390 F.3d 433 (6th Cir. 2004)
In the 1960s, GenCorp, Inc. and Olin Corporation began a business relationship that defies easy categorization. By agreement of the parties, Olin built a manufacturing plant with GenCorp’s assistance that for more than a decade supplied Gen-Corp with a chemical used to produce urethane foam. The manufacturing plant also produced hazardous waste. Plant operators hauled the waste to an offsite landfill, which eventually landed on the EPA’s National Priority List of hazardous waste facilities and which generated clean-up costs of more than $65 million.
Olin sued GenCorp under the Comprehensive Environmental Response, Compensation and Liability Act (CERCLA), 42 U.S.C. § 9601 et seq., to recover a percentage of the clean-up costs and to obtain a declaratory judgment that GenCorp bore responsibility for that same percentage of future costs. The district court granted the relief, concluding that GenCorp was a “[c]overed person[]” under § 9607(a) of the statute because it “arranged for disposal ... of hazardous substances” and concluding that it was responsible for 70% of the clean-up costs at the landfill. We agree with that conclusion and reject each of the parties’ other challenges to the district court’s decision except one: Olin’s claim that a declaratory judgment should have been entered. Accordingly, we affirm the district court’s judgment as to GenCorp’s contribution liability and the apportionment of liability between the parties but remand the case to consider Olin’s request for a declaratory judgment.
I.
A.
During the 1960s, GenCorp (once known as “General Tire and Rubber Company”) manufactured urethane foam — a material used in carpets and mattresses, among other products — at a plant (the “PVC Plant”) located in Ashtabula, Ohio. Among other ingredients, the production of the foam required toluene di-isocyanate (“TDI”), which is a “critical” chemical ingredient in the process. JA 1319. To minimize the financial risks of urethane production, GenCorp sought a reliable and cost-effective source of TDI.
Olin Corporation claimed to have a supply answer to this demand. Having developed a cost-effective process for producing TDI, which involved highly toxic chemicals, including phosgene or “mustard” gas, it began discussions with GenCorp about building and operating a plant for manufacturing the chemical (the “TDI Plant”). On June 21, 1962, after lengthy negotiations, the parties concluded a final agreement (the “1962 Agreement”).
As the contractual cornerstone of the 1962 Agreement, Olin agreed to “erect or cause to be erected, at its sole cost and expense” the TDI Plant. JA 660. The parties agreed to locate the plant on land owned by GenCorp, which was adjacent to GenCorp’s own PVC Plant (where the TDI from Olin’s new plant would eventually be converted into urethane foam), and agreed that Olin would pay GenCorp $10 a year for the lease of the land. Olin also agreed to build a second plant at its own expense to produce toluene diamine or “TDA” — a chemical needed in the TDI manufacturing process — and to supply enough TDA to meet the TDI Plant’s requirements.
In return, GenCorp agreed to purchase 50% of the TDI Plant’s output at cost, which included variable costs, fixed costs for the TDI Plant and “related facilities,” and a “charge-in” for the TDA supplied by Olin to the plant. JA 671. Both parties remained free to sell TDI to third parties without sharing any profits from these sales. GenCorp also agreed to purchase hydrochloric acid — an additional byproduct of the TDI manufacturing process — from Olin and to supply steam from its PVC Plant to meet the TDI Plant requirements.
Olin initially retained title to the TDI Plant under the 1962 Agreement and “owned and operated” the plant. JA 660, 663. The parties agreed that at some later date Olin would “sell, convey and transfer” title to the TDI Plant to GenCorp at Olin’s then-current book value for the plant. JA 678-79. GenCorp agreed to advance Olin up to one half of its total capital costs for the plant as “earnest money [] towards its obligation to purchase and acquire the TDI Plant.” JA 680. That amount, it was agreed, would be “credited by Olin against the ultimate sale and purchase price” due at the future sale.
The parties affirmed in a separate agreement that Gen-Corp had a “definite obligation to take title” to the TDI Plant and had committed itself to this purchase through a “substantial deposit against the ultimate purchase price.” JA 699. In accordance with these future plans for ownership of the plant, the engineering specifications for the plant were “subject to the approval” of both GenCorp and Olin, JA 662, and Olin agreed not to make any capital expenditures without GenCorp’s approval after the plant’s start-up. Olin also agreed to insure the TDI Plant for “all risk of loss, liability and damage to property and facilities.” JA 679.
To the ends of “oversee[ing] the construction, operation and management of the TDI plant,” the contract created a four-member committee (the “TDI Committee”) composed of two representatives of GenCorp and two representatives of Olin. In consultation with the committee, Olin agreed to select a Plant Manager and other staff “to supervise the day-to-day-operation of the plant.” JA 663. Olin also agreed to appoint a Committee Manager, who was the sole party authorized to instruct the Plant Manager on behalf of the Committee. The 1962 Agreement contained no guidelines for committee governance — i.e., specifying whether decision-making would occur by consensus or by majority vote, or how a deadlock would be resolved — but rather allowed the Committee to devise its own management rules (which apparently were never created).
Although not detailed in the agreement, GenCorp supplied all of the TDI Plant’s hourly workers, keeping them on its own payroll and negotiating a collective bargaining agreement for them. Olin, on the other hand, supplied and paid many of the plant’s salaried, supervisory employees, including the Plant Manager and other departmental heads, although GenCorp filled some management positions with its own employees. All employees ultimately reported to the Olin-appointed Plant Manager.
