This case raises a question left in the wake of Immigration & Naturalization Service v. Chadha, 462 U.S. 919, 103 S.Ct. 2764, 77 L.Ed.2d 317 (1983), invalidating as violative of separation of powers principles the device of the legislative veto of administrative regulations. The context is Congress' pre-Chadha deregulation of the airline industry. The specific issue is whether an unconstitutional legislative veto provision contained in section 43(f) of the Airline Deregulation Act of 1978, Pub.L. No. 95-504, Sec. 43(f), 92 Stat. 1705, 1750 (1978), is severable from the remainder of the airline employee protection provisions of that statute. We conclude that the veto provision should be severed; we therefore reverse the judgment of the District Court, 594 F.Supp. 92, which held the veto provision inseverable. Inasmuch as the District Court did not have occasion to address the other issues raised by the airlines, we remand the case for further consideration.

* The Airline Deregulation Act of 1978 contains an employee protection program for the benefit of employees displaced through the impact of deregulation on the shifting fortunes of individual air carriers. See id. Sec. 43 at 92 Stat. 1750, codified at 49 U.S.C. app. Sec. 1552 (1982). As originally crafted, the program contained two prongs. The first provided monetary support and assistance in relocation to individuals who had lost their employment or suffered a diminution in compensation as a result of a qualifying industry dislocation. See 49 U.S.C. app. Sec. 1552(a)-(c). While that provision is still on the books, Congress has never funded its implementation; inasmuch as that provision's operation was expressly made subject to Congress' appropriation of funds for financial assistance, see 49 U.S.C. app. Sec. 1552(a), this portion of the employee protection program is inoperative and, accordingly, no implementing regulations have been promulgated under it.

The employee protection program's second prong triggered the litigation now before us. That portion consists of a "first hire" requirement. That is to say, a person who had been employed for at least four years prior to October 24, 1978 (the effective date of the Act) by an air carrier holding a certificate under the pre-deregulation regime and who was furloughed or terminated (other than for cause) prior to the Act's effective date had a first right of hire by other pre-deregulation certificated air carriers. See 49 U.S.C. app. Sec. 1552(d), (h)(1). This right of first hire was without regard to age but applied only to airlines hiring within the individual's occupational specialty; moreover, airlines could lawfully recall their own furloughed employees before hiring those displaced from other carriers. See 49 U.S.C. app. Sec. 1552(d).

The Secretary of Labor was granted authority to issue, amend and repeal rules and regulations necessary to administer the employee protection plan. See 49 U.S.C. app. Sec. 1552(f)(1), (h)(3). However, that authority was limited by the following provisions:

The Secretary shall not issue any rule or regulation as a final rule or regulation under this section until 30 legislative days after it has been submitted to the Committee on Commerce, Science, and Transportation of the Senate and the Committee on Public Works and Transportation of the House of Representatives. Any rule or regulation issued by the Secretary under this section as a final rule or regulation shall be submitted to the Congress and shall become effective 60 legislative days after the date of such submission, unless during that 60-day period either House adopts a resolution stating that that House disapproves such rules or regulations, except that such rules or regulations may become effective on the date, during such 60-day period, that a resolution has been adopted by both House stating that the Congress approves of them.

49 U.S.C. app. Sec. 1552(f)(3). Exceptionally, this section combines the "report and wait" provision (found in the first sentence of the quoted language) with a one-House legislative veto1 (set forth in the second sentence).

Pursuant to his statutory authority, the Secretary of Labor published regulations for the administration of both prongs of the protection plan in March 1979. See 44 Fed.Reg. 19,146 (1979). A revision of the proposed regulations, covering only the first hire provisions, was published in September 1982. See 47 Fed.Reg. 41,304 (1982). Final regulations were published on November 22, 1983, see 48 Fed.Reg. 52,854 (1983), were duly transmitted to Congress, and were to become effective after sixty legislative days.

In February 1984, however, before the regulations became effective, Alaska Airlines, Inc. and various other airlines filed a complaint in United States District Court for the District of Columbia. They alleged that the first hire provisions of the Act were invalid because of the inclusion of an unconstitutional legislative veto in section 43(f)(3). In addition, they challenged the regulations themselves as violative of the Due Process Clause and, on non-constitutional grounds, as arbitrary and capricious, an abuse of agency discretion and otherwise not in accordance with law. The Department of Labor conceded below, see Alaska Airlines, Inc. v. Donovan, 594 F.Supp. 92, 94 (D.D.C.1984), and does not contest here, that the legislative veto provision is unconstitutional under Chadha. The Government argued, however, that the invalid veto provision is severable from the remainder of the Act and that the remainder of the employee protection plan, and the rules and regulations adopted thereunder, are lawful and valid.

