McLouth Steel Products Corporation v. Lee M. Thomas, Administrator, and U.S. Environmental Protection AgencyMcLouth Steel Products Corporation v. Lee M. Thomas, Administrator, and U.S. Environmental Protection Agency
Opinion for the Court filed by Circuit Judge WILLIAMS.
This case presents a challenge by McLouth Steel Products Corporation to the Environmental Protection Agency’s denial of its petition to exclude waste generated at its steel-making facility from EPA’s list
Subtitle C of RCRA requires EPA to promulgate regulations establishing a comprehensive federal management system to protect human health and the environment from hazardous wastes.
Once EPA lists a waste as hazardous, a party may petition EPA for delisting — exclusion of its specific waste from the generic listing.
On September 25, 1981, McLouth filed a petition requesting delisting for the sludge from its Trenton, Michigan plant. Joint Appendix at 57. On November 18, 1986, EPA denied the petition on the ground that McLouth had not substantiated its claim that the waste was non-hazardous. 51 Fed. Reg. 41,624 (Nov. 18, 1986).
In arriving at that conclusion, EPA used its VHS model (referring to “vertical and horizontal spread”) to predict the “leach-ate” levels of the hazardous components of McLouth’s waste. 1 The model “estimates the ability of an aquifer to dilute the toxi-cants from a specific volume of waste, and predicts toxicant levels at a receptor well.” 50 Fed.Reg. 48,886/3 (Nov. 27, 1985). It is not intended to predict contamination based on conditions at the specific site of disposal. Rather, it is based on certain “reasonable worst case assumptions,” 50 Fed.Reg. 7,882, 7,883/2 (Feb. 26, 1985), because, EPA reasons, “it cannot be guaranteed that the site-specific circumstances will not change.” 50 Fed.Reg. at 48,907/3. A user of the model feeds into it data as to the actual leachate concentrations of the constituents of a specific waste, and the amount of waste generated at the site, and the model predicts contamination levels. EPA then compares these predictions with heаlth-based standards for each constituent to determine whether a waste should be delisted.
The VHS model predicted that significant levels of two hazardous constituents, lead and cadium, would leach from McLouth’s waste and contaminate the groundwater. 51 Fed.Reg. 41,624-625 (Nov. 18, 1986). McLouth filed suit in this court under
EPA argues that the VHS model is not subject to
This court recently confronted the rule/policy distinction in
Community Nutrition Institute v. Young,
In practice, there appears some overlap in the
Community Nutrition
criteria; the second criterion may well swallow the first. If a statement denies the decisionmаker discretion in the area of its coverage, so that he, she or they will automatically decline to entertain challenges to the statement’s position, then the statement is binding, and creates rights or obligations, in the sense those terms are used in
Community Nutrition.
The question for purposes of
We find that the VHS model meets the definition of a legislative rule under the
Community Nutrition
criteria. For openers, EPA’s current claim that “it does not consider itself ... bound by [the VHS] model” (Br. at 20) is obviously of little weight. The agency’s past characterizations, and more important, the nature of its past applications of the model, are what count.
See Community Nutrition,
It is true that in the Federal Register notices announcing EPA’s intent to employ the VHS model, EPA indicated that it retained discretion to deviate from its use.
See, e.g.,
50 Fed.Reg. at 7,883 (VHS model is “one factor to determine the potential impact of unregulated disposal of petitioned waste on human health and the environment”);
id.
at 7,886 (if waste fails VHS model “the waste could be considered nonhazardous”);
id.
at 48,910 (“the VHS model results, while providing importаnt input to a delisting decision, are not necessarily the sole basis for such a decision”). Yet other language in that notice strongly suggests that EPA will treat the model as a binding norm. For example, in its first reference to the VHS model, EPA referred to it as “the quantitative approach” that
“will
be used to predict the level of the various toxicants which could migrate to environmental receptors.”
Id.
at 7,882, 7,896 (emphasis added). The use of the word “will” suggests the rigor of a rule, not
EPA has stated that factors not included in the model “may be considered in the delisting decision where the petitioner can make a
compelling
case that they are germane to a particular petition.”
