dissenting:
I.
, Soon after learning of a bid by an outside entity to take over Cone Mills Corporation, the Board of Directors of Cone Mills met to evaluate the alternatives available to the company. Of the alternatives considered, the preferred option was a leveraged buy-out (LBO) by the management. Chairman Trog-don and other senior executive management presented to the Board an offer to acquire all outstanding shares of common stock in Cone Mills, subject to the arrangement of satisfactory financing, government approval, and approval by Cone Mill’s shareholders.
As could be expected, news of the hostile bid proposal and the possible LBO by management created concern among the employees about their job security, their pension benefits, and the future of the company. To still the disquiet and keep the employees informed, Cone Mills, through Chairman Trogdon, communicated with its employees— by letter, newsletter, and video presentation — about various changes that would be made in their employee benefit plans if the LBO was approved. Trogdon repeatedly assured Cone Mills’s employees and shareholders that management would protect their interests and keep them informed of any changes that would occur.
One of the changes contemplated by management in the course of implementing the LBO was the revision of three preexisting pension plans and the adoption of an Employee Stock Ownership Plan (the 1983 ESOP). On December 12, 1983, Trogdon wrote a letter to all Cone Mills employees, assuring employees that their benefits under the existing pension plan (ERP) would be kept in place, and that under the new structure, “you can receive no less than the full amount of your pension benefits.” The letter went on to say that “subject to legal and tax rulings, it appears that in the first two years (1984-85), over fifty million dollars of stock could be contributed to your ESOP.”
Two days later Trogdon wrote another letter, addressed to salaried employees only, assuring them that the company’s ERP would be kept in place and would operate in tandem with the new 1983 ESOP.
The letter informed the employees that the ERP was over-funded, and that if the LBO went through, management planned to contribute the surplus or its equivalent in company stock to the ESOP. Indeed, the letter told the employees that there was an “agreement among management, and the banks,” that, if the LBO was approved, the surplus from the ERP or its equivalent in company stock would be contributed to the ESOP. “When the transaction is executed and the *1041contribution is made, you, I, and all other Cone employees will ‘take title’ to a substantial asset in which we currently have no rights or ownership.”
On March 26, 1984, Cone Mills shareholders approved the management LBO by a nearly unanimous vote. Ninety-nine percent of Cone Mills’ salaried employees voted for the LBO. Salaried employees owned 71,582 shares, or 1.3% of Cone Mills common stock, and the management group controlled 288,-532, or 5.23%.
Not until April 2, 1984 did management execute the 1983 ESOP plan documents. The executed documents made no mention of the promised contribution to the ESOP of the pension reversion surplus or its equivalent. The documents laid out a formula for contribution to the fund whereby Cone Mills would contribute stock to the fund on the. basis of a “10/10/1 formula.” The surplus reversion of the pension fund, when collected by Cone Mills, totaled $69 million. Cone Mills paid $54.8 million of this amount into the ESOP, or $14.2 million less than the full amount of the surplus.
II.
The majority opinion has taken care to outline the reasons why the plaintiff class has no cause of action under the terms of ERISA to enforce the representations made by Cone Mills to its employees regarding the pension reversion surplus. Building upon the basic tenet of ERISA that employee benefits plans should be reduced to written instruments, the majority reasons that any promise not contained in a written plan document need not be enforced under ERISA. It points out that the 1983 ESOP plan documents were not executed until April 2, 1984 — well after the letters promising that the surplus would be contributed to the 1983 ESOP — and as adopted, did not mention the surplus. Instead, the official plan documents adopted a “10/10/1 formula” for contribution to the ESOP, which it has scrupulously followed. Thus, according to the majority, the defendants had no fiduciary duty to honor the pension reversion surplus promise and in fact were prohibited from honoring it because of their fiduciary duty to adhere to the official plan documents. But see Fischer v. Philadelphia Elec. Co., 994 F.2d 130 (3rd Cir.1993) (ERISA’s fiduciary standards do not permit a plan administrator to make affirmative material misrepresentations to plan participants about changes that are being considered in its employee benefits plan).
