Amicus Curiae Brief — Shell Oil Co. v. United States (Nos. 07-1607, 07-1601)

\D No. AD) a RS as

Naene

IN THE i a Sa

Supreme Court of the United States —s

THE BURLINGTON NORTHERN AND SANTA FE RAILWAY

COMPANY, AND UNION PACIFIC RATI.ROAD COMPANY,

Petitioners,

Vv.

UNITED STATES OF AMERICA, AND DEPARTMENT OF

TOXIC SUBSTANCES CONTROL, STATE OF CATJFORNIA,

Respondents.

SHET.1.OIL COMPANY,

Petitioner,

Vv.

UNITED STATES OF AMERICA, et al.,

Respondents.

On Writ of Certiorari to the United States

Court of Appeals for the Ninth Circuit

BRIEF OF AMICUS CURIAE

ASSOCIATION OF AMERICAN RAILROADS

IN SUPPORT OF PETITIONERS

LOUIS P. WARCHOT CARTER G. PHII.1.1PS*

MICHAEL J. RUSH G. PAUL MOATES

ASSOCIATION OF SAMUEL I. GUTTER

AMERICAN RAILROADS ERIC A. SHUMSKY

LAW DEPARTMENT NAOMI SCHOENBAUM

50 F Street, N.W. STACEY L. WILSON

Washington, D.C. 20001 SIDLEY AUSTIN 1.1.P

(202) 639-2505 1501 K Street, N.W.

Washington, D.C. 20005

(202) 736-8000

Counsel for Amicus Curiae

November 24, 2008 * Counsel of Record

WILSON-EPES PRINTING Co.. INC. — (202) 789-0096 — WasninGron, D.C. 20002

both YIBRZO

3722720

<n ee

TABLE OF CONTENTS

Page

ARs Fe BF EE OO ocidiscsonsccsvsseseccenccosscees li

INTEREST OF AMICUS CURIAE ASSOCIA-

TION OF AMERICAN RAILROADGB................ 1

EERE NE ee 2

ii telinsiciniiinienninaiicsalbsinineneinipainssintsnnen 4

I. COMMON-LAW PRINCIPLES FAVOR

APPORTIONMENT IN CASES LIKE

Il. AN UNYIELDING APPLICATION OF

JOINT AND SEVERAL LIABILITY WILL

HAVE EXTRAORDINARY AND INEQUI-

TABLE EFFECTS.......... eineldceiadiabpebtdiaaiiniduitiy 15

ITT cisiaitniianennesitiniicdgibednindintnniiiiuabubsienincniee 24

(1)

‘3

TABLE OF AUTHORITIES

CASES ' Page

Amoco Oil Co. v. Borden, Inc., 889 F.2d 664

IIE Ere er Oren 17

Arthur Richards, Inc. v. 79th Fifth Avenue

Co., 441 N.E.2d 1114 (N.Y. 1982).............. 22

Cal. Orange Co. v. Riverside Portland Ce-

ment Co., 195 P. 694 (Cal. Ct. App.

ERI aeons Renee Oe SO ne ae ee 8, 10

Chipman v. Palmer, 77 N.Y. 51 (1879)......... 7,10

Clifford v. Atl. Cotton Mills, 15 N.E. 84

a a 22

Dent v. Beazer Materials & Sers., Inc., 993

F. Supp. 923 (D.S.C. 1995), affd, 156

F.3d GBS (46h Cir. 1906)... .ccccccccccccccccccccccces 10

Desert Palace v. Costa, 539 U.S. 90 (2003)... 12

Eckman v. Lehigh & Wilkes-Barre Coal

Co., 50 Pa. Super. 427 (1911)...................4. 7,10

Edgar v. Walker, 32 S.E. 582 (Ga. 1899)...... 21

Gallick v. Balt. & Ohio R.R., 372 U.S. 108

EERE ii ea tn nore Rec ee 12

Harley v. Merrill Brick Co., 48 N.W. 1000

A a dati isthe accion melanin 7,10

Hill v. Chappel Bros. of Mont., Inc., 18 P.2d

I lel asta 7

Jackson v. Virginia, 443 U.S. 307 (1979)..... 14

Kamb v. U.S. Coast Guard, 869 F. Supp.

Pe I i ccitiitencctinndntncsctiinidencennes 12

Leo Sheep Co. v. United States, 440 U.S.

SEILER tee eee a ee 18

Merchs.’ Mut. Ins. i v. Baring Bros. &

ee. I dctiictttdcetssccncsncecsceees 12

Michalic v. Cleveland Tankers, Inc., 364

Ee ee RE a 12

Ogden v. Lucas, 48 [Il]. 492 (1868) ................ 8

ili

TABLE OF AUTHORITIES -— continued

Page

Quaker State Minit-Lube, Inc. v. Fireman’s

Fund Ins. Co., 52 F.3d 1522 (10th Cir.

SET ccinsishicsithaciiernestebiiacedeiskililaiandiiad ibidiuiabadisiiiiisiemenene 17

Ralston v. United Verde Copper Co., 37

F.2d 180 (D. Ariz. 1929), affd, 46 F.2d 1

EE AR nn 10

Reeves v. Sanderson Plumbing Prods., Inc.,

I a esnenene 14

Sellick v. Hall, 47 Conn. 260 (1879) ............. 7

Thomas v. Ohio Coal Co., 199 Ill. App. 50

Sit nidiliatiidienid-ieeiacadeannisaaiabtittemibihineseciindsese 10

United States v. Alcan Aluminum. Corp.,

964 F.2d 252 (3d Cir. 1992) ...................000 17

United States v. Alcan Aluminum Corp.,

315 F.3d 179 (2d Cir. 2003) .................. 4, 16,17

United States v. Atchison, Topeka & Santa

Fe Ry., Nos. CV-F-92-5068 OWW, CV-F-

96-6226 OWW, -6228 OWW, 2003 WL

25518047 (E.D. Cal. July 15, 2008)........... 8,9

United States v. Atl. Research Corp., 127 S.

ti ale Ra 4,16

United States v. Bestfoods, 524 U.S. 51

RISES ae oe Oe oa ee ee 4,13

United States v. Burlington N. R.R., 200

F.3d 679 (10th Cir. 1999).......................006. 12

United States v. Burlington N. & Santa Fe

Ry., 520 F.3d 918 (9th Cir. 2008) .......... passim

United States v. Chem-Dyne Corp., 572 F.