Plaintiffs-appellees moved for summary judgment, and the Department of Labor moved to dismiss or for an order affirming the validity of the regulations. On May 17, 1984, the day the regulations were to become effective, the District Court granted the airlines' motion, holding that section 43 was invalid in its entirety due to the inclusion of what the court deemed to be an inseverable legislative veto provision. The court's opinion issued the following day. Neither the order nor the opinion addressed the other issues raised in the airlines' complaint.

Before reaching the principal issue presented on this appeal, namely the severability of the legislative veto from the remainder of section 43 of the Airline Deregulation Act, several preliminary issues must be addressed.

* The first question is whether we have jurisdiction over this appeal. The issue is raised only in a footnote in the Alaska Airlines Brief, see Appellees' Brief at 9 n. *, where appellees question, in passing, whether jurisdiction in fact lies in this court; because of the obvious importance of that question, the subject merits our analysis at the outset. A basic jurisdictional provision of Title 28 is that the Supreme Court, not the Courts of Appeals, is vested with jurisdiction over appeals from District Court decisions holding an Act of Congress unconstitutional in any civil action in which the United States, or an agency, official or employee thereof, is a party. 28 U.S.C. Sec. 1252 (1982). In keeping with this provision, the several Courts of Appeals enjoy jurisdiction over appeals from final decisions of the district courts "except where a direct review may be had in the Supreme Court." 28 U.S.C. Sec. 1291 (1982) (emphasis added). The airlines suggest that since the District Court held the entire section unconstitutional, exclusive review of that judgment should lie in the Supreme Court.

Those two Title 28 provisions might be read to support the airlines' position. In fact, the appeal in this action was originally filed in the Supreme Court, but due to a recent Supreme Court decision was thereafter refiled in this court. See Appellant Secretary of Labor's Brief at 10 n. 5. To determine whether the shift in tribunals was proper and jurisdiction actually lies here we examine that recent Supreme Court disposition.

In Equal Employment Opportunity Commission v. Allstate Insurance Co., --- U.S. ----, 104 S.Ct. 3499, 82 L.Ed.2d 810 (1984), the Court, by order, held that it had no jurisdiction in a legislative veto case. The obvious difficulty in analyzing the Supreme Court's decision in Allstate is that the dismissal for lack of jurisdiction was without opinion. There was, however, a dissent by the Chief Justice, joined by Justice O'Connor, which presents at least two Justices' view of the Court's reasoning. The district court in that case, 570 F.Supp. 1224 (S.D.Miss.1983), had held the Reorganization Act of 1977, Pub.L. No. 95-17, 91 Stat. 29 (1977), unconstitutional because of the presence of a non-severable legislative veto provision. That statute contained the provision pursuant to which the Equal Employment Opportunity Commission had been authorized to enforce the Equal Pay Act, Pub.L. No. 88-38, 77 Stat. 56 (1963). After ruling that the Reorganization Act failed in its entirety, the court held that the EEOC lacked authority to enforce the Equal Pay Act and granted summary judgment in favor of Allstate.2

An appeal thereafter taken directly to the Supreme Court was summarily dismissed for want of jurisdiction. Appearing to find the Court's approach overbroad, the Chief Justice stated:

Had the District Court simply determined, as a matter of statutory construction, that appellant cannot exercise the authority to enforce the Equal Pay Act because Congress did not wish the Act to be operative absent the veto provision, I would agree that dismissal would clearly be compelled under Heckler v. Edwards [465 U.S. 870, 104 S.Ct. 1532, 79 L.Ed.2d 878 (1984) ]. Under these circumstances, it would be clear, as in Heckler, that appellant was not challenging the constitutional holding of the District Court since it concedes the validity of the court's holding on the legislative veto provision. But this is not the holding of the District Court. And the mere assertion now by the Solicitor General, and implicitly by the Court, that the holding was in reality one of statutory construction cannot change the fact that the District Court explicitly and unambiguously held the entire Act unconstitutional.

Allstate, supra, 104 S.Ct. at 3502 (emphasis in original).