The majority supports its argument that the pension reversion surplus promise may not be enforced by extending the rule in Pizlo v. Bethlehem Steel Corp., 884 F.2d 116 (4th Cir.1989), to the instant facts. In Pizlo, we held that existing, written plan documents may not be modified by oral or informal and unauthorized amendments. Pizlo, 884 F.2d at 120. See also Nachwalter v. Christie, 805 F.2d 956, 960 (11th Cir.1986); Moore v. Metropolitan Life Insurance Co., 856 F.2d 488, 489 (2nd Cir.1988); Alday v. Container Corp. of America, 906 F.2d 660, 665 (11th Cir.1990). The purpose of the Pizlo rule against informal amendments is to provide certainty in the provision of benefits to employees. Without a rule requiring adherence to the written formal document, employees .would be unable to ascertain with certainty their benefits, individual employees might receive benefits intended to benefit employees as a group, and employers might be discouraged from informing their employees about proposed changes or new plans.
The majority extends the Pizlo rule to preclude the enforcement of any and all extra-plan promises made to employees, including promises formally extended by authorized company officials in written form. It construes the rule that ERISA fiduciaries have no obligation to fulfill promises not contained in a written plan to apply even where no written plan exists at the time the promises are made.
Yet there is some irrationality in applying the same rule that prohibits enforcement of a promise extended after the adoption of plan documents to promises made prior to the adoption of the official plan documents, especially when the pre-plan promise pertains to what employees can expect when the proposed plan takes effect. Application of the Pizlo rule assumes an existing,-available for inspection, formal plan, and charges employ*1042er and employee with the duty to compare its terms with those of the promise. The instant circumstances differ in a most significant respect: 'there was no existing plan with which the promise made could be compared. The ability of the employees to ascertain their benefits and act accordingly was thus frustrated. The rule announced in Pizlo against informal modifications of an extant written plan has vitality in the circumstances where a plan document exists at the time of the promise. But here it is clear that Trog-don’s letters did not modify or amend or contradict an existing plan document; rather, they made false promises to employees about a proposed new plan that failed to make good on the promise when, months later, it finally came into existence.
After reversing the district court and holding that the employer had no fiduciary duty to fulfill the pension reversion surplus promise, the majority goes even further and rejects the district court’s holding that the employees may enforce the promise on an equitable estoppel or third party beneficiary contract theory.1 The majority holds, in other words, that ERISA’s coverage is not broad enough to prohibit the misrepresentation concerning the' proposed benefits plan because only a formal written plan may be referred to when ascertaining fiduciary duties, but is so exclusive that it preempts any state common law actions and even precludes the court from applying common law principles to provide equitable relief.
However, it seems to me that the authority the majority relies upon to reject the use of common law equitable and contract theories is distinguishable. None of the cases cited deal with a situation similar to the one we here encounter — an employer materially misleading plan participants about the terms of the plan to be adopted and thereby securing good will, employee stability and their desired votes for the leveraged buy-out. Instead, the cases cited by the majority deal with an attempt by employees to hbld employers to extra-plan benefits promises made after a plan was adopted. See, e.g., Singer v. Black & Decker Corp., 964 F.2d 1449 (4th Cir.1992) (employees sought enhanced early retirement benefits not provided for in the existing benefits plan, based on the contention that offers by the company of a limited special retirement program that temporarily operated to amend the existing plan constituted an implied representation that similar benefits would be offered in the future); Coleman v. Nationwide Life Ins. Co., 969 F.2d 54 (4th Cir.1992) (plaintiff sought benefits under group health insurance plan where company personnel had represented to her that she was covered, when in fact her eligibility had automatically terminated under the terms of the written plan), cert. denied, — U.S. -, 113 S.Ct. 1051, 122 L.Ed.2d 359 (1993).
