PRECEDENTIAL
    UNITED STATES COURT OF APPEALS
         FOR THE THIRD CIRCUIT
              _____________

             Nos. 16-1994 & 16-2244
                 _____________

         TRINITY INDUSTRIES, INC.;
TRINITY INDUSTRIES RAILCAR CORPORATION
                     Appellants in No. 16-2244

                        v.

     GREENLEASE HOLDING COMPANY;
    AMPCO-PITTSBURGH CORPORATION

                       Greenlease Holding Company,
                         Appellant in No. 16-1994

                _______________

  On Appeal from the United States District Court
     for the Western District of Pennsylvania
           (W.D. Pa. No. 2-08-cv-01498)
     District Judge: Hon. Joy Flowers Conti
                _______________
                           Argued
                      September 5, 2017

Before: CHAGARES, JORDAN, and HARDIMAN, Circuit
                    Judges.

                 (Filed: September 11, 2018)
                      _______________

Steven F. Baicker-McKee [ARGUED]
Mark K. Dausch
Marc J. Felezzola
Babst Calland
603 Stanwix Street
Two Gateway Center, 6th Floor
Pittsburgh, PA 15222
      Counsel for Greenlease Holding Co.

Frederick W. Addison, III
Nolan C. Knight [ARGUED]
Munsch Hardt Kopf Harr & Dinan
3800 Lincoln Plaza
500 North Akard Street
Dallas, TX 75201
      Counsel for Trinity Industries, Inc. and Trinity
      Industries Railcar Corp.

Paul D. Steinman [ARGUED]
Jessica S. Thompson
Eckert Seamans Cherin & Mellott
600 Grant Street, 44th Floor
Pittsburgh, PA 15219
      Counsel Ampco-Pittsburgh Corp.




                               2
                      _______________

                 OPINION OF THE COURT
                     _______________

JORDAN, Circuit Judge.

       This is a dispute about the proper allocation of costs to
remediate a contaminated manufacturing site in Greenville,
Pennsylvania. From 1910 until 1986, Greenlease Holding Co.
(“Greenlease”),1 a subsidiary of the Ampco-Pittsburgh
Corporation (“Ampco”), owned the site and operated railcar
manufacturing facilities there. Trinity Industries, Inc. and its
wholly-owned subsidiary, Trinity Industries Railcar Co.
(together referred to as “Trinity”), acquired the site from
Greenlease in 1986 and continued to manufacture railcars there
until 2000. An investigation by the Commonwealth of
Pennsylvania into Trinity’s waste disposal activities resulted in
a criminal prosecution and eventual plea-bargained consent
decree which required, in relevant part, that Trinity remediate
the contaminated land. That effort cost Trinity nearly $9
million.
       This appeal arises out of the District Court’s
determination that, under the Comprehensive Environmental
Response, Compensation, and Liability Act, 42 U.S.C. § 9601
et seq., (“CERCLA”), and Pennsylvania’s Hazardous Sites
Cleanup Act, 35 Pa. Stat. § 6020.101 et seq., (“HSCA”),
Trinity is entitled to contribution from Greenlease for

       1
        Greenlease was known first as the Greenville Metal
Products Company and then as the Greenville Steel Car
Company. For purposes of this opinion, we refer to all
Greenlease and Greenville entities as “Greenlease.”




                               3
remediation costs. After eight years of litigation, and having
sorted through a century of historical records, the District
Court allocated 62% of the total cleanup costs to Greenlease
and the remainder to Trinity. The parties filed cross-appeals
challenging a number of the District Court’s rulings, including
its ultimate allocation of cleanup costs. For the reasons that
follow, we will affirm the District Court’s pre-trial rulings on
dispositive motions; we will vacate its cost allocation
determination; and we will remand for further proceedings
consistent with this opinion.

 I.    FACTUAL BACKGROUND2

       The site in question, known by the parties as the “North
Plant,” is a tract of land that was used as a manufacturing site
by a succession of companies. Greenlease and Trinity also, at
different times, operated facilities on a nearby tract of land
called the “South Plant,” though that property does not figure
prominently in this appeal. Over time, the footprint of the
North Plant grew from eleven to thirty-four acres. That
industrial development, as well as the many years of
manufacturing activity that occurred there, resulted in multiple
releases of hazardous materials – primarily lead – into the
ground.

       A. The North Plant – 1898 to 1986

       From at least 1898 until sometime before Greenlease’s
acquisition of the North Plant in 1910, Shelby Steel Tube

       2
          The facts recounted here are taken from the District
Court’s post-trial findings of fact or from facts in the record
that are undisputed.




                               4
Company owned and operated a steel tube factory on eleven
acres of land that is now part of the North Plant. Over the
course of its ownership, Shelby Steel deposited historic fill as
it was constructing its manufacturing facilities. According to
the District Court, “[h]istoric fill is ‘a soil mixed with various
non-native materials, including construction demolition debris,
concrete, asphalt, or it could be industrial materials such as slag
or ash.’” (App. at 186.) Unfortunately, historic fill often
contains lead and other contaminants.

       Greenlease began its manufacturing activities at the
North Plant soon after acquiring the property. Between 1911
and 1922, it significantly expanded the North Plant to support
its growing business of building and repairing railcars. During
that expansion, Greenlease used historic fill in the foundations
supporting the new structures and rail lines. Operations at the
North Plant included two shops to paint the railcars, and
Greenlease used a variety of toxic chemicals and lead paint
during the painting process, without doing anything
meaningful to collect or contain the runoff.

       B. Relationship Between Greenlease and Ampco

       In 1983, Ampco acquired Greenlease,3 but their
relationship predated that acquisition. They had had three
overlapping board members since 1979 and continued to do so
until 1986. Other than those three shared board members and

       3
          Greenlease’s stock was first acquired in 1937 by
another company, the Pittsburgh Forging Co. Ampco then
acquired all of the stock of the Pittsburgh Forging Co., and,
through a series of transactions, became the sole shareholder
of Greenlease.




                                5
one shared officer, no other persons were employees of both
Ampco and Greenlease. Greenlease employees alone “were
responsible for all day-to-day operations at the North Plant,
including any waste disposal, waste handling, painting,
abrasive blasting, welding, and fabrication operations.” (App.
at 81-82.) Those employees coordinated disposal with outside
contractors and communicated with the Pennsylvania
Department of Environmental Protection (“PADEP”) on
environmental matters. Indeed, Ampco “did not employ any
engineers or persons with technical experience in
manufacturing that could make decisions for [Greenlease] with
respect to environmental compliance or waste management.”
(App. at 82.) Instead, “Ampco employed only a professional
staff, such as accountants, actuaries, and lawyers[.]” (App. at
82.) Ampco did provide Greenlease with advice regarding the
laws and regulations related to Greenlease’s waste generation,
and Ampco monitored that waste generation.

       The cooperation between parent and subsidiary was
complete enough that Greenlease adopted a resolution
declaring that any action taken by Ampco that it “may think
necessary and desirable to take on behalf of [Greenlease] shall
be deemed to be the action of [Greenlease’s Board].” (App. at
72 (citation omitted).) Ampco also asserted the right to
approve Greenlease’s expenditures that exceeded a certain
amount, though Greenlease was solely responsible for placing
and paying any purchase orders. In addition, Ampco provided
certain services to Greenlease to minimize costs, including
overseeing a single retirement plan and providing centralized
financial planning and master insurance policies.




                              6
      C. Trinity’s Acquisition of the North Plant

       In 1986, Ampco authorized the Greenlease board of
directors to sell the North Plant to Trinity. The Purchase and
Sale Agreement between Trinity and Greenlease (the
“Agreement”) included a clause declaring that Greenlease
“makes no representation or warranty regarding compliance
with the Environmental Protection Act, any other
environmental laws or regulations or any hazardous waste laws
or regulations (collectively, ‘Environmental Laws’).” (App. at
199.)      Mutual indemnification provisions specific to
environmental liabilities provided, in pertinent part:

      [Greenlease] agrees to indemnify and hold
      harmless [Trinity] against Damages arising out
      of or related to violations of Environmental
      Laws, which were caused by [Greenlease] or its
      predecessors in title to the assets at the [North
      Plant] on or prior to the date of Closing. [Trinity]
      agrees to indemnify and hold harmless
      [Greenlease] against Damages arising out of or
      related to violations of Environmental Laws,
      which are caused by [Trinity] or its successors in
      title to the assets at the [North Plant] after the
      date of the Closing. It is the intention of the
      parties that liability under this Section for any
      condition that is caused by the acts of
      [Greenlease] or its predecessors in title to the
      assets prior to the date of the Closing and by the
      acts of [Trinity] or its successors in title to the
      assets after the date of Closing shall be allocated
      between the parties in a just manner taking into




                               7
       account degree of fault, period of violation and
       other relevant factors.

(App. at 61 (some alterations in original).) Those indemnities
were stated to be effective for only three years after the closing
of the property sale. The Agreement further provided that
Trinity “has not assumed, and expressly denies assumption
hereby of, any other liability, obligation or commitment of
[Greenlease] other than as set forth above or otherwise
expressly set forth herein.” (App. at 60-61 (alteration in
original).) Finally, a “[n]on-waiver of [r]emedies” clause in
the Agreement provided that “[t]he rights and remedies herein
provided are cumulative and are not exclusive of any rights or
remedies which the parties hereto may otherwise have at law
or in equity.” (App. at 62.)

        Following the 1986 sale of the North Plant to Trinity,
Greenlease continued to exist only as a “shell holding company
without any [employees,] business activities, for profit
activities, or other commercial undertakings[.]” (App. at 89.)
Its assets decreased at the end of each year following the sale
of the North Plant, from about $51 million in 1987 to $658,594
in 1990. In the third and fourth years following the sale of the
North Plant to Trinity, Greenlease issued dividends to Ampco,
leaving Greenlease with only a $250,000 reserve for liabilities.
At that time, Greenlease had no known liabilities beyond the
reserve. The executive vice president and chief administrative
officer for Ampco, who was also an officer and director of
Greenlease, stated that it was common for dividends to be
made from a subsidiary to Ampco after an indemnification
period ended. An environmental reserve was placed on




                                8
Greenlease’s books when Trinity sued Greenlease and
Ampco.4

       D. The North Plant – 1987 to 2004

        After purchasing the North Plant, Trinity continued the
manufacture of railcars there. In one of the paint shops, it
installed concrete floors and used tar paper to capture paint
drippage. Beginning in late 1987, it implemented a policy
preventing the use of metal-containing paints at the North
Plant. In 1994, Trinity removed the second paint shop,
excavated the old dirt floors, and dumped the soil onto a field
at the South Plant. Trinity then erected a new paint shop at the
North Plant.