Using this explanation as guidance to the Court's position on the jurisdictional issue, it appears that the appeal in the case at hand appropriately lies in this court. The District Court here did not break any new ground with respect to a constitutional issue. To the contrary, its decision that the legislative veto provision was unconstitutional was squarely rooted in Chadha, as evidenced by the fact that neither side contests that portion of the District Court's decision. Thus, the constitutional decision is, in truth, not a part of this appeal. What is at issue on appeal, rather, is the severability vel non of the remainder of the section in which the legislative veto appears. That question, as we shall see, reduces to a matter of statutory interpretation, see generally Stern, Separability and Separability Clauses in the Supreme Court, 51 HARV.L.REV. 76, 115 (1937). As even the Allstate dissenters opined, jurisdiction would not under those circumstances lie in the Supreme Court but would instead be vested in this court. See also Heckler v. Edwards, supra, 104 S.Ct. at 1536 ("a party does not have a right to direct review in the Supreme Court under 28 U.S.C. Sec. 1252 unless the holding of federal statutory unconstitutionality is in issue"). Accordingly, we are persuaded that the present appeal is properly before us.

The second preliminary issue is whether this dispute is ripe for review, inasmuch as Congress has not attempted to exercise the legislative-veto power purportedly conferred by the Airline Deregulation Act. In a case suggesting the possibility of a lack of ripeness, Clark v. Valeo, 559 F.2d 642 (D.C.Cir.), aff'd sub nom. Clark v. Kimmitt, 431 U.S. 950, 97 S.Ct. 2667, 53 L.Ed.2d 267 (1977), this court in a pre-Chadha ruling stated that "[u]ntil Congress exercises the one-house veto, it may be difficult to present a case with sufficient concreteness as to standing and ripeness to justify judicial resolution of the pervasive constitutional issue which the one-house veto provision involves." Id. at 649. However, this court also stated, importantly for our purposes, that "[a] contention that there are no real considerations of ripeness here can only rest on a view of the merits that a one-house veto is so patently unconstitutional that nothing more is needed to inform the judgment of the court." Id. at 649 n. 8. This court, at that time, was unwilling to deem the legislative veto device patently unconstitutional; however, in Chadha 's wake, this sort of veto provision is manifestly unconstitutional, and thus under Clark v. Valeo 's teaching the issue before us may well be ripe for adjudication.

On the other hand, the proposition that the issue is not ripe finds its strongest support in a decision from another circuit, Muller Optical Co. v. Equal Employment Opportunity Commission, 743 F.2d 380 (6th Cir.1984). In that case, the legislative veto provision of the Reorganization Act of 1977 was once again in issue. The Sixth Circuit concluded that "the existence of a one-House veto provision in a statute does not render the statute invalid but only renders the act of Congress, if it attempts to exercise its one-House veto[,] invalid...." Id. at 388. The Reorganization Plan No. 1 of 1978, which transferred authority to the EEOC to enforce the Age Discrimination in Employment Act, see infra note 5, had long since been reviewed by Congress, and Congress had chosen not to exercise its legislative veto. The court stated:

To inquire at this point in time whether Congress would have passed the Reorganizational [sic] Act without the one-House veto provision and, thereby, whether the substantive provisions of the Act are effective absent the provision, makes no sense when Congress has already declined to use the veto provision and thus has already approved the plan.

Id.

In Muller, the Reorganization Plan in question was promulgated, and Congress elected not to veto it, prior to Chadha. Congress thus tacitly approved the plan, thereby strengthening the argument that the constitutionality vel non of the unutilized veto provision should not have been adjudicated. Here, however, the Department of Labor promulgated its regulations after Chadha was handed down. At that point, Congress did not, in truth, enjoy the option of vetoing the rules; indeed, to have attempted to do so, in light of Chadha, would have been an exercise in futility. Hence, the Congressional approval found tacitly in Muller cannot, in reason, be read into the situation at hand. In the post-Chadha era in which Congress knows that it may not lawfully exercise the purported veto authority, to hold that a court cannot reach the issue of the constitutionality of such a device unless Congress has exercised the provision would be, in effect, to shield all agency action from review as to whether Congress would have provided the underlying authority without its veto oversight. This we decline to do.

The final preliminary question is whether Chadha should be applied retroactively. The test for retroactivity of a judicial decision is laid out in Northern Pipeline Construction Co. v. Marathon Pipe Line Co., 458 U.S. 50, 102 S.Ct. 2858, 73 L.Ed.2d 598 (1982), and Chevron Oil Co. v. Huson, 404 U.S. 97, 92 S.Ct. 349, 30 L.Ed.2d 296 (1971). Northern Pipeline states that there are

three considerations ... properly bearing upon the issue of retroactivity. They are, first, whether the holding in question "decid[ed] an issue of first impression whose resolution was not clearly foreshadowed" by earlier cases, ... second, "whether retrospective operation will further or retard [the] operation" of the holding in question, ... and third, whether retroactive application "could produce substantial inequitable results" in individual cases....

