Esmark, Inc. petitions for review of an order of the National Labor Relations Board (the "NLRB" or "Board") finding Esmark guilty of violations of sections 8(a)(3) and 8(a)(5) of the National Labor Relations Act (the "NLRA"), 29 U.S.C. section 158(a)(3), (a)(5). The Board's order is reported at 289 N.L.R.B. No. 51 (June 29, 1988). We grant Esmark's petition for review and remand the case to the Board for further proceedings.

Esmark is a holding company. Prior to 1981, one of its wholly-owned subsidiaries was Swift & Company, one of America's great meatpacking firms.1 By 1980, Esmark officials had become dissatisfied with Swift's fresh meats division. The fresh meats division required large amounts of cash for its daily operations, was perceived poorly in the financial community and, Esmark believed, was not competitive because of the high labor costs attributable to the company's master agreement with the United Food & Commercial Workers International Union (the "Union").

On April 25, 1980, Esmark's president Donald Kelly asked John Copeland, head of Swift's fresh meats division, to develop a plan for disposition of the division. After the Union rejected an employee stock ownership plan, Esmark began to explore the possibility of selling the fresh meats division as a separate corporation to outside interests. On June 26, 1980, Esmark's board of directors approved a restructuring plan under which several of Swift's fresh meats plants were to be closed, while the remainder would be included in a new corporation and sold.

On June 30, 1980, notices of closing were sent out covering the Moultrie, Georgia and Guymon, Oklahoma plants, which are at the center of this litigation. Originally, Moultrie and Guymon were not to be included in the newly formed corporation, but were to be closed permanently. Copeland subsequently concluded that the Moultrie and Guymon plants were competitive, except for their labor costs, and therefore should be included in the planned stand-alone fresh meats corporation. However, he remained concerned about the wage rates at the two plants. Throughout the latter half of 1980 and early 1981, Copeland and other Swift officials met with the Union's representative, Lewie Anderson, and attempted to persuade Anderson to agree to changes to the master agreement at Moultrie and Guymon. Since the collective agreement was not scheduled to expire until September 1, 1982, Anderson rejected all proposed modifications. Copeland informed Anderson that, if the Union would not agree to concessions, Moultrie and Guymon would be closed before the sale to the independent concern; after the sale, the new entity would reopen the plants, and would be free to set terms of employment at variance with those contained in the master agreement. (Copeland expected that only the Moultrie and Guymon plants would be reopened outside the terms of the existing labor contract--apparently the master agreement wage rates were considered more competitive in other parts of the country.)

During August, September and October of 1980, various transactions were completed in preparation for the divestiture of the fresh meats division. Although these transactions may seem somewhat byzantine, it should be recalled that all of these transactions were carefully orchestrated to achieve maximum tax advantages, as well as for other business reasons, using highly competent legal and financial counsel. First, Swift set up its own wholly-owned subsidiary known as Transitory Food Processors. On October 15, the name of Transitory was changed to Swift & Company, effective October 24. On October 21, (the original) Swift transferred all of its assets, except for the fresh meats division, to Transitory, effective October 27. On October 24, the name of the original Swift & Co. was changed to Swift Independent Packing Company (Sipco).2 To sum up: as of October 27, 1980, Sipco (old Swift) directly owned only the fresh meats operations which had been separated by Copeland; Sipco also owned a subsidiary, (new) Swift & Company, which in turn owned all of old Swift's other assets (principally those connected with processed meats operations). Then, on October 27, Sipco declared a dividend of all of the (new) Swift stock to Esmark. Thus, after October 27, Sipco owned only the fresh meats operations; the other aspects of the old Swift were owned by Esmark, through its wholly-owned subsidiary (old name, new corporation) Swift & Company.

Sipco was to be sold in principal part to outside interests through a public stock offering. But first the structure of the corporation to be sold was further altered. On January 26, 1981, Swift Independent Corporation ("Sic") was incorporated. On February 23, 1981, Esmark and Sic entered into an agreement whereby Esmark sold Sipco's stock to Sic in exchange for a promissory note for $100 million. On April 21, all of Sic's stock was transferred to Esmark, together with a $35 million note payable to Esmark, in exchange for Esmark's surrender of the existing $100 million Sic note. Thus, Esmark now held all of Sic's stock, Sic held all of Sipco's stock, and Sipco owned the fresh meats operations it had always owned, including the Moultrie and Guymon plants. Sic's stock, rather than Sipco's, would be sold to the public.

But the corporate maneuverings were not yet completed. Since only Moultrie and Guymon were to be reopened outside of the master agreement, Sipco officials decided to create a new wholly-owned subsidiary of Sic, New Sipco, Inc., which would hold only the Moultrie and Guymon facilities. (It is unclear precisely what this final transaction was meant to accomplish.)3 The Moultrie and Guymon plants were transferred to New Sipco in late April, 1981.

While still under Sipco's control, however, Moultrie and Guymon were closed on April 17, 1981. As the ALJ noted,

Benefits were paid to separated employees at Moultrie and Guymon at special offices away from the plants.... The purpose was apparently to emphasize the fact that Sipco and the old Fresh Meat Division of Swift & Company were entirely different companies than New Sipco, Inc. [and] ... that New Sipco, Inc., ... for whom most of the employees would work if and when [re]hired ..., should be recognized as a new employer.

ALJ op. at 29.

In early 1981, Esmark and Sipco officials prepared a prospectus for the sale of Sic stock. This prospectus (which Anderson first saw in draft form in February), boldly proclaims Esmark's intentions with regard to its subsidiaries' labor contracts:

Sipco's Guymon, Oklahoma and Moultrie, Georgia plants are in the opinion of management, efficient facilities capable of successful and competitive operations. However, labor costs at these plants under the Master Agreement ... were significantly higher than those prevailing at many plants in their respective areas. A new subsidiary of the Company intends to open Guymon and Moultrie shortly after completion of this offering with the objective of achieving competitive labor costs at these facilities.... [T]he possible reduction of such costs at the Guymon and Moultrie plants will add further strength to Sipco's ability to compete effectively and profitably in the fresh meats industry.

