POSNER, Circuit Judge.
These appeals present a variety of interesting questions arising out of a protracted effort by the Equal Employment Opportunity Commission to fix personal liability on an employer’s officers for a practice not authoritatively determined to be discriminatory until years after they committed it, and on a successor of the employer.
The three individual defendants, whom we shall refer to as the “Vucitech group,” were (together with a fourth person, now dead) the officers and shareholders of MTC Gear Corporation, a closely held corporation engaged in the manufacture of gears and other mechanical devices used in motor vehicles. MTC employed between 75 and 100 workers, who were represented by a union. In 1979 the Vucitech group negotiated on behalf of MTC a new collective bargaining agreement with the union. After counsel advised that the recently enacted Pregnancy Discrimination Act of 1978, 42 U.S.C. § 2000e(k) (amending Title VII of the Civil Rights Act of 1964), did not require the payment of maternity benefits as part of the fringe benefits for employees' dependents (as opposed to the employees themselves), the Vucitech group offered the union, in lieu of dependents’ maternity benefits, a “baby bonus” plan. Under the plan every male employee would receive a lump sum of $750 for every baby that his wife gave birth to while he was employed by the corporation. The union agreed, and agreed again in the 1982 collective bargaining negotiations with the Vucitech group, when the baby bonus was raised to $1,000. In between the two negotiations a district court held that the Pregnancy Discrimination Act did not require the payment of maternity benefits to employees’ wives. We affirmed that decision in 1983, in EEOC v. Joslyn Mfg. & Supply Co., 706 F.2d 1469 (7th Cir.1983), but, for reasons to appear presently, we soon vacated our opinion, see 724 F.2d 52 (7th Cir.1983) (per curiam).
As early as 1981, male employees of MTC had complained to the EEOC about the denial of maternity benefits for their wives. At first glance the complaint may seem an odd one. The baby bonus is a maternity benefit, though denominated as being in lieu of a maternity benefit; and people are not usually held to violate the law merely because they use confusing verbiage. It is true that $750 surely, and $1,000 probably, was too little in the years in question to cover the full medical costs of an average pregnancy, but medical benefits often have deductibles, coinsurance, ceilings, or other conditions that operate to limit the employer’s or insurer’s liability for the full extent of the employee’s (or dependent’s) medical expenses. However, maybe because MTC’s female employees, as distinct from the wives of its male em[939]*939ployees, received maternity benefits that were both expressly so designated and larger than the baby bonuses received by male employees for their wives, the Vuci-tech group does not argue that the baby bonus was a maternity benefit — although the district court did deduct from its award the baby bonuses that had been paid. At all events, all of the parties assume, and so shall we, that once the Pregnancy Discrimination Act had been interpreted to apply to dependents’ benefits, the baby bonus plan was unquestionably in violation of the Act. The plan may have violated Title VII even without the Pregnancy Discrimination Act, as we shall see.
The Commission notified MTC of the charges. Conciliation efforts, with the Vu-citech group representing the company, followed, but they failed in March 1982. The Commission decided, however, to postpone formal action until the Supreme Court resolved the question of benefits for pregnant dependents, which it did a few months after our Joslyn decision, holding in Newport News Shipbuilding & Dry Dock Co. v. EEOC, 462 U.S. 669, 103 S.Ct. 2622, 77 L.Ed.2d 89 (1983), that the Act forbids discrimination not only with respect to the wages and benefits of pregnant (necessarily female) employees but also with respect to benefits for pregnant dependents of male employees. We quickly vacated Jos-lyn, and in May 1984 the EEOC filed suit under Title VII against MTC, charging sex discrimination in the failure to pay maternity benefits for the wives of ten (later reduced to eight) employees of MTC but not naming any individuals as defendants.