        Six years later, in 2000, Trinity ceased the North Plant
operations. It sold the property in 2004 to a third-party (the
“Buyer”). In connection with that sale, Trinity did not conduct
an environmental assessment to determine whether the soil was
contaminated, and it prohibited the Buyer from performing
such testing without its consent. The Buyer demolished almost
all of the existing buildings at the North Plant to sell the scrap
steel for profit. Trinity maintains that, at some point, the Buyer
dumped onto the North Plant property hazardous chemicals
and waste that had been produced by the demolition of the
North Plant buildings, exacerbating the pre-existing
environmental harm.




       4
        In 2008, that reserve was $150,000, and in 2009, it
was $282,500.




                                9
       E. The Commonwealth’s Investigation and the
          Consent Decree

       In 2004, the Commonwealth of Pennsylvania and
PADEP began an investigation into allegations that Trinity had
improperly disposed of hazardous waste at the North Plant.
The Commonwealth filed a criminal complaint against Trinity
in 2006, raising three felony counts and eight misdemeanor
counts related to the illegal handling and disposal of hazardous
waste. Trinity entered into a plea agreement with the
Commonwealth that required the repayment of investigative
costs, payment of a fine, contribution to a nonprofit
organization, and, pursuant to a consent decree authorized by
PADEP (the “Consent Decree”), the remediation of
environmental contamination.

       The Consent Decree stated that further investigation of
the North Plant was “necessary to fully identify the nature and
extent of the release of hazardous substances at and/or
potentially migrating from the North Plant … and to determine
the Response Actions necessary to remediate the hazardous
substances at and/or potentially migrating from [the North]
Plant.” (App. at 513.) The cleanup was governed by
Pennsylvania’s Land Recycling and Environmental
Remediation Standards Act, 35 Pa. Stat. § 6026.101 et seq.,
commonly known as “Act 2,” and the associated investigation
was not limited to the time during which Trinity owned and
operated the North Plant.

       Trinity was on a short leash. It was ordered to get
approval from PADEP before it took any “significant step”
pertaining to the property, and it was required to submit to
PADEP “an investigation work plan, a supplemental




                              10
investigation work plan, a notice of intent to remediate, a
remedial investigation report, a proposed cleanup work plan, a
supplemental cleanup work plan, and a final report.” (App. at
213-14.) Those additional mandates increased the difficulty
and expense of the remediation project. The remediation
efforts were also affected by the fact that “[t]he North Plant
was a ‘high profile, high visibility location’” and is bordered
by residential communities on three sides. (App. at 218
(citation omitted).)

       PADEP approved Trinity’s remedial investigation work
plan in 2007. Trinity later sent Greenlease a pre-suit notice
describing the contamination and its legal position that
Greenlease had contributed to the pollution.

      F. Trinity’s Cleanup of the North Plant

       To perform the necessary cleanup, Trinity had to buy
back the North Plant. It then selected Golder Associates, Inc.
(“Golder”) to perform, direct, and supervise the cleanup
operations. PADEP approved that selection. Trinity did not
employ a competitive bidding process to select Golder because
it had been impressed by Golder’s cleanup operations at
several other sites and because the Consent Decree’s deadlines
created an urgency to get a remediation consultant in place as
soon as possible. Trinity and Golder agreed to an “open
billing” process that provided Golder would be paid only for
the work it ultimately needed to perform. (App. at 218-19.)
Billing was on a “cost plus 10 percent” basis, which gave
Golder a ten percent markup on the expenses it incurred. (App.
at 219.)




                              11
        Golder’s cleanup efforts required it to first identify
areas of the property that were of concern. It analyzed
available historical information concerning construction and
manufacturing activities that had taken place at the North
Plant. It then conducted soil sampling to further identify areas
requiring remediation. Golder ultimately divided the North
Plant into twenty impact areas that required remediation.
Thirteen of the twenty impact areas were primarily
contaminated by lead. The remaining impact areas were
primarily contaminated by volatile and semi-volatile organic
compounds and a variety of other hazardous substances. Major
remediation activities included excavating contaminated soil,
refilling excavated areas with clean material, chemically
treating contaminated soil, transporting excavated soil to
appropriate landfills, and placing asphalt caps over parts of the
North Plant. In total, Golder disposed of approximately 39,000
tons of soil off-site and capped about 15,000 tons of soil with
asphalt.

       Those efforts cost nearly $9,000,000 and made the
property usable again. Parts of the North Plant with asphalt
caps are suitable for use as a parking lot. Other areas are
suitable for industrial or commercial use. There is ongoing
work at the North Plant to ensure that the safety mechanisms
created as part of the environmental remediation continue to
function.5




       5
        According to the District Court’s findings of fact, the
work includes maintaining the asphalt caps and continued
ground water monitoring.




                               12
II.    PROCEDURAL HISTORY

       Invoking federal and state laws, Trinity filed a
complaint against Greenlease and Ampco in 2008 to defray the
North Plant remediation costs. More specifically, Trinity
sought cost recovery under CERCLA pursuant to 42 U.S.C.
§ 9607, cost recovery under the Resource Conservation and
Recovery Act (“RCRA”) pursuant to 42 U.S.C.
§ 6972(a)(1)(B), and contribution under CERCLA pursuant to
42 U.S.C. §§ 9613(f)(1) and 9613(f)(3)(B). It also brought cost
recovery and contribution claims under the HSCA, as well as
state common law claims for contribution and negligence per
se.

       A. Pre-Trial Motions and Rulings

      Trinity’s claims against Ampco were premised on
Ampco’s alleged direct or derivative liability for Greenlease’s
conduct at the North Plant. Upon cross motions for summary
judgment on that issue, the District Court concluded that
Ampco was not directly or derivatively liable for pollution at
the North Plant.

        Greenlease also moved for judgment on the pleadings,
arguing that Trinity’s claims were barred by the
indemnification provisions of their Agreement. It claimed that
once the mutual indemnities expired, neither party was entitled
to seek compensation from the other. The District Court
rejected that argument, ruling that the existence and expiration
of the indemnification provisions did not prevent Trinity from
seeking other remedies available at law or in equity.




                              13
        Greenlease and Trinity later filed cross motions for
summary judgment on Trinity’s CERCLA, RCRA, HSCA, and
common law claims. The District Court granted partial
summary judgment for Trinity, holding as a matter of law that
Greenlease was a potentially responsible person under
CERCLA and the HSCA. It also granted Greenlease’s cross-
motion in part, granting it summary judgment on all of
Trinity’s claims other than those for contribution under 42
U.S.C. § 9613(f)(3)(B) and 35 Pa. Stat. § 6020.705(c)(2). The
litigation proceeded to a bench trial to determine the equitable
allocation of cleanup costs between the parties.

       Prior to trial, Trinity tried to recoup costs associated
with its cleanup of the South Plant, but the District Court
concluded that Trinity was not entitled to those costs because
Greenlease had never owned or operated that property or
disposed of any hazardous waste at the South Plant.

       B. The Parties’ Cost Allocation Proposals

       Trinity’s and Greenlease’s experts each provided the
District Court with a proposal for the equitable allocation of
cleanup costs between the parties. Trinity’s expert, Joseph B.
Gormley, Jr., relied on available historical information to
identify three sources of contamination at the North Plant:
volatile chemicals used in manufacturing operations; general
dispersions caused by painting; and historic fill used for
construction.    He then employed that same historical
information to assign each party a percentage of responsibility
for the contamination found within each impact area. Next,
Gormley analyzed the major remediation activities and
associated costs required to clean up each impact area. To
arrive at a total cost allocation for the major remediation




                              14
activities, he multiplied the percentage of responsibility for
each specific impact area by the major remediation activity
costs in that specific area and added those results together.
That produced an overall percentage allocation. Gormley
applied that same overall percentage to general project costs
not tied to any specific impact area. Ultimately, he allocated
99% of the costs to Greenlease and 1% to Trinity.

        Not surprisingly, Greenlease’s expert, Steven Gerritsen,
proposed a very different cost allocation. He concluded that
most of the lead present at the North Plant was caused by the
use of historic fill rather than Greenlease’s operations at the
facility. He calculated that Greenlease was responsible for
depositing fill on only 2.8 acres of the thirty-four acre North
Plant. He opined that the rest of the fill predated Greenlease’s
purchase of the property and was thus not Greenlease’s
responsibility. Gerristen also suggested that much of Golder’s
work was unreasonable and unnecessary and thus that Trinity
had spent more money than it should have to perform the
cleanup. Gerristen ultimately concluded that Greenlease
should be allocated only 12-13% of the cleanup costs.

       C. The District Court’s Cost Allocation Opinion

        In an admirably thorough opinion, the District Court
endeavored to make sense of the extensive record, including
the competing expert contentions. It first concluded that
Greenlease was not responsible for any of the contamination
attributable to Shelby Steel or any other non-party because
Trinity had failed to show that those parties were “unknown,
insolvent, or otherwise immune from suit.”6 (App. at 351.)

       6
           A court may equitably allocate among the parties




                              15
The Court, however, rejected Greenlease’s contention that
Golder incurred unreasonable or excessive costs when
performing its cleanup at the North Plant.

        To assign each party a percentage of responsibility for
the contamination within each impact area, the District Court
relied heavily on historic maps and schematics of the North
Plant. For many impact areas, the Court agreed with
Greenlease that the lead contamination could be attributed
solely to Shelby Steel’s use of historic fill, and therefore should
not be a source of liability for Greenlease. For other impact
areas, the Court found that Greenlease was responsible for the
deposit of historic fill, or was solely responsible for the use of
volatile chemicals, and that Greenlease should thus bear full
responsibility for the pollution. For the remaining impact
areas, the District Court split responsibility between the parties
based on the number of years that each had owned the property
or on various other considerations such as known use of a
specific chemical contaminant.

       After determining the percentages of responsibility
within each impact area, the District Court considered the
major remediation activities that took place in each impact area


before it the share of hazardous waste contamination belonging
to responsible third-party entities not before it (such allocated
amounts being known as “orphan shares”). But it can typically
only do so if such orphan shares belong to entities that are
unknown, insolvent, or immune from suit. See Litgo N.J. Inc.
v. Comm’r N.J. Dep’t of Envtl. Prot., 

, 380 n.4 (3d
Cir. 2013) (permitting equitable allocation of orphan shares
among liable parties at the court’s discretion). As found by the
District Court, that is not the case here.




                                16
to determine an overall allocation of cost. Though it purported
to follow Gormley’s methodology, the Court departed from it
in an important respect: Gormley’s methodology accounted for
the fact that different remediation activities cost different
amounts of money, whereas the District Court’s methodology
did not. To arrive at its cost allocation, the Court multiplied
the percentage of responsibility it attributed to Greenlease by
the square footage or cubic yardage involved in each
remediation activity. The District Court then added the results
and divided by the total square footage and cubic yardage for
all remediation activities at the North Plant to arrive at the
overall cost allocation percentage. By those calculations, it
concluded that Greenlease was responsible for 83% of the total
costs, while Trinity was responsible for 17%.