(emphasis supplied).

The main event occurred on April 22, 1981. On the 21st, as noted, Esmark had exchanged its $100 million note for a $35 million version, together with all of Sic's stock. On the 22nd, Esmark sold 65% of Sic's outstanding stock to the public, retaining only a 35% minority interest. On April 30, a Sipco official wrote two letters to Anderson. The first, on Sipco stationery, informed Anderson of the stock sale, and that Moultrie and Guymon were now owned by New Sipco. The second letter, on New Sipco stationery (the distinctness of the two corporate entities apparently being viewed as critically important to the success of the "close and then reopen" strategy), informed Anderson that New Sipco, as a successor employer, would unilaterally set wage rates for Moultrie and Guymon when they reopened. The letter also stated that New Sipco would be pleased to receive applications from Sipco's former employees.

Interviews were conducted beginning about May 3. In order to preserve its image as a "new" employer New Sipco required each applicant to complete an application form, take a physical, etc. Despite this application process "almost all of the newly hired employees ... had [previously] worked for Sipco." ALJ op. at 31.4 The master agreement was not applied at either Moultrie or Guymon.

The ALJ concluded that the conduct described above had violated neither section 8(a)(3) (which generally prohibits discrimination in employment based on an employee's union activities or affiliation), nor 8(a)(5) (prohibiting an employer from refusing to bargain with the union representing its employees or repudiating a collective bargaining agreement). The ALJ noted that Esmark had entered all of these transactions with an economic motivation; further, the judge found that Esmark was not a single or joint employer with its subsidiaries, and therefore could not be held liable for their actions. The judge also found that Sipco and New Sipco were "successor employers," and consequently were not obligated to apply the master agreement at Moultrie and Guymon. Therefore no 8(a)(5) violation had occurred when Sipco and New Sipco refused to adhere to the terms of the master agreement. Further, no plan to evade the master agreement by closing and reopening the plants in violation of section 8(a)(3) had been carried out. Esmark (the party disposing of the plants) and Sipco (the party reopening them) were independent entities, acting for their own lawful reasons; they had not jointly participated in an unlawful plan to evade the master agreement.

The Board reversed the ALJ's 8(a)(3) and 8(a)(5) rulings. Under the NLRB's stock sale cases, a collective bargaining agreement generally continues in effect despite a stock sale. In the Board's view, a stock sale was all that had occurred between the plant closings on April 17 and the reopenings in May. Therefore, Sipco remained bound by the master agreement (which it had adopted before the stock sale), and its repudiation of the contract violated section 8(a)(5). As to the 8(a)(3) charge, the Board stressed that, from the first discussions regarding the restructuring, Esmark and its subsidiaries had intended to close and then reopen the Moultrie and Guymon plants as a means of evading their obligations under the master agreement. Therefore, the closing of the plants on April 17, and the consequent discharge of all of the plants' employees, was "inherently destructive" of important rights of the work force--specifically, the right to the benefits of a collective bargaining agreement during its term. Thus, the closing and employee discharges had violated section 8(a)(3). The Board also extended liability for the unfair labor practices to Esmark, finding that it had been an "active participant" in the unfair labor practices committed by its subsidiaries. Esmark petitions for review. The Board has filed a cross-application for enforcement of its order.

Esmark first argues that the Union's unfair labor practice charge was untimely. Section 10(b) of the NLRA provides that "no complaint shall issue based upon any unfair labor practice occurring more than six months prior to the filing of the charge with the Board." 29 U.S.C. Sec. 160(b). Section 10(b)'s limitations period commences when the aggrieved individual has actual notice that an unfair labor practice has been committed.5 In addition, the NLRB has held that, at least in some contexts, the 10(b) period may begin to run when the charging party receives unequivocal notice of an adverse decision, even though the decision will be implemented in the future.6

The 10(b) period begins when the victim of an unfair labor practice receives unequivocal notice of a final adverse decision. Rumors or suspicions will not do;7 nor is it relevant when an unlawful decision was actually made, if the decision was not communicated to the affected party until later.8 Moreover, the decision must be final, and not subject to further change; knowledge that another party might commit an unfair labor practice when the time is right will not start the 10(b) period.9 While the victims of an unfair labor practice should be encouraged to file a charge with the NLRB as soon as possible, individuals should not be forced to file anticipatory or premature charges, challenging tentative or merely hypothetical decisions, in order to protect their statutory rights. The 10(b) period does not commence until an aggrieved party has knowledge of the facts necessary to support a present, ripe, unfair labor practice charge.

The Union filed its unfair labor practice charge on May 29, 1981. Therefore, the charge is timely so long as the Union did not receive notice of the plan to evade the collective bargaining agreement before November 29, 1980. Esmark argues that the decision to close the Moultrie and Guymon plants, and reopen them in the guise of a "successor" employer, was made on June 26, 1980, and therefore the Union's charge is untimely. It is undisputed that the Union received formal notice of the plant closures on June 30, 1980. But the Union does not allege that the decision to close the plants was an unfair labor practice when considered in isolation. What the Union challenges is Esmark's plan to close and then reopen the Moultrie and Guymon facilities in a manner calculated to free the Esmark group of its obligations under the collective agreement. The 10(b) period did not begin until the Union learned of all aspects of the reorganization, and specifically that the plant closings were part of a larger plan involving evasion of the master agreement. It is irrelevant when Esmark actually made the decision to commit an unfair labor practice if that decision was not disclosed to the Union until much later.

