delivered the opinion of the Court.
The United States brought this action for the costs of cleaning up industrial waste generated by a chemical plant. The issue before us, under the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CERCLA), 94 Stat. 2767, as amended, 42 U. S. C. §9601 et seq., is whether a parent corporation that actively participated in, and exercised control over, the operations of a subsidiary may, without more, be held liable as an operator of a polluting facility owned or operated by the subsidiary. We answer no, unless the corporate veil may be pierced. But a corporate parent that actively participated in, and exercised control over, the operations of the facility itself may be held directly liable in its own right as an operator of the facility.
I
In 1980, CERCLA was enacted in response to the serious environmental and health risks posed by industrial pollution. See
Exxon
Corp. v. Hunt,
II
In 1957, Ott Chemical Co. (Ott I) began manufacturing chemicals at a plant near Muskegon, Michigan, and its intentional and unintentional dumping of hazardous substances significantly polluted the soil and ground water at the site. In 1965, respondent CPC International Inc. 3 incorporated a wholly owned subsidiary to buy Ott I’s assets in exchange for CPC stock. The new company, also dubbed Ott Chemical Co. (Ott II), continued chemical manufacturing at the site, and continued to pollute its surroundings. CPC kept the *57 managers of Ott I, including its founder, president, and principal shareholder, Arnold Ott, on board as officers of Ott II. Arnold Ott and several other Ott II officers and directors were also given positions at CPC, and they performed duties for both corporations.
In 1972, CPC sold Ott II to Story Chemical Company, which operated the Muskegon plant until its bankruptcy in 1977. Shortly thereafter, when respondent Michigan Department of Natural Resources (MDNR) 4 examined the site for environmental damage, it found the land littered with thousands of leaking and even exploding drums of waste, and the soil and water saturated with noxious chemicals. MDNR sought a buyer for the property who would be willing to contribute toward its cleanup, and after extensive negotiations, respondent Aerojet-General Corp. arranged for transfer of the site from the Story bankruptcy trustee in 1977. Aerojet created a wholly owned California subsidiary, Cordova Chemical Company (Cordova/California), to purchase the property, and Cordova/California in turn created a wholly owned Michigan subsidiary, Cordova Chemical Company of Michigan (Cordova/Michigan), which manufactured chemicals at the site until 1986. 5
By 1981, the federal Environmental Protection Agency had undertaken to see the site cleaned up, and its long-term remedial plan called for expenditures well into the tens of millions of dollars. To recover some of that money, the *58 United States filed this action under §107 in 1989, naming five defendants as responsible parties: CPC, Aerojet, Cordova/California, Cordova/Miehigan, and Arnold Ott. 6 (By that time, Ott I and Ott II were defunct.) After the parties (and MDNR) had launched a flurry of contribution claims, counterclaims, and cross-claims, the District Court consolidated the eases for trial in three phases: liability, remedy, and insurance coverage. So far, only the first phase has been completed; in 1991, the District Court held a 15-day bench trial on the issue of liability. Because the parties stipulated that the Muskegon plant was a “facility” within the meaning of 42 U. S. C. § 9601(9), that hazardous substances had been released at the facility, and that the United States had incurred reimbursable response costs to clean up the site, the trial focused on the issues of whether CPC and Aerojet, as the parent corporations of Ott II and the Cordova companies, had “owned or operated” the facility within the meaning of § 107(a)(2).
The District Court said that operator liability may attach to a parent corporation both directly, when the parent itself operates the facility, and indirectly, when the corporate veil can be pierced under state law. See CPC Int’l, Inc. v. Aerojet-General Corp., 777 F. Supp. 549, 572 (WD Mich. 1991). The court explained that, while CERCLA imposes direct liability in situations in which the corporate veil cannot be pierced under traditional concepts of corporate law, “the statute and its legislative history do not suggest that CERCLA rejects entirely the crucial limits to liability that are inherent to corporate law.” Id., at 573. As the District Court put it:
“a parent corporation is directly liable under section 107(a)(2) as an operator only when it has exerted power or influence over its subsidiary by actively participating in and exercising control over the subsidiary’s business *59 during a period of disposal of hazardous waste. A parent’s actual participation in and control over a subsidiary’s functions and decision-making creates ‘operator’ liability under CERCLA; a parent’s mere oversight of a subsidiary’s business in a manner appropriate and consistent with the investment relationship between a parent and its wholly owned subsidiary does not.” Ibid.