       The District Court then considered a variety of equitable
factors to ensure the fairness of the overall cost allocation. It
ultimately reduced Greenlease’s percentage of responsibility,
based on three equitable factors.

       First, it found that at least a portion of Trinity’s
remediation costs were attributable to the actions of the third-
party Buyer and, in particular, the Buyer’s decision to demolish
buildings at the North Plant. The Court said that Trinity failed
to “specify the amount of response costs it incurred to
remediate the waste left at the North Plant by [the Buyer].”
(App. at 380.) Therefore, “there [was] an equitable need to
reduce Greenlease’s percentage of responsibility for response
costs to reflect an amount attributable to [the Buyer].” (App.
at 380.) Accordingly, the Court reduced Greenlease’s
responsibility by 6%.




                               17
        Second, it concluded that the existence of the
indemnification provisions demonstrated the parties’ intent to
shift liability, so it further reduced Greenlease’s share of
responsibility by 5%.

       Third, it recognized that the property value of the North
Plant had increased as a result of remediation since the land
was now suitable for some commercial or industrial uses. The
Court concluded that an additional 10% reduction in
Greenlease’s responsibility was appropriate to account for that
increased market value that would inhere to Trinity.

       After accounting for those equitable deductions, the
District Court determined that Greenlease was responsible for
62% of “all response costs incurred by … Trinity … for the
cleanup at the North Plant[.]” (App. at 388-89.)

III.   DISCUSSION7

       A. Statutory Background

       Congress enacted CERCLA in 1980 “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.” Burlington N. & Santa Fe Ry. v.
United States, 

, 602 (2009) (internal quotation

       7
           The District Court had jurisdiction over Trinity’s
federal law claims under 42 U.S.C. §§ 6972(a) and 9613(b) and
28 U.S.C. § 1331. It had supplemental jurisdiction over
Trinity’s state law claims under 28 U.S.C. § 1367. We have
jurisdiction pursuant to 28 U.S.C. § 1291.




                              18
marks and citation omitted). Under CERCLA, a party who has
paid for environmental remediation may seek to hold other
potentially responsible parties (“PRPs”) liable through the cost
recovery mechanisms of § 107(a) or the contribution
mechanisms of § 113(f) of that statute.8 Agere Sys., Inc. v.
Advanced Envtl. Tech. Corp., 

, 216-18 (3d Cir.
2010). The remedies under those two provisions are distinct.

(citing United States v. Atl. Research Corp., 

, 138 (2007)). While § 107(a) authorizes complete cost
recovery under a joint and several liability theory, § 113(f)
permits a party to seek contribution from other PRPs following
a CERCLA suit brought by a governmental authority against
that first party, or after that party has resolved its “liability to
the United States or an individual State through an
administratively or judicially approved settlement.” 

enacted the HSCA in 1988 to
provide additional statutory tools to deal with the improper
disposal of hazardous waste within the Commonwealth. 35 Pa.
Stat. § 6020.102; Gen. Elec. Envtl. Servs., Inc. v. Envirotech
Corp., 

, 115 (M.D. Pa. 1991).

       Although Trinity initially sought both cost recovery and
contribution from Greenlease, the only claims remaining on
appeal are claims for contribution pursuant to CERCLA
subsection § 113(f)(3)(B), and the analogous section of the
HSCA, 35 Pa. Stat. § 6020.705(c)(2). See also Trinity Indus.,
Inc. v. Chi. Bridge & Iron Co., 

, 136 (3d Cir.
2013) (holding that a party who enters into a consent decree
under state law is entitled to seek contribution under
§ 113(f)(3)(B)). Because a party’s “liability under the HSCA

       8
           As cited earlier, those sections of CERCLA are
codified at 42 U.S.C. §§ 9607(a) and 9613(f), respectively.




                                19
mirrors liability under CERCLA” and “the cost recovery and
contribution provisions in HSCA are virtually identical to
those in CERCLA,” Agere Sys., 

, our
resolution of Trinity’s claim for contribution under CERCLA
is determinative of its companion HSCA claim.

       B. Greenlease’s Appeal

       Greenlease raises three primary issues on appeal. First,
it appeals the District Court’s determination that the
indemnification provisions of the Agreement between it and
Trinity do not preclude Trinity from seeking contribution. We
will affirm because the language of the Agreement better
supports the District Court’s conclusion. Second, Greenlease
appeals the ruling that the costs Trinity and Golder incurred in
cleaning up the North Plant were all necessary and reasonable
under CERCLA. We will affirm because those costs have the
requisite nexus to remedying environmental harm at the North
Plant and because the record does not support Greenlease’s
contention that Trinity incurred excessive costs. Third,
Greenlease challenges the overall cost allocation ordered by
the District Court. We agree with Greenlease that the Court’s
cost allocation analysis was flawed, and we will therefore
vacate the judgment and remand for further proceedings.

              1. The Agreement’s Indemnification Provisions
                 Do Not Preclude Trinity from Seeking
                 Contribution from Greenlease.

       Greenlease argues that, at the conclusion of the three-
year mutual indemnification period stated in its Agreement
with Trinity, the parties were released from any subsequent
statutory or common law responsibility to one another.




                              20
Greenlease thus asserts that it was error to deny its motion for
judgment on the pleadings. Our review of a motion for
judgment on the pleadings is plenary. Caprio v. Healthcare
Revenue Recovery Grp., LLC, 

, 146 (3d Cir.
2013). Such a motion should not be granted unless the moving
party has established that there is no material issue of fact to
resolve, and that it is entitled to judgment as a matter of law.
Rosenau v. Unifund Corp., 

, 221 (3d Cir. 2008).
We also exercise plenary review over questions of contract
interpretation. Great Am. Ins. Co. v. Norwin Sch. Dist., 

, 243 (3d Cir. 2008).

        CERCLA allows parties to utilize indemnification
agreements “to shift the ultimate financial loss” for
environmental cleanup costs. Hatco Corp. v. W.R. Grace &
Co. Conn., 

, 404 (3d Cir. 1995). The statute says
plainly that it does not “bar any agreement to insure, hold
harmless, or indemnify a party to such agreement for any
liability under this section.” 42 U.S.C. § 9607(e)(1). Whether
the expiration of the indemnification provisions at issue here
effectively shifted all financial burden for CERCLA cleanup
costs to Trinity thus turns on the proper interpretation of the
Agreement.        “[A]greements among private parties …
addressing the allocation of responsibility for CERCLA claims
are to be interpreted by incorporating state … law.” 

. Here, that means Pennsylvania law.

        When a contract is clear and unambiguous,
Pennsylvania binds the parties to the intent contained within
the writing itself. Wert v. Manorcare of Carlisle PA, LLC, 

, 1259 (Pa. 2015). “The whole instrument must be
taken together in arriving at contractual intent.” Great Am.

(quoting Murphy v. Duquesne Univ. of the




                              21
Holy Ghost, 

, 429 (Pa. 2001)). Courts are not to
interpret one provision of the contract in a way that annuls a
different provision of it, Capek v. Devito, 

, 1050
(Pa. 2001), and “when specific or exact provisions seem to
conflict with broader or more general terms, the specific
provisions are more likely to reflect the intent of the parties[,]”
Musko v. Musko, 

, 256 (Pa. 1997). Those
interpretive rules lead us to conclude that the Agreement at
issue reserved Trinity’s right to seek contribution from
Greenlease for environmental cleanup costs.

         The Agreement’s indemnification provisions stated, in
relevant part, that each party indemnified the other for any
“[d]amages arising out of or related to violations of
Environmental Laws” and that liability for any such violations
would be “allocated between the [parties] in a just manner
taking into account degree of fault, period of violation and
other relevant factors.” (App. at 599-600.) It is true that the
mutual indemnification expired after three years. The
Agreement did not, however, contain language expressing the
parties’ intent that Trinity would assume all of Greenlease’s
obligations and liabilities after that three-year period. Rather,
the Agreement contained explicit “non-assumption of
liabilities” and “non-waiver of remedies” clauses. The “non-
assumption of liabilities” clause provided that Trinity “has not
assumed, and expressly denies assumption hereby of, any other
liability, obligation or commitment of [Greenlease] other than
as set forth above or otherwise expressly set forth herein.”
(App. at 567.) It is reading far too much into the words “any
other liability” to think they meant that the prominent risk of
environmental liability was the one thing the parties meant for
Trinity to be stuck with. Moreover, the “non-waiver of
remedies” clause plainly provided that “[t]he rights and




                                22
remedies herein provided are cumulative and are not exclusive
of any rights or remedies which the parties hereto may
otherwise have at law or in equity.” (App. at 612-13.) The
express language of the contract, therefore, provides both that
Trinity did not assume any of Greenlease’s liabilities or
obligations following the three-year mutual indemnification
period, and that Trinity did not waive its statutory rights under
CERCLA and the HSCA to seek contribution from Greenlease.
In short, while the contractual right to indemnification ended,
all other rights remained.

       Greenlease’s three primary arguments to the contrary do
not persuade us.          First, Greenlease argues that the
indemnification provision should control our interpretation of
the entire Agreement because it is more specific than the “non-
waiver of remedies” clause. That reasoning, however, puts too
high a premium on specificity. Yes, the contractual indemnity
is specific. But the non-assumption of liabilities and non-
waiver of remedies provisions are plain enough for us to
discern the intent of the parties, and that intent was to preserve
non-contractual rights. Besides, there is a sense in which the
indemnification language is not more specific than the other
relevant provisions: it does not address the parties’ liabilities
after the first three years following the sale. The “non-
assumption of liabilities” and “non-waiver of remedies”
clauses do. They are not time limited and therefore can be
understood as specifically addressing the time period after the
expiration of the contractual indemnities. We will not construe
the indemnification provision to cover time periods that, by the
plain language of the contract, it does not cover. See Jacobs
Constructors, Inc. v. NPS Energy Servs., Inc., 

,
373 (3d Cir. 2001) (“[B]ecause the nature and purpose of any
indemnity agreement involves the shifting and voluntary




                               23
assumption of legal obligations, they are to be narrowly
construed.”).

        Second, Greenlease argues that allowing Trinity to seek
contribution against it pursuant to the “non-waiver of
remedies” clause “renders the environmental indemnity
provision meaningless[.]” (Green. Opening Br. at 36.) But
that argument again ignores the critical fact that the parties, by
agreeing to the three-year mutual indemnification provision,
granted to each other certain contractual rights separate and
distinct from any statutory, legal, or equitable rights or
remedies. The “non-waiver of remedies” clause is perfectly
clear in that regard, reserving to both parties “any rights or
remedies which the parties … may otherwise have at law or in
equity.” (App. at 613.) As the District Court concluded, the
contractual remedies created by the indemnification provision
were, by the terms of the Agreement, “cumulative” and not
“exclusive” of the remedies available at law or in equity. (App.
at 66, 613.) Greenlease could have bargained for a provision
in the Agreement whereby Trinity would have assumed all of
Greenlease’s obligations and liabilities following the
expiration of the three-year indemnification provision. But it
did not.

