chem-Dyne decision, which the Burlington Northern court followed, is an
opinion resolving arranger defendants’ motion for partial summary
judgment regarding their joint and several liability, in which the defendants
only presented their legal argument that “because joint and severally[sic]
liability is not expressly provided for in CERCLA, there is no basis for its
imposition.”110 As to divisibility, the court only found “an insufficient
evidentiary basis, with unresolved factual questions, which precludes the
resolution of this case in the form of a summary judgment motion.”111
The Supreme Court’s endorsement of the district court’s independent
apportionment (in the sense of divisibility of the damages by causation)
also avoids serious constitutional difficulties that would be presented were
CERCLA to impose disproportionate, retroactive liability for all damages
where an arranger defendant’s proven maximum contribution to the harm is
comparatively slight. The Court’s endorsement protects the proverbial
“one-drum” contributor to a massive hazardous waste disposal facility
(8th Cir. 2006) (ultimately concluding that Hercules did not establish geographic divisibility by operable
units of the remediation).
108. Hercules, 247 F.3d at 718 (quoting Alcan Aluminum Corp., 990 F.2d at 722), cited in
Burlington Northern, 129 S. Ct. at 1881.
109. So long as there is a “genuine issue of material fact” as to divisibility of the damages, the
plaintiff’s motion for summary judgment must be denied even though the defendant bears the ultimate
burden of persuasion at trial. FED. R. CIV. P. 56(c).
110. United States v. Chem-Dyne Corp., 572 F. Supp. 802, 810 (S.D. Ohio 1983).
111. Id. at 811. But cf. Transport Leasing Co. v. Cal. (Caltrans), 861 F. Supp. 931, 939–40 (C.D.
Cal. 1992) (showing divisibility of harm could not be resolved in the context of pre-trial motions); Am.
Int’l, Inc. v. DataCard Corp., 146 B.R. 391, 402 (N.D. Ill. 1992) (adopting report and recommendations
of magistrate judge denying summary judgment and finding that divisibility issue involved unresolved
questions of fact); Hatco Corp. v. W.R. Grace & Co., 801 F. Supp. 1309, 1327–31 (D.N.J. 1992)
(denying plaintiff’s summary judgment motion on divisibility because defendant raised “sufficient fact
questions as to the relative degrees of harm caused by [the plaintiff]”); Weyerhaeuser Corp. v. Koppers
Co., 771 F. Supp. 1406, 1416 (D. Md. 1991) (declining to rule on apportionment issue at summary
judgment phase); United States v. Marisol, Inc., 725 F. Supp. 833, 841–42 (M.D. Pa. 1989) (defendants
are at the very least entitled to conduct some discovery of defenses alleging divisibility of the harm).
2009]
Restatement for Arranger Liability Under CERCLA
393
requiring cleanup. The Department of Justice consistently maintained in
Congressional hearings during the CERCLA reauthorization process in
1985 that it did not intend and did not think it could “impose liability for
one hundred percent of the costs on a de minimis generator.”112 The
Assistant Attorney General for Environment and Natural Resources
explained, “Congress indicated in the legislative history in 1980 that joint
and several liability was deleted expressly from the statute because it
should be left to the courts to deal with on a case-by-case basis so that it
would not be oppressively or unfairly applied in inappropriate
circumstances.”113 Even if “CERCLA . . . does not require causation as a
prerequisite to liability,”114 a less stringent attitude toward the burden
regarding apportionment permits courts to avoid violating the requirement
of “fair play and substantial justice” embodied in the Due Process Clause of
the Fifth Amendment.115 If a generator’s maximum contribution is small
enough, its apportioned share of the liability may be nothing, particularly in
light of section 107(o), added to CERCLA in 2002, presumptively
exempting persons who arranged for disposal or treatment of less than 110
gallons of liquid materials or less than 200 pounds of solid wastes from
liability for response costs at a facility on the National Priorities List.116
112. Insurance Issues and Superfund: Hearing before the Comm. on Environment and Public
Works, United States Senate, 99th Cong. 13 (1985) (testimony of Assistant Attorney General Henry
Habicht); see generally Alfred R. Light, “Minimum Contacts” Under CERCLA: Joint and Several
“Generator” Liability and the Fifth Amendment, 24 TOXICS L. REP. (BNA) 545, 547–49 (2009)
(detailing the government’s view of joint and several liability in the 1985 Superfund Amendments and
Reauthorization Act (SARA) hearings).
113. Superfund Improvements Act of 1985: Hearings Before the Comm. on the Judiciary, United
States Senate, 99th Cong. 71 (1985) (statement of Henry Habicht).
114. United States v. Burlington N. & Santa Fe Ry. Co., 520 F.3d 918, 937 (9th Cir. 2008), rev’d
on other grounds, Burlington N. & Santa Fe Ry. Co. v. United States, 129 S. Ct. 1870 (2009).
115. See generally Light, supra note 112, at 546, 548–50 (exploring this requirement through a
CERCLA hypothetical based on the facts regarding Shell Oil Co. in Burlington Northern).
116. 42 U.S.C. § 9607(o) (2006), added to CERCLA by the Small Business Liability Relief and
Brownfields Revitalization Act, Pub. L. 107-118, 115 Stat. 2356 (2002).
EXAMINING CONTRACTUAL MODELS FOR TRANSFERRING
ENVIRONMENTAL LIABILITY: HOW THEY WORK AND
WHERE THEY ARE HEADED
Kristin L. Hines*
TABLE OF CONTENTS
Introduction ............................................................................................... 395
I. Background to Private Indemnification Agreements ............................ 401
A. CERCLA’s Liability Scheme .......................................................... 401
B. Insured Fixed-Price Cleanups .......................................................... 405
II. Case Studies......................................................................................... 413
A. New York Brownfield Redevelopment ........................................... 414
B. California Superfund Site ................................................................ 415
III. Legal Implications & Projected Growth............................................. 416
Conclusion................................................................................................. 418
INTRODUCTION
Seller (S) wants out. S wishes to sell his property to a prospective
buyer (B), but S’s property is contaminated with hazardous waste. Under
federal hazardous waste law, S is liable for the costs of cleaning up the
property to regulatory standards. S is also liable for any damage this
contamination caused to third parties, neighboring properties, or natural
resources—in the past as well as in the future.1
Even though B has not participated in any of the activities causing the
contamination, B would assume the same liability S presently has if B buys
the property without performing due diligence on its environmental
condition. If the government has not yet brought an enforcement action
against S to force him to remediate the property or to recover its costs for
cleaning it, S may be disinclined to transfer the property for fear B will
* J.D. candidate, Vermont Law School, 2010; M.A., University of Minnesota, 2007; B.A.,
Bates College, 2002.
1. Based on liability under the Comprehensive Environmental Response, Compensation, and
Liability Act, 42 U.S.C. §§ 9601–9675 (2006).
396
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
discover and report the contamination. Consequently, the property remains
as is, untreated and potentially causing environmental and public health
harm. At best, it is not put into productive use as it might have been if sold
to B.
Here is where the private liability buyout firm (LBF) comes in. The
firm assesses the site and bids to contractually assume S’s liability for a
one-time-payable, fixed price. The firm estimates the cost to remediate the
site to the proper regulatory-use standards and monitor and maintain it
there in perpetuity. The LBF then adds a premium based on the extent of
the exposure it would be assuming, especially into the future. If utilizing an
insurance-backed contract, the firm would simultaneously negotiate with an
insurer to purchase site-specific environmental insurance policies to cover
any cleanup cost overruns or future expenses associated with the site, thus
limiting the firm’s own financial exposure.
Depending on S’s needs and desire to sell the property, transferring his
environmental liability to the firm may be a highly attractive option. If S
decides to “cash out,” the firm structures a contract between it and S that
indemnifies S from future actions by the government and third parties. The
firm is motivated by a variety of financial incentives, discussed in detail
below. In practice, these environmental risk transfers have been productive
transactions for the parties. They enable S to more or less “walk away”
and give B peace of mind while facilitating the re-development of
contaminated properties.2
*
Environmental liabilities can arise through remediation obligations,
environmental law violations, or litigation.3 These liabilities may have a
deleterious effect on the holder by complicating or impeding property
conveyances, or by incurring additional costs through noncompliance
penalties, causing reputational damage to the business, or creating a drag on
earnings if remediation costs are charged incrementally against net income
2. The author would like to note that this paper is primarily descriptive in nature and does not
purport to provide an objective quantitative assessment of the remediation industry’s environmental
results to date. Because these contracts are private, the available data is largely provided by the industry
itself. Further, insufficient time has elapsed to assess the long-term performance of private cleanups
against those administered by the Environmental Protection Agency (EPA). This paper, instead,
explores when fixed-price contracts may be appropriate and what makes them successful, including the
particular statutory framework that first enabled them.