The 1962 Agreement did not address hazardous waste disposal. The TDI manufacturing process generated at least two byproducts that amount to “hazardous waste” for CERCLA purposes: a tar-like residue waste and spent vacuum-pump oils, both of which contain the toxic chemicals TDI, TDA and chlorobenzene. Early plant designs contemplated generation and disposal of the TDI residue. The TDI Plant engineer told Phillip Sayre, a Gen-Corp employee who reviewed the design specifications, that the TDI chemical residue would be “drum[med][] off and bur[ied].” JA 1323. Despite Sayre’s concerns about the dangers of this plan, which he conveyed to his supervisors at Gen-Corp, GenCorp approved the plant designs and capital appropriation requests by Olin.
After approving the plant designs, the TDI Committee discussed TDI residue disposal and the GenCorp committee members researched and recommended offsite locations for disposal sites in view of their greater familiarity with the Ashtabula area. D. Ct. Op. of July 14, 1999, at 40 [hereinafter D. Ct. Op.]. TDI employees disposed of some waste at the plant site itself, which the parties agree caused a “release” into the waterway known as “Fields Brook” that ran between the TDI and PVC Plants on GenCorp’s land (a site that is the subject of this litigation but not this appeal).
Benjamin Brenkus, a hauler apparently hired by the Olin Plant Manager, carried other waste offsite. From the beginning of plant operations in 1964 through February 1976, Brenkus hauled waste to the Big D site, all of which the TDI Plant hourly employees had placed in metal drums and loaded onto Brenkus’ truck for disposal. D. Ct. Op. at 41. Payment for Brenkus’ hauling efforts required completion of a service requisition form, which on occasion GenCorp employees in intermediate management positions signed.
Over the years of the TDI Plant’s operation, the Committee addressed the issue of waste disposal in a variety of ways. TDI residue disposal (i.e., the hauling costs) appeared as a line item on Olin’s budget for the TDI Plant, which the Committee approved, and which GenCorp paid for as part of the variable costs incorporated into the TDI price. The Committee approved (apparently without dissension) capital expenditures aimed at reducing the volume of waste generated by the plant. D. Ct. Op. at 41. GenCorp Committee member Robert Laundrie visited the Big D site more than once, and he issued his criticisms of disposal methods to Olin’s Plant Manager, at 40-41 (although the record contains no evidence as to whether Olin acted upon these concerns). And GenCorp and Olin jointly funded “Golden TDI” research, a program aimed (fruitlessly, it turns out) at finding commercial uses for the TDI residue in order to reduce the amount of waste shipped offsite. at 41.
Despite the title-transfer provision of the 1962 Agreement, GenCorp never purchased the TDI Plant from Olin. Instead, on September 27, 1971, the parties signed a new agreement (the “1971 Agreement”) that terminated the 1962 Agreement effective October 1, 1973. In the new agreement GenCorp relinquished its right to purchase the TDI Plant in exchange for a $1.65 million payment from Olin as reimbursement for its earlier “earnest money.” JA 929. GenCorp also extended its lease of the land on which the plant was located for three additional twenty-year terms. In a separate agreement signed on the same day, the parties agreed that GenCorp would purchase “not less than 80%” of its yearly TDI requirements at a price determined under a new formula supplied by the agreement (essentially the lesser of market price or cost). JA 931. At the same time they signed these new agreements, the parties unwound earlier employment arrangements by transferring the GenCorp employees working at the TDI Plant to Olin’s payroll and by dissolving the TDI Committee. D. Ct. Op. at 15.
B.
In the late 1970s, Olin began investigating toxic contamination at the Big D site, identifying the property as an “area of concern requiring further study.” D. Ct. Op. at 16. In May of 1980, the Ohio EPA requested information from Olin about disposal of waste from the plant. And in June of 1981, the matter came before the federal EPA, after Olin filed a notification required by CERCLA that it had disposed of 28,000 tons of waste at the Big D site. The TDI Plant ceased operations later that year, and Olin began dismantling the facilities as well as performing clean up and decontamination of the plant site.
At its own initiative, but in consultation with state environmental officials, Olin performed an erosion-control project at the Big D site in 1982. The project increased the slope gradient of the landfill through the addition of topsoil and clay in order to redirect storm water run-off, at a cost of over $780,000, D. Ct. Op. at 18— none of which Olin seeks to recover from GenCorp. Olin informed the state and federal environmental agencies about its efforts, claiming that “the [Big D site] is not a significant threat to human health or the environment and no further corrective action is required or warranted.” JA 1110.
The federal EPA disagreed with this assessment. On September 8, 1983, after a notice-and-comment period, it placed both the Big D and Fields Brook sites on the National Priority List for clean-up. In 1985, the federal agency notified Olin that it was a “potential responsible party” for the site and requested that Olin provide information concerning the site and conduct a “remedial investigation and feasibility study” for the landfill. D. Ct. Op. at 19. Negotiations for a remedial action plan for Big D took place during 1989 and 1990. When Olin refused to yield on certain points in the negotiations, the federal EPA issued a unilateral administrative order on March 27, 1990, under CERCLA § 106, 42 U.S.C. § 9606, requiring Olin to implement the agency’s remedial action plan for the site. D. Ct. Op. at 20-21. Olin formed a “Big D Site team” and in late 1992 began implementing the remedial action plan. at 21-23.
C.
Anticipating that Olin would seek contribution for these remediation costs, Gen-Corp filed a declaratory judgment action against Olin on October 25, 1993, in the United States District Court for the Northern District of Ohio. GenCorp sought a declaration that it was not liable under CERCLA for the clean-up costs at Big D and Fields Brook, among other sites. Olin counterclaimed, seeking its own declaration either that GenCorp was jointly and severally liable with Olin for the response costs or owed Olin contribution for at least one half of the costs. The parties voluntarily dismissed GenCorp’s claim in 1995, but continued litigating Olin’s counterclaim. GenCorp eventually amended its response to Olin’s counterclaim, claiming “breach of contract to insure” the TDI plant, among other claims. JA 161.
