386 F.2d 551
NATIONAL LABOR RELATIONS BOARD, Petitioner,
SEHON STEVENSON & CO., Inc., Respondent.
United States Court of Appeals Fourth Circuit.
Argued Feb. 8, 1966.
Decided Oct. 27, 1967.
Peter M. Giesey, Atty., N.L.R.B. (Arnold Ordman, Gen. Counsel, Dominick L. Manoli, Associate Gen. Counsel, Marcel Mallet-Prevost, Asst. Gen. Counsel, and Warren M. Davison, Atty., N.L.R.B., on brief) for petitioner.
C. Robert Schaub, Huntington, W. Va. (Jenkins & Jenkins, Huntington, W. Va., on brief) for respondent.
Before HAYNSWORTH, Chief Judge, and SOBELOFF and BOREMAN, Circuit judges.
HAYNSWORTH, Chief Judge:
We enforce a Board order to bargain in this case, though the union's majority has not been established by an election, because the employer conducted his own investigation, which confirmed the union's claim, and left the employer with no basis for a good faith doubt of it.
In support of a demand for bargaining upon a claim of majority representation, the union1 submitted to the employer 19 signed authorization cards. There had been 29 employees in the unit, though three of them had just been discharged, two of whom were reinstated two weeks later. Omitting the one employee lawfully discharged, the union had 18 authorization cards from a unit of 28. If the two wrongfully discharged employees, shortly reinstated, are excluded there were 16 signed cards of 26 employees in the unit.
Representatives of the employer received the cards and checked them against a list of employees. Objection was made to the cards signed by the three recently discharged men, but otherwise there was no objection to the cards and no expression of doubt of the union's claim of majority status. The employer stated that the question of recognition would be discussed with his directors.
Thereafter, the employer polled all of the employees. Each was asked about his signing a union authorization card, and his answer was noted on a list of employees. There was evidence that two of these conversations ranged widely and included threats of discharge and of plant closure.
Three days after the employer's check of the union's cards, the union filed charges of violations of 8(a)(1), (3) and (5) of the Act. Twelve days later, the employer filed a petition for an election under 9(c), and, on the same day, the employer's lawyer wrote a letter to the union in which it was stated the cards did not appear in order, the signatures did not check out completely and that several employees 'are understood to have stated' they signed under some misunderstanding. The record contains no subsequent reference to the claims of disorder in the cards or their signatures.
The employer's petition for an election was denied because of the pendency of the unfair labor practice charges. A hearing was held on those charges resulting in findings of violations of 8(a)(1), (3) and (5).
At the hearing, there was proof of the polling of the employees and of the instances of threats of discharges and plant closing. The finding of the threats supports the conclusion of a violation of 8(a)(1), if the polling did not. There was also evidence of circumstances surrounding the discharge of two of the three men fired three days before the card check which clearly supports the inference that those two discharges were discriminatory. Those were the two men who were reinstated two weeks later with pay for 80 hours of work. The Board contends that the question of those discharges is not moot, however, for those employees normally worked a longer work week than 40 hours and each lost a week's vacation with pay. The resultant 8(a)(3) finding is obviously correct and enforceable.
In NLRB v. Logan Packing Co., 4 Cir., 386 F.2d 562, decided this day, we have considered at some length the serious unreliability of signed authorization cards as an indication of the wishes of a majority of the employees. Because of that unreliability, they should not be accepted as proof of the union's claim of majority status, and an employer is entitled to doubt the union's claim as long as it is so unreliably founded.
The presentation of such a claim places an employer in a difficult dilemma, for, thus far, the Board has given him no assurance comparable to that we now proffer. Yet, if he undertakes to inform himself, he runs great risk of being charged with and held responsible for unfair labor practices, just as occurred here.2 It is almost impossible for an employer, in such a situation, to steer a course of inquiry, safely and free of such grave risk.3 And, if he is found to have committed an unfair labor practice in the course of that perilous run, the Board will employ the fact to foreclose assertion of the doubt the employer sought to resolve.
Under the circumstances, the representation issue should have been treated as such. The employer is unconcerned about the 8(a)(1) and 8(a)(3) charges. It indicated its willingness to make the two discharged employees whole; it is a matter of a very few dollars only. There is every indication it would have posted appropriate notices, and a valid, reliable election could have been held long ago. The controversy is a result of the bargaining order and insistence upon bypassing the much swifter, more reliable election process.
After the hearing, however, we are not at liberty to disregard the evidence of the result of the employer's own investigation and its bearing upon the presence of a question concerning representation.