We agree with the NLRB that Esmark's "plans for the plants were inchoate and imprecise" until well into the 10(b) period. Slip op. at 16 n. 11. As the activation of Sic and New Sipco in the spring of 1981 demonstrates, it was not clear before the limitations period commenced exactly what form the sale transaction would take--whether the transaction would be structured as an asset or stock sale, and exactly whose stock or assets were to be sold (Swift and Company, Sipco, New Sipco or Sic). Moreover, although Esmark argues that it told the Union in the fall of 1980 that the new owners of Moultrie and Guymon would "undoubtedly" repudiate the master agreement, it was also Esmark's position in its discussions with the Union that the new owner would be an independent concern, and would make its own decisions concerning the business' operations: whether or not to reopen Moultrie and Guymon, whether or not to adhere to the master agreement and which employees to hire.10 As our discussion below should indicate, the form of an acquisition transaction, the nature of the new employer's operations and the composition of its workforce are pivotal considerations in determining the extent of a new employer's obligations to a predecessor's workers. Yet none of these highly relevant facts were known by Esmark, much less the Union, at the time Esmark insists the Union should have filed an unfair labor practice charge. Any charge the Union could have filed would also have been premature. Before the 10(b) period commenced the Union knew only that the new owner would probably repudiate the master agreement. But knowledge that another party is likely to commit an unfair labor practice is not sufficient to commence the 10(b) period.

Until Anderson saw the draft prospectus for the stock sale in February 1981, and observed that Sipco had adopted the labor agreement by its conduct, the Union could not have known that the plan to evade the collective bargaining agreement was illegal. And, until the prospectus distribution was completed and the stock sale imminent, it was unclear exactly what corporate maneuverings were to take place. It is asking too much to expect the Union to have had the prescience to file an unfair labor practice charge sooner than it did, given the imprecise development and partial concealment of Esmark's reorganization "plan" at the time the 10(b) period opened. We will not disturb the NLRB's conclusion that the Union's charge was timely.

Esmark argues that the Board incorrectly found that it had violated section 8(a)(3) by participating in the plan to close and then reopen the Moultrie and Guymon plants, and thereby repudiate the master agreement. Esmark contends that it was essential for the Board to demonstrate that Esmark acted with an unlawful antiunion motive in making the plant closing decision and that the record will not support a finding that Esmark acted with the requisite state of mind.

Esmark is correct that, in general, the Board must find antiunion animus in order to establish that an employer has violated section 8(a)(3). Metropolitan Edison Co. v. NLRB, 460 U.S. 693, 700, 103 S. Ct. 1467, 1472, 75 L. Ed. 2d 387 (1983). However, in a limited class of cases the Board may find an 8(a)(3) violation without establishing that the employer acted with an unlawful motive.

Some conduct [ ] is so inherently destructive of employee interests that it may be deemed proscribed [by section 8(a)(3) ] without need for proof of an underlying improper motive. That is, some conduct carries with it unavoidable consequences which the employer not only foresaw but which he must have intended and thus bears its own indicia of intent. If the conduct in question falls within this inherently destructive category, the employer has the burden of explaining away, justifying or characterizing his actions as something different than they appear on their face, and if he fails, an unfair labor practice charge is made out. And even if the employer does come forward with counter explanations for his conduct in this situation, the Board may nevertheless draw an inference of improper motive from the conduct itself and exercise its duty to strike the proper balance between the asserted business justifications and the invasion of employee rights in light of the Act and its policy.

NLRB v. Great Dane Trailers, Inc., 388 U.S. 26, 34-35, 87 S. Ct. 1792, 1797-1798, 18 L. Ed. 2d 1027 (1967) (citations omitted).

The Supreme Court has not provided a precise definition of "inherently destructive" conduct. However, it is clear that the label "inherently destructive" may be applied only to conduct which exhibits hostility to the process of collective bargaining itself; actions which merely further an employer's substantive bargaining position in a particular contract negotiation are not "inherently destructive" as long as the employer respects the employees' right to engage in concerted activity.11 Inherently destructive conduct is that conduct which has "far reaching effects which would hinder future bargaining"; i.e., that conduct which "creat[es] visible and continuing obstacles to the future exercise of employee rights."12 However, where an employer's conduct is of temporary duration, and seeks to put pressure on union members to accept a particular management proposal, but does not attempt to prevent the employees from bargaining collectively, it is not unlawful without proof of antiunion motivation.13

Two types of acts are considered "inherently destructive." First are actions which distinguish among workers based on their participation (or lack of participation) in a particular concerted action (such as a strike).14 On the other hand, conduct may be inherently destructive even though it does not divide the work force into antagonistic factions, but instead "discourages collective bargaining in the sense of making it seem a futile exercise in the eyes of employees."15

We believe that the Board could permissibly find that Esmark's conduct falls into the latter category, and therefore no showing of antiunion motive was required to support an 8(a)(3) finding. The Board could have concluded that Esmark's repudiation of the master agreement, after only a two-week hiatus in operations, communicated to the affected employees that collective bargaining was "a futile exercise." The NLRB could conclude that the calculated repudiation of a collective bargaining agreement and prompt institution of less favorable terms sends a signal to employees that despite their diligent efforts to organize and bargain collectively, their contract may be disregarded. Workers' contract rights would be reduced to a cause of action under section 301 of the Labor Management Relations Act, 29 U.S.C. section 185, whose outcome could not be assured (given the shuffling of corporate operators) and would in any event be deferred. Workers could wonder, the Board might conclude, why collective representation, with its attendant costs, is worthwhile if their employer can manipulate things so easily by selling assets and restructuring the holding company hierarchy. As the Fifth Circuit explained in a remarkably similar case,