By consent of the parties, the district court transferred the case to Magistrate Judge David Perelman for trial. The court initially denied both parties’ motions for summary judgment, determining that genuine issues of fact precluded any pretrial disposition of the claims. Before trial, however, the district court reconsidered its decision in light of intervening case law and entered summary judgment against Olin on one of its counterclaims. It determined that Olin could not maintain its claim for joint and several liability under CERCLA § 107 because it was a potentially liable party itself and thus could seek only contribution from GenCorp.
With respect to Olin’s remaining contribution claim under § 113(f), the district court conducted trial proceedings in three phases. Phase I dealt with the initial question of liability; Phase II allocated liability between the parties on a percentage basis; and Phase III assigned a dollar amount to these percentages. After a bench trial addressing the Phase I and II issues, the district court determined that GenCorp was liable as an “arranger” of hazardous waste disposal under CERCLA either by virtue of its affiliation with Olin in an “association” or “commercial entity” (but not a joint venture) — such that the two companies constituted a single “person” under 42 U.S.C. § 9601 — or by virtue of its activities carried out in its own name. D. Ct. Op. at 47. The court also determined that GenCorp was liable as an “owner” and an “operator” of the land on which the plant was built. In addressing the parties’ respective shares of liability, the district court concluded that GenCorp should bear 30% of the costs of Big D (and Olin 70%) and GenCorp should bear 40% of the costs of the Fields Brook site (and Olin 60%). D. Ct. Op. at 60. During Phase III, the district court determined that these liability shares translated into costs to GenCorp of approximately $19 million (plus more than $9 million in prejudgment interest) and entered a judgment in that amount. D. Ct. Op. of May 10, 2002 at 77-78; Judgment Order at JA 597.
The district court did not finally resolve GenCorp’s breach-of-contract claims based on Olin’s alleged promise to insure. It tabled the ultimate resolution of the claim in light of a pending dispute in the Southern District of New York between Olin and its insurers over the insurers’ responsibility to pay for the hazardous waste clean-up costs. The court determined that Olin had satisfied its contractual obligation to Gen-Corp, and if Olin obtained any recovery from its insurers in the New York proceeding, that recovery would “inure to the benefit of ... [GenCorp]” as well. D. Ct. Op. of May 10, 2002, at 77. Because that parallel proceeding had not concluded, the district court “h[e]ld in abeyance a determination as to what share of any recovery [in the New York litigation] would be Gen-Corp’s,” but entered a final judgment in the remainder of the suit under Federal Rule of Civil Procedure 54(b).
The appeal and cross-appeal from the district court’s judgment raise a host of questions: (1) does this Court have jurisdiction over the suit? (2) does the statute of limitations bar Olin’s claim? (3) is Gen-Corp liable for response costs related to the Big D site as an “arranger” for the disposal of hazardous waste generated by the TDI Plant? (4) did the district court abuse its discretion in determining each parties’ proportionate responsibility for the clean-up costs? (5) did the district court err in awarding prejudgment interest? and (6) should the district court have granted Olin’s request for declaratory relief regarding future clean-up costs?
II.
As a matter of first priority, the Court has jurisdiction over this appeal. Under Rule 54(b) of the Federal Rules of Civil Procedure, a district court may, “[w]hen more than one claim for relief is presented in an action, ... direct the entry of a final judgment as to one or more but fewer than all of the claims,” provided that the court makes “an express determination that there is no just reason for delay and upon express direction for the entry of judgment.” Compliance with the rule permits immediate appellate review of a district court’s judgment even though the lawsuit contains unresolved claims.
To comply with the rule, a district court initially must expressly “direct the entry of final judgment as to one or more but fewer than all of the claims” in the case. It then must determine whether “the needs of the parties” outweigh the efficiency of having one appeal at the conclusion of the case in its entirety, and it must spell out its reasons for concluding that prompt review is preferable. On appeal, we review compliance with the first requirement de novo, and compliance with the second requirement for an abuse of discretion; see also Curtiss-Wright Corp. v. Gen. Elec. Corp., 446 U.S. 1, 10 (1980).
As to the first requirement, this case involves more than one claim and the district court permissibly entered final judgment as to all but one of the claims— GenCorp’s contract claim that Olin failed to obtain insurance for the TDI Plant. A “claim” under Rule 54(b) “denotes the aggregate of operative facts which give rise to a right enforceable in the courts” even if the party has raised different theories of relief. Gen. Acquisition, Inc. v. GenCorp, Inc., 23 F.3d 1022, 1028 (6th Cir.1994) (quotation marks omitted). In this case, the district court had before it Olin’s contribution claim as well as GenCorp’s breach-of-contract and other claims. Olin’s contribution claim arises under CERCLA while GenCorp’s contract claim arises under state law governing contractual rights and fiduciary responsibilities. The claims also do not share a single “aggregate of operative facts.” Olin’s CERCLA claim arises from the disposal of hazardous waste at the Big D landfill, while GenCorp’s claim arises out of Olin’s alleged duty to insure for risks related to the TDI Plant. Nor does the fact that Olin’s recovery on its claim could be set off in full or in part by GenCorp’s recovery on its claim impede the separability of these claims for appeal purposes. See Curtiss-Wright, 446 U.S. 1 (reversing court-of-appeals determination that certification of claim was improper because pending counterclaim could ultimately provide set-off to liability). Lastly, the district court’s resolution of the CERCLA claim satisfies the finality requirement of Rule 54(b) and 28 U.S.C. § 1291, as the court provided an “ultimate disposition” of the claim and addressed all forms of relief requested by Olin. Compare Liberty Mut. Ins. Co. v. Wetzel, 424 U.S. 737 (1976) (concluding that district court’s finding of liability, which did not address any of the relief requested by the plaintiff, did not satisfy Rule 54(b)); Gen. Acquisition, 23 F.3d at 1027 (determining that district court’s judgment resolving damages issue but not liability was not “final”).
