Be-Lo Stores v. NLRB

Filed: November 4, 1997 UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT Nos. 96-1575(L) (11-CA-14586(L)) Be-Lo Stores, Petitioner, versus National Labor Relations Board, Respondent. O R D E R The Court amends its opinion filed September 16, 1997, as follows: On page 3, second full paragraph, line 2 -- the word "which" is changed to read "that." On page 11, first full paragraph, line 4 -- the hyphen in "clearly relevant" is deleted. For the Court - By Direction /s/ Patricia S. Connor Clerk Volume 1 of 2 PUBLISHED UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT BE-LO STORES, Petitioner, v. NATIONAL LABOR RELATIONS BOARD, No. 96-1575 Respondent, UNITED FOOD AND COMMERCIAL WORKERS INTERNATIONAL UNION, Local 400, AFL-CIO, CLC, Intervenor. NATIONAL LABOR RELATIONS BOARD, Petitioner, UNITED FOOD AND COMMERCIAL WORKERS INTERNATIONAL UNION, Local 400, AFL-CIO, CLC, No. 96-1657 Intervenor, v. BE-LO STORES, Respondent. On Petition for Review and Cross-Application for Enforcement of an Order of the National Labor Relations Board. (11-CA-14586, 11-CA-14775-1, 11-CA-14775-1, 11-CA-14775-2, 11-CA-14775-3, 11-CA-14775-4, 11-CA-14775-5, 11-CA-14775-6, 11-CA-14775-7, 11-CA-14775-8, 11-CA-14775-9, 11-CA-14775-10, 11-CA-14712, 11-CA-14793-1, 11-CA-14793-2, 11-CA-14793-3, 11-CA-14793-4, 11-CA-14793-5, 11-CA-14811, 11-RC-5823) Argued: January 27, 1997 Decided: September 16, 1997 Before ERVIN and LUTTIG, Circuit Judges, and HILTON, United States District Judge for the Eastern District of Virginia, sitting by designation. Volume 1 Opinion by Judge Luttig Volume 2 Dissent by Judge Ervin _________________________________________________________________ Reversed in part, affirmed in part, and remanded by published opin- ion. Judge Luttig wrote the majority opinion, in which Judge Hilton joined. Judge Ervin wrote a dissenting opinion. _________________________________________________________________ COUNSEL ARGUED: Stanley Graves Barr, Jr., Charles Vincent McPhillips, KAUFMAN & CANOLES, P.C., Norfolk, Virginia, for Be-Lo Stores. John Emad Arbab, NATIONAL LABOR RELATIONS BOARD, Washington, D.C., for NLRB. Carey Robert Butsavage, BUTSAV- AGE & ASSOCIATES, P.C., Washington, D.C., for Intervenor. ON BRIEF: Arlene F. Klinedinst, Ashley L. Taylor, Jr., KAUFMAN & CANOLES, P.C., Norfolk, Virginia, for Be-Lo Stores. Frederick L. Feinstein, General Counsel, Linda Sher, Associate General Counsel, Aileen A. Armstrong, Deputy Associate General Counsel, Peter Winkler, Supervisory Attorney, William A. Baudler, NATIONAL LABOR RELATIONS BOARD, Washington, D.C., for NLRB. George Wiszynski, BUTSAVAGE & ASSOCIATES, P.C., Washing- ton, D.C., for Intervenor. _________________________________________________________________ 2 OPINION LUTTIG, Circuit Judge: "[O]ur nation's labor policies have never included a preference for imposing a collective bargaining representative upon those who have not affirmatively chosen that representative by election." N.L.R.B. v. Appletree Chevrolet, Inc., 671 F.2d 838, 840 (4th Cir. 1982) (Appletree Chevrolet II). Since "an election, not a bargaining order, remains the traditional, as well as the preferred, method for determin- ing the bargaining agent for employees," N.L.R.B. v. Appletree Chev- rolet, Inc., 608 F.2d 988, 996 (4th Cir. 1979) (Appletree Chevrolet I), the "extraordinary and drastic remedy" of forced bargaining pursuant to N.L.R.B. v. Gissel, 395 U.S. 575 (1969), is reserved for only the most "unusual cases," N.L.R.B. v. J. Coty Messenger Serv., Inc., 763 F.2d 92, 99 (2nd Cir. 1985). Because such an order is disfavored, we have admonished the National Labor Relations Board that it must undertake a comprehensive analysis for each of Gissel's necessary predicate findings, and support its order with detailed record evi- dence. It is, in other words, "manifestly insufficient" for the Board to satisfy the Gissel predicates by "simply engaging in perfunctory or boiler-plate language, or using a litany, reciting conclusions by rote without factual explication." Appletree Chevrolet I 608 F.2d at 997 (internal quotation marks omitted). Notwithstanding our repeated warnings, and similar warnings from our sister circuits, "even a cur- sory examination of the decisions applying Gissel . . . reveals that the Board has declined repeatedly to assist the courts .. . by revealing [its] reasons for issuing Gissel bargaining orders." Red Oaks Nursing Homes, Inc. v. N.L.R.B., 633 F.2d 503, 508 (7th Cir. 1980) (citing, inter alia, Appletree Chevrolet I, 608 F.2d at 996). Here, notwithstanding our repeated admonitions, the Board imposed a far-reaching mandatory bargaining order that would require the company to bargain with the United Food and Commercial Workers Union, although significant questions exist as to whether the Union ever achieved majority status among the company's employ- ees; some six years have elapsed since the company's election vic- tory; more than two-thirds of the company's work force employed at the time of the company's unfair labor practices are no longer even employed by the company; there were relatively few violations and 3 they occurred in less than one-half of the company's thirty stores; and no evidence whatsoever exists that a fair election could not be held today. Moreover, the Board imposed its mandatory bargaining order on the strength of little more than "perfunctory or boiler-plate lan- guage" lacking substantive "factual explication." Appletree Chevrolet I, 608 F.2d at 997 (internal quotation marks omitted). Because a mandatory bargaining order should not have issued under the facts of this case, and such an order would not have been sustainable on the basis of the kind of ipse dixit relied upon by the Board here in any event, we reverse the Board's entry of the manda- tory bargaining order. We also reverse, as unsupported by substantial evidence, the Board's conclusions that Be-Lo violated the National Labor Relations Act, 29 U.S.C. § 151, et seq. ("the Act") by (1) dis- tributing a "pink slip" flyer, (2) denying Union picketers access to its property, and (3) dismissing or failing to recall five pro-union employees. We affirm, however, for reasons discussed, a number of the Board's other findings with regard to individual personnel actions. I. In May 1990, intervenor United Food and Commercial Workers Union, Local 400 (the "Union") began organizing at several of Be- Lo's retail grocery stores. In February 1991, following a lengthy recruitment campaign, the Union demanded recognition, advising Be- Lo that a majority of the company's employees had signed union authorization cards designating the Union as their official bargaining representative. Be-Lo challenged the Union's claim of majority and declined to recognize the Union, and an election to resolve the matter was scheduled for March 21, 1991. During the weeks preceding the election, the company openly cam- paigned against unionization, holding meetings, circulating memo- randa, showing videotapes, and sending out flyers, including a flyer in the form of a mock "pink slip" purportedly given to employees of unionized stores that had been forced to close following unionization. Be-Lo also took "employment actions" against a number of employ- ees who favored unionization of the company. 4 The Union lost the March 21 election 377 to 220. The remaining 159 of the 756 total eligible voters either did not vote or had their nondeterminative ballots contested. Following the election, the Union picketed several Be-Lo locations. Be-Lo responded by seeking five separate injunctions to exclude the picketers from its premises under the company's no-solicitation pol- icy. In all five cases, the courts of the Commonwealth of Virginia granted Be-Lo's requested injunctions. Following the filing of charges by the Union and the filing of com- plaints by the NLRB Regional Director against Be-Lo alleging viola- tions of various provisions of the National Labor Relations Act, a hearing was held before an administrative law judge. The ALJ found that, overall, Be-Lo's campaign was conducted in accordance with the Act. See J.A. Vol. I, at 360. The ALJ specifically rejected the NLRB General Counsel's claims that Be-Lo's pink slip flyer and other anti- union campaign messages violated section 8(a)(1), and that Be-Lo was discriminatorily enforcing its no-solicitation policy. The ALJ held, however, that a number of statements made to individual employees by certain of Be-Lo's store managers violated section 8(a)(1), and that sixteen individual employment actions, of which thirteen were various forms of discharges, taken by Be-Lo during the election campaign violated section 8(a)(3). On the basis of these vio- lations, the ALJ issued a Gissel mandatory bargaining order. On appeal from the ALJ's decision, the Board reversed two of Be- Lo's employee-specific section 8(a)(3) discriminatory discharge vio- lations and affirmed eleven others, and affirmed the ALJ's findings of employee-specific section 8(a)(1) violations. The Board also held, however, contrary to the ALJ, that Be-Lo threatened its employees with job loss in violation of section 8(a)(1) by distributing the pink slip flyer and by denying Union picketers access to the sidewalks and parking lots at the specified Be-Lo Stores following the election. See J.A. Vol. I, at 341, 348-50. Based on these violations, the Board affirmed the ALJ's mandatory bargaining order. II. "`[A]n election, not a bargaining order, remains the traditional, as well as preferred, method' for determining the employee's bargaining 5 agent." N.L.R.B. v. So-Lo Foods, Inc., 985 F.2d 123, 126 (4th Cir. 1993) (quoting Appletree Chevrolet I, 608 F.2d at 996). While we accord the Board respect as to its choice of remedies because of its presumed expertise, see So-Lo Foods, 985 F.2d at 126 (citing Gissel, 395 U.S. at 612 n.32), we "exercise less deference" and require scru- pulous specificity from the Board when it issues mandatory bargain- ing orders on the authority of NLRB v. Gissel Packing. See N.L.R.B. v. Rexair, Inc., 646 F.2d 249, 250 (6th Cir. 1981) (courts closely scru- tinize the Board "when it has imposed the very strong remedy of issu- ing a bargaining order without holding a new election"). A Gissel order "is not a snake-oil cure for whatever ails the workplace; it is an extreme remedy that must be applied with commensurate care." Skyline Distributors v. N.L.R.B., 99 F.3d 403, 410 (D.C. Cir. 1996) (internal quotation marks omitted). Gissel orders may enter in essentially two types of cases -- so- called Category I cases, where "exceptional," "outrageous," and "per- vasive" unfair labor practices have occurred and the coercive effects of such practices "cannot be eliminated by the application of tradi- tional remedies," Gissel, 395 U.S. at 613-14, and, in less extraordi- nary cases (Category II), where the Board has found "(1) that the Union once had a majority status, (2) that such status had been dissi- pated by pervasive misconduct on the part of the employer, (3) `that the possibility of erasing the effects of (these) past (pervasive) prac- tices and ensuring a fair election . . . is slight,' and (4) `that employee sentiment . . . would, on balance, be better protected by a bargaining order.'" Appletree Chevrolet I, 608 F.2d at 996 (quoting Gissel, 395 U.S. at 614). Here, the Board did not specify whether it was imposing a Gissel Category I or Gissel Category II order, stating that it was unnecessary to decide the issue because the case "clearly satisfies the greater bur- den imposed under . . . Category II." J.A. Vol. I, at 350 n.32. We therefore deem the order to have been a Gissel Category II order, as we do when the Board declines to specify the nature of its action, yet conducts only a Category II analysis. See Appletree Chevrolet I, 608 F.2d at 996 n.8. Scrutinizing the Board's order as such, it is apparent that the order cannot stand. Not only was the Board's (and ALJ's) fundamental determination that the Union enjoyed majority status flawed, which alone would serve to invalidate the sweeping remedy 6 ordered, see, e.g., N.L.R.B. v. Cell Agricultural Manu. Co., 41 F.3d 389, 397 (8th Cir. 1994), but even if that determination were unassail- able, the Board's conclusion that a fair rerun election would be impossible was in error, given the large turnover in Be-Lo's work force and the six-years that have elapsed since the alleged violations occurred. A. 1. As a threshold matter, the Board erred when it concluded that the Union produced a sufficient number of authorization cards (403) to establish majority status among Be-Lo's work force as of March 20, 1991. In reaching this conclusion, the Board presumed as a matter of law (for reasons discussed infra) that the employees whose names appeared on a voter eligibility list current as of January 26, 1991 (hereafter "GC-63") were the names of eligible voters in the bargain- ing unit as of March 20, 1991, and it calculated the number of autho- rization cards required for the Union to establish majority status (379) based upon the number of employees on this list. At the same time that the Board presumed there had been no changes in the company's work force between January 26 and March 20, the Board allowed the NLRB General Counsel to introduce authorization cards signed dur- ing this two-month period by employees favoring Union representa- tion in order to establish that the Union enjoyed majority status as of election day, but disallowed introduction of evidence by Be-Lo that forty-four of the employees who were listed on the GC-63 and who signed cards prior to January 26, left Be-Lo's employ between Janu- ary 26 and March 20. As well, the Board disallowed Be-Lo from introducing evidence that there were seventy-five new employees who had been hired between January 26 and March 20, virtually none of whom had signed cards. See J.A. Vol. I, at 344; J.A. Vol. III, at 1114-43, 1160-80; J.A. Vol. IV, at 1606-13. If Be-Lo's proffered evi- dence had been considered and accepted by the Board and the ALJ, it would have shown that the Union did not actually enjoy majority status on March 20, on January 26, or on any date in between; in other words, it was only because the Board provided the Union, but not Be- Lo, the benefit of the period between January 26 and March 20 for 7 purposes of gauging Union support that the Union was able to estab- lish majority status. The Board accepted the January 26 voter eligibility list as repre- senting the Be-Lo employees eligible to vote in the election as of March 20, rather than allowing evidence to be presented on the actual number and identity of persons employed by Be-Lo as of that date, because it found that Be-Lo induced reliance upon the GC-63 "as the final voter eligibility list for the purpose of proving majority status." J.A. Vol. I, at 351. Having found that Be-Lo induced reliance upon the GC-63, the Board excluded all evidence challenging that list, under the authority of its decision in Bannon Mills, 146 NLRB 611, 633-35 (1964). In this, the Board clearly erred. The Board, while recognizing that the holding of Bannon Mills does not apply to Be-Lo's actions,1 nonetheless affirmed exclusion of Be-Lo's evidence on the authority of Bannon Mills, reasoning that "the principle [of that case], the protection of the integrity of the hear- ing process" applied equally in this context as in the context where a party has withheld evidence in response to a subpoena. J.A. Vol. I, at 352 n.35. Even were we to recognize such an extension of Bannon Mills, which we do not, such an extension would not serve to validate the Board's exclusion of Be-Lo's evidence as to the changes in its work force between January 26 and March 20, because in no sense can it be said that Be-Lo induced reliance upon the GC-63. The Board and ALJ essentially estopped Be-Lo from arguing that changes had occurred in its work force, on the strength of a single, cryptic statement by Be-Lo's counsel made in the course of an exchange on a motion to partially quash a NLRB subpoena directed to Be-Lo for payroll and related documents. During that exchange, counsel for the NLRB stated that he was seeking _________________________________________________________________ 1 In Bannon Mills, the Board held that a party may be sanctioned with exclusion of evidence when it withholds evidence in response to a sub- poena request. Because Be-Lo fully complied with its subpoena request by producing all of the business and payroll records properly requested by the General Counsel, the Board conceded, as it must, that Bannon Mills (which is not the law of this circuit in any event) is, strictly speak- ing, inapplicable here. See J.A. at 352 n.35. 8 a list[ ] of all bargaining unit employees for each of those payroll periods [in February and March]. In other words, a listing like they provided to the Board for the excelsior list for the election [GC-63]. . . . [because] the only thing avail- able to the Regional Director to ascertain the majority that we are looking for in the Gissel complaint was the excelsior list. J.A. Vol. II, at 438. In response, Be-Lo's counsel stated that, the election was March 21, and there were certainly some deletions from the excelsior list [GC-63] for those employ- ees who were not on the payroll as of the date of the elec- tion. Those deletions were made and the board has that list as of March 21. They have it. Id. at 439. On the basis of this latter statement, the Board held that Be-Lo induced reliance upon the GC-63. We can appreciate the source of the ALJ's and the Board's confu- sion regarding the response by Be-Lo's counsel. On its face, the state- ment could be understood as a representation that the General Counsel had been provided a revised Excelsior list, which included the deletions and additions of names for those employees leaving and entering Be-Lo's employ during the period January 26 to March 20. However, when this statement is considered in the context of the whole record, it is fairly clear that Be-Lo's counsel was saying in this response only that the General Counsel and Union had in their posses- sion a separate list of employees who either had been terminated by Be-Lo since creation of GC-63 or whose names, for one reason or another, had not appeared on that list, not as saying that the General Counsel and the Union had in their possession a revised GC-63 incor- porating these personnel changes. As Be-Lo explains in its brief, and as the record confirms, the General Counsel and the Union had in fact been provided with what even the Union itself describes as a "list of employees who: a) are active employees not on the Excelsior list (Exhibit "A"), or b) employees who were terminated and should not be on the Excelsior list (Exhibit "B")." J.A. Vol. 7, at 2435 (Letter of March 4, 1991 from Carey R. Butsavage to Sherrie Black of National 9 Labor Relations Board, attaching Exhibits A & B).2 That Be-Lo's counsel was referring only to the list of changes and not to a new, integrated Excelsior list, is further supported, we believe, by the fact that the Union finds it necessary to misquote Be-Lo's counsel in this important colloquy. The Union quotes Be-Lo not as saying (as Be- Lo's counsel did) that "there were certainly some deletions from the excelsior list for those employees who were not on the payroll as of the date of the election . . . and the Board has that list as of March 21," but rather, as saying that, "[i]f you want a list of employees for purposes of determining the Union's majority status, you have the Excelsior list [GC-63]. You have it." See Intervenor's Br. at 12 (citing J.A. Vol. II, at 439). Be-Lo's counsel never made any such statement. Ultimately, though, we need not determine whether (as the Board contends) the statement by Be-Lo's counsel was intended as confir- mation that the GC-63 was an accurate list of the company's employ- ees on March 20 or whether (as Be-Lo contends) the statement referenced only the separate list of employees who had departed Be- Lo or whose names had not appeared on the original Excelsior list. For, when considered against the backdrop of the three thousand pages of record in this case -- a significant portion of which is devoted to the issue of determining the Union's majority status -- this single statement is manifestly insufficient to support a conclusion that Be-Lo induced reliance on the GC-63 such as to bar the company from introducing evidence on the inaccuracy of that list as of March 20. Not only did disputes over the GC-63 prompt both parties to con- test over a hundred ballots before the final election on March 21; Be- Lo steadfastly maintained throughout the entire post-election proceed- ing giving rise to this appeal that the original Excelsior list (GC-63) required additions and deletions in order to make it an accurate list as of the March 20 majority date. Contrary to the holding of the Board that Be-Lo only contested the Union's reliance on GC-63 "some 5 months into the hearing and after the [Union] had rested," see J.A. Vol. I, at 251, Be-Lo vehemently and repeatedly challenged the accu- racy of GC-63 before the close of the General Counsel's case. See J.A. Vol. III, at 1114-117. _________________________________________________________________ 2 An Excelsior list is a list of employees provided by the employer for use by a union during an election. The Excelsior list in this case became, with minor alterations not relevant to this dispute, the GC-63. 