Browning-Ferris and several other companies have brought a suit for contribution under the Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA--the Superfund statute). The suit is against Richard Ter Maat and two corporations of which he is (or was--one of the corporations has been sold) the president and principal shareholder; they are M.I.G. Investments, Inc. and AAA Disposal Systems, Inc.

Back in 1971 the owners of a landfill had leased it to a predecessor of Browning-Ferris, which operated it until the fall of 1975. Between then and 1988 it was operated by M.I.G. and AAA. In June of that year, after AAA was sold and Ter Maat moved to Florida, M.I.G. abandoned the landfill without covering it properly. For tax reasons, M.I.G. had been operated with very little capital, and it lacked funds for a proper cover. Two years after the abandonment, the EPA placed the site on the National Priorities List, the list of the toxic waste sites that the Superfund statute requires be cleaned up, see 42 U.S.C. sec.sec. 9605(8)(B), 9616(d), (e), and shortly afterward Browning-Ferris and the other plaintiffs, which shared responsibility for some of the pollution at the site, agreed to clean it up.

Section 113(f)(1) of the Superfund law authorizes any person who incurs costs in cleaning up a toxic-waste site to "seek contribution from any other person who is liable or potentially liable under section 9607(a) of this title. . . . In resolving contribution claims, the court may allocate response costs among liable parties using such equitable factors as the court determines are appropriate." 42 U.S.C. sec. 9613(f)(1). Section 107(a)(1), 42 U.S.C. sec. 9607(a)(1), a part of the statutory provision to which section 113(f)(1) refers, includes in the set of potentially liable persons anyone who owned or operated a landfill when a hazardous substance was deposited in it, and this set is conceded to include both M.I.G. and AAA. The district judge held, however, that Ter Maat was not himself a potentially liable person, because he had done nothing that would subject him to liability on a "piercing the corporate veil" theory for the actions of the two corporations. So far as corporate liability for clean-up costs was concerned, the judge ruled that of the 55 percent of those costs that he deemed allocable to transporters and operators (the other 45 percent he allocated to the owners of the landfill and the generators of the toxic wastes dumped in it), 40 percent was the responsibility of Browning-Ferris and the other 60 percent the responsibility of M.I.G. and AAA. As between those two, the judge allocated responsibility equally, holding that, although the two corporations had operated the landfill jointly, the statute required him to allocate liability severally rather than jointly.

Browning-Ferris and the other companies that have incurred clean-up costs at the site of the former landfill have appealed. All of them join in making the following three arguments: the corporate veil should be pierced two ways, one to make Richard Ter Maat liable for the conduct of both "his" corporations, AAA and M.I.G., and the other to make AAA liable for M.I.G.'s conduct as an affiliate; in any event Ter Maat is directly as distinct from derivatively liable for contribution, as he was personally an operator of the landfill; the Superfund statute does permit joint liability in a contribution suit and it would be equitable to make AAA pay for the whole amount allocated to the two corporations, since they were jointly liable and M.I.G. is assetless (or may be--it has some insurance). Finally, Browning-Ferris argues that the district court allocated too much of the liability for the pollution at the site to it relative to M.I.G. and AAA.