Applying that test to the facts of this ease, the District Court held both CPC and Aerojet liable under § 107(a)(2) as operators. As to CPC, the court found it particularly telling that CPC selected Ott IPs board of directors and populated its executive ranks with CPC officials, and that a CPC official, G. R. D. Williams, played a significant role in shaping Ott II’s environmental compliance policy.
After a divided panel of the Court of Appeals for the Sixth Circuit reversed in part,
United States
v.
Cordova/Michigan,
“[Wjhere a parent corporation is sought to be held fia-ble as an operator pursuant to 42 U. S. C. § 9607(a)(2) based upon the extent of its control of its subsidiary which owns the facility, the parent will be liable only when the requirements necessary to pierce the corporate veil [under state law] are met. In other words,... whether the parent will be liable as an operator depends *60 upon whether the degree to which it controls its subsidiary and the extent and manner of its involvement with the facility, amount to the abuse of the corporate form that will warrant piercing the corporate veil and disregarding the separate corporate entities of .the parent and subsidiary.” Id., at 580.
Applying Michigan veil-piercing law, the Court of Appeals decided that neither CPC nor Aerojet 7 was liable for controlling the actions of its subsidiaries, since the parent and subsidiary corporations maintained separate personalities and the parents did not utilize the subsidiary corporate form to perpetrate fraud or subvert justice.
We granted certiorari,
*61 J — { H-1 H
It is a general principle of corporate law deeply “ingrained in our economic and legal systems” that a parent corporation (so-called because of control through ownership of another corporation’s stock) is not liable for the acts of its subsidiaries. Douglas & Shanks, Insulation from Liability Through Subsidiary Corporations, 39 Yale L. J. 193 (1929) (hereinafter Douglas); see also,
e. g., Buechner
v.
Farbenfabriken Bayer Aktiengesellschaft,
But there is an equally fundamental principle of corporate law, applicable to the parent-subsidiary relationship as well as generally, that the corporate veil may be pierced and the shareholder held liable for the corporation’s conduct when,
inter alia,
the corporate form would otherwise be misused to accomplish certain wrongful purposes, most notably fraud, on the shareholder’s behalf. See,
e. g., Anderson
v.
Abbott, supra,
at 362 (“[Tjhere are occasions when the limited liability sought to be obtained through the corporation will be qualified or denied”);
Chicago, M. & St. P. R. Co.
v.
Minneapolis Civic and Commerce Assn.,
IV
A
If the Act rested liability entirely on ownership of a polluting facility, this opinion might end here; but CERCLA liability may turn on operation as well as ownership, and nothing in the statute’s terms bars a parent corporation from direct liability for its own actions in operating a facility owned by its subsidiary. As Justice (then-Professor) Douglas noted almost 70 years ago, derivative liability cases are to be distinguished from those in which “the alleged wrong can seemingly be traced to the parent through the conduit of its own personnel and management” and “the parent is directly a participant in the wrong complained of.” Douglas 207,208.
11
*65
In such instances, the parent is directly liable for its own actions. See H. Henn & J. Alexander, Laws of Corporations 347 (3d ed. 1983) (hereinafter Henn
&
Alexander) (“Apart from corporation law principles, a shareholder, whether a natural person or a corporation, may be liable on the ground that such shareholder’s activity resulted in the liability”). The fact that a corporate subsidiary happens to own a polluting facility operated by its parent does nothing, then, to displace the rule that the parent “corporation is [itself] responsible for the wrongs committed by its agents in the course of its business,”
Mine Workers
v.
Coronado Coal Co.,
Under the plain language of the statute, any person who operates a polluting facility is directly liable for the costs of cleaning up the pollution. See 42 U. S. C. § 9607(a)(2). This is so regardless of whether that person is the facility’s owner, the owner’s parent corporation or business partner, or even a saboteur who sneaks into the facility at night to discharge its poisons out of malice. If any such act of operating a corporate subsidiary’s facility is done on behalf of a parent corporation, the existence of the parent-subsidiary relationship under state corporate law is simply irrelevant to the issue of direct liability. See
Riverside Market Dev. Corp.
v.