10 Even if Be-Lo could be said to have induced reasonable reliance on GC-63 through its counsel's single statement, reliance upon this statement by the General Counsel was not so prejudicial as to warrant exclusion of Be-Lo's clearly relevant evidence. The General Counsel had subpoenaed additional documents from Be-Lo because it was dis- satisfied with GC-63 and wished to make its own determination of which employees were eligible on March 20. See J.A. at Vol. II, 438- 40. Thus, the Board had in its possession, through subpoenaed pro- duction, all of the information necessary to know that GC-63 was not an accurate list of Be-Lo's employees as of March 20. Apparently, the NLRB itself even intended, at least initially, to challenge the accuracy of the list. Moreover, as explained, the General Counsel and the Union had in their possession a list of Be-Lo's employees who either had departed the company since January 26 or whose names had been omitted from the original Excelsior list. Under these circumstances, we are at a loss to understand how the General Counsel could have been prejudiced. 2. Apart from the Board's erroneous exclusion of Be-Lo's evidence of attrition and new hires prior to March 20, it is also apparent that the Board counted many authorization cards that were at least of questionable validity (even if not invalid) in the course of its finding that the Union had attained -- by only 24 cards-- a majority of sup- port as of March 20. For example, Be-Lo challenged 40 cards as invalid because they were obtained under the false pretenses that the sole purpose for the cards was to authorize an election.3 Be-Lo pres- ented the testimony of forty-three current and former employees who testified that they signed cards only after they were falsely assured that the only purpose of the cards was to seek an election, rather than _________________________________________________________________ 3 If the language on a card is "deliberately and clearly cancelled by a union adherent with words calculated to direct the signer to disregard and forget the language above his signature," see Gissel, 395 U.S. at 606-07, then a pro-union authorization card may not be counted. Such words of cancellation obviously include statements "that the card [is] to be used solely for the purpose of obtaining an election." Id. at 584; see also Burlington Industries, Inc. v. N.L.R.B., 680 F.2d 974, 976 (4th Cir. 1982). 11 to record their preference for union representation. This testimony was directly corroborated by the testimony of a Union witness, Kim Howell, who testified on cross-examination that she signed her card after she was told that the card's only purpose was to seek an election and that she, in turn, likewise misrepresented the purpose of the authorization cards to numerous other employees: BE-LO's COUNSEL: [I] want to know what [Juanita Frid- ley] told you. Did she tell you that the card was necessary in order to have an election? HOWELL: Yes. BE-LO's COUNSEL: Did she tell you that the card would be used only to get an election? HOWELL: Yes. At the end of re-cross-examination, Judge Linsky elicited the follow- ing additional testimony from Howell: JUDGE LINSKY: The only reason for you to sign the card was to get an election? HOWELL: Yes. JUDGE LINSKY: No other reason? HOWELL: Just to get an election. JUDGE LINSKY: Okay, and what did you tell these people you got to sign cards? HOWELL: I told them it was just for election. .... JUDGE LINSKY: So, you told -- you asked people to sign the card to get an election only? 12 HOWELL: Yes. JUDGE LINSKY: And no other reason? HOWELL: Just for election. JUDGE LINSKY: All right. J.A. Vol. I, at 204; see also J.A. Vol. II, at 602-03; J.A. Vol. III, at 1024. Notwithstanding this testimony, the ALJ, in the following perfunc- tory footnote, credited all of the Union's thirty-seven organizer- witnesses who claimed that they never misrepresented the purpose of any cards, see J.A. Vol. I, at 377, and discredited every single one of Be-Lo's 43 witnesses: In so far as there are inconsistencies, I credit some witnesses over others: I credit Organizer Karen Gompers over Lucille Thornburg and Linda Urguelles; I credit Organizer Fred Carter over Julie Price and Debbie White; I credit Organizer Dudley Saunders over Joyce Riddick and Roy Ray; I credit Organizer Berry Morrisette over William McCain; I credit Organizer Danny Murray over Lola Ford, Allen Butcher, Katherine Schuman, and Adelle Williams; I credit Organizer Cynthia Allgood over Dorethea Holley and Venus McAl- lister, I credit Organizer Charles Garber over Jamie Turner, I credit Organizer Juanita Fridley over Richard Davis, Deb- orah Ann Moser, and Will March Jr.; I credit Organizer Pat- rick Burgwin over Pamela LeJeune; I credit Organizer Russell Wise over Michael Gray, Ronald Taylor, and Harry Fitzpatrick; I credit Organizer Jim Green over Barbara Ridd- ick, Mary Johnson, Shirley Nease, Barbara Wilson, Larry Parker, Crystal Frazier, and Frank Davis; I credit Organizer Donna McNutt over Terry Huffstetler, John Mehan, Sandra Carvey, William McConnell, and Frank Davis; I credit Organizer Steven Henry over Aurelia Watts, Cordelia Beas- ley, and Julie Borman; I credit Organizer Tom Rogers over Evelyn Chappell, Hatti Edwards; Shelia Morgan, Pauline 13 Pawhit, Frank Flora, Gloria Love, Theresa Robles, and Charlene Augson; I credit Organizer Vera Harrison over Shirletta Pope, Shirley Luter, and Jessica Goode; I credit Organizer Jim Jarboe over Janet Spence; I credit Organizer Mark Frederici over Catherin Lattuga; I credit Organizer Paul Evans over Vickie Brown; I credit Organizer Chad Yound over Harry Parker, I credit Organizer Willie Snow over Grace Calloway; I credit Organizer Lynn Colbert over Vickie Shatzoff, Rowmesa Hoyen, Chris Thomas, and Vanessa Devlin; and I credit Organizer Michael Heflin over Yvonne Giles. See J.A. Vol. I, at 377 n.3. The ALJ stated conclusorily in the text in which this footnote appears just that "[m]any of [Be-Lo's] employees . . . were so terrified about their job security. . . that they were ready, willing, and able to help Be-Lo's case by [committing perjury]. . . . Others were merely mistaken. . . ." See id. at 377. The Board just as perfunctorily affirmed the ALJ in but a single sentence within a foot- note. See id. at 352 n.36. Where an ALJ provides no more than a generalized, conclusory statement purportedly incorporating a host of individual comparative credibility determinations with respect to multiple witnesses, we refuse to indulge the presumption that its findings are entitled to the ordinary deference. Cf. Burlington Industries, Inc. v. N.L.R.B., 680 F.2d 974, 977 (4th Cir. 1982) ("We are not, however, required to accept [the] ALJ's credibility determinations where they are not sup- ported by substantial evidence."). Otherwise, savvy ALJ's could sim- ply ground their judgments in broad, categorical statements that they credit all of one party's witnesses and discredit all of the other party's witnesses, and thereby effectively insulate their decisions from mean- ingful judicial review. Here, for example, we suppose it is possible that the ALJ's finding that each and every one of the Union's thirty-seven witnesses was more credible on the stand than each and every one of the company's forty-three witnesses was "simply the happenstance result of a dispas- sionate consideration of the substantive testimony of the [eighty] wit- nesses in question." Fieldcrest Cannon, Inc. v. N.L.R.B., 97 F.3d 65, 78 (4th Cir. 1996) (Luttig, J., concurring in part and dissenting in 14 part). However, the far more likely explanation, particularly given the testimony by Kim Howell, is simply that Be-Lo was never provided a fair chance to dispute the validity of the Union's cards on the grounds that they were obtained under false pretenses. The ALJ counted other highly questionable Union authorization cards as well. For example, in N.L.R.B. v. Heck's, Inc., 386 F.2d 317 (4th Cir. 1967), we held that cards obtained by supervisors cannot be counted in determining a union's majority status unless there is "other proof of voluntariness," reasoning that, [a supervisor's] day to day authority over employees pro- vide[s] a basis for potential tyranny when improperly exer- cised by a supervisor thwarted in his aim to obtain union recognition, and an employee properly could doubt his abil- ity to obtain protection by appeals to higher company authority. Id. at 322. Despite our clear holding in Heck's, the ALJ counted 15 cards that were indisputably obtained by supervisors. The ALJ also counted 13 cards not authenticated by any witness, see J.A. Vol. I, at 315-16, based upon its own comparison of the signatures on the cards with the signatures on the respective employees' W-4 forms, even though Federal Rule of Evidence 901(b)(3) allows authentication by comparison only where the trier of fact has made the comparison with "specimens" which themselves "have been authenticated."4 Even in _________________________________________________________________ 4 At one point during a colloquy between the ALJ and the Union with regard to the challenged authorization cards, the Union offered to call a lay witness to compare the signatures on the authorization cards with the signatures from the respective employees' W-4 forms. The ALJ stated that that would not be necessary, and the company agreed that such would be unnecessary, that the judge could make the comparison without lay assistance. After the company agreed that a lay witness was unneces- sary to compare the sets of signatures, the ALJ said, in an impromptu statement, that the parties had stipulated that the W-4s were the authentic W-4s from the company's files. Insofar as we can discern, the parties never stipulated to anything. However, even if the company's silence in response to the ALJ's off-hand remark, which remark did not call for any response from the company's counsel, were understood as an affirmative 15 Action Auto Stores, 298 NLRB 875 (1990), enfd. mem. 951 F.2d 349 (6th Cir. 1991), upon which the Board relied in affirming the ALJ's admission of these particular cards, see J.A. Vol. I, at 352 n.36, the ALJ compared the signatures on the cards "with the same authenti- cated specimens utilized by the expert witnesses," which were "known to bear the [employees'] signatures." Action Auto, 298 NLRB at 879; see also N.L.R.B. v. General Wood Preserving Co., 905 F.2d 803, 811 (4th Cir. 1990) (noting that authenticity was proven by com- paring signatures "with what the undisputed evidence established to be the authentic specimens"). In the aggregate, these actions by the ALJ, even if not error for rea- sons that are not apparent in the record, only contribute to our concern that the Union's majority status was one of agency construct, rather than grassroots support.5 B. Even if the Board had properly found that pro-Union employees constituted a majority on March 20, the Board nonetheless erred in _________________________________________________________________ stipulation, it was a stipulation at most to the authenticity of the forms generally, not a stipulation to the signatures on those forms. Not even the Union's counsel seems to have believed that the parties stipulated to the authenticity of the signatures on the W-4 Forms. When, subsequently, he separately introduced into evidence each W-4 Form, he repeatedly and consistently recited that each form only "purportedly" was signed by the named individual, a careful representation that obviously carried with it the implicit acknowledgment that the signature had never actually been authenticated. See J.A. at 1082-97. In fact, the ALJ himself referenced the W-4s in the same way, clearly suggesting that not even he believed that the parties had stipulated to the authenticity of the signa- tures. Id. at 1087 ("Why don't you state what it is though, you know, the number and the authorization card of so and so and the W-4 purportedly from the same person." (emphasis added)). 5 Because the General Counsel never made the requisite showing that the Union achieved majority status, it follows a fortiori that the Board erred in holding that Be-Lo violated section 8(a)(5) by declining to bar- gain with the Union after March 20, 1991. 16 imposing a bargaining order on the facts in this case. "[Section] 8(a)(1) violations alone are not enough to support a Gissel order," Appletree Chevrolet II, 671 F.2d at 841; rather, the second part of the Appletree Chevrolet I four-part test imposes, in effect, two additional requirements. First, the employer's misconduct must have been wide- spread and far-reaching, and second, this pervasive misconduct must have had the demonstrable effect of dissolving the majority once pos- sessed by the Union. Although, in imposing its bargaining order, the Board characterized Be-Lo's violations as "hallmark," see J.A. Vol. I, at 353, "[n]ot every `hallmark' violation amounts to exceptional misconduct under Gissel," Kinney Drugs, Inc. v. N.L.R.B., 74 F.3d 1419 (2nd Cir. 1996), due in no small part to the Board's increasing practice of characterizing even relatively minor unfair labor practices as "hallmark." Here, the record belies any assertion either that the company's violations were widespread or that they eroded the Union's majority status. In order to be considered "pervasive," a company's unfair labor practices must, as the word connotes, be felt throughout all, or virtu- ally all, of the bargaining unit. Be-Lo operated thirty retail grocery stores in the bargaining unit during the period in question and, accord- ing to the Board's findings (a number of which we reverse herein as erroneous), a total of only some forty-three individual section 8(a)(1) violations were committed in approximately half of Be-Lo's stores,6 and individual section 8(a)(3) violations occurred in only eight of the _________________________________________________________________ 6 In its brief before this court, the Union alternatively claimed that the ALJ held that Be-Lo committed "some 50" individual section 8(a)(1) violations, Intervenor's Br. at 6, and "over 70" violations of section 8(a)(1), id. at 7. Similarly, the Board claimed that the ALJ found "more than 70" individual violations of section 8(a)(1), Appellee's Br. at 12, and "at least 80-some" violations, id. at 41. The confusion of the Board and Union with respect to the number of section 8(a)(1) violations found by the ALJ is doubtless attributable in part to the fact that neither the ALJ nor the Board in their opinions totaled the number of violations that were found by the ALJ, preferring instead simply to refer to the section 8(a)(1) and section 8(a)(3) violations collectively as "repeated" or "nu- merous," J.A. Vol. I, at 360, 374 (ALJ's opinion), and "over fifty," id. at 352 (Board's opinion). In our review of the ALJ's opinion, however, we have identified only some forty-three section 8(a)(1) violations aris- ing from Be-Lo's treatment of individual employees. 17 stores.7 Despite the Board's reflexive characterization of these viola- tions as "hallmark," it is questionable whether a number of these were violations at all. The ALJ held, for example, that Be-Lo supervisor Cheryl Perras "unlawfully interrogated [Evelyn] Keyes and threatened store closure, loss of jobs, and reduced hours for employees if the Union was selected" when Perras merely "asked Keyes what she (Keyes) thought about the Union" and "went on to tell Keyes that if the Union got in and wages were raised some stores might close and people lose their jobs or hours would be cut." See J.A. Vol. I, at 285. The ALJ also found that the company violated section 8(a)(1) when a co-manager of one of Be-Lo's stores simply mentioned to an employee that he had seen the employee with a Union organizer. As the ALJ described the violation: In January 1991 Cottrell was approached by union organizer Karen Gompers and shortly thereafter co-manager of store 102 David Rodriguez told Cottrell he had seen her with the Union organizer. This statement by Rodriguez to Cottrell creates the impression that Cottrell's union activity was under surveillance and being monitored by management in violation of Section 8(a)(1) of the Act. Id. at 289. And there are numerous similar examples throughout the ALJ's opinion. Even were the content of the remarks themselves more serious, these violations, collectively, would still not qualify as pervasive. Vir- tually all of the forty-three section 8(a)(1) violations and the fourteen section 8(a)(3) violations found by the Board arise from a statement or an action by a single supervisor to a single employee outside the presence of others. As a consequence, only some 40 of Be-Lo's 756 employees (less than six percent of Be-Lo's work force) were directly affected by these individual violations. Unfair labor practices that impact such a small portion of the bargaining unit do not qualify as _________________________________________________________________ 7 The Board, reversing the ALJ, also justified its pervasiveness finding with the fact that Be-Lo violated the Act by distributing its pink slip flyer to all of its employees shortly before the election. As we explain infra Section IV, we agree with the ALJ that the company's distribution of the pink slip flyer did not constitute a violation of section 8(a)(1). 