Northern Pipeline, supra, 458 U.S. at 88, 102 S.Ct. at 2880 (quoting Chevron, supra, 404 U.S. at 106-07, 92 S.Ct. at 355).

With regard to the first factor, the Association of Flight Attendants argues before us that Chadha erected a new principle of law that could not reasonably have been anticipated. In response, the airlines point to fifty years of political and scholarly debate and speculation as to the constitutionality of such provisions; over the years various commentators have questioned the constitutionality of the legislative veto device,3 while others have rushed to its defense.4 Furthermore, this court in Consumer Energy Council of America v. FERC, 673 F.2d 425 (D.C.Cir.1982), aff'd mem. sub nom. Process Gas Consumers Group v. Consumer Energy Council of America, 463 U.S. 1216, 103 S.Ct. 3556, 77 L.Ed.2d 1402, 77 L.Ed.2d 1403, 77 L.Ed.2d 1413 (1983), held prior to Chadha that the legislative veto provision of the Natural Gas Policy Act of 1978, Pub.L. No. 95-621, 92 Stat. 3350 (1978), was unconstitutional. In view of all this, at the very least, some considerable doubt had been raised prior to Chadha with respect to the constitutionality of legislative veto provisions.

The second factor, the furtherance or retarding of the decision's operation, speaks more strongly in favor of Chadha 's retroactive application. That decision, of course, pertained to a bedrock issue of the structure of the National Government, namely the separation of powers enshrined by the Framers in Philadelphia in the summer of 1787. The fundamental concern over fidelity to that principle of governance of a free people would scarcely be furthered by shielding from challenge a separation-of-powers violation simply by virtue of the fact that the unconstitutional provision was passed prior to Chadha. While it may be argued that regulations promulgated and not vetoed (or even promulgated but vetoed) prior to Chadha should not be subjected to retroactive application, it seems strained to argue that the provision itself should be protected from Chadha 's retroactive application. The veto device works a potential violation of separation of powers principles; that provision indisputably ripens into a very real violation when regulations are purportedly vetoed. An alternative problem arises in this post-Chadha day and age in which Congress is plainly forbidden from exercising a veto. If we refuse to apply Chadha to the post-Chadha adoption of regulations under a pre-Chadha statute that includes a legislative veto, the Executive branch might thereby be permitted to exercise power that Congress would not have delegated had it not retained some power of review through the legislative veto device. In short, while the second factor might shield from retroactive application those regulations fully in place prior to Chadha, that factor should not shield the legislative veto provisions themselves from constitutional scrutiny.

The third factor, substantial inequitable results from retroactive application, weighs heavily on neither side. The Association of Flight Attendants argues that great inequity would result from Chadha's retroactive application in that employees would no longer enjoy the protective mechanisms carefully crafted by a Congress which recognized "an almost moral obligation" to protect airline employees. See Association of Flight Attendants' Brief at 41. However, whatever inequity the flight attendants may discern arises not entirely from the retroactive application of Chadha but, more precisely, from any judicial decision not to sever the remainder of section 43. Their arguments speak to the Congressional intent important to the severance decision; they do not speak to the retroactivity vel non of Chadha.

There is, as yet, little case law on Chadha 's retroactive application.5 Three pre-Chadha cases in this circuit struck down legislative veto provisions as unconstitutional, see American Federation of Government Employees v. Pierce, 697 F.2d 303 (D.C.Cir.1982); Consumers Union of the United States, Inc. v. Federal Trade Commission, 691 F.2d 575 (D.C.Cir.1982), aff'd mem. sub nom. Process Gas Consumer Group v. Consumer Energy Council of America, 463 U.S. 1216, 103 S.Ct. 3556, 77 L.Ed.2d 1402, 77 L.Ed.2d 1403, 77 L.Ed.2d 1413 (1983); Consumer Energy Council of America v. FERC, supra, 673 F.2d 425. Inasmuch as all three decisions were pre-Chadha, they are not of direct relevance to the issue of the retroactivity of Chadha. Furthermore, while Pierce relies on Consumer Energy Council and Consumers Union, the veto clause found unconstitutional in Pierce was enacted after this court's decision in Consumer Energy Council, hence the reliance on Consumer Energy Council was not a retroactive application. On the other hand, the Supreme Court's summary affirmance of Consumers Union and Consumer Energy Council was rendered after Chadha. That summary disposition, while obviously enjoying less precedential significance than a case afforded plenary consideration, see, e.g., Mandel v. Bradley, 432 U.S. 173, 17677 (1977); Fusari v. Steinberg, 419 U.S. 379, 391-92, 95 S.Ct. 533, 540-41, 42 L.Ed.2d 521 (1975) (Burger, C.J., concurring), might be viewed as assuming Chadha 's retroactive application.

