The opinion
SNEED, Circuit Judge
(with whom
CHOY, ANDERSON and FARRIS, Circuit Judges, concur) dissenting:
As with many legal issues, where one ends up in this case depends upon where one starts. While I have no disagreement with the court’s holding that union signatory clauses are secondary and within the scope of section 8(e), I respectfully dissent from its holding that such clauses are invariably within the construction industry pro *1325 viso to section 8(e) when a collective bargaining relationship exists between the employer and the union seeking the clause. The majority reaches its conclusion by commencing with the view that the construction industry presents unique problems in labor relations that require the legitimation of union signatory clauses. Although admitting that the legislative history is not compelling, the court finds that legitimation was provided by the construction industry proviso. My starting point, on the other hand, is with Connell Construction Co. v. Plumbers & Steamfitters Local 100, 421 U.S. 616 , 95 S.Ct. 1830 , 44 L.Ed.2d 418 (1975), and its concern with the “top-down” organizing effect of union signatory clauses. This concern led the panel before which this case was heard originally to hold as follows:
“We hold that the construction industry proviso extends shelter only when a collective bargaining relationship exists and even then only when the employer or his subcontractor has employees who are members of the signatory union at work at some time at the jobsite at which the employer wishes to engage a nonunion subcontractor.” Pacific Northwest Chapter of the Associated Builders & Contractors, Inc. v. NLRB, 609 F.2d 1341, 1347 (9th Cir. 1979).
I would adhere to that holding and to the reasons given by the panel therefor. Id. at 1348-50 .
The court rejects the panel’s result because it is said to represent an “even narrower” view of the construction industry proviso than that adopted by the Supreme Court in Ccnnell. 654 F.2d 1320 . Alleviation of “the frictions that may arise when union men work continuously alongside nonunion men on the same construction site,” 421 U.S. at 630 , 95 S.Ct. at 1839 , is said not to have been a concern of the Supreme Court in Connell because in that case “the union did not purport to represent any of the interests served by the proviso.” 654 F.2d 1320 . Presumably this is because there was no collective bargaining agreement between the employer and union in that case.
As I see it, the crucial language in Connell is as follows:
“We therefore hold that this agreement, which is outside the context of a collective-bargaining relationship and not restricted to a particular jobsite, but which nonetheless obligates Connell to subcontract work only to firms that have a contract with Local 100, may be the basis of a federal antitrust suit because it has a potential for restraining competition in the business market in ways that would not follow naturally from elimination of competition over wages and working conditions.” 421 U.S. at . 635, 95 S.Ct. at 1841 .
This language, as the majority acknowledges, is not entirely unambiguous. It is, however, not inconsistent with the panel’s holding as set out above; and, more importantly, that holding is more empathetic with Connell’s concern with the “top-down” organizing effect of union signatory clauses than is that of the majority;
It is true, as the majority states, that any union signatory clause has some “top-down” organizing effect. The difference between the panel’s holding and that of the en banc court lies in the scope of this effect. The en banc court permits a clause that forbids subcontracting work covered by the collective bargaining agreement to any firm not having a contract with the signatory union. It matters not whether any of the employees of the signatory employer, who are members of the signatory union, are employed on the jobsite. It matters not that, in the absence of the union signatory clause, the work could have been subcontracted to a firm which, even though adhering to union standards governing wages and working conditions, was more efficient than its competitors.
It must be acknowledged that the panel’s holding would not greatly diminish the “top-down” organizing effect of the union *1326 signatory clause. It would permit the choice of the more efficient firm only when to do so would entail no jobsite friction between members of the signatory union and workers performing work covered by the collective bargaining agreement. Thus, a provision in the agreement precluding subcontracting of all the covered work would prevent choice of the more efficient firm and thus provide a substantial “top-down” organizing effect. A clause prohibiting subcontracting is permissible. National Woodwork Manufacturers Association v. NLRB, 386 U.S. 612, 644-46 , 87 S.Ct. 1250, 1268-69 , 18 L.Ed.2d 357 (1967); see also Fibreboard Paper Products Corp. v. NLRB, 379 U.S. 203 , 85 S.Ct. 398 , 13 L.Ed.2d 233 (1964); NLRB v. Johnson, 368 F.2d 549 (9th Cir. 1966). However, the panel’s holding would afford an opportunity by which an employer could accept a union signatory clause without sacrificing all right to choose the more efficient subcontractor. Preservation of this opportunity would be more consistent with the primary concern of Connell than is the position of the en banc court. I doubt that the en banc court seriously contends otherwise. It does believe that the preservation of this opportunity would provide a “strong incentive for anti-union discrimination by employers.” 654 F.2d 1322 . I am dubious. Under either view an anti-union employer will strive to avoid a union signatory clause. That an employer under the panel’s view would have a limited opportunity to subcontract to a firm not having an agreement with the signatory union is not likely to make it more anti-union.
