concurring:
I concur in the court’s judgment. I do not think, however, that we are actually bound by the “law of the case” to affirm the Authority’s adoption of the “vitally affects” test, still less its application of that test in these cases. In our prior opinions, we instructed the Authority to justify its failure to adopt the “vitally affects” test used in the private sector. See American Federation of Government Employees, Local 32 v. FLRA, 853 F.2d 986, 992-93 (D.C.Cir.1988); American Federation of Government Employees, Local 32 v. FLRA, 774 F.2d 498, 503-04 (D.C.Cir.1985). But we did not actually hold that it was obliged to embrace that standard. Nor do I believe that we lack “jurisdiction” to consider petitioners’ arguments on the ground that they did not intervene earlier to urge the FLRA to maintain its position in the face of this court’s not so subtle prodding. Technically, petitioners were not aggrieved until after our second remand, when the FLRA changed its position and declared negotiable the proposals of the American Federation of Government Employees, Local 32 and the National Treasury Employees Union. See 33 F.L.R.A. 335 (1988).1 But, after two opinions of our court have pushed the FLRA into exactly the position it has adopted, I think it would not only be institutionally unseemly for a new panel to switch course, I do not think we are free to do so because we must defer under Chevron, U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984), to the agency’s reasonable interpretation of its statute, which imposes on management officials of federal agencies a duty to bargain with their employees’ unions regarding “conditions of employment.” See 5 U.S.C. § 7103(a)(12). I write separately, nevertheless, because I think that we exceeded our legitimate scope of review in those two previous panel opinions — certainly in tone. I also believe that the substantive rule of labor law we have pressed the FLRA into accepting has dubious applicability to the federal sector.
The problem this case presents — our pri- or opinions to the contrary — does not really *331have its analog in the private sector. That is so because the Federal Labor Relations Act imposes a number of limits on negotiability of proposals that deal with working conditions of employees not found under the National Labor Relations Act. See Fort Stewart Schools v. FLRA, — U.S. -, -, 110 S.Ct. 2043, 2047, 109 L.Ed.2d 659 (1990); 5 U.S.C. §§ 7103(a)(14)(C), 7106(a) and 7117(a). One of those limitations is that government agencies and unions are not free to bargain over proposals that are inconsistent with government-wide regulations. See 5 U.S.C. § 7117(a)(1). In other words, theoretically at least, the Executive Branch can take virtually any proposal off the bargaining table by adopting a government-wide regulation covering the subject (which often makes government petitions for review to the courts of appeal from FLRA decisions seem rather unnecessary — unless one views the time of Article III judges as somewhat expendable).
Here we face a government-wide regulation that bears on the negotiability of the union’s proposal. OPM’s regulations dealing with the enormously sensitive subject of reductions in force (RIF) provide that employees compete — based on factors such as tenure, grade level, and veterans preferences — against each other within a “competitive area.” That concept is definable by an agency in organizational and geographical terms but — and this is the nub of the problem — the competitive area “must include all employees within the competitive area so defined.” 5 C.F.R. § 351.402(b). Thus, although the agency has limited discretion to pick the bureau, major command, directorate, or other equivalent major subdivisions of an agency within the local “competitive area,” see id., whatever definition it chooses will cover all employees within the area — those in different bargaining units, as well as non-union employees. Because of OPM’s regulation, the agency may not agree to separate competitive areas for different bargaining units. And, therefore, if an agency were to agree to a union’s proposal for a definition of “competitive area,” it would perforce cover employees not represented by the union and, presumably, employees with different interests.
