concurring in part and dissenting in part:
I concur with the majority in all but the final part of its excellent and thorough opinion. I cannot agree that specific numerical objectives for reducing reimbursed services in a particular geographic area “are best classed as ‘general statements of policy.’ ” Majority opinion (“Maj. op.”) at 1056. I therefore would not exempt these numerical objectives from the notice and comment requirements of the Administrative Procedure Act (APA).
The majority recognizes that these numerical objectives can and are intended to produce significant effects. It acknowledges that the numerical objectives are “highly specific,” requiring a PRO, for example, to “reduce 5,425 unnecessary or inappropriate admissions in 22 specific hospitals.” Maj. op. at 1055. By following such strict guidelines, a PRO may end up denying reimbursements for services that qualify as “medically necessary” under medicare law. In the end, however, the majority closes its eyes to this possibility, preferring to “await the sharpened facts that come from the actual workings of the regulation.” Maj. op. at 1056. Because I cannot blink away the likely consequences of such specific objectives, and because the APA ensures participation in rulemaking before, rather than after, the fact, I dissent.
I.
The majority’s reasons for adopting a “wait and see” approach are unpersuasive. Above all, the majority uncritically accepts an agency’s own description of its action as a “policy statement” rather than a “substantive rule.” The majority says its “concerns are allayed” because the Department of Health and Human Services (HHS) “insist[s] that these [numerical] norms are merely hortatory.” Maj. op. at 1056. While claiming it is only “giving weight” to HHS’ characterization, maj. op. at 1056, the majority defers to an extent that I believe is incompatible with this Circuit's precedents.
In Community Nutrition Institute v. Young, 818 F.2d 943 (D.C.Cir.1987), this court confronted a claim by the Food and Drug Administration that certain of its rules were interpretive, rather than substantive, and thus exempt from notice and comment. Far from acquiescing in that assertion, the court gave it only “some, albeit ‘not overwhelming,’ deference” and ultimately reversed the agency’s finding. Id. at 946 (quoting Brock v. Cathedral Bluffs Shale Oil Co., 796 F.2d 533, 537 (D.C.Cir.1986)). The court took the same tack in Environmental Defense Fund, Inc. v. Gorsuch, 713 F.2d 802, 816 (D.C.Cir. 1983). And, in Citizens to Save Spencer County v. EPA, 600 F.2d 844 (D.C.Cir. 1979), our Circuit found that the “label that the particular agency puts upon its given exercise of administrative power is not, for our purposes, conclusive; rather it is what the agency does in fact.” Id. at 879 n. 171 (quoting Lewis-Mota v. Secretary of Labor, 469 F.2d 478, 481-82 (2d Cir.1972)).
This approach conforms with more general APA precedents. Because of the “salutary effect of the [APA’s] public comment procedures,” courts should “only reluctantly ... recognize exemptions therefrom.” Humana of South Carolina v. Califano, 590 F.2d 1070, 1082 (D.C.Cir. 1978). As this court further noted in American Bus Ass’n v. United States, 627 F.2d 525 (D.C.Cir.1980), “Congress was alert to the possibility that these exceptions [from notice and comment] might, if broadly defined and indiscriminately used, defeat the section's purpose. Thus, the legislative history of the section is scattered with warnings that various of the exceptions are not to be used to escape the requirements of section 553.” Id. at 528.
Nothing in this policy of giving courts the final word when construing the APA has been altered by Chevron U.S.A., Inc. v. Natural Resources Defense Council, 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984). The Chevron Court counseled greater deference, but only when a court “reviews an agency's construction of the statute which it administers ” — when an agency makes “policy and ... rules to fill any gap left, implicitly or explicitly, by *1059Congress.” Id. at 842, 843 (emphasis added).
II.
The majority’s deference to HHS’ characterization of its numerical objectives as “merely hortatory” is especially inappropriate because it is so at odds with the facts. Those facts demonstrate that (1) the numerical objectives are designed to reduce reimbursement levels and (2) because of this intended effect, they may exclude “medically necessary” services from medicare coverage. It is this risk of unwarranted exclusions that necessitates notice and comment rulemaking.
A.
As the majority concedes, Congress’ purpose in introducing Peer Review Organizations (PROs) into the medicare statute in 1982 was to “crack down on excessive reimbursements to hospitals.” Maj. op. at 1041. Numerical objectives supplied the means for attaining that end. Congress found the prior system of professional standards review organizations to be defective, because “widespread inappropriate usage of costly health care services” persisted. S.Rep. No. 494, 97th Cong., 2d Sess. 41 (1982), U.S. Code Cong. & Admin.News 1982, pp. 781, 817. The inauguration of PROs was therefore promoted as a way of getting “entities who have proven their effectiveness to enter into 'performance based contracts for the conduct of review.” Id. (emphasis added).