3. Michael O. Hill & C. Gregory Rogers, Using Insured Fixed-Price Cleanups to Respond to
New Accounting Standards, Gain Tax Savings, and Lower Cleanup Costs While Increasing Cost
Certainty, 202 BUREAU OF NAT’L AFFAIRS, ENVTL. DUE DILIGENCE GUIDE, 231:2065, 2070 (Nov.
2008).
2009]
Examining Contractual Models for Transferring Liability
397
on a company’s balance sheets.4 Although environmental insurance
policies are now available through a number of insurance companies,5
environmental liability transfers (ELTs) may still be the best strategy for
managing these risks in certain types of situations. For reasons discussed
below, this Note focuses on liability transfers used to remediate
contaminated sites. Significantly, however, other applications for ELTs are
gaining prominence as well.6 In staying with this purpose, this Note
examines the mechanism used most successfully to transfer liability in the
context of site cleanups: the insured fixed-price cleanup (IFC) contract.7
The concept of using insured fixed-price contracts to transfer
environmental liability from the party financially responsible for the
contamination8 to an LBF was pioneered in the late 1990s by private
remediation companies.9 The first IFC debuted in 1999 when a private
liability buyout firm assumed the liability of hundreds of potentially
responsible parties (PRPs) who had contributed hazardous waste to Maine’s
largest Superfund site, the Portland-Bangor Waste Oil Site Facility, and
were being held financially responsible for the cleanup costs.10 Like S in
the introductory hypothetical, these PRPs sought relief from ongoing
obligations and elected to pay a one-time fee for a full release of their
liability.11 This basic model, applicable in a variety of scenarios, has
4. Id. at 2066.
5. Id. at 2066–67. Insurance companies added pollution exclusions to their standard
commercial general liability insurance policies in response to the new statutory environmental liabilities
created by CERCLA’s enactment in 1980. As a result, pollution liabilities were generally not insured
against until the 1990s. Id. at 2066.
6. Most notably, accounting requirements now mandate disclosing certain corporate
liabilities. As discussed in more detail below, this may create a market for using ELTs as a legal and
responsible way to remove environmental liabilities from corporate balance sheets. See id. at 2069–70.
7. IFCs, or variants thereof, are also referred to as guaranteed fixed-price remediation
(GFPR), performance-based contracts (PBCs), and liability buyouts (LBOs). Id. at 2065 n.2.
8. Transferors could include potentially responsible parties (PRPs) (if the liability arose from
an enforcement action under CERCLA) or an owner wishing to sell a contaminated property, among
others. See 42 U.S.C. § 9607(a)(1)–(4) (2006) (enumerating the parties that could be held liable for
property contamination).
9. The Travelers Research Corporation (TRC), a publicly held company, was the first to
register a proprietary IFC model, “The Exit Strategy®,” in 1998. INTERNATIONAL DIRECTORY OF
COMPANY HISTORIES, (Jay P. Pederson & Taylor Grant eds., 32d ed. 2000) (citing section entitled “TRC
Companies, Inc.—Company History”), available at http://www.fundinguniverse.com/companyhistories/TRC-Companies-Inc-Company-History.html.
10. See John J. Fialka, Maine Experiment May Point the Way to Ending Tangle of Litigation
Around U.S. Superfund Law, WALL ST. J., Apr. 29, 1998, at A28.
11. See Michael O. Hill, A Tale of Two Sites: How Insured Fixed-Price Cleanups Expedite
Protections, Reduce Costs, and Help the EPA, the SEC, and the Public, 18 NAT’L ENVTL.
ENFORCEMENT J. 3, 3 (Sept. 2003), available at http://www.naag.org/assets/files/pdf/neej-2003september.pdf.
398
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
become the basis for the liability transfer industry. Though initially
considered a “radical experiment,”12 ELTs have become well-accepted
devices used to manage environmental liability risk and allocation.13 A
number of remediation firms now compete for these contracts.14
Thus, the ELT industry has created a valuable niche market in less than
ten years.15 Starting in 2008, however, the economic climate has changed
significantly. Indeed, much ink has been spilled over the role of debt swaps
in what has been termed a global financial “meltdown,”16 as well as the
ensuing government bailouts of key players in several industries, including
insurers that underwrite fixed-price cleanup contracts.17 While still
uncertain, the negative ramifications of the current economic crisis on the
ELT industry thus far appear to be contained to the potential loss of some of
the insurers that presently provide insurance products used to secure fixedprice cleanup contracts.18 As is explored in a later section of this Note,
depressed economic conditions may actually promote increased use of
ELTs because of their potential utility in bankruptcies as well as in
12. TRC Companies, Inc., Exit Strategy Program, www.trcsolutions.com (follow “Services”
hyperlink; then “Environmental” hyperlink; then “Real Estate” hyperlink; then “Exit Strategy Program”
hyperlink) (last visited Dec. 23, 2009).
13. Joshua A. Bloom, Environmental-Liability Buyouts: How to Know When It’s the Real
Thing, 20 NAT. RESOURCES & ENV’T 37, 37 (2006) (“Years ago, this then-novel and fairly unknown
[ELT] concept was viewed with suspicion. Today, however, under the right conditions and structured to
truly benefit all parties, these transactions have an important and secure place . . . .”).
14. Nationwide, there are approximately twenty to thirty firms that offer a version of these
services, including the aforementioned TRC; CH2M Hill Ltd.; Dames & Moore, Inc.; EMCON;
Environmental Elements; Environmental Resources Management; Harding Lawson Associates Group,
Inc.; ICF Kaiser International, Inc.; IDM International; IT Group; Roy F. Weston; Safety-Kleen Corp.;
Thermo Terra Tech; Thermo Retec; Waste Management, Inc. INTERNATIONAL DIRECTORY OF COMPANY
HISTORIES, supra note 9. However, only six or seven of these firms offer a full environmental liability
transfer using a more secure version of an IFC.
15. TRC’s annual revenue, for example, was approximately $450 million per year as of 2007.
Eliminating Toxic Waste Liabilities—A General Counsel’s Dream, METRO. CORPORATE COUNSEL, Dec.
2007, at 38, available at http://www.metrocorpcounsel.com/pdf/2007/December/38.pdf [hereinafter
Eliminating Toxic Waste Liabilities].
16. See, e.g., Louise Story, Wall Street Report Tries to Dissect Financial Meltdown, N.Y.
TIMES, Aug. 7, 2008, available at http://www.nytimes.com/2008/08/07/business/07report.html.
17. American International Group (AIG) is a major provider of insurance policies for IFCs. In
September of 2009, the Government Accountability Office (GAO) reported that the federal
government’s bailout of AIG, totaling approximately $182 billion as of July 2009, appears to have
stabilized the insurance giant at this time. The federal government now owns approximately eighty
percent of AIG. American International Group Inc., N.Y. TIMES, Business, Oct. 31, 2009, available at
http://topics.nytimes.com/top/news/business/companies/american_international_group/index.html.
While the ELT industry does not appear imperiled on the whole, the ramifications for AIG as a future
underwriter of IFCs is unclear.
18. Other insurers that are able to offer comparable products are not facing the current
difficulties of AIG. See generally Edmund L. Andrews & Peter Baker, At A.I.G., Huge Bonuses After
$170 Billion Bailout, N.Y. TIMES, Mar. 15, 2009, at A1 (outlining the financial crisis as applies to AIG).