The airlines advance one additional argument in favor of retroactive application. They note observations in both Justice Powell's concurrence in Chadha, supra, 462 U.S. at 959, 103 S.Ct. at 2788 ("The Court's decision ... apparently will invalidate every use of the legislative veto.") (Powell, J., concurring), and Justice White's dissent, id. at 967, 103 S.Ct. at 2792 ("Today the Court ... sounds the death knell for nearly 200 other statutory provisions in which Congress has reserved a 'legislative veto.' ") (White, J., dissenting), that the Court's ruling would invalidate all legislative veto provisions. The airlines emphasize that the opinion of the Court did not respond to these observations, while it did respond to others offered by the dissent; the airlines draw from this an indication that the Court has indicated that Chadha should have retroactive effect. This is a weak reed on which to rest. Failure to respond to these comments is, at bottom, probative of nothing. Nonetheless, when all is said and done, we agree, by virtue of the analysis previously set forth, that Chadha should be given retroactive effect.

Having disposed of these preliminary inquiries, we turn now to the question of the severability vel non of the remainder of section 43 from the unconstitutional legislative veto provision.6 Since it is, of course, a veto provision whose severability is in question, the severability analysis set forth in Chadha itself is particularly instructive for us. In Chadha, the Court stated that "the invalid portions of a statute are to be severed ' "[u]nless it is evident that the Legislature would not have enacted those provisions which are within its power, independently of that which is not." ' " Chadha, supra, 462 U.S. at 931-32, 103 S.Ct. at 2774 (quoting Buckley v. Valeo, 424 U.S. 1, 108, 96 S.Ct. 612, 677, 46 L.Ed.2d 659 (1976) (quoting Champlin Refining Co. v. Corporation Comm'n, 286 U.S. 210, 234, 52 S.Ct. 559, 564, 76 L.Ed. 1062 (1932)). "A provision is further presumed severable if what remains after severance 'is fully operable as a law.' " Chadha, supra, 462 U.S. at 934, 103 S.Ct. at 2775 (quoting Champlin, supra, 286 U.S. at 234, 52 S.Ct. at 564). While the immigration statute at issue in Chadha contained a severability clause, thus distinguishing it from the Airline Deregulation Act before us,7 the analytical framework enunciated by Chadha applies to the severability issue which we are called upon to resolve.

Chadha presumes severability of the offensive provision if what remains after severance is fully operable as law. To this end, the burden is placed squarely on the party arguing against severability to demonstrate that Congress would not have enacted the provision without the severed portion.8 This presumption articulated by Chadha was echoed in a later Supreme Court severability case: "Whether an unconstitutional provision is severable from the remainder of the statute in which it appears is largely a question of legislative intent, but the presumption is in favor of severability." Regan v. Time, Inc., --- U.S. ----, 104 S.Ct. 3262, 3269, 82 L.Ed.2d 487 (1984). The Court has thus clearly laid down a presumption in favor of severability if what remains is operative as law; in a word, severance lies unless it is "evident" that what remains would not have been enacted.

Our charge is to save as much of the statute as we can,9 consistent of course with the underlying legislative intent. Only if we conclude that Congress would not have included a provision absent the constitutionally flawed portion is that provision to fall. The issue cannot be whether Congress preferred the statute with the unconstitutional provision over the same statute without that provision. Manifestly, Congress' preference is abundantly clear from its inclusion of the unconstitutional provision. Nor is the question whether Congress would have passed some alternative version of the statute if it knew that it could not lawfully have included the offending provision. That is, "the question is not whether Congress would have enacted th[is] exact statute[ ] had it known at the time of enactment that the legislative veto provisions were invalid, but rather, whether Congress would have preferred th[is] statute[ ], after severance of the legislative veto provision[ ], to no statute[ ] at all." Gulf Oil Corp. v. Dyke, 734 F.2d 797, 804 (T.E.C.A.) (emphasis in original), cert. denied, --- U.S. ----, 105 S.Ct. 173, 83 L.Ed.2d 108 (1984).