Supp. 802 (S.D. Ohio 1983)...........cccccc-.e0-+- 4

United States v. Hercules, Inc., 247 F.3d

Ee rae 11

United States v. Twp. of Brighton, 153 F.3d

NS BIIIE vinrisecidnctnisitcsenininasindsiniasenninen 10, 11

lv

TABLE OF AUTHORITIES — continued

Page

U.S. EPA vy. Sequa Corp. (In re Bell Petro-

leum Servs., Inc.), 3 F.3d 889 (5th Cir.

I iinnsiicestiei hile can hail passim

Woodland v. Portneuf Marsh Valley Irriga-

tion Co., 146 P. 1106 (Idaho 1915)............. 10

CONSTITUTION AND STATUTES _

eile I WI Bits OF Ola Gs crceciecccccccccesecsoece 13

ie ee i irececcsevscvveccccscescesscassees 1

SS aa eae 13, 14, 17

Te 13, 16

a Nc scceicecosunsinnetinelis 13

RULE

SS ae ee 13

SCHOLARLY AUTHORITIES

G. Boston, Apportionment of Harm in Tort

Law: A Proposed Restatement, 21 U.

Dayton L. Rev. 267 (1996) ...................0008+- g

W. Page Keeton et al., Prosser and Keeton

on the Law of Torts (5th ed. 1984)............ 5, 6,9

W. Prosser, Handbook of the Law of Torts

| 5 EN a 8

W. Prosser, Handbook of the Law of Torts

EES Se ee ee 21

W. Prosser, Joint Torts and Several Liabil-

ity, 25 Cal. L. Rev. 413 (1937).........c.c0c000-+. 8, 11

OTHER AUTHORITIES

Restatement (Second) of Torts (1965) ...... 5, 9, 11,

14, 21

Vv

TABLE OF AUTHORITIES — continued

Page

Restatement (Third) of Torts: Apportion-

ment of Liability (2000) ..................... 5, 6, 8, 11

Kathleen Segerson, Reason Found., Policy

Study No. 187, Redesigning CERCLA Li-

abilty: An Analysis of the Issues (Apr.

1995), available at http://www.reason.

ee Riccnistecsactniebecntanniotinnneun 23

INTEREST OF AMICUS CURIAE

ASSOCIATION OF AMERICAN RAILROADS!

The Association of American Railroads (“AAR”) is

an incorporated, non-profit trade association repre-

senting the nation’s major freight railroads. AAR

members operate approximately 72% of the rail in-

dustry’s line-haul mileage, produce 95% of its freight

revenues, and employ 92% of rail workers. Amtrak,

which carries virtually all intercity rail passenger

traffic, is a member of AAR, as are several commuter

railroads. In matters of significant interest to its

mer’ ers, AAR frequently appears before Congress,

the courts and administrative agencies on behalf of

the railroad industry, including by participating as

amicus curiae in cases that raise issues of vital con-

cern to its membership and the judicial system. AAR

filed a brief in this case in support of the petition for

certiorari.

This is a case of great concern to the proper ad-

ministration of the Comprehensive Environmental

Response, Compensation, and Liability Act (CER-

CLA), 42 U.S.C. § 9601 et seq. This is true both gen-

erally and as it particularly affects the railroad in-

dustry. For historical reasons, the Nation’s railroads

are exposed to substantial potential liability under

this regime. Many of them possess large land hold-

ings with unbroken fee title stretching back a century

or more, which they hold as passive landowners, and

which lessees and adjacent owners have polluted.

‘Pursuant to Supreme Court Rule 37.6, AAR states that no

counsel for any party authored this brief in whole or in part and

that no entity or person, aside from amicus curiae, its members,

and its counsel, made any monetary contribution toward the

preparation or submission of this brief. The parties have con-

sented to the filing of this brief and letters reflecting their con-

sent have been filed with the Clerk of Court.

2

The decision below threatens to impose untold liabil-

ity on passive landowners, like these railroads, that

Congress neither foresaw nor intended. Perversely,

the decision below will make apportionment easier to

obtain for truly culpable polluters, but leave passive

landowners jointly and severally liable for cleanup

costs for contamination in which they played no part

and that did not even occur on their own land. The

AAR submits this brief to explain why nothing in the

law requires this bizarre result.

INTRODUCTION

As petitioners have thoroughly explained, the deci-

sion below radically departs from basic common-law

principles that CERCLA did not displace. AAR sub-

mits this brief to supplement that submission with

regard to two related issues that are fundamental to

the error in the decision below.

First, the decision below represents a marked de-

parture from basic principles of liability apportion-

ment. CERCLA incorporates common-law principles

in areas like this one where Congress did not speak,

and the common law favors apportionment in circum-

stances like these. Both the Second and Third Re-

statements of Torts provide for apportionment among

causes when there is any “reasonable basis” to do so.

This well-accepted test reflects a long tradition of

permitting apportionment upon even rough approxi-

mations of comparative causation, and particularly

where multiple pollutants or polluters are involved.

Indeed, the Restatements highlight pollution and

public nuisance as circumstances in which appor-

tionment is especially appropriate, and numerous

common-law decisions over the past century and

more have applied this accepted rule. The district

court's well-reasoned and detailed apportionment

3

analysis falls squarely within this common-law tradi-

tion, and the Ninth Circuit's decision to overturn it is

a substantial misstep that this Court should correct.