       Third, Greenlease relies on Keywell Corporation v.
Weinstein, 

(2d Cir. 1994), a decision by the United
States Court of Appeals for the Second Circuit, to argue that
all CERCLA and HSCA liability automatically transferred to
Trinity after the expiration of the three-year mutual
indemnification provision. There are, though, important
differences between the contract at issue in Keywell and the
Agreement here that are sufficient to make that case inapposite.
The corporate plaintiff in Keywell sought to recover CERCLA




                               24
cleanup costs from two individual defendants, who had been
officers of the corporation that sold the relevant piece of land
to the plaintiff. 

The purchase agreement for the
land included a two-year indemnification provision
guaranteeing to hold the plaintiff harmless for any damages
arising out of “any breach of warranty or representation” by the
selling entity “or its management stockholders” and for “any
liabilities or obligations of [s]eller” not explicitly listed in the
purchase agreement. 

The plaintiff then entered into
a separate thirty-year indemnification agreement with the
corporate seller that guaranteed to hold the plaintiff harmless
for any damages that “arose or existed” prior to the purchase
agreement. 

thirty-year indemnification
agreement stated that only the corporate entity would be held
to the longer indemnification period, not its individual officers.

to seeking to recover CERCLA cleanup
costs from the individual defendants, the plaintiff had entered
into yet another contract, this last one “unconditionally
releas[ing]” the corporate entity’s former “Management
Group,” which included the individual defendants, from any
claims the plaintiff might have had under the purchase
agreement. 

set of facts, the Second Circuit held
that the plaintiff could not recover CERCLA cleanup costs
from the individual defendants because the relevant contractual
documents unequivocally expressed the parties’ intent to shift
any and all liability away from the individual officers of the
corporate entity after the initial two-year indemnification
period. 

        In contrast, the Agreement between Trinity and
Greenlease does not demonstrate an unequivocal intent to shift
liability away from Greenlease after the three-year contractual
indemnification period expired. On the contrary, rather than




                                25
releasing Greenlease from liability, the Agreement states that
Trinity did not assume any of Greenlease’s liabilities or
obligations, unless otherwise expressly provided by the
Agreement. Greenlease’s reliance on Keywell is therefore
misplaced, and we will affirm the denial of its motion for
judgment on the pleadings.

              2.       The Costs Trinity Incurred Were
                       Necessary and Reasonable.

       Greenlease next argues that the District Court
impermissibly allocated to it costs that Trinity unnecessarily
incurred by failing to impose cost controls on the remediation
work at the North Plant. We review the District Court’s factual
findings for clear error, but review de novo its interpretation of
CERCLA. Agere Sys., 

.

        A plaintiff can obtain contribution from a PRP under
§ 113(f)(3)(B) of CERCLA only if it first demonstrates a prima
facie case of liability under § 107(a). See N.J. Tpk. Auth. v.
PPG Indus., Inc., 

, 104 (3d Cir. 1999). Here, that
requires Trinity to demonstrate the following: first, that the
North Plant is a facility; second, that Greenlease is a PRP;
third, that “the release or threatened release of a hazardous
substance has occurred”; and fourth, that Trinity incurred
“necessary response costs consistent with the [National
Contingency Plan.]”9 Chevron Mining Inc. v. United States,

       9
          The National Contingency Plan provides a set of
standards governing environmental cleanup activities,
including “‘methods and criteria for determining the
appropriate extent of removal, remedy, and other measures,’
42 U.S.C. § 9605(a)(3), and ‘means of assuring that remedial




                               26


, 1269 (10th Cir. 2017) (internal quotation marks
and citation omitted). Greenlease does not dispute the District
Court’s conclusions on the first three points. It only argues that
the District Court erred by determining, as a legal matter, that
Trinity’s response costs were per se necessary because they
were undertaken in compliance with the Consent Decree. That
argument, however, even if it had merit, is irrelevant, since the
record is clear that Trinity’s response costs were in fact
necessary under CERCLA. We thus need not address whether
response costs undertaken in compliance with a consent decree
should be considered necessary per se.

       A cost is considered “necessary” and hence subject to
shared liability if there is “some nexus between [it] and an
actual effort to respond to environmental contamination.”10


action measures are cost-effective.’ [42 U.S.C.] § 9605(a)(7).”
United States v. E.I. Dupont De Nemours & Co. Inc., 

, 168 (3d Cir. 2005) (en banc).
        10
            The case law that has developed around CERCLA
has interpreted the term “necessary” to refer to a more elastic
concept than how the word is typically understood. For
example, CERCLA case law defines a “necessary” cost as one
that has some “nexus” to the cleanup of environmental harm,
not as a cost without which the cleanup would not have been
possible. Compare Young v. United States, 

, 863
(10th Cir. 2005) (interpreting the term “necessary cost” in the
CERCLA context to refer to a cost that has a “nexus” to an
environmental cleanup), with NECESSARY, Black’s Law
Dictionary (10th ed. 2014) (defining “necessary” as something
“[t]hat is needed for some purpose or reason; essential”). We
have undertaken our analysis of what costs were or were not
necessary in this case in light of CERCLA precedent. Our




                               27
Young v. United States, 

, 863 (10th Cir. 2005); cf.
Black Horse Lane Assoc., L.P. v. Dow Chem. Corp., 

, 297 (3d Cir. 2000) (determining that a plaintiff did not
meet its burden to demonstrate the necessity of a response
action because it “did not relate to any remedial or response
action at the” relevant site). It must be, in other words, a
response cost, and CERCLA broadly defines a “response” to a
hazardous release to include a wide variety of investigative,
removal, and remedial actions. See 42 U.S.C. § 9601(23)-(25)
(providing a non-exhaustive list of “response” actions); W.R.
Grace & Co.-Conn. v. Zotos Int’l, Inc., 

, 92 (2d Cir.
2009) (noting that “response costs are liberally construed under
CERCLA”). The District Court’s detailed factual findings
make clear that there was a nexus between the costs Trinity
incurred and its effort to investigate and remediate the
contamination at the North Plant.

       The cleanup activities at the North Plant were guided by
the Consent Decree’s requirement that those efforts be
undertaken pursuant to the dictates of Pennsylvania’s Act 2.
That statute requires that remediation activities meet one of
three standards: a background standard comparing
contaminated areas to unaffected areas; a uniform statewide
health standard set by a state agency, which differs depending
on whether the site is meant for residential or commercial use;
or a site-specific standard “based on a site-specific risk
assessment so that any substantial present or probable future
risk to human health and the environment is eliminated or
reduced” so that the site could be utilized in accordance with
its “present or currently planned future use[.]” (App. at 212


opinion does not address how the term “necessary” should be
interpreted in contexts outside of CERCLA.




                              28
(citing 35 Pa. Stat. § 6026.301(a)).) Trinity used the statewide
health standard to determine which areas required “some type
of response action” and then used the site-specific standard to
guide the actual “soil cleanup.”11 (App. at 223.) It did not use
the background standard.

        During the investigation phase of Trinity’s cleanup
activities, its consultant Golder used soil sampling to determine
the areas of concern requiring remediation. That necessitated
the establishment of a “standard action level,” which is the
numerical threshold for determining when soil is contaminated
to an extent requiring treatment. For example, to determine
whether areas contaminated by lead – the primary contaminant
of concern – required treatment, Golder originally selected a
standard action level of 1000 milligrams of lead per kilogram
of soil. That was not a random choice. It selected that standard
because it had observed that, at a threshold level of 1500
mg/kg, some soil samples passed toxicity testing, while others
failed. At the more exacting 1000 mg/kg level, Golder was
confident that it would catch all of the soil requiring
remediation.

       But Golder was also cost conscious on that point. The
selection of an accurate standard was important because failure
to adequately remove all of the contaminated soil would
require Golder to put in place more costly hazardous waste
caps that could leave the land unusable. It initially chose the
1000 mg/kg standard for the reasons just noted, but when,
during the cleanup process, it discovered that a significant

       11
         The site-specific standard also required Trinity and
Golder to engage the local community and to accept public
comments about the cleanup efforts.




                               29
amount of soil exceeded the 1000 mg/kg standard yet could
still safely remain in place because it was going to “be capped
anyway as part of the approved remedy” (App. at 234), it
conducted a “site characterization study” to determine whether
there was a more appropriate standard action level (App. at
234-35). Golder settled on a 2500 parts per million standard
that was approved by PADEP. The record accordingly
establishes an appropriate cost sensitivity and a nexus between
Golder’s (and hence Trinity’s) investigative efforts and the
purpose of remedying environmental harms.

       The same is true with regard to the activities Golder
undertook to remediate the contaminated areas. It used three
primary response actions: first, simply consolidating
contaminated soil and placing an asphalt cap atop that soil;
second, excavating and chemically treating contaminated soil
to render it nonhazardous and then placing an asphalt cap over
the remediated area; and third, transporting contaminated soil
to an appropriate landfill.12 Golder’s soil excavation efforts
allowed it to use simple asphalt caps to cover the excavated
areas, as opposed to what are called Subtitle C caps. Subtitle
C of RCRA regulates the precise manner in which a hazardous
waste cap is put in place and maintained. Installing and
maintaining a cap in compliance with Subtitle C is more

       12
          Certain contaminated soil was amenable to chemical
treatment that rendered it nonhazardous; other soil was not
amenable to such treatment and remained hazardous prior to
disposal. The soil that was chemically treated could be
transported to a nonhazardous waste landfill, which was two to
four times cheaper than disposal at a hazardous waste landfill.
The soil remaining hazardous had to be transported to a
hazardous waste landfill.




                              30
difficult, complex, and expensive than installing and
maintaining a simple asphalt cap. The District Court found that
use of a Subtitle C cap would have made the North Plant site
look like a “landfill,” would not have been “consistent with the
residential character of Greenville,” (App. at 241), and would
have rendered much of the North Plant unusable for any
purpose. Those factual findings reinforce that Golder’s
activities had the required nexus to the stated purpose of
remedying environmental harms. The response costs Trinity
incurred were therefore necessary under CERCLA.

       Although Greenlease is correct that “[t]he cleanup at the
North Plant was more difficult, inclusive, and expensive
because it was done pursuant to the consent order and with
oversight by … PADEP,” (App. at 225), we do not agree that
those extra costs were consequently unnecessary. The Consent
Decree required compliance with state environmental
standards. To ensure that those statutory requirements were
met, Trinity and Golder had to get PADEP’s approval for each
step of the cleanup. The costs incurred to comply with the
Consent Decree were thus aimed directly at satisfying state
environmental standards and are appropriately classified as
“necessary to the containment and cleanup of hazardous
releases.” Redland Soccer Club, Inc. v. Dep’t of Army of U.S.,

, 850 (3d Cir. 1995) (citation omitted).

