In labor negotiations, as in any negotiations, the party with more information has an edge. So when Nielsen Lithographing Company claimed during wage negotiations with its workers' union that it had to reduce wages and benefits in order to remain competitive with other printers, the union asked the company to let its accountants examine the books and other records--financial statements, tax returns, records of compensation paid managerial and supervisory personnel--that would substantiate (or refute) the company's claim of need. At argument, the union's counsel said the union wanted to see all of the company's books and records, including purchasing records; although the scope of the request is not directly in issue, we remind the Board of its own asseveration that requests must be reasonable, balancing pertinence of the information sought with burden on the employer in producing it. See, e.g., Washington Materials, Inc., 276 N.L.R.B. 839, 854 (1985), enforced in part and denied enforcement in part, 803 F.2d 1333 (4th Cir.1986); see also Teleprompter Corp. v. NLRB, 570 F.2d 4, 11 (1st Cir.1977).

The company refused the demand for access to its books, and the union struck, and later filed charges with the Labor Board, which held that Nielsen's refusal to open its books to the union was an unfair labor practice and ordered Nielsen to cease the practice and rehire the workers who had struck and whom Nielsen had fired. 274 N.L.R.B. No. 118 (1986). Nielsen asks us to set aside the order, and the Board and the union ask us to enforce it. Although Nielsen's plant is in Ohio, and all the events relevant to this case, including the negotiations and the strike, occurred there, it transacts business in this circuit and is therefore entitled to seek judicial review of the Board's order here. See 29 U.S.C. Secs. 160(e), (f); S.L. Industries, Inc. v. NLRB, 673 F.2d 1, 3 (1st Cir.1982).

In NLRB v. Truitt Mfg. Co., 351 U.S. 149, 76 S. Ct. 753, 100 L. Ed. 1027 (1956), the Supreme Court held that it is an unfair labor practice for an employer who claims to be financially incapable of paying a wage increase requested by a union to refuse to let the union see the employer's books for purposes of verifying its claim. Otherwise labor negotiations would involve an even greater element of bluff, guesswork, and sheer gambling than they inevitably do, because the union would be put to the Hobson's choice of acceding to a quite possibly exaggerated claim of poverty or risking its members' jobs. The Court didn't think that forcing the union to play Russian roulette was the epitome of good faith bargaining.

Nielsen, however, never claimed that it was unable to pay the existing scale of wages and benefits. It admitted to being profitable but said it wanted to bring its wage bill into line with the wages paid by competitors to whom it was losing sales. A company can survive, certainly in the short run and often in the long run, even though it is paying higher wages than its competitors. The company may have some other cost advantage; its competitors may price above their costs; the market may be expanding rapidly. The company will grow less rapidly than if its costs were lower and may stagnate or decline, but it need not die. There is thus no contradiction in a company's stating on the one hand that it is profitable and on the other hand that its costs are higher than its competitors' and it wants to reduce them. The Board concedes that if this is all Nielsen said, Nielsen had no duty to open its books to the union. See Washington Materials, Inc. v. NLRB, 803 F.2d 1333, 1338-39 (4th Cir.1986) (but see NLRB v. Western Wirebound Box Co., 356 F.2d 88, 91 (9th Cir.1966), which held that an employer's claim of competitive disadvantage required the employer, on the union's demand, to present substantiating data). A need is objective; it can be substantiated. But how do you substantiate a want? If a company says it wants to make higher profits by reducing its labor costs, what data would falsify its statement?

The Board found, however, that Nielsen had done more than express a desire for lower costs and higher profits; that it had said the wage cuts were necessary if the company was to remain competitive and reverse a trend of losing business to lower-cost competitors. The company's president had told the workers that "to survive we must be able to compete. Our business ... and jobs are at stake if we can not.... If we don't [compete] the recent trend of losing even greater amounts of work to other companies will continue and the jobs of our employees will be lost." The Board held that this statement, and others like it, were sufficient under Truitt to create a duty of substantiation. Cf. Armored Transport of Calif., Inc., 288 N.L.R.B. No. 70 (1988).