2009]
Examining Contractual Models for Transferring Liability
399
facilitating compliance with accounting standards that require disclosing
environmental liabilities.19
So far, when appropriate and done correctly, ELTs have been
considered win-win solutions.20 Foremost, ELTs bring cost-savings: IFCs
are cheaper than time-and-materials cleanups, as demonstrated by a recent
study by the U.S. Army showing that fixed-price contracts had an average
savings of twenty-two percent below estimates submitted at the project’s
outset.21 Moreover, the ability to transfer environmental liabilities adds
value by facilitating the redevelopment of contaminated properties, often
benefiting the community as well as the private parties involved.22 To the
extent that ELTs avoid cleanups in which the government bears the initial
response cost, the public also benefits by not having to pay for the cleanup
out of taxpayer dollars in the event that the government is unable to recover
its expenditures or the site is abandoned in bankruptcy. ELTs also eliminate
the transactional costs associated with litigating environmental enforcement
actions,23 conserving money spent on legal resources for both the
government and private parties. Finally, IFCs are significantly faster than
traditional cleanups.24 The benefits and track record to date explain why
practitioners predict that ELTs will be increasingly used in the future.25
19. See Hill & Rogers, supra note 3, at 2065–66 (“[R]ecent accounting pronouncements that
initially will take effect Dec. 15, 2008, require certain environmental liabilities to be recorded at their
market value . . . and proposed disclosure rules would require companies to provide detailed quantitative
and qualitative information about their environmental obligations . . . . [B]ecause the risk transfer is so
complete, IFCs . . . provide a viable way to remove recognized environmental liabilities from the
balance sheet.”).
20. See Bloom, supra note 13, at 42 (concluding that “all benefit” when liability buyouts are
appropriate and conducted correctly).
21. U.S. ARMY ENVTL. CTR., TRACKING PERFORMANCE ON THE ARMY’S PERFORMANCEBASED CONTRACTS, at 4 (May 16, 2006), available at http://aec.army.mil/usaec/cleanup/pba00.html
(follow “Tracking Performance” hyperlink) (showing cost avoidance ranging from 21.3% to 33.8%
percent). According to Hill and Rogers, this approximate twenty percent margin includes the cost of
adding insurance. Hill & Rogers, supra note 3, at 2066.
22. See Bloom, supra note 13 (“Under the right circumstances, the environmental-liability
assumption transaction can fix what had been to that point an uncertain contingent cost, can fuel a
transaction that would not have occurred but for the liability buyout, can free up funds otherwise tied up
as reserve, and can reestablish human capital to once again pay attention to the company’s core
business.”).
23. See Hill, supra note 11, at 7 (citing a study by the GAO estimating that PRPs spend half of
what the total cleanup costs again in litigation and transaction costs, whereas IFC can be completed
without any litigation and minimal legal fees).
24. See id. (citing an estimate by the GAO that the EPA takes an average of seventeen to
nineteen years to list and cleanup a Superfund site as compared to nineteen months, the time it took to
complete the cleanup of the first Superfund site to use an IFC, the Portland-Bangor Waste Oil Facility,
which was completed with no litigation and forty percent below estimated costs).
25. See ARTHUR J. HARRINGTON, ENVIRONMENTAL LIABILITY AND THE EMERGING ART FORM
OF TRANSFERENCE: A COMPREHENSIVE GUIDE TO TRANSFERRING FINANCIAL RISK TO THIRD PARTIES
400
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
In examining the industry’s development and the projections for its
growth, this Note pays particular attention to the contractual mechanisms
and regulatory structure that have enabled these transfers’ success. Though
ELTs have several current applications briefly noted here, this paper
primarily considers ELTs in the context of the Comprehensive
Environmental Response, Compensation, and Liability Act (CERCLA)
liability,26 which remains the dominant analytic paradigm.27 This focus is
due, in part, to the fact that ELTs were contemplated by a provision in
CERCLA itself28 and later arose in direct response to the inefficiencies
produced by implementing CERCLA, as discussed in detail in Part II. State
hazardous waste laws also frequently mirror CERCLA’s liability structure,
providing an additional rationale for this approach. It is worthwhile to note,
however, that the ELT concept can be extended to encompass nonCERCLA applications as well. Several of these applications are mentioned
briefly in Part IV, though an in-depth discussion is outside the scope of this
paper.
Part III presents two case studies illustrating the basics of an IFC
transfer in different settings. Part IV discusses the legal status of completed
ELTs, an area flagged as potentially problematic by the literature at the
fledgling industry’s outset.29 This section concludes by speculating on what
has allowed the industry to achieve such success to date, as well as what
lessons may be learned, if any, for structuring future legislation to enable
similarly effective market solutions to what is both a private and, in some
ways, public problem.
WHEN BUYING & SELLING CONTAMINATED PROPERTY, at 98–99 (2006) (predicting that future
environmental liability exposures (for example, from natural resource damages, chemical activities
previously thought safe, or public demand for more financial transparency with contingent liabilities)
will provide strong incentives for hedging these risks through innovative liability transfer devices); see
also Hill & Rogers, supra note 3, at 2073 (suggesting several reasons why “IFCs should be poised for a
significant increase in popularity”).
26. 42 U.S.C. §§ 9601–9675 (2006).
27. See generally Matthew J. Lawlor, Comment, Super Settlements for Superfund: A New
Paradigm for Voluntary Settlement?, 27 B.C. ENVTL. AFF. L. REV. 123, 124 (1999) (providing an
excellent discussion of the emergence of the liability transfer concept in anticipation of concluding the
first such IFC, or “super settlement,” in 1999); see also Kenneth F. Gray, “Super Settlements”: Early
Release for all PRPs at Multiparty Superfund Sites?, 12 NAT. RESOURCES & ENV’T 298, 298 (1998)
(outlining the general structure of a “super settlement,” an early term for IFCs).
28. 42 U.S.C. § 9607(e)(1) (2006) (“Nothing in this subsection shall bar any agreement to
insure, hold harmless, or indemnify a party to such agreement for any liability under this section.”).
29. See Lawlor, supra note 27, at 151–58 (questioning whether liability transfers will withstand
judicial scrutiny, as well as whether the federal-state relationship between the EPA and stateadministered site cleanups will prove problematic).
2009]
Examining Contractual Models for Transferring Liability
401
I. BACKGROUND TO PRIVATE INDEMNIFICATION AGREEMENTS
By the mid-1990s, dissatisfaction with remediation efforts under
CERCLA created a potentially lucrative market opening for private
remediation companies.30 Congress’s failure in 1995 to reauthorize funding
for a critical piece of CERCLA’s liability scheme31 opened the door for
private-sector innovation able to realize CERCLA’s public-interest goals in
the process of turning a profit. The ELT industry sprang up to capitalize on
inefficiencies produced by CERCLA’s implementation and, in doing so,
created a mechanism for structuring transfers that continues to find new
applications today.
A. CERCLA’s Liability Scheme
Congress enacted CERCLA in 1980 for the principle purpose of forcing
cleanup of leaking hazardous waste sites, in large part as a response to
growing public concern over a number of highly publicized hazardous
waste exposures.32 CERCLA operates under the “polluter pays” principle
by casting a wide net of liability around persons who have contributed to
the release of hazardous substances into the environment. This net ensnares
not only hazardous substance generators and transporters, but also the
owners and operators of property used to treat, store, or dispose of
hazardous substances.33 Under section 107(a), these potentially responsible
parties are liable for all response and remediation costs associated with
removing hazardous substances from a site until it achieves compliance
with a certain regulatory-use standard, thereby “closing” the case.34
CERCLA’s section 107 imposes strict liability on identified PRPs,35 and
federal courts have held PRPs jointly and severally liable unless the harm
caused by the release of the hazardous substance is divisible and the parties
30. Id. at 125.
31. Superfund Not Funded, Congressional Investigator Reports, ENVTL. NEWS SERV., Feb. 23,
2004, http://www.ens-newswire.com/ens/feb2004/2004-02-23-11.asp.
32. The 1978 Love Canal emergency, which required that an entire neighborhood in New York
evacuate their homes upon the discovery of underground hazardous waste contamination, is the iconic,
but by no means isolated, incident.
33. See § 9607(a)(1)–(4).
34. See id. (a)–(c). PRPs are also liable for damages to natural resources, though few actions
have been brought for natural resource damages to date. But see Kathleen Chandler Schmid, The
Depletion of the Superfund and Natural Resource Damages, 16 N.Y.U. ENVTL. L.J. 483, 508, 510
(2008) (arguing that while federal funds are still necessary for most states to bring natural resource
damage claims, some more affluent states have begun pursuing such claims on their own).