As to the second requirement, the district court did not abuse its discretion in concluding that it would be efficient presently to dispose of all of the claims, save for GenCorp’s insurance-contract claim. In arguing to the contrary, GenCorp contends that the district court “failed to make the required findings or to balance the ... factors that should be considered.” GenCorp’s Mot. to Dismiss at 2. In fact, however, the district court did explain why there was no “just reason for delay.” D. Ct. Op. of Nov. 21, 2002, at 10. And its explanations for certifying the partial judgment for appeal in this decade-long suit made ample sense given that the only question remaining on GenCorp’s claim was one that would be resolved by another court and given the likelihood that Gen-Corp would appeal its CERCLA liability (in view of its possible preclusive effect) even if the amount owed was wholly offset by an insurance recovery.
III.
GenCorp next argues that Olin failed to comply with the applicable statute of limitations in bringing its CERCLA contribution claim. We disagree.
As an initial point of reference, the parties share common ground that § 113(g)(2) of CERCLA, 42 U.S.C. § 9613(g)(2), applies to this action. They agree with the district court that § 113(g)(2) — not § 113(g)(3) (entitled “Contribution”) — applies to this case because Olin incurred clean-up costs at Big D not as the result of a court judgment or settlement with the government but as the result of a unilateral administrative order issued by the EPA under § 9606. For this reason, the district court concluded, Olin’s counterclaim constituted “an initial action for recovery of [] costs” under § 113(g)(2).
Having agreed that § 113(g)(2) governs this case, the parties dispute which portion of § 113(g)(2) applies. Section 113(g)(2) contains two different time limitations, and provides:
An initial action for recovery of [] costs ... must be commenced — (A) for a removal action, within 3 years after completion of the removal action ...; and (B) for a remedial action, within 6 years after initiation of physical on-site construction of the remedial action, except that, if the remedial action is initiated within 3 years after the completion of the removal action, costs incurred in the removal action may be recovered in the cost recovery action.
42 U.S.C. § 9613(g)(2). In GenCorp’s view, the actions taken by Olin in the early 1980s at the Big D site — changing the slope of the land and adding topsoil at a cost of $780,000 — triggered the “initiation” of “remedial action” for the statute of limitations, which means that the time for bringing a remedial-recovery lawsuit has long since passed. In Olin’s view, the limitations period did not begin running in the 1980s (1) because these earlier efforts constituted “removal” but not “remedial” actions and (2) because the work done in the early 1980s was “separate and distinct from, and could not start the limitations period for, the work done in the 1990s in response to the EPA’s orders” even if the earlier work was “remedial” in nature. Olin Br. at 53. (Olin also argues that GenCorp waived its statute-of-limitations defense, but we need not reach that question.)
In resolving this issue, the district court looked to the National Contingency Plan regulations defining removal actions, and determined that Olin’s “work performed in the early 1980s falls within the parameters of removal actions.” D. Ct. Op. at 28. The district court noted that the short duration of the project and its relatively low cost indicated that the action constituted removal, not remediation. The work done in the 1980s, the court added, did not meet the test for remedial actions because it took place before the “remedial investigation and feasibility study” and before the EPA’s unilateral administrative order implementing the final plan for the site. at 29.
We agree with this analysis — with one elaboration. The salient point here is that CERCLA’s definition of “remedial action” imposes a separate requirement on the term: “remedial action” encompasses only “those actions consistent with [the] permanent remedy taken ... to prevent or minimize the release of hazardous substances,” and unless “remedial action” meeting this requirement has begun, the statute-of-limitations-triggering event cannot occur. 42 U.S.C. § 9601(24) (emphasis added). Critically, Olin’s earlier actions at the landfill were not consistent with the “permanent remedy” for the site — namely, the final remedial plan implemented by the federal EPA. Indeed, in adding topsoil and clay covers to prevent soil erosion and in changing the slope gradient of the site, Olin’s interim measure was inconsistent with the federal EPA’s remedial plan. The final plan called for Olin to dismantle its previous work at the site by removing the clay cap, excavating the site and removing and incinerating waste and contaminated soil from the site. D. Ct. Op. at 29. For this reason, “the prior work was not ‘permanent’ as would be required for a remedial action, nor was it consistent with the remedy mandated by the EPA in the 1990 [unilateral order].”
In addressing this issue, the Ninth Circuit has concluded that “the initiation of physical on-site construction of the remedial action” under 42 U.S.C. § 9601(24) and § 9613(g)(2) “can only occur after the final remediation plan is adopted” and that action taken before the plan adoption cannot constitute “remedial action” for statute-of-limitations purposes. California ex rel. Cal. Dep’t of Toxic Substances Control v. Neville Chem. Co., 358 F.3d 661, 667 (9th Cir.2004) (emphasis added); see also 42 U.S.C. § 9617 (outlining procedures for proposing and adopting a final remedial action plan). While other circuits have rejected this bright-line rule, they adhere to the principle that for a “remedial action” to begin, the work must be “consistent with [the] permanent remedy.” See Geraghty & Miller, 234 F.3d 917; United States v. Navistar Int’l Trans. Corp., 152 F.3d 702 (7th Cir.1998). We need not decide today whether to adopt the Ninth Circuit’s bright-line rule because it would not affect the outcome of this case. In this instance, Olin’s work in the 1980s at Big D was not consistent with the EPA-ordered final remedial plan, irrespective of the timing of the work, and GenCorp offers no argument otherwise. Under these circumstances, Olin’s actions in the 1980s did not trigger the statute of limitations for recovery of its costs related to its “remedial” actions over a decade later.