MTC was a much different entity in 1984 from what it had been at the beginning of 1982. For shortly after the conciliation efforts had broken down, the Vucitech group, whose members were in their sixties and wanted to retire, had sold all their stock in MTC to Muzzamil Niazi, who as part of the deal assumed all of MTC’s liabilities, expressly including any liability growing out of the charges of sex discrimination that had been filed with the EEOC. The Vucitech group retired, but their quiet life was soon interrupted. Niazi was a crook. He embezzled money from MTC on a grand scale, driving the company into the ground. In November 1983 the secured creditors seized all of MTC’s assets and shut it down. Niazi exited, still owing the Vucitech group $2.8 million for the sale of their stock to him. As the members of the group had been counting on this money to finance their retirement, the Niazi fiasco forced them out of retirement. In December 1983, using the proceeds of a $2.8 million loan that they obtained from a bank in Chicago, they bought for $3 million all of MTC’s machinery and equipment at public auction and placed these assets in a new corporation, Profile Gear Corporation, which they owned together with two former employees of MTC and which opened for business in January 1984. Except for the change in name and partial change in ownership, Profile was essentially the same entity (in nature of business, and in customers and employees, though naturally there was some turnover in both of these, groups) as MTC, now defunct. So when the EEOC sued MTC in 1984 it named Profile as an additional defendant.
One year later the EEOC amended its complaint to add the members of the Vuci-tech group as defendants too. The case proceeded to trial. The district court held that the members of the group, having been personally involved in the discriminatory acts, should be held liable along with the employer itself, see 42 U.S.C. § 2000e(b); Musikiwamba v. ESSI, Inc., 760 F.2d 740, 753 (7th Cir.1985), and refused to hold that Newport News, decided after the discriminatory acts committed by these defendants, should be applied only prospectively or that the EEOC had been guilty of laches for failing to sue the Vuci-tech group until 1985. The court also held that Profile was not liable as a successor company for discriminatory acts by MTC, and therefore entered judgment against the Vucitech group only, for a shade under $14,000. The Vucitech group appeals, and the EEOC cross-appeals.
Title VII is an equitable statute in the technical sense; and the Vucitech group makes a powerful if incomplete argument that it was horribly inequitable for the [940]*940EEOC to haul them out of retirement to defend against practices that were lawful when committed and (they argue) in any event harmed no one; to reach into their depleted pockets and make them personally liable for what is really a corporate liability; and to delay the proceedings until MTC was broke so that the judgment would indeed come out of their personal assets. And they say that the principles that judicial decisions need not always be applied retroactively and that laches (delay, in prosecuting a claim, that is both unreasonable and prejudicial) is a defense to an equity suit provide between them all the authority we need to correct this exhibition of bureaucratic lassitude and injustice.
The situation is far more complex. In General Electric Co. v. Gilbert, 429 U.S. 125, 97 S.Ct. 401, 50 L.Ed.2d 343 (1976), the Supreme Court had held that an employee benefit plan which excludes expenses related to pregnancy is not sex discrimination under Title VII, because the group discriminated against consists of pregnant persons, not women as such. Congress overruled this unpopular decision in the Pregnancy Discrimination Act of 1978, which amended Title VII by defining the terms “because of sex” and “on the basis of sex” to include “because of or on the basis of pregnancy,” adding: “and women affected by pregnancy ... shall be treated the same for all employment-related purposes, including receipt of benefits under fringe benefit programs, as other persons not so affected.” 42 U.S.C. § 2000e(k). This language is ambiguous. The first clause could be read to cover employees with pregnant wives, because exclusion of maternity benefits from those employees’ dependent coverage would be an exclusion because of sex; the second clause could not be so read (it protects pregnant employees ). The Vucitech group was apprised of the new statute by MTC’s counsel, who advised them that maternity benefits for employees’ benefits were not covered but also noted that the EEOC disagreed. Thus the Vucitech group knowingly assumed a nontrivial risk that their counsel’s interpretation would not prevail. No doubt they took the risk, in part at least, because of the small potential liability.
Years later this court, joined by some (but not all) other courts that considered the issue, agreed with the position that MTC’s counsel had taken. But the relevance of this fact is obscure. Counsel could not have been relying on cases not yet decided in advising MTC that it could go ahead with the baby bonus plan without violating the law. Nor is the language of the Pregnancy Discrimination Act so pellucid that counsel could be confident of its meaning in the absence of authoritative interpretation. Dependent coverage could easily be thought embraced by the broadly worded first clause, and by the spirit of the Act. To deny maternity benefits for an employee’s wife is in some sense to discriminate (that is, treat differently) on account of pregnancy, since if instead of being pregnant the wife had incurred the same medical expenses as a result of a broken leg she would have been covered. Moreover, if we treat the baby bonus as a de facto maternity benefit, MTC was paying higher benefits in regard to pregnancy to its female than to its male employees, and thus may have been violating Title VII without regard to the Pregnancy Discrimination Act. MTC was not excluding pregnancy from the class of medical conditions or disabilities covered by employer benefits; it was merely treating male and female employees differently with respect to pregnancy; and that is, or at least could be thought, sex discrimination, albeit discrimination against male rather than against female employees.