I am loath to conclude that a company may, with impunity, make representations to its employees about a benefits plan — with the implicit purpose of assuring employees that their benefits will be enhanced if a proposed leveraged buy-out is successful — and then renege on those promises when the LBO is effectuated and the plan is adopted. The majority’s approach allows (and perhaps invites) employers to make promises to employees about a plan still in gestation thereby eliciting desired behavior, and then disavow those promises and hide behind the protection of a written plan which does not incorporate the promises made. Such a result would make a mockery of the most fundamental canon of the ERISA statutory scheme — that “a fiduciary shall discharge his duties with respect to a plan solely in the interest of the participants and beneficiaries,” and shall exercise those duties with “the care, skill, prudence, and diligence” of a “prudent man acting in like capacity.” 29 U.S.C. § 1104. See Massachusetts Mut. Life Ins. Co. v. Russell, 473 U.S. 134, 157, 105 S.Ct. 3085, 3098, 87 L.Ed.2d 96 (1985) (“[Cjourts must always bear in mind the *1043ultimate consideration whether allowance or disallowance of such relief would better effectuate the underlying purposes of ERISA— enforcement of strict fiduciary standards of care -in the administration of all aspects of pension plans and promotion of the best interest of participants and beneficiaries.”) (Brennan, J., concurring); see also Berlin v. Michigan Bell Telephone Co., 858 F.2d 1154, 1168 (6th Cir.1988) (rejecting the view that any communications or representations .made prior to a decision to adopt a plan are nonfiduciary); Fischer, 994 F.2d at 133-135 (allegations that employer breached fiduciary duty by knowingly or negligently making material misrepresentations regarding a change to a plan during period when change was under serious consideration stated a claim that could not be dismissed on summary judgment). Moreover, as explained below, the statute itself, and our own Fourth Circuit precedent indicate that this court would be well within the discretion granted to it by the Congress under the ERISA statute in providing equitable relief to the employers by looking to common law estop-pel or contract theories.2
III.
Besides granting federal courts the power to enjoin any act that violates a provision of ERISA or the terms of a benefits plan, ERISA specifically permits federal courts to grant “other appropriate equitable relief.” See 29 U.S.C. § 1132(a)(3)(B). We have recently recognized that, in enacting ERISA, Congress “intended that the courts would develop a federal common law of rights and obligations under ERISA-regulated plans.” Singer v. Black & Decker Corp., 964 F.2d 1449, 1452 (4th Cir.1992) (quoting Firestone Tire & Rubber Co. v. Bruch, 489 U.S. 101, 109 S.Ct. 948, 103 L.Ed.2d 80 (1989)) (quotations and internal citations omitted). See also Holland v. Burlington Indus. Inc., 772 F.2d 1140, 1147 n. 5 (4th Cir.1985) (Congress intended for courts to borrow from state law when appropriate in fashioning federal common law to govern ERISA), aff'd sub nom Brooks v. Burlington Indus. Inc., 477 U.S. 901, 106 S.Ct. 3267, 91 L.Ed.2d 559 and cert. denied sub nom. Slack v. Burlington Indus., Inc., 477 U.S. 903, 106 S.Ct. 3271, 91 L.Ed.2d 562 (1986). In Singer, we acknowledged the difficulty courts may encounter in delineating the situations where federal common law would be a “desirable companion to ERISA,” from those where it would unduly expand rights and remedies established by the statutory scheme. But we also admonished the district court for “paint[ing] with too broad a brush in stating categorically that courts should' not look to state common-law causes of action that have been preempted by ERISA in fashioning federal common law.” Singer, 964 F.2d at 1452. Indeed, the Singer court defended the use by a federal court of common law doctrines “to assist in shaping the federal common law of ERISA,” by noting that use of common law doctrines “to further[] the goals of ERISA” would advance, not frustrate, the intention of Congress that ERISA subject plans and plan sponsors to a uniform body of benefits law. Id. at 1452-53. An obvious source for. federal common law would be existing state common law applicable in analogous situations. Id.