IV.
Turning to Olin’s contribution claim, the parties agree that several preconditions for CERCLA liability have been established: (1) the Big D site is a “facility” under 42 U.S.C. § 9601(9); (2) a “release” of “hazardous substances” (the TDI residue) occurred; and (3) the release caused Olin to incur “necessary costs of response,” see 42 U.S.C. § 9607(a). See Stip. of Parties at JA 121; Centerior, 153 F.3d at 347-48. The only element of CERCLA liability in dispute is whether GenCorp falls into one of four categories of “potentially responsible parties” described in CERCLA § 107(a)(1)-(4).
Invoking § 107(a)(3) of CERCLA, the district court concluded that GenCorp was an “arranger” of hazardous waste disposal: a “person who by contract, agreement, or otherwise arranged for disposal or treatment, or arranged with a transporter for transport for disposal or treatment, of hazardous substances owned or possessed by such person, by any other party or entity, at any facility or incineration vessel owned or operated by another party or entity and containing such hazardous substances.” 42 U.S.C. § 9607(a)(3). In reaching this conclusion, the district court embraced two alternative theories of “arranger” liability: (1) that the relationship Olin formed with GenCorp created a single “person” under 42 U.S.C. § 9601(21) — as an “association” or a “commercial entity” — such that Olin’s actions regarding hazardous waste disposal could be attributed to GenCorp; and (2) GenCorp “arranged for disposal” of hazardous waste in its own right as a separate potentially liable “person.” D. Ct. Op. at 30, 47.
GenCorp challenges each of these conclusions on appeal. Because we conclude that GenCorp “arranged for” hazardous waste disposal in its own capacity, we need not address the separate question whether the two companies formed an “association” or “commercial entity” under 42 U.S.C. § 9601(21).
A.
CERCLA imposes liability on any person who “arrange[s]” “by contract, agreement or otherwise” for the “disposal or treatment ... [or] for transport for disposal or treatment” of “hazardous substances” that is “owned or possessed” by that person. 42 U.S.C. § 9607(a)(3). While the legislation does not define what it means to “arrange” for disposal of waste, traditional definitions of the word, its statutory context and case law supply ample clues. To “arrange” for something means to “plan or prepare” for it, though not necessarily to implement the plan. Webster’s II New College Dictionary 62 (2001). Making such preparations, moreover, does not require a formal disposal agreement, as the statute provides that a person may arrange for hazardous waste disposal “by contract, agreement or otherwise.” 42 U.S.C. § 9607(a)(3) (emphasis added). Neither must the arrangements for waste disposal stem from a discrete event; they may arise from a broader “transaction.” See United States v. Cello-Foil, Prods., Inc., 100 F.3d 1227, 1231 (6th Cir.1996).
Circuit precedent also makes clear that one may not become an arranger through inadvertence. The party must have some intent to make preparations for the disposal of hazardous waste, though that intent goes to the matter of disposing waste generally, not to disposing of it in a particular manner or at a particular location. See id. at 1231-32 (“[O]nce it has been demonstrated that a party possessed the requisite intent to be an arranger, the party cannot escape liability by claiming that it had no intent to have the waste disposed in a particular manner or at a particular site.”); United States v. Aceto Agricultural Chems. Corp., 872 F.2d 1373, 1380 (8th Cir.1989); see also CPC Int’l, Inc. v. Aerojet-Gen. Corp., 759 F.Supp. 1269, 1279 (W.D.Mich.1991); cf. United States v. Township of Brighton, 153 F.3d 307, 315 (6th Cir.1998) (noting that it is “no defense” to operator liability that the “actor [was not] responsible for a particular hazard”). The requisite intent may be “inferred from the totality of the circumstances,” Cello-Foil, 100 F.3d at 1231, and need not depend on the parties’ own characterizations of their transactions, Aceto, 872 F.2d at 1381 — all of which makes this inquiry highly fact-driven, Cello-Foil, 100 F.3d at 1231.
Measured by these requirements, the district court correctly concluded that Gen-Corp arranged for the disposal of hazardous waste from the TDI Plant. GenCorp and Olin entered into the 1962 Agreement to build a manufacturing plant that would convert TDA (which Olin would produce) into TDI (which GenCorp would purchase and use in its own production of urethane foam). The generation of toxic waste was a natural byproduct of this manufacturing process. And even if GenCorp somehow did not realize that this process would generate hazardous waste when it entered into the contract (because Olin held the patent for the process), it necessarily appreciated this reality when it approved the plant design specifications and capital expenditure requests. The construction plans specifically provided that the hazardous waste generated by the TDI Plant would be placed in drums and buried at an offsite location.
Other facts found by the district court show that the parties’ business arrangements for the TDI Plant encompassed plans or preparations for waste disposal: (1) the TDI Committee (whose membership included equal numbers of representatives from Olin and GenCorp) discussed TDI residue disposal; (2) GenCorp Committee members in particular researched and recommended waste disposal locations; (3) the TDI Committee approved methods to reduce the volume of waste sent offsite (i.e., to the Big D landfill) through capital improvements to the plant; and (4) Olin and GenCorp both funded research aimed at further reducing the volume of residue waste that needed to be disposed offsite. See D. Ct. Op. at 45-46. Considered together, these facts amply show that GenCorp “intended to” (and actually did) “enter into a transaction that included an ‘arrangement for’ the disposal of hazardous substances.” Cello-Foil, 100 F.3d at 1231; accord Carter-Jones Lumber Co. v. Dixie Distributing Co., 166 F.3d 840, 845 (6th Cir.1999).