It is possible that no one was really hurt by the baby bonus plan. The union may have been trading maternity benefits (more precisely, the difference between those benefits and the baby bonus) for some other form of benefit (or wage) worth more to the employees, including married male employees; in that event there may have been no net discrimination in favor of the female employees, who, as we have just noted, received maternity benefits more generous than the baby bonuses. But this is not an equitable argument; it is an argu[941]*941ment that the Pregnancy Discrimination Act, at least in the context of explicitly bargained-for wages and fringe benefits, is a gratuitous and perhaps idle interference with labor markets — which it may be, but that is not our business.
The argument that retroactive application of the Supreme Court’s interpretation of the Act would be unjust is a separate argument, but also lacks merit. Judicial decisions normally are retroactive; that is, they apply to conduct that occurred before the decision was rendered. This result is due not to the myth that courts do not make, but only find, law, so that their decisions merely declare what was the law all the time, but to a sense that as between two surprised parties the burden of surprise should rest on the party who guessed wrong about the direction in which the law was moving; to recognition that retroactive application is an important check on courts’ innovating too rapidly (they are made to consider the costs of discontinuity); and perhaps to a sense that since statutes themselves frequently upset expectations, see, e.g., Bradley v. School Board of City of Richmond, 416 U.S. 696, 94 S.Ct. 2006, 40 L.Ed.2d 476 (1974), we should not be unduly dismayed when decisions interpreting statutes do so as well.
The major exception to the principle that a judicial ruling applies to conduct engaged in before the ruling was made is for decisions that overrule established precedents on which substantial reliance had been placed. Newport News did not overrule any decision by the Supreme Court, but it did overrule our decision in Joslyn. This would be a highly pertinent consideration, see Saint Francis College v. Al-Khazraji, — U.S. —, 107 S.Ct. 2022, 2025-26, 95 L.Ed.2d 582 (1987); Anton v. Lehpamer, 787 F.2d 1141, 1143-44 (7th Cir.1986), and would serve to distinguish EEOC v. Texas Industries, Inc., 782 F.2d 547 (5th Cir.1986), which applied Newport News retroactively in a circuit that did not have a Joslyn -type precedent — but for the fact that Joslyn was decided after the Vu-citech group negotiated the collective bargaining agreements on which their liability for sex discrimination is based. In almost identical circumstances, the Ninth Circuit applied Newport News retroactively. See EEOC v. Puget Sound Log Scaling & Grading Bureau, 752 F.2d 1389, 1392-94 (9th Cir.1985); see also Goodman v. Lukens Steel Co., — U.S. —, 107 S.Ct. 2617, 2621-22, 96 L.Ed.2d 572 (1987).
Occasionally, it is true, decisions are applied only prospectively even though they have not overruled previous decisions but merely resolved novel and difficult questions. See, e.g., Allen v. State Board of Elections, 393 U.S. 544, 571-72, 89 S.Ct. 817, 834-35, 22 L.Ed.2d 1 (1969). So any decision which (1) “establishes] a new principle of law by overruling clear past precedent on which litigants may have relied or by deciding an issue of first impression whose resolution was not clearly foreshadowed” is a candidate for prospective application, provided that such application (2) will serve (or at least not disserve) the operation of the new principle of law and (3) retroactive application would be inequitable. Chevron Oil Co. v. Huson, 404 U.S. 97, 106-07, 92 S.Ct. 349, 355, 30 L.Ed.2d 296 (1971).