It would be a complete contradiction to state that [repudiation of a collective bargaining agreement] did not jeopardize the Union's position as bargaining agent or diminish its ability effectively to represent [its members]. Furthermore, no conduct could more efficaciously convey to the employees the futility of engaging in concerted activity, and thereby directly and unambiguously deter the exercise of that right, the guarantee most fundamentally protected by the Act. From the [workers'] standpoint, it would be futile to engage in collective bargaining through a representative if the Company would repudiate any resulting agreement at will. Accordingly, the Company's conduct was inherently destructive of important employee rights, and no proof of antiunion motivation is required.16

Esmark also argues that the Board erred when it found that Esmark violated section 8(a)(5) by participating in the unlawful repudiation of Sipco's collective bargaining agreement. Esmark contends that Sipco's contractual obligations after the stock sale should have been determined under the "successorship" doctrine, which governs the labor obligations of arms-length purchasers of portions of a business enterprise. Contrary to Esmark's protestations, we agree with the Union that treating Sipco as a successor corporation, bound to bargain with the Union but not to adhere to the existing collective agreement, would be to hold that Sipco was "a successor to itself." We cannot countenance such a novel reading of the Supreme Court's successorship cases, and therefore reject Esmark's arguments.

In its successorship cases the Supreme Court has held that an arms-length purchaser of portions of a prior business may be required to bargain with the union representing the predecessor's workers if there is "substantial continuity" between the two employer's operations.17 However the Court has also held that the successor is not bound to a collective bargaining agreement executed by the prior employer.

A potential employer may be willing to take over a moribund business only if he can make changes in corporate structure, composition of the labor force, work location, task assignment, and nature of supervision. Saddling such an employer with the terms and conditions of employment contained in the old collective-bargaining contract may make these changes impossible and may discourage and inhibit the transfer of capital.

NLRB v. Burns Int'l Security Servs., 406 U.S. 272, 279, 287-88, 92 S. Ct. 1571, 1577, 1582-83, 32 L. Ed. 2d 61 (1972); see also Howard Johnson Co. v. Hotel Employees Int'l Union, 417 U.S. 249, 261, 94 S. Ct. 2236, 2242, 41 L. Ed. 2d 46 (1974).

The successorship doctrine is limited to situations in which the predecessor and successor are unrelated entities and the new employer does not assume the contractual obligations of the prior employer. In these circumstances no argument can be made that the successor is in reality a continuation of the employing entity. In Burns the Court specifically noted the successor did not "become liable for [any] of [the predecessor's] financial obligations." 406 U.S. at 286, 92 S.Ct. at 1581. The Court further noted that a collective bargaining agreement might remain in force "in a variety of circumstances involving a merger, stock acquisition, reorganization or assets purchase." Id. at 291, 92 S.Ct. at 1584. In Howard Johnson the Court was even more explicit regarding the limited scope of the successorship doctrine:

It is important to emphasize that this is not a case where the successor corporation is the "alter ego" of the predecessor, where it is "merely a disguised continuance of the old employer." Such cases involve a mere technical change in the structure or identity of the employing entity ... without any substantial change in ownership or management. In these circumstances, the courts have had little difficulty holding that the successor is in reality the same employer and is subject to all the legal and contractual obligations of the predecessor.

417 U.S. at 259 n. 5, 94 S. Ct. at 2242 n. 5 (emphasis added) (citation omitted).18

It is significant that, in Burns and related cases, the Supreme Court distinguished, but did not overrule, John Wiley & Sons, Inc. v. Livingston, 376 U.S. 543, 84 S. Ct. 909, 11 L. Ed. 2d 898 (1964). In John Wiley the employer's separate corporate existence was extinguished in a merger, and all of its obligations and employees were assumed by the entity which was the product of the merger. In this circumstance, the Supreme Court held that the "new" employer was bound to arbitrate grievances under the collective agreement with the union representing the pre-merger workforce. The Court noted that it was the general corporate law rule that contractual obligations survive a merger (or, one might add, a stock sale). Writing for a unanimous court, Justice Harlan explained:

The objectives of national labor policy ... require that the rightful prerogative of owners independently to rearrange their businesses and even eliminate themselves as employers be balanced by some protection to the employees from a sudden change in the employment relationship.... It would derogate from "the federal policy of settling labor disputes by arbitration" if a change in the corporate structure or ownership of a business enterprise had the automatic consequence of removing a duty to arbitrate previously established; and this is so as much in cases like the present, where the contracting employer disappears into another by merger, as in those in which one owner replaces another but the business entity remains the same.

Id. at 549, 84 S.Ct. at 914 (emphasis added).

The successorship doctrine is simply inapplicable to a stock sale transaction. In the successor context, the question is whether two different corporate entities, one succeeding to the other's business, are bound in like measure by a contract executed by only one of them; however "the stock transfer involves no break or hiatus between two legal entities, but is, rather, the continuing existence of a legal entity, albeit under new ownership."19 The successorship doctrine is therefore of no relevance to the present case, which involves a transfer of stock ownership in the employing corporation from one party to another. And the Board has long held that a corporate entity remains liable after a stock sale for labor obligations which accrued prior to the sale.20 This rule represents a rational construction of the labor laws. Although a labor contract has no greater binding effect than any other type of contract, see, e.g., Burns, 406 U.S. at 285-86, 92 S.Ct. at 1581-82, neither should a collective bargaining agreement be easier to avoid than other contractual obligations. And the usual corporate law rule is that the contracts of a corporation remain in effect despite a sale of the corporation's stock. Esmark is essentially asking that we pierce the corporate veil to find the corporation equivalent to its shareholders, so that, when the shareholders change, the "employing enterprise" has been altered, triggering Burns' successorship analysis. But the accepted rule is that the corporate veil will only be pierced to protect the interests of third parties; the separate corporate entity will not be disregarded to allow the corporation to escape its obligations.