The employer admitted that it polled all of the employees and made notations of the answers. The witness first claimed that he made no tally of the results, though later he admitted he did, and it was silly to have denied it in the first place, for the tally was the sole purpose of the interrogations. He did not say what those results actually were, but the employer's strong interest in countering the union's claim and the reluctance of the witness to reveal the results of his investigation, compels the inference that the interrogations confirmed the union's claim.
There is no reason to attribute unreliability to the results of the employer's inquiry. If the interrogation of each employee, as disclosed on this record, was less searching than it might have been there is no indication that the employer, after embarking upon the inquiry, felt under any constraint about the range of the questions. If he delved no more deeply than the record indicates, it was apparently because he had gone far enough to obtain answers he thought acceptable and reliable.
The employer had expressed no doubt of the union's claim when the cards were checked. He may well have had independent information of the union's strength. It was not until two weeks later that the employer's lawyer referred vaguely to unspecified misunderstandings of card signers. At the hearing, the employer's president testified that he had heard that some of the older employees did not want the union, a circumstance not shown to have been inconsistent with the union's claim, and that he 'had some information * * * that more than one person' thought he was signing an application for free insurance rather than a union authorization card. That claim was not otherwise supported.
When the employer has conducted his own investigation, the results of which confirm the union's claim, we think it reliably establishes the union's majority and the employer's lack of any good faith doubt of it. Resurrection of that doubt, after the investigation, requires much more than a very general unspecific and unelaborated suggestion that some of the employees may have thought the union's authorization cards were applications for insurance which would cost them nothing.
If we had no more in this case than the cards and the finding of violations of 8(a)(1) and (3), the Board's order to bargain would not be enforced for the reasons we state today in Logan. The employer's investigation, however, so conclusively answers the question of the union's majority and the question of the employer's lack of doubt of it, we think it may reasonably be said to have settled the representation question. If that question is settled, if its answer is not dependent upon a choice of permissible inferences, it is within the Board's power to order bargaining without resort to a 9(c) election, the machinery for the resolution of representation questions.
SOBELOFF, Circuit Judge (concurring specially):
While I concur in the result reached by the majority, I feel obligated to record my disagreement with some of its reasoning. With my colleagues, I conclude that the Labor Board's order should be enforced because the union represented a majority of the employees in an appropriate unit and the employer had no good faith doubt of this fact. Florence Printing Co. v. NLRB, 333 F.2d 289 (4th Cir. 1964); NLRB v. Overnite Transportation Co., 308 F.2d 279 (4th Cir. 1962).
My concern is with the court's over-broad dicta regarding the unreliability of signed authorization cards and its gratuitous assertion that in the absence of the employer's poll, the union's majority would not have been satisfactorily established. I think the majority opinion, as well as the Logan1 case, decided today, upon which it relies, overstates the unreliability of authorization cards and undermines the useful functions they may properly serve. Since the Logan opinion, with its elaborate discussion of authorization cards, is incorporated by reference into the majority opinion of this case, it is necessary to examine Logan closely.
Oddly, in Logan the court observes with singular impartiality that there is nothing more unreliable than cards in determining the desires of a majority of employees unless it be an employer's poll. Yet in the instant case, the court joins these two 'unreliable' devices and 'conclusively' establishes the union's majority. I agree that a survey taken by an employer is likely to be an extremely poor barometer of the employee's attitude toward unionization. The primary force of the poll conducted by this employer is, in light of its revelations, to show convincingly the employer's lack of good faith. Only secondarily, and insubstantially, does the poll buttress the conclusion that the union represented a majority of employees as established by the authorization cards.
There is no gainsaying that abuses may attend the solicitation of authorization cards.2 Nor can it be denied that a valid secret-ballot election is normally the preferred method of selecting a bargaining representative. Nevertheless, the cards are not inherently unreliable, and it cannot be laid down as a flat rule of law that they are unreliable regardless of circumstances. The fact is that the only statistical study of cards, upon which the Logan opinion relies heavily, concludes: 'Authorization cards do have validity.'3 In reaching the same conclusion almost from the inception of the Wagner Act, this circuit for almost thirty years has repeatedly held cards to be 'ample proof of majority representation.' NLRB v. Greensboro Coca Cola Bottling Co., 180 F.2d 840, 844 (4th Cir. 1950); Bilton Insulation, Inc. v. NLRB, 297 F.2d 141 (4th Cir. 1961); Peoples Motor Express Co. v. NLRB, 165 F.2d 903 (4th Cir. 1948); NLRB v. Schmict Baking Co., 122 F.2d 162 (4th Cir. 1941). In addition, the Supreme Court, both before and after the passage of the Taft-Hartley Act, had sanctioned the card count as an alternative means to an election for achieving exclusive representative status. See United Mine Workers v. Arkansas Oak Flooring, 351 U.S. 62, 76 S.Ct. 559, 100 L.Ed. 941 (1956); Franks Bros. Co. v. NLRB, 321 U.S. 702, 64 S.Ct. 817, 88 L.Ed. 1020 (1944).