B.
GenCorp’s arguments to the contrary amount to variations on two themes: (1) GenCorp did not actively participate in arranging for the waste disposal at Big D and (2) GenCorp did not “own or possess” the hazardous waste. Neither argument is convincing.
In advancing its first argument, Gen-Corp leans heavily on United States v. Bestfoods, 524 U.S. 51 (1998), and this Court’s interpretation of the decision in Carter-Jones, 166 F.3d 840, and in Township of Brighton, 153 F.3d 307. To GenCorp’s way of thinking, Bestfoods, Carter-Jones and Township of Brighton require a party “to actually dispose of hazardous waste,” something that GenCorp allegedly never did. GenCorp Br. at 24.
This trio of cases does not bear the weight that GenCorp places on them. Bestfoods dealt with a parent company’s potential liability under 42 U.S.C. § 9607(a)(2) as an “operator” of a polluting facility owned and operated by its subsidiary. 524 U.S. at 55. The Court rejected the proposition that a parent corporation’s participation in and control over its subsidiary’s operations “without more” sufficed for operator liability. The Court instead held that for liability to attach either the corporate veil must be pierced or the parent must independently meet the test for operator liability through actions in its own capacity, not its subsidiary’s.
Carter-Jones took the logical next step of extending Bestfoods to an individual corporate shareholder facing potential arranger liability under CERCLA § 107(a)(3). 166 F.3d at 846-47. In doing so, it concluded that the district court’s findings “satisf[ied] the Bestfoods requirement that an officer be actively involved in the arrangements for disposal before individual liability may be imposed.”
Township of Brighton applied the Bestfoods principle in the context of a governmental entity’s potential liability as an “operator” of a waste facility. It concluded that “we apply the same plain-meaning standard articulated in Bestfoods to any context, whether the defendant is acting in a corporate, governmental, or any other capacity.” 153 F.3d at 314 n. 7.
The central lessons of these cases are (1) that CERCLA does not alter the corporate-law principle that, in the absence of a valid reason for piercing the corporate veil, a parent corporation is not liable for the acts of its subsidiaries, and (2) an alleged violator must independently satisfy the test for CERCLA liability— whether for operator liability, as in Bestfoods, or for arranger liability, as in Carter-Jones. The cases, however, do not alter the basic principles for ascertaining “arranger” liability. The references in these cases to “active involvement” or “actual control” relate either to the substantive standard for operator liability, Bestfoods, 524 U.S. at 66-67; Township of Brighton, 153 F.3d at 314, or refer to the notion that an entity must itself be involved as an “operator” or an “arranger,” Carter-Jones, 166 F.3d at 846-47. Nowhere in these cases is there any suggestion that a party must have been “actively involved” in arranging to dispose of hazardous waste at a particular location and in a particular manner. All of this means that even if Olin rather than Gen-Corp specifically engaged Benjamin Brenkus’ hauling services to carry the hazardous waste to Big D, see GenCorp Br. at 30-31, that fact does not insulate GenCorp from liability for “arrang[ing] for” the disposal of hazardous waste.
GenCorp’s second argument fares no better. In claiming that it cannot be liable as an “arranger” unless it “owned or possessed” the hazardous waste (the contaminated residue and vacuum-pump oils), Gen-Corp points out that the district court never found that it had title to or physically held the waste but found only that it had a “nexus” to the waste. As an initial matter, however, the company bears some responsibility for this omission (if indeed it was an omission). GenCorp did not squarely join this issue at the district court level, and it cannot fairly criticize the district court for failing to address an issue it never directly raised. See GenCorp Post-Trial Br. at JA 222-24 (arguing that disposal activities of hourly TDI Plant employees on its payroll did not satisfy arranger liability requirements); PACCAR Inc. v. TeleScan Technologies, L.L.C., 319 F.3d 243, 258 (6th Cir.2003).
The legal argument, at any rate, is mistaken. Section 107(a)(3) of CERCLA, it is true, imposes liability on “any person who ... arranged for disposal ... of hazardous substances owned or possessed by such person.” 42 U.S.C. § 9607(a)(3) (emphasis added). But to say that the statute requires ownership or possession of the waste does not establish what evidence will satisfy the requirement or, more particularly, whether constructive ownership or possession will suffice.
When Congress enacted CERCLA in 1980, it did so against a statutory and common-law backdrop that had long dignified constructive ownership as a legitimate means by which to establish criminal and other forms of liability for individuals and entities that did not have title to or physically possess certain items. See, e.g., United States v. Crippen, 459 F.2d 1387, 1388 (3d Cir.1972) (“constructive” possession — “dominion and control over the [] substance” — may satisfy the possession requirement in 21 U.S.C. § 841); United States v. Hutchinson, 488 F.2d 484, 488 (8th Cir.1973) (possession in § 841 may be “actual or constructive”); United States v. Poore, 594 F.2d 39, 43 (4th Cir.1979) (accepting constructive possession as satisfying 26 U.S.C. § 5861(d)’s possession requirement); State v. Wolery, 46 Ohio St.2d 316, 348 N.E.2d 351, 360 (1976) (“The prevailing rule at common law and in most jurisdictions is that actual physical possession is not a requisite of receiving [stolen property]. Possession may be constructive.”) (quotation marks omitted); People v. Mumford, 60 Mich.App. 279, 230 N.W.2d 395, 396-97 (1975) (“possession” element in a drug offense “is to be construed in its commonly understood sense and may encompass both actual and constructive possession”); Duchac v. State, 505 S.W.2d 237, 240-41 (Tenn.1973) (it “appears to be the general rule throughout the United States” that “effective control” over burglary tools satisfies the possession requirement for that offense).