18 pervasive. See Somerset Welding & Steel, Inc. v. N.L.R.B., 987 F.2d 777, 780 (D.C. Cir. 1993) (refusing to enforce bargaining order where only ten percent of the employees were directly affected by the employer's unfair labor practices); Avecor, Inc. v. N.L.R.B., 931 F.2d 924 (D.C. Cir. 1991) (refusing to enforce bargaining order; six per- cent of labor force directly affected). Finally, even if the violations here had been of both the number and seriousness that could support a finding of pervasive violation, there is no evidence that Be-Lo's practices "dissipat[ed] a previous union majority." Appletree Chevrolet II, 671 F.2d at 841. In Appletree Chevrolet II, we held that concrete evidence, not "pure[ ] speculat- [ion]," must support a Board determination that a union's majority status was dissipated by pervasive unfair labor practices. Id. Viola- tions of the kind and number at issue here simply could not have dis- solved the Union's majority status in Be-Lo's 756 member employee bargaining unit absent wide dissemination, of which there is no evi- dence in this record. The record reflects that Be-Lo's work force is transient by nature, and that its employees are widely dispersed throughout thirty separate stores in seven different cities. Even at the same store, employees may have little contact with their fellow employees who work different shifts. Common sense suggests that this is not the type of corporate workplace in which a relatively few statements made to, or employment actions taken against, single indi- viduals in particular stores are likely to affect Union support company-wide. And the record in this case confirms this common- sense assessment that the effects of the individual violations were insignificant. During the election campaign period of January 26 to March 20, the Union obtained 69 new authorization cards, of which 50, or 72%, were procured from employees working at the fifteen stores in which Be-Lo committed its "unfair labor practices." Thus, if the unfair practices had any real effect at all (which we doubt), it would appear that they galvanized support for the Union. Cf. Red Oaks Nursing Home, 633 F.2d at 510 ("[T]he most compelling evi- dence that a bargaining order was not warranted is the clear evidence in the record that the employer's unfair labor practices did not drive the employees directly involved to abandon the union."). C. Even if the incidents of impermissible conduct are pervasive and erode a union's majority status, the Board must still carefully analyze, 19 in record findings, the "continuing impact" of those violations. See Appletree Chevrolet II, 671 F.2d at 841 ("The continuing impact of [pervasive unfair labor practices] is the important matter."); Appletree Chevrolet I, 608 F.2d at 996-97. Or, as the Sixth Circuit has held, "in reaching its determination to issue a bargaining order, the Board must make factual findings and must support its conclusion that there is a causal connection between the unfair labor practices and the probabil- ity that no fair election could be held." Henry Bierce Co. v. N.L.R.B., 23 F.3d 1101, 1110 (6th Cir. 1994) (emphasis in original) (internal quotation marks omitted). Moreover, the Board also "must make a detailed analysis assessing . . . the potential effectiveness of ordinary remedies." Appletree Chevrolet I, 608 F.2d at 997 (internal quotation marks omitted). When the Board issues a bargaining order without analyzing and "explain[ing] with specificity" why more traditional remedies such as a cease and desist order, posted notice, or similar less drastic remedies would not be equally effective, that order cannot stand. Id. at 997 n.10 (quoting N.L.R.B. v. Craw, 565 F.2d 1267, 1272 (3d Cir. 1977)); see also id. at 997-98; Red Oaks Nursing Home, 633 F.2d at 509 ([C]onsideration of the effectiveness of ordinary remedies . . . [is] prerequisite for a bargaining order). Plainly, the Board did not undertake the requisite comprehensive analysis here. The Board's "continuing impact" analysis was limited to a summary of Be-Lo's violations and management's involvement in those violations, followed by the conclusory comment, unsupported by any evidence, that "`the cloud created by these violations [is] likely to linger' and cannot be dispersed by a traditional cease-and- desist order." J.A. Vol. I at 354 (quoting Avecor, 931 F.2d at 938 (dicta)). This is not the "detailed analysis" of the "continuing effect of misconduct, and the potential effectiveness of ordinary remedies" that we require under Appletree Chevrolet I, 608 F.2d at 997 (internal quotation marks omitted). If ever a fair rerun election would be possible, then it is here, given the substantial turnover in Be-Lo's work force and other changed cir- cumstances since the violations occurred. An inordinate amount of time has passed since Be-Lo's unfair labor practices were committed and, as one of our sister circuits has noted, the adverse effects of all but the most egregious unfair labor practices become "dissipated" "[w]ith the passage of time," "particularly when that conduct 20 amounted to insubstantial violations in the first place." Laverdiere's Enterprises, 933 F.2d 1045, 1055 (1st Cir. 1991). The election cam- paign in which almost all of Be-Lo's unfair labor practices occurred was conducted in February and March of 1991. The ALJ did not issue its decision until September 15, 1993, and the Board did not issue its order until July 15, 1995. Thus, two years elapsed between the elec- tion and the ALJ's decision, and another two years elapsed before the Board rendered its decision affirming the ALJ. "[I]nordinate delay attributable to the Board `cannot be ignored in developing a remedy. . . .'" Id. at 1055 (quoting Texas Petrochemicals Corp. v. N.L.R.B., 923 F.2d 398, 406 (5th Cir. 1991)). It strains credulity to believe that Be-Lo's unfair labor practices, such as they were, had such long last- ing effects that a fair rerun election could not have been held four years later, much less today, some six years after the original viola- tions occurred. Even more important than the passage of time, the vast majority of Be-Lo's workers who were present when Be-Lo committed the labor violations are no longer even employed at Be-Lo. We have long held, contrary to the Board's view, see, e.g., Q-1 Motor Express, 308 N.L.R.B. 1267 (1992); Salvation Army Residence, 293 N.L.R.B. 944, 945 (1989), that "significant employee turnover through normal attri- tion" is highly relevant to determining the necessity of a bargaining order and well "may make a bargaining order inappropriate." So-Lo Foods, 985 F.2d at 128-29; see also Appletree Chevrolet II, 671 F.2d at 841-42, and we reaffirm those decisions today. Significant employee turnover often (if not usually) renders a bargaining order inappropriate. Not only is the possibility of a fair rerun election great when "many of the intimidated employees have moved on" and been replaced by new workers who have not witnessed the company's unfair labor practices, Avecor, 931 F.2d at 936, but "the issuance of a bargaining order in the face of significant employee turnover risks unjustly binding new employees to the choices made by former ones," J.L.M., Inc. v. N.L.R.B., 31 F.3d 79, 84 (2nd Cir. 1994). See Appletree Chevrolet II, 671 F.2d at 841-42 (declining to enforce a Gissel order in light of a 75% employee turnover). Over sixty-six percent of Be-Lo's employees turned over between the March 21, 1991, election and entry of the Board's order. See J.A. Vol. I, at 353. At oral argument, neither counsel for the Board nor 21 counsel for the Union could cite us to any case in which a Gissel Cat- egory II order was enforced in the face of employee turnover so high.8 On the other hand, numerous courts have denied enforcement of Gissel Category II orders where workplace turnover was considerably less than that at Be-Lo. See, e.g., N.L.R.B. v. Marion Rohr Corp., 714 F.2d 228, 231 (2nd Cir. 1983) (order not enforced in light of 35% turnover); N.L.R.B. v. Chester Valley, Inc., 652 F.2d 263, 273 (2nd Cir. 1981) (order not enforced in light of 34-60% turnover); Red Oaks Nursing Home, 633 F.2d at 510 (order not enforced in light of 35- 64% turnover). Of course, high employee turnover is an even more important consideration where, as here, it was unlikely, because of separate work shifts and multiple business locations, that the unfair labor practices had much impact company-wide in the first place. While acknowledging that only one third of the employees employed when the violations occurred still remained with the com- pany when the mandatory bargaining order issued, the Board dis- missed the employee turnover evidence with the terse observation that "`[p]ractices may live on in the lore of the shop and continue to repress employee sentiment long after most, or even all, original par- ticipants have departed.'" J.A. Vol. I, at 353 (quoting Bandag, Inc., 583 F.2d 765, 772 (5th Cir. 1978) (dicta)). Absent substantial eviden- tiary support that the effects of unlawful practices have in fact contin- ued to be felt in the workplace, we believe that such inferences as to the likely effect of "lore of the shop" have no place in the calculus of whether a mandatory bargaining order is warranted. Under such a speculative and indeterminate standard, the Board could decide in every case that "the possibility of a fair rerun election is slight," even _________________________________________________________________ 8 In So-Lo Foods, although we noted that "turnover through normal attrition may make a bargaining order inappropriate and is relevant to th[e] inquiry" of whether to issue a mandatory bargaining order, 985 F.2d at 128-29, we declined to allow the company to supplement the record before us so as to include an affidavit in which the company's Director of Personnel represented that there had been a 73% employee turnover. We did so, however, because all but "perhaps a small portion of the turn- over" occurred after the Board's decision. See id. at 128 n.6. Indeed, the turnover that had occurred prior to the ALJ's decision was so minimal that the ALJ refused even to hear the evidence, a decision with which the company itself implicitly agreed. Id. 22 if the entire work force had turned over since commission of the unfair labor practices, thus eviscerating one of the most important of the heightened requirements for a Gissel Category II mandatory bar- gaining order. Just as the Board's reliance on boiler-plate language discussing egregious misconduct in Appletree Chevrolet I would have led "in effect [to] the automatic issuance of bargaining orders," 608 F.2d at 998, so also here would talismanic invocation of the "lore of the shop," if approved, serve automatically to validate the issuance of bargaining orders. The perniciousness of such a blind acceptance of the "lore of the shop" is evident here, as there was no evidence what- soever that the effects of Be-Lo's unfair labor practices were "living on in the lore of Be-Lo's shops." Indeed, the highly transient nature of Be-Lo's work force, and the fact that the company does business in some 30 different locations in seven different cities, underscore the likelihood that whatever "lore" existed in Be-Lo's "shops" would have had little chance of "living on" for as many years as have elapsed. In the end, it is apparent that the Board here -- much like the Board in Appletree Chevrolet -- simply "list[ed] unfair labor prac- tices" and "followed [with a] conclusory statement that a fair election is no longer possible," "disregard[ing the] factors that militate against enforced bargaining," Marion Rohr, 714 F.2d at 231 (citing Appletree Chevrolet I, 608 F.2d at 997-98), and undertaking only the most superficial analysis of alternative remedies. Mandatory bargaining simply cannot be ordered on such a basis. III. The Board, reversing a contrary holding by the ALJ, held that Be- Lo had discriminatorily denied Union members access to its property in violation of section 8(a)(1) when it sought and maintained state court injunctions to remove paid non-employee Union picketers from the sidewalks and parking lots of certain stores in the period follow- ing the election. The Board erred in this conclusion also. The Board held that Be-Lo's efforts to exclude non-employee Union picketers infringed the employees' rights under section 7 of the Act, which provides that "[e]mployees shall have the right to self- organize, to form, join or assist labor organizations." 29 U.S.C. § 157. 23 The Supreme Court has held, however, that, once an employer's pri- vate property rights are established under state law, the employer gen- erally "cannot be compelled to allow distribution of union literature by non-employee organizers on his property," Lechmere, Inc. v. N.L.R.B., 502 U.S. 527, 533 (1992), even under the authority of sec- tion 7. The Board recognized as much, see J.A. Vol. I, at 348, but nonetheless held that an employer may forfeit its private property rights and violate section 8(a)(1) of the Act by denying access to Union picketers while allowing other distribution -- whether such other distribution be charitable, political, or civic -- because such conduct "discriminate[s] against the union" in a manner forbidden by N.L.R.B. v. Babcock & Wilcox Co., 351 U.S. 105, 112 (1956). See J.A. Vol. I, at 348. Because nonemployees' claims to access to an employer's private property are at their nadir when the nonemployees wish to engage in protest or economic activities, as opposed to organi- zational activities, see U.F.C.W. v. N.L.R.B., 74 F.3d 292, 300 (D.C. Cir. 1996), we seriously doubt, as do our colleagues in other circuits, that the Babcock & Wilcox disparate treatment exception, post- Lechmere, applies to nonemployees who do not propose to engage in organizational activities. See, e.g., Cleveland Real Estate Partners v. N.L.R.B., 95 F.3d 457, 465 (6th Cir. 1996). If it does, we further doubt that an employer's approval of limited charitable or civic distri- bution while excluding union distribution constitutes discrimination. See id. ("No relevant labor policies are advanced by requiring employers to prohibit charitable solicitations in order to preserve the right to exclude nonemployee distribution of union literature when access to the target audience is otherwise available."). We need not decide this issue today, however, because the record in this case amply reflects that Be-Lo had an enforced policy of disallowing solic- itation and distribution (as even appellees must concede) and that, as the ALJ found, the solicitation that occurred at the several of Be-Lo's stores was only "isolated and sporadic," J.A. Vol. I, at 375. Be-Lo has had in place a broad no-solicitation policy in all of its stores for more than twenty years, and, in furtherance of this policy, the company has always posted no-solicitation signs in the front win- dows of its stores. See J.A. Vol. III, at 1067-74, 1124-59, 1190-92; J.A. Vol. IV., at 1402-04. As the ALJ found, "Be-Lo's policy was a very strong and broad policy which prohibited all solicitations on its property." Id. at 375. While the record reflects that some groups did 24 occasionally engage in distribution at a handful of the company's stores without the knowledge or approval of Be-Lo's owners, execu- tives, or store-level management, these sporadic activities do not evi- dence that Be-Lo took a "laissez-faire attitude" toward enforcement of its no-solicitation policy, which was the equivalent of purposeful discrimination. The record reflects, as stated by the Board: that Muslims selling oils and incense were present on a "pretty constant" basis in front of Store 232 and were pres- ent on a "regular" basis in front of Store 236. Further, an "occasional" Jehovah's Witness distributed the Watchtower magazine at Store 148 and on one occasion a local Lions Club solicited at that store. Also, Lyndon LaRouche follow- ers on a "couple of occasions" handed out literature at Stores 28 and 120. In addition, a person sold a cookbook inside Store 102 and "occasional[ly]" individuals sold Girl Scout cookies and greeting cards inside Store 232. J.A. Vol. I, at 348. These few solicitations, which occurred in but a few of Be-Lo's thirty stores over the period of a year and a half, are no more than could be expected at any large retail chain that was zeal- ously defending its property rights. Certainly, they cannot be said to constitute evidence that Be-Lo "allowed other distribution" as the Court intended that phrase in Babcock. To affirm the Board's contrary finding on this record would be tantamount to a holding that if an employer ever allows the distribution of literature on any of its prop- erty, then it must open its property to paid nonemployee union picket- ers. We are confident that the Supreme Court never intended such a result. Cf. Cleveland Real Estate, 95 F.3d at 465 ("To discriminate in the enforcement of a no-solicitation policy cannot mean that an employer commits an unfair labor practice if it allows the Girl Scouts to sell cookies, but is shielded from the effect of the Act if it prohibits them from doing so."). The Board therefore simply erred when it held that Be-Lo violated section 8(a)(1) by seeking state court injunctions to enforce its no-solicitation policy against the Union's picketers.9 _________________________________________________________________ 9 Because we hold that Be-Lo did not violate the Act by seeking to obtain and enforce state court injunctions to exclude the Union's picket- ers, it again follows a fortiori that the Board also erred when, through a retroactive application of its own decision in Loehmann's Plaza, 305 N.L.R.B. No. 81 (1991), the Board held Be-Lo liable for the Union's attorneys' fees and costs in defending the injunction action. 25 IV. The Board further erred, we are convinced, in reversing the ALJ's conclusion that the distribution of Be-Lo's pink slip flyer did not vio- late section 8(a)(1). As Be-Lo notes, the Board selectively quoted the flyer, the language of which, when read out of context, might well support a conclusion that the pink slip constituted a "threat of repri- sal." The Board only quoted the text of the mock pink slip purportedly signed by other grocery stores that had closed following unionization in the nearby area, together with a portion of a sentence from the flyer which read, "you may want to look at what the [Union] got for their former dues payer's in this area -- a pink slip." J.A. Vol. I, at 341. The pink slip read as follows: Dear Unionized Employees: I regret to inform you that because we have lost our ability to compete in this extremely competitive market, we shall be forced to close this store and put you out of work. J.A. Vol. IX, at 3238-39. The Board held that the distribution of this flyer "threatened that [Be-Lo's employees] would get a `PINK SLIP' if they voted in the Union and became `Unionized Employees,'" because the recipients could easily identify with the employees of the other stores who had lost jobs and thus "reasonably conclude that if they voted in the Union, they would be the `Unionized Employees' addressed in the `Pink Slip' who would suffer store closure and job loss." J.A. Vol. I, at 341. Even assuming that the employees who received the pink slip could have understood the company's message as one that closure of stores might result were the company to unionize, the distribution of this flyer would nonetheless be permissible under the Act, as the ALJ held. As the ALJ explained, an employer has the right to "make a pre- diction as to the precise effects he believes unionization will have on his company . . . [and such a prediction may] convey an employer's belief as to demonstrably probable consequences beyond his control." 26 See Gissel, 395 U.S. at 618. At worst, the pink slip flyer was, as the ALJ held, a prediction, based upon the experiences of similarly- situated stores, of the probable economic consequences for Be-Lo and its employees if Be-Lo were to become unionized, made in response to the Union's repeated claims that unionization represented "job security for all," see, e.g., J.A. Vol. V, at 1690, and that Be-Lo was "telling half truths," "outright lies," "distorting the facts," and "spread- ing propaganda." See, e.g., id. at 1706. The flyer was not a threat that the company would close its doors in retaliation if the Union pre- vailed in the election. See J.A. Vol. I, at 360. In fact, the portions of the flyer omitted by the Board in discussing the flyer confirm as much: Dear Be-Lo Associate, For several months you have been hearing sweet- sounding promises from the out-of-town union organizers, about higher wages, job security, grievance procedures, seniority rights and other expensive benefits which are easy to promise but hard to deliver. At the same time they have also fed you militant anti- BeLo distortions charging this company - your company - with "half truths, lies and propaganda". Half-truths? Lies? Propaganda? When you're trying to figure out who is giving you the facts and who is promising you the moon, you may want to take another look at what the UNITED FOOD AND COMMERCIAL WORKERS UNION got for their former dues payers in the area- A PINK SLIP. We will do everything we can to prevent this from hap- pening to you, whether or not the union is voted in, but you should consider what the union has DELIVERED in the past before you believe what they PROMISE in the future . J.A. Vol. IX, at 3238 (emphasis added). When the full text of the flyer is read, and the flyer is understood in the context of the claims being 27 made by the Union against Be-Lo, it is evident that Be-Lo was not threatening to "take action solely on [its] own initiative for reasons unrelated to economic necessities." See Gissel, 395 U.S. at 618. Even the Board is hardpressed to characterize the flyer as more than a "carefully phrased" "prediction" "as to demonstrably probable conse- quences beyond [Be-Lo's] control." See Gissel, 395 U.S. at 618. It can say only that the flyer "reasonably tended to cause company employees to infer that they would be the `Unionized Employees' addressed in the pink slip, who would suffer the announced store clo- sure and job losses if they voted for the Union." Appellee's Br. at 10. Accordingly, as the ALJ held, the flyer "constitute[d] free speech and legitimate propaganda," see J.A. Vol. I, at 279, protected under sec- tion 8(c) of the Act. See DTR Industries, Inc. v. N.L.R.B., 39 F.3d 106, 113-14 (6th Cir. 1994). V. The Board held that Be-Lo discharged or constructively discharged eleven individual employees in violation of section 8(a)(3). Although we affirm the Board's conclusions that Be-Lo violated section 8(a)(3) with respect to six of these individuals, we hold that the Board erred in affirming the ALJ's finding that the remaining five individual employment actions taken by Be-Lo were unlawful. In holding that Be-Lo committed violations when it did