In light of the prevalent, judicially-approved4 practice of using authorization cards during the twenty years since the 1947 statute, I am unconvinced by the strained statutory interpretation fashioned in Logan that Taft-Hartley was meant to eliminate cards. Section 9(a) specifically provides that exclusive bargaining representatives may be 'designated or selected' by a majority of employees. Section 8(a)(5), with which we are here concerned, makes it a violation to refuse to bargain with a representative chosen in accordance with section 9(a), which does not require an election. While it is true that section 9(c) now specifies that the only method the Board can employ for certification is a secretballot election, the statute nowhere limits employees to this procedure and it does not make compliance with section 9(c) a prerequisite to an 8(a)(5) violation. In fact, the House version of the Hartley Bill which would have required certification as a sine qua non of recognition was rejected in conference. See H.R. 3020, 80th Cong., 1st Sess. 21, 81 8(a)(5) (May 13, 1947); H.R.Conf.Rep. No. 510, 80th Cong., 1st Sess. 41 (June 3, 1947). See also Lesnick, Establishment of Bargaining Rights Without an NLRB Election, 65 Mich.L.Rev. 851, 861 (1967).
The proper solution of the problem arising from the possible abuse of this long-established procedure lies in a case-by-case analysis of the use made of the cards and not in their outright repudiation. Evaluation of the reliability of authorization cards should be made in two distinct contexts. One is where there is no election and no employer unfair labor practice other than a refusal to bargain; the other is where there has been an election which is vitiated by an employer violation.
The first category, illustrated by the instant case, is where the union submits to the employer cards from a majority of employees and demands recognition and collective bargaining. The majority opinion here and the Logan opinion indicate that on these facts alone the employer would be justified in refusing to bargain, basing a claim of doubt on the asserted inherent unreliability of cards. This smacks of policy-making by stereotype.
The facts of the instant case are instructive. The employer was presented with authorization cards from nearly 70% (19 out of 29) of the employees in the unit. The statistics relied upon in Logan to denigrate cards point out that when 70% Of the employees have signed up, three out of four times the union will be successful in a subsequent election. In addition, in this case there was no allegation of any ambiguity in the cards or any misrepresentation or threat or any of the other abuses so painstakingly catalogued in Logan. Notwithstanding the acknowledged statistical chances and the apparent propriety of this judicially-approved practice, the court suggests that the employer may without further showing assert a good faith doubt about the union's majority. I submit that such a contention by the employer would be baseless. Under the circumstances set forth above, it would be incumbent upon the employer to show some irregularity justifying his doubt.
This test was indeed met in Logan, which could have been more appropriately decided on narrower grounds than that chosen by the court. In that case, the employer received cards signed by barely more than 50% (43 out of 80) of the eligible employees who had four years earlier rejected the union in a consent election. Upon investigation, the employer discovered evidence of widespread union coercion designed to secure signatures. Armed with four affidavits depicting economic threats made to employees by the union, the employer filed unfair labor practice charges before the union petitioned the Board for a bargaining order. In the light of these facts, it was not unreasonable to infer that the union lacked a validly obtained majority and that the employer had a bona fide doubt. A refusal to enforce the Board's order on these facts rather than a wholesale denunciation of authorization cards would have been the more defensible course.
The Logan dicta undermine authorization cards in the second set of circumstances in which we must evaluate them. In this category the cards are vital. The situation arises when, after being presented with cards, the employer commits unfair labor practices which vitiate the subsequent Board-supervised election. In this context we are faced with a question of remedies: is a re-run election adequate to redress the wrong done or must an order to bargain based on the card majority be issued?
Just as in determining whether an employer had a good faith doubt, so in reviewing a remedy fashioned by the Board to expunge the consequences of unfair labor practices, an ad hoc investigation appears in order. In each case the ultimate question will be: given the unfair labor practice, will a re-run election be a more reliable test of the employees' desires than the card count taken before the commission of the violation? The answer will depend, of course, on the nature of the employer infraction as well as upon the circumstances attending the card check.
A good recent example of the necessity for card checks as an antidote to unfair labor practices is NLRB v. Ralph Printing and Lithograph Co., 379 F.2d 687 (8th Cir. 1967). In that case, the union submitted cards from 93% Of the employees and demanded recognition. there was no oral statement of disbelief in the union's majority and no allegation of ambiguity in the text of the cards or of threats or misrepresentations in their obtention. Nevertheless the employer refused to submit the cards to a neutral third party. After losing the election by a narrow margin, the union filed objections based on coercive interrogation and surveillance. While the investigation of these objections was pending, the company granted a ten-cent an hour wage increase, with an implied promise of more benefits if the workers should continue to vote 'correctly.' The Eighth Circuit properly enforced the Board's order to bargain. Under these circumstances, could there by any doubt that the substantial card majority furnished a more reliable guide to the employees' true feelings than a re-run election? This is a far cry from the employer's 'minimal infractions' in Logan, which could perhaps be cured by a second election.