Congress made these contracts “performance based” by requiring that numerical objectives be negotiated for each PRO. See 42 U.S.C. § 1320c-2(c)(7) (1982). These objectives were thus the crucial mechanism for reducing reimbursements; they were intended to have that effect. Indeed, the sample PRO contract in this case unequivocally states that “[t]he contractor shall achieve the admission and procedure objectives” that it has negotiated with the Health Care Financing Administration. J.A. at 959 (emphasis added). I fail to see, then, how the majority can conclude that HHS “will put no stock in the absolute reduction of given medical procedures as a criterion for renewal of PRO contracts.” Maj. op. at 1056. The medicare law requires that a PRO’s performance “will be judged” against these objectives. 42 U.S. C. § 1320c-2(c)(7) (1982). HHS would be derelict in its duties if it “put no stock” in a PRO’s compliance.
B.
Since numerical objectives are thus intended to have a significant effect on reimbursements, the crucial question is whether they may sometimes prove too effective. A PRO, knowing it will be measured against such firm benchmarks, may — unconsciously or otherwise — deny reimbursement for some “medically necessary” services, in order to meet its objectives. I believe this result is entirely plausible. And, if reimbursements are in fact denied for eligible services, hospitals will be among the affected parties.
The fact that numerical objectives may have this type of impact on hospitals and other parties strongly suggests that the objectives require notice and comment, see Batterton v. Marshall, 648 F.2d 694, 708 (D.C.Cir.1980), although it does not compel that conclusion. Cf. American Postal Workers Union v. United States Postal Service, 707 F.2d 548, 560 (D.C.Cir.1983), cert. denied, 465 U.S. 1100, 104 S.Ct. 1594, 80 L.Ed.2d 126 (1984). As the Batterton court stated, “[t]he critical question is whether the agency action jeopardizes the rights and interests of parties, for if it does, it must be subject to public comment prior to taking effect.” 648 F.2d at 708 (footnote omitted). Specifically, the Bat-terton court held that a new method of calculating unemployment statistics was not a “policy statement,” or otherwise exempt from notice and comment, because “recipients of CETA emergency job program monies [were] subject to a new method for determining ... allocation^].” Id. In other words, statutory entitlements were affected by the agency action.
In addition to their possible impact on medicare entitlements, the numerical objec*1060tives have been negotiated pursuant to an explicit legislative directive. This is another crucial factor that suggests agency action is a substantive rule rather than a statement of policy. American Postal Workers, supra, at 558.
The majority tries to sidestep these factors — reciting, for example, that “the real dividing line between regulations and general statements of policy is publication in the Code of Federal Regulations.” Maj. op. at 1056. But this claim is just another way of deferring to HHS’ own views. Obviously, an agency that contends its rule is not substantive is unlikely to publish that rule in the CFR. This fact, however, adds nothing to the underlying contention. If the agency's action is in reality a substantive rule, it is no less so for remaining unpublished.
It is conceivable, of course, that all of the medical services targeted for nonreim-bursement under the numerical objectives will, in fact, qualify as “unnecessary.” In other words, the numerical objectives in PRO contracts may simply strengthen enforcement of the statutory standard that already specifies which medical services should not be reimbursed. But, there are strong reasons for doubting that the impact of numerical objectives will be so confined. These doubts are fed by Congress’ clear intent to reduce medicare costs, by the competitive bidding process that generates the numerical objectives, by the ultimate use of those objectives to judge PRO performance when contracts are renewed, by the malleable character of the “medically necessary” reimbursement standard, and by the fact that a PRO’s application of that standard to medicare claims will usually be final. See 42 U.S.C. § 1320c-3(a)(2) (1982). Thus, the numerical objectives may well reduce the level of reimbursements below what the statute intends. I conclude that the affected hospitals are entitled to participate in the formulation of the numerical objectives.
This conclusion is supported by a recent decision by the Ninth Circuit that very closely resembles the case before us. In W.C. v. Bowen, 807 F.2d 1502 (9th Cir. 1987), the court confronted a regulation that affected the plaintiffs in the same indirect but palpable way that the PRO objectives affect hospitals. W.C. involved Congress’ new policy of having the Social Security Administration (SSA) review administrative law judge decisions granting disability benefits. Unable to review all such benefits awards, the SSA promulgated a regulation limiting its review to those awards made by ALJs who granted seventy percent or more of the disability claims they adjudicated.
The W. C. trial court found, and the Ninth Circuit agreed, that the regulation that targeted review was substantive and therefore subject to notice and comment requirements of the APA. The Ninth Circuit noted that the regulation not only changed existing policy but was also promulgated pursuant to legislative authority in the form of the Bellmon Amendment. That Amendment, reflecting Congress’ concern about rising disability payments, directed the Secretary of Health and Human Services to begin reviewing ALJ decisions “on his own motion.”