not recall employee Erwin Hatchett and when it discharged employees Shirley Terry, Pamela Jackson, Gwen Andrews, and Michael Salazar, the ALJ relied exclu- sively on evidence that these employees were strongly pro-Union and that they were discharged at a "suspicious time" (meaning during the election campaign period). It has long been the rule that "the mere fact that an employee -- even a union activist-- [is] discharged dur- ing a union campaign, standing alone, will not support a discrimina- tory finding if there is a `supportable cause' for the discharge." Appletree Chevrolet I, 608 F.2d at 993 n.5 (emphasis added); see also, e.g., McLean Trucking Co. v. N.L.R.B., 719 F.2d 1226, 1228-33 (4th Cir. 1983) (reversing violation based on evidence showing that a known union activist was discharged at a suspicious time, where other evidence showed that "supportable cause" existed because of the employee's absenteeism); Wellington Mill Div., West Point Mfg. Co. v. N.L.R.B., 330 F.2d 579, 586 (4th Cir. 1964). With regard to each of these five personnel, Be-Lo presented unrebutted evidence of a 28 "supportable cause" for the discharge. Therefore, the ALJ erred in holding that these personnel actions violated the Act. As noted by both the ALJ and the Board, Be-Lo initially laid off employee Erwin Hatchett because of the general decline in business related to the deployment of troops in the Persian Gulf. The unrebut- ted evidence supports Be-Lo's contention that it did not recall Hat- chett because no new meat cutter positions were available at the stores in Hatchett's area. See J.A. Vol. I, at 345-46; Vol. II, at 671-72. The ALJ and the Board nonetheless held that Be-Lo's failure to recall Hatchett violated section 8(a)(3) solely because there existed evidence that Hatchett was "at least a competent employee, and [Be-Lo] knew of his support for the Union, and . . . failed to recall him as job open- ings arose." J.A. Vol. I, at 346. With the exception of the character- ization of Hatchett as a competent worker, which adds nothing of substance to the inquiry, this is the same type of evidence which we held insufficient in Appletree Chevrolet I to support the Board's find- ing of section 8(a)(3) violations. The Board likewise erred in holding that Be-Lo violated the Act by discharging employees Terry, Jackson, Andrews, and Salazar. The Board held that Shirley Terry's discharge was unlawful solely on the ground that she was fired after Terry's "pro-Union stance was well- known to [Be-Lo]," and after her picture appeared in a Union flyer. J.A. Vol. I, at 343; see also id. at 366. In the face of Be-Lo's unrebut- ted evidence that Terry was discharged for admittedly giving away food to a co-employee without store authorization, see J.A. Vol. IV, at 1507, the Board and the ALJ were unjustified under Appletree Chevrolet I in concluding that Terry's dismissal was an act of "harass- ment" because of her "prounion posture." J.A. Vol. I, at 343-344. Similarly, the Board was not justified in concluding, solely from the evidence that Be-Lo was "well aware" of Pamela Jackson's prounion activities and that Jackson had received a prior favorable work perfor- mance evaluation, J.A. Vol. I, at 373, that Jackson was discharged for discriminatory reasons, when the company presented unrebutted evi- dence that Jackson was discharged for failing to comply with the company's request to provide a doctor's note after she had called in sick under suspicious circumstances. See J.A. Vol. I, at 258-59; J.A. Vol. III, at 790-97; J.A. Vol. V, at 1659-60. In holding that Gwen Andrews, who was "expelled" from her job by a pro-union manager, 29 was illegally discharged, the ALJ yet again relied solely on the legally insufficient evidence that Andrews was discharged two weeks before the election and was a known union supporter, see J.A. Vol. I, at 371, when the company presented unrebutted evidence that Andrews was expelled for refusing to follow her manager's instructions. See id. The Board's conclusion that Be-Lo violated the Act when it discharged Michael Salazar is, for the same reason, unsustainable. Although Be- Lo clearly knew of Salazar's prounion feelings and indisputably dis- charged him during the election campaign, Be-Lo's evidence that it fired Salazar because of his pervasive absenteeism stood unrebutted. See J.A. Vol. IV, at 1383, 1386. We uphold, however, the remaining three violations challenged by Be-Lo on appeal -- the constructive discharges of Jaime Cottrell and Kelly Call and the discharge of Sabrina Frazier. In those cases, although Be-Lo presented compelling evidence that the employees at issue either voluntarily quit or good cause existed, there was substan- tial evidence in the record, in addition to the employee's pro-union stance and the timing of his or her departure, which supports the ALJ's finding of a Section 8(a)(3) violation. There is direct evidence in the record, for example, that Cottrell was transferred from store 102 to store 67 as punishment for prounion activities and that she was unable to work as a result of this transfer. Cottrell testified that Be- Lo's district manager told her that since "he could not fire her" he would send her to store 67, "an antiunion store," where "[a]ll of my employees hate the union" and "want to know who the hell Jaime [Cottrell] is," and also that he would "be on her like [a] fl[y] on shit." J.A. Vol. I, at 342, 366; J.A. Vol. II at 545-58. And while Be-Lo claims that Kelly Call quit voluntarily, there was direct evidence that Kelly Call's produce manager expelled her solely for wearing a Union t-shirt, and that thereafter she was informed by a store manager that she had "quit." See J.A. Vol. I, at 370. And although the ALJ relied in part on the suspicious timing of Sabrina Frazier's discharge in holding that discharge to be unlawful, see J.A. Vol. I, at 372, this decision does not run afoul of Appletree Chevrolet I because the record contains evidence that a Be-Lo manager told Frazier she was "digging her own grave" by appearing in a pro-Union flyer shortly before her discharge, and that the reason that store officials were "rid- ing her ass" was because of her support for the Union. J.A. Vol. II, at 719, 729. The additional direct evidence of discriminatory dis- 30 charge in these three cases sets them apart from the other five chal- lenged on appeal, and provides sufficient, albeit not conclusive, support for the ALJ's findings. We also affirm the Board's holdings, expressly uncontested by Be- Lo on appeal, that Be-Lo violated Section 8(a)(3) by discharging employees Angela Cox, Kim Howell, and Lavonne Billups. CONCLUSION The Board's order of mandatory bargaining is reversed, as are its conclusions that Be-Lo violated section 8(a)(1) by distributing the pink slip flyer and by seeking injunctions to prevent nonemployee picketing, and that the company violated section 8(a)(3) by not recall- ing Erwin Hatchett and discharging Shirley Terry, Pamela Jackson, Gwen Andrews, and Michael Salazar. We affirm the Board's findings that Be-Lo violated section 8(a)(3) when it discharged employees Jaime Cottrell, Kelly Call, Sabrina Frazier, Angela Cox, Kim Howell, and Lavonne Billups. The case is remanded to the Board for entry of such order as remains appropriate in light of the opinion herein. REVERSED IN PART, AFFIRMED IN PART, AND REMANDED 31 Volume 2 of 2 PUBLISHED UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT BE-LO STORES, Petitioner, v. NATIONAL LABOR RELATIONS BOARD, No. 96-1575 Respondent, UNITED FOOD AND COMMERCIAL WORKERS INTERNATIONAL UNION, Local 400, AFL-CIO, CLC, Intervenor. NATIONAL LABOR RELATIONS BOARD, Petitioner, UNITED FOOD AND COMMERCIAL WORKERS INTERNATIONAL UNION, Local 400, AFL-CIO, CLC, No. 96-1657 Intervenor, v. BE-LO STORES, Respondent. On Petition for Review and Cross-Application for Enforcement of an Order of the National Labor Relations Board. (11-CA-14586, 11-CA-14775-1, 11-CA-14775-1, 11-CA-14775-2, 11-CA-14775-3, 11-CA-14775-4, 11-CA-14775-5, 11-CA-14775-6, 11-CA-14775-7, 11-CA-14775-8, 11-CA-14775-9, 11-CA-14775-10, 11-CA-14712, 11-CA-14793-1, 11-CA-14793-2, 11-CA-14793-3, 11-CA-14793-4, 11-CA-14793-5, 11-CA-14811, 11-RC-5823) Argued: January 27, 1997 Decided: September 16, 1997 Before ERVIN and LUTTIG, Circuit Judges, and HILTON, United States District Judge for the Eastern District of Virginia, sitting by designation. Volume 1 Opinion by Judge Luttig Volume 2 Dissent by Judge Ervin _________________________________________________________________ Reversed in part, affirmed in part, and remanded by published opin- ion. Judge Luttig wrote the majority opinion, in which Judge Hilton joined. Judge Ervin wrote a dissenting opinion. _________________________________________________________________ COUNSEL ARGUED: Stanley Graves Barr, Jr., Charles Vincent McPhillips, KAUFMAN & CANOLES, P.C., Norfolk, Virginia, for Be-Lo Stores. John Emad Arbab, NATIONAL LABOR RELATIONS BOARD, Washington, D.C., for NLRB. Carey Robert Butsavage, BUTSAV- AGE & ASSOCIATES, P.C., Washington, D.C., for Intervenor. ON BRIEF: Arlene F. Klinedinst, Ashley L. Taylor, Jr., KAUFMAN & CANOLES, P.C., Norfolk, Virginia, for Be-Lo Stores. Frederick L. Feinstein, General Counsel, Linda Sher, Associate General Counsel, Aileen A. Armstrong, Deputy Associate General Counsel, Peter Winkler, Supervisory Attorney, William A. Baudler, NATIONAL LABOR RELATIONS BOARD, Washington, D.C., for NLRB. George Wiszynski, BUTSAVAGE & ASSOCIATES, P.C., Washing- ton, D.C., for Intervenor. _________________________________________________________________ 34 ERVIN, Circuit Judge, dissenting: If the facts of this case cannot sustain a bargaining order, then almost none will. Substantial evidence plainly supports the Board's findings of nearly 80 violations, many of them so-called "hallmark" violations, the vast majority of them unbriefed by Be-Lo and not even discussed by the majority here. Among the particulars that the major- ity does discuss, substantial evidence supports the Board's finding that Be-Lo violated the Act when it sent its employees a mock "pink slip" just before the election, which the Board found to constitute an actual threat of job loss if the employees voted for the Union. In addi- tion, substantial evidence supports the Board's finding that Be-Lo violated the Act by disparately enforcing its no-solicitation policy to deny Union picketers access to Be-Lo's property. There ought to be really little question that the Board properly issued a bargaining order in this case. Indeed, because so many of the violations are hallmark violations, being, on the whole, "outrageous" and "pervasive," the bargaining order was warranted even without a showing of majority support for the Union. Although neither the ALJ nor the Board classified whether the bargaining order fell into the first or second category of appropriate orders under NLRB v. Gissel Pack- ing Co., 395 U.S. 575 (1969), more than sufficient evidence supports a determination that a Gissel order was warranted under the first cate- gory. In any event, the Board properly determined that the Union did, in fact, enjoy majority support. Be-Lo's attempt to introduce evidence showing that the Union did not possess majority support, which was rejected by both the ALJ and the Board, fails here, too, despite the majority's characterization of the matter. Be-Lo failed to fully comply with a subpoena requesting certain payroll records and cannot com- plain, ex post, that the General Counsel relied on an amended Excelsior list to establish a Gissel majority. Even if the bargaining order is thus seen as falling within Gissel's second category, its issu- ance was proper. Because I believe that the majority ignores or twists precedent, our proper standard of review, and the facts in this case, I believe that Be- Lo's petition for review should be denied and the Board's cross- application for enforcement of its order should be granted in its entirety. I therefore respectfully dissent. 35 I. Notably absent from the majority's opinion is any discussion of this court's proper standard of review of Board decisions and orders. The Board's decision must be upheld if its factual findings are sup- ported by substantial evidence in the record. See Universal Camera Corp. v. NLRB, 340 U.S. 474, 490-91 (1951); FPC Holdings, Inc. v. NLRB, 64 F.3d 935, 942 (4th Cir. 1995); 29 U.S.C. § 160(e). This is so "even though we might have reached a different result had we heard the evidence in the first instance." NLRB v. General Wood Pre- serving Co., 905 F.2d 803, 810 (4th Cir.) (citing NLRB v. Daniel Con- str. Co., 731 F.2d 191, 193 (4th Cir. 1984)) (internal quotation marks omitted), cert. denied, 498 U.S. 1016 (1990). "Substantial evidence has been held to mean such relevant evidence as a reasonable mind might accept as adequate to support a conclusion." Id. (internal quota- tion marks and citation omitted). We have previously advised that our inquiry should end, and the Board's order be enforced, particularly in those circumstances where, as here, "the record is fraught with con- flicting testimony and essential credibility determinations have been made." NLRB v. Nueva Eng'g, Inc., 761 F.2d 961, 965 (4th Cir. 1985) (citing NLRB v. Air Prods. and Chems., Inc., 717 F.2d 141, 145 (4th Cir. 1983); Dubin-Haskell Lining Corp. v. NLRB, 375 F.2d 568, 571 (4th Cir. 1967)). Throughout its opinion the majority simply ignores this well-established standard of review and engages in its own de novo review of the evidence, substituting its own judgment for that of the Board at almost every turn. The majority thus refuses to acknowledge that this court sits only as an appellate court of limited jurisdiction whose powers to review Board proceedings are not only not plenary but have been greatly circumscribed by Congress. Moreover, we have previously recognized that the Board possesses a broad discretion to choose an appropriate remedy. NLRB v. Williams Enters., 50 F.3d 1280, 1289 (4th Cir. 1995). We have therefore reviewed that selection deferentially for an abuse of the discretion lodged in the Board. Indeed, in Williams we stated that we "must enforce [the Board's] choice `unless it can be shown that the order is a patent attempt to achieve ends other than those which can fairly be said to effectuate the policies of the NLRA.'" Id. (quoting Virginia Elec. & Power Co. v. NLRB, 319 U.S. 533 (1943), and citing Overnite Transp. Co. v. NLRB, 372 F.2d 765, 770 (4th Cir.), cert. denied, 389 36 U.S. 838 (1967)) (emphasis added).1 The majority fails entirely to link this standard of review to the Board's choice of remedies in this case. Applying these standards of review appropriately to the case at hand, a case involving numerous egregious violations of the NLRA, many of them hallmark violations of the most serious kind, one can only conclude that the Board's findings are plainly supported by sub- stantial evidence in the record as a whole and that the Board did not abuse its discretion in its selection of remedies. II. In its refusal to enforce the Board's bargaining order, the majority has focused on the forest of the Board's remedy but ignored most of the trees that make that remedy appropriate in this case. The Board found nearly 80 violations of the NLRA. Be-Lo contested only a handful of those violations before us, however, and the majority con- centrates its energies on only that handful. Although I agree with the majority about the confusion surround- ing the number of violations, see slip op. at 17 n.6, I disagree greatly about the final tally. By my own careful, yet conservative, count, I believe the Board found a total of 77 violations.2 In this total number are several that the majority refuses to sustain, including the Board's specific reversal of the ALJ's findings that Be-Lo's own election campaign was not a violation of the Act and that Be-Lo had not dis- criminatorily enforced its no-solicitation policy against the picketers. The Board instead found that Be-Lo's threats to employees of store closure and job loss if the Union won the election, by means of a pre- election memorandum and "pink slip," was a violation of § 8(a)(1) and that Be-Lo had otherwise ignored its no-solicitation policy with _________________________________________________________________ 1 The Supreme Court has recently reiterated that, because Congress has "delegate[d] to the Board the primary responsibility for making remedial decisions," its "consistent appraisal of the Board's remedial authority" recognizes that the "Board's views merit the greatest deference." ABF Freight Sys., Inc. v. NLRB, 510 U.S. 317, 323-24 (1994). 2 Because of the manner in which the ALJ used his language, I have not counted at least six additional § 8(a)(1) violations that he may actually have found. 37 respect to non-Union-related solicitors and thus violated § 8(a)(1) by denying the picketers access to Be-Lo's property. The Board thus had found two more § 8(a)(1) violations than the ALJ had, see J.A. at 340- 41, 348-50, both of which the majority, in turn, reverses. Also among the total of 77 are the Board's adoption of the ALJ's findings that Be- Lo violated § 8(a)(1) by maintaining the state court injunction actions after the General Counsel had issued complaints alleging that Be-Lo's denial of access to the Union was unlawful and that Be-Lo violated § 8(a)(5) by failing to bargain with the Union after March 20, 1991, when the Union represented a majority of unit employees. The major- ity also reverses these. In addition, the Board adopted the ALJ's find- ings of 73 additional violations of § 8(a)(1) and (3) by, inter alia, threatening employees with store closure, interrogating employees, threatening employees with discharge, threatening reductions in working hours, threatening layoffs, and discharging several employ- ees. See J.A. at 340, 360-74. My tally does not include the Board's reversal of the ALJ's findings that two employees, Tom DeYarmon and Colleen Hitt, had been discharged in violation of § 8(a)(3). These 73 remaining violations are comprised of 59 § 8(a)(1) violations and 14 § 8(a)(3) violations against individual employees.3 Of the § 8(a)(3) violations, Be-Lo concedes three of these, the majority upholds three others, see slip op. at 28, 30-31, and neither Be-Lo nor the majority discusses yet three more. The majority quibbles over whether several of the § 8(a)(1) viola- tions that the Board found were violations at all. See slip op. at 16-17. For example, the ALJ found four § 8(a)(1) violations when coman- ager Perras met with employee Keyes, a known union supporter, shortly before the election: Perras showed Keyes some documents regarding the union _________________________________________________________________ 3 These 59 § 8(a)(1) violations were found to occur at 18 different stores as follows: Store 28 (3); Store 37 (8); Store 62 (5); Store 73 (1); Store 96 (3); Store 102 (4); Store 107 (1); Store 110 (1); Store 121 (2); Store 122 (2); Store 126 (6); Store 144 (2); Store 145 (3); Store 148 (5); Store 232 (2); Store 234 (7); Store 235 (1); Store 236 (3). The 14 § 8(a)(3) violations were found to occur at eight different stores as fol- lows: Store 37 (1); Store 102 (1); Store 110 (1); Store 126 (3); Store 145 (1); Store 148 (2); Store 232 (2); Store 235 (2). 38 campaign and after Keyes read the documents Perras asked Keyes what she (Keyes) thought about the Union. Perras went on to tell Keyes that if the Union got in and wages were raised some stores might close and people lose their jobs or hours would be cut. In other words Perras unlawfully interrogated Keyes and threatened store closure, loss of jobs, and reduced hours for employees if the Union was selected all in violation of Section 8(a)(1) of the Act. J.A. at 363. Although I believe these are among the weaker of the vio- lations the Board upheld, the ALJ specifically credited Keyes over Perras. When this episode is viewed in light of Be-Lo's vehement anti-union campaign, I do not think that this court can say that sub- stantial evidence in the whole record does not support these findings, "even though we might have reached a different result had we heard the evidence in the first instance." NLRB v. General Wood Preserving Co., 905 F.2d 803, 810 (4th Cir.) (internal quotation marks and cita- tion omitted), cert. denied, 498 U.S. 1016 (1990). Similarly, the majority finds suspect an episode of potential sur- veillance: In January 1991 [employee] Cottrell was approached by Union Organizer Karen Gompers and shortly thereafter the comanager of store 102, David Rodriguez, told Cottrell he had seen her with the union organizer. This statement by Rodriquez to Cottrell creates the impression that Cottrell's union activity was under surveillance and being monitored by management in violation of Section 8(a)(1) of the Act. Rodriguez never specifically denied this took place. J.A. at 365. However, we have previously held that "[i]t is an unfair labor practice for an employer to create in the minds of employees an impression that he is closely observing union organizational activity." J.P. Stevens & Co. v. NLRB, 638 F.2d 676, 683 (4th Cir. 1980); see also NLRB v. Nueva Eng'g, Inc., 761 F.2d 961, 967 (4th Cir. 1985). Again, it is difficult to see how this finding is not supported by sub- stantial evidence in the record as a whole. In neither of these episodes does the majority actually reverse the Board's finding of violations. 