Professor Pollitt's study of Labor Board re-run elections indicates that the Ralph Printing case is not an isolated exception. Premising his observations on 20,153 elections held between 1960 and 1962. Pollitt concludes that in over twothirds of the cases, the party who destroyed the election process in the first instance wins in the re-run. Pollitt, NLRB Re-Run Elections: A Study, 41 N.Car.L.Rev. 209, 212 (1963). Moreover, his analysis suggests that certain unfair labor practices are more irremediable by re-run elections than others. For example, threats to move the plant appear highly successful and enduring in their effect, since in only 29% Of the re-runs was a different result achieved. On the other hand, threats to eliminate benefits or to refuse to deal with the union if elected seem ineffectual, as indicated by a changed result 75% Of the time.
It is true that these statistics are based on a limited sampling and are manipulatable. While not conclusive, they do indicate that unfair labor practices irremediable by a re-run election are more frequent than the Logan opinion suggests. Absent the possibility of a valid re-run election, the only effective recourse left the Board is to order bargaining on the basis of the card majority. As the Second Circuit recently observed:
'Such card majorities must by necessity be deemed evidence of the status quo ante where the employer's conduct has been so flagrantly hostile to the organizing efforts of a union that a secret election has undoubtedly been corrupted as a result of the employer's militant opposition.'
NLRB v. Flomatic Corp., 347 F.2d 74 (2d Cir. 1965).
'Crocodile tears' shed by an employer over the loss of his employees' free and untrammelled choice after he has violated either section 8(a)(1) or (3) or both should not impress us. Nor do I share my brethren's concern that the employees may end up being represented by a minority union. In the first place, this is not a statistical probability as evidenced by the figures cited in Logan. Secondly, the Supreme Court has made it clear that there are other values-- such as the need to deter employer unfair practices-- in our labor policy which may on occasion override the desideratum of majority representation. See, e.g., Brooks v. NLRB, 348 U.S. 96, 75 S.Ct. 176, 99 L.Ed. 125 (1954); Franks Bros. v. NLRB, 321 U.S. 702, 64 S.Ct. 817, 88 L.Ed. 1020 (1944). Finally, a minority union facing a hostile employer can cause the employees little harm while having strong incentive to do them much good. See, Bok, The Regulation of Campaign Tactics in Representative Elections Under the National Labor Relations Act, 78 Harv.L.Rev. 38, 134-5 (1964). As Professor Bok writes,
'There is every reason for the union to negotiate a contract that will satisfy the majority, for the union will surely realize that it must win the support of the employees, in the face of a hostile employer, in order to survive the threat of a decertification election after a year has passed.'
Bok, supra, at 135.
In summary, authorization cards, although susceptible to abuse, have always been considered an acceptable means of selecting a bargaining agent. Their validity may not be judged without investigation into the particular card employed and the overall atmosphere in which signatures were procured. When textually unambiguous and secured without fraud or coercion, cards may be a reliable indicator of employees' union desires. More importantly, where employer unfair labor practices have destroyed the possibility of a valid election, a card majority is the only reliable guide to the wishes of the employees. I regret this court's unwarranted and unnecessary weakening of an established and occasionally essential device in the collective bargaining process.
Food Store Employees Union, Local 347, Amalgamated Meat Cutters and Butcher Workers of North America, AFL-CIO
The Board found the whole course of the polling to be a violation of 8(a) (1)
See the discussion in Union Authorization Cards, 75 Yale L.J. 805, 823-8 (1966); see also Judge Timber's dissent in NLRB v. Gotham Shoe Manufacturing Co., 359 F.2d 684, 687 (2d Cir. 1966)
1962 Proceedings: Section of Labor Relations Law, American Bar Association, 18
See, e.g., NLRB v. Southbridge Sheet Metal Works, Inc., 1 Cir., 380 F.2d 851; Jas. H. Matthews & Co. v. NLRB, 354 F.2d 432 (8th Cir. 1965); NLRB v. Winn-Dixie Stores, Inc., 341 F.2d 750 (6th Cir. 1965); NLRB v. American Manufacturing Co. of Texas, 351 F.2d 74 (5th Cir. 1965); NLRB v. Philamon Laboratories, Inc., 298 F.2d 176 (2d Cir. 1962)