I agree with the en banc court that the legislative history of the construction industry proviso is not conclusive. We have previously considered the legislative history of the proviso. ACCO Construction Equipment, Inc. v. NLRB, 511 F.2d 848, 851 (9th Cir. 1975). In surveying the legislative history, we concluded that the dominant motive of Congress in adopting the proviso was to permit agreements that aimed at eliminating jobsite friction between union and nonunion workers. Id. at 851 ; accord, Drivers Local 695 v. NLRB, 361 F.2d 547 , 553 (D.C.Cir.1966); Essex County District Council of Carpenters v. NLRB, 332 F.2d 636, 640 (3d Cir. 1964). And the National Labor Relations Board previously believed: “[T]he 8(e) proviso was intended to prevent labor strife among nonunion and union employees at the same jobsite.” International Union of Operating Engineers, Locals 542, 542-A and 542-B, 216 N.L.R.B. 408 , 409 (1975), enforced, 532 F.2d 902 (3d Cir. 1976), cert. denied, 429 U.S. 1072 , 97 S.Ct. 808 , 50 L.Ed.2d 789 (1977). I see no reason to depart from our earlier analysis of the purpose of the proviso.
Admittedly, Senator Kennedy thought that the construction industry proviso would permit subcontracting clauses such as those challenged in these cases. Others argued that the proviso was intended to permit agreements not to use prefabricated pipe unless union workers did the fabricating. 105 Cong.Rec. 19785-86, 19809 (1959); II Leg.Hist. 1815-16 (extension of remarks of Senator McNamara and Representative Thompson); contra 105 Cong.Rec. 20004-05 (1959); II Leg.Hist. 1861 (extension remarks of Representative Kearns). But neither of these purposes were clearly understood as the reason for the proviso. As Connell and ACCO concluded, the legislative history supports the conclusion that most congressmen thought the proviso was intended to address the problem of jobsite friction. See, e. g., I Leg.Hist. 82 (message from President Eisenhower to Congress); 105 Cong.Rec. 15852, II Leg.Hist. 1684 (statement of Congressman Goodell); 105 Cong.Rec. 17333-34, II Leg.Hist. 1383-84 (proposed instructions for Senate conferees). Also, as I read the excerpt from the memorandum by Representatives Thompson and Udall quoted by the en banc court, 654 F.2d 1319 , it too strongly supports the conclusion that the proviso was aimed at the jobsite friction problem. The' references to employees working “side by side” on a “single building” so indicate. 105 Cong.Rec. 15541 (1959); II Leg.Hist. 1577.
The en banc court notes that some subcontractor clauses similar to those chai *1327 lenged in these cases existed in 1959. Since they conclude that the Congress intended to preserve the law relating to the construction industry as it then existed, the court argues that Congress intended that the proviso would protect broad subcontractor clauses. I believe this reasoning gives far too much weight to the bare reference to the pattern of collective bargaining referred to in the Committee Report. See Connell, supra, 421 U.S. at 628-29 , 95 S.Ct. at 1838 . Furthermore, the lawfulness and permissible scope of subcontractor agreements had not been conclusively determined in 1959. And the proviso clearly changed the law by explicitly limiting its protection to jobsite work. That limitation at least partially overruled the Sand Door case, Local 1976, United Brotherhood of Carpenters v. NLRB, 357 U.S. 93 , 78 S.Ct. 1011 , 2 L.Ed.2d 1186 (1958). Thus the most that can be said is that Congress intended to preserve the existing law within the scope of the proviso. That proposition, however, begs rather than answers the question of what the proper scope of the proviso is. That is what this case is about. Our decision, whatever it is, can draw but little comfort from the legislative history.
The majority also argues that practical considerations favor the result they reach. But the practical problems of following the panel’s decision would not have been severe. If broad subcontractor clauses are approved, unions must still police construction sites to ensure compliance. No more is required by a rule limiting subcontractor clauses to jobsites at which members of the signatory union are employed. The serious practical problems raised by the majority would arise only if subcontractor clauses were required to be renegotiated on a job-site by jobsite basis, but that would not be required by the panel’s interpretation of the proviso.
Thus, we come to the nub of the matter. Which interpretation of the proviso is more faithful to curbing “top-down” organizing, which, it must be remembered, was Congress’s objective in adopting the 1959 Act which included section 8(e). Connell, supra, 421 U.S. at 632-33 , 95 S.Ct. at 1840 . The holding of the panel gets my nod. 1
. Because the panel concluded that the broad subcontractor clause in the AGC agreement violated section 8(e), it did not address the lawfulness of the agreement’s self-help enforcement provision. 609 F.2d at 1351 . I agree with the majority of this en banc court that the self-help enforcement provision must be interpreted to mean what it says and that Congress clearly intended that subcontractor clauses were to be enforced judicially rather than by strikes or picketing. Therefore, I agree that the self-help clause in the AGC contract is invalid.