39 Essentially, then, there are 59 § 8(a)(1) and 9 § 8(a)(3) violations that this court must accept as substantiated. On all but three of the § 8(a)(3) violations, Be-Lo did not even brief the court or make any argument. Be-Lo belatedly urged us that it had not waived its right to contest these violations, claiming that "the page limitations and the overwhelming size of the appendix prevented a complete discussion of all of the alleged violations of 8(a)(1)." Reply Br. of Pet'r at 13. Be-Lo, however, did not request from us a page limit extension, nor did it point out this waiver issue in its opening brief. Moreover, we have never shied away from examining thoroughly the record evi- dence before us, no matter how voluminous. See, e.g., Fieldcrest Can- non, Inc. v. NLRB, 97 F.3d 65, 69 (4th Cir. 1996) (dealing with more than 7800 transcript pages, more than twice that presented to us in the instant case). Although we have not previously dealt with this precise situation,4 it is nevertheless well-settled that where an employer offers "no basis or argument for overturning the Board's findings and con- clusions with regard to the §§ 8(a)(1) and (3) violations . . . the employer has effectively waived the right to object to these findings as erroneous." NLRB v. Horizon Air Servs., Inc., 761 F.2d 22, 26 (1st Cir. 1985) (citing case law from four other circuits). Having examined the material, I believe there is substantial evidence, when viewing the record as a whole, to support the Board's findings as to these unbrie- fed violations. The remaining violations, therefore, must be viewed against this backdrop of a company hell-bent on preventing unioniza- tion at nearly any cost. III. I turn now to the "pink slip" episode, reversed by the majority, slip op. at 26-28. The Board, reversing the ALJ, found that Be-Lo had vio- lated § 8(a)(1) by sending all bargaining unit employees a memoran- dum letter and accompanying "pink slip" shortly before the election that conveyed the message that stores would close and jobs would be _________________________________________________________________ 4 The case the Board relies on, NLRB v. Daniel Constr. Co., 731 F.2d 191, 198 (4th Cir. 1984), is inapposite, for there, unlike here, the employer failed altogether to object initially to the Board in accordance with 29 U.S.C. § 160(e) and the Board's rules pursuant to 29 C.F.R. § 102.46(b) (1983). 40 lost if the Union won the election. See J.A. at 341. The letter stated in its entirety: Dear Be-Lo Associate, For several months you have been hearing sweet- sounding promises from the out-of-town union organizers, about higher wages, job security, grievance procedures, seniority rights and other expensive benefits which are easy to promise but hard to deliver. At the same time they have also fed you militant anti- BeLo distortions charging this company--your company-- with "half truths, lies and propaganda". Half-truths? Lies? Propaganda? When you're trying to figure out who is giving you the facts and who is promising you the moon, you may want to take another look at what the UNITED FOOD AND COMMERCIAL WORKERS UNION got for their former dues payers in this area--A PINK SLIP. We will do everything we can to prevent this from hap- pening to you, whether or not the union is voted in, but you should consider what the union has DELIVERED in the past before you believe what they PROMISE in the future. (See the attachment) J.A. at 3238 (emphases in original). Attached to the letter was a pink piece of paper with the logos of grocery store chains Big Star Foods, Safeway, and Colonial Stores at the top that read: Dear Unionized Employees, I regret to inform you that because we have lost our abil- ity to compete in this extremely competitive market, we shall be forced to close this store and put you out of work. 41 Sincerely, Colonial Stores/Big Star Safeway J.A. at 3239. The ALJ concluded that these materials, like other campaign mate- rials distributed by Be-Lo, constituted free speech and legitimate pro- paganda under § 8(c) of the Act, 29 U.S.C. § 158(c), because they did "not rise to the level of a threat of reprisal if the employees select[ed] the union as their collective bargaining representative." J.A. at 279, 360. The Board disagreed, reasoning that the employees knew it had been Be-Lo that created and sent the "pink slip"; that the employees could reasonably infer that they would be the "unionized employees" who would suffer store closure and job loss if they voted for the Union; and, most importantly, given the numerous other § 8(a)(1) vio- lations in which Be-Lo did directly threaten employees with store clo- sure and job loss, that these employees were, in fact, threatened that they would get a "pink slip" if they voted for the Union. See J.A. at 341. A threat of store closure if employees select union representation is a § 8(a)(1) violation. See NLRB v. Nueva Eng'g, Inc., 761 F.2d 961, 966 (4th Cir. 1985). The Supreme Court has balanced the free speech rights of employers, recognized in 29 U.S.C. § 158(c), with the eco- nomic dependence of employees who are susceptible to credit readily an employer's implications: Thus, an employer is free to communicate to his employees any of his general views about unionism or any of his spe- cific views about a particular union, so long as the commu- nications do not contain a "threat of reprisal or force or promise of benefit." He may even make a prediction as to the precise effects he believes unionization will have on his company. In such a case, however, the prediction must be carefully phrased on the basis of objective fact to convey an employer's belief as to demonstrably probable conse- quences beyond his control or to convey a management decision already arrived at to close the plant in case of unionization. If there is any implication that any employer 42 may or may not take action solely on his own initiative for reasons unrelated to economic necessities and known only to him, the statement is no longer a reasonable prediction based on available facts but a threat of retaliation based on misrepresentation and coercion. NLRB v. Gissel Packing Co., 395 U.S. 575, 618 (1969) (emphases added) (citation omitted). Be-Lo argues that these materials must be seen in the context of the entire election campaign. In particular, the letter and "pink slip" were in direct response to Union promises of job security and a claim that any attempted refutation by Be-Lo would be a "half-truth, lie and pro- paganda." Reply Br. of Pet'r at 15. It may well be that one could rea- sonably argue that the language in the letter stating that "[w]e will do everything we can to prevent this from happening to you, whether or not the union is voted in" saves Be-Lo in this instance. However, a reasonable mind could nevertheless conclude, see NLRB v. General Wood Preserving Co., 905 F.2d 803, 810 (4th Cir.) (noting that "[s]ubstantial evidence has been held to mean such relevant evidence as a reasonable mind might accept as adequate to support a conclu- sion" (internal quotation marks and citation omitted)), cert. denied, 498 U.S. 1016 (1990), as the Board did here, that Be-Lo failed to accompany this communication with any objective substantiation of the precise effects or demonstrably probable consequences of union- ization. Furthermore, a reasonable mind could conclude that, given the substantial number of times that Be-Lo did, in fact, directly threaten store closure and job loss, this was nothing but a thinly- veiled threat of retaliation. Since it is within the special expertise of the Board to adjudge whether coercive language has been used, see Gissel, 395 U.S. at 620; Nueva Eng'g, 761 F.2d at 966, the court ought to have affirmed the Board's finding that the letter and "pink slip" amount to a violation of § 8(a)(1). The majority totally fails to link its reversal to the standard of review or to recognize the Board's special expertise in this area. By substituting its own opinion for that of the Board's expertise, the majority engages in a de novo review that contravenes our well- established precedent. See supra part I. This I cannot accept. 43 IV. The Board found that Be-Lo had discriminatorily enforced its no- solicitation policy in violation of § 8(a)(1) and therefore ordered Be- Lo to reimburse the Union for certain legal expenses incurred in con- nection with the state injunction cases. The majority reverses this finding and order. See slip op. at 23-25. I cannot agree. A. As a general rule, an employer cannot be compelled to allow non- employee union agents to solicit on its property. Lechmere, Inc. v. NLRB, 502 U.S. 527, 533, 537 (1992); NLRB v. Babcock & Wilcox Co., 351 U.S. 105, 112 (1956). However, there are two exceptions to this rule, an "inaccessibility" exception and a "discrimination" excep- tion, of which only the latter is implicated here. An employer may prohibit nonemployee distribution of union literature if it "does not discriminate against the union by allowing other distribution." Babcock & Wilcox, 351 U.S. at 112; see also Sears, Roebuck & Co. v. San Diego County Dist. Council of Carpenters, 436 U.S. 180, 205 (1978). Lechmere is not to the contrary, being principally concerned with reaffirming the general rule and rejecting the Board's attempt to recast it as a multifactor balancing test even in cases of nonemployee activities. See Lechmere, 502 U.S. at 535-38. The majority suggests, but does not hold, that nonemployee picket- ers protesting unfair labor practices would fail the Babcock & Wilcox discrimination exception following Lechmere. The majority relies on Cleveland Real Estate Partners v. NLRB, 95 F.3d 457 (6th Cir. 1996), see slip op. at 24, which in turn criticized the Board's decision in this case, Be-Lo Stores, 318 N.L.R.B. 1 (1995). Although the Cleveland court at least emphasized that its holding only applied to "nonem- ployee" handbillers, see Cleveland, 95 F.3d at 464, both that court and the majority here totally ignore the fact that the ALJ, and the Board by adoption, specifically noted that the picketers in this case included "employee[s] or former employees some of whom had been dis- charged in violation of Section 8(a)(3) of the Act." J.A. at 374. Pursu- ant to 29 U.S.C. § 152(3), an unlawfully discharged worker retains her status as an "employee." Thus the Board in this case found that Be-Lo had violated the Act not just by refusing access to nonem- 44 ployee Union organizers but by also denying access to employee picketers. There is little doubt that Supreme Court precedent recog- nizes that picketing by employees to protest unfair labor practices is a § 7 right. See Davis Supermarkets, Inc. v. NLRB, 2 F.3d 1162, 1177 (D.C. Cir. 1993), cert. denied, 511 U.S. 1003 (1994). Indeed, it is a right that can trump an employer's private property rights. See Republic Aviation Corp. v. NLRB, 324 U.S. 793, 803-04 (1945); see also Babcock & Wilcox, 351 U.S. at 109-11. Here, the Board, in reversing the ALJ, found that Be-Lo had other- wise ignored its no-solicitation policy with respect to non-Union- related solicitors and thus violated § 8(a)(1) by denying the Union's picketers, including employees, access to Be-Lo's property. This case is further complicated by the fact that employee picketers were never asked to leave Be-Lo's property unless non-employee Union organiz- ers were present as well. Nevertheless, the Board made it clear that employees could not lose their § 7 rights merely by associating with non-employees in these circumstances. See J.A. at 348 n.25. More- over, at least one court has upheld a § 8(a)(1) violation in similar cir- cumstances where six of twelve picketers were employees or unlawfully discharged employees. See Davis, 2 F.3d at 1177-78. The majority not only ignores the presence of employee picketers but mischaracterizes the evidence when it asserts that any solicitation that occurred did so "without the knowledge or approval of Be-Lo's owners, executives, or store-level management." Slip op. at 25. Not only did the comanager of Store 102 know of the cookbook salesman, who solicited inside the store, but he spoke with him and affirma- tively accepted a cookbook sample so that other employees could look at it and determine if they wanted to purchase one. See J.A. at 551. The majority also mischaracterizes the "few solicitations" as occurring over a period of a year and a half, see slip op. at 25, when the Board expressly refused to consider Union-proffered evidence of additional incidents outside the few months of the spring and summer of 1991, believing it was unnecessary to consider that additional evi- dence to reach its conclusion, see J.A. at 348 n.29. The Board was disturbed by the disparate enforcement of the no- solicitation rule, allowing sundry solicitations precisely at the same time that Be-Lo "consistently and rigorously enforced its no- 45 solicitation rule on a corporatewide basis" against the picketers and handbillers. J.A. at 349. The Board ultimately concluded that Be-Lo "took a laissez-faire approach to the religious and political groups, while ejecting the union protesters through lawsuits and threats of arrest," and thereby "follow[ed] a policy of `discriminat[ing] against the [U]nion' within the meaning of" Babcock & Wilcox. J.A. at 349 (quoting Babcock & Wilcox, 351 U.S. at 112). As I must reiterate again, although "we might have reached a different result had we heard the evidence in the first instance," NLRB v. General Wood Pre- serving Co., 905 F.2d 803, 810 (4th Cir.) (internal quotation marks and citation omitted), cert. denied, 498 U.S. 1016 (1990), substantial evidence does indeed support the Board's finding that "antiunion dis- criminatory motives" animated Be-Lo in "ejecting union agents while tolerating intermittent, but not infrequent, solicitations on its property by purveyors of incense and religious literature." J.A. at 349. See also Riesbeck Food Markets, Inc. v. NLRB, 91 F.3d 132, 1996 WL 405224, at **4 (4th Cir. July 19, 1996) (unpublished disposition) (cit- ing favorably the Board's opinion in the instant case, 318 N.L.R.B. No. 1 (1995), that the prohibition on union distribution was discrimi- natory). I would sustain the Board's conclusion that Be-Lo violated § 8(a)(1) by disparately enforcing its no-solicitation rule. B. Because the majority reverses the Board's conclusion on the dispa- rate enforcement of its no-solicitation rule, it also reverses the Board's order awarding the Union certain attorneys' fees and costs in defending the various state court injunction actions. See slip op. at 25 n.9. Because I would find a violation, I briefly discuss the Board's decision to apply Loehmann's Plaza [I], 305 N.L.R.B. 663, 1991 WL 251696 (1991), decision supplemented by Loehmann's Plaza [II], 316 N.L.R.B. 109 (1995), review denied sub nom. United Food and Com- mercial Workers, Local No. 880 v. NLRB, 74 F.3d 292 (D.C. Cir.), cert. denied sub nom. International Bhd. of Teamsters, Local No. 243 v. NLRB, 117 S. Ct. 52 (1996), retroactively to require Be-Lo to reim- burse the Union for legal expenses from the time the General Counsel issued its complaint in September 1991 until Be-Lo moved to stay the injunctions in December 1991. Great deference is normally given to the Board's judgment on retroactivity: unless manifest injustice can be shown, the decision 46 should be upheld. Oakwood Hosp. v. NLRB, 983 F.2d 698, 703 (6th Cir. 1993); but cf. ARA Servs., Inc. v. NLRB, 71 F.3d 129, 135 (4th Cir. 1995) (determining that denial of enforcement, rather than remand, is appropriate where retroactive application of a new jurisdic- tional rule would work a "manifest injustice"). The Board determined there was no manifest injustice in this case, relying on its reasoning in Loehmann's Plaza I. See J.A. at 350 n.30. In Loehmann's Plaza I, the Board held, in a part not reversed by Loehmann's Plaza II, that "once the General Counsel decides to initiate a formal adjudicatory proceeding, the Board's jurisdiction is invoked and it becomes the exclusive forum for an adjudication of a[n employer's] property rights." Loehmann's Plaza I, 1991 WL 251696 at *12. Therefore, whenever there is an outstanding injunction when the General Coun- sel issues a complaint, the employer must take affirmative action within seven days of the issuance of the complaint to have the injunc- tion withdrawn. Id. at *13. The Board went on to apply that holding against the parties before it, declaring: [W]e cannot conclude that, at the time the conduct at issue in this case occurred, union charging parties had settled expectations that state court lawsuits to enjoin protected trespassory activities would be found to be unfair labor practices from their inception and would call for make- whole remedies extending throughout the pendency of the suit. Neither can we conclude that employer/property own- ers have settled expectations that they could prosecute such lawsuits with impunity. There is, therefore, nothing mani- festly unjust in applying to the parties [these] rulings . . . . Id. In the instant case, the General Counsel issued the first complaint on September 12, 1991, Loehmann's Plaza itself was decided on November 21, the Board's Regional Director notified Be-Lo of the Loehmann's Plaza decision on December 3, and Be-Lo moved to stay the injunction on December 9. Be-Lo is thus in no different a position than the parties in Loehmann's Plaza to whom the rule was applied retroactively. Other than state in a conclusory fashion that retroactive application would be manifestly unjust, see Br. of Pet'r at 33, Be-Lo has made no affirmative showing that it would be unjustly harmed by 47 the Board's order. I believe we should therefore enforce that part of the order. Although perhaps a little unfair given these circumstances, it is not manifestly unjust. Be-Lo has made at best only a cursory, half-hearted effort to show manifest injustice and, therefore, has not really carried its burden to convince me otherwise. V. The majority, either expressly or implicitly, upholds the Board's determination of nine § 8(a)(3) violations. The majority, however, rejects the Board's conclusion that five individuals were either dis- charged or constructively discharged unlawfully. For the following reasons, I must disagree in each case. There is no question that Erwin Hatchett was initially laid off for a "supportable cause," a decline in business due to the Persian Gulf War. It was Be-Lo's failure to recall Hatchett when openings later arose, however, that the Board determined to be a § 8(a)(3) violation. After his layoff, Hatchett became an active picketer, a circumstance of which Be-Lo was aware. Be-Lo claimed, and the majority accepts, that Hatchett was not called back because there were no openings for meat cutters in the four stores in Hatchett's immediate area. See slip op. at 29. The Board expressly rejected Be-Lo's claim, finding instead that Be-Lo had implicitly conceded that there were job openings at other stores to which Hatchett could have been called back. The Board's conclusion is logically and legally sound: "[Be-Lo] lawfully cannot refuse to recall a laid-off employee because of his support for the Union and then seek to justify the failure to recall by limiting the employee's right to recall to stores at which there are no openings." J.A. at 346. The majority relies on dicta in Appletree Chevrolet I to the effect that the discharge of even a union activist requires a "plus" if the employer has articulated a "supportable cause" for the dis- charge. See NLRB v. Appletree Chevrolet, Inc., 608 F.2d 988, 994 n.5 (4th Cir. 1979). The majority conveniently ignores, however, the lan- guage in the text of the opinion which emphatically recognizes that