Section 43 of the ADA is an extensive and elaborate provision, yet only one part of one subsection, namely section 43(f)(3), is unconstitutional under Chadha. That section, codified at 49 U.S.C. app. Sec. 1552(f)(3), contains two provisions, as we have seen. The first, the so-called "report and wait" provision, is left unaffected in Chadha 's wake; the second is, of course, the legislative veto provision constitutionally eviscerated by Chadha. If the veto provision were severed, the remainder of the section (including the first sentence of the subsection) would be fully operable as law. The employee protection program would still be spelled out in great detail, and the Secretary of Labor would still retain authority to promulgate the necessary rules and regulations. The sole difference would be that the Secretary would be required only to submit the rules and regulations to the appropriate Congressional committees on a "report and wait" basis rather than expose those regulations to the gauntlet of possible invalidation by way of an exercise of the purported veto authority.

To overcome the presumption of severability with respect to this provision, the airlines must show, as we have seen, that it is evident that Congress would not have passed the section had Congress known that the veto provision would disappear on account of its unconstitutionality. Or, as the Temporary Emergency Court of Appeals would put it, the challengers must show that it is evident that Congress would have preferred no airline employee protection provision at all to the existing provision sans the veto provision. Not only is evidence for that position lacking, what evidence there is indicates that Congress had a strong, humanitarian desire to provide for airline employee protection; in contrast to this clearly expressed intent to erect protective safeguards for displaced workers, Congress expressed relatively little concern about the veto provision. That is to say, the evidence, in our view, positively indicates that Congress would have preferred the employee protection plan with the veto provision severed out to a deregulatory statute stripped of any employee protection plan. It is to that evidence that we now turn.

The employee protection provisions contained in the Airline Deregulation Act had their genesis in the Senate bill. The Senate Report is thus particularly instructive as to the importance Congress attributed to that program. Noting the concern expressed by airline employees over the security of their jobs once deregulation was effected, the Senate Report stated:

The Committee felt that this concern [of airline employees] should not be ignored and that careful consideration of the question of employee protection programs was warranted. The Committee concluded that the kind of employee dislocations that might occur as a result of the new regulatory structure should be dealt with by a statutory employee protection program, as has been done in certain cases in the past.... The Committee believes that the Congress ... must insure that the benefits to the public which result from its decision to alter substantially the regulation of air transportation are not paid for by a minority--the airline employees and their families who have relied on the present system.

S.Rep. No. 631, 95th Cong., 2d Sess. 113-14 (1978). To that end, the Senate bill, S. 2493, included an employee protection program very much like the regime that was ultimately adopted. See Conference Report, H.R.Rep. No. 1779, 95th Cong., 2d Sess. 105 (1978), U.S.Code Cong. & Admin.News 1978, pp. 3737, 3813 (observing that the Conference substitute was "basically the same as the Senate bill" with certain stated exceptions, including some reworking of the subsection containing the veto provision).

The House bill, H.R. 12611, also contained an employee protection program. As explained by the House Report:

The bill stipulate[d] that no authority granted by th[e] act shall be exercised by any carrier unless the Secretary of Labor has certified to the CAB that interests of employees have been adequately protected by fair and equitable arrangements, with benefits no less than those established pursuant to section 5(2)(f) of the Interstate Commerce Act and section 405 of the Rail Passenger Service Act.

H.R.Rep. No. 1211, 95th Cong., 2d Sess. 26 (1978), U.S.Code Cong. & Admin.News 1978, pp. 3737, 3762. That this provision was viewed both as an important feature of the bill and as being stronger than the Senate plan is evidenced in the various comments disparaging the perceived reduction in protection resulting from the Conference substitute.10

The airlines argue, however, that "Congress was uncertain about the need for additional labor protection beyond that already available through the CAB." Appellees' Brief at 41. This is quite true. Indeed, statements in the legislative history reasonably can be read as evidencing a sense that employee protection was not in fact vital to the crafting of a comprehensive deregulatory regime. Those statements stemmed not from legislative oversight or, worse yet, callousness to the potential plight of displaced workers, but reflected, rather, the sincerely held belief in various legislative quarters that loss of employment in a deregulated industry was unlikely.11 While that sentiment is certainly present in the debate, as the airlines correctly and understandably emphasize, the sentiment does not, upon analysis, truly speak to the precise issue at hand. For, inherent in the notion of increased competition in the industry is the reality that some carriers will likely lose business, while other carriers will gain. Thus, in Congress' view at the time, while total employment within the industry, once deregulated, might well remain stable or grow, individual employees might nonetheless lose their jobs due to deregulation; and, on the other hand, jobs with other carriers would likely become available. The preferred hiring status provisions fashioned by Congress would thus move those employees who lost their jobs into the newly opened or created positions.