       A clearer way to understand Greenlease’s contentions is
to see them as challenging the reasonableness of Trinity’s
expenditures, not their necessity. Greenlease does not point to
any specific activity that was not “necessary.” Rather, it
complains that Trinity incurred excessive costs because the
Consent Decree lacked meaningful cost control mechanisms,
because Trinity hired Golder without competitive bidding, and




                              31
because Trinity agreed to a “cost-plus” billing arrangement
with Golder. Those arguments fare poorly precisely because
they do not address necessity, as that concept is applied in the
context of CERCLA.

       Greenlease’s arguments fall flat in light of the District
Court’s factual findings that we have already recounted in
some detail. Greenlease does not point to any record evidence
demonstrating how any of those facts resulted in unreasonably
excessive spending. In contrast, as the Court found, Trinity
and Golder worked together “to try to control costs or pay only
reasonable costs,” (App. at 218), and worked with PADEP “to
reduce the amount of work [Trinity] had to do to comply with
the” Consent Decree (App. at 225). The District Court credited
expert testimony that the billing methods used by Trinity
“contributed to the cost efficiency of the response work at the
North Plant” and “prevented Golder from up-charging
[Trinity.]” (App. at 219-20.) Greenlease has given us no sound
reason to disagree with that assessment.13

       13
            Even if Greenlease’s argument had merit, and the
matter were in equipoise, we might yet be inclined to affirm
the District Court’s finding that the costs Trinity incurred were
reasonable. That is because Trinity incurred those costs in
furtherance of the Consent Decree. Although we need not, and
do not, decide here whether costs incurred by a private party in
compliance with a state consent decree are presumed
reasonable under CERCLA, we note that similar costs incurred
by a government party are presumed reasonable. For example,
it is black letter CERCLA law that when a government’s
actions are not inconsistent with the National Contingency
Plan, its costs are presumed reasonable, E.I. 

, and are recoverable against PRPs, 42 U.S.C.




                               32
      We will therefore affirm the District Court’s
determination that Trinity’s response costs were necessary and
reasonable.

             3. The District Court Erred in Allocating Costs
                Between Trinity and Greenlease.

        Greenlease argues that the District Court used a purely
speculative methodology, different from the methodology
proposed by Trinity’s expert witness Gormley to allocate costs
between the parties.14 In particular, the criticism is that the
District Court relied on “volumes and surface areas … as a
proxy for the costs Trinity incurred at each impact area[.]”
(Green. Opening Br. at 22.) Greenlease contends that that
methodology was arbitrary because it failed to account for the
reality that different units of measure are not interchangeable
and because volumetric data cannot reliably serve as a proxy
for costs when some remediation activities cost more than


§ 9607(a)(4)(A). Since compliance with a consent decree
entered pursuant to state law “establishe[s] … compliance with
the National Contingency Plan,” Niagara Mohawk Power
Corp. v. Chevron U.S.A., Inc., 

, 137 (2d Cir.
2010), costs incurred by a government party in compliance
with such a decree should be presumed reasonable. There may
be a related principle warranting a similar presumption in a
context like this.
      14
           Greenlease’s expert incorporated Gormley’s cost
allocation methodology into his own cost allocation analysis,
so Gormley’s methodology was the only one presented to the
District Court.




                              33
others. According to Greenlease, the District Court was forced
to resort to a methodology based on volumetric data alone
because Trinity failed to present sufficient evidence
documenting how much it cost to undertake each of the major
remediation activities within each impact area. Greenlease’s
position is thus that the District Court’s cost allocation
methodology cannot stand, given the Court’s failure to include
actual costs in its analysis. We agree that the Court materially
deviated from the methodology presented by Gormley and so
arrived at a speculative cost allocation methodology that must
be corrected.

         CERCLA provides PRPs with a right to contribution for
remediation expenses. Atl. Research 

.
A district court “may allocate response costs among liable
parties using such equitable factors as the court determines are
appropriate.” 42 U.S.C. § 9613(f)(1). “[T]he law does not
command mathematical preciseness from the evidence in
finding damages. Instead, all that is required is that sufficient
facts ... be introduced so that a court can arrive at an intelligent
estimate without speculation or conjecture.” Scully v. US
WATS, Inc., 

, 515 (3d Cir. 2001) (alterations in
original) (internal quotation marks and citations omitted). We
review an allocation of CERCLA damages for abuse of
discretion. Agere Sys., 

. A district court
abuses its discretion when its decision depends “upon a clearly
erroneous finding of fact, an errant conclusion of law or an
improper application of law to fact.” 

       The parties and their experts in this case placed the
District Court in an unenviable position. Each of the parties
staked out extreme positions on cost allocation, with Trinity’s
expert Gormley opining that Greenlease should be held




                                34
responsible for 99% of all cleanup costs and Greenlease’s
expert opining that, despite Greenlease’s 76 years of building
and manufacturing activity at the North Plant, Trinity should
be held responsible for nearly 90% of all cleanup costs. The
record became even more difficult to sort out when, on direct
examination, Gormley gave testimony that was unclear at best
and departed from the methodology contained in his expert
report. Although we commend the District Court’s painstaking
effort to analyze nearly a century of building and
manufacturing activity by multiple parties to allocate costs
equitably between Greenlease and Trinity, the attempt to
untangle the evidentiary knot presented by the parties fell
short.

       Before addressing the District Court’s cost allocation
methodology, we begin with the methodology that Gormley
proposed in his expert report and explained somewhat at trial.
Gormley’s report presented a six-step approach to allocating
costs. First, using “historical information and investigation
findings,” Gormley assigned a percentage of responsibility to
each party for contamination in each area of concern, (D.I. 285-
2 at 10), and he applied those percentages to the impact areas
within each area of concern. He documented that step in
Tables 4-1 and 6-2. Second, he calculated the quantity of
material in each impact area that was subject to specific major
remediation activities. That step was documented in Table 6-
2. Third, he multiplied the estimated quantities of material
used for (or remediated by) major remediation activities by
each party’s percentage of responsibility for contaminating
each impact area. Fourth, he summed results from step three
to develop Trinity’s and Greenlease’s respective responsibility
percentages “for each major remediation activity[.]” (Id.)
Fifth, he multiplied the percentage of responsibility for each




                              35
major remediation activity by the cost of each such activity to
determine how to allocate the costs for each. Finally, Gormley
totaled how much in costs each party was responsible for
across all major remediation activities “to calculate a total
percent cost allocation for the major remediation activities.”
(D.I. 285-2 at 10.) Steps five and six were documented in
Table 7-1. The report opined that the final percentage
calculated at step six could be used to allocate the “general
construction costs” (i.e., costs that were incurred on a project-
wide basis that were not tied to a specific impact area) between
both parties.     Gormley’s expert report presented his
methodology as a single analysis with multiple steps.15

       Gormley’s testimony at trial, however, muddied his
otherwise straightforward methodology. At trial, he described
his methodology as a “three-stage process.” (D.I. 340 at 108.)
Stage 1, termed the “AOC-by-AOC percentage allocation,”
involved creating a percentage allocation specific to each area
of concern; stage 2, termed the “IA-by-IA percentage
allocation,” involved creating a percentage allocation for each
impact area; and stage 3, termed “major remediation
allocation,” involved creating a specific allocation for each
major remediation activity. (D.I. 340 at 108-11.) Trinity’s
counsel, in a perhaps confusingly worded set of questions,
asked if each stage was meant “to be mutually exclusive” of
the other stages, (D.I. 340 at 111), by which he appears to have
been asking if each “stage” was a separate and distinct
methodology that could be used to allocate costs, as opposed
to steps in a single methodology. In a truly confusing answer,

       15
           While the report did not break the methodology into
the six discrete steps we describe here, it did have each of the
steps, and identifying them separately is, we believe, helpful.




                               36
Gormley stated that the three stages “weren’t supposed to be
mutually exclusive,” and he went on to testify that “[t]he first
[stage] could be taken on its own,” but that the second and third
stages built on the first stage. (D.I. 340 at 111.) He ultimately
agreed with Trinity’s counsel, however, that each of his three
stages “could be used by someone who was trying to develop
their own logical or fair means to allocate responsibility for the
contamination at the North Plant[.]” (D.I. 340 at 111.)
Gormley’s testimony departed from his expert report in a
crucial way – his report made clear that each step in the
methodology built on those that came before it, and that they
were not independent means to come up with a cost allocation.
His testimony, however, was less than clear as to whether the
“stages” of his methodology were each independent analytical
means to allocate costs or steps that built on one another.

       Led by the unclear testimony, the District Court chose
Gormley’s “stage 3” – divorced from the analytical
foundations for that stage in the earlier steps of Gormley’s
analysis – to guide its cost allocation analysis.16 That at least

       16
              The District Court interpreted Gormley’s trial
testimony as establishing that “[e]ach of the three methods
used by [him] could be used on its own—without considering
the other two methods—to allocate responsibility for the
contamination at the North Plant.” (App. at 249.) Although
that conclusion was understandable based on Gormley’s
testimony, it was mistaken. While Gormley agreed that
“someone who was trying to develop their own logical or fair
means to allocate responsibility” could incorporate any one of
his stages into an allocation methodology, (D.I. 340 at 111), he
never testified that someone could separate out one stage, and
then use that stage’s allocation methodology alone to allocate




                               37
appears to have been the Court’s approach because it titled its
allocation analysis, “Overall Allocation of Responsibility
based upon Major Remediation Activity”; it stated that it
“determined an overall allocation based upon the extent of each
major remediation activity in each [impact area]”; it explicitly
listed the major remediation activities it “considered … in its
calculation”; and it cited Table 7-1 – the table corresponding
to Gormley’s stage 3 – when reaching its allocation
determination. (App. at 375-77.) The District Court, however,
materially deviated from Gormley’s suggested major
remediation activity allocation methodology by focusing only
on the quantity of material involved in all major remediation
activities, without distinguishing between activities and
without regard to cost. That was despite Gormley’s testimony
confirming “that a central feature of the analysis … reflected
in [Table] 7-1 is [the] notion of the … costs[.]” (D.I. 341 at
33.)

       The District Court’s allocation methodology proceeded
in four steps. First, it made its own factual determinations
regarding the percentage of responsibility each party bore for
contamination in each specific impact area.17 Second, it
totaled, for each impact area, the quantity of material used or
remediated by major remediation activities. The Court’s
analysis did not differentiate between remediation activities.
For example, it treated placing asphalt caps and placing topsoil


all cleanup costs for remediating contamination at the North
Plant.
       17
          We find no error in the Court’s underlying factual
findings with regard to the contamination in each specific
impact area.