Two issues are relatively simple and we address them first. One is whether an individual can shield himself from liability for operating a hazardous-waste facility merely by being an officer or shareholder of a corporation that also operates the facility. The answer is no. The principle of limited liability shields a shareholder from liability for the debts (including debts arising from tortious conduct) of the corporation in which he owns shares (with the exception discussed later for "veil piercing" situations), but not for his personal debts, including debts arising from torts that he commits himself. In other words, the status of being a shareholder does not immunize a person for liability for his, as distinct from the corporation's, acts. E.g., Sidney S. Arst Co. v. Pipefitters Welfare Education Fund, 25 F.3d 417, 420 (7th Cir. 1994); Spartech Corp. v. Opper, 890 F.2d 949, 953 (7th Cir. 1989); Redwing Carriers, Inc. v. Saraland Apartments, 94 F.3d 1489, 1503 (11th Cir. 1996); Aronson v. Price, 644 N.E.2d 864, 867 (Ind. 1994). There is no liability shield at all for an officer. If he commits an act that is outside the scope of his official duties, his employer may not be liable; but he is whether or not the act was within that scope. E.g., Itofca, Inc. v. Hellhake, 8 F.3d 1202, 1204(7th Cir. 1993); Carter-Jones Lumber Co. v. Dixie Distributing Co., 166 F.3d 840, 846-47 (6th Cir. 1999); United States v. Northeastern Pharmaceutical & Chemical Co., 810 F.2d 726, 744 (8th Cir. 1986). Which is not to say, however, that the officer is automatically liable for the acts of the corporation; there is no doctrine of "superiors' liability," comparable to the doctrine of respondeat superior, that is, the employer's strict liability for torts of the employee committed within the scope of his employment.

So if Ter Maat operated the landfill personally, rather than merely directing the business of the corporations of which he was the president and which either formally, or jointly with him (as well as with each other), operated it, he is personally liable. E.g., United States v. Bestfoods, 524 U.S. 51, 55 (1998); Sidney S. Arst Co. v. Pipefitters Welfare Education Fund, supra, 25 F.3d at 421-22; Riverside Market Development Corp. v. International Bldg. Products, Inc., 931 F.2d 327, 330 (5th Cir. 1991) (per curiam). The line between a personal act and an act that is purely an act of the corporation (or of some other employee) and so not imputed to the president or to other corporate officers is sometimes a fine one, but often it is clear on which side of the line a particular act falls. If an individual is hit by a negligently operated train, the railroad is liable in tort to him but the president of the railroad is not. Or rather, not usually; had the president been driving the train when it hit the plaintiff, or had been sitting beside the driver and ordered him to exceed the speed limit, he would be jointly liable with the railroad. If Ter Maat did not merely direct the general operations of M.I.G. and AAA, as in id. at 330, or specific operations unrelated to pollution, United States v. Bestfoods, supra, 524 U.S. at 66-67; United States v. Township of Brighton, 153 F.3d 307, 313-15 (6th Cir. 1998), but supervised the day- to-day operations of the landfill--for example, negotiating waste-dumping contracts with the owners of the wastes or directing where the wastes were to be dumped or designing or directing measures for preventing toxic substances in the wastes from leeching into the ground and thence into the groundwater--then he would be deemed the operator, jointly with his companies, of the site itself. E.g., Control Data Corp. v. S.C.S.C. Corp., 53 F.3d 930, 937 (8th Cir. 1995); Witco Corp. v. Beekhuis, 38 F.3d 682, 692 (3d Cir. 1994); FMC Corp. v. Aero Industries, Inc., 998 F.2d 842, 846 (10th Cir. 1993); United States v. Northeastern Pharmaceutical & Chemical Co., supra, 810 F.2d at 743. Unfortunately the district court did not consider this possibility, although urged to do so by the plaintiffs, and so a remand is necessary to determine Ter Maat's status.

It is also clear we think that CERCLA does not preclude the imposition of joint as distinct from several liability in a suit for contribution. This is, or at least should be considered, a case of first impression at the appellate level. Carter-Jones Lumber Co. v. Dixie Distributing Co., supra, 166 F.3d at 847-48, does hold that CERCLA permits joint liability in a contribution suit, but on the unsatisfactory ground that state law so required, when it is plain that actions for CERCLA contribution are governed by federal law. Section 113(f), 42 U.S.C. sec. 9613(f)(1), the CERCLA contribution provision, provides expressly that actions under it are to be governed by federal law, with the federal court to "allocate response costs among liable parties using such equitable factors as the court determines are appropriate." There is a savings provision in CERCLA, 42 U.S.C. sec. 9652(d), but the plaintiffs' claims in this case are based exclusively on the federal contribution provision, and not on state law.