The first condition is satisfied here: the Supreme Court’s interpretation of the Pregnancy Discrimination Act in Newport News was one of two equally plausible interpretations of an ambiguous statute. Of course, the first condition, stated as broadly as it is stated in Allen and some other cases, such as Keith Fulton & Sons, Inc. v. New England Teamsters & Trucking Industry Pension Fund, Inc., 762 F.2d 1124, 1137 (1st Cir.1984), will often be satisfied; and this has worried some courts, see, e.g., Budinich v. Becton Dickinson & Co., 807 F.2d 155, 159 (10th Cir.1986) (per curiam), cert. granted, — U.S. —, 108 S.Ct. 226, 98 L.Ed.2d 185 (1987), but their worry is misplaced. The first condition is just a threshold. If the ruling was not foreseen or readily foreseeable, the court should at least consider (if asked to) the possibility that retroactive application would cause so much turmoil that the ruling should have only future application. The presumption remains [942]*942strongly in favor of retroactive application, however, and is not rebutted by anything in the record of this case. Prospective-only application of the Pregnancy Discrimination Act would leave some of the Act’s intended beneficiaries without remedy, while retroactive application would neither injure innocent third parties nor be inequitable to the defendants, who have shown no substantial detrimental reliance on what turned out to be an erroneous view of the law. No doubt they would not have paid the baby bonuses if they had known what construction the Supreme Court would place on the Pregnancy Discrimination Act, but the district court allowed them to deduct those bonuses from the award to the EEOC. They have made no attempt to show that the loss of $14,000 — the entire liability of the Vucitech group — a sum they have already deposited with the district court against the possibility of our affirming the judgment, would work unusual hardship upon them; and we note that, although there is no general right of contribution under Title VII, see Northwest Airlines, Inc. v. Transport Workers Union of America, 451 U.S. 77, 101 S.Ct. 1571, 67 L.Ed.2d 750 (1981), the district court can order contribution among parties actually named as defendants in the Title VII suit, In re Burlington Northern, Inc. Employment Practices Litigation, 810 F.2d 601, 610 (7th Cir.1986). If particular members of the Vucitech group are too impecunious to bear their shares of the joint liability, this was something for them to show; no such showing was attempted.
Laches (the Vucitech group’s next argument) is usually invoked in situations where delay in prosecuting a claim has made the claim harder to defend against, because of the death of witnesses or other developments. See Goodman v. McDonnell Douglas Corp., 606 F.2d 800, 808 and n. 17 (8th Cir.1979). There was no such harm here, as ever since Newport News the defendants have had by their own acknowledgment no defense against the charge of sex discrimination. A dictum in Cannon v. University of Health Sciences/The Chicago Medical School, 710 F.2d 351, 361 (7th Cir.1983), states that the impairment of a defendant’s ability to defend is the only kind of harm that is relevant to a defense of laches against a suit by the EEOC. The court thought this result compelled by a remark in Occidental Life Ins. Co. v. EEOC, 432 U.S. 355, 373, 97 S.Ct. 2447, 2458, 53 L.Ed.2d 402 (1977), that if “a defendant in a Title VII enforcement action [was] significantly handicapped in making his defense because of an inordinate EEOC delay in filing the action after exhausting its conciliation efforts,” the court could “restrict or even deny backpay relief.” Although Occidental does not mention laches by name, the allusion seems unmistakable, since the case holds that there is no statute of limitations in suits by the EEOC. But nothing in the opinion suggests that the Court wanted to confine the district court’s power to instances where the defendant had been harmed in mounting a defense rather than in some other way. That is not a conventional limitation on the defense. If the defendant in a suit for trademark infringement had been led to believe that the plaintiff would not sue, and in reliance expended large sums of money on exploiting the allegedly infringing trademark, he can plead laches if the plaintiff later sues, even if the delay has not impaired his ability to defend in the slightest. See, e.g., NAACP v. N.A.A.C.P. Legal Defense & Educational Fund, Inc., 753 F.2d 131, 137 (D.C.Cir.1985); Citizens & Landowners Against the Miles City/New Underwood Powerline v. Secretary, United States Dept. of Energy, 683 F.2d 1171, 1177 (8th Cir.1982).