In developing a federal common law of rights and obligations under ERISA-regulat-ed plans, courts must be cautious. The doctrines of equitable or promissory estoppel or third party beneficiary or. unilateral contract theory should not be used to alter the terms of established benefits plans or to undermine the purposes of the federal statutory scheme. See id. at 1453-54 (Wilkinson, J. concurring). We recently determined in Coleman, 969 *1044F.2d 54, that estoppel principles could not be used to enforce an informal statement to an employee that contradicted an explicit term of the extant benefit plan. In Coleman, the plaintiff argued that her employer was es-topped from denying her requested medical benefits because it had verbally represented to her that she was still covered by the policy when, in fact, by the terms of the plan, her coverage had been automatically terminated. The court denied relief, noting that “[u]se of estoppel principles to effect a modification of a written employee benefit plan would conflict with ‘ERISA’s emphatic preference for written agreements.’ ” Coleman, 969 F.2d at 58 (quoting Began v. Ford Motor Co., 869 F.2d 889, 895 (5th Cir.1989)).
The Coleman rule is a sound one. Equitable estoppel should not be employed to give force to unauthorized, oral promises that contradict the written terms of an established plan. I do not, however, agree with the majority that Coleman and Singer preclude the application of estoppel theory in the instant case. Neither case involved a scenario similar to the one we here encounter. In particular, neither case, it seems to me, instructs us to deny equitable relief where no plan exists when a promise is made and relied upon, and the promise pertains to what the contemplated plan would provide.
I also disagree with the majority’s conclusion that our prior precedent precludes the employees from going forward on a third-party beneficiary theory, or indeed upon a theory of unilateral contract. No case within our circuit has held that the application of common law contract principles is inappropriate in every circumstance. On the contrary, in Provident Life & Acc. Ins. Co. v. Waller, 906 F.2d 985 (4th Cir.1990), we explicitly concluded that “fashioning a federal common law rule of unjust enrichment” was appropriate “in the circumstances of this case.” Id. at 993. In making that determination, the Provident Life court looked to the “plan contract itself, the statutory policies of ERISA, and state law,” and decided that the enrichment doctrine would further the clear intent of the plan, was consistent with the statutory provisions of ERISA, and would “fit the archetypal unjust enrichment scenario.” Id. at 992-93.
My reading of Provident Life, Singer, and Coleman leads me to conclude that they do not prohibit the application of common law principles such as estoppel, third-party beneficiary, or unilateral contract theory in ERISA cases, but rather counsel caution in the use of those doctrines. Other circuits are in accord-. See, e.g., Fischer, 994 F.2d at 136 (permitting plaintiffs to proceed on their claim under ERISA on theory of equitable estoppel where employees retired in reliance upon a misrepresentation by employer that a retirement ‘swe'etener’ was not being considered for adoption); Kane v. Aetna Life Ins., 893 F.2d 1283 (11th Cir.1990) (holding that although estoppel may not be invoked to amend, modify, enlarge, or extend coverage, it may be used to hold insurer to its agent’s representations regarding the interpretation of an ambiguous plan provision), cert. denied, 498 U.S. 890, 111 S.Ct. 232, 112 L.Ed.2d 192 (1990); Cleary v. Graphic Communications Int'l Union, 841 F.2d 444, 447-49 (1st Cir.1988) (recognizing the applicability of estop-pel principles, but holding that informal representations of the plan administrator did not constitute a representation that benefits would be paid); Haeberle v. Board of Trustees, 624 F.2d 1132, 1139 (2nd Cir.1980) (recognizing the applicability of estoppel to ERISA actions, but requiring strict proof of its elements). Happily, Singer provides several principles that “have emerged as guides for the courts in demarking those situations in which the development of federal common law is inappropriate,” and, by extension, those situations in which the use of common law principles are warranted. Singer, 964 F.2d at 1452. Singer indicates that it is appropriate to apply federal common law when the application will not run counter to the policies of ERISA or the terms of an extant writtén plan:
Importantly, courts must be conscientious to fashion federal common law only when it is necessary to effectuate the purposes of ERISA. Thus, resort to federal common law generally is inappropriate when its application would conflict with the statutory provisions of ERISA, discourage employers from implementing plans governed *1045by ERISA, or threaten to override the explicit terms of an established ERISA benefit plan. And, courts should remain circumspect to utilize federal common law-to address issues that bear at most a tangential relationship to the purposes of ERISA.