This is not an irrational extension of Truitt. A rational businessman is concerned with the long run as well as the short run. He wants to maximize the present value of his company's earnings; and if the company has a dismal future, its expected future earnings, and hence its present value, will be depressed as a result. An employer that in negotiations with its union claims that its wages are out of line with those of its competitors and as a result its future is bleak--however rosy the present may seem--makes a serious and factual claim, one that if true must cause the union to give serious consideration to making the concessions the employer is demanding, or at least making some concessions. Informed bargaining over the issue requires that the union have access to the data from which the company has projected its bleak future.

So at least the Board could reason within the analytical framework established in Truitt--so it did reason in this case. The problem is that right after it ruled in favor of the union we decided NLRB v. Harvstone Mfg. Corp., 785 F.2d 570 (7th Cir.1986), a case similar to the present one, against the Board. Following earlier circuit precedent (United Fire Proof Warehouse Co. v. NLRB, 356 F.2d 494, 498 (7th Cir.1966)), we held that predictions that a business will falter--even that it will close--are "nothing more than truisms," id. at 577, and do not trigger the duty of disclosure under Truitt, a duty that we deemed limited to inability to pay during "the term [ordinarily three years] of the new collective bargaining agreement" being negotiated, id.

The Labor Board in its decision in this case had actually relied on the very decision (Harvstone Mfg. Corp., 272 N.L.R.B. 939 (1984)) that we reversed. Nielsen therefore asked for reconsideration but the Board refused, saying: "Respondent filed with the Board a motion for reconsideration contending ... that the Board's order ... directly conflicts with a certain decision of the United States [Court of] Appeals for the Seventh Circuit. The Board is of the opinion that the Respondent does not assert any matter not previously considered and that the Respondent's contentions are without merit." Period.

The decision alluded to so coyly was, of course, our Harvstone decision. How the Board could say that the motion for reconsideration did "not assert any matter not previously considered" by the Board perplexes us, since we did not think the Board gifted with prevision and therefore able to read and evaluate our decisions before they are rendered. In any event there is no reasoned discussion of our decision.

Yet if either in the present case or in Harvstone the Board had anticipated the substance of the analysis in our opinion in Harvstone, then perhaps its statement in the present case that the petition for reconsideration "did not assert any matter not previously considered" by the Board could be interpreted to mean that our opinion contained no argument that the Board had not considered. But a comparison of the Board's opinions in this case and in Harvstone with our opinion in Harvstone (see 785 F.2d at 575-77) scotches any such contention--which anyway the Board does not make: at argument the Board's counsel speculated that the Board had not discussed our case only because it didn't think Nielsen would seek judicial review in this circuit. This, if so, was careless, given the wide venue (of which more shortly) for proceedings to review the Board's decisions. In any event counsel's speculation about the Board's motives is not especially plausible, since even if the Board had known that Nielsen was going to the Seventh Circuit it might have ignored our decision in Harvstone all the same; its "flagrant disregard of judicial precedent" is well known. Mary Thompson Hospital, Inc. v. NLRB, 621 F.2d 858, 864 (7th Cir.1980); see also NLRB v. Ashkenazy Property Management Corp., 817 F.2d 74 (9th Cir.1987) (per curiam); Yellow Taxi Co. v. NLRB, 721 F.2d 366, 382-83 (D.C. Cir.1983); Kitchen Fresh, Inc. v. NLRB, 716 F.2d 351, 357 (6th Cir.1983); Ithaca College v. NLRB, 623 F.2d 224, 228-29 (2d Cir.1980); Allegheny General Hospital v. NLRB, 608 F.2d 965, 968-70 (3d Cir.1979).