Second, not only is the Ninth Circuit’s heavy bias

against apportionment insupportable, it would have

extraordinary and untoward practical repercussions

for landowners who have done everything that could

reasonably be expected of them. For while Congress

plainly intended CERCLA to place burdens on certain

classes of defendants, these burdens are not without

limits. For instance, CERCLA combines a strict li-

ability regime with a modified causation standard

and retroactive liability — factors which, taken to-

gether, can result in massive cleanup costs. CERCLA

thus works harsh effects on passive landowners, who

“may be required to pay huge amounts for damages to

which their acts did not contribute.” U.S. EPA v. Se-

qua Corp. (In re Bell Petroleum Servs., Inc.), 3 F.3d

889, 897 (5th Cir. 1993); see also id. (CERCLA often

imposes “liability ... for conduct predating the enact-

ment of CERCLA, ... even for conduct that was not

illegal, unethical, or immoral at the time it oc-

curred”).

But the burdens imposed by the statute were not

meant to be unlimited, and the Ninth Circuit erred in

concluding that Congress intended such a result.

There is no textual basis for concluding that CER-

CLA’s already extreme exposure to liability should be

multiplied by raising the bar for apportionment so

high that joint and several liability for landowners is

a near certainty. In no way does “the basic structure

of the CERCLA statutory scheme” require that land-

owners bear these costs in full. United States v. Bur-

lington N. & Santa Fe Ry., 520 F.3d 918, 941 (9th

Cir. 2008). The Ninth Circuit reached its conclusion

only by mistakenly departing from basic common-law

4

rules, mistakenly inferring a supposed legislative

purpose that landowners rather than taxpayers al-

ways should bear cleanup costs, and imposing eviden-

tiary burdens that are unmeetable and have no statu-

tory basis. This Court should reject the bald inequity

embraced by the decision below, which does violence

to the common-law principles that inform a proper

reading of the statute.

ARGUMENT

I. COMMON-LAW PRINCIPLES FAVOR AP-

PORTIONMENT IN CASES LIKE THESE.

A. CERCLA is silent on the proper method for ap-

portioning damages among defendants, and so the

lower courts have relied on “traditional and evolving

common law principles,” especially as reflected in the

Restatements of Torts. See Jn re Bell Petroleum

Serus., Inc., 3 F.3d at 895; United States v. Alcan

Aluminum Corp., 315 F.3d 179, 185 (2d Cir. 2003);

United States v. Chem-Dyne Corp., 572 F. Supp. 802,

810 (S.D. Ohio 1983).2 This approach derives from

the basic rule that “in order to abrogate a common-

law principle, the statute must speak directly to the

question addressed by the common law.” United

States v. Bestfoods, 524 U.S. 51, 63 (1998) (internal

quotation marks and alteration omitted). And, rele-

vant here, the common law broadly authorizes the

apportionment of damages based upon causation.

The familiar rule under the Second Restatement is

that harm should be apportioned by cause when there

is any “reasonable basis” to do so:

2? This Court has reserved the question whether CERCLA cre-

ates joint and several liability. United States v. Atl. Research

Corp., 127 S. Ct. 2331, 2339 n.7 (2007).

5

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.

Restatement (Second) of Torts § 433A(1) (1965) (em-

phasis added); see also id. § 881 cmt. a (“apportion-

ment is made” for “harms ... that afford a reasonable

basis for division”). The Third Restatement endorses

the same rule, as do numerous standard authorities.

See, e.g., Restatement (Third) of Torts: Apportion-

ment of Liability § 26(b) (2000) (damages to be di-

vided by cause when there is a “reasonable basis” to

do so); W. Page Keeton et al., Prosser and Keeton on

the Law of Torts § 52, at 346, 350 (5th ed. 1984) (“en-

tire liability is imposed only where there is no factual

basis for holding that one wrongdoer’s conduct was

not a cause in fact of part of the harm” and “emphasis

is placed on the possibility of reasonable apportion-

ment”). The rule embodies simple justice: “No party

should be liable for harm it did not cause ....” Re-

statement (Third) of Torts § 26 cmt. a.

Notwithstanding this basic common-law principle,

the Ninth Circuit now requires “precision” in appor-

tionment and detailed “records” to satisfy the burden

of apportionment. 520 F.3d at 944. To apportion

damages, however, it never has been essential that

particular harms be capable of precise attribution to

specific causes. Neither “precision” nor particular

“records” ever have been thought necessary. See id.

at 952 (Bea, J., dissenting) (“The panel applies CER-

CLA in a novel and unprecedented way to impose im-

possible-to-satisfy burdens on CERCLA defendants.”

(footnote omitted)).

6

Rather, it is well established that the proper ques-

tion is whether “a factual basis can be found for some

rough practical apportionment.” Keeton et al., supra,

§ 52, at 345 (emphasis added); see also id. § 52, at 350

(“The difficulty of any complete and exact proof ...

has not been regarded as sufficient justification for

entire liability.”). As the Third Restatement explains,

“[a]s long as there is some evidence that would permit

[apportionment], courts should permit the factfinder”

to apportion damages. Restatement (Third) of Torts

§ 26 rptrs’ note cmt. h (emphasis added); id. § 26 cmt.

f (“The fact that the magnitude of each indivisible

component cannot be determined with precision does

not mean that the damages are indivisible. All that

is required is a reasonable basis for dividing the

damages.”).

As the Fifth Circuit properly has explained in this

very context:

If the expert testimony and other evidence estab-

lishes a factual basis for making a reasonable es-

timate that will fairly apportion liability, joint

and several liability should not be imposed in the

absence of exceptional circumstances. The fact

that apportionment may be difficult, because

each defendant's exact contribution to the harm

cannot be proved to an absolute certainty, or the

fact that it will require weighing the evidence

and making credibility determinations, are in-

adequate grounds upon which to impose joint

and several liability.

In re Bell Petroleum Servs., Inc., 3 F.3d at 903.