An earlier Sixth Circuit case had tossed off a casual dictum that contribution liability under CERCLA is several, not joint, Centerior Service Co. v. Acme Scrap Iron & Metal Corp., 153 F.3d 344, 348 (6th Cir. 1998), but the dictum is defunct after Carter-Jones. A similar dictum in Pinal Creek Group v. Newmont Mining Corp., 118 F.3d 1298, 1303 (9th Cir. 1997), is, when read in context, limited to cases in which potentially responsible parties are suing under a provision of CERCLA, sec. 107(a), 42 U.S.C. sec. 9607(a), that imposes automatic joint liability on defendants; the cases hold that the parties' rights are nevertheless subject to allocation under section 113(f), which does not include that automatic feature. See id. at 1306; Akzo Coatings, Inc. v. Aigner Corp., 30 F.3d 761, 764 (7th Cir. 1994); Centerior Service Co. v. Acme Scrap Iron & Metal Corp., supra, 153 F.3d at 349- 53. Joint liability thus is optional, rather than unavailable.

In traditional common law, when two or more persons inflict an indivisible injury each is fully liable for the injury. That is, the plaintiff can if he wants sue one of the tortfeasors for the entire damages and let the other go, and the one who is sued has no remedy against the one who got off scot-free. CERCLA modifies the traditional common law rule (as many other statutes do and as many state courts have done by modifying the common law) by allowing one liable party to sue another for contribution. It does not follow that if, as in this case, contribution is sought from more than one party, the defendants cannot be held jointly liable. It is up to the district judge, guided only by equitable considerations--a broad and loose standard, see Kerr-McGee Chemical Corp. v. Lefton Iron & Metal Co., 14 F.3d 321, 326 (7th Cir. 1994); Environmental Transportation Systems, Inc. v. ENSCO, Inc., 969 F.2d 503, 509 (7th Cir. 1992); United States v. Colorado & Eastern R.R., 50 F.3d 1530, 1536 (10th Cir. 1995)--to decide, and it is easy to imagine cases, of which this may be one, where such considerations weigh heavily in favor of joint liability.

Suppose, to alter the facts for simplicity's sake, that Browning-Ferris had been made to clean up the entire site even though it had made only a small (say, 1 percent) contribution to its toxicity. Suppose M.I.G. and AAA were the bad actors jointly responsible for the other 99 percent. Suppose that for tax or other reasons M.I.G. had no assets. Under the view of the district court, even though M.I.G. and AAA had combined to inflict an indivisible injury (the contamination for which they were jointly responsible as joint operators), AAA would have to pay only 50 percent of the contamination for which it and M.I.G. were jointly liable, or 49.5 percent of the total clean-up cost (remember that we're assuming that the two corporations are jointly responsible for 99 percent of the total contamination), while Browning-Ferris would have to pay 50.5 percent of the total clean-up cost even though it was responsible for only 1 percent of that cost. Cf. Carter-Jones Lumber Co. v. Dixie Distributing Co., supra. These are not our facts, but they show that a rule against ever holding contribution defendants jointly liable would be inconsistent with the statutory direction that the district court allocate liability equitably among the liable parties. Cf. Centerior Service Co. v. Acme Scrap Iron & Metal Corp., supra, 153 F.3d at 354 n. 12; Sun Co. v. Browning-Ferris, Inc., 124 F.3d 1187, 1193 (10th Cir. 1997); Pinal Creek Group v. Newmont Mining Corp., supra, 118 F.3d at 1303. The judge did not make such an allocation, mistakenly believing himself constrained to allocate liability equally among joint polluters, and so this is another issue requiring further consideration on remand.

The next issue is whether the district judge allocated too large a share (40 percent) of responsibility for the cost of the clean up to Browning-Ferris relative to the defendants, who had operated the landfill for a lot longer time and had dumped a much larger quantity of wastes in it. The judge allocated as large a share as he did to Browning-Ferris because he found that it had operated the landfill poorly and had dumped particularly toxic wastes from a nearby Chrysler plant in violation of its operating permit, and the liquid character of the wastes had hastened their absorption into groundwater. Browning- Ferris argues both that these findings are erroneous and that, in any event, there is no evidence that the wastes from the Chrysler plant increased the cost of cleaning up the site and anyway the amount dumped in the landfill was not as great as the district judge found. From evidence that a considerable portion of the Chrysler wastes were dumped elsewhere, Browning- Ferris argues that defendants' expert had exaggerated the amount deposited in the landfill. Browning-Ferris may be correct on all these factual points, but we cannot say that the district court committed any clear errors in finding as it did, and that of course is our criterion.