where employees are particularly active in the union, an inference of improper motivation is not only warranted but may serve as just such a "plus factor." See id. at 993-94 (remarking that there are "an infinite number of precedents" that sustain a discriminatory discharge in such circumstances). The Board determined that Be-Lo's reason for not 48 recalling Hatchett, no openings, was pretextual and that Hatchett was, in fact, not recalled because of his union activities. Substantial evi- dence supports the Board's conclusion that Hatchett was construc- tively discharged in violation of § 8(a)(3). The majority shortchanges the evidence the Board relied upon in finding Shirley Terry's discharge unlawful. Terry, a five-year employee, was very prounion: her picture was distributed on a union flyer and she joined the Union's picket lines in her off-hours while still employed. Be-Lo was well aware of her activity. The Board determined that Be-Lo harassed her by subjecting her to heightened scrutiny from the time her picture appeared in the flyer in January 1991, waiting for a misstep. The Board rejected Be-Lo's proffered justification for her May 6 discharge, that she gave away food to a fel- low co-worker, as pretextual. In fact, Terry did give employee Gard- ner some food scraps for his dog that she otherwise would have thrown away and marked them as "no charge." The Board found that Be-Lo escalated its investigation of Terry, interviewing her in a closed car and coaxing her into admitting that she had discounted food in the past. The ALJ and Board, however, credited her testimony that she had discounted food in the past only with her manager's approval, an exoneration that Be-Lo refused to countenance. The Board concluded that, on the basis of these circumstances and Be- Lo's repeated harassment of Terry, Be-Lo "was out to get Terry because of her support for the Union and seized upon this opportunity to terminate her." Courts have found similar discharges pretextual, see, e.g., McLane/Western, Inc. v. NLRB, 827 F.2d 1423, 1425 (10th Cir. 1987) (upholding a § 8(a)(3) violation where a pro-union grocery warehouse employee was discharged for eating a broken cracker from an unsalvable case in alleged violation of the company's anti-pilfering rule), and have looked askance at discriminatorily-motivated investi- gations that would not have occurred but for the employer's desire to get rid of a pro-union employee, see, e.g., Signal Oil and Gas Co. v. NLRB, 390 F.2d 338, 340-42 (9th Cir. 1968) (upholding a § 8(a)(3) violation where an employer discharged an employee who made a pro-strike remark, allegedly for "poor attitude," but only after delving into the employee's record after being prompted by the remark). See also American Thread Co. v. NLRB, 631 F.2d 316, 321-23 (4th Cir. 1980) (concluding that employee would not have been discharged, although he may have been subjected to lesser punishment, but for his 49 union activity). Similarly, substantial evidence supports the Board's finding of a § 8(a)(3) violation in this case. Pamela Jackson also picketed the company on her off-duty hours, an activity of which Be-Lo was well aware. Jackson took a sick day, apparently to visit her brother who had recently been shot, when the store manager refused to give her the day off. The district manager subsequently refused to let her return to work without a doctor's excuse. The Board found, and Be-Lo even conceded before it, that Be-Lo's established policy was to treat such an absence as unexcused and issue a written warning for the infraction, not to discharge the employee. The ALJ noted that Jackson had specifically told her dis- trict manager that she had not seen a doctor. See J.A. at 373. The Board therefore concluded that when Be-Lo set the production of a doctor's note as the condition for her return, a condition it knew she could not meet, Be-Lo violated its own admitted sickness excuse pol- icy in retaliation for her protected union activities. See J.A. at 347. Seen in the context of Be-Lo's hostility toward picketing and a rash of other discharges of known union activists, substantial evidence supports the Board's conclusion that Jackson was constructively dis- charged in violation of § 8(a)(3). Gwen Andrews, a known union supporter, was constructively dis- charged following a minor dispute with her supervisor. The majority mischaracterizes the evidence when it asserts that Be-Lo presented "unrebutted evidence" that Andrews was "expelled" for refusing to follow her "pro-union" supervisor's instructions. See slip op. at 29-30. The Board expressly determined that there was no evidence that Andrew's supervisor was pro-union at the time of her constructive discharge. Instead, the evidence suggests that Andrew's supervisor had become firmly allied with management, as his presence during an unlawful interrogation of Andrews reflects. See J.A. at 346. More- over, it is difficult to see how the evidence against Andrews stood "unrebutted" when only Andrews testified, not her supervisor. Andrews testified that while she was wrapping cornish game hens, her supervisor said "I need you to price and weigh chicken legs." She replied "okay" and attempted to finish wrapping the hens. Her super- visor repeated his request twice more, and when Andrews said she had heard him the first two times, he ran to Andrews, got in her face, and asked her "what the hell" her problem was. He then told her to 50 go home and not to come back. See J.A. at 346, 371. The ALJ specifi- cally credited Andrews' testimony over that of two company officials. The Board determined that the timing of this minor dispute on March 5, 1991, following a February incident in which Andrews was unlaw- fully interrogated and threatened, and just two weeks before the elec- tion, turned out to be a convenient way for Be-Lo to rid itself of a pro- union employee. See J.A. at 346-47. Substantial evidence supports the conclusion that Andrews was unlawfully constructively discharged when a Be-Lo supervisor told her not to come back. The majority also mischaracterizes the evidence when it claims Michael Salazar was fired because of the "unrebutted" evidence of his "pervasive absenteeism." See slip op. at 30. Even Be-Lo does not claim Salazar was pervasively absent, only that Salazar "was late for work many times and was late for work by more than an hour on sev- eral occasions." Br. of Pet'r at 36. The Board, adopting the ALJ's findings, noted that while Salazar's "attendance record was less than outstanding," Be-Lo's reason for his discharge under the circum- stances was pretextual pursuant to a Wright Line analysis. Those cir- cumstances were that Salazar was mugged and stabbed on the way home from work on March 8, 1991. Nevertheless, Salazar attempted to return to work on March 9, but the store manager released him to go home when he suffered dizziness. Salazar, who possessed no med- ical insurance, was too sick to return to work on March 10. The Board found that Salazar, who did not own a phone, went to a store near his home and attempted to call his manager to let him know he could not come in that day, but the phone was out of order. When Salazar did return to work on March 11, his manager fired him for his absence the day before, despite knowing of the stabbing and sending him home on March 9. The Board concluded that Be-Lo, which was "un- alterably opposed" to unionization, "seized on this opportunity--just 10 days before the election--to fire this prounion employee." J.A. at 362. Indeed, the Board had found that this manager had previously unlawfully interrogated Salazar and that Salazar was known to be a union supporter who had solicited other employees to sign authoriza- tion cards in the presence of Be-Lo officials. Substantial evidence on the record as a whole supports the Board's conclusion that Salazar would not have been discharged for his March 10 absence but for his prounion position. 51 VI. A. In NLRB v. Gissel Packing Co., 395 U.S. 575 (1969), the Supreme Court upheld the authority of the Board to issue a bargaining order in appropriate circumstances. See Gissel, 395 U.S. at 610-16. The Court there recognized a tripartite categorization of unfair labor prac- tices. The first category comprises those instances where the unfair labor practices are "outrageous" and "pervasive" such that a bargain- ing order is the only effective remedy. There is no need to inquire into majority status in such cases. See id. at 613-14. The second category comprises "less extraordinary cases marked by less pervasive prac- tices which nonetheless still have a tendency to undermine majority strength and impede the election process." Id. at 614. The third cate- gory, which will not sustain a bargaining order, comprises those minor and less extensive unfair labor practices that have a minimal impact on electioneering. See id. at 615. For the second category of cases, the Court discussed a number of factors the Board should consider in determining the propriety of a bargaining order. See id. at 614-15. We have previously stated those criteria as depending on the following Board findings: (1) that the Union once had majority status, (2) that such sta- tus had been dissipated by pervasive misconduct on the part of the employer, (3) "that the possibility of erasing the effects of [these] past [pervasive] practices and ensuring a fair election (or a fair rerun) by the use of traditional reme- dies, though present, is slight" and (4) "that employee senti- ment once expressed through cards would, on balance, be better protected by a bargaining order." NLRB v. Appletree Chevrolet, Inc., 608 F.2d 988, 996 (4th Cir. 1979) (Appletree Chevrolet I) (quoting Gissel, 395 U.S. at 614-15); see also NLRB v. So-Lo Foods, Inc., 985 F.2d 123, 126 (4th Cir. 1992). In Appletree Chevrolet I we also required the Board, in this second cate- gory, not just to find the existence of unfair labor practices but "to make specific findings as to the immediate and residual impact of the unfair labor practices on the election process"; in other words, the 52 Board "must make a detailed analysis assessing the possibility of holding a fair election in terms of any continuing effect of miscon- duct, and the potential effectiveness of ordinary remedies." Appletree Chevrolet I, 608 F.2d at 997 (internal quotation marks and citations omitted). The Gissel Court, however, also noted that the Board "draws on a fund of knowledge and expertise all its own" in fashion- ing remedies and thus that reviewing courts must give special respect to its choice of remedy. Gissel, 395 U.S. at 612 n.32, 616. We have therefore noted that we review the selection of remedy for abuse of discretion. See NLRB v. Williams Enters., 50 F.3d 1280, 1289 (4th Cir. 1995). There is thus a tension between Appletree Chevrolet I's requirements to sustain a bargaining order and the Board's broad dis- cretion in initially choosing it that we have yet to resolve and that the majority ignores altogether. The threat of a store closure is a "hallmark" violation that, in itself --absent significant mitigating circumstances evidencing that the vio- lative conduct is not as serious as it might seem--can support a Gissel order. See So-Lo Foods, 985 F.2d at 126-27 & 127 n.5. Hallmark vio- lations, which also include threats of job loss and discharge of union adherents in violation of § 8(a)(3), fall within Gissel's first category. NLRB v. General Wood Preserving Co., 905 F.2d 803, 822 (4th Cir.) (citing NLRB v. Jamaica Towing, Inc., 632 F.2d 208, 212-13 (2d Cir. 1980)), cert. denied, 498 U.S. 1016 (1990). Therefore, they do not require a showing of majority support for the union. The ALJ declined to determine whether the bargaining order he issued fell into the first or second of Gissel's categories, for it was "clear" that "the Union enjoyed majority support, the unfair labor practices committed by Be-Lo were `hallmark' violations, the effect of these unfair labor practices rendered a fair rerun election impossi- ble, and the interests of justice cry out for the issuance of a bargaining order." J.A. at 327, 380 (emphases added). The Board agreed that the facts of this case supported a bargaining order even under the more stringent requirements of Gissel's second category and thus also refused to classify the bargaining order it issued as "Category I" or "Category II." See J.A. at 350 & n.32. Although the Board expressly refused to classify its bargaining order, the majority "deems" it to be a Category II order, and, on that 53 basis, finding the Board's analysis erroneous or lacking, refuses to sustain it. See slip op. at 6. The majority never considers whether the order is sustainable as a Category I bargaining order. I believe the majority's analysis is fundamentally flawed in several respects. First, there is no warrant for subjecting the Board's order in this case only to the harsher requirements of a Category II analysis. Second, the facts of this case plainly support the issuance of a Category I bargain- ing order. Finally, even under the more stringent burden necessary to sustain a Category II order, the Board, as evidenced by its own "de- tailed analysis" of the matter, did not abuse its discretion. B. The majority cites Appletree Chevrolet I as settled law that where the Board declines to specify whether its bargaining order falls into the first or second of Gissel's categories, yet conducts only a Cate- gory II analysis, then we "deem the order to have been a Gissel Cate- gory II order." Slip op. at 6 (citing Appletree Chevrolet I, 608 F.2d at 996 n.8). In that case, however, we independently concluded that "this is the proper categorization for this case." Appletree Chevrolet I, 608 F.2d at 996 n.8. That such a categorization was warranted was evident from the facts that that case involved only four § 8(a)(3) vio- lations arising from the discharge of four employees and three addi- tional § 8(a)(1) violations and that we noted that the Board apparently had misgivings about the propriety of a bargaining order in any event. See id. at 991, 1001. Moreover, although somewhat unclear, it appears we rejected all the violations except one § 8(a)(1) violation of unlawful interrogation. See id. at 996, 998-99. In the instant case, even the majority sustains nine § 8(a)(3) viola- tions for unlawful discharge, six expressly and three implicitly, as well as dozens of § 8(a)(1) violations. It is thus not self-evident that this case is necessarily a Category II case. Moreover, judicial bodies routinely refuse to expend resources categorizing the nature of claims or remedies when the evidence or analysis is deemed to satisfy the strictest requirements or greatest burden. Indeed, courts do not even classify Gissel bargaining orders as Category I or II when a case clearly satisfies the greater burden of Category II. See, e.g., Davis Supermarkets, Inc. v. NLRB, 2 F.3d 1162, 1171 (D.C. Cir. 1993), cert. denied, 511 U.S. 1003 (1994); NLRB v. Horizon Air Servs., Inc., 761 54 F.2d 22, 29 & n.5 (1st Cir. 1985). In fact, the Board specifically relied on Davis Supermarkets when it decided that it was unnecessary to specify whether the instant case fell into Category I or II. See J.A. at 350 n.32. The majority therefore erred when it "deemed" the bargaining order in this case to be a Category II order and "scrutinized" it only on those terms. C. The facts of this case can plainly sustain a Gissel Category I bar- gaining order. I would affirm the Board's finding of at least 77 viola- tions. The "pink slip" episode and the various employee discharges are all hallmark violations, as are at least 50 other of the § 8(a)(1) vio- lations, as the Board found. See J.A. at 352. These numerous egre- gious violations evince that there are no significant circumstances mitigating the seriousness of Be-Lo's activities here and are more than sufficient to support the Board's issuance of a bargaining order under the first category. Moreover, there is certainly substantial evi- dence in the record to support the conclusion that the extent and nature of Be-Lo's unfair labor practices show them to be "outrageous" and "pervasive" within the meaning of Gissel and thus "crying out" for a bargaining order. Even under the majority's approach, there remain 59 § 8(a)(1) vio- lations, the vast majority hallmark violations, and 9 hallmark § 8(a)(3) discharges. Indeed, many of these § 8(a)(1) violations involve threats of store closure, which, standing alone, are sufficient to support a Cat- egory I bargaining order. See So-Lo Foods, 985 F.2d at 127 (citing Gissel, 395 U.S. at 615). To support a Gissel Category I bargaining order, the Board need not undertake a detailed analysis but "instead must only make `mini- mal findings' of the lasting effect of unfair labor practices." Power Inc. v. NLRB, 40 F.3d 409, 422 (D.C. Cir. 1994). In Power, the Dis- trict of Columbia Circuit upheld a Gissel Category I bargaining order where the employer permanently laid off thirteen employees and repeatedly threatened plant closure in the event of unionization. The 55 court held that the Board satisfied the minimal fact-finding require- ment to support a Category I order when it found Power's repeated, numerous, and persistent threats of plant closure, which "reached, directly, nearly every employee in the bargaining unit," "among the most serious, flagrant and long-lasting forms of interference with employee rights," acting "to destroy employee support for a union and the pos- sibility of a fair election." The Board also concluded that it is "difficult to imagine any act of management better calcu- lated to chill union support" than the layoff of thirteen union supporters just days before the election. Id. at 423 (citations omitted). In the instant case, because the Board thought its analysis was suf- ficient to support the more demanding standards of a Category II order, there should really be little question that it satisfied the minimal findings necessary for a Category I order. Without detailing that basis here,5 it is sufficient to note that the Board specifically determined that of the "serious" and "extensive" unfair labor practices, more than 30 hallmark violations were committed in the weeks just prior to the election; that Be-Lo discharged six union adherents in March 1991 on the eve of the election; that Be-Lo's "misconduct includes the type of severe and pervasive coercion that has lingering effects and is not readily dispelled by the passage of time"; that "management from top to bottom" was engaged in the antiunion campaign and committed the serious hallmark violations; that the pervasive effect of Be-Lo's mis- conduct precluded the "possibility of a fair rerun election"; that after the election Be-Lo took no action "to eradicate the effects of its unlawful threats of closure and discharge" and, in fact, continued its "extensive antiunion campaign"; and that "the cloud created by these violations is likely to linger and cannot be dispersed by a traditional cease-and-desist order." J.A. at 352-54. These substantial findings are more than adequate to support a Category I bargaining order. _________________________________________________________________ 5 I will treat that basis in more detail below when discussing the fact that the Board's analysis is sufficiently detailed to support even a Cate- gory II order. 56 D. Even under the stricter requirements necessary to support a Cate- gory II order, the bargaining order in the instant case was fully war- ranted. 1. Substantial evidence supports the Board's finding that the Union enjoyed majority support on March 20, 1991, for the Union offered into evidence 403 proper authorization cards out of 756 eligible employees as of that date. See J.A. at 351. The parties argue vigor- ously over whether the Board erred in relying on the ALJ's finding after the ALJ refused Be-Lo's proffer of certain rebuttal evidence. Be- Lo asserts that the Excelsior list it generated was improperly relied upon as an accurate roll of final eligible voters,6 and the majority inexplicably and erroneously accepts Be-Lo's sleight-of-hand argu- ments. See slip op. at 10-11. The Excelsior list in this case contained the names of 765 employees who were on the payroll as of January 26, 1991, as amended by the parties at a preelection conference on March 19, a number which the ALJ reduced by nine. Be-Lo contends that once these numbers are adjusted to reflect employees who were terminated before March 20 and new hires who were added after Jan- uary 26, then only 362 of 769 eligible employees had signed authori- zation cards, less than a majority. Be-Lo made this same argument to the Board, which properly rejected it. The ALJ and the Board both found that the General Coun- sel had reasonably relied in good faith on the Excelsior list to estab- lish a Gissel majority at trial. On the first day of the hearing before the ALJ, February 18, 1992, Be-Lo moved to quash the General Counsel's subpoena duces tecum that requested "[p]ayroll records reflecting all bargaining unit employees employed by Be-Lo for each pay period, from February 1, 1991 through April 1, 1991." See J.A. at 147 (subpoena duces tecum, item 27). The General Counsel explained that what it was seeking with its request "is a listing of all _________________________________________________________________ 6 An Excelsior list contains the names and addresses of employees and is furnished by the employer to the union for use in an election. See NLRB v. Wyman-Gordon Co., 394 U.S. 759, 761-62 (1969). 