Esmark deliberately chose to dispose of Swift's fresh meats division as an ongoing, self-sufficient enterprise, rather than to sell its physical assets piecemeal. Esmark undoubtedly believed that the form of the transaction would be advantageous, due to the tax consequences and because Sipco's contracts (including favorable loan agreements, leases and contracts with suppliers and customers) would continue in force. Having chosen to preserve Sipco as a viable enterprise, Esmark may not now elect to ignore only those agreements it finds disagreeable. See EPE, Inc. v. NLRB, 845 F.2d 483, 488 (4th Cir.1988); United Food & Commercial Workers Int'l Union, Local 152 v. NLRB, 768 F.2d 1463, 1471 (D.C.Cir.1985); Miami Foundry Corp. v. NLRB, 682 F.2d 587, 589 (6th Cir.1982).

Adoption of Esmark's position would subject workers' contractual rights to the vagaries of corporate control transactions which have little or no effect on the workers' jobs, other than perhaps to alter the name that appears on the workers' paychecks. As Justice Harlan noted in John Wiley, this would undermine the fundamental goal of the federal labor laws, which is to ensure industrial peace and stability. An employee should not have to wonder whether his employer will continue to adhere to his contractual obligations every time the employer's stock is traded on a stock exchange. See EPE, Inc., 284 N.L.R.B. No. 21, slip op. at 12 (June 11, 1987), enf'd, 845 F.2d 483 (4th Cir.1988). Nor is there any good reason why a change in corporate control should be relevant to corporate contract or labor relations matters. The NLRB's stock sale doctrine, which holds an employer to his contractual obligations despite a change in stock ownership, is fully consistent not only with elementary corporate law but with the purposes of the labor laws.

Therefore Sipco remained bound by the master agreement at Moultrie and Guymon after the stock sale.21 And, where a collective bargaining agreement is in effect, an employer violates section 8(a)(5) by renouncing the contract and unilaterally imposing less favorable terms of employment. See Sec. 8(d) of the NLRA, 29 U.S.C. Sec. 158(d). The Board's 8(a)(5) finding is fully supported by the record and applicable law.

Section 8(a)(5) makes it an unfair labor practice for an employer "to refuse to bargain collectively with the representatives of his employees." (emphasis added). Esmark argues that it could not have violated section 8(a)(5), directly, since it was not the employer of the Union members injured by the allegedly unlawful conduct here. Nor could Esmark be liable for "aiding and abetting" its subsidiaries' violations of section 8(a)(5), since the NLRA apparently does not incorporate any general principle of aider and abettor liability. Cf. 18 U.S.C. Sec. 2; United States v. Pino-Perez, 870 F.2d 1230 (7th Cir.1989) (en banc). Esmark also argues that it could not be derivatively or vicariously liable for the 8(a)(5) violations of its subsidiaries. Corporate law generally provides that shareholders have only limited liability for misdeeds of the corporation; the Board did not rely upon the exceptions to this rule of limited liability in its order and, argues Esmark, these exceptions are not applicable here in any event. This is clearly the most challenging issue in this appeal. While the panel members agree that the Board's theory of derivative liability is legally sound, they disagree as to whether or not the Board's factual findings are sufficient to support application of that theory in the present case. Judge Cudahy believes the NLRB's findings are adequate, and would therefore enforce the Board's order in its entirety. Judges Flaum and Easterbrook, however, are of the view that the Board's findings provide an insufficient basis to impose liability on Esmark. The case will therefore be remanded to the NLRB for further proceedings.22

As the Supreme Court observed when reviewing an attempt by the Board to hold a parent company responsible for the unfair labor practices of its subsidiary, "[t]he insulation of a stockholder from the debts and obligations of his corporation is the norm, not the exception." NLRB v. Deena Artware, Inc., 361 U.S. 398, 402-03, 80 S. Ct. 441, 443-44, 4 L. Ed. 2d 400 (1960). However the separation of parent and subsidiary is not absolute. As the Court noted in Deena Artware, there may be a variety of situations in which it is appropriate to hold a parent corporation liable for the sins of its subsidiary. In the field of labor relations, the NLRB has been vested with primary authority to determine those situations in which "piercing the corporate veil" to hold the parent liable for its subsidiary's misdeeds is necessary and appropriate to effectuate the purposes of the federal labor laws. Id. at 403, 80 S.Ct. at 443.

But the Board may not disregard settled principles of corporate law. And, as Esmark argues at length, the general rule is that a shareholder is not liable for the obligations of a corporation unless the shareholder exercised "complete domination" over the corporation's decisionmaking, treating the corporation as a "mere instrumentality" or "alter ego" to advance the shareholder's personal interests. See generally Steven v. Roscoe Turner Aeronautical Corp., 324 F.2d 157, 160-61 (7th Cir.1963); Secon Serv. Sys., Inc. v. St. Joseph Bank & Trust Co., 855 F.2d 406, 413-16 (7th Cir.1988); Van Dorn Co. v. Future Chem. & Oil Co., 753 F.2d 565, 569-73 (7th Cir.1985). Thus, in general, a parent corporation may not be held to account for the liabilities of a subsidiary unless the legal separateness of parent and subsidiary has been disregarded in a wide range of corporate matters.

The NLRB has recognized this basis for piercing the corporate veil in its "single employer" doctrine.

[I]n determining the relevant employer, the Board considers several nominally separate business entities to be a single employer where they comprise an integrated enterprise. The controlling criteria, set out and elaborated in Board decisions, are interrelation of operations, common management, centralized control of labor relations and common ownership.

Radio Broadcast Technicians Local 1264 v. Broadcast Serv. of Mobile, Inc., 380 U.S. 255, 256, 85 S. Ct. 876, 877, 13 L. Ed. 2d 789 (1965) (per curiam).23 As this court has recently emphasized, no one of these factors is conclusive; instead, the Board must weigh the totality of the circumstances and determine whether the parent exercised such pervasive control of the subsidiary at the policymaking level that the purposes of the labor laws are served by treating the two entities as one. NLRB v. Emsing's Supermarket, Inc., 872 F.2d 1279, 1288-89 (7th Cir.1989).