The dictum in Cannon is also in tension with our decision in Lingenfelter v. Keystone Consolidated Industries, Inc., 691 F.2d 339 (7th Cir.1982) (per curiam), upholding a defense of laches in a suit under the Veteran’s Reemployment Rights Act, where the principal harm to the defendant was the expense of the wages he had paid substitute workers and would now (if the suit succeeded) have to pay over again to the plaintiff, in the form of backpay. Lin-genfelter was not a suit by the EEOC and did not arise under Title VII, but why either point should make a difference is [943]*943unclear when as in this case the EEOC is suing as the representative of employees who could have brought their own case, see 42 U.S.C. § 2000e-5(f)(1); Occidental Life Ins. Co. v. EEOC, supra, 432 U.S. at 360-61, 97 S.Ct. at 2451-52. And there are decisions that apply the same notion of laches as in IAngenfelter to cases seeking backpay under Title VII, including our own Kamberos v. GTE Automatic Electric, Inc., 603 F.2d 598, 603 (7th Cir.1979), although the discussion of the point in Kamberos is not as clear as one might wish.
So, to put our own dictum against that of the panel in Cannon, if the defendants in this case could show that as a result of the EEOC’s delay in suing they reasonably had expended large sums of money on employee benefits that unexpectedly — unpredictably — turned out to be illegal, resulting in a staggering unanticipated further liability, we do not see why they could not plead laches. But no more than in Cannon need the issue be decided in this case, and we have gone into it only to make clear that it is still an open question in this circuit. Among many other differences between the hypothetical case just put and the present case, the baby bonus plan was held, not illegal, but insufficient, for remember that the defendants have been allowed to deduct the baby bonuses from the payments that the district court has ordered them to make to the EEOC. Moreover, the EEOC’s delay in suing, the focus of a laches defense, caused no harm of any sort to the Vucitech group. Compare In re Bohart, 743 F.2d 313, 324-29 (5th Cir.1984). The group may even have benefited, from delay in having to pay the judgment; but we cannot be sure of this, because the district court’s award to the EEOC includes prejudgment interest at rates ranging from 6 to 20 percent.
The defendants say that the EEOC, had it sued MTC in 1982, might have obtained and executed a judgment before MTC collapsed in the wake of Niazi’s malfeasance. This argument might seem to be saying no more than that if only the EEOC had acted faster, an entity not in fact responsible for the unlawful conduct (at* least not primarily responsible — for Nia-zi could have repudiated as illegal the baby bonus provision of the collective bargaining agreement that he inherited, and he didn’t do so) — MTC in Niazi’s hands — might nevertheless have been forced to bear the costs of it. But this characterization is unfair to the members of the Vucitech group. MTC had, after all, expressly assumed their liability for sex discrimination, and presumably they had compensated Nia-zi in one form or another to do this. Still, the argument that the Vucitech group would not have been left holding the bag if the EEOC had acted faster cannot carry the day. This is not only because MTC might have fought off a judgment till its collapse, and not only because any further depletion of MTC’s assets would simply have increased the amount of money that the Vucitech group had to pay to get back into the business, so that they would have paid for the judgment indirectly; it is also because there was nothing unreasonable in the EEOC’s delay. The Commission acted reasonably in waiting until Newport News was decided before flooding the courts with suits over maternity benefits; the federal courts have enough business without encouraging government agencies to file protective lawsuits designed to avoid findings of laches. That brought matters up to June 1983. By this time MTC was only five months from collapse, and it is absurd to suggest that the, EEOC should have sued and obtained a judgment within those five months. Thereafter, whether the EEOC waited 18 days or (as it did) 18 months to name the Vucitech group as defendants made no difference to them, except to delay the payment of the judgment by that much more time. Nor was the 18-month delay unreasonable, since it was taken up by efforts to settle with MTC and, when those efforts failed, to determine who had been the individuals responsible for the discrimination practiced by the corporation.
The further argument that the EEOC’s delay in naming the members of the Vuci-tech group as defendants prevented them from settling with the Commission at the [944]*944conciliation stage fails on two grounds: there is no indication that the Commission would have settled for less than the modest amount eventually awarded by the district court, and the unwillingness of the group to settle on MTC’s behalf makes it unlikely that they would have settled on their own behalf. If the prospect of personal liability might be more intimidating than that of corporate liability, still settlement of personal liability would have required an immediate outlay from the group’s personal funds, and it is doubtful that they would have been willing to make such an outlay. We point out in this connection that persons not named in EEOC charges can sometimes be sued in private actions, see Eggleston v. Chicago Journeyman Plumbers’ Local Union No. 130, 657 F.2d 890, 905-06 (7th Cir.1981), so perhaps the Vuci-tech group should have known at the time that it was negotiating with the EEOC over MTC’s liability that they might be sued, either by the employees they had discriminated against or by the EEOC on behalf of those employees.