35. See § 9607(a).
402
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
can show a reasonable basis for apportioning damages.36 The result is
draconian but effective from the view of the government and the public: at
least one party is on the hook for the cleanup costs, either by financing the
remediation directly or reimbursing the government for its costs. The
paying party or parties can then bring “contribution actions” against other
PRPs for their individual shares, thus relieving the government of having to
join all PRPs in the suit or prove their individual contributions to apportion
damages.37 In the event that liable PRPs could not be identified or were
financially insolvent, orphaned sites or contributory shares were to be
covered by CERLCA’s “Superfund,” a trust fund that was created from
taxes on the oil and chemical industries but, as noted above, never
adequately funded.
Two pieces of legislation enacted post-CERCLA significantly evolved
its liability regime. To mitigate the harshness of strict liability, the
Superfund Amendments and Reauthorization Act of 1986 (SARA) created
an “innocent landowners” defense. Under this provision, “owners,” as
defined by CERCLA’s section 107(a), are exempted from liability if they
either involuntarily acquired contaminated property, or voluntarily acquired
property without “reason to know” that hazardous substances had been
disposed of on site.38
Then, in 2002, the Small Business Liability Relief and Brownfields
Revitalization Act (Brownfields Amendments) widened the exclusions by
creating a “bona fide prospective purchaser” defense.39 This defense
protects owners of contaminated properties from liability if all disposal
happened prior to their purchasing the land; if the owner neither contributed
to the contamination nor is affiliated with any party who is potentially liable
for the government’s cleanup response costs; and if the purchaser
36. See, e.g., United States v. Monsanto Co., 858 F.2d 160, 166–71 (4th Cir. 1988) (holding
that defendants who fell within one of the four PRP classes were strictly liable once the requisite nexus
with the waste site was established, and that joint and several liability is appropriate under CERLCA
when the harm is “indivisible”); United States v. Alcan Aluminum Corp., 964 F.2d 252, 268–69 (3d Cir.
1992) (inferring from CERCLA’s strict liability scheme that holding known polluters jointly and
severally liable is appropriate under the purpose of the statute).
37. The Superfund Amendments and Reauthorization Act of 1986 (SARA) amended CERCLA
to provide an explicit right to contribution; prior to SARA, courts had recognized an implied right under
CERCLA. See Pub. L. No. 99–499, § 106, 100 Stat. 1648 (1999).
38. SARA affected this change by expanding section 107(b)(3)’s third-party exemption
through redefining the term “contractual relationship” under section 101(35) rather than altering the
definition of “owner” under section 107(a). See Pub. L. No. 99–499, 100 Stat. 1616(35)(A) (1999).
39. The Small Business Liability Relief and Brownfields Revitalization Act, Pub. L. No. 107118, § 222 (2002).
2009]
Examining Contractual Models for Transferring Liability
403
performed due diligence prior to the sale.40 Both amendments sought to
facilitate the redevelopment of contaminated properties, to which
CERCLA’s strict liability for innocent landowners had acted as a powerful
deterrent.
Settlements were envisioned as a key element in CERCLA’s remedial
scheme, both to expedite cleanup and avoid expensive, protracted
litigation.41 CERCLA’s section 122 provides a number of devices to
promote settlement and reduce transaction costs, including: de minimis
settlements (expedited settlements for small-volume waste contributors);
nonbinding, preliminary allocations of responsibility issued by the EPA;
and covenants protecting PRPs who settle from suit by the federal
government in any future action related to the site.42 CERCLA’s section
113(f)(2) on contribution provides an important complement to section
122’s settlement provisions by absolving parties who settle their liability
with the federal or state government from being held liable for contribution
claims from other PRPs on the same matter.43 These provisions reflect the
emphasis placed on CERCLA’s settlement mechanism as the preferred
means of resolving liability and recovering or allocating cleanup costs.
Despite the hopes hung on CERCLA’s imposition of strict liability and
the continued need for cleanup of hazardous waste sites, the ambitions that
drove CERCLA’s creation and repeated amendment have not come to
fruition. Separate from criticisms of CERCLA’s design are the realities of
its performance: high cleanup costs,44 high transactional costs,45 and
extremely slow progress to site closure.46 Chronic under-funding continues
40. The statute requires buyers to have made “all appropriate inquiry” within the meaning of
CERCLA’s section 101(35). 42 U.S.C. § 9601(35)(B) (2006).
41. See id. § 9622(a) (directing settlement agreements “[w]henever practicable and within the
public interest”).
42. ROGER W. FINDLEY & DANIEL A. FARBER, ENVIRONMENTAL LAW IN A NUTSHELL 202 (6th
ed. 2004).
43. 42 U.S.C. § 9613(f)(2) (2006) (“A person who has resolved its liability to the United States
or a State in an administrative or judicially approved settlement shall not be liable for claims for
contribution regarding matters addressed in the settlement. Such settlement does not discharge any of
the other potentially liable persons unless its terms so provide, but it reduces the potential liability of the
others by the amount of the settlement.”).
44. The average site cleanup cost was estimated at between $15.38 million and $18.25 million
in 1992 dollars. See ZYGMUNT J.B. PLATER ET AL., ENVIRONMENTAL LAW AND POLICY: NATURE, LAW
AND SOCIETY 852 (1998).
45. Id. at 855 (relying on a 1992 study by the Rand Corporation, which estimated that legal
transaction costs accounted for approximately twenty percent of total private party Superfund
expenditures between 1984 and 1989); see also Hill, supra note 11, at 7 (citing a 1994 GAO study
estimating that PRPs spend as much as one dollar in litigation and transaction costs for every two dollars
they spend on the actual cleanup).
46. See GAO, SUPERFUND, TIMES TO COMPLETE SITE LISTING AND CLEANUP, GAO/RCED98-74 (Feb. 4, 1998), available at www.gao.gov/archive/1998/rc98074t.pdf.
404
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
to cripple CERCLA’s effectiveness, with no end in sight: CERCLA’s
dedicated taxes from oil and gas expired in 1994 and reauthorization has
lapsed since. Though appropriations from the federal budget have been
used to fund the Superfund, this substitution falls short on principle and
money alike.47 Relying on taxpayer monies not only defeats the “polluter
pays” principle, but has proved insufficient to cover the expenditures
needed to comply with remediation schedules for sites currently on the
National Priorities List (NPL).48 Properties that should be added to the NPL
are not, and recent federal practice has been to decline to pursue site
remediation where financially solvent PRPs are not available to assume the
cost.49 The result is that many contaminated sites remain untreated. In
addition to precluding redevelopment and the opportunity to generate
positive community benefits, this stagnation has obvious public health and
environmental ramifications.
Apart from the issue of CERCLA’s implementation, CERCLA’s
structure is itself paradoxically problematic and enabling. On the one hand,
CERCLA’s liability scheme inherently creates an incentive for PRPs to sue
other PRPs in order to spread the costs. Historically, Superfund cleanups
lagged as lawsuit after lawsuit was filed, taking 9.4 years on average to
achieve site closure after being listed on the NPL.50 This inefficiency is a
byproduct of CERCLA’s design, and it ultimately created the opportunity
for private entrepreneurship to find a market solution to costly and
intractable lawsuits.
At the same time, however, CERCLA expressly allowed for this
possibility in its indemnification provision.51 Section 107(e)(1) states that
“[n]othing in this subsection shall bar any agreement to insure, hold
harmless, or indemnify a party to such agreement for any liability under this
section.”52 Indeed, by the 1990s private companies were beginning to
experiment with new liability arrangements that were compatible with the
47. See Schmid, supra note 34, at 511–12 (examining Superfund’s budget statistics and federal
appropriations from 1995 to 2004 to show that discretionary appropriations have been less than the
revenues historically generated through the pre-1994 taxes).
48. Id. at 512 (stating that $1.22 billion was allocated to the Superfund in 2007 and that costs
for current NPL sites range from between $1.4 and $1.7 billion annually).
49. Id. at 512–13 (citing NAT’L ADVISORY COUNCIL FOR ENVTL POLICY & TECH., SUPERFUND
FINAL REPORT 21 (2004)).
50. See GAO, SUPERFUND, TIMES TO COMPLETE, supra note 46.
51. 42 U.S.C. § 9607(e)(1) (2006).
52. Id. (referring to the beginning of subpart (e)(1), which establishes the government’s right to
pursue the owner of a contaminated site despite any private indemnification agreement). See Hill &
Rogers, supra note 3, at 2068 (“Thus, while an IFC generally will not legally bar future government
suits, CERCLA expressly allows private party indemnities such as those provided by an IFC, and, to
date, none of these indemnities appears to have failed.”).