                              38
as functionally the same for its cost allocation analysis despite
the fact that those two activities’ costs vary significantly.
Third, it multiplied, on an impact area-specific basis, each
party’s percentage of responsibility for contamination with the
total quantity of material used or remediated. Fourth, it used
the resulting numbers to determine the percentage of material,
in total, for which each party was responsible. That calculation
led the Court to attribute to Greenlease 83% of responsibility
for the contamination of the North Plant and to Trinity 17%.
The Court, citing Gormley’s testimony and expert report, used
those percentages to allocate “all response costs …, including
responsibility investigation, removal and remedial past costs
incurred through February 2015, for general construction costs,
… and future construction costs for ongoing operations and
maintenance work.” (App. at 377 (emphasis omitted).) Those
percentages, however, were too speculative for two reasons.
First, the Court’s methodology failed to differentiate between
different remediation activities and their varied costs, and,
second, the methodology, as applied, treated data measured in
square feet as equivalent to data measured in cubic yards.

       Although the District Court’s reliance on volumetric
data as the key factor in allocating response costs is not without
support in our case law, its use here was flawed.18 In Agere
Systems, Inc. v. Advanced Environmental Technology
Corporation, we endorsed a volumetric-centered approach to

       18
           Although we determine that the District Court erred
in utilizing the cost allocation methodology that it did, the
Court’s use of volumetric data and a focus on major
remediation activity as a means to determine the allocation of
response costs was reasonable and within the Court’s
discretion.




                               39
allocating CERCLA costs because, in that case, “volume
allocation likely reflect[ed] the dollar amounts” at 

. We clarify here that such a volumetric-centered
approach is only appropriate where the evidence supports a
finding that one standardized volumetric unit correlates with a
standardized per unit measure of cost. That may often be the
case when a CERCLA cleanup involves only one impact area,
or when a cleanup involves one primary major remediation
activity. But when, as here, an environmental cleanup involves
many impact areas and remediation activities with varying
costs, a volumetric-centered approach that fails to account for
cost differences will very likely lead to an allocation that is
inequitable because it is divorced from the record evidence and
analytically unsound. When, as a hypothetical example, 100
units of material that costs $1 per unit to remediate are treated
the same as 100 units of material that costs $10 per unit to
remediate, the analysis will be hard to justify.

       That kind of error occurred here and was compounded
when the District Court treated conceptually distinct units of
measurement as equal. It added together data measured in
square feet – a unit of surface area – with data measured in
cubic yards – a unit of volume. Performing such a calculation
was, as Greenlease contends, like comparing “apples to
oranges.”19 (Green. Opening Br. at 53.) Without pure


       19
           “Cubic measures and square measures represent
fundamentally different things. A cubic measure is always a
three-dimensional unit of volume: length times width times
height. A square measure is always a two-dimensional unit of
area: length times width.” Chris Magyar, Cubic Yards to
Square Feet Conversion, SCIENCING (Mar. 13, 2018),




                               40
speculation as to the depths at issue for the square footage
measurements, or record evidence establishing those depths, it
would not have been possible for the District Court to equate
cubic yards to square feet. The Court’s findings of fact and
conclusions of law do not reflect any such analysis.

        Those problematic deviations from Gormley’s
methodology compel us to conclude that there was an abuse of
discretion and that we must vacate the District Court’s
judgment as to the allocation of costs between Greenlease and
Trinity. If the District Court was persuaded by Gormley’s
analytical approach, then, on remand, it should adhere to the
cost allocation methodology he set forth in his expert report –
a methodology that both experts relied upon in coming to their
respective cost allocation estimates. That methodology will
require the Court to conduct a separate cost allocation analysis
for each major remediation activity. Much of the information
needed for that is readily available in the record, but additional
fact-finding by the District Court may be needed.20



https://sciencing.com/cubic-yards-square-feet-conversion-
8641439.html (last visited Aug. 21, 2018).
       20
           We reiterate that any cost allocation methodology
must differentiate between major remediation activities and
account for the varying costs across those activities.
Exactitude is not required. Indeed, at this late date it is
probably not even possible. It is enough for the Court to make
a reasonable estimate of costs based on an appropriate record.
See 

(explaining that the law only
requires that district courts “arrive at an intelligent estimate” of
CERCLA damages “without speculation or conjecture”; it does




                                41
        To apply Gormley’s methodology properly, the District
Court must use volumetric and cost data specific to the
remediation activities. For every major remediation activity,
then, the Court should calculate how much of that activity each
party was responsible for. It can then apply that percentage
breakdown to the total cost of that specific activity at the North
Plant. Once it assigns each party a cost allocation for every
major remediation activity, the Court will be able to add the
parties’ respective shares of costs together. From those totals,
the Court can calculate the overall percentages to use in
determining an equitable allocation of costs between
Greenlease and Trinity. The District Court remains free to
exercise its discretion to adjust those percentages, subject to
the guidance provided herein. It is also free to reopen the
record, should it determine that it is necessary to do so to carry
out the kind of analysis we have described.21
        C. Trinity’s Cross-Appeal

        Trinity raises three primary issues in its cross-appeal.
First, it appeals the District Court’s factual determination of

not require courts to arrive at a “mathematical[ly] precise[]”
figure (citations omitted)).

       21
          Because we must remand this case, we do not address
whether Trinity met its burden to prove damages. However, if
the Court chooses to reopen the record on remand, we
encourage it to permit the parties to address whether some
South Plant costs were impermissibly included in the Court’s
prior allocation of costs at the North Plant. It may also allow
the parties to introduce evidence quantifying the costs incurred
in remediating contamination caused by third parties.




                               42
responsibility for the lead contamination at the North Plant.
We will affirm because we cannot say that the Court abused its
discretion, given the evidentiary record before it. Second,
Trinity challenges the District Court’s decision to grant
Greenlease equitable deductions to account for the
Agreement’s indemnification provisions and for the purported
increase in value of the North Plant following the cleanup. We
agree that the District Court erred in the manner in which it
applied those equitable deductions. We emphasize, however,
that the District Court is free on remand to apply equitable
deductions in accordance with the principles discussed in this
opinion.     Third, Trinity appeals the District Court’s
determination that Ampco is not liable for the conduct of
Greenlease. We will affirm on that point because Trinity
cannot demonstrate that Ampco is either directly or
derivatively liable for Greenlease’s conduct at the North Plant.

              1. The District Court’s Allocation of
                 Responsibility for Lead Contamination was
                 Not an Abuse of Discretion.

       Trinity challenges the District Court’s determination
that Greenlease’s painting operations did not contribute to lead
contamination requiring remediation. It contends that it is
undisputed that Greenlease’s painting operations at the North
Plant resulted in lead runoff seeping into the ground. We
review an allocation of CERCLA damages for abuse of
discretion. Agere Sys., 

. Given the
evidence and expert testimony in the record supporting the
District Court’s determination, we do not agree that there was
an abuse of discretion.




                              43
        The District Court did not, as Trinity suggests,
“disregard the co-contributing effects of Greenlease’s lead
paint releases.” (Trinity Opening Br. at 64.) Rather, as the
Court explained, it found that the historic fill utilized at the
North Plant by various parties over the years was “the source
of the lead contamination that required remediation[.]” (App.
at 402.) In other words, the District Court found that any
contamination by lead paint alone would not have resulted in
contamination requiring remediation.        The Court then
incorporated “the overall percentage of responsibility for the
lead contamination that required remediation” in its equitable
cost allocation analysis. (App. at 402.)

       The District Court’s finding that historic fill and not
lead paint was the source of the contamination requiring
remediation was adequately supported by Greenlease’s expert
Gerritsen. He supported his conclusion by studying soil
samples and observing no correlation between painting
operations and lead contamination. In particular, Greenlease’s
expert observed that lead exceeding PADEP standards was
consistently present in historic fill rather than native soil. That
Trinity’s expert reached a different conclusion – without
conducting an analysis of soil samples – is of no import. The
District Court was entitled to believe Greenlease’s expert
analysis, as it had adequate support to be admissible. See
United States v. Allegheny Ludlum Corp., 

, 184
(3d Cir. 2004) (“[W]hen presented with two sound but
conflicting expert opinions, a district court has discretion to
credit one over the other.”). Accordingly, we will affirm the
conclusion that Greenlease’s paint operations did not result in
lead contamination requiring remediation.




                                44
              2. The District Court Abused Its Discretion
                 When Granting Equitable Deductions
                 Premised on the Indemnification Provisions
                 and the Purported Increased Value of the
                 North Plant.

        CERCLA grants trial courts broad discretion to
“allocate response costs among liable parties using such
equitable factors as the court determines are appropriate.” 42
U.S.C. § 9613(f)(1). “Congress intended to grant the district
courts significant flexibility in determining equitable
allocations of response costs, without requiring the courts to
prioritize, much less consider, any specific factor.” Beazer E.,
Inc. v. Mead Corp., 

, 446 (3d Cir. 2005).
However, “[w]e do not simply ‘rubber-stamp’ a district court’s
equitable allocation[.]” Lockheed Martin Corp. v. United
States, 

, 234 (D.C. Cir. 2016) (citation omitted).
Rather, we review the equitable allocation of environmental
cleanup costs for abuse of discretion. Agere Sys., 

; 

n.18.

       Trinity argues that the District Court’s 5% equitable
deduction in favor of Greenlease due to the contractual
indemnification provisions, and its 10% equitable deduction in
favor of Greenlease due to the purported increased value of the
North Plant, were improper. We agree, and so too does
Greenlease, which acknowledges that the District Court’s
“percentage reductions were completely arbitrary and
speculative.” (Green. Opening Br. at 24.) The District Court
abused its discretion when it applied the 5% equitable
deduction because it erroneously interpreted our precedent. It
also abused its discretion when it applied the 10% equitable




                              45
deduction because it failed to explain how it arrived at that
figure, and we can discern no basis for the figure in the record.

                     i.   The 5% Indemnification Provisions
                          Deduction

        The District Court relied on our opinion in Beazer East,
Inc. v. Mead Corporation when it took into consideration the
Agreement’s indemnification provisions to reduce
Greenlease’s percentage of responsibility by 5%. It concluded
that “it would be error” to not incorporate the parties’ intent, as
manifested by the three-year limit on the indemnification
provisions, into its equitable allocation. (App. at 383.) It
reached that conclusion because, in Beazer, we held that it was
error for a district court to fail to incorporate the relevant
parties’ mutual intent when entering a contract as part of its
equitable 

. In that case, the district
court had failed to give “significant consideration” to the
parties’ intent when equitably allocating CERCLA costs, 

that both parties had intended that the
defendant-seller “would not bear any environmental liability
following the … sale,” 

The district court had
reasoned that, because the contract at issue did not
“demonstrate[] a clear and unambiguous intent to transfer all
CERCLA liability,” as required by the relevant state law, the
parties’ intent to shift liability should be a subordinate factor to
the “polluter pays” principle embedded in CERCLA. 

We said that the district court erred because the legal
interpretation of the contract did not prevent the court from
giving, as a matter of equity, significant consideration to “the
intent of the parties, which [was] manifested by their actions
and in the written agreement[.]” 