Id. (internal quotations and citations omitted).
A.
In applying Singer's principles to the instant case, we must first determine whether, in this case, estoppel, unilateral contract, or third party contract theory would conflict with the statutory provisions of ERISA. According to the majority, application of estop-pel theory to enforce the pre-plan promises made by Trogdon would undermine the ERISA provision that employee benefits plans be established and administered pursuant to a written plan, see 20 U.S.C. § 1102(a)(1), and the subsequent conclusion that the written plan be the sole benchmark in an ERISA action. See Coleman, 969 F.2d at 59.
The majority’s emphasis on adherence to the written plan documents as its justification for ignoring any and all employer promises regarding prospective plans is unwarranted. The rationale apparently derives from language in Coleman and the concurring opinion in Singer, both of which were concerned with extant written plans. Coleman referred to ERISA’s “emphatic preference” for written agreements, and Singer stated that ERISA “demands adherence” to the written terms of an ■ employee benefit plan. See Coleman, 969 F.2d at 58; Singer, 964 F.2d at 1453 (Wilkinson, J. concurring). ERISA does indeed indicate a preference for written plans. But the preference does not constitute a strict statutory prerequisite for coverage under the act. ERISA does not insist on a written plan when the promise of' bringing it about is still in gestation. The preference for written plans over oral promises has been gleaned from §§ 1102(a)(1) and (b)(3) of the ERISA statute, which provide, in pertinent part, that “[ejvery employee benefit plan shall be established and maintained pursuant to a written instrument ... [and shall] (3) provide a procedure for amending such plan ...” These provisions, however, are merely directives to plan administrators and fiduciaries of plans covered by ERISA:
ERISA does not, however, require a formal, written plan_ There is no requirement of a formal written plan in either ERISA’s coverage section ERISA § 4(a), 29 U.S.C. § 1003(a), or -its definitions section, ERISA .§ 3(1), 29 U.S.C. § 1002(1). Once it is determined that ERISA covers a plan, the Act’s fiduciary and reporting provisions do require the plan to be established pursuant to a written instrument, ERISA §§ 102 and 402, 29 U.S.C. §§ 1022 and 1102; but clearly these are only the responsibilities of administrators ... and are not prerequisites to coverage under the Act.
Donovan, 688 F.2d at 1372. See also Scott v. Gulf Oil, 754 F.2d 1499, 1503 (9th Cir.1985), Clearly, the fiduciary strictures of ERISA are as applicable to benefits “plans” that-are, through breach of promise, not formally adopted or not accurately reduced to written documents as they are to formally adopted written plans. See Donovan, 688 F.2d at 1372. Thus the written plan does not play such an exalted role in the statutory scheme that its mere subsequent existence and a fiduciary’s adherence to its terms provide an excuse for permitting serious prior misrepresentations on the part of the fiduciary.
Similarly, the conclusion derived from the preference for written plans — that informal modifications of established plans are unen-foreeable — does not necessarily lead to the conclusion that promises that relate to benefits plans and are made prior to official adoption of a plan may never -be enforced. The fact that a written plan is eventually adopted does not cure the previous misrepresentation, especially if employees have already acted in -reliance upon the -misrepresentation. The later adopted plan cannot, in practical terms, operate retroactively to provide timely notice to the employees of the discrepancy between the promise and the plan. Why, then, should eventual adoption of a plan protect the company against charges that it materially misled its employees or breached *1046a unilateral contract for which consideration was given?