This sort of problem simply does not arise in the private sector because no private employer has the authority to impose uniformity of any aspect of wages, hours, or working conditions among all employees, union and non-union, or among different bargaining units. Under the National Labor Relations Act, a somewhat different issue involving an employer’s obligation vel non to bargain over matters that have impact on interests outside any employment relationship has surfaced. In Allied Chemical and Alkali Workers of America v. Pittsburgh Plate Glass Co., 404 U.S. 157, 92 S.Ct. 383, 30 L.Ed.2d 341 (1971), the Supreme Court held that an employer did not have to bargain with its union over health benefits of retired employees (who were no longer employees) because the issue did not have a “vital effect” on bargaining unit employees. It was thought speculative whether the interests of the bargaining unit's medical costs would be affected positively or negatively by the union’s effort to extend its bargaining to retirees. Although the opinion is unclear on this point, it seems that the court was also concerned over a conflict of interest between active and retired employees. In any event, the case gave rise to the notion that, under the National Labor Relations Act, an employer is obliged to bargain with a union representing bargaining unit employees over an issue directly affecting non-employees if the question “vitally affects” the bargaining unit. It has never, as far as I can tell, been applied to a situation where two groups of employees of the same employer — bargaining unit and non-bargaining unit (or union and non-union) — are affected by a union negotiating proposal.2
*332As the majority describes, we remanded this case twice because the Authority did not adequately explain what appeared to us to be its inconsistent treatment of two cases involving the question of the negotiability of union proposals relating to competitive areas. But we went further than simply asking for a reconciliation of the two cases; we quite clearly indicated that we disapproved of the Authority’s analysis and seemingly directed the Authority — in the context of asking for an adequate explanation — to adopt the “vitally affects” test drawn from Pittsburgh Plate Glass. The Authority had instead sought to adopt a test which focused on the interests of the non-bargaining unit employees. A proposal — such as one that sought to define the competitive area — was non-negotiable if
the nature and degree of the impact of the proposal is so intrinsically related to the working conditions of nonunit employees so as to invade the purview of other unit representatives or require the agency to act in a way that will have a significant effect on the rights of employees not represented by the union offering the disputed proposal.
22 F.L.R.A. 478, 482 (1986).
We, in effect, disapproved of that standard and broadly suggested that the Authority adopt a test for negotiability (the “vitally affects” test) which emphasizes the interests of bargaining unit employees and ignores the concerns of other employees who are swept up by the proposed clause. To be sure, it can be argued that the Supreme Court in Pittsburgh Plate Glass, its only application of the “vitally affects” test, actually did decisively weigh the interests of the affected non-employees (the retired persons), and therefore the Authority could have adopted the “vitally affects” test and still determined that the proposals were not negotiable. But the tenor of our prior opinions certainly pointed the FLRA in the other direction — that an application of that test would result in a determination that the proposals were negotiable. As should be apparent, I believe we went too far in twice remanding to the Authority. Our language was, in my view, too unneutral for a remand for inadequate explanation — particularly the second time. That the majority treats the Authority’s adoption of the “vitally affects” test — and its application to these facts — as the “law of the case” makes my point; that formulation necessarily suggests that the FLRA was obliged to adopt our interpretation of the FLRA. The Authority understandably, therefore, seems to have understood our prior opinions as more a direction than a request for explanation. I regret the need to criticize the prior panel opinions, but in a post-Chevron era (requiring heightened deference to agency interpretation of imprecise legislative language) we must take care not to cajole agencies, through remand for inadequate explanations, into adopting legislative interpretations we lack authority to require.3 See Atari Games Corp. v. Oman, 888 F.2d 878, 887 (D.C.Cir.1989) (Silberman, J., concurring in the judgment). And agencies should maintain the courage of their convictions — although a little hard thinking would help as well.
Unfortunately, in this case, we pushed the Authority into acceptance of a National Labor Relations Act proposition — which itself is not all that crystalline — developed to deal with quite different kinds of issues. Now that the Authority has adopted as its own interpretation of the FLRA the very construction of its statute we urged upon it, we have no choice but to defer, even if we overreached in our prior opinions and the FLRA’S assimilation of our view is not all that persuasive. Perhaps the government could have headed off this result if it had bestirred itself previously — but it is too late now. I take solace in supposing that OPM can cure the problem by merely modifying its government-wide regulations.
. Thus, petitioners had no “objection” to urge to the FLRA until then. And, by that time, the Authority was well apprised of the issue whether the "vitally affects" test applied to the cases before it. See American Federation of Government Employees v. FLRA, 878 F.2d 460, 466 (D.C.Cir.1989).
. In Teamsters Union v. Oliver, 358 U.S. 283, 79 S.Ct. 297, 3 L.Ed.2d 312 (1959) and Fibreboard Paper Products v. NLRB, 379 U.S. 203, 85 S.Ct. 398, 13 L.Ed.2d 233 (1964), referred to by the Supreme Court in Pittsburgh Plate Glass, employers were held to be required to bargain over their subcontracting outside the bargaining unit and outside the employment relationship, which *332had a negative impact on bargaining unit employees.
. The majority "resists" my claim that the prior panels exceeded the appropriate scope of review, see Maj.Op. at 330 n. 1, but does not dispute it other than to defend the panels' motive — which, of course, I do not impugn.