International Bldg. Prods., Inc.,
This much is easy to say: the difficulty comes in defining actions sufficient to constitute direct parental “operation.” Here of course we may again rue the uselessness of CERCLA’s definition of a facility’s “operator” as “any person ... operating” the facility, 42 U. S. C. § 9601(20)(A)(ii), which leaves us to do the best we can to give the term its “ordinary or natural meaning.”
Bailey
v.
United States,
B
With this understanding, we are satisfied that the Court of Appeals correctly rejected the District Court’s analysis of direct liability. But we also think that the appeals court erred in limiting direct liability under the statute to a parent’s sole or joint venture operation, so as to eliminate any possible finding that CPC is hable as an operator on the facts of this case.
1
By emphasizing that “CPC is directly hable under section 107(a)(2) as an operator because CPC actively participated in and exerted significant control over Ott II’s business and decision-making,”
The well-taken objection to the actual control test, however, is its fusion of direct and indirect habihty; the test is administered by asking a question about the relationship between the two corporations (an issue going to indirect liability) instead of a question about the parent’s interaction with the subsidiary’s faeihty (the source of any direct habihty). If, however, direct habihty for the parent’s operation of the faeihty is to be kept distinct from derivative habihty for the subsidiary’s own operation, the focus of the enquiry must
*68
necessarily be different under the two tests. “The question is not whether the parent operates the subsidiary, but rather whether it operates the facility, and that operation is evidenced by participation in the activities of the facility, not the subsidiary. Control of the subsidiary, if extensive enough, gives rise to indirect liability under piercing doctrine, not direct liability under the statutory language.” Oswald 269; see also
Schiavone
v.
Pearce,
In addition to (and perhaps as a reflection of) the erroneous focus on the relationship between CPC and Ott II, even those findings of the District Court that might be taken to speak to the extent of CPC’s activity at the facility itself are flawed, for the District Court wrongly assumed that the actions of the joint officers and directors are necessarily attributable to CPC. The District Court emphasized the facts that CPC placed its own high-level officials on Ott II’s board of directors and in key management positions at Ott II, and that those individuals made major policy decisions and conducted day-to-day operations at the facility: “Although Ott II corporate officers set the day-to-day operating policies for the company without any need to obtain formal approval from CPC, CPC actively participated in this decision-making because high-ranking CPC officers served in Ott II management positions.” Id., at 559; see also id., at 575 (relying on “CPC’s involvement in major decision-making and day-today operations through CPC officials who served within Ott II management, including the positions of president and chief *69 executive officer,” and on “the conduct of CPC officials with respect to Ott II affairs, particularly Arnold Ott”); id., at 558 (“CPC actively participated in, and at times controlled, the policy-making decisions of its subsidiary through its representation on the Ott II board of directors”); id., at 559 (“CPC also actively participated in and exerted control over day-to-day decision-making at Ott II through representation in the highest levels of the subsidiary’s management”).
In imposing direct liability on these grounds, the District Court failed to recognize that “it is entirely appropriate for directors of a parent corporation to serve as directors of its subsidiary, and that fact alone may not serve to expose the parent corporation to liability for its subsidiary’s acts.”
American Protein Corp.
v.
AB Volvo,
This recognition that the corporate personalities remain distinct has its corollary in the “well established principle [of corporate law] that directors and officers holding positions with a parent and its subsidiary can and do ‘change hats’ to represent the two corporations separately, despite their common ownership.”
Lusk
v.
Foxmeyer Health Corp.,
In sum, the District Court’s focus on the relationship between parent and subsidiary (rather than parent and facility), combined with its automatic attribution of the actions of dual officers and directors to the corporate parent, erroneously, even if unintentionally, treated CERCLA as though it displaced or fundamentally altered common-law standards of limited liability. Indeed, if the evidence of common corporate personnel acting at management and directorial levels were enough to support a finding of a parent corporation’s direct operator liability under CERCLA, then the possibility of resort to veil piercing to establish indirect, derivative liability for the subsidiary’s violations would be academic. There would in essence be a relaxed, CERCLA-speeifie rule of derivative liability that would banish traditional standards and expectations from the law of CERCLA liability. But, as we have said, such a rule does not arise from congressional silence, and CERCLA’s silence is dispositive.