The trickier question, which returns us to the issue of the district court's equitable discretion in allocating liability among polluters, is whether the court must find a causal relation between a party's pollution and the actual cost of cleaning up the site. To answer this question we have to distinguish between a necessary condition (or "but-for cause") and a sufficient condition. If event A is a necessary condition of event B, this means that, without A, B will not occur. If A is a sufficient condition of B, this means that, if A occurs, B will occur. If A is that the murder weapon was loaded and B is the murder, then A is a necessary condition. If A is shooting a person through the heart and B is the death of the shooting victim, then A is a sufficient condition of B but not a necessary condition, because a wound to another part of the victim's body might have been fatal as well.

This distinction may sometimes be important in the pollution context. It is easy to imagine a case in which, had X not polluted a site, no clean-up costs would have been incurred; X's pollution would be a necessary condition of those costs and it would be natural to think that he should pay at least a part of them. But suppose that even if X had not polluted the site, it would have to be cleaned up--and at the same cost--because of the amount of pollution by Y. (That would be a case, perhaps rare, in which the clean-up costs were sensitive neither to the amount of pollution nor to any synergistic interaction between the different pollutants.) Then X's pollution would not be a necessary condition of the clean up, or of any of the costs incurred in the clean up. But that should not necessarily let X off the hook. For suppose that if X had not polluted the site at all, there still would have been enough pollution from Y to require a clean up, but that if Y had not polluted the site X's pollution would have been sufficient to require the clean up. In that case, the conduct of X and the conduct of Y would each be a sufficient but not a necessary condition of the clean up, and it would be entirely arbitrary to let either (or, even worse, both) off the hook on this basis. So far as appears, this is such a case; Browning-Ferris's pollution was serious enough (if indeed it dumped a large quantity of Chrysler's particularly toxic wastes) to require that the site be cleaned up, but the other pollution at the site was also enough. If Browning-Ferris's conduct was thus a sufficient though not a necessary condition of the clean up, it is not inequitable to make it contribute substantially to the cost.

We do not suggest that this is one of the presumably rare cases in which the total costs of clean up are unaffected by the number of polluters or the specific amounts or types of pollution contributed by each. Browning-Ferris's pollution was not, so far as appears, so serious all by itself as to have required the incurring of all the clean-up costs that were incurred. Even so, no principle of law, logic, or common sense required the court to allocate those total costs among the polluters on the basis of the volume of wastes alone. Not only do wastes differ in their toxicity, harm to the environment, and costs of cleaning up, and so relative volume is not a reliable guide to the marginal costs imposed by each polluter; but polluters differ in the blameworthiness of the decisions or omissions that led to the pollution, and blameworthiness is relevant to an equitable allocation of joint costs. (Presumably it would not entitle the judge to make one polluter pay for separable costs wholly imposed by other polluters.) The district judge did not abuse his discretion in deciding that all these factors warranted making Browning- Ferris bear more than its proportional volumetric share of the pollution.

It remains to consider two issues of derivative liability. If Ter Maat is deemed on remand to be personally liable as an operator of the landfill, and especially if the judge allocates 100 percent of the joint liability of M.I.G. and AAA to AAA, the solvent one of the pair (there is no suggestion that the change of ownership gets AAA off the liability hook), then Browning-Ferris and the other plaintiffs will be able to collect the amount of contribution to which they are entitled from the defendants. But these are big "ifs" (how big is for the district judge to decide, in the first instance, on remand). The judge may find that Ter Maat was not an operator and that AAA should not bear the entirety of its joint liability with M.I.G. In that event, it will become important whether Ter Maat and AAA are derivatively liable for the conduct of M.I.G., and as the issue is fully briefed we shall resolve it now and hope to head off a further appeal.