57 bargaining unit employees for each of those payroll periods. In other words, a listing like they provided to the Board for the [E]xcelsior list for the election, that's what we are talking about." J.A. at 438. Be-Lo in response stated: It seems to me that the Gissel[ ] remedy can only be as of the date that the union claims it has majority status, what- ever date that was, and it was certainly before the election. [. . .] Now, the election was March 21, and there were certainly some deletions from the [E]xcelsior list for those employees who were not on the payroll as of the date of the election. Those deletions were made and the Board has that list as of March 21. They have it. J.A. at 439. Based on these assurances that it had the proper final voter eligibility list, the General Counsel relied on the amended Excelsior list to establish the Gissel majority.7 Only after the General Counsel rested its case some five months into the hearing did Be-Lo first assert on July 8, 1992, that the Excelsior list was inaccurate and could not be relied upon to show majority support. See J.A. at 1113- 43, 1160-80. The majority concocts a fanciful explanation about the existence of two lists. See slip op. at 9-10. However, the majority's imagination does serious violence to the ordinary meanings of words and gram- matical structure. Be-Lo nowhere stated in the colloquy before the ALJ that there were two lists. Be-Lo's counsel talked only of "dele- tions from the [E]xcelsior list" in the first sentence of the second paragraph quoted above. The next sentence plainly and unambigu- ously states that "[t]hose deletions were made and the Board has that list." "That list" can only refer to the amended Excelsior list, the adjective "that" modifying "list" and designating the list mentioned in the previous sentence. The final sentence, "They have it." conclu- _________________________________________________________________ 7 The Board also noted that the Union as well relied on this Excelsior list, for it only called as witnesses card signers who appeared on it. See J.A. at 351 n.34. 58 sively establishes that Be-Lo was talking only about one thing, the "it" being the amended Excelsior list mentioned in the first sentence and referenced in the second. In any event, it is plain that the majority continuously wants to make itself into the fact-finder in this case. If there now appears to be confusion as to what Be-Lo meant, the ALJ evidently determined that it was clear enough at the time and that Be- Lo was assuring both the General Counsel and the ALJ himself that the General Counsel already possessed the amended Excelsior list. The Board also found, despite Be-Lo's claims to the contrary, that Be-Lo had not fully complied with the subpoena. The majority clearly mischaracterizes this matter. See slip op. at 8 n.1, 11. The General Counsel had requested payroll records; Be-Lo provided work sched- ules and time sheets. The Board judged this proffered evidence to be "seriously defective" because employees could be out of work during the relevant time period for various reasons, such as maternity leave, and therefore not listed on the work schedules and timesheets even though they would still be in [Be-Lo's] employ and eli- gible to vote. In addition, we observe that [Be-Lo] did not proffer work schedules and timesheets for all the stores in the bargaining unit during the relevant time period. J.A. at 351. I have examined this material and believe substantial evi- dence supports the Board's conclusion that Be-Lo's "belatedly prof- fered evidence is seriously flawed, and that the judge did not err in finding the `best evidence' of the number of employees in the bar- gaining unit as of March 20" to be the Excelsior list. J.A. at 351. In support of this conclusion, the Board cited Bannon Mills, Inc., 146 N.L.R.B. 611 (1964). Be-Lo argues that Bannon Mills stands only for the proposition that where the employer fails to respond to the General Counsel's subpoena can the ALJ properly exclude the employer's evidence based on that unprovided material. See Br. of Pet'r at 10, 12; Reply Br. of Pet'r at 8-9. In Bannon Mills, the employer refused to comply with the General Counsel's subpoena duces tecum requesting payroll records. Consequently, the General Counsel produced secondary evidence to prove majority support for the union. When the employer objected that this was not the "best evi- 59 dence," but still refused to produce that alleged "best evidence," the objection was overruled. After the General Counsel rested, the employer attempted to introduce those subpoenaed, but unproduced, payroll records. The ALJ ruled the records inadmissible, Bannon Mills, 146 N.L.R.B. at 633-34, and the Board affirmed, id. at 613 n.4. Be-Lo asserts that because it did comply with the subpoena, as the ALJ found, Bannon Mills is inapplicable. The Board recognized that the instant case is distinguishable from Bannon Mills because the ALJ had found that Be-Lo did comply. Nevertheless, the Board stated: [W]e find that the principle that Bannon Mills stands for, the protection of the integrity of the hearing process, is properly invoked here where the General Counsel relied in good faith on [Be-Lo's] affirmation on the first day of the hearing that the final voter eligibility list was, in effect, the best evidence of the number of employees in the bargaining unit prior to the election for the purpose of establishing a Gissel major- ity. J.A. at 352 n.35.; cf. NLRB v. American Art Indus., Inc., 415 F.2d 1223, 1229-30 (5th Cir. 1969) (upholding the Bannon Mills principle and recognizing that "to maintain the integrity of the hearing process" the ALJ properly refused to admit secondary evidence proffered by the employer which it refused to produce in the face of a valid subpoena duces tecum), cert. denied, 397 U.S. 990 (1970), and cert. denied, 401 U.S. 912 (1971). Be-Lo wants to make the ALJ's finding of compliance final. However, as discussed above, the Board found that Be-Lo's proffered evidence was "seriously defective" and incom- plete and thus that Be-Lo had not fully complied with the subpoena. While Be-Lo did not attempt to introduce evidence that it did not pro- duce, work schedules and time sheets are not payroll records. Be-Lo did not comply with the subpoena and the General Counsel relied on Be-Lo's assurances that the amended Excelsior list contained only those names of employees on the payroll as of the election date. I must also point out that while Be-Lo only wants to use its num- bers of employee terminations and new hires to show that the Union did not have a majority of authorization cards, the ALJ specifically noted that "during the union organizing campaign the Union secured 60 many more authorization cards than 403." J.A. at 312, 376. The Union had only placed in evidence cards from those whose names appeared on the amended Excelsior list. This crucial fact is totally ignored by the majority. The Union was therefore every bit as bound and hindered by the reliance on the amended Excelsior list as was Be- Lo. Be-Lo, in effect, wants to relitigate this matter, but all litigation must come to an end at some point. The ALJ determined that the amended Excelsior list was the "critical document," id. at 313 n.2, 376 n.2, and the Board agreed. I can see absolutely no basis for this court to determine that the Board or the ALJ abused their discretion in rejecting Be-Lo's other belatedly proffered evidence. 2. Be-Lo also contends that, even on the basis of the amended Excelsior list, the Union lacked majority support. These contentions are meritless. The Board properly refused to pass on Be-Lo's claims that certain cards should have been excluded because they were obtained by misrepresentation or by supervisors. Even excluding those cards, the Union still enjoyed the support of an uncoerced majority.8 See J.A. at 352. _________________________________________________________________ 8 Although Be-Lo claims before us, and the majority accepts, see slip op. at 15, that 15 cards were improperly obtained by supervisors, before the Board Be-Lo excepted to only 14 supervisor-obtained cards. See J.A. at 352 n.36. More significantly, the majority highlights the testimony of Kim How- ell, a union witness, for the purpose of showing that certain cards were obtained through misrepresentation. See slip op. at 12-14. Notably absent from the majority's discussion is the number of cards allegedly affected. The Board treated that number as only seven, see J.A. at 352 n.36, and therefore counted in that total not only Howell's own card, but also the "5 to 6 other employees" to whom she allegedly misrepresented their purpose. See Br. of Resp't at 41 n.10. Most importantly of all, however, it is apparent that the ALJ never counted any of the cards Howell obtained. The ALJ was careful to list all 37 organizers who put the cards he counted into evidence and the number of cards put into evidence by each. See J.A. at 377. Only from this list of 535 cards placed in evidence (the ALJ apparently made an arithmetical mistake by tallying only 534) did the ALJ consider whether 61 Be-Lo also asserts that 43 additional cards should be excluded because the signers of those cards testified that they were told the card would only be used to gain an election. As Gissel makes clear, employees should be bound by the clear language of what they sign unless that language is deliberately and clearly canceled by a union adherent with words calculated to direct the signer to disregard and forget the language above his signature. There is nothing inconsistent in handing an employee a card that says the signer authorizes the union to represent him and then telling him that the card will proba- bly be used first to get an election. Gissel, 395 U.S. at 606. In this case, it cannot be gainsaid that the authorization card language was self-evidently clear, stating: I hereby authorize United Food and Commercial Workers International Union, AFL-CIO, or its chartered Local Union to represent me for purposes of collective bargaining, respecting rates of pay, wages, hours of employment, or other conditions of employment, in accordance with applica- ble law. See, e.g., J.A. at 3146. The Board accepted the ALJ's credibility find- ings on this matter: The organizers were a group of men and women, some of whom were young and some not so young. Some of the organizers were white and some were black. I found all 37 organizers to be honest and competent individuals. Without exception they impressed me by their demeanor on the stand as individuals exceedingly worthy of belief. Not one of these organizers, I find, ever said to an employee they were _________________________________________________________________ the Union possessed majority support. Howell's name does not appear on this list. The careful questioning that the majority shows the ALJ engaged in with respect to Howell leads to the inescapable conclusion that the ALJ recognized the cards she obtained were tainted and therefore did not count them. 62 soliciting to sign a card that the signing of the union authori- zation card was for the sole purpose of setting an election or that an individual who signed a card was just indicating that he or she wanted an election. In virtually all instances the persons who signed the card were given the opportunity to read the card, were encouraged to do so, and appeared to do so. In those instances where the Union received the card from a "key worker," i.e. an employee in the store who was helping to organize the employees, one of the 37 organizers would verify the signature on the card and ask the employee if he or she understood the card and had any questions. [. . .] Not only don't I believe [Be-Lo's 43] witnesses . . . but I'm convinced that the fact that these otherwise intelligent, decent people would have come into a court room and tes- tify as they did is evidence that the unfair labor practices of Be-Lo have not been dissipated by the passage of time and that a fair rerun election is not possible in this case. Many of these employees, I conclude, were so terrified about their job security and staying on Be-Lo's good side that they were ready, willing, and able to help Be-Lo's case by saying that union organizers said things to them in soliciting them to sign authorization cards that were not said. Others were merely mistaken when they claim what the person soliciting the card said to them. A number of them claimed all they remembered being said was that an election could be held if enough employees signed a card and that they didn't read the card. J.A. at 377-78 (emphases in original) (footnote omitted); id. at 352 n.36. I find it simply incredible that the majority would merely cite the entire footnote in which the ALJ made specific credibility determinations,9 greatly excise the vast majority of the ALJ's textual discussion of the basis of those findings, and then condemn the ALJ for a "perfunctory footnote," for "stat[ing] conclusorily" his basis, and _________________________________________________________________ 9 Again, notably absent from the footnote is any mention of Kim How- ell, in whom the majority inexplicably invests so much. 63 for "provid[ing] no more than a generalized, conclusory statement." Slip op. at 14. As we, and most, if not all, of our sister circuits, have noted, "It is well settled that[,] absent exceptional circumstances, the ALJ's credi- bility findings, `when adopted by the Board[,] are to be accepted by the [reviewing] court.'" NLRB v. Air Prods. and Chems., Inc., 717 F.2d 141, 145 (4th Cir. 1983) (quoting Dubin-Haskell Lining Corp. v. NLRB, 375 F.2d 568, 571 (4th Cir.), cert. denied, 393 U.S. 824 (1967)). See, e.g., Exxel/Atmos, Inc. v. NLRB, 28 F.3d 1243, 1246 (D.C. Cir. 1994) ("[W]e must accept the ALJ's credibility determina- tions, as adopted by the Board, unless they are patently insupportable." (internal quotation marks and citation omitted) (emphasis added)); NLRB v. Harding Glass Co., 80 F.3d 7, 10 n.2 (1st Cir. 1996) (noting that "credibility determinations, of course, are for the Board rather than for us to make, and they stand unless beyond the `bounds of reason'" (internal quotation marks and citations omit- ted) (emphasis added)); Kinney Drugs, Inc. v. NLRB, 74 F.3d 1419, 1427 (2d Cir. 1996) (stating that Board findings, "based on the ALJ's assessment of the credibility of witnesses, . . . will not be overturned unless they are hopelessly incredible or they flatly contradict either the law of nature or undisputed documentary testimony" (internal quotation marks and citation omitted) (emphases added); NLRB v. Florida Medical Ctr., Inc., 576 F.2d 666, 671 (5th Cir. 1978) (stating that credibility findings will be upheld unless "self-contradictory" (emphasis added); Retlaw Broad. Co. v. NLRB, 53 F.3d 1002, 1006 (9th Cir. 1995) (stating that credibility determinations are upheld unless "inherently incredible or patently unreasonable" (internal quo- tation marks and citation omitted) (emphasis added)); Ready Mixed Concrete Co. v. NLRB, 81 F.3d 1546, 1551 (10th Cir. 1996) (stating that the court "refuse[s] to substitute our judgment on the credibility of witnesses for that of the ALJ, absent `extraordinary circumstances'" and that the court will not sit "as a super trial exam- iner," weighing "the credibility of one witness against another" nor searching for "contradictory inferences" (internal quotation marks and citations omitted) (emphasis added)). In Fieldcrest Cannon, Inc. v. NLRB, 97 F.3d 65 (4th Cir. 1996), we very recently rejected an almost identical challenge to an ALJ's credi- bility determinations. In that case, the ALJ credited all 83 Board wit- 64 nesses and discredited all 154 of Fieldcrest's. There we faithfully applied our Air Products precedent and refused to find "exceptional circumstances," noting that the ALJ's careful fact-finding was entitled to respect. See id. at 69-70. In particular, we allowed the ALJ's assessments of demeanor to be general, since "the balancing of wit- nesses' testimony is at the heart of the factfinding process, and it is normally not the role of the reviewing courts to second-guess a fact- finder's determinations about who appeared more `truthful' or `credi- ble.'" Id. at 71. Indeed, we noted the important context that "[a]n atmosphere of reprisal and recrimination against union supporters" provided as a basis for evaluating the testimony of witnesses and observed that it undoubtedly served to enhance "the credibility of those who were accusing the company of violations." Id. at 72. Rather than applying our own clear precedent in Air Products and Fieldcrest, which is fully consonant with that of our sister circuits, the majority, instead, misapplies our earlier language in Burlington Indus., Inc. v. NLRB, 680 F.2d 974 (4th Cir. 1982). In context, we there said: As we have stated many times, we are in no position, when viewing the cold record, to evaluate the credibility of wit- nesses. For this reason, we heavily defer to the ALJ's deter- minations of witness credibility. We are not, however, required to accept the ALJ's credibility determinations where they are not supported by substantial evidence. Id. at 977 (citation omitted). Although in Burlington we did, in fact, reject one of the ALJ's credibility determinations, we did so by spe- cifically reviewing the record and finding substantial evidence for that determination to be lacking. Indeed, we suggested that credibility res- olutions could only fall by application of the substantial evidence test, and, in that case, because it was clear that the Board's order would remain unaffected, we refused to engage in any additional analysis. See id. at 978. In the instant case, the majority not only ignores the overwhelming precedent recognizing the heavy deference given to credibility determinations, but abjures its responsibility to engage in even one substantial evidence review of one of these determinations, let alone the dozens that the ALJ responsibly made. That, I submit, is "perfunctory" and "conclusory" and totally inappropriate for a 65 review of Board findings that is normally deferential by express Con- gressional design. If the majority wishes to overturn these dozens of credibility determinations, or at least enough of them to show that the Union did not enjoy majority support, then it must assess each one individually to determine whether it is supported by substantial evi- dence. I believe, in contrast, that such "exceptional circumstances," Air Products, 717 F.2d at 145, as would warrant us in overturning these credibility determinations do not exist here. The ALJ's reasoned explanation for why it credited the organizers' testimony over Be- Lo's terrified witnesses is not "patently insupportable," "beyond the bounds of reason," "hopelessly incredible," "self-contradictory," nor "patently unreasonable." The ALJ and Board therefore properly counted these 43 cards. Finally, I am totally dumbfounded by the majority's rejection of 13 authorization cards on the ground that the ALJ could not authenticate them. See slip op. at 15-16. Federal Rule of Evidence 901(b)(3), by its own terms, allows the trier of fact to authenticate by comparison "with specimens which have been authenticated." In 14 cases, the ALJ compared employee signatures on the authorization cards with their signatures on their respective W-4 forms which were stipulated to be authentic. The majority, with unfathomable logic, admits that the W-4 forms themselves were authentic but declares that the signa- tures on those forms were not. Slip op. at 15-16 n.4. What was being stipulated to be authentic if not the signed W-4 forms which came from the company's files? Although the majority may be confused, certainly none of the parties nor the ALJ were, as the following collo- quy demonstrates: [Union counsel]: Your Honor, I think the issue is whether, at this point in time in the trial, we need to put a witness on, a lay witness, to say, yes, the signature on the W-4 looks like the signature on the card. It seems to me that you're going to make that decision anyway regardless-- [ALJ]: A lay witness to do that, no. You're not going-- you don't need to put a lay witness on. 66 [General Counsel]: And, that's all we're stating. [Be-Lo counsel]: Yeah, we agree you don't have to put a lay witness on. [ALJ]: Okay. So, it's stipulated that these are authentic W- 4s from the Company's files for the people whose signature appears on the W-4s . . . . J.A. at 1082 (emphasis added). As this exchange proves, Be-Lo itself stipulated that the signed W-4 forms from its own files were authentic and that other witnesses would not have to be put on to authenticate the signatures, allowing the ALJ to so determine by comparison. To posit that a completed W-4 form is stipulated to be authentic but not the signature thereon makes the above colloquy totally meaningless. I could find no authority, and the majority cites none, in which a reviewing court has determined that a signature on a document stipu- lated to be authentic is itself unauthenticated. It is obvious that the point of the colloquy was not to establish a basis for later challenging the authenticity of the W-4 signatures but rather was merely to pre- vent unnecessary delay in the trial by having laywitness after laywit- ness authenticate signatures, pursuant to Fed. R. Evid. 901(b)(2), when there was no question about their genuineness, having come from Be-Lo's own files. See also 28 U.S.C. § 1731 ("The admitted or proved handwriting of any person shall be admissible, for purposes of comparison, to determine the genuineness of other handwriting attributed to such person."). In any event, the ALJ did, in fact, carefully compare the signatures on the authorization cards with those on the W-4 forms. Indeed, he rejected one of the 14 where the "handwriting was visibly and obvi- ously different" but counted the remaining 13 since the two signa- tures, in each case, "clearly appear to have been made by the same person." J.A. at 378. Neither the ALJ nor the Board abused its discre- tion in counting these 13 cards. Because the majority either ignores or improperly applies the appropriate standard of review with respect to the Board's resolution of Be-Lo's exceptions to the authorization cards, I would conclude that the Union possessed majority support on March 20, 1991. 