In our view, legislative comments with respect to the overall employment picture within the industry speak more to the provision for payment of benefits from federal coffers than to the first hire provisions. Congress deemed it unlikely that the former provision would, in fact, come into play; that sense of the National Legislature is eloquently evidenced by the fact that Congress to this day has not funded that program. These comments thus do not speak, as the airlines would read them, to the "first hire" provisions. Indeed, the "first hire" provisions could, in theory, play a constructive fiscal role in keeping potentially costly federal relief provisions from being triggered in the first instance. These comments, then, do nothing to counter the evidence that Congress placed great importance on the existence of the employee protection plan as an integral part of a comprehensive deregulatory regime.

In stark contrast to the numerous comments indicating the importance of the employee protection plan, the only discussion cited by the parties with respect to the veto provision is a solitary comment by then-Representative Elliott Levitas of Georgia expressing approval of the presence of a legislative veto in the provision.12 Not only is this sole source a thin reed for eviscerating an entire remedial system fashioned out of humane Congressional concern for adversely affected airline employees but, what is more, Rep. Levitas was a vigorous advocate generally for the proposition, eviscerated by Chadha, that Congress must retain control over administrative agencies via the legislative veto device.13 Now to be sure, Rep. Levitas was, as the airlines rightly emphasize, a member of the Conference Committee for the Airline Deregulation Act and of the House Committee on Public Works and Transportation, the committee to which the House bill was referred. Thus, his statement was not one merely of a zealous advocate championing his well-known position in favor of lavishly sprinkling regulatory (or deregulatory) legislation with legislative veto devices. However, there is not a shred of evidence that Rep. Levitas' statement represented the view of Congress as a whole (or even his full Committee), nor is there any indication that Congressman Levitas himself attached especial significance to this particular veto provision, as opposed to his keen interest generally in seeing the incorporation of such devices in legislation enacted by the Congress.

Failing to find any direct statement evidencing a Congressional intent of inseverability, appellees advance various items of circumstantial evidence of what they perceive to be such an intent. Appellees first point to the fact that only section 43 of the Act is subject to the veto provision. That fact, however, is hardly surprising. The Act was, of course, a deregulatory statute, as its very name suggests, whereas section 43 was, for lack of a better term, a regulatory provision. It is, as a matter of common sense, the regulatory provision which would be the most likely candidate for inclusion of a legislative veto device. That fact, however, still does not speak to whether Congress would have preferred no employee protection plan at all, as opposed to a plan containing a "report and wait" mechanism but lacking a veto provision.

The airlines also argue that the legislative veto provision in section 43 is "distinctive among the almost 200 legislative veto provisions which Congress has included in federal statutes." Appellees' Brief at 22. Admittedly, the provision reflects, as a whole, a unique combination of features. For one thing, as we have seen, a "report and wait" requirement exists alongside the veto, see 49 U.S.C. app. Sec. 1552(f)(3); in addition, a requirement is imposed upon the Secretary of Labor to promulgate rules and regulations necessary to carry out the section within six months after October 24, 1978, see 49 U.S.C. app. Sec. 1552(f)(2). It is not clear to us that anything of decisional import is to be drawn from the inclusion of these features.

The "report and wait" requirement simply directs the submission of rules and regulations, prior to issuance as final rules, to particular oversight committees of Congress. This mechanism may reasonably be viewed as the assertion of an interest by committees not normally charged with the review of labor regulations but directly concerned with the airline industry over labor regulations affecting deregulation of that industry. This understandable interest does not indicate that the veto is inseverable, particularly since the "report and wait" clause remains inviolate notwithstanding Chadha.

Neither is it clear what conclusion is to be drawn from the six-month requirement imposed on the Secretary. The airlines suggest that this subsection was designed to insure that the regulations would be reviewed by essentially the same members of Congress who had shaped the Airline Deregulation Act. While an election would have intervened between the passage of the Act and the required time of submission, the historical experience of reelections of incumbents might presumably lend some credence to this interpretation. Nonetheless, like election predictions, this approach is all purely speculative. The provision could just as readily be seen as an indication of the high importance Congress attached to the employee protection plan, which Congress therefore mandated be implemented as quickly as possible. Furthermore, even assuming arguendo appellees' point, that view fails, at bottom, to speak to the issue at hand. From it, we cannot conclude that Congress, or even the committees involved, would have preferred no employee protection plan to the adopted plan divorced of its legislative veto provision.