Under these circumstances, courts quite appropriately infer that Congress meant the phrase “ownership or possession” to include constructive ownership or possession. See Cmty. for Creative Non-Violence v. Reid, 490 U.S. 730, 739 (1989) (“[W]here Congress uses terms that have accumulated settled meaning under ... the common law, a court must infer, unless the statute otherwise dictates, that Congress means to incorporate the established meaning of these terms.”) (quotation omitted); cf. Overstreet v. North Shore Corp., 318 U.S. 125 (1943) (interpreting statutory phrase by reference to its meaning in another statute); see generally 2B Norman J. Singer, Sutherland on Statutory Construction § 53.03 (5th ed. 1992) (“It is well established that the interpretation of a doubtful statute may be influenced by language of other statutes which are not specifically related, but which apply to similar persons, things, or relationships.”).
In imposing arranger liability on companies under CERCLA, other circuits have accepted less than legal title or physical possession. See United States v. Shell Oil Co., 294 F.3d 1045, 1058 (9th Cir.2002) (accepting the premise that “parties who did not literally own or physically possess hazardous waste at the time it was disposed of or released” may be liable); Catellus Dev. Corp. v. United States, 34 F.3d 748, 752 (9th Cir.1994) (considering ownership or control to be “evidence” of, but not a precondition for, arranger liability); United States v. Northeastern Pharm. & Chem. Co., Inc., 810 F.2d 726, 748 (8th Cir.1986) (“[R]equiring proof of personal ownership or actual physical possession of hazardous substances ... would be inconsistent with the broad remedial purposes of CERCLA.”); see also Geraghty & Miller, 234 F.3d at 929 (stating that a “nexus” — i.e., an obligation to control the hazardous waste — must exist). As these cases indicate, control over the hazardous waste may suffice where literal ownership or possession falls short. See Shell Oil, 294 F.3d at 1058 (summarizing that courts have not imposed arranger liability without either ownership of the waste or “authority to control or duty to dispose of” hazardous waste) (citing United States v. Iron Mt. Mines, Inc., 881 F.Supp. 1432, 1451 (E.D.Colo.1995)); Northeastern Pharm., 810 F.2d at 743-44 (individual corporate officer’s “actual control” satisfied the statute’s possession requirement); CPC Int’l, 759 F.Supp. at 1278 (“section 107(a)(3) requires the assumption of responsibility for or control over the disposition of hazardous waste” rather than legal title).
For several reasons, Gen-Corp’s control over the hazardous waste suffices to establish constructive ownership and possession. First, even if GenCorp never held title to the TDI Plant, it had an active interest in the facility through its option to buy the plant, secured by its “earnest money” — its contribution of one half of the construction costs. Second, GenCorp had equal representation on the Committee that oversaw “the construction, operation and management of the TDI Plant.” Third, the Committee approved the design plans, capital appropriations requests and the budgets of the TDI Plant, all of which contemplated continued offsite waste disposal. Indeed, on the basis of these facts, among others, the district court permissibly concluded that GenCorp was an “operator” of the TDI Plant by “manag[ing], direct[ing], or conduct[ing] operations specifically related to pollution,” D. Ct. Op. at 42-43. Surely a person who “manages” activities “specifically related to pollution” at a plant exercises control over the waste disposal process.
These facts and the district court’s fact finding also show that GenCorp and Olin did not have a traditional buyer-seller relationship. Contrary to GenCorp’s concerns, in other words, our decision does not mean that mere customers of waste-generating manufacturers or processors will face “arranger” liability under CERCLA — unless, that is, they exercise control over the hazardous waste and its disposal as GenCorp did in implementing this unique business relationship. See Morton Int’l, Inc. v. A.E. Staley Mfg., Co., 343 F.3d 669, 683 (3d Cir.2003); Shell Oil, 294 F.3d at 1057; Concrete Sales & Servs., Inc. v. Blue Bird Body Co., 211 F.3d 1333, 1339 (11th Cir.2000); Gen. Electric Co. v. AAMCO Transmissions, Inc., 962 F.2d 281, 287 (2d Cir.1992).
GenCorp, lastly, appears to argue that even if the evidence establishes its “arranger” status under the 1962 Agreement with Olin, the evidence does not support the district court’s finding of arranger liability after the parties unwound the agreement (effective October 1, 1973). See Gen-Corp Br. at 40. While that argument may well have force — particularly since Gen-Corp’s involvement in oversight of the TDI Plant ended in 1973 — GenCorp does not address the district court’s conclusion that post-1973 activities would not have made a material difference to the cost allocation in view of the extent of contamination at Big D that had already taken place by that time. See D. Ct. Op. at 55. Accordingly, we reject this claim.
V.
Both parties challenge the district court’s allocation of costs. In its cross-appeal, Olin challenges the district court’s allocation of liability for the Big D clean-up (70% Olin, 30% GenCorp), claiming that it should have evenly divided the costs between the two companies. According to Olin, the district court abused its discretion in allocating costs, by improperly considering two factors: (1) Olin’s several-year delay in giving notice to GenCorp of the federal EPA’s involvement at the Big D site, and (2) Olin’s relatively greater authority over the TDI Plant under the parties’ 1962 Agreement.