67 3. Even under the stricter requirements necessary to support a Cate- gory II bargaining order, the Board thoroughly explained its rationale in issuing the order, including the causal connection between the unfair labor practices and the improbability that a fair rerun could be held, as well as the issue of employee turnover, and substantial evi- dence supports its conclusion that the bargaining order was more than warranted. See J.A. at 352-54. The majority's rejection of the Board's detailed analysis essentially boils down to its belief that the violations were not widespread or pervasive and that, due to turnover in the workforce, the effects of the violations have dissipated. See slip op. at 16-23. The 59 § 8(a)(1) violations occurred in 18 of the unit's 30 stores and the 14 § 8(a)(3) violations occurred in 8 stores. According to the ALJ's analysis, the Union enjoyed majority support in 19 of the 30 stores. See J.A. at 378-80. Not surprisingly, Be-Lo's unlawful anti- union zeal manifested itself most frequently in those stores where the Union enjoyed the greatest support.10 The ALJ found § 8(a)(1) viola- tions to occur at only three stores where the Union did not enjoy majority support11; no § 8(a)(3) violations occurred at such stores. At _________________________________________________________________ 10 Thus, for example, at Store 37, where 37 of 47 employees signed authorization cards, the ALJ found 8 § 8(a)(1) violations and 1 § 8(a)(3) violation; at Store 62, where 10 of 16 employees signed authorization cards, the ALJ found 5 § 8(a)(1) violations; at Store 102, where 6 of 6 employees signed authorization cards, the ALJ found 4 § 8(a)(1) viola- tions and 1 § 8(a)(3) violation; at Store 126, where 15 of 20 employees signed authorization cards, the ALJ found 6 § 8(a)(1) violations and 3 § 8(a)(3) violations; at Store 148, where 14 of 21 employees signed authorization cards, the ALJ found 5 § 8(a)(1) violations and 2 § 8(a)(3) violations; at Store 234, where 37 of 48 employees signed authorization cards, the ALJ found 7 § 8(a)(1) violations; and at Store 235, where 12 of 18 employees signed authorization cards, the ALJ found 1 § 8(a)(1) violation and 3 § 8(a)(3) violations. 11 It is of course possible that Be-Lo's vigorous anti-union campaign dissipated majority support at these stores even before the authorization cards were signed. Thus it could well be that its violations paid their divi- dends. 68 only one store where the Union did enjoy majority support were no individual violations found. The majority contends that these numerous egregious violations are not "pervasive" because they were not "felt throughout all, or virtually all, of the bargaining unit." Slip op. at 17. No support is provided for this assertion. In the circumstances of this case, it is ludicrous to assert that where violations occur at 60 percent of the stores--and 95 percent of those stores where the Union exhibits majority support-- that such violations are not "pervasive" within the meaning of that word for Gissel Category I purposes, let alone "extensive" within the meaning of that word for Gissel Category II purposes, a category spe- cifically designed for "less pervasive practices." See Gissel, 395 U.S. at 613-14. It flies in the face of common sense to expect an anti-union employer to attack where it already enjoys presumptive favor. The majority further attempts to bolster its determination of lack of pervasiveness by asserting that "only some 40 of Be-Lo's 756 employees (less than six percent of Be-Lo's work force) were directly affected by these individual violations." Slip op. at 18. The Board obviously made no such finding. Moreover, the test for the effect of violations has never been directly individualized. The majority's approach would require hundreds of violations in order to be "perva- sive" under the facts of this case. Because such an approach plainly ignores the dynamics of the workplace, we have previously noted that the natural "assumption is that employer threats, especially threats of plant closure, are rapidly disseminated among the employee group." Air Prods., 717 F.2d at 146 (citations omitted). Here at least 20 repeated threats of store closure, not counting the company-wide "pink slip" episode, were made at 14 separate stores prior to the election.12 Applying this presumption just to the 14 stores at which the threats of store closure were made (i.e. assuming that none of these threats of store closure, nor any of the other threats, were communi- cated by affected employees to any employees at the other 16 stores), still results in the conclusion that nearly 400 employees (384) were affected by these threats. Thus, more than 50 percent of Be-Lo's _________________________________________________________________ 12 The ALJ found repeated threats of store closure, a hallmark § 8(a)(1) violation, to be made at Stores 27, 37, 73, 102, 107, 110, 122, 126, 144, 145, 148, 232, 234, and 236. See J.A. at 360-74. 69 employees were affected just by the threats of store closure alone, far more than the 20 percent we found to be sufficient to satisfy Appletree Chevrolet I requirements in Air Products.13 See id. Because of the seriousness of threats of closure, courts routinely enforce Gissel Category II bargaining orders, without ever inquiring into the percent- age of workers directly affected by the violations. See, e.g., Indiana Cal-Pro, Inc. v. NLRB, 863 F.2d 1292, 1301-02 (6th Cir. 1988); NLRB v. Horizon Air Servs., Inc., 761 F.2d 22, 31 (1st Cir. 1985). The majority also rests its decision that a fair rerun election is pos- sible on the grounds that employee turnover and the passage of time have dissipated the effects of the unfair labor practices. It appears the majority wants to create some level below the four year period between the election campaign and the Board's decision and the 66 percent employee turnover during that time into mechanistic thresh- olds above which bargaining orders can never be supported. For example, the majority attaches portentous significance to the fact that "[a]t oral argument, neither counsel for the Board nor counsel for the Union could cite us to any case in which a Gissel Category II order was enforced in the face of [66 percent] employee turnover." Slip op. at 21-27. This is simply false. In response to our query, counsel noted at oral argument that the Board, in its written opinion, cited Action Auto Stores, 298 N.L.R.B. 875 (1990), enf'd mem. 951 F.2d 349 (6th Cir. 1991), in which a bargaining order was enforced where there was 75 percent turnover. See J.A. at 353 n.41. The majority is also forced to distinguish our decision in NLRB v. So-Lo Foods, Inc., 985 F.2d _________________________________________________________________ 13 This analysis does not rely on the "pink slip" which was distributed to every employee. Because I believe substantial evidence supports the Board's conclusion that the "pink slip" episode constitutes a § 8(a)(1) threat of store closure, a fortiori, I would conclude that there can be no question but that the impact of Be-Lo's unlawful conduct was "perva- sive" or "extensive" within the meaning of either category of Gissel order. In addition, this analysis does not rely on Be-Lo's disparate enforce- ment of its no-solicitation policy. Uncounted numbers of employees must necessarily have seen the picketers, including some of their fellow employees. The effect of the disparate enforcement and the injunctions against the picketing must have been widely felt. 70 123 (4th Cir. 1992), in which we enforced a bargaining order in the face of 73 percent turnover.14 Most courts do not take such a jaded view of the exigencies of the facts before them and the realities of the workplace. By actually examining the circumstances of the unfair labor practices, courts will, and have, upheld Gissel Category II orders in the face of very high, see, e.g., Amazing Stores, Inc. v. NLRB, 887 F.2d 328, 329-31 (D.C. Cir. 1989) (95 percent turnover), cert. denied, 494 U.S. 1029 (1990), indeed, even universal turnover, see, e.g., NLRB v. Gordon, 792 F.2d 29, 33-34 (2d Cir.) (100 percent turnover), cert. denied, 479 U.S. 931 (1986). In fact, it appears that as the pervasiveness of the unfair labor practices under a Category II analysis approaches the pervasiveness that would support a Category I order, notwithstanding the "outra- geousness" that also must still be satisfied in such cases, courts require much more minimal findings on dissipation due to turnover. See Amazing Stores, 887 F.2d at 331. Courts are also savvy to the litigation tactics of anti-union employ- ers. For example, the District of Columbia Circuit expressly rejects the need for an "elaborate turnover inquiry" where a company has a naturally high turnover rate, for otherwise the employer "would be tempted to engage in unfair practices, knowing that over the course of protracted administrative proceedings they can outlast any union activists and effectively stall existing momentum for unionization." Id.; see also St. Francis Fed'n of Nurses and Health Professionals v. NLRB, 729 F.2d 844, 856 (D.C. Cir. 1984) (stating that "normal employee turnover [does not] justify displacing the Board's determi- nation that a bargaining order is appropriate" (citation omitted)). _________________________________________________________________ 14 The majority also ignores our requirement in So-Lo Foods that an employer bears the burden of showing that turnover occurred through normal attrition. Without such evidence, the court is warranted in specu- lating, contrary to the majority's presumptions here, that "the significant turnover is attributable to management maneuvering from a Company intent on dissipating a pro-union group of employees." So-Lo Foods, 985 F.2d at 129. Be-Lo presented no such evidence here. The majority thus violates our own admonition that "we would be remiss in considering such turnover evidence where it is not clearly established that the turn- over occurred innocently." Id. (emphasis added). 71 Other courts are in agreement. See, e.g., Justak Bros. & Co. v. NLRB, 664 F.2d 1074, 1082 (7th Cir. 1981) (disallowing an employer to seize upon turnover, for otherwise "an employer could engage in a scheme of unfair labor practices and yet escape a bargaining order by delaying and waiting for employee turnover"); Chromalloy Mining & Minerals Alaska Div., Chromalloy American Corp. v. NLRB, 620 F.2d 1120, 1132-33 (5th Cir. 1980) (noting that"an employer with a rapid turnover rate could commit unfair labor practices freely, . . . [and ultimately] avoid any bargaining obligation indefinitely, and the union would assume the role of Sisyphus, condemned in Hades con- tinually to roll a stone up a hill, only to find that it slides down again before reaching the top"). The Board did, in fact, analyze the turnover issue. Accepting for the sake of Be-Lo's argument that only one-third of the original bar- gaining unit remained, the Board was persuaded that the more than 250 remaining employees had been directly affected by serious unfair labor practices and that, as a result, new employees would be affected by the prior unlawful conduct. The majority dismisses the Board's reasoning as nothing more than a "talismanic invocation of the `lore of the shop.'" Slip op. at 23. The majority plainly mischaracterizes the Board's analysis, totally ignoring the Board's extensive, stated rea- sons for rejecting the turnover argument: [Be-Lo] points to no action that it took to eradicate the effects of its unlawful threats of closure and discharge, and its actual discharge of union adherents. The evidence, indeed, is to the contrary. After the election,[Be-Lo] unlaw- fully discharged four more employees and unlawfully failed to recall one employee from layoff. In addition, beginning in April, [Be-Lo] discriminatorily applied its no-solicitation rule to deny the Union access to its property for the purpose of peaceful picketing and threatened picketers with arrest. Finally, . . . [Be-Lo] unlawfully maintained state court tres- pass actions against the Union even after the Board had issued a complaint against [Be-Lo] and required [Be-Lo] to stay those actions. Thus, it is clear that in the months after the election [Be-Lo's] management was intent on continuing its extensive antiunion campaign. . . . [H]igh officials, who were themselves responsible for serious violations of the 72 Act, remain, including Manual Saunders, the owner of three stores and member of [Be-Lo's] board of directors; the Har- rells, who own four stores; and Scott, [Be-Lo's] vice presi- dent for operations. The continuing presence of these owners and high officials can serve only to reinforce in the minds of the employees the lingering effects of[Be-Lo's] violations. Further, certain middle and lower level officials still remain who committed a large number of the unlawful acts found by the judge. Thus, the numerous and serious violations found, the participation of management from top to bottom, and the continuation of [Be-Lo's] conduct after the election, show that [Be-Lo] is deeply committed to its antiunion position, a commitment from which it is not likely to retreat. J.A. at 353-54 (internal quotation marks and footnotes omitted).15 Unlike the majority, I find this analysis to be neither "terse," "specula- tive," "indeterminate," nor "blind." See slip op. at 22. Indeed, it is more than even the "minimal findings" that ought to suffice in these circumstances. See Amazing Stores, 887 F.2d at 331. In fact, it is more extensive than the analysis we upheld in Air Products under our Appletree Chevrolet I specificity requirement. See Air Prods., 717 F.2d at 146. For similar reasons, the passage of time must be examined on a case-by-case basis. As a general matter, courts have refused to find the lapse of time alone sufficient to overturn a bargaining order. See, e.g., St. Francis, 729 F.2d at 856 (stating that "holding a rerun elec- tion simply because of the passage of time rewards employer recalci- trance and offers no deterrence to future unfair labor practices" (internal quotation marks and citation omitted)); NLRB v. L.B. Foster Co., 418 F.2d 1, 4 (9th Cir. 1969), cert. denied, 397 U.S. 990 (1970). We have recently enforced a bargaining order in which four years elapsed from the time of the representation election and the Board's order, the same as here. See Fieldcrest Cannon, Inc. v. NLRB, 97 F.3d 65 (1996). Even longer periods of time are not unusual. See, e.g., Davis Supermarkets, Inc. v. NLRB, 2 F.3d 1162 (D.C. Cir. 1993) _________________________________________________________________ 15 The ALJ specifically found that 28 officials and supervisors who had committed unfair labor practices remained at Be-Lo. See J.A. at 380. 73 (nearly six years elapsed from time of filing of complaint and Board's bargaining order), cert. denied, 511 U.S. 1003 (1994). Given the com- plexity of this case, the number of violations alleged and found, the length of the trial (57 days over a seven month period), and the fact that Be-Lo contested every violation found against it, to conclude that this passage of time necessarily dissipated the effects of the multitudi- nous unfair labor practices merely rewards Be-Lo for a litigation strat- egy designed to increase time and turnover. There is absolutely no evidence that Be-Lo has been harmed by the four year administrative proceedings; indeed, as the majority decides the matter, they have been positively benefited. I refuse either to encourage or to turn a blind eye to the practice of anti-union employers deciding it is eco- nomically advantageous to pay attorneys' fees rather than negotiate with a union. The Board thoroughly explicated its rationale that because Be-Lo's unlawful misconduct was severe and pervasive--including more than 50 hallmark violations at fully half of the stores, more than 30 of them just before election--its lingering effects would not be dispelled by the passage of time. Moreover, the Board was particularly convinced that the possibility of a fair rerun election would be unlikely because the most serious violations were committed throughout Be-Lo's hier- archy, from the president, store owners, and other high officers on down to district managers, merchandisers, store managers, coman- agers, and assistant managers and because Be-Lo maintained its vehe- ment anti-union campaign for months after the election. See J.A. at 352-54. Substantial evidence supports the Board's conclusion that a fair rerun election is impossible. The Board's detailed analysis fully comports with our requirements under Appletree Chevrolet I. See Appletree Chevrolet I, 608 F.2d at 996-97; cf. Air Prods., 717 F.2d at 146 (enforcing a bargaining order, and sustaining the Board's anal- ysis under our Appletree Chevrolet I requirements, in circumstances where the violations were less pervasive and egregious than here and the Board's analysis less detailed). I remain at a loss to understand how the majority could possibly conclude that the Board abused its discretion in issuing a bargaining order in this case. The record in this case plainly supports the issuance of either a Gissel Category I or Category II bargaining order. The Board respon- 74 sibly performed its duties such that either category of order should be enforced. VII. The majority's opinion is high on rhetoric and hyperbole. Unfortu- nately, it is bereft of a thoughtful consideration of our precedent as intertwined with our limited scope of review. The majority is only able to reach its decision by sitting as a super trial examiner, substitut- ing its own judgment throughout for the Board's and mischaracteriz- ing the evidence and facts. I cannot agree with the approach the majority takes, let alone its results, and I must respectfully dissent. 75

Be-Lo Stores v. NLRB

Related Cases