The argument is that the corporate veil should be pierced, and Ter Maat, as shareholder of M.I.G. and AAA, and AAA as an affiliate of M.I.G. (both dominated by Ter Maat, and AAA in addition a 30 percent shareholder of M.I.G.), should be held liable for M.I.G.'s debt to the plaintiffs. Although there is a split of authority on whether federal or state law governs the piercing of the corporate veil in CERCLA cases (the Supreme Court expressly left the question open in United States v. Bestfoods, supra, 524 U.S. at 63 n. 9), that is not a problem here because the parties have assumed that state law (Illinois state law, specifically) governs and so any contention that federal law should instead govern has been waived. Lloyd v. Loeffler, 694 F.2d 489, 495 (7th Cir. 1982).

The general rule, of course, in Illinois as elsewhere, is that a shareholder qua shareholder, and a parent, subsidiary, or other affiliate, qua affiliate, is not liable for a corporation's debts. E.g., In re Rehabilitation of Centaur Ins. Co., 632 N.E.2d 1015, 1017 (Ill. 1994); Itofca, Inc. v. Hellhake, supra, 8 F.3d at 1204 (applying Illinois law). That is the principle of limited liability and it serves the important social purpose of encouraging investment by individuals who are risk averse and therefore will not invest (or will insist on a much higher return) in an enterprise if by doing so they expose their entire wealth to the hazards of litigation. But in some circumstances the corporate veil is pierced and the corporation's debtor allowed to collect his debt from the shareholder or affiliate.

In the case of a voluntary creditor, for example someone who had lent money to the corporation, the strongest case for piercing the veil is presented when the corporation had led potential creditors to believe that it was more solvent than it really was. Holland v. Joy Candy Mfg. Corp., 145 N.E.2d 101 (Ill. App. 1957); In re Xonics Photochemical, Inc., 841 F.2d 198, 201 (7th Cir. 1988); In re Kaiser, 791 F.2d 73, 75 (7th Cir. 1986). For example, the nonbank subsidiary of a bank holding company might hold itself out as being the bank, leading creditors to believe that they were dealing with a more substantial corporation than was in fact the case; that would be a good case for piercing the veil. FMC Finance Corp. v. Murphree, 632 F.2d 413, 423 (5th Cir. 1980); cf. Fentress v. Triple Mining, Inc., 635 N.E.2d 102, 108 (Ill. App. 1994); Hystro Products, Inc. v. MNP Corp., 18 F.3d 1384, 1390-91 (7th Cir. 1994) (Illinois law). Although Illinois formulates the test for piercing the veil in more traditional terms--with much talk of "unity of interest and ownership," "commingling of assets," disappearance of the separate "personalities" of the corporation and its owner or affiliate, and the desirability of avoiding "injustice," e.g., id. at 1388-91--it is hard to see how a voluntary creditor can complain if he knows that his debtor lacks sufficient assets to be certain to be able to pay the debt when it comes due.

Analysis is more difficult in the case of an involuntary creditor, such as the plaintiffs here, who wish to be compensated for having in effect "lent" the money to clean up the site of the former landfill but lent under compulsion, having been forced to clean it up by the Superfund law rather than pursuant to a contract with the "debtors," such as M.I.G. See generally Secon Service System, Inc. v. St. Joseph Bank & Trust Co., 855 F.2d 406, 413-14 (7th Cir. 1988). In such a case there is no issue of protecting reliance induced by misrepresentations by the debtor. The plaintiffs in dumping toxic wastes to the landfill obviously were not relying on M.I.G.'s appearing to have greater assets than it actually had. In these circumstances we can think of only two arguments for piercing the corporate veil. The first is that the owners may have so far neglected the legal requirements (requirements not intended solely for the protection of creditors) for operating in the corporate form that they should be taken to have forfeited its protections, forfeiture thus operating to enforce the legal requirements that the state has seen fit to impose on investors who want the benefits of limited liability. Moreover, if the formalities have been flouted, it becomes hard to see how the investors could reasonably have relied on the protections of limited liability; they would have known they were skating on thin ice. In such a case the investment-encouraging function of limited liability is attenuated.