The Board's "alter ego" doctrine is similar: generally, one corporation is the alter ego of another where the factors necessary to support a "single employer" finding are met and, in addition, the Board finds that the second corporation was a "disguised continuance" of the employing enterprise, resulting in evasion of the employer's obligations under the labor laws.24

In order for the "single employer" or "alter ego" theories to apply, the General Counsel must establish that the stockholder and corporation were in effect a single entity. The focus is broad, not limited to the specific transaction alleged to constitute an unfair labor practice. The Board must determine whether the corporation was a bona fide, separate enterprise, with its own decisionmaking apparatus and financial structure. As Esmark notes, the ALJ's factual findings in this case would apparently not support a conclusion that Esmark was the alter ego of, or a single employer with, the subsidiaries which directly employed the Union's members.25 Moreover, the Board's order specifically disclaims any reliance on these theories:

[W]e find Esmark to be liable because the actual repudiation of the master agreement was ... the fulfillment of an unlawful plan in which Esmark was an active participant, to close and then reopen these plants for the purpose of evading the application of the master agreement.... Esmark's direct participation in this unlawful conduct at pertinent times forecloses Esmark from evading liability under the criteria set out in [the "single employer" and "alter ego" doctrines].

* * *

We ... reject the contention of Respondent Esmark that it is not liable for closing of these plants on the grounds that the decision to close was made in June 1980 solely by Respondent SIPCO, through John Copeland, and that Esmark, as a mere parent corporation, did not participate in nor have any control over the labor relations decisions of its (then) wholly owned subsidiary. In rejecting this contention, we note that the record convincingly establishes that Esmark was a principal actor in the closures. We first note that at the time the decision was made to close these plants in June 1980 Copeland was reporting directly to Esmark's president, Donald Kelly. Copeland's plans for reorganizing the former fresh meats operations and for closing Moultrie and Guymon were made at the direction of and in consultation with Kelly. Esmark admits that its board of directors approved the decision to close the plants.... As Esmark was a principal actor in the unlawful plant closures, its liability for these unfair labor practices does not require any finding that it had the status of alter ego/single employer, etc., with the SIPCO group.

Slip op. at 16, 17, 21 n. 18 (emphasis added).

The NLRB's refusal to rely on the single employer or alter ego doctrines is not surprising since the ALJ's findings would not support the conclusion that the SIPCO subsidiaries were so integrated with Esmark that the separate corporations should be considered one enterprise. What, then, is the legal basis for the NLRB's ruling that Esmark is liable for the unfair labor practices committed here? Although not specifically invoked in its order, the NLRB's finding that Esmark was liable as a "direct participant" may be justifiable under a well-established exception to the general rule that the corporate veil will not be pierced in the absence of large-scale disregard of the separate existence of a subsidiary corporation.26 Writing in 1929, then-Professor William O. Douglas counseled that four actions would generally insure that a parent corporation would not be held liable for its subsidiaries' wrongdoing:

(1) A separate financial unit should be set up and maintained. That unit should be sufficiently financed so as to carry the normal strains upon it.... (2) The day to day business of the two units should be kept separate.... (3) The formal barriers between the two management structures should be maintained. The ritual of separate meetings should be religiously observed.... (4) The two units should not be represented as being one unit.

Douglas and Shanks, Insulation from Liability through Subsidiary Corporations, 39 Yale L.J. 193, 196-97 (1929). So far as appears from the record here, Esmark followed the four steps recommended by Justice Douglas to avoid liability for its subsidiaries' obligations. However, Douglas also noted that "[t]here is a group of cases where liability is imposed upon the parent for torts of the subsidiary, even though the four standards of organization and operation which have been discussed above are ... met." Id. at 205. As an example, Justice Douglas noted that liability has been imposed in

instances where the parent is directly a participant in the wrong complained of. The parent has been held liable in a tort action for inducing the subsidiary by means of its stock ownership to breach a contract with the plaintiff. Stock ownership was not enough. But the use of the latent power incident to stock ownership to accomplish a specific result made the parent a participator in or doer of the act. Again, there was interference in the internal management of the subsidiary; an overriding of the discretion of the managers of the subsidiary.

Id. at 208-09 (footnote omitted) (emphasis added). Later in the same article Douglas noted that "[d]irect intervention or intermeddling by the parent in the affairs of the subsidiary and more particularly in the transaction involved, to the disregard of the normal and orderly procedure of corporate control carried out through the election of the desired directors and officers of the subsidiary and the handling by them of the direction of its affairs, seems to have been determinative in some cases to holding the parent liable." Id. at 218.

Also writing in 1929, Judge Learned Hand similarly recognized that a parent corporation could be held liable for the actions of its subsidiaries where the parent directly supervised the conduct of a specific transaction.

Control through the ownership of shares does not fuse the corporations, even when the directors are common to each. One corporation may, however, become an actor in a given transaction, or in part of a business, or in a whole business, and, when it has, will be legally responsible. To become so it must take immediate direction of the transaction through its officers, by whom alone it can act at all.... [L]iability ordinarily must depend upon the parent's direct intervention in the transaction, ignoring the subsidiary's paraphernalia of incorporation, directors and officers. The test is therefore rather in the form than in the substance of the control; in whether it is exercised immediately, or by means of a board of directors and officers, left to their own initiative and responsibility in respect of each transaction as it arises.

Kingston Dry Dock Co. v. Lake Champlain Transp. Co., 31 F.2d 265, 267 (2d Cir.1929) (citations omitted).