In sum, the rule that estoppel, unilateral contract, or third party contract principles should not be used when they conflict with the ERISA preference for written contracts or violate the edict against informal modification of established plans does not prevent their application under the facts of the instant case. The more pertinent inquiry, as I see it, is whether, given the peculiarities of the instant case, the application of estoppel or contract principles might be “necessary to effectuate the purposes of ERISA.” Provident Life, 906 F.2d at 992 (quoting U.S. Steel Mining Co. v. District 17, United Mine Workers of America, 897 F.2d 149, 153 (4th Cir.1990)).
A review of cases discussing the policies and purposes of the ERISA statutory scheme reveals several recurring themes. ERISA was enacted, first and foremost, to protect employees and to secure promised benefits. ERISA aspires to preserve the financial integrity of plan funds for the benefit of equitable distribution to all plan participants; it favors a uniform body of federal law ápplicable to all employee benefits plans; it seeks to avoid confusion over the terms of benefits plans by compelling fiduciaries to establish and maintain plans pursuant to one official written document; it endorses full disclosure to plan participants; and it holds plan fiduciaries to the highest standards of care in the management of employee benefits plans. See, e.g., Donovan v. Dillingham, 688 F,2d 1367, 1372 (“the policy of ERISA is to safeguard the well-being and security of working men and women and to apprise them of their rights and obligations under any employee benefit plan”); Coleman, 969 F.2d at 60 (“ERISA is designed ‘to promote the interests of employees and their beneficiaries in employee benefits plans’ ”); Pohl v. Nat’l Benefits Consultants, Inc., 956 F.2d 126, 128 (7th Cir.1992) (“one of ERISA’s purposes is to protect the financial integrity of pension and welfare plans”); Provident Life, 906 F.2d at 993 (ERISA “indicates a desire to ensure that plan funds are administered equitably and that no one party, not even plan beneficiaries, should unjustly profit”); Cleary, 841 F.2d at 447 n. 5 (pension fund stability is of major concern, given the strong policy in ERISA of protecting plan funds for the sake of employee participants); Haeberle, 624 F.2d at 1139 (ERISA provisions that, establish standards of fiduciary responsibility, require reporting and disclosure to employees of current information that affect rights, and" mandate a written plan naming fiduciaries, are “clearly designed to protect pension funds from improper and improvident acts by those entrusted with responsibility for their investment”); Massachusetts Mut. Life Ins. Co. v. Russell, 473 U.S. 134, 157, 105 S.Ct. 3085, 3098, 87 L.Ed.2d 96 (1985) (the underlying purposes of ERISA are “enforcement of strict fiduciary standards of care in the administration of all aspects of pension plans and promotion of the best interest of participants and beneficiaries”) (Brennan, J. concurring).
Holding a company to its written promises about the terms of a soon-to-be-adopted ESOP plan, especially when the promises are made during a time when employee goodwill and favorable voting attitudes are of utmost importance to and sought by the company, seems to me to further the goals of the statutory scheme. Enforcement of a pre-plan promise concerning a proposed plan would not contravene it, it not yet being in existence. Such enforcement would be consistent with the directive that plan administrators be held to strict fiduciary standards and consonant with ERISA’s mandate that employers fully disclose information to its employees in succinct, correct, and easy to understand terms. Such enforcement would also further, rather than frustrate, the goal of clarity over what statements about plan terms rule. If companies may, with impunity, make promises about expected plan terms and then ignore them when the documents themselves are drawn up, confusion on the part of the employees is bound to result. If, on the. other hand, a company is held to its formal pre-plan relied-upon promises or pre-plan promises for which consideration has been given, it is less likely to misrepresent its intentions. Its pre-plan statements will be in conformity with the formal plan documents and no confusion need ensue. Finally, the use of estoppel, unilateral contract, or *1047third party beneficiary doctrine to enforce the promise to contribute the pension reversion surplus into the 1983 ESOP would not favor any individual plan participants to the detriment of the participants as a group and would threaten no damage to, but would rather enhance the financial integrity or stability of the fund since the assets paid to employees will come from outside the plan.
B.