2
We accordingly agree with the Court of Appeals that a partieipation-and-control test looking to the parent’s supervi *71 sion over the subsidiary, especially one that assumes that dual officers always act on behalf of the parent, cannot be used to identify operation of a facility resulting in direct parental liability. Nonetheless, a return to the ordinary meaning of the word “operate” in the organizational sense will indicate why we think that the Sixth Circuit stopped short when it confined its examples of direct parental operation to exclusive or joint ventures, and declined to find at least the possibility of direct operation by CPC in this case.
In our enquiry into the meaning Congress presumably had in mind when it used the verb “to operate,” we recognized that the statute obviously meant something more than mere mechanical activation of pumps and valves, and must be read to contemplate “operation” as including the exercise of direction over the facility’s activities. See supra, at 66-67. The Court of Appeals recognized this by indicating that a parent can be held directly liable when the parent operates the facility in the stead of its subsidiary or alongside the subsidiary in some sort of a joint venture. See
Identifying such an occurrence calls for line-drawing yet again, since the acts of direct operation that give rise to parental liability must necessarily be distinguished from the interference that stems from the normal relationship between parent and subsidiary. Again norms of corporate behavior (undisturbed by any CERCLA provision) are crucial reference points. Just as we may look to such norms in identifying the limits of the presumption that a dual office *72 holder acts in his ostensible capacity, so here we may refer to them in distinguishing a parental officer’s oversight of a subsidiary from such an officer’s control over the operation of the subsidiary’s facility, “[Ajctivities that involve the facility but which are consistent 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 liability.” Oswald 282. The critical question is whether, in degree and detail, actions directed to the facility by an agent of the parent alone are eccentric under accepted norms of parental oversight of a subsidiary’s facility.
There is, in fact, some evidence that CPC engaged in just this type and degree of activity at the Muskegon plant. The District Court’s opinion speaks of an agent of CPC alone who played a conspicuous part in dealing with the toxic risks emanating from the operation of the plant. G. R. D. Williams worked only for CPC; he was not an employee, officer, or director of Ott II, see Tr. of Oral Arg. 7, and thus, his actions were of necessity taken only on behalf of CPC. The District Court found that “CPC became directly involved in environmental and regulatory matters through the work of... Williams, CPC’s governmental and environmental affairs director. Williams ... became heavily involved in environmental issues at Ott II.”
We think that these findings are enough to raise an issue of CPC’s operation of the facility through Williams’s actions, though we would draw no ultimate conclusion from these findings at this point. Not only would we be deciding in the first instance an issue on which the trial and appellate courts did not focus, but the very fact that the District Court did not see the ease as we do suggests that there may be still *73 more to be known about Williams’s activities. Indeed, evén as the factual findings stand, the trial court offered little in the way of concrete detail for its conclusions about Williams’s role in Ott II’s environmental affairs, and the parties vigorously dispute the extent of Williams’s involvement. Prudence thus counsels us to remand, on the theory of direct operation set out here, for reevaluation of Williams’s role, and of the role of any other CPC agent who might be said to have had a part in operating the Muskegon facility. 14
V
The judgment of the Court of Appeals for the Sixth Circuit is vacated, and the ease is remanded with instructions to return it to the District Court for further proceedings consistent with this opinion.
It is so ordered.
Notes
“CERCLA . . . imposes the costs of the cleanup on those responsible for the contamination.”
Pennsylvania
v.
Union Gas Co.,
“The term ‘facility’ means (A) any building, structure, installation, equipment, pipe or pipeline (including any pipe into a sewer or publicly owned treatment works), well, pit, pond, lagoon, impoundment, ditch, landfill, storage container, motor vehicle, rolling stock, or aircraft, or (B) any site or area where a hazardous substance has been deposited, stored, disposed of, or placed, or otherwise come to be located; but does not include any consumer product in consumer use or any vessel.”
CPC has recently changed its name to Bestfoods. Consistently with the briefs and the opinions below, we use the name CPC herein.