Granted, "flagrant disregard of judicial precedent" must not be confused with refusing to knuckle under to the first court of appeals (or the second, or even the twelfth) to rule adversely to the Board. See United States v. Mendoza, 464 U.S. 154, 104 S. Ct. 568, 78 L. Ed. 2d 379 (1984). The Supreme Court, not this circuit or even all twelve circuits that have jurisdiction to review orders of the Labor Board, is the supreme arbiter of the meaning of the laws enforced by the Board--a precept especially apt given the extraordinarily broad venue for proceedings to review Board orders, which virtually invites forum shopping of the kind exhibited in this case, compare NLRB v. A. Duie Pyle, Inc., 730 F.2d 119, 128 (3d Cir.1984), and may in consequence cause a skewing of precedents at the circuit level. Although no circuit has rejected our decision in Harvstone, no circuit has followed it either (it is only two years old), and there are previous cases, besides Western Wirebound, that agree with the Board's position. See United Steelworkers of America v. NLRB, 401 F.2d 434, 436 (D.C.Cir.1968); International Telephone & Telegraph Corp. v. NLRB, 382 F.2d 366, 370-71 (3d Cir.1967). This circuit is not authorized to interpret the labor laws with binding effect throughout the whole country, and the Board therefore is not obliged to accept our interpretation.

So the Board could have decided to stick by its guns and not bow to our decision in Harvstone--though it could also have defeated the company's effort at forum shopping simply by filing its petition to enforce its order in the Sixth Circuit (the natural venue for this case) and asking us to transfer the company's petition to review the order to that circuit, pursuant to 28 U.S.C. Sec. 2112(a). See Farah Mfg. Co. v. NLRB, 481 F.2d 1143, 1145 (8th Cir.1973) (per curiam). We doubt whether, as some courts have suggested, see id. at 1145; U.S. Electrical Motors v. NLRB, 722 F.2d 315, 319 (6th Cir.1983) (per curiam), we also have a nonstatutory "inherent" power to transfer a petition to review to another circuit, in order to defeat forum shopping. But that issue need not be decided here. The Board has not asked us to exercise such power; more to the point, it was content to file its petition for enforcement in this circuit, after the company had filed its petition for review here. The union, it is true, asked us to transfer the case to the Sixth Circuit. But the union filed the motion late; the Board had already acquiesced in review by this court; and since no petition for review or enforcement had been filed in the Sixth Circuit, the union was relying solely on our supposed inherent power to transfer petitions to review where necessary to prevent forum shopping.

If the Board wanted to stand pat on its Harvstone decision and try to take a decision by us adhering to our Harvstone decision to the Supreme Court, it could have done so merely by stating in response to Nielsen's petition for reconsideration that it had read our decision and disagreed with it. Instead it refused to acknowledge the pertinence of our decision to this case--and this despite the fact that the Board itself has been inconsistent with regard to the issue presented in this case; compare its decision in Harvstone to its decision in Washington Materials, Inc., supra, 276 N.L.R.B. at 840-41, which is consistent with our Harvstone decision, but not the Board's.

The decisions we cited earlier that accuse the Board of "flagrant disregard of judicial precedent" were not accusing the Board of refusing to accept circuit precedent as binding, but of dealing with judicial precedent in a disingenuous, evasive, and in short dishonest manner. The Board's response to our Harvstone decision must be regarded in the same light. Faced with a conflict both in the circuits and its own decisions, the Board had the duty to take a stance, to explain which decisions it agreed with and why, and to explore the possibility of intermediate solutions. (For it is possible that the facts of this case place it halfway between Truitt and Harvstone. The company's ambiguous statements could be interpreted as a veiled threat of layoffs in the near if not the immediate future, thus upping the ante compared to Harvstone and perhaps bringing the case within the gravitational field of Truitt.) We do not follow stare decisis inflexibly; if the Board gives us a good reason to do so, we shall be happy to reexamine Harvstone. It has not done so. It has acted arbitrarily in its treatment of administrative as well as judicial precedent--and this is a pattern, too, see, e.g., Continental Web Press, Inc. v. NLRB, 742 F.2d 1087 (7th Cir.1984).

The petition for review is granted, the Board's order is set aside, and the matter is returned to the Board for such further proceedings as may be consistent with this opinion. We need not determine our course if the Board declines to follow Harvstone, its decision is again appealed to us and again reversed, and the Board on the third round decides to continue to stand fast; but presumably our present decision would stand as the law of the case in any subsequent proceeding in this court. See Frattaroli v. NLRB, 590 F.2d 15, 17 (1st Cir.1978).