                                46
         Critical to our holding in Beazer was the fact that the
district court had determined that both parties expressed a
mutual intent to shift CERCLA liabilities following the
relevant sale. It was only a nuanced application of state
contract law that prevented the parties’ mutual intent from
being enforced as a matter of law. Therefore, in that case,
equity demanded that the district court give significant
consideration to the parties’ shared intent. Here, in contrast,
the District Court’s findings make clear that there was no
mutual intent, as expressed by the written agreement or by the
actions of both parties, to shift CERCLA liability following the
sale of the North Plant. It was, at most, only Greenlease’s
subjective intent to shed all CERCLA liability following the
expiration of the three-year indemnification period. A party’s
subjective intent to avoid liability, which contradicts the
agreement at issue, should not be given significant
consideration when equitably allocating environmental
cleanup costs. Because it appears that the District Court here
mistook Beazer to permit Greenlease’s subjective intent to be
given substantial weight, its 5% equitable deduction in favor
of Greenlease was an abuse of discretion.

        Nothing we have said here should be interpreted as
altering the principle set out in Beazer that, as a matter of
equity, trial courts can take into consideration “the intent of the
parties … [as] manifested by their actions and in the written
agreement[.]” 

But when the intent resulting in the
equitable deduction is not shared by both parties and appears
contrary to provisions of the contract, a district court must
explain why, as a matter of equity, it is nevertheless appropriate
to award an equitable deduction. Because we view the District
Court as having misapplied Beazer, we remand for it to take a
fresh look at whether it is appropriate, on the record before the




                                47
Court, to award Greenlease an equitable deduction premised
on the contractual indemnification provisions.

                   ii.   The 10% Property Value Increase
                         Deduction

       The District Court concluded that a 10% equitable
deduction in favor of Greenlease was appropriate because the
North Plant’s value had increased since the remediation work
transformed the site from being unsuitable for any productive
purpose to being usable as a site for some commercial or
industrial purposes. Although we agree with the District
Court’s identification of the increased value of a remediated
site as an appropriate equitable factor to consider when
allocating cleanup costs, we cannot agree with its application
of that principle here because the record did not contain any
evidence concerning the fair market value of the North Plant,
either before or after the remediation.

        If a landowner successfully seeks contribution from
others for environmental cleanup costs, that owner should
likely be required to share the benefits of any increase in value
brought about by the cleanup. Courts have thus taken the
increased market value of a remediated property into
consideration when allocating response costs. See, e.g., Litgo
N.J. Inc. v. Comm’r N.J. Dep’t of Envtl. Prot., 

,
387 (3d Cir. 2013) (discussing the increased value of
remediated land); Minyard Enters., Inc. v. Se. Chem. & Solvent
Co., 

, 387 (4th Cir. 1999) (directing a lower court
to take into consideration “the fact that the [p]roperty may
appreciate following its remediation”); Farmland Indus., Inc.
v. Col. & E. R.R. Co., 

, 1500-01 (D. Colo.
1996) (concluding that “it would be inequitable” not to take




                               48
into account the fact that the former owner “garner[s] no
tangible benefit from the cleanup of land it no longer owns”).
Limiting a party’s ability to benefit from an economic windfall
comports with “CERCLA’s general policy against double
recovery[.]” 

.

        The problem with the District Court’s 10% deduction,
then, was not in the decision to consider the increased market
value of the North Plant as an equitable factor but rather in the
application of that factor without any record evidence
concerning the North Plant’s value. It is only appropriate to
take increased value into consideration when there is evidence
concerning an actual increase, such as proof of the fair market
value of the property before and after the cleanup. See N.Y.
State Elec. & Gas Corp. v. FirstEnergy Corp., 

,
239 (2d Cir. 2014) (refusing to take into consideration “the
economic benefit of the cleanup” because the party seeking the
equitable deduction “fail[ed] to offer evidence about any
increase in the value of the land”). Because the District Court
may reopen the record for purposes already discussed, 

subsection III.B.3, it may also receive additional
evidence concerning the fair market value of the North Plant
site, both before and after the remediation activities, to allow it
to come to a reasoned percentage reduction premised on the
increased fair market value, if any, of the North Plant site.

              3. The District Court Did Not Err in Deciding
                 that Ampco Is Neither Directly Nor
                 Derivatively Liable for the Contamination at
                 the North Plant.

      Trinity argues that the District Court erred in
determining that Ampco was not liable for Greenlease’s share




                                49
of environmental cleanup costs. As Trinity sees it, the
evidence it presented demonstrated genuine issues of material
fact that were sufficient to entitle it to a trial on the question of
Ampco’s liability. It advances two closely related theories to
support its position that Ampco is legally responsible for
Greenlease’s conduct at the North Plant. First, Trinity
contends that Ampco is directly liable because it qualifies
under CERCLA as an “operator” of the North Plant. Second,
it asserts that, under a veil-piercing theory, Ampco is
derivatively liable for Greenlease’s operation of the North
Plant. Direct and derivative liability are two analytically
distinct bases for holding a parent company liable for
environmental cleanup costs resulting from a subsidiary’s
conduct. United States v. Bestfoods, 

, 67-68
(1998).

        We review the District Court’s grant of summary
judgment de novo. Shelton v. Bledsoe, 

, 559 (3d
Cir. 2015). Summary judgment is appropriate only if, after
drawing all reasonable inferences in favor of the non-moving
party, there exists “no genuine dispute as to any material fact.”
Shuker v. Smith & Nephew, PLC, 

, 770 (3d Cir.
2018) (quoting Fed. R. Civ. P. 56(a)). After our own
independent assessment of the record evidence, we agree with
the District Court that Ampco is not liable for Greenlease’s
conduct at the North Plant, and we will therefore affirm the
grant of summary judgment in favor of Ampco.

                     i.   Ampco Is Not Directly Liable for
                          Greenlease’s Share of Responsibility
                          for Contamination at the North Plant.




                                 50
        CERCLA holds an “operator” of a facility “directly
liable for the costs of cleaning up the pollution.” 

. Direct liability attaches to a parent company
whose subsidiary owns a facility only if the “act of operating a
corporate subsidiary’s facility is done on behalf of a parent
corporation[.]” 

“operate” is read according to its
“ordinary or natural meaning” to refer to “someone who directs
the workings of, manages, or conducts the affairs of a facility.”

(citation omitted). To be directly liable, “an operator
must 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.” 

-67. Whether
Ampco is directly liable, therefore, must be based on its
“participation in the activities of the” North Plant. 

Trinity cannot hold Ampco liable for the environmental
cleanup costs merely by showing that “dual officers and
directors made policy decisions and supervised activities at the
facility.” 

Direct liability will only exist if there
is evidence that Ampco managed the day-to-day activities of
the North Plant in a manner that exceeds “the interference that
stems from the normal relationship between parent and
subsidiary.” 

As the Supreme Court instructed in
United States v. Bestfoods, the relevant inquiry for direct
liability focuses on the relationship between the parent entity
and the polluting facility, not the parent’s relationship to its
subsidiary. 

      The District Court rightly determined that the record
here would not permit a reasonable fact-finder to conclude that
Ampco’s involvement in the day-to-day operations of the
North Plant exceeded “the normal relationship between parent
and subsidiary,” 

in a manner that would support




                               51
holding Ampco directly liable for Greenlease’s conduct. The
undisputed facts establish, rather, that “[Greenlease]
employees were responsible for all day-to-day operations at the
North Plant, including any waste disposal, waste handling,
painting, abrasive blasting, welding, and fabrication
operations.” (App. 81-82.) Greenlease employees, not Ampco
employees, coordinated disposal with outside contractors and
communicated with PADEP on environmental matters. In fact,
Ampco “did not employ any engineers or persons with
technical experience in manufacturing that could make
decisions for [Greenlease] with respect to environmental
compliance or waste management.” (App. at 82.) Instead,
“Ampco employed only a professional staff, such as
accountants, actuaries, and lawyers[.]” (App. at 82.) Helping
with administrative work is consistent with a typical parent-
subsidiary relationship, and certainly does not establish
Ampco’s direct involvement with the North Plant, which
Bestfoods demands to hold a parent directly liable for
environmental cleanup costs.

        Trinity maintains that Ampco crossed the line into
operating the North Plant. According to Trinity, Ampco did so
through individuals who advised Greenlease with regard to
environmental laws and regulations, monitored Greenlease’s
activities, provided Greenlease with legal advice regarding
compliance with environmental laws, and were involved with
Greenlease’s plans to increase the North Plant’s production
capacity and to modernize its operations. Trinity does not,
however, explain how any of those activities, even if one
accepts Trinity’s take on the evidence, turns Ampco’s
supervision of Greenlease into anything other than a typical
parent-subsidiary relationship. Bestfoods makes clear that
“[a]ctivities that involve the facility but which are consistent




                              52
with the parent’s investor status, such as monitoring of the
subsidiary’s performance, supervision of the subsidiary’s
finance and capital budget decisions, and articulation of
general policies and procedures, should not give rise to direct

(citation omitted). That the policies
Ampco advised on may have included environmental issues
does not, on this record, change the calculus.

       Accordingly, we agree with the District Court’s
conclusion that Ampco’s actions with respect to the North
Plant did not fall outside the bounds of typical “parental
oversight of a subsidiary’s facility,” 

are not a
basis for direct liability.




       22
            Trinity argues that “substantial factual similarities”
between Bestfoods and the facts here support its argument that
Ampco is directly liable for Greenlease’s operation of the
North Plant. (Trinity Opening Br. at 78.) It contends that
“[e]very fact referenced by the Supreme Court in Bestfoods has
a parallel in this case.” (Trinity Reply Br. at 20.) We disagree.
In Bestfoods, the Supreme Court vacated a judgment in favor
of a parent corporation because one of its employees “played a
conspicuous part in dealing with the toxic risks emanating from
the operation of the [subsidiary’s] 

.
Here, by contrast, Trinity has not pointed to record evidence
that any officer, director, or employee of Ampco played a
significant, let alone conspicuous, role in the operation of the
North Plant.




                               53
                   ii.   Ampco Is Not Derivatively Liable for
                         Greenlease’s Share of Responsibility.

        A parent corporation can be held derivatively liable
under CERCLA for its subsidiary’s actions “only when[] the
corporate veil may be pierced[.]” 

And “the corporate
veil may be pierced” only in extraordinary circumstances, such
as when “the corporate form would otherwise be misused to
accomplish certain wrongful purposes[.]” 

see also
Wedner v. Unemp’t Comp. Bd. of Review, 

, 794
(Pa. 1972) (“The corporate entity or personality will be
disregarded [o]nly when the entity is used to defeat public
convenience, justify wrong, protect fraud or defend crime.”
(citation omitted)). In such circumstances, the law permits a
subsidiary to be deemed an “alter ego” of its parent so that the
parent can be held liable for the actions of its subsidiary.
Pearson v. Component Tech. Corp., 

, 484 (3d Cir.
2001). Piercing the corporate veil is a limited exception to the
“general principle of corporate law deeply ingrained in our
economic and legal systems that a parent corporation … is not
liable for the acts of its subsidiaries.” 