B. These authorities reflect an abiding tradition in

which common-law courts have permitted apportion-

ment based on the fact-finder’s approximation of rela-

tive causation. Presented with “evidence which rea-

7

sonably tends to show the relative proportion” of the

tortfeasors’ causal responsibility, Eckman v. Lehigh

& Wilkes-Barre Coal Co., 50 Pa. Super. 427, 432

(1911), courts “permit the jury, as reasonable men, to

make from the evidence the best possible estimate,”

Hill v. Chappel Bros. of Mont., Inc., 18 P.2d 1106,

1110 (Mont. 1932). As early as 1879, the Connecticut

Supreme Court expressed a strong preference for ap-

portionment — even rough apportionment — as more

appropriate than the inherently inequitable alterna-

tive:

It may be wery difficult for a jury to determine

just how mw: h damage the defendant is liable for

and how much should be left for the city to an-

swer for; but this is no more difficult of ascer-

tainment than many questions which juries are

called upon to decide. They must use their best

judgment, and make their result, if not an abso-

lutely accurate one, an approximation to accu-

racy.... If the plaintiff is entitled to damages and

the defendant liable for them, the one is not to be

denied all damages, nor the other loaded with

damages to which he is not legally liable, simply

because the exact ascertainment of the proper

amount is a matter of practical difficulty.

Sellick v. Hall, 47 Conn. 260, 274 (1879).3

Indeed, far from imposing rigid proof requirements

like the specific historical records that the Ninth Cir-

cuit demanded here, 520 F.3d at 944, courts often

3 Accord Harley v. Merrill Brick Co., 48 N.W. 1000, 1002 (lowa

1891) (“the fact that it is difficult to measure accurately the

damage which was caused by the wrongful act of each contribu-

tor to the aggregate result does not ... make any one liable for

the acts of others”); Chipman v. Palmer, 77 N.Y. 51, 53-54

(1879).

8

have “relaxed the quality and quantity of evidence

required to meet the burden of production.” Re-

statement (Third) of Torts § 26 rptrs’ note cmt. h. As

Prosser explained in his seminal article:

The difficulty of assessing separate damages ...

is not regarded as sufficient justification for en-

tire liability....

... It has been said that no very exact proof will

be required, and that general evidence as to the

proportion in which the defendants contributed

to the result will be sufficient to support separate

verdicts.

W. Prosser, Joint Torts and Several Liability, 25 Cal.

L. Rev. 413, 438-39 (1937).4

The district court in the decision below far sur-

passed the “rough practical apportionment” or “best

possible estimate” that common-law courts routinely

have accepted as sufficient. It held a bench trial that

stretched over multiple weeks, after which it enter-

tained multiple motions to amend its findings of fact

and conclusions of law. See United States v. Atchi-

4 See also W. Prosser, Handbook of the Law of Torts § 47, at

328, 334-35 (1st ed. 1941) (whenever “a logical basis [could] be

found for some rough practical apportionment, which limits a

defendant's liability to that part of the harm which he has in

fact caused, it [could] be expected that the division [would] be

made”); Ogden v. Lucas, 48 Ill. 492, 404 (1868); Cal. Orange Co.

v. Riverside Portland Cement Co., 195 P. 694, 695 (Cal. Ct. App.

1920) (“[t]hough in cases of this sort entire accuracy is impossi-

ble, and the difficulty of accurately proportioning and assessing

the damage done by defendant's mill is great” “the trial court

was at liberty to cstimate as best it could, from the evidence be-

fore it, how much of the total damage caused by the operations

of the two cement companies was occasioned by defendant's

plant, and in doing so might measure with a liberal hand the

amount of damage caused by defendant's mill”).

9

son, Topeka & Santa Fe Ry., Nos. CV-F-92-5068

OWW, CV-F-96-6226 OWW, -6228 OWW, 2003 WL

25518047, at *1 (E.D. Cal. July 15, 2003). It ulti-

mately issued a lengthy order that contained nearly

300 individual findings of fact, id. at *2-43, and en-

gaged in a detailed apportionment analysis, in the

course of which it evaluated the relative contributions

of the various parcels of land, id. at *80-96; see No.

07-1601 Pet. 21-22. If this is a “meat-axe” approach,

520 F.3d at 944, the common law has been operating

a butcher shop for more than a century.

C. Of particular relevance here, authorities long

have recognized that environmental torts (and ear-

lier, so-called “public nuisances”) are particularly

susceptible of apportionment. See, e.g., G. Boston,

Apportionment of Harm in Tort Law: A Proposed Re-

statement, 21 U. Dayton L. Rev. 267, 301 (1996)

(“[P]rinciples of comparative causation are especially

relevant in the toxic torts area, where measures of

toxicity can be applied to determine apportionment.”)

So, for instance, Prosser explains that

[nJuisance cases, in particular, have tended to

result in apportionment of the damages, largely

because the interference with the plaintiffs use

of land has tended to be severable in terms of

quantity, percentage, or degree. Thus defen-

dants who independently pollute the same

stream, or who flood the plaintiffs land from

separate sources, are liable only severally for the

damages individually caused, and the same is

true as to nuisances due to noise, or pollution of

the air.

Keeton et al., supra, § 52, at 349 (footnotes omitted;

citing numerous cases); Restatement (Second) of

Torte § 433A cmts. c, d, illus. 5; id. § 881 cmt c., illus.

1, 2. Apportionment therefore repeatedly has been

10

used in cases closely analogous to this one, in which

pollution by multiple tortfeasors harmed a plaintiffs

person or property. E.g., Ralston v. United Verde

Copper Co., 37 F.2d 180, 184 (D. Ariz. 1929), aff'd, 46

F.2d 1 (9th Cir. 1931); Thomas v. Ohio Coal Co., 199

Ill. App. 50, 56-57 (1916); see also Woodland v. Port-

neuf Marsh Valley Irrigation Co., 146 P. 1106, 1106-

07 (idaho 1915); Harley v. Merrill Brick Co., 48 N.W.