The powers and responsibilities of MDNR have since been transferred to the Michigan Department of Environmental Quality.
Cordova/California and MDNR entered into a contract under which Cordova/California agreed to undertake certain cleanup actions, and MDNR agreed to share in the funding of those actions and to indemnify Cordova/California for various expenses. The Michigan Court of Appeals has held that this agreement requires MDNR to indemnify Aerojet and its Cordova subsidiaries for any GERCLA liability that they may incur in connection with their activities at the Muskegon facility. See
Cordova Chemical Co.
v.
MDNR,
Arnold Ott settled out of court with the Government on the eve of trial.
Unlike CPC, Aerojet does not base its defense in this Court on a claim that, absent unusual circumstances, a parent company can be held liable as an operator of a facility only by piercing the corporate veil. Rather, Aerojet denies liability by claiming that (I) neither it nor its subsidiaries disposed of hazardous substances during their operation of the facility, see Brief for Respondents Aerojet-General Corp. et al. 27-36, and (2) it is entitled to a third-party defense under § 107(b)(8) of CERCLA, 42 U. S. C. § 9607(b)(3), see Brief for Respondents Aerojet-General Corp. et al. 38-46. The Court of Appeals expressed some measure of agreement with Aerojet on these points and instructed the District Court to consider them on remand. See
Compare
United States
v.
Cordova/Michigan,
There is significant disagreement among courts and commentators over whether, in enforcing CERCLA’s indirect liability, courts should borrow state law, or instead apply a federal common law of veil piercing. Compare,
e. g.,
Some courts and commentators have suggested that this indirect, veil-piercing approach can subject a parent corporation to liability only as an owner, and not as an operator. See, e.g., Lansford-Coaldale Joint Water Auth. v. Tonolli Corp., supra, at 1220; Oswald, Bifurcation of the Owner and Operator Analysis under CERCLA, 72 Wash. U. L. Q. 223,281-282 (1994) (hereinafter Oswald). We think it is otherwise, however. If a subsidiary that operates, but does not own, a facility is so pervasively controlled by its parent for a sufficiently improper purpose to warrant veil piercing, the parent may be held derivatively liable for the subsidiary’s acts as an operator.
While this article was written together with Professor Shanks, the passages quoted in this opinion were written solely by Justice Douglas. See Douglas 193, n. *.
See Oswald 257 (“There are . . . instances ... in which the parent has not sufficiently overstepped the bounds of corporate separateness to warrant piercing, yet is involved enough in the facility’s activities that it should be held liable as an operator. Imagine, for example, a parent who strictly observed corporate formalities, avoided intertwining officers and directors, and adequately capitalized its subsidiary, yet provided active, daily supervision and control over hazardous waste disposal activities of the subsidiary. Such a parent should not escape liability just because its activities do not justify a piercing of the subsidiary’s veü”).
We do not attempt to recite the ways in which the Government could show that dual officers or directors were in fact acting on behalf of the parent. Here, it is prudent to say only that the presumption that an act is taken on behalf of the corporation for whom the officer claims to act is strongest when the act is perfectly consistent with the norms of corporate behavior, but wanes as the distance from those accepted norms approaches the point of action by a dual officer plainly contrary to the interests of the subsidiary yet nonetheless advantageous to the parent.
There are some passages in the District Court's opinion that might suggest that, without reference to Williams, some of Ott IFs actions in operating the facility were in fact dictated by, and thus taken on behalf of, CPC. See, e. g., 777 F. Supp., at 561 (“CPC officials engaged in . . . missions to Ott II in which Ott II officials received instructions on how to improve and change”); id, at 559 (“CPC executives who were not Ott II board members also occasionally attended Ott II board meetings”). But nothing in the District Court’s findings of fact, as written, even comes dose to overcoming the presumption that Ott II offidals made their ded-sions and performed their acts as agents of Ott II. Indeed, the finding that "Ott II corporate officers set the day-to-day operating polides for the company without any need to obtain formal approval from CPC,” ibid., indicates just the opposite. Still, the Government is, of course, free on remand to point to any additional evidence, not dted by the District Court, that would tend to establish that Ott II’s decisionmakers acted on specific orders from CPC.