The observations made by Justice Douglas and Judge Hand sixty years ago are still valid today--a parent corporation may be held liable for the wrongdoing of a subsidiary where the parent directly participated in the subsidiary's unlawful actions.27 Under this "transaction-specific" theory of direct participation, parent companies have been held liable for a wide variety of misconduct by subsidiaries: patent or copyright infringement,28 false advertising,29 fraud,30 conversion,31 and the creation of a nuisance.32 See also P. Blumberg, The Law of Corporate Groups: Problems of Parent & Subsidiary Corporations under Statutory Law of General Application 273-74, 298-99, 321-22, 338-42 (1989) (detailing use of "direct participation" theory of intercorporate liability in antitrust and patent, trademark and copyright infringement cases). One district court has found a parent corporation liable for the breach of a collective bargaining agreement by its subsidiary where the parent specifically directed the subsidiary to disregard its obligations under the NLRA. International Union, United Auto. Workers v. Cardwell Mfg. Co., 416 F. Supp. 1267, 1283-84, 1287-89 (D.Kan.1976).

In these cases the shareholder or parent corporation was not held "directly" liable for its own independently wrongful acts. The parent corporation did not, side-by-side or concurrently with its subsidiary, directly cause an injury to the third-party/victim. Cf. Corey v. Havener, 182 Mass. 250, 65 N.E. 69 (1902). There was no allegation that the parent did anything directly to the third party. Instead the parent only acted against the third party's interests through the agency of the subsidiary. The owner's liability was based on its control of its subsidiaries' actions from "behind the scenes." Thus the parent was not held "directly liable"; it was liable derivatively for transactions of its subsidiary in which the parent interposed a guiding hand.

Nor was liability imposed on the parent because it "aided and abetted" the subsidiary's tortious conduct or breach of contract. Imposing liability on Esmark in this case will not result in broad "aider and abettor" liability for unfair labor practices, extending potentially to lawyers, management consultants or investment bankers who "actively participate" in a corporation's decision to commit an unfair labor practice. Aiders and abettors may suggest a course of conduct and attempt to persuade a corporation to adopt the recommended policy. But an outsider may only advise; it may not compel a corporation to act. A third party confronts the corporation as an independent entity, with its own decisionmaking apparatus geared towards advancing the subsidiary's independent interests. A parent corporation, by virtue of its ownership interest, may direct a subsidiary's actions, and the subsidiary will have no choice but to obey. A parent corporation may disregard the subsidiary's independence at its whim. Cf. Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752, 104 S. Ct. 2731, 81 L. Ed. 2d 628 (1984) (parent and wholly-owned subsidiary may not enter "conspiracy" within meaning of section 1 of Sherman Antitrust Act, since subsidiary always subject to potential, if not actual, control by parent).

Therefore holding a parent corporation liable for "directly participating" in its subsidiary's decisionmaking is fundamentally different from holding an unrelated third party liable for acting in concert with, or aiding and abetting, the subsidiary's misconduct. Where the parent specifically directs the actions of its subsidiary, using its ownership interest to command rather than merely cajole, the possibility of its violating the federal labor laws is present. The NLRB could permissibly find that a parent corporation should not be permitted to act through its subsidiaries to the detriment of the subsidiaries' workforce and yet escape liability for acts which it has mandated. The situation is markedly different from liability of a mere adviser whose counsel happens to be accepted by the corporation in a particular situation.

It is solely where a parent disregards the separate legal personality of its subsidiary (and the subsidiary's own decisionmaking "paraphernalia"), and exercises direct control over a specific transaction, that derivative liability for the subsidiary's unfair labor practices will be imposed under the theory adopted by the Board in the present case. Thus, although 8(a)(5) liability may only be imposed for an employer's conduct towards "his employees," Esmark may still be liable for the unfair labor practices committed here. The "employer of these employees" (Sipco) is the entity which actually acted in derogation of its employees' rights; however Esmark was also present, guiding and directing Sipco at the time of the charged conduct.

The Board has previously recognized that "direct participation" by a shareholder in the unfair labor practices of a corporation may result in individual liability for the violations. In the leading case of Riley Aeronautics Corp., 178 N.L.R.B. 495 (1969), the Board stated:

[T]he corporate veil will be pierced whenever it is employed to perpetrate fraud, evade existing obligations, or circumvent a statute. Thus, in the field of labor relations, the courts and Board have looked beyond organizational form where an individual or corporate employer ... was in active concert or participation in a scheme or plan of evasion....

Id. at 501 (citations omitted) (emphasis added).

The Board has stressed that a shareholder will not be found liable for "direct participation" where the shareholder's actions were pursuant to his or her proper decisionmaking function as an officer or director of the corporation.33 However, where the shareholder participates in corporate decisionmaking outside of the normal channels of corporate control, personal liability has been imposed in exceptional cases. D & I Trucking, Inc., 237 N.L.R.B. 55 (1978), is instructive in this regard. In that proceeding, Dorothy and Irvin Cornett were sole owners of the D & I and Tipprell Trucking corporations. The General Counsel alleged that the Cornetts "deactivated" D & I after the employees of that corporation voted for union representation and thereafter continued their trucking business through Tipprell. The General Counsel argued that the Cornetts had violated section 8(a)(5) and should be held personally liable for the closing and reopening transactions executed by their wholly-owned corporations. It should be emphasized that the Cornetts could have made the same argument Esmark makes here--that they were not "employers" of the affected employees, since their signatures did not appear (in their individual capacities) on the collective bargaining agreement. Nevertheless the NLRB held the Cornetts personally liable, relying on the Riley Aeronautics decision. Since "[t]he Cornetts ... personally and actively engaged in and participated in [a] plan and scheme of evasion" by transferring their trucking operations to a "new" corporation allegedly not constrained by the predecessor's labor contract, they were held personally liable for the unfair labor practices nominally committed by their corporations. Id. at 64.34

From the foregoing it appears that the NLRB has, in extraordinary cases, employed a theory of "direct participation" to hold a stockholder liable for the unfair labor practices nominally carried out by a corporation. This theory, initially derived in Riley from corporate law decisions by state courts (and generally consistent with current state case law), holds a shareholder to be at one with the statutory "employer" where the shareholder actively participated in the decision to commit the particular wrong, and the shareholder's power of control was exercised "directly," outside the normal channels of corporate decisionmaking. The question then becomes whether the Board's factual findings in this case are sufficient to establish liability under this limited, transaction-by-transaction theory of piercing the corporate veil.