1000, 1001-02 (lowa 1891); Chipman v. Palmer, 77

N.Y. 51, 53-54 (1879); Cal. Orange Co. v. Riverside

Portland Cement Co., 195 P. 694, 695 (Cal. Ct. App.

1920); Eckman, 50 Pa. Super. at 432.

Courts in applying these common-law principles in

the context of CERCLA have properly declined to set

the bar anywhere near as high as the Ninth Circuit

did in specifying what constitutes a “reasonable ba-

sis” for apportionment. The Sixth Circuit, for in-

stance, has held that courts “should be receptive to

any argument for divisibility that provides a reason-

able basis” for apportioning harm. United States v.

Twp. of Brighton, 153 F.3d 307, 320 (6th Cir. 1998).

In line with the common-law authorities, the Fifth

Circuit likewise has authorized apportionment on the

basis of “a reasonable estimate.” In re Bell Petroleum

Serus., Inc., 3 F.3d at 903.5

What is more, apportionment has been performed

using precisely the factors that the district court re-

lied upon here. The Ninth Circuit held that a “simple

fraction based on the time that the Railroads owned

the land cannot be a basis for apportionment.” 520

5 See also Dent v. Beazer Materials & Serus., Inc., 993 F. Supp.

923, 946 (D.S.C. 1995) (apportioning CERCLA damages among

former owners by distinguishing types of hazardous products

and applying causal analysis), affd, 156 F.3d 523, 530-31 (4th

Cir. 1998).

11

F.3d at 945. Time, however, long has been recognized

as an appropriate basis for apportioning pollution-

based harms:

The harm inflicted may be conveniently sever-

able in point of time. Thus if two defendants, in-

dependently operating the same plant, pollute a

stream over successive periods, it is clear that

each has caused a separate amount of harm, lim-

ited in time, and that neither has any responsi-

bility for the harm caused by the other.

Restatement (Second) of Torts § 433A cmt. c; Prosser,

25 Cal. L. Rev. at 434-45 (same; “[i]n such cases there

is available a logical basis for the apportionment of

the loss”). Accord Restatement (Third) of Torts § 26

rptr’s note cmt. f.

In Jn re Bell Petroleum Services, likewise, the Fifth

Circuit permitted CERCLA apportionment based on a

combination of periods of ownership and evidence

about what occurred during those periods. 3 F.3d at

903-04. The Sixth Circuit did as well in Township of

Brighton. See 153 F.3d at 320 (“if [defendant] can

show that it was only an operator after a particular

year, and that only a certain percentage of the haz-

ardous material was introduced into the property af-

ter that year, it can show divisibility and be held li-

able only for the releases occurring after it became an

operator’); see also United States v. Hercules, 247

F.3d 706, 719 (8th Cir. 2001) (chronological evidence

can prove divisibility of harm). That is closely analo-

gous to what the district court did here. See No. 07-

1601 Pet. 21; No. 07-1601 Pet. App. 253a.

Geography likewise supplies a reasonable basis for

apportionment of CERCLA liability. See Twp. of

Brighton, 153 F.3d at 320 (if an entity can show that

its “activities were completely limited to a discrete

12

and measurable section of the property, and that the

releases onto or from that section represented a dis-

crete and measurable harm, this would provide a rea-

sonable basis for apportionment”); United States v.

Burlington N. R.R., 200 F.3d 679, 700 (10th Cir.

1999) (affirming decision dividing credit for a settle-

ment agreement that apportioned liability based on

geography); see also Kamb v. U.S. Coast Guard, 869

F. Supp. 793, 799 (N.D. Cal. 1994) (apportioning

based on geography and volume of contaminant).

D. The Ninth Circuit further departed from basic

common-law principles when it demanded “records”

evincing with “precision” highly particularized proof

such as “the proportion of the amount of chemicals

stored, poured from one container to another, or

spilled on each parcel.” 520 F.3d at 944. This ex-

trastatutory rule not only would impermissibly favor

direct evidence over circumstantial evidence, but in-

deed would require the use of a specific kind of direct

evidence. Such an inflexible proof requirement runs

contrary to the fundamental rule, embodied through-

out the law, that any competent evidence will suffice.

“(D]irect evidence of a fact is not required.” Michalic

v. Cleveland Tankers, Inc., 364 U.S. 325, 330 (1960).

On the contrary, it is a “conventional rule of civil liti-

gation” that proof by “either direct or circumstantial

evidence” is permissible. Desert Palace v. Costa, 539

U.S. 90, 99-100 (2003) (internal quotation marks and

alterations omitted); Merchs.’ Mut. Ins. Co. v. Baring

Bros. & Co., 87 U.S. 159, 161 (1873) (“[c]lompetent

evidence may be ... direct or circumstantial”). And to

“demand|{] either direct evidence ... or else more sub-

stantial circumstantial evidence” improperly usurps

the factfinder’s duty and power to draw reasonable

inferences from the evidence. Gallick v. Balt. & Ohio

R.R., 372 U.S. 108, 114 (1963).

13

Precisely because this basic rule holds true

throughout the law, statutes and rules are specific

when they depart from it by requiring particular

forms of proof. The examples of this are familiar.

The statute of frauds and its statutory descendents

require that certain contract terms be in writing.

See, e.g., Cal. Civ. Code § 1624(a) (conveyance of an

interest in land requires a writing signed by the

party to be charged). The best evidence rule requires

that, “To prove the content of a writing, recording, or

photograph, the original writing, recording, or photo-

graph is required, except as otherwise provided in

these rules or by Act of Congress.” Fed. R. Evid.

1002. Cf. also-U.S. Const. art. III, § 3, cl. 1 (requiring

that treason be proved “on the Testimony of two Wit-

nesses to the same overt Act, or on Confession in

open Court”).