2009]
Examining Contractual Models for Transferring Liability
405
existing legal and regulatory structure, but which could facilitate the
settlements CERCLA had envisioned while streamlining the remediation
process. 53 The ELT concept thus hatched. Its realization, however, took
more than apparent legal feasibility. Market forces and developments in the
remediation and insurance industries converged to create an environment
ripe for the emergence of IFCs.54
B. Insured Fixed-Price Cleanups
As the remediation market matured throughout the 1980s and early
1990s, the companies performing remediation services on a time-andmaterials basis gradually gained enough experience to start confidently
estimating a site’s total cleanup costs at the outset.55 These companies’
ability to accurately assess costs, sometimes even prior to any regulatory
action, enabled the liability buyout firms to offer parties a fixed-fee cleanup
price.56 Notably, no transferor in an IFC to date has ever had to pay more
than the contract price.57 Some buyout firms also developed or acquired
remediation companies in order to better control the remediation process
and offer packaged services.58
ELTs have been especially effective in several situations, one of which
is the seller-buyer hypothetical in the introduction.59 The other primary
application is with enforcement actions for cleanup costs of a contaminated
site where the regulatory agency is required to deal with a large number of
low-volume contributing PRPs. In this instance, consolidating liability in
the buyout firm (to the extent that individual PRPs agree to transfer their
liability) saves litigation costs for all parties, significantly streamlines site
53. See Lawlor, supra note 27, at 125 (introducing the “super” settlement [ELT] concept in
1999 as a then-recent response to the gridlock of traditional Superfund negotiations).
54. HARRINGTON, supra note 25, at 81–83 (discussing contributory factors such as the
adoption of new regulatory use-based standards, market demand, and the emergence of remediation
insurance policies).
55. Id. at 81.
56. Id.
57. Hill & Rogers, supra note 3, at 2066 (basing assertion on professional familiarity with
hundreds of IFCs). Because IFCs are private transactions, not all data is publicly available.
58. For example, the liability buyout firm TRC has added subsidiaries to widen its service base
to include preparing remedial designs and performing environmental assessments, site cleanups, and
monitoring services. INTERNATIONAL DIRECTORY OF COMPANY HISTORIES, supra note 9.
59. See HARRINGTON, supra note 25, at 84–85 (outlining “ideal” risk scenarios where
quantifying the remediation risk allows it to be factored into a sales transaction between a property seller
and buyer). This quantification enables the parties to the sale to sidestep the problem of regulatory
delay in approving the remediation plan, which would typically assist the parties in “fixing” the liability
risk. A fixed-fee contract provided by a private remediation firm allows the parties either to modify the
sale price accordingly, or transfer the liability to the remediation firm itself.
406
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
administration, and often presents less risk of insolvency among the
parties.60 The fact that LBFs can obtain insurance policies to cover cleanup
cost overruns as well as their ongoing exposure—policies that are much
less likely to be issued to individual PRPs—adds a layer of financial
security.61 From the government’s perspective, these financial assurances
favor transferring the risk to the buyout firm.
It is important to note that IFCs are not a panacea, however. Insured,
fixed-price cleanups are best suited to sites where the expected cleanup will
approach or exceed approximately five million dollars, since below this
amount the margin is not great enough for the buyout firm to justify taking
the risk.62 Situations that present high transaction costs, such as a large
number of low-volume PRPs, also favor using an IFC for cost savings and
efficiency.63 The need for cost certainty in order to facilitate an outside
transaction may also be a factor, as in the case of a merger, acquisition, or
sale.64
Whether using an IFC is appropriate depends on the specifics of the site
and the needs of the parties. Alternatives include using a fixed cleanup
price without the insurance for smaller jobs, which firms may be willing to
do if they can “fold the risks into an existing umbrella or other portfoliotype of policy,” or are willing to self-insure in low-risk, low-cost cleanups.65
Another, more limited, option is to structure private-party indemnities,
usually used in a single seller-buyer situation.66 Neither of these
alternatives generates the same benefits that an IFC does, and a more
detailed analysis of the pros and cons of each is beyond the scope of this
paper. The key point is that IFCs are not appropriate for all sites, but when
they are well-suited and done properly, these devices have produced
excellent results.
1. Incentives Driving the Industry
ELTs are driven by business interests. The buyout firms are motivated
by profit opportunities, and the transferors by a desire to shed a risk that is
60. These advantages are most readily realized when up-front payments are required.
61. HARRINGTON, supra note 25, at 83.
62. Hill, supra note 11, at 8. Insurance can also be hard to obtain under this amount, and the
premium the firm would pay plus the costs of due diligence in assessing the site may make it financially
unviable. Hill & Rogers, supra note 3, at 2073.
63. Hill, supra note 11, at 8.
64. Id.
65. Hill & Rogers, supra note 3, at 2073.
66. Id.
2009]
Examining Contractual Models for Transferring Liability
407
adversely affecting their business in some way.67 From the transferor’s
perspective, property conveyance opportunities, bankruptcy and dissolution
needs, merger and acquisition transactions, and regulatory enforcement
exposure can all drive transfers and provide situations amenable to liability
buyouts.68 In the most comprehensive risk transfers, or those in which the
buyout firm assumes all of the transferor’s environmental liability for the
site,69 the LBF will fully indemnify the transferor (the PRP or the seller) for
all future liabilities associated with the site. This includes unforeseen
exposures, such as the discovery of unknown contaminants or damage to
other properties from migrating toxins. This indemnification—in addition
to the set cleanup cost—is the major benefit to using an insured fixed-price
contract.
Such “full-service” ELTs provide a regulatory or “case closure” result
rather than a limited scope of remediation work, and they include whatever
services are required to meet regulatory-use standards and maintain
compliance.70 In some cases, the cost for achieving regulatory closure has
been reduced significantly because of the willingness of environmental
agencies to adopt cleanup standards contingent on the site’s future use.
This change from the previous mandate, which required that all traces of
contamination be removed regardless of the site’s intended end-use and
apart from the risk actually posed to public health, has resulted in
significant savings in remediation costs and more brownfield
redevelopment.71
A maximum ELT includes “all remedial and long-term monitoring,
operation, and maintenance requirements, covering both known and
unknown pollution conditions existing at the time the buyout transaction is
executed.”72 Significantly, the buyout firm also indemnifies the transferor
from any liability in the event of a regulatory “re-opener,” an option the
regulatory agency retains if additional contaminates are discovered or the
standards of cleanup become more stringent after the regulatory sign-off.73
67. Bloom, supra note 13, at 38.
68. Id.
69. The transferor may choose to transfer all or part of its environmental liability, depending on
its needs and cash flow. Id. at 39.
70. Id. at 38.
71. HARRINGTON, supra note 25, at 81–83.
72. Bloom, supra note 13, at 39.
73. Though important from an environmental and public health standpoint, statutory
“reopener” clauses have proven problematic by scaring PRPs away from transferring their liability for
fear that such transfers will be futile despite the indemnification provided by the liability buyout firm.
Though the EPA sometimes issues “comfort letters” assuring PRPs of the unlikelihood of further
regulatory action, EPA still retains its right to “reopen” the case. Some state environmental agencies,
such as New Jersey’s Department of Environmental Protection, have recognized the beneficial function
408
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
As a technical matter, even an IFC, the most comprehensive transfer
mechanism, does not legally bar the government from pursuing the
transferor at a later date. In other words, “despite the contractor’s
indemnity and the insurer’s policy, the government retains the right to
pursue the owner” under CERLCA section 107(e)(1).74 However, reopeners
are rare. According to a 2003 survey, only twelve reopeners were found out
of 11,497 sites that had achieved “no further action” status.75 Consequently,
ELTs still offer the transferor a functional “walk-away,”76 thereby freeing it
to focus on its business activities.