The parallel between the facts in W.C. and in the case before this court is striking. The Bellmon Amendment intensified the scrutiny of disability claims without altering the standards by which those reviews would be conducted. The numerical objectives for PRO review of hospital reimbursement claims serve the same cost-cutting function in the same manner. In W.C., even though the standards remained unchanged, the court concluded that “[t]he review program affects existing rights” in part because “it was designed to alter ALJ decisions.” Id. at 1505 (emphasis added). As the Ninth Circuit thus recognized, where there is strong congressional intent to reduce benefits payments, there is a concomitant risk of overshooting the mark. Because the numerical objectives in the present case are similarly designed to reduce hospital reimbursements, we should provide the same buffer against “overshooting” by permitting notice and comment.
*1061C.
The majority’s attempt to distinguish W. C. is unpersuasive and, not surprisingly, relegated to a closing footnote. The footnote argues that this case differs from W.C. in two “profound” respects. First, the majority says “there has been no finding that HHS’ objectives have in fact caused, sub rosa, a shift in the underlying substantive standard of Medicare review.” Maj. op. at 1057 n. 4. It is true that the court rulings in W. C. occurred long enough after the Bellmon regulation was promulgated that the courts could take judicial notice of a decline in actual disability awards. But nowhere did the opinions indicate that this factual finding was essential to the holding. On the contrary, both the district and circuit courts in W. C. relied on the fact that the intent of the Bellmon program was to reduce disability awards. W.C. v. Heckler, 629 F.Supp. 791, 798 (W.D.Wash.1985) (“the Bellmon Review Program was not a procedural rule because the ultimate purpose of the program was to change the outcome of agency decisions”); W.C. v. Bowen, 807 F.2d at 1505 (to same effect).
This court, of course, has no data before it to gauge the success of the PRO objectives in reducing medicare reimbursements. This cannot be determinative, however, of the objectives’ rulemaking status. The majority’s approach means that, even when an agency action otherwise possesses the attributes of a substantive rule, the court must wait to determine its true impact before deciding whether notice and comment should have been accorded at the outset. “In an ideal world,” the majority rationalizes, “one could ascertain at the outset the need for notice and comment.” Maj. op. at 1056. By resigning itself to an imperfect world, the majority misses the real issue. Granted that our foresight is not ideal, the key question is how the risks of our imperfect vision should be allocated. I believe that doubts about a rule’s future impact should be resolved in favor of participatory rulemaking. The majority is content, instead, to make plaintiffs wait out “the actual workings of the regulation.” Maj. op. at 1056.
Delaying the determination that a rule is substantive allows a rule to establish itself as the norm, only to be thrown out later as improperly formulated. This is unfair to affected parties and disrupts orderly government. Such an approach also runs counter to the very purpose of notice and comment, which is to assure that the correct rules are established in the first instance. As this court has explained, “public participation assures that the agency will have before it the facts and information relevant to a particular administrative problem ... [and] increase[s] the likelihood of administrative responsiveness to the needs and concerns of those affected.” Guardian Federal Savings & Loan v. FSLIC, 589 F.2d 658, 662 (D.C.Cir.1978).
The majority offers one other ground for distinguishing W.C. from this case. In W.C., we are told, the Social Security Administration imposed a “tough new substantive barrier” between claimants and their disability awards. Maj. op. at 1057 n. 4. Notwithstanding the majority’s forceful adjectives, the “barriers” imposed in these two cases are really quite similar. In W.C., Congress created a new program for reviewing disability claims. So, too, in this case, Congress replaced the former system of professional standards review organizations with tougher watchdogs: the new PROs. In W.C., the agency went beyond the statute’s requirement of heightened review, targeting the disability decisions of particular AUs. So, too, in this case, HHS went beyond the statute’s simple requirement of “negotiated objectives,” establishing numerical goals for reducing particular categories of medical services that HHS devised on its own.
In short, these two cases overflow with similarities rather than differences. And, in each case, participatory rulemaking is warranted because “the purpose of § 553 is ‘to enable[] the agency promulgating the rule to educate itself before establishing ... procedures which have a substantial impact on those regulated.’ ” National Tour Brokers Ass’n v. United States, 591 F.2d 896, 902 (D.C.Cir.1978) (footnote omit*1062ted). By denying notice and comment, we defeat that process. The impact of the type of “policy” statement HHS has issued in this case can numb the ability of regulated institutions to challenge the agency’s policy decision. If an agency never needs to hear from its constituency, it can fail to carry out its statutory mission and never become the wiser. I would hold that the numerical objectives contained in PRO contracts should be subject to notice and comment, at least with respect to the hospitals located within a PRO’s jurisdiction.