Appellees also seek support in the treatment of the veto provision, and the protection plan as a whole, in the deliberations of the Conference Committee. They argue that

the House accepted the Senate's novel concept of an [employee protection plan] that it had not theretofore embraced, but only with a significantly strengthened legislative review provision. The terms of this compromise constitute strong evidence of nonseverability because, as in [American Federation of Government Employees v. Pierce, supra ], one house accepted the overall thrust of the other house's conflicting legislation, but only with the veto-related restrictions incorporated.

Appellees' Brief at 33 (footnote omitted). The airlines thus argue that the decision here should be influenced by our decision in Pierce, a case in which a veto provision was held inseverable. However, Pierce is readily distinguishable from the case at hand. Pierce concerned a provision of an appropriations act for the Department of Housing and Urban Development, Pub.L. No. 97-272, 96 Stat. 1160, 1164 (1982), which provided that funds could not be used for any reorganization by HUD, prior to January 1, 1983, without prior approval of the two Appropriations Committees. The district court had assumed the invalidity of the approval clause and severed it from the remainder of the provision. This court reversed, finding intent of nonseverability. However, as we shall now see, the Conference Committee compromise providing the evidence of inseverability in Pierce differed totally from the compromise reached in conference with respect to the Airline Deregulation Act.

In Pierce, the House had "never put forward a provision without a committee approval clause tied to it." Pierce, supra, 697 F.2d at 307. On the other hand, "the Senate initially opposed any restraint on the Department ... [and] the Senate Committee[ ] clearly expressed reservations about the House version of the bill." Id. The Conference Committee compromise gave control to the Appropriations Committees, but only for a limited duration. Neither House, it was abundantly clear in Pierce, would have accepted a complete ban on reorganizing HUD--yet that is precisely the result severance would have brought about.

The Conference Committee compromise reached here bears no resemblance at all to the situation in Pierce. As to the airline deregulation bill, the Conference adopted, for the most part, the Senate bill's employee protection plan. That plan contained a veto clause, to which the Conference Committee added the "report and wait" provision. Even if the "report and wait" provision may be viewed as designed to strengthen the legislative veto provision, we have found no evidence that the House's agreement to the Senate version of the protection plan was given in exchange for beefing up the veto device. In fact, the House's protection plan was viewed, at least by several Representatives, as being much stronger than the Senate's plan. See supra note 10. It seems to us unlikely that a program offering less protection would have occasioned a call for even greater legislative control over the bureaucracy. Nor have we found any indication that the Senate's agreement to the minor modifications in its version of the plan was premised on a "strengthening" of the veto provision. In short, we conclude that Pierce is wholly inapposite to the case at hand.

We emerge from our trek through the legislative history of the Airline Deregulation Act far short of the destination that must be reached for the airlines to prevail. We fail to find satisfied here the exacting inseverability standard that it be "evident" that Congress would have preferred no employee protection program over the program passed by Congress shorn of its veto provision. Ample evidence persuades us that the employee protection program, whatever its merit, was deemed by Congress to be an important aspect of the Act. In contrast, there is not a shred of evidence that the veto provision was deemed to be a vital feature of the protection plan.

In sum, we hold the legislative veto provision here to be severable from the remainder of the employee protection provisions; the remainder of section 43 therefore remains viable. However, since the District Court did not have occasion to rule on the other grounds advanced by the airlines in their challenge to the rules adopted by the Secretary, the case is remanded to that court for consideration and adjudication of those issues.

Reversed and remanded.

GINSBURG, Circuit Judge, concurring in the judgment.

The court's intricate discussion of preliminary issues seems to me distracting. As I view this case, only one genuine issue appears: would Congress have preferred excision of the veto provision to demolition of the airline employee protective prescriptions. Before the district court, the Department conceded the unconstitutionality of the veto, and does not contest that matter here. Therefore, we face no appeal from decision of a constitutional question; instead, a ripe statutory interpretation controversy--a construction of legislation dispute securely within our appellate jurisdiction--is before us. Furthermore, as the court recognizes, slip op. at ---- - ---- & n. 5, a ruling that the labor protective prescriptions are severable strips discourse on Chadha 's "retroactive application" of "any importance."

On the severability question, I am in full agreement with the court's judgment, which undertakes a moderate salvage operation. Deletion of the veto preserves section 43 and thereby gives effect to the dominant intent of Congress. A declaration of inseverability, tearing down the section in its entirety, would be far more destructive of the legislature's will. Cf. Buckley v. Valeo, 424 U.S. 1, 108-09, 96 S.Ct. 612, 677, 46 L.Ed.2d 659 (1976) (per curiam).