From the perspective of the liability buyout firm, IFCs have the
potential to be lucrative deals. Most immediately, IFCs offer the firm a
chance to recover savings if it can achieve the remediation goals for less
than the contract’s fixed price.77 However, competition in the industry
keeps firms from padding their estimates, and the transferor is free to reject
bids that are too high. The structure of a true IFC acts as a check on moral
hazards. The firm deposits the client’s money into an escrow account, less
a small portion of working capital upfront, and is only reimbursed as the
remediation work is performed, and then only for reasonable and necessary
expenses as approved by the regulators in charge of the site
administration.78 If the firm’s cleanup costs exceed the escrow funds, the
firm’s insurer will cover spillover costs up to the limit of the insurance
policy that the firm purchased for the cleanup. However, if this occurs, the
firm makes no profit from the cleanup and jeopardizes the availability of
project coverage in the future.79
that ELTs serve and have accordingly started providing more satisfactory “no further action” assurances
to PRPs. See Dennis M. Toft & Todd W. Terhune, NJDEP Approves the Use of Risk Transfer Tools to
Facilitate
Brownfields
Transactions,
N.J.
LAWYER
MAG.,
available
at
http://wolffsamson.client.tagonline.com/news/articles/publish/article_48.shtml.
74. Hill & Rogers, supra note 3, at 2067–68.
75. Robert A. Simmons et al., Quantifying Long-term Environmental Regulatory Risk for
Brownfields: Are Reopeners Really an Issue?, 46 ENVTL. PLANNING & MGMT. 257, 266 (2003).
76. Bloom, supra note 13, at 37–38 (explaining that a PRP’s statutory liability “can never truly
be absolved,” but that through a series of contractual indemnities and insurance mechanisms, a
functional equivalent of a true “walk away” can be achieved).
77. HARRINGTON, supra note 25, at 84 (“Basically, if the remediation firm is able to realize
innovative techniques to obtain closure for a cost less than the fixed-fee contract, the benefit of the
contract price over the actual costs inures to the engineering firm as a reward for its innovative and
entrepreneurial activity.”).
78. Bloom, supra note 13, at 41 (distinguishing these guaranteed contracts held in trust or by
an escrow agent from “self-insured retention” policies where the client pays the remediation contractor
directly). In the latter scenario, the contractor uses the funds to pay its own insurance premium to secure
spill-over coverage above $x amount in expenditures, takes its profit cut, and then performs the services
until it runs out of money. This structure creates no incentive for the contractor to perform remediation
services quickly or efficiently since it has already paid itself and the insurer will absorb spillover.
79. HARRINGTON, supra note 25, at 83.
2009]
Examining Contractual Models for Transferring Liability
409
IFCs can be profitable in other ways as well. If the firm acquires title
to the contaminated property, it may be able to resell the property after
cleanup for substantially more than its pre-remediation value. Firms may
also receive tax credits for redeveloping brownfield sites, which may be
transferrable depending on the project’s home-state laws.80 Though not all
projects offer these additional profit opportunities, the tax credits in
particular incentivize redevelopment.
That IFCs are profit-motivated does not mean that they are any less
effective, legal, or productive for the parties involved.81 The regulatory
sign-off certifying that the site meets the proper cleanup standard serves as
quality assurance—as does the fact that the firm remains on the line for any
future obligations associated with the site. Moreover, because any profit is
withheld until certification, the buyout firm has an incentive to conclude the
remediation swiftly. This result is desirable from the perspective of the
state or federal agency involved in administering the site, as well as from a
redevelopment angle. When structured correctly, ELTs thus have the rare
potential to benefit all parties involved.
2. Contract Provisions
ELT contracts can vary greatly because they are tailored to the specific
needs of the particular transaction and site. Many IFCs share some general
features, however. The following chart summarizes some typical contract
provisions, as outlined by attorney Arthur J. Harrington in a recent
practitioner’s guide to risk transfers in the context of real estate
transactions.82
80. U.S. ENVTL. PROT. AGENCY, EPA 560-F-08-312, BROWNFIELDS TAX INCENTIVES
GUIDELINES (Nov. 2008), available at http://www.epa.gov/brownfields/tax/index.htm.
81. See Hill, supra note 11, at 6 (“IFCs are not only legal—as evinced by the court’s entry of
the PBWO [Portland Bangor Waste Oil Facility] decree—but they promote existing EPA and SEC
policies and goals.”).
82. See generally HARRINGTON, supra note 25, at 87–93 (providing detailed explanations and
examples of key contract provisions).
410
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
Sample Contract Provisions83
Provision
“Baseline
Remediation
Definition”
Function
Defines the desired
remediation result
sufficient
to
support the end use
for the site in
compliance with
regulatory-use
standards.
“Definition
of Specifies
the
Case Closure”
regulatory
event
that will conclude
the
active
remediation work.
“Definition
‘Preexisting
Pollution
Conditions’”
83. Id.
of Defines the scope
of
contaminants
covered.
Sample language
“‘Baseline Remediation’ means
. . . perform remediation of
preexisting pollution conditions
under applicable environmental
law to the extent necessary to
support commercial uses with
exposure scenarios equivalent to
those of a [e.g.] shopping center
. . . installed and maintained to
act as the functional equivalent
of a barrier to human exposure
. . . .”
“‘Case
Closure’
means
obtaining a case closure letter
from the [state regulatory
agency] in accordance with and
defined in [cite the case closure
regulatory requirement] specific
to
contamination
release
incidents, confirming that no
further action is necessary for
achieving
baseline
site
conditions.”
“‘Preexisting
Pollution
Conditions’ means the pollution
conditions existing on the
effective date, including without
limitation, the effects of
continuing release or passive
migration . . .
and
without
regard
to
whether
such
conditions were known or
unknown to the remediation
contractor on the effective date.”
2009]
Examining Contractual Models for Transferring Liability
“Remediation
Guarantee
Contractor”
Guarantees
the
of contract’s
performance at a
fixed fee sufficient
to
obtain
site
closure regardless
of costs.
Defines the manner
“Early
Termination of the in
which
the
Contract”
contract can be
terminated prior to
project completion
and contemplates
potential remedies
for both client and
contractor.
Providing
“Indemnification
Obligations”
indemnity for the
benefit of the client
against specified
claims.
411
“Contractor guarantees that it
will fulfill all of its obligations
under this agreement as
necessary to achieve project
completion, regardless of the
expense, difficulty or time to
complete same, or the limits,
expiration, cancellation or lapse
of the policy.”
“The client should have the
ability
to
terminate
the
agreement for a material breach
and receive all payments made
but not expended to a third party
remediation account . . . .”
[Contractor agrees to indemnify
client for the following]: “b.
Any breach of the agreement by
the contractor . . . .
....
d. Any failure of the contractor
to perform its services covered
by the fixed-fee remediation
contract . . . .”
These sample provisions are merely skeletal. Other necessary
provisions include payment procedures, dispute resolution procedures
(usually mediation or arbitration), client obligations pertaining to site access
and payment, and security for the firm’s performance. Restrictive
covenants may also run with the property, such as maintaining an
impervious surface over contamination.84 If the regulatory sign-off required
the use of engineering controls such as a parking lot or building to provide a
physical cap, or institutional controls such as deed restrictions, the firm is
obligated to maintain these containment mechanisms in perpetuity.85
84. Id. at 88.
85. See generally Toft, supra note 73.
412
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
The firm’s obligations can thus be extensive and ongoing. As a result,
the liability buyout firm is required to provide additional financial
assurances prior to the regulator’s approval of the ELT.86 Though a
thorough discussion of the rules governing these financial assurances is
beyond the scope of this paper, it is significant to note that this is currently
the only regulated part of these transactions. The ELT itself is purely
contractual and is not regulated on either the state or federal level.
3. The Firm’s Environmental Insurance Policies
As previously noted, the buyout firm negotiates for a site-specific
environmental insurance policy from a major insurer as part of structuring
the IFC.87 The purpose of this policy is to collateralize the risk the buyout
firm assumes. When underwritten properly, an insurance policy provides
assurance that the buyout firm will have sufficient funds available to it for
the full performance of its contractual obligations.88 Ideally, the scope of
the coverage tracks the scope of liability transferred, but the two
transactions are independent.