Under the Chenery doctrine, an agency's determination may only be upheld by a reviewing court based on the rationales and factual findings actually adopted by the agency, as reflected in the agency's decision. The legal basis for the agency's action must be consistent with governing statutory and constitutional principles; "[i]f those [explicitly stated] grounds [of decision] are inadequate or improper, the court is powerless to affirm the administrative action by substituting what it considers to be a more adequate or proper basis."35 Beyond merely providing an adequate legal basis for its decisions, "[t]he agency must [also] make findings that support its decision, and those findings must be supported by substantial evidence."36

The members of the panel deciding this case agree that the "direct participation" theory, when properly limited to control exercised in disregard of corporate formalities, is a viable theory of intercorporate liability which may be employed by the Board to hold a parent corporation (or other shareholder) liable for the unfair labor practices of a subsidiary corporation. The panelists do not agree, however, whether the Board's order contains factual findings sufficient to support liability under this theory in the present case.

Judge Cudahy believes that, although the question is close, Chenery does not require a remand here. Although Sipco was a nominally separate corporate entity, the Board's order makes clear that Esmark exercised direct control over all aspects of the corporate reorganization, including specifically the decision to close and reopen the Moultrie and Guymon plants to evade the terms of the master agreement. Throughout this transaction, Sipco's officer Copeland reported directly to Esmark's president Donald Kelly, and the Esmark board of directors was kept fully apprised of, and approved, relevant aspects of the transaction. Moreover, although Esmark argues that the decision to reopen Moultrie and Guymon at wage rates set by management was made after the stock sale by Sipco's board of directors, acting independently, the prospectus for the stock sale (prepared in large part by officers of Esmark) lays out in detail what will happen and why it will happen months in advance of the events themselves. Esmark's securities counsel apparently required a fullness of disclosure which is now embarrassing to its labor counsel. Further, throughout the discussions concerning the reorganization between Esmark and Sipco officials, and between Sipco and the Union, it was recognized that Moultrie and Guymon would be closed before the stock sale, and reopened afterwards with new wage rates set by the new, "successor" employer. (The incorporation of New Sipco, and the transfer to it of the Moultrie and Guymon facilities, was also intended to insure the legality of repudiating the master agreement after the stock sale.) In these circumstances, it is difficult to argue that Esmark was merely a minority shareholder, or an "innocent bystander," at the time of the decision to commit the unfair labor practices. The Board also specifically noted that Esmark paid the substantial closing costs involved in the shut down of Moultrie and Guymon, an action which Esmark has nowhere explained, but which would suggest that the entire transaction was intended to benefit Esmark directly, and that Esmark undertook a direct obligation to the displaced workers. Moreover, Judge Cudahy would also note that a corporate reorganization is, by its nature, motivated by concerns "external" to the subsidiary. A subsidiary does not "independently" decide to consolidate, merge, go public or private. These are decisions made from "outside" the affected corporation. Only the owner of a corporation decides which corporate transformations are in the owner's (not necessarily the corporation's) best interests. A reorganization is motivated by concerns foreign to the subsidiary as an independent enterprise. The nature of the transaction involved in the current case provides additional support for the Board's conclusion that Esmark exercised an uncommon, and inappropriate, degree of control over its subsidiaries' actions. Under the "substantial evidence" standard of review, Judge Cudahy therefore believes that the facts of record adequately support the Board's conclusion that Esmark should be liable, together with its subsidiaries, for the unlawful acts which occurred here.

Judges Flaum and Easterbrook hold a contrary view, which is hence the view of the panel majority. In order to prevent the "direct participation" theory of liability from extinguishing the general rule of parent corporation immunity from the obligations of a subsidiary, that theory must be carefully limited to situations in which the parent corporation's control over particular transactions is exercised in disregard of the separate corporate identity of the subsidiary. Parents and dominant shareholders are almost always "active participants" in the affairs of an owned corporation. And, in the usual case, the exercise of such control over a subsidiary's actions is entirely permissible, and does not result in the owner's personal liability. See Secon Serv. Sys., Inc. v. St. Joseph Bank & Trust Co., 855 F.2d 406, 415 (7th Cir.1988). In order to pierce the corporate veil and displace the general rule of a shareholder's limited liability, the Board must make specific findings that a stockholder's power of control was exercised through improper means. In the present case, Judges Flaum and Easterbrook find no mention in the NLRB's order of the manner in which Esmark exercised control: whether control was exercised through (perhaps specific) directions to Sipco's (nominally independent) board of directors (in which case Esmark would not be liable for the resulting harm to Sipco's employees); or whether, instead, Esmark made Sipco's decisions for it, without any concern for Sipco's separate legal identity. In other words, the NLRB's order failed to draw the critical distinction between permissible "active" participation by a parent in the affairs of a subsidiary, and impermissible "direct" control (i.e., that intervention which is accompanied by disregard of the subsidiary's separate corporate form). Judges Flaum and Easterbrook believe that it is the Board's responsibility to make explicit and comprehensive findings that Esmark ignored its subsidiaries' separate decisionmaking "paraphernalia," if liability is to be imposed on Esmark under the "direct participation" theory. Since the Board's findings on this central issue are inadequate, the case must be remanded to the NLRB for further factual development.

For the foregoing reasons Esmark's petition for review is granted, the Board's cross-application for enforcement is denied, and the case is remanded to the NLRB for further proceedings.