CERCLA is no different: When Congress means to

require specific proof under CERCLA, it does so ex-

pressly. This is true of the “innocent purchaser” de-

fense, which exempts from liability defendants who

purchased already-contaminated land, and did not

know of and had no reason to know of the contamina-

tion. 42 U.S.C. §§ 9601(35)(A)(G), 9607(b)(3). In es-

tablishing this defense, Congress set forth detailed

evidentiary requirements. For example, to establish

the “no reason to know” element, the landowner must

“demonstrate” that it conducted inquiries of, among

other things, “historical sources, such as chain of title

documents, aerial photographs, building department

records, and land use records,” “recorded environ-

6 Cf. Bestfoods, 524 U.S. at 62-63 (CERCLA incorporates

common-law principles when Congress is silent); Desert Palace,

539 U.S. at 99 (declining to depart from conventional rules of

civil litigation because, among other reasons, the statute was

silent).

14

mental cleanup liens against the facility,” and “gov-

ernment records, waste disposal records, under-

ground storage tank records, and hazardous waste

handling, generation, treatment, disposal, and spill

records, concerning contamination at or near the fa-

cility.” Id. § 9601(35)(B)(i), @ii)(), (IV), (V).

That Congress imposed such requirements in cer-

tain provisions of CERCLA speaks volumes about its

failure to do so elsewhere. And here, nothing in

CERCLA’s text imposes any particularized eviden-

tiary requirement or burden to support apportion-

ment. Nowhere does CERCLA demand the use of re-

cords like the Ninth Circuit sought here. See 520

F.3d at 958 & n.16 (Bea, J., dissenting). Nor does ap-

portionment generally require any such specialized

proof; if anything, common-law evidentiary burdens

regarding apportionment are relaxed. See supra at 6-

9.

Importantly, the special evidentiary requirement

adopted by the Ninth Circuit was critical to its deci-

sion; it was only by establishing this standard that

the court of appeals could hold categorically insuffi-

cient the evidence upon which the district court had

relied, and so overturn the district court’s reasonable

apportionment. 520 F.3d at 936 n.18. This is be-

cause, as the Ninth Circuit properly recognized, the

question of whether harm is apportionable in a par-

ticular case is a factual question that is reviewed only

for clear error. Jd. at 942; see Restatement (Second)

Torts § 434(2) & cmt. d. It therefore was squarely

within the province of the factfinder — the district

court judge who heard the evidence — to “draw rea-

sonable inferences from basic facts to ultimate facts.”

Jackson v. Virginia, 443 U.S. 307, 319 (1979); accord

Reeves v. Sanderson Plumbing Prods., Inc., 530 U.S.

133, 150 (2000) (“drawing of legitimate inferences

15

from the facts” is the task of the factfinder (internal

quotation marks omitted)). The Ninti Circuit could

overturn the district court’s thoughtful apportion-

ment analysis only by holding that the evidence upon

which it relied did not “suffice” as a matter of law.

520 F.3d at 943.7 This conclusion amounts to nothing

more than an evasion of the admittedly applicable

standard of review. The district court apportioned

damages only after a lengthy bench trial that re-

sulted in hundreds of individual findings of fact and a

carefully supported apportionment analysis. This

manifestly was not clearly erroneous, and nothing in

the law of apportionment or the rules of evidence

prevented the district court from relying on the com-

petent evidence that was before it as a reasonable ba-

sis for apportionment.

Il. AN UNYIELDING APPLICATION OF JOINT

AND SEVERATI. LIABILITY WILL HAVE

EXTRAORDINARY AND INEQUITABLE

EFFECTS.

The Ninth Circuit’s “unreasonable” apportionment

standard, see id. at 953 (Bea, J., dissenting), has sig-

nificant consequences for a category of defendants

who already face exceptionally broad liability under

the statute: landowners, including passive landown-

ers who have in no way contributed to the need for

cleanup. The decision below would multiply those

harsh effects, in a way that “Congress did not... in-

tend.” Jd. The burden that would be placed on rail-

roads exemplifies the draconian outcome that results

from imposing vigorous and unyielding proof re-

quirements for apportionment.

7 See also 520 F.3d at 945 (“many of the district court's con-

clusions were factually correct but legally insufficient”).

16

A. Any discussion of CERCLA liability must begin

with a recognition of the broad ways in which the

statute imposes liability. First, liability is effectively

strict. It affects four classes of parties including,

relevant here, the current owner or operator of the

facility, and one who, at the time of disposal, owned

or operated the facility at which the hazardous sub-

stances were disposed of. 42 U.S.C. § 9607(a)(1), (2).8

Parties who fall into these classes are commonly re-

ferred to as “potentially responsible parties,” or

“PRPs.” The defenses available to PRPs are exceed-

ingly limited — showing that the release from the fa-

cility was caused solely by an act of God, an act of

war, or the act of a third party with whom the PRP

had no direct or indirect contractual relationship. Id.

§ 9607(b). As a result — and while the statute never

uses the actual words — courts have held that CER-

CLA imposes strict liability. See, e.g., United States

v. Atl. Research Corp., 127 S. Ct. 2331, 2336 (2007);

Alcan Aluminum Corp., 315 F.3d at 184; In re Bell

Petroleum Servs., Inc., 3 F.3d at 897. At its most

fundamental level, this means that a landowner can

be held liable even if the tenant caused the disposal

on the lessor’s property, and even if all defendants

exercised due care.

These results are exacerbated by the exceptionally

broad manner in which EPA has purported to define

the term “facility,” which is a sine qua non of CER-

CLA liability for owners and operators. See generally

42 U.S.C. § 9607(a)(1), (2) (imposing liability on own-

ers and operators of a “facility”). Much of the land

owned by railroads that is at issue in this and similar

8In addition, liability is imposed upon any person that ar-

ranged to dispose of its hazardous substances at the facility, and

any transporter who delivered hazardous substances to the facil-

ity. 42 U.S.C. § 9607(a)(3), (4).