We turn to the cross-appeal. It may seem moot, the Vucitech group having deposited with the district court enough money to satisfy the judgment. But it is not, because, as noted earlier, two of the employees whom the EEOC is representing in this litigation were denied maternity benefits after Niazi took over MTC and the Vucitech group retired. The members of the group were not personally involved in the discrimination against these two workers; the only possible defendant, with MTC down the tubes, is Profile.
The entire issue of successor liability, which is so important in regard both to common law torts (especially of the delayed-action variety, where the culpable act and the injury may be separated by many years) and to statutory torts such as discrimination in violation of Title VII, is dreadfully tangled, reflecting the difficulty of striking the right balance between the competing interests at stake. See generally Roe, Mergers, Acquisitions, and Tort: A Comment on the Problem of Successor Corporation Liability, 70 Va.L.Rev. 1559 (1984); Green, Successor Liability: The Superiority of Statutory Reform to Protect Products Liability Claimants, 72 Com.L.Rev. 17 (1986). In favor of successor liability is the interest in preventing tortfeasors from externalizing the costs of their misconduct by selling their assets free of any liabilities and distributing the proceeds to their shareholders. See Silici-ano, Corporate Behavior and the Social Efficiency of Tort Law, 85 Mich.L.Rev. 1820, 1854-55 (1987). Against is the interest in a fluid market in corporate assets, which is impeded if purchasers acquire along with the assets legal liabilities of unknown, sometimes unknowable, dimensions. The latter consideration dominated common law thinking until recent years, producing a rule, now eroding (see, e.g., Phillips, Product Line Continuity and Successor Corporation Liability, 58 N.Y.U.L.Rev. 906 (1983)), that in a sale of assets (as here), as distinct from a merger or consolidation, the purchaser took free of any liabilities not expressly assumed, including tort liabilities. See, e.g., Forest Laboratories, Inc. v. Pillsbury Co., 452 F.2d 621, 625-26 (7th Cir.1971); Note, Purchase of Assets and Successor Liability: A Necessarily Arbitrary Limit, 11 Del.J. Corp.L. 791 (1986).
A similar but looser approach, in which the focus is on the continuity between the predecessor’s and successor’s businesses and the latter’s notice of the former’s acts, has long been followed in labor cases in which the issue is the successor’s duty to honor the obligations assumed by its predecessor in a collective bargaining agreement. See, e.g., Fall River Dyeing & Finishing Corp. v. NLRB, — U.S. —, 107 S.Ct. 2225, 2235-37, 96 L.Ed.2d 22 (1987); NLRB v. Burns International Security Services, Inc., 406 U.S. 272, 278-81 and n. 4, 286, 92 S.Ct. 1571, 1577-79 and n. 4, 1581, 32 L.Ed.2d 61 (1972). These cases have in turn been thought to provide an appropriate model for dealing with succes-sorship questions in Title VII cases. See, e.g., Bates v. Pacific Maritime Ass’n, 744 F.2d 705 (9th Cir.1984); EEOC v. MacMillan Bloedel Containers, Inc., 503 F.2d 1086, 1089-92 (6th Cir.1974). A liability is [945]*945a liability, whatever its legal source; and while circumstances alter cases, the basic issue in every successorship case is how to strike a proper balance between on the one hand preventing wrongdoers from escaping liability and on the other hand facilitating the transfer of corporate assets to their most valuable uses.
There is, however, the following difference between the tort and labor contexts: in the latter, there is no question that the successor knows of any collective bargaining agreements that his predecessor has signed; in the former, the successor may be ignorant of the predecessor’s liability. Employment discrimination, a tort in a labor context, is thus a mixed case. Another mixed case is the successor’s liability for a predecessor’s unfair labor practice; and here as we would expect there must be a showing of notice to the successor. See Golden State Bottling Co. v. NLRB, 414 U.S. 168, 185, 94 S.Ct. 414, 425, 38 L.Ed.2d 388 (1973).