In addition to the mandate that the use of common law doctrines in ERISA actions effectuate rather than obstruct the policies and provisions of the statute, Singer directs courts to investigate whether the use of a common law theory to enforce an extra-plan promise would “threaten to override the explicit terms of an established ERISA benefit plan.” Singer, 964 F.2d at 1452. Here, of course, there was no “established ERISA benefit plan” at the timé the pension reversion surplus promise was made or when the vote took place. The chronology of events was critical. The fact that Cone Mills eventually did adopt a plan that did not contain the promise did not transform the pre-plan misrepresentation into a “threat to the explicit terms” of the later adopted plan. Rather, the failure of Cone Mills to honor its pre-plan promise by contributing the surplus money to the fund laid bare the earlier misrepresentation. Moreover, the plan document itself, as finally adopted, does not by its terms bar Cone Mills from honoring its earlier promise. The “10/10/1” contribution formula formally adopted may be read as merely a guarantee of a minimum contribution to the fund which could easily be supplemented by the pension reversion surplus contribution without alteration, amendment or threat to the terms of the 1983 ESOP plan.
Coleman does not mandate a different .conclusion. The promise in Coleman was a statement assuring an employee that she was covered under a group health insurance plan, when in fact, by the terms of the established plan, she was not. Enforcement of that extra-plan statement would indeed have threatened an explicit term of the extant written plan document which was crucial to maintaining the financial stability of the fund and the equitable distribution of the benefits. Any payments made to Coleman would necessarily have depleted the fund to the detriment of those participants who were covered by the plan’s terms. The pre-plan promise in the instant case does not in any way threaten the ESOP fund itself, or the rights of the plan participants as a whole, for the surplus contribution at issue will be contributed to the fund as a whole from a source outside the fund. Accordingly, I conclude that the use of estoppel or contract principles to enforce the pension reversion promise, made well before the plan documents came into formal existence, would not threaten the terms of the 1983 ESOP plan.
C.
Singer also cautions that common law principles should not be applied where their application would discourage employers from implementing benefits plans. There are myriad variables that weigh in an employer’s calculation of whether to provide pension and benefits plans. Equitable relief granted on the facts of the instant case is hardly likely to be a significant factor. There may be some legitimate concern that a rule holding plan administrators to promises about a plan made before the plan. itself is officially adopted will discourage employers from communicating with employees about the plan. Truthful communications about prospective plans are, obviously, to be encouraged. However, communications that are misleading, unreliable,.unrealistic, or premature,.and which, furthermore, extract reliance or a performance by employees, should, just as obviously, be discouraged. Under ERISA the duty of administrators and employers to be forthright and honest in whatever communications they have with present or future plan participants, at all stages of the development of the plan, is paramount.3
*1048IV.
It is a fact of modern life that employees make a variety of employment, personal, and financial decisions based upon the expectations they hold with regard to their benefits plans. Here, it appears that they went further, and.supported an LBO on the basis of their expectations. It is not unreasonable to suspect that a management group in the position of Cone Mills — threatened by a hostile takeover, scrambling to save the company and their jobs, anxious to stabilize • the workforce and keep employees calm and assured — may be inclined to be less than completely responsible in communications to employees if such irresporisibility is permitted by the courts. I would affirm the determination of the district court that the common law theories of equitable estoppel and third party
contract are viable theories under the circumstances of the case. I would remand to allow investigation not only of the elements of one or both of those two legal concepts but would permit, if the employees chose to pursue it, the doctrine of unilateral contract.
I respectfully dissent.
. The district court’s finding with respect to the federal common law equitable and contract theories was a legal conclusion certified to this court for interlocutory appeal. -The district • court deferred receiving testimony on the issue of reliance, an element of both theories, until after the interlocutory appeal. Affirmance of the district court’s legal conclusion that the employees could recover on the basis of these two theories would necessitate a remand for findings of fact on the issue of reliance.