17

cases is made up of thin strips of property that, be-

cause of common-carrier obligations, are appurtenant

to industrial facilities. CERCLA’s statutory defini-

tion of facility is loose, see generally id. § 9601(9)(B),

and EPA has in application defined facilities in

sweeping ways that often do not track property own-

ership. See Pet. Br. 48-50; e.g., Quaker State Minit-

Lube, Inc. v. Fireman’s Fund Ins. Co., 52 F.3d 1522,

1524-25 (10th Cir. 1995). Simply put, EPA can sweep

in an adjacent landowner that has engaged in no cul-

pable conduct whatsoever merely by defining broadly

the boundaries of a particular facility, and it has

every incentive to do so when such a landowner is a

solvent company with deep pockets.

Piling on to this harsh scheme, courts have held

that a CERCLA plaintiff need not “establish a specific

causal connection between [the] defendant’s hazard-

ous substances and the release or the plaintiffs in-

currence of response costs.” United States v. Alcan

Aluminum Corp., 964 F.2d 252, 264 (3d Cir. 1992);

see Amoco Oil Co. v. Borden, Inc., 889 F.2d 664, 670

n.8 (5th Cir. 1989). Furthermore, courts have con-

cluded that CERCLA liability is retroactive — a PRP

will be held responsible for cleanup costs stemming

from conduct that predates enactment of the law in

1980. Alcan Aluminum Corp., 315 F.3d at 188 (col-

lecting cases).

The weight of this liability regime sits heaviest on

the shoulders of the landowner. Unlike the catego-

ries of PRPs in (a)(3) and (a)(4), for instance — “ar-

rangers’ and “transporters” who may at least have

had some role in disposal of the waste — the owner

may be liable for wholly passive conduct — 1.e., doing

nothing more than owning land that someone once

polluted. The government's proof is reduced to the

simple question of fee ownership, and absent any rea-

18

sonable form of apportionment, the owner is left

“holding the bag.” 520 F.3d at 941.

B. Railroads, which are quintessential passive

landowners, serve as a vivid illustration of the ex-

tremely harsh manner in which these aspects of

CERCLA operate. Due in large measure to the na-

ture of their business, railroads possess substantial

landholdings, much of which follows the 140,000 or

more miles of right-of-way railroad track throughout

the country. In the mid-19th Century, for instance,

certain railroads (including predecessors to the peti-

tioners here) received land grants from the federal

government (in exchange for discounted hauling of

federal freight, among other things) in order to facili-

tate national economic development. See generally

Leo Sheep Co. v. United States, 440 U.S. 668, 670-77

(1979). These landholdings — many of which took the

form of checkerboard parcels of land — are broadly

geographically dispersed. See id. at 672.

The railroads often are, and they historically have

been, passive landowners with regard to such parcels.

In short, the railroads have thousands of lessees.

AAR’s members collectively have more than 75,000

land leases. One member railroad alone has 7800

land leases in 28 states, with (largely incomplete) re-

cords of an additional 28,000 known historical leases.

Another member has more than 23,000 land leases,

the majority of which are smaller than one acre.

What is more, because much of this property is in

proximity to rail lines, the railroads’ lessees are often

commercial and industrial concerns that require ac-

cess to these transportation facilities.

For railroads, therefore, all of the harshest ele-

ments of CERCLA liability line up against them.

They are fee owners of thousands of pieces of com-

mercial and industrial property, with title stretching

19

back decades and beyond. (The railroad land grants,

for instance, took place between 1850 and 1871.) The

odds that one of those tenants spilled or disposed of

hazardous substances are high, as is the possibility

that the offending tenant is no longer in existence.

For all of these reasons, CERCLA works particularly

harsh effects on landowners like railroads.

C. That Congress contemplated these particular

aspects of the CERCLA regime is clear. What is

equally clear is that the strict evidentiary reqnuire-

ments imposed by the decision below find no support

in CERCLA or in the common law. And, they would

work extraordinary and inequitable hardship on rail-

roads and other landowners like them. The Ninth

Circuit rejected the district court’s thoughtful appor-

tionment among causes, and instead demanded evi-

dence of “the proportion of the amount of chemicals

stored, poured from one container to another, or

spilled on each parcel.” 520 F.3d at 944. For in-

stance, it sought “adequate records” by which “to es-

timate the amount of leakage attributable to activi-

ties on the Railroad parcel, how that leakage traveled

to and contaminated the soil and groundwater under

the Arvin parcel, and the cost of cleaning up that con-

tamination.” Id.

It comes as no surprise that “none of this data is in

the record,” id., because such data rarely exists. The

panel recognized as much, conceding that “[iJt may

well be that such information is, as a practical mat-

ter, not available for periods long in the past, when

future environmental cleanup was not contemplated.”

Id. This is an understatement. Because CERCLA

operates retroactively, it may impose massive liabil-

ity on passive landowners whose title to landholdings

stretches back a century or more, to a time when “fu-

ture environmental cleanup” indeed “was not con-

20

templated.” There would have been no cause for any-

one to keep the records demanded by the Ninth Cir-

cuit — if such records ever existed — and certainly no

cause to believe that a landowner/lessor would have a

lessee’s records of its activities.

The same is true for more recent activities as well,

because commercial leases do not typically require

tenants to retain and disclose the detailed records

that the Ninth Circuit now demands. Much less do

they require this for adjacent properties, as in this

case. And therein lies the problem. The passive

landowner is the least likely party to have access to

any such detailed records that — if they ever were

kept at all — belonged to the site operator. Of course,

it is implausible to think that many (if any) commer-

cial enterprises kept detailed records of spills or leaks

in the years before modern environmental regula-

tions required such information. And it is unthink-

able that the lessor would have done so. The Ninth

Circuit was totally at ease with this inequitable re-

sult. 520 F.3d at 945 (“The net result of our approach

to apportionment of liability, consequently, may be

that landowner PRPs, who typically have the least

direct involvement in generating the contamination,

will be the least able to prove divisibility.”). But

there is not one piece of evidence that Congress in-

tended any such result.