Ideally, the insurance component is comprised of two types of policies:
a “cleanup cost cap” and a “pollution legal liability” policy.89 The cost cap
protects the remediation company if costs exceed their original estimation,
providing a financial backstop to the risk assessment. Typically, the firm
will insure the project for no less than a one hundred percent cost overrun.90
Thus, for example, if the firm calculates that the remediation will cost five
million dollars, it will purchase at least the same in insurance, thereby
ensuring it has available at least ten million, or double what it anticipates
needing. In a true result-oriented remediation contract, this “cost cap”
insurance also covers revisions to the work plan driven by regulatory
changes as required to reach case closure.91
86. Currently, the federal or state agency involved in the site administration must consent to the
LBF becoming a permittee or signatory to a consent order in order for the firm to contractually assume a
PRP’s obligations. Typically, these decrees contain specific financial assurance requirements to ensure
that the firm has sufficient funds available to it to fulfill the regulatory obligations attached to the
consent order. See Hill, supra note 11, at 7 (noting that the state, the EPA, or sometimes a court must
approve an IFC prior to its performance).
87. Four insurance companies typically underwrite these policies: AIG, XL, Zurich, and
Chubb. HARRINGTON, supra note 25, at 65.
88. Bloom, supra note 13, at 40.
89. The proprietary names for these products vary. “Cleanup cost caps” are also commonly
referred to as “remediation stop loss” policies, for example. See HARRINGTON, supra note 25, at 65.
90. Bloom, supra note 13, at 40.
91. Id.
2009]
Examining Contractual Models for Transferring Liability
413
Pollution legal liability policies, by contrast, are entirely independent of
the underlying remediation contract. Instead, the buyout firm purchases
insurance to cover other environmental liabilities associated with the site as
a whole.92 Coverage often extends to any cleanup activities outside the
scope of work necessary to achieve regulatory closure for the area of the
site under contract. These policies can act as “catch-alls” and may include
coverage for third-party cleanup, bodily injury, and property damage
claims, as well as for business interruption, liability for offsite disposal and
migrating pollutants, and for natural resource damages.93 Significantly,
pollution policies may also be written to insure against regulatory
reopeners, as discussed previously.94 This protection has greatly appealed
to transferors looking to shed their future exposure.
The combined coverage of both types of insurance policies has been
integral in facilitating the use of ELTs. These products protect both parties
from the likelihood that cost overruns will overwhelm the firm and thereby
jeopardize its ability to perform its contractual obligations. For the firm,
these policies sufficiently limit its financial risk so as to keep the incentive
structure behind IFCs intact. While the continued underwriting ability of at
least one major insurer remains somewhat uncertain,95 it appears
improbable that the environmental insurance industry will disappear. So
long as comparable products remain available, IFCs are likely to continue
apace.
II. CASE STUDIES
IFCs are increasingly used in a variety of settings and include clients as
big as the U.S. Army, the U.S. Air Force, and the Department of Defense.96
Popular transaction deals include risk transfers for real estate or brownfield
sites; Superfund or complex litigation cost recoveries; discontinued
industrial operations; and bankruptcies.97 A brief description of two recent
IFC projects illustrates a few of its different applications.
92. The coverage is contingent on the scope of the liability transferred to the LBF. See id.
(describing various aspects of a liability buy-out structure).
93. Id.
94. Id. at 41.
95. See Andrews, supra note 18.
96. See Hill & Rogers, supra note 3, at 2065–74. In 2007, one liability buyout firm alone had
98 active remediation sites, $520 million in aggregate clean-up obligations, and approximately $1.25
billion in insurance backing, estimated at about $6 billion in asset value. Eliminating Toxic Waste
Liabilities, supra note 15.
97. Eliminating Toxic Waste Liabilities, supra note 15.
414
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
A. New York Brownfield Redevelopment
In one of the biggest brownfield redevelopment projects in the country,
an LBF facilitated a real estate transaction between a public utility and a
prospective purchaser by assuming complete environmental liability for a
contaminated ten-acre, one billion-dollar piece of waterfront property in
Manhattan.98 The owner, a major public utility, had operated a steam and
power generating plant on the site for more than 150 years, resulting in
major contamination from coal and fuel oil, urban fill, and PCBs.99
Described as a “developer’s dream,” the prospective buyer was nevertheless
unable to accept the uncertain but substantial liability risks associated with
the site.100 Moreover, neither the owner nor the prospective buyer
anticipated being able to resolve the environmental liabilities swiftly
enough to secure regulatory approval in time to meet the developer’s
schedule.101
After assessing the site, the firm negotiated an insured fixed-price
cleanup with the utility that guaranteed compliance with the regulatory
standards for a mixed-use development.102 Covered by a cost-cap insurance
policy,103 the firm assumed all remediation costs, including those associated
with decommissioning the site, removing asbestos and lead paint, and
demolishing the plant.104 In this case, the case closure “event” concluding
the active remediation phase was defined in the contract as the regulatory
review and delivery of the “clean” property to the developer.105
The maintenance phase for this IFC includes indemnifying the
developer from all future pollution-liability exposures. To mitigate the
threat of a reopener to the developer, the firm negotiated a voluntary
consent order with the regulatory agencies whereby the agencies agreed to
pursue the firm, and not the developer, in the event additional remediation
98. Id.
99. TRC Exit Strategy Program, Site Solutions, Corporate Divestiture and Redevelopment of
Prime
Manhattan
Real
Estate
Made
Possible
by
TRC
Exit
Strategy,
http://www.trcexitstrategy.com/site_solutions_4a.html (last visited Dec. 21, 2009) [hereinafter
Corporate Divestiture].
100. Id.
101. Id.
102. Id.
103. Environmental Expert.com, US $50.8m Brownfield Remediation Contracts Signed,
BUSINESS
WIRE,
(Dec.
28,
2008),
http://www.environmentalexpert.com/resultEachPressRelease.aspx?cid=28518&codi=32621&lr=1&word=brownfield%2bremedia
tion.
104. Corporate Divestiture, supra note 99.
105. Id.
2009]
Examining Contractual Models for Transferring Liability
415
becomes necessary.106 Under its contract with the utility, the firm must
continue annual monitoring and perform any maintenance required for the
site to remain in regulatory compliance. By involving the local Community
Board and New York City Council, the firm also ensured that its closure
plans had municipal support.107 When finished, a massive brownfield in a
prime location will be converted into five million square feet of office,
residential, and retail space. This redevelopment demonstrates an IFC’s
utility when well suited to the site.
B. California Superfund Site
This particular IFC arose in response to CERCLA liability. In this
situation, adjoining landfills had managed the disposal of municipal solid
and hazardous wastes, chemicals, and oil for almost half a century.108 As
new housing developments eventually cropped up in the surrounding area,
contamination from the landfill site was discovered leaking into the
groundwater. Pursuant to state hazardous waste laws, the State of
California initiated an enforcement action against the landfill owners and
operators, as well as the businesses that had disposed of industrial waste at
the site.109 Prior to the buyout firm’s offer, progress on the site had been
deadlocked for ten years due to the complexity of the liability and the
number of lawsuits between the PRPs.110
Using an IFC, the buyout firm assumed the liability of all the PRPs and
was able to begin site cleanup. In this case, the most effective solution was
for the firm itself to obtain title to the property. This streamlined the State’s
site administration and gave the firm complete control over the remediation.
Within a year a physical barrier capping the landfill had been engineered,
eliminating further leakage.111
In conjunction with the State and
municipality, the firm continues to develop a plan to remediate the preexisting contamination and redevelop the site for a designated end-use.112
In this situation, preventing further leakage was the immediate
objective. A complete buyout through a statutory settlement procedure
enabled the firm to become the sole responsible party, which in turn gave it
106. Id.
107. Id.
108. TRC Companies, Inc., California Superfund Site, Solid & Hazardous Waste Landfill,
http://www.trcsolutions.com/content/project.aspx?csuid=96c1bcae-97b6-4b47-b7d7db415c9df2a1&pud=36930502-5b2f-4010-ae92-7cff9abf80e9 (last visited Dec. 21, 2009).
109. Id.
110. Id.
111. Id.
112. Id.
416
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
the flexibility necessary to implement an initial stopgap remediation plan.
Though every remediation is unique and requires structuring the contract
accordingly, these two projects exemplify the scenarios that most frequently
lend themselves to effective resolution through the use of IFCs.
III. LEGAL IMPLICATIONS & PROJECTED GROWTH
Now a decade old, ELTs appear to have glided over the speculative
legal bumps in the road projected at their outset.113 No IFC indemnity has
yet failed,114 and no legal challenges have arisen to date. Whether a court
would find a claim following a transfer of ownership valid would likely
depend on the firm’s ability to see the remediation project to completion.