Whatever the precise formulation of the rule should be in employment-discrimination cases, there can be little doubt that Profile is liable as MTC’s successor. Although the particular discrimination in question occurred after the Vucitech group had temporarily departed from the scene, it was merely a continuation of the baby bonus policy which the group had adopted and, as it were, bequeathed to its successor, Niazi. The collective bargaining agreement that the group negotiated in 1982 and that contained the $1,000 baby bonus did not expire until February 1985. By that time the group was back in control, and knew or should have known that the original charges of sex discrimination had not been resolved — and upon inquiry of the EEOC about their status would quickly have learned that two additional charges had been filed. So there was no surprise; or at least there should not have been, and of course the Vucitech group’s knowledge, actual or constructive, must be imputed to Profile, which they controlled. In addition, the Vucitech group was at least somewhat implicated in Niazi’s violations of the Pregnancy Discrimination Act, which were pursuant to the collective bargaining agreement that the group had negotiated in 1982. The amount in controversy is again very small — a mere $2,000 — making us wonder at the motives behind the defendants’ steadfast resistance to paying the maternity benefits that they unlawfully denied their employees’ wives. And these amounts were known from the outset; there was no open-ended liability as in a tort suit.
Finally, there is the “substantial continuity” between predecessor and successor of which the cases speak; indeed, the facts in Fall River Dyeing are rather similar to those of the present case. This bears not only on the issue of notice, but also on the issue of maintaining the fluidity of corporate assets, which is impeded if assets sold piecemeal are each encumbered by the liabilities of their previous owner. When the successor company knows about its predecessor’s liability, knows the precise extent of that liability, and knows that the predecessor itself would not be able to pay a judgment obtained against it, the presumption should be in favor of successor liability, even if the successor (as here) purchased the assets of its predecessor rather than merged the predecessor into it or consolidated with it. This solution prevents the externalizing of the liability without disappointing the reasonable expectations of investors (and hence impeding the market for corporate assets). The successor, if he knows of his potential liability, will demand compensation in the form of a lower price for the assets, and in this way the burden of liability will be shifted back to the owners of those assets, where it belongs.
In rejecting successor liability, the district court attached great and in our view undue weight to the fact that because MTC had gone broke, there was an element of windfall in making Profile liable as its successor; had it not been for the succession, the two employees involved in this aspect of the case would have been out of luck. Two recent decisions of this court, Musikiwamba v. ESSI, Inc., supra, and Wheeler v. Snyder Buick, Inc., 794 F.2d 1228, 1236 [946]*946(7th Cir.1986), say that successor liability is inappropriate where prior to the succession the predecessor lost its ability to pay a judgment, so that as in this case, but for the succession, the plaintiffs would depart empty-handed. However, we do not understand these decisions to have imposed an ironclad requirement in all cases of successor liability. Certainly in an equity case, as Wheeler was and this case is (as the Vuci-tech group tirelessly emphasizes), the proper approach to the issue of successor liability is not to erect a set of hoops to force plaintiffs to jump through but to ask whether such liability would strike a reasonable balance between the interest in fully sanctioning unlawful conduct and the interest in facilitating the market in corporate and other productive assets. Cf. John Wiley & Sons, Inc. v. Livingston, 376 U.S. 543, 548-51, 84 S.Ct. 909, 913-15, 11 L.Ed.2d 898 (1964); Barksdale, Successor Liability Under the National Labor Relations Act and Title VII, 54 Tex.L.Rev. 707, 734 (1976). The disappearance of the tort-feasor (the predecessor in liability) before the succession is relevant, because it reduces the likelihood both that the succession was a device for escaping the consequences of liability and that the successor had notice of its predecessor’s legal liability. MTC did not go broke as part of some elaborate scheme to escape the trivial liability arising from the failure to pay maternity benefits, and to this extent the point emphasized by the district court and our previous decisions does indeed weigh against successor liability. But the point is outweighed by the fact that MTC’s discrimination was a legacy of the Vucitech group, which operating in the corporate form of Profile took over the company once again after Niazi’s defalcation. The Niazi interlude is an irrelevance; the discriminations, built as they were into the collective bargaining agreement that the Vucitech group had negotiated, would have occurred even if the group had never sold their stock to Niazi, in which event MTC rather than Profile would be liable to the two employees in question. We conclude that Profile is liable as a successor employer, and hence that the EEOC must succeed on its cross-appeal.
Affirmed in Part and Reversed in Part.