D. Compounding the harm embodied in its core

ruling embracing inflexible joint and several liability,

the Ninth Circuit imposed special additional re-

quirements on landowners. It held that whereas Re-

statement principles “work[] nicely” when “several

defendants are all polluters themselves,” a landowner

can only “establish divisibility by demonstrating a

reasonable basis for concluding that a certain propor-

tion of the contamination did not originate on the

21

portion of the facility that the landowner owned at

the time of the disposal.” Jd. at 937, 938. This in-

verts logic; it would place the greatest burden on the

least culpable party, who also is the least likely to

possess or control the information that the court now

would demand.

Not only is this illogical; it is fundamentally incon-

sistent with the basic common-law principle that les-

sors are not usually held responsible for the actions of

lessees. See Restatement (Second) of Torts § 355.

The common law recognized that a iessee assumes

the obligations of ownership, with few exceptions.

And, of particular relevance here, lessors were not

held liable for nuisances created by lessees unless the

intended purpose of the lease would necessarily re-

sult in a nuisance. As Prosser explained, the lessor is

“not responsible for activities to which he did not con-

sent and which he had no reason to contemplate or

expect, or for dangers created by the tenant which he

has not authorized.” W. Prosser, Handbook of the

Law of Torts § 63, at 403 (4th ed. 1971) (footnote

omitted). Here, there is nothing in the record to re-

flect, nor any basis to believe, that by leasing land to

an agricultural chemical storage and distribution fa-

cility, a nuisance necessarily would occur. As Holmes

put it, “the landlord will not be liable for the use of

the premises in such a way as to do harm, merely be-

cause there was a manifest possibility of their being

9 Accord Edgar v. Walker, 32 S.E. 582, 584 (Ga. 1899) (“As a

general rule, a landlord is not liable to third persons for any in-

jury they sustain, occasioned by the wrongful act of his tenant in

keeping the rented premises in a dangerous or unhealthy condi-

tion. The only exceptions to this rule are (1) When the landlord

has contracted with the tenant, to repair; (2) when he has let the

premises in a ruinous condition; (3) when he has expressly lLi-

censed the tenant to do acts amounting to a nuisance.”).

22

used in such a way.” Clifford v. Atl. Cotton Mills, 15

N.E. 84, 87 (Mass. 1888).'°

Regardless of CERCLA’s modification of the base

common-law rule regarding liability, there is no justi-

fication for adopting a rule that landowner/lessors —

who would not have been liable at all under the

common law — now will be subject to special burdens

when apportionment among causes is at issue. There

is nothing in CERCLA to support the Ninth Circuit’s

notion that because the statute’s liability regime is

“expansive” in some regards, it is without limitation

in all other respects. On the contrary, Congress did

not make the statute expansive in this regard; it did

not speak to this issue at all. 520 F.3d at 957 n.12

(Bea, J., dissenting). Through its silence, Congress

“left it to the courts to fashion some rules that will, in

appropriate instances, ameliorate this harshness.” In

re Bell Petroleum Serus., Inc., 3 F.3d at 897.

The practical result of the Ninth Circuit’s rule is

that joint and several liability is a near certainty.

Accordingly, any “orphan shares” attributable to de-

funct parties will become the responsibility of the re-

maining viable PRPs. In a case like this one, the

railroad, whose only act was to lease its land, ends up

jointly responsible for all cleanup costs, picking up

the shares not only of former tenants that actually

caused the problem, but potentially the shares of all

nearby landowners whose land falls within EPA’s

definition of the “facility” on which disposals oc-

10 Furthermore, contrary to the approach taken by the Ninth

Circuit, even when a lessor is held liable, liability may be appor-

tioned between the lessor and lessee. E.g., Arthur Richards, Inc.

v. 79th Fifth Ave. Co., 441 N.E.2d 1114 (N_Y. 1982)

23

curred.!! An appropriate division of harm among

PRPs, by contrast, provides some measure of propor-

tionality. It does not in the normal situation excul-

pate a PRP entirely, and indeed in this case the dis-

trict court’s apportionment would leave the railroads

responsible for millions of dollars in past and pro-

jected cleanup costs. See Pet. Br. 13 n.4. Reasonable

apportionment simply holds out for the landowner

the possibility that its final responsibility will in

some measure reflect its relationship to the site. The

decision below, by contrast, departs from common-

law principles, imposes an unyielding and over-

whelming form of liability that Congress never wrote

into the statute, and would impose untold liability

where it is entirely unfair.

11 The opportunity to seek contribution from other PRPs —

held out by the Ninth Circuit as a means to blunt these inequi-

ties, 520 F.3d at 940-41, 945 — provides small comfort in these

situations. Contribution allocates liability among viable PRPs.

But when the landowner is one of the few (or only) viable PRPs,

it takes on the orphan shares of defunct, culpable parties.

In fact, the government's practice of picking the easy targets

further compounds the inequities. See Kathleen Segerson, Rea-

son Found., Policy Study No. 187, Redesigning CERCLA Li-

abilty: An Analysis of the Issues § V1.B.1 (Apr. 1995), available

at http://www.reason.org/ps187.html (government targets cer-

tain PRPs for expediency, particularly based on their ability to

pay). Empowered by joint and several liability, the U.S. can sue

the landowner and others whose liability is easily proved, shift-

ing to those defendants the burden of prosecuting contribution

actions against the PRPs, against whom recovery is far less

likely, or indeed impossible.

24

CONCLUSION

For the foregoing reasons, the decision below

should be reversed.

Respectfully submitted,

LOUIS P. WARCHOT CARTER G. PHILLIPS*

MICHAEL J. RUSH G. PAUL MOATES

ASSOCIATION OF SAMUEL I. GUTTER

AMERICAN RAILROADS ERIC A. SHUMSKY

LAW DEPARTMENT NAOMI SCHOENBAUM

50 F Street, N.W. STACEY L. WILSON

Washington, D.C. 20001 SIDILEY AUSTIN LIP

(202) 639-2505 1501 K Street, N.W.

Washington, D.C. 20005

(202) 736-8000

Counsel for Amicus Curiae

November 24, 2008 * Counsel of Record