Second, it could be argued that enterprises engaged in potentially hazardous activities should be prevented from externalizing the costs of those activities, by being required to maintain or at least endeavor to maintain a sufficient capital cushion to be answerable in a tort suit should its activities cause harm for which liability would attach, on pain of its shareholders' and affiliates' losing their limited liability should the corporation fail to do this. This argument has not carried the day in any jurisdiction that we are aware of, presumably because of the risks that it would impose on shareholders and because the potential victims of the corporation's hazardous activities can be protected without making inroads into limited liability by requiring enterprises engaged in such activities to post a bond large enough to assure that any judgment against the corporation will be collectible. Courts do, it is true, frequently mention "undercapitalization" as a separate ground from neglect of corporate formalities for piercing the corporate veil. Hystro Products, Inc. v. MNP Corp., supra, 18 F.3d at 1389-91; Sea-Land Services, Inc. v. Pepper Source, 941 F.2d 519, 522 (7th Cir. 1991); People v. V & M Industries, Inc., 700 N.E.2d 746, 751 (Ill. App. 1998); Jacobson v. Buffalo Rock Shooters Supply Inc., 664 N.E.2d 328, 332 (Ill. App. 1996); Fentress v. Triple Mining, Inc., supra, 635 N.E.2d at 107; McCracken v. Olson Cos., 500 N.E.2d 487, 492 (Ill. App. 1986). They do not do so on the basis of unusual risks to potential tort victims or other involuntary creditors, however, though conceivably such concerns are in the background of their thinking.

There is no evidence that either M.I.G. or AAA failed to comply with the legal requirements for operating in the corporate form, except with regard to keeping minutes of their corporate meetings--not a failure significant enough to warrant forfeiture of limited liability, given that penalties should be proportioned to the gravity of the misconduct being penalized. The plaintiffs further argue, however, that M.I.G. was undercapitalized. The clearest case--here merging, though, with neglect of corporate formalities--for forfeiture of limited liability on this ground is where the corporation has failed to maintain the minimum capitalization required by law. But such cases are few (the only Illinois case we can find is Gallagher v. Reconco Builders, Inc., 415 N.E.2d 560, 564 (Ill. App. 1980)), and do not bear on the "pierceability" of the corporate veil of M.I.G., which is not accused of ever having maintained a lower capitalization than the law required. The fact that its motivation in operating with little capital was to reduce taxes does not argue for piercing the corporate veil, since taking advantage of lawful opportunities for avoiding taxes is not wrongful activity.

There is of course a difference between assets available to pay a judgment and taxable income; the former might be ample while the latter was slight. But our point is only that the fact that M.I.G. might have been "undercapitalized" for tax reasons is not a reason for piercing the corporate veil. The cases in which undercapitalization has figured in the decision to pierce the corporate veil are ones in which the corporation had so little money that it could not and did not actually operate its nominal business on its own. An example is the corporation in the Fentress case, which had no capital at all that the court could find, and no bank account. Even there the court did not pierce the veil on the basis of undercapitalization alone, but cited also a persistent neglect of corporate formalities. Undercapitalization is rarely if ever the sole factor in a decision to pierce the corporate veil, William P. Hackney & Tracey G. Benson, "Shareholder Liability for Inadequate Capital," 43 U. Pitt. L. Rev. 837, 885-87 (1982), and we think is best regarded simply as a factor helpful in identifying a corporation as a pure shell, which M.I.G. was not.

So on remand, to summarize, there will be no issues of veil piercing to consider but the court will have to decide Ter Maat's personal liability and whether AAA should pay more than 50 percent of the two corporations' liability to the plaintiffs.

Affirmed in Part, Reversed in Part, and Remanded.