However, as positive experiences with IFCs continue to accumulate,
regulatory agencies appear increasingly willing to work with buyout firms
to facilitate transfers.115 That the EPA and various state departments of
environmental protection have been willing to recognize existing IFCs
suggests that they are becoming an enduring part of the legal landscape, and
for good reason. Regulators may be more flexible in letting the industry
structure IFCs because they are gaining a solvent entity (the buyout firm)
where none may have existed before, and this entity voluntarily assumes the
liability for a contaminated site.
Environmental liability transfers are currently unregulated. Whether
they should be regulated and to what degree, if so, remain open questions.
Because the type and scope of the risk being transferred dictates the
contract price, the buyout firm has a strong incentive to uncover the
liabilities associated with the site in order to adequately gauge the firm’s
potential exposure. IFCs thus require transparency and comprehensive due
diligence.116 The nature of the transaction promotes a fair valuation of the
113. See Lawlor, supra note 27, at 150, 152 (questioning whether liability transfers will
withstand judicial scrutiny, and predicting that “in the end . . . the Super Settlement [ELT] concept is
likely to be viewed as merely an innovative and potentially very useful twist on [CERCLA’s] settlement
process”).
114. Hill & Rogers, supra note 3, at 2068.
115. See Toft, supra note 73, at 52; see also Dennis Harnish, From the State’s Perspective,
NAT’L ENVTL ENFORCEMENT J. 11 (Sept. 2003) (recommending that “our sister states and the U.S. EPA
consider such an [IFC] approach in an appropriate case”). Dennis Harnish is an Assistant Attorney
General in the State of Maine Attorney General’s Office representing Maine’s Department of
Environmental Protection. Working with buyout firms may mean, in part, taking affirmative steps prior
to the completion of the remediation, such as issuing “comfort” letters or other no-further-action
assurances to reassure the parties. Many PRPs may demand such assurance before considering an IFC a
viable option for managing their exposure.
116. See Toft, supra note 73, at 51.
2009]
Examining Contractual Models for Transferring Liability
417
liability,117 which is necessarily site-specific. The dynamic between players
in an IFC is fundamentally different than those involved in bundling and
reselling mortgage securities because the same party that performs the
valuation also incurs the risk. Thus, there is no incentive to improperly
valuate the liability, since the firm cannot pass the risk through to other
parties. However, while regulation may not be necessary to avoid moral
hazards, some measure of oversight may create a more uniform process for
structuring IFCs from state to state.
ELTs are poised for new applications, including managing liability for
natural resource damages as well as environmental contingencies on
corporate balance sheets.118 With regard to natural resource damages, PRPs
may be liable for contaminated riverbeds and groundwater, destroyed
habitat, species loss, and other harms to natural resources under CERCLA
or some state hazardous waste laws.119 Original to CERCLA,120 this cause
of action enables trustees such as federal and state governments and Indian
tribes to sue for such damages, which are determined by the amount
required to restore the natural resource to its previous state or replace it if
restoration proves impossible.121 Though this liability has existed since
CERCLA’s enactment, these damages have been historically
underutilized.122 There is some evidence that this trend is changing,
however.123 Like any other type of liability, an uptick in regulatory
enforcement actions creates a market. To the extent that natural resource
damages are increasingly used as remedies, the ELT industry will track this
growth.
The more prominent application, however, likely concerns the
maturation of accounting standards post-Enron.124 Though the 2002
Sarbanes-Oxley Act does not explicitly address corporate environmental
117. Hill & Rogers, supra note 3, at 2070 (“The fair value of a liability is the price that would
be paid to transfer the liability in an orderly transaction between market participants. . . . A quoted price
for the identical liability in an active market is the best evidence of fair value.”).
118. See HARRINGTON, supra note 25, at 97–99 (discussing emerging liabilities).
119. See Schmid, supra note 34, at 483 (contrasting natural resource damages to common law
remedies, which only require polluters to stop the release of the harmful substance and compensate the
public for its economic loss). Economic damages rarely capture the full spectrum of loss and are thus
usually inadequate to allow for restoration.
120. 42 U.S.C. § 9607(a)(4)(A) (2006).
121. See Schmid, supra note 34, at 484.
122. Id. at 519–20 (noting an increase in the number and award size of natural resource damage
claims after 1995).
123. HARRINGTON, supra note 25, at 97 (“States such as New Jersey are actually engaging
lawyers in the private sector to prosecute natural resource damage claims against responsible parties for
contamination and their revenue enhancement for such states.”).
124. See id. at 99 (stating that public desire for fiscal transparency will likely be a moving force
behind new forms of environmental liability risks); see also Bloom, supra note 13, at 38.
418
VERMONT JOURNAL OF ENVIRONMENTAL LAW
[Vol. 11
disclosures, practitioners now stress the importance of viewing these
disclosure obligations in light of the Enron-era scandals.125 Recent
enforcement actions by the Securities and Exchange Commission support
this view.126 Currently, certain environmental liabilities must be recorded at
their market value on corporate balance sheets,127 and proposed new
disclosure standards could expand this scope to include all environmental
contingencies.128 By allowing for fair market valuation and transfer, IFCs
could enable a company to comply with accounting requirements while
ridding its balance sheet of these liabilities. As accounting standards
evolve, IFCs may become an increasingly attractive, transparent way to
manage corporate environmental contingencies.
CONCLUSION
In the late 1990s those studying the concept of environmental liability
transfer anticipated that the practice would become increasingly popular.129
This prediction has indeed borne out. Increased use of IFCs in particular is
likely to continue because of the advantages they offer, so long as cost-cap
and pollution liability insurance policies remain available. Bringing in a
solvent party willing to assume the liability is too good a backstop for
regulators to pass up. IFCs will continue to thrive so long as they remain
the most efficient solution to a problem affecting business interests.
Other than watching how current economic forces shape new
applications, the apparent success of ELTs may be instructive in other ways.
Foremost, the industry’s rise invites further inquiry into what factors enable
125. See Mitchell F. Crusto, Endangered Green Reports: “Cumulative Materiality” in
Corporate Environmental Disclosure After Sarbanes-Oxley, 42 HARV. J. ON LEGIS. 483, 484 (2005)
(“Although Sarbanes-Oxley has not yet mandated that the U.S. Securities and Exchange Commission
. . . amend the environmental disclosure rules, the legal context within which they must now be read has
changed significantly.”) (citing, e.g., Andrew N. Davis & Stephen J. Humes, Environmental Disclosures
After Sarbanes-Oxley, PRAC. LAW, June 2004, at 19, 20); see also Philip E. Karmel, SEC Disclosure
Requirements for Environmental Liabilities and the Impact of the Sarbanes-Oxley Act, in PRACTICING
LAW INSTITUTE, REAL ESTATE LAW AND PRACTICE COURSE HANDBOOK SERIES, NEW SOLUTIONS TO
ENVIRONMENTAL PROBLEMS IN BUSINESS & REAL ESTATE DEALS 293, 298 (Practicing Law Institute
ed., 2003).
126. Hill & Rogers, supra note 3, at 2069 (noting that there have been three recent SEC
enforcement actions against public companies for alleged manipulation of environmental reserves in
their corporate accounting).
127. Id. at 2070 (referring to new merger and acquisition accounting rules beginning in 2009).
128. Id. (referencing the June 2008 Proposed Statement of Accounting Standards, Disclosure of
Certain Loss Contingencies).
129. See Lawlor, supra note 27, at 160 (“It is likely that the Super Settlement concept will be
put into widespread use across the United States as its advantages become better known.”).
2009]
Examining Contractual Models for Transferring Liability
419
market solutions to respond to space carved out by particular statutory
structures and regulatory flexibility. Is it desirable, or possible, to design
for these features in other legislative contexts? In the case of ELTs, while
CERCLA’s inefficiencies were not intended, section 107(e)(1) did expressly
provide for the private indemnification agreements that laid the legal
foundation for transfers. As the industry matures, both the contracts’
durability and the degree of satisfaction with the remediation outcomes will
undoubtedly continue to inform this discussion.
Whether the story of the ELT industry yields transferable lessons, such
as how to incentivize transparency or structure future legislation to realize
market solutions, remains to be seen. Nevertheless, ELTs, especially those
conducted through insured fixed-price cleanup contracts, provide an
interesting model for study and an effective practical tool. Creative and
ethical use of these devices will determine the bounds of their utility.