Olin must run uphill in making this argument. Abuse-of-discretion review applies to the district court’s allocation of costs. See United States v. Consolidation Coal Co., 345 F.3d 409, 412 (6th Cir.2003). And the statute augments that discretion by saying that the court “may allocate response costs ... using such equitable factors as the court determines are appropriate,” 42 U.S.C. § 9613(f)(1) (emphasis added). Even if Olin is correct that it had no legal duty to notify GenCorp before beginning to clean up the site, see Olin Br. at 63, the district court permissibly considered the fact that “Olin foreclosed Gen-Corp from cooperating with the regulatory authorities during years of negotiation ... and then, after extensive clean-up work was underway ... called upon GenCorp to be a partner in cost-sharing as a fait accompli.” D. Ct. Op. at 53. The nature of the remediation plan and the costs associated with it assuredly could have been affected by the participation of a third party who knew the facility first hand. That we will never know the answer to that question with certitude, however, is precisely why the defense of laches exists and why this factor bears a similarity to that “equitable” defense. See United States v. R.W. Meyer, Inc., 932 F.2d 568, 572 (6th Cir.1991) (noting that under CERCLA § 113(f)(1) a court may consider “any factor it deems in the interest of justice in allocating contribution recovery .... No exhaustive list of criteria ... should be formulated .... [but] [a] court may consider ... any traditional equitable defenses”) (emphasis added).
GenCorp similarly challenges the district court’s allocation of liability with respect to the Fields Brook site (60% Olin, 40% GenCorp). But because GenCorp raised this issue for the first time in its reply brief and because GenCorp failed even to develop the argument there, we need not consider it here. See Priddy v. Edelman, 883 F.2d 438, 446 (6th Cir.1989).
VI.
GenCorp next argues that even if we affirm the district court’s judgment as to its contribution liability (which we do), we must reverse the award of prejudgment interest to Olin. In doing so, it raises two challenges to the district court’s analysis: (1) the trial court erroneously believed that prejudgment interest in a CERCLA contribution action is mandatory, and (2) Olin’s “inexplicable delay” in bringing the suit “eliminated” its claim for such interest. GenCorp Br. at 60.
The terms of the statute and recent case law foreclose both arguments. An award of prejudgment interest — whether in a joint and several liability action under § 107 or a contribution action under § 113 of CERCLA — is mandatory and the statute makes no exception for late claims. See 42 U.S.C. § 9607(a) (“The amounts recoverable in an action under this section shall include interest on the amounts recoverable under [other provisions].”) (emphasis added); Consolidation Coal, 345 F.3d at 415 (deciding, in the context of a contribution action, that “[a]n award of prejudgment interest ... is mandatory”); see also Township of Brighton, 153 F.3d at 321 (interest is mandatory in a § 107 action). The district court accordingly did not abuse its discretion in awarding prejudgment interest on this claim.
VII.
In its cross-appeal, Olin also claims that the district court should have “enter[ed] a declaratory judgment finding [GenCorp] liable for Olin’s allowable response costs incurred after December 31, 1999, at the previously-determined percentage share for each site.” D. Ct. Op. of Jan. 22, 2003, at JA 602. Section 113(g)(2), the relevant statute, provides as follows: “In any [] action described in this subsection, the court shall enter a declaratory judgment on liability for response costs or damages that will be binding on any subsequent ... actions to recover further response costs or damages.” 42 U.S.C. § 9613(g)(2) (emphasis added). Case law confirms that the words mean what they say. In the context of a § 107 action, this Court has held that “[t]he entry of declaratory judgment as to liability is mandatory.” Kelley v. E.I. DuPont de Nemours & Co., 17 F.3d 836, 844 (6th Cir.1994). GenCorp responds that whether declaratory relief as to future liability in a contribution suit is mandatory is an open question in this circuit — one on which other courts have divided — and that awarding declaratory relief is inherently a “discretionary” function of district courts. GenCorp Reply Br. at 53.
We agree with Olin insofar as it claims that requests for declaratory judgments concerning future response costs in § 107 and § 113(f) suits must be treated alike. The sections governing contribution “work in conjunction” with the liability provisions of § 107(a), as “parties seeking contribution” must turn to “§ 107 to establish the basis and elements of liability of the defendants.” Centerior, 153 F.3d at 350; see also Dent v. Beazer Materials & Servs., Inc., 156 F.3d 523, 531-32 (4th Cir.1998) (determining, in a contribution action, that a court must enter declaratory relief as to future response costs where liability for existing costs is established); AlliedSignal v. Amcast Int’l Corp., 177 F.Supp.2d 713, 757 (S.D.Ohio 2001).
Entry of a declaratory judgment, however, cannot be fully mandatory. Under Article III of the United States Constitution, “Congress cannot create a right of action where no case or controversy otherwise exists.” Kelley, 17 F.3d at 844. In order to satisfy the “case or controversy” requirement, “a party seeking declaratory relief must allege facts to support a likelihood” that it will incur future costs recoverable under CERCLA. Id. at 845. Based on the record before us, we are unable to find sufficient evidence of future response costs to satisfy the “case or controversy” requirement. Therefore, we remand this jurisdictional issue to the district court for initial consideration of whether a “case or controversy” exists, and if so, for the entry of a declaratory judgment for future response costs.
VIII.
For these reasons, we affirm the district court’s judgment as to GenCorp’s contribution liability and its (and Olin’s) relative shares of the remediation costs. We remand, however, for a determination by the district court whether a “case or controversy” exists with respect to continuing clean-up costs, and if that jurisdictional requirement is satisfied, for the entry of a declaratory judgment.