(internal quotation marks and citations omitted); see also
Lumax Indus., Inc. v. Aultman, 

, 895 (Pa. 1995)
(“[T]here is a strong presumption in Pennsylvania against
piercing the corporate veil.”).

       Trinity seeks to use both federal law and state law to
pierce the corporate veil. The federal law principles we have
articulated for when a subsidiary is merely an alter ego of its
parent are substantially similar to the principles set forth in
Pennsylvania case law. Our analysis of both, therefore, can
largely proceed in tandem, though we do specifically note
Trinity’s state law-specific arguments. Under either theory,




                              54
Trinity has failed to adduce sufficient evidence to create a
triable issue of fact.

        We have identified several factors helpful in
determining whether, as a matter of federal common law, a
subsidiary is merely an alter ego of its parent. Those factors
include “gross undercapitalization, failure to observe corporate
formalities, nonpayment of dividends, insolvency of
[subsidiary] corporation, siphoning of funds from the
[subsidiary] corporation by the dominant stockholder,
nonfunctioning of officers and directors, absence of corporate
records, and whether the corporation is merely a façade for the
operations of the dominant stockholder.” 

.23 No single factor is dispositive, and we consider
whether veil piercing is appropriate in light of the totality of
the circumstances. Cf. Trs. of Nat’l Elevator Indus. Pension,
Health Benefit & Educ. Funds v. Lutyk, 

, 194 (3d
Cir. 2003) (explaining that the alter ego test factors do not
comprise “a rigid test”); Am. Bell Inc. v. Fed’n of Tel. Workers
of Pa., 

, 887 (3d Cir. 1984) (requiring “specific,
unusual circumstances” before piercing the corporate veil
(citation omitted)).24

       23
            Factors considered by Pennsylvania state courts
include “undercapitalization, failure to adhere to corporate
formalities, substantial intermingling of corporate and personal
affairs and use of the corporate form to perpetrate a fraud.”

.
       24
          See also Advanced Tel. Sys., Inc. v. Com-Net Prof’l
Mobile Radio, LLC, 

, 1281 (Pa. Super. Ct. 2004)
(looking to the “totality of circumstances” when conducting
corporate veil piercing analysis).




                              55
         Proving that a corporation is merely an alter ego is a
burden that “is notoriously difficult for plaintiffs to meet.”

. “[I]n order to succeed on
an alter ego theory of liability, plaintiffs must essentially
demonstrate that in all aspects of the business, the two
corporations actually functioned as a single entity and should
be treated as such.” Id.25 Under our precedent, the basis for
piercing the corporate veil must be “shown by clear and
convincing evidence.” 

(quoting Kaplan
v. First Options of Chi., Inc., 

, 1522 (3d Cir.
1994)).26


       25
           See also E. Minerals & Chem. Co. v. Mahan, 

, 333 n.6 (3d Cir. 2000) (explaining that Pennsylvania
law “require[s] a threshold showing that the controlled
corporation acted robot- or puppet-like in mechanical response
to the controller’s tugs on its strings or pressure on its buttons”
before allowing a plaintiff to pierce the corporate veil (citation
omitted)); Culbreth v. Amosa (Pty) Ltd., 

, 14 (3d
Cir. 1990) (interpreting Pennsylvania law to require that “the
party seeking to pierce the corporate veil on an alter-ego theory
establish[] that the controlling corporation wholly ignored the
separate status of the controlled corporation and so dominated
and controlled its affairs that its separate existence was a mere
sham”).
       26
           Trinity presents three arguments for why the District
Court erred by applying the clear and convincing evidence
standard to its alter-ego analysis: that it is always improper for
a district court to apply that standard to a motion for summary
judgment, that federal law does not incorporate the clear and




                                56
       Trinity appears to agree that most of the traditional
factors we look to when determining whether to pierce the
corporate veil are either inapplicable to this case or favor
Ampco. Its primary arguments for piercing the corporate veil
are that “Greenlease became undercapitalized when Ampco


convincing standard when the plaintiff does not allege fraud,
and that Pennsylvania applies a preponderance of the evidence
standard to its alter-ego analysis.
       First, Trinity is incorrect as a matter of law that “under
no circumstances” is a clear and convincing standard
“appropriate for summary judgment purposes.” (Trinity
Opening Br. at 68.) “[T]he determination of whether a given
factual dispute requires submission to a jury must be guided by
the substantive evidentiary standards that apply to the case.”
Anderson v. Liberty Lobby, Inc., 

, 255 (1986).
“Consequently, where the clear and convincing evidence
standard applies, the trial judge [at summary judgment] must
inquire whether the evidence presented is such that a jury
applying that evidentiary standard could find only for one
side.” Justofin v. Metro. Life Ins. Co., 

, 522 (3d
Cir. 2004). Second, our precedent is clear, as a matter of
federal common law in this Circuit, that “[b]ecause alter ego is
akin to and has elements of fraud theory, … it … must be
shown by clear and convincing evidence.” 

(citation omitted). Trinity has not presented any
compelling argument to revisit that longstanding proposition.
Third, we do not need to address the standard of proof we think
Pennsylvania applies to its alter-ego analysis because, whether
we apply a preponderance of the evidence standard or a clear
and convincing evidence standard to the state law analysis, our
ultimate conclusion is the same – no reasonable fact-finder
could justify piercing the corporate veil on this record.




                               57
siphoned off Greenlease’s assets,” that “Ampco and
Greenlease’s interactions exceeded norms that characterize
parent/subsidiary relationships,” (Trinity Opening Br. at 74),
that the equities tilt in its favor under Pennsylvania’s alter-ego
test, and that public policy favors holding Ampco responsible.

                             a. Greenlease      Was    Not
                                Undercapitalized and Ampco
                                Did Not Siphon Funds From
                                Greenlease.

         Trinity argues that Greenlease’s issuing to Ampco some
$50 million dollars in dividends in the years following the sale
of the North Plant, leaving only $250,000 in reserve for
liabilities, favors piercing the corporate veil. But “the inquiry
into corporate capitalization is most relevant for the inference
it provides into whether the corporation was established to
defraud its creditors or [an]other improper purpose such as
avoiding the risks known to be attendant to a type of business.”

. There is no basis in the record to
suggest that Greenlease was undercapitalized while operating
the North Plant. Instead, Trinity suggests only that Greenlease
lacked funds after Greenlease’s operations of the North Plant
had effectively stopped. That is of “little relevancy to
determining whether piercing the corporate veil [is] justified
here.” 

also no evidence that Greenlease issued
dividends to Ampco with awareness of its liability to Trinity or
to escape subsequent liability. See Zubik v. Zubik, 

, 273 (3d Cir. 1967) (“Unless done deliberately, with
specific intent to escape liability for a specific tort or class of
torts, the cause of justice does not require disregarding the




                                58
corporate entity.”). As the District Court noted, “it would be
unreasonable for Ampco to leave Greenlease’s earnings from
the sale of the North Plant in an account when at the time the
dividends were issued Greenlease was a nonoperating
company with no known liabilities.” (App. 91 (emphasis
removed).)

                            b. Greenlease and Ampco’s
                               Relationship Was a Typical
                               Parent-Subsidiary
                               Relationship.

       Trinity emphasizes that there was significant overlap
between the boards of Ampco and Greenlease and argues that
Ampco dominated Greenlease to an unusual extent. But
“duplication of some or all of the directors or executive
officers” is not fatal to maintaining legally distinct corporate
forms. 

(citation omitted); see also
Am. 

(noting that “there must be specific,
unusual circumstances” to justify veil piercing, and mere
control and participation in management is inadequate).
Greenlease ran the North Plant and hired all of the employees
on the ground. Although Ampco was required to approve large
decisions, Greenlease generally functioned with autonomy on
decisions     concerning      manufacturing,     environmental
compliance, and disposal of waste. We have already said and
now repeat that the District Court rightly determined that the
record simply does not support Trinity’s position that
Greenlease’s relationship with Ampco was materially different
than a normal parent-subsidiary relationship.




                              59
                            c. Trinity Cannot Pierce the
                               Corporate     Veil  Under
                               Pennsylvania Law.

        Trinity argues that the District Court erred by
disregarding the “equitable underpinnings” of Pennsylvania’s
alter-ego framework. (Trinity Reply Br. at 6.) It maintains that
Pennsylvania disregards the legal fiction of separate corporate
entities “whenever justice or public policy demand[s]” it.
(Trinity Reply Br. at 7 (quoting Ashley v. Ashley, 

,
641 (Pa. 1978).) According to Trinity, permitting Ampco to
reap the benefits of the over $50 million in dividends from
Greenlease without being held accountable for Greenlease’s
conduct is an injustice.         But Trinity overlooks that
Pennsylvania requires a plaintiff seeking to pierce the
corporate veil to make “a threshold showing that the controlled
corporation acted robot- or puppet-like in mechanical
response” to the controlling shareholder’s demands. E.
Minerals & Chem. Co. v. Mahan, 

, 333 n.6 (3d
Cir. 2000) (citation omitted). Trinity has not made that
showing here.

        Pennsylvania law is also clear that courts are not to
disregard the legal fiction of separate corporate entities if it
would render “the theory of the corporate entity … useless.”

; see also 

(“Care should be taken on all occasions to avoid making the
entire theory of the corporate entity … useless.” (internal
quotation marks omitted) (quoting 

)).
To permit Trinity to pierce the corporate veil in this instance,
in the face of all the objective criteria favoring Ampco, would,
in essence, result in rendering useless Ampco’s legitimate use
of the corporate form when setting up Greenlease as a




                              60
subsidiary. The record is devoid of evidence that Ampco
misused separate corporate entities for some nefarious
purpose. To pierce the corporate veil would thus fly in the face
of Pennsylvania’s “strong presumption … against piercing the
corporate veil.” 

.

                             d. Public Policy Considerations
                                Do Not Favor Trinity.

         Finally, Trinity argues that the District Court failed to
consider public policy justifications for piercing the corporate
veil to ensure that the “polluter pays.” (Trinity Opening Br. at
74.)     As discussed above, however, both federal and
Pennsylvania law favor maintaining the legal fiction of
separate corporate entities. Because the evidence does not
suggest that there was fraud or an attempt to use a corporate
façade as an alter ego, public policy first favors upholding the
integrity of the corporate form. Trinity has not presented any
public policy consideration sufficiently compelling to
overcome the strong presumption against veil piercing.

IV.    CONCLUSION

       For the foregoing reasons, we will affirm in part but will
vacate the District Court’s cost allocation determination and
remand for further proceedings consistent with this opinion.




                               61