. For purposes of reviewing the legal conclusion reached by the district court that "equitable es-toppel and third-party beneficiary theories apply to this case,” I assume, without deciding, that. employees could prove the elements necessary to succeed on those theories. We are not in the position to make any such determination, since the district court deferred receiving testimony on the issue o'f reliance. However, the facts also present to me the possibility of the employees prevailing upon a theory of unilateral contract, i.e., when A makes a promise to be fulfilled upon B’s performance of an act, B’s performance binds A to the promise. The employees’ performance — their support and votes for the LBO— was in response to a promise that the pension surplus would be contributed to the proposed 1983 ESOP, and so constituted the consideration necessary to bind Cone Mills to its promise.
. In footnote 14, the majority takes issue with my suggestion that the surplus contribution promise should be enforced according to principles of contract law, adopted into the federal common law of ERISA. The majority asserts that ERISA principles preempt the application of contract theory because the promise was one made "concerning an employee pension benefit plan." See *1048Majority Opinion at page 1034 n. 14. The majority then goes on to suggest in footnote 16 that the fiduciary standards mandated by ERISA do not apply to the surplus contribution promise because those duties "only appl[y] to actions taken by a plan fiduciary in accordance with his duty to administer the employee benefit plan,” and, at the time of the promise, no plan had yet been adopted. See Majority Opinion at page 1036 n. 16. The majority thus tries to have it both ways. I accept that the promise was one concerning an employee benefits plan, and that thus state common law claims are preempted. However, I cannot accept the proposition that, when an employer (during formulation of the plan) makes a promise that “concerns an employee benefits plan," and the employees change their position because of that promise, that the promise and breach thereof are not "actions taken by a plan fiduciary in accordance with his duty to administer the employee' benefit plan." See Fischer v. Philadelphia Elec. Co., 994 F.2d 130, 133-35 (3rd Cir.1993) (fiduciaiy standards of ERISA do not allow employer to make affirmative misrepresentations about its intentions as regards a benefit plan during the formation stage of the plan); Berlin v. Michigan Bell Telephone Co., 858 F.2d 1154 (6th Cir.1988) (employer breached fiduciary duty under ERISA by misleading employees into believing that no early retirement would be adopted when in fact a plan was being seriously considered).
The majority goes on to contend that design of a benefit plan is strictly a business decision left to the discretion of employers. See Majority Opinion at 1036 n. 16, citing Belade v. ITT Corp., 909 F.2d 736, 738 (2nd Cir.1990). Again, while that statement may be assumed to be correct, I do not believe that the freedom ERISA grants employers during formulation stages to design and adopt a plan translates into the freedom to make a promise to employees about a plan, and then renege on it after the employees have relied upon the promise and changed their position. Perhaps it is true, as the majority contends, that Cone Mills, when it made the promise, was not intentionally misrepresenting its plans, and that thus its actions do not amount to fraud. But its change of mind about the surplus after the employees relied on the promise and supported the LBO is certainly not performance of a kind to be tolerated in a fiduciaiy. See Fischer, 994 F.2d at 133 ("Put simply, when a plan administrator speaks, it must speak truthfully.”). The majority defines the scope of ERISA's fiduciaiy duties so very narrowly as to strip them of any real meaning. See Firestone Tire and Rubber Co. v. Bruch, 489 U.S. 101, 110-11, 109 S.Ct. 948, 954, 103 L.Ed.2d 80 (1989) (in considering issues under ERISA that are not covered explicitly by the Act, a court should be guided by principles of trust law); Biggers v. Wittek, 4 F.3d 291, 294 (4th Cir.1993) (same). If ERISA's fiduciaiy duties are to be so impoverished that they permit an employer to hold out a promise of a contribution of $14.2 million to its employees in exchange for the employee's support of a management leveraged buy-out, and then, after the employees have given their support, permits the employer to renege by withdrawing the promise, then it is doubly necessaiy, in order that the goals of ERISA be effectuated, that the concepts of equitable estoppel and contract theory be incorporated into the ERISA federal common law. See Fischer, 994 F.2d at 136 (holding estoppel to be available as a federal common law remedy under ERISA to recover benefits which the plaintiffs did not receive as a result of the defendants' misleading statements).