More critically than EPA’s language adopting the model, its later conduct applying it confirms its binding character. In its final decision оn McLouth’s delisting petition, it stated that “even if the listed wastes do not exhibit the characteristics of hazardous waste, the wastes must also pass the VHS model evaluation ... in order to be delisted.” 51 Fed.Reg. 41,624/3 (Nov. 18, 1985) (emphasis added). Further, it said that “the VHS model was made final on November 27, 1986 ... and all comments received on the proposal for the model were incorporated. Generally these comments are no longer entertained by the agency.” Id. at 41,626.
EPA was even more close-minded and dismissive in its denial of a delisting petition by LTV Steel Company:
LTV had several criticisms of individual parameters used in the VHS model itself (e.g., the dilution predicted by the model for large volume wastes and the lack of attenuation for metal species). Since the VHS landfill model was made final on November 27, 1985 (see 50 FR 48886, Appendix), and all comments received in the proposal for the model were incorporаted, these comments will not be entertained.”
51 Fed.Reg. 41,317-18 (Nov. 14, 1986) (emphasis added).
The language quoted above is similar to the “type of mandatory, definitive language” that was “a powerful, even potentially dispositive, factor” in this court’s determination that the FDA action levels in
Community Nutrition
were legislative rules.
As evidence of its flexible use of the VHS model, EPA calls our attention to four instances in which it rejected the model’s predictions. These four come out of approximately 100 final delisting decisions sincе its appearance (see Exhibit J). (EPA evidently ran the model in all delisting applications filed after its adoption and which were not rejected as incomplete. See Reply Brief at 4 n. 4.) But these four rejections in no way reflect any EPA willingness to question the model itself. In three EPA made adjustments simply because the model had been fed faulty data, see 51 Fed.Reg. 16,865-66 (May 7, 1986); id. at 16,068-70 (April 30, 1986); id. at 36,247-49 (Oct. 9, 1986). In the fourth, the variability of a waste treatment facility’s raw materia] led EPA to grant a delisting thаt was in fact based on the VHS model, as it was conditioned on the applicant’s testing each batch of waste under the model, and, as to any batch that failed the model, providing new treatment or disposition under RCRA’s management system, see id. at 36,976-79 (Oct. 16, 1986). If a rule says, “If A, then B,” an agency does not show its open-mindedness when it refuses to find B after being persuaded that A is not the case.
Thus despite its claim that it is open to “new approachеs” to delisting decisions,
see
Brief for Respondent at 24, EPA has evidenced almost no readiness to reexamine the basic propositions that make up the VHS model,
i.e.,
propositions about the numerical relationship between leachate concentrations and waste quantities on one
EPA’s claim to have been open to consideration of other factors does not make the VHS model any less of a rule.
See
50 Fed.Reg. 7,896 (stating VHS model is a quantitative tool for assessing 6 of the 11 factors that must be considered under the delisting regulations). If a disposition turns on affirmative answers to two questions, and an agency adopts a rule giving conclusive answers to the first question оnce certain data are supplied, the rule is a rule even though it does not purport to answer the second question.
Cf. Pickus v. United States Board of Parole,
The EPA has rested on the policy statement exception to
While the exact definition of an interpretive rule is unclear, 3 it is clear that wе don’t have one here. The VHS model meets Community Nutrition’s affirmative definition of a legislative rule: it substantially curtails EPA’s discretion in delisting decisions and accordingly has present binding effect. While not all rules of thumb are legislative rules, they are when they have these characteristics.
II. Failure to Provide Adequate Notice and Opportunity to Comment
Perhaps paradoxically, EPA argues that while “at no time did it claim to be proposing the modеl as a rule,” Brief for Respondent at 21-22, the model should, if found a rule, be upheld because its promulgation satisfied the requirements of
The story, however, is not so simple. Although the notice’s heading spoke of a “Proposed rule and request for comment,” а reader going on to the “Summary” would find no reference to the VHS model, only to EPA’s proposed treatment of the six specific delisting petitions (which EPA addresses in the format of a notice-and-comment rule-making,
see
An agency may not introduce a proposed rule in this crabwise fashion. No earlier eases finding inadequacy of notice are exactly parallel, but they suggest the courts’ concern that notice be clear and to the point.
See AFL-CIO v. Donovan, 757
F.2d 330, 339 (D.C.Cir.1985) (“no notice, much less adequate notice” where agency reprinted entire set of regulation in 40 pages of the Federal Register, including the proposed change but not identifying it in the preamble, which highlighted other proposed changes);
National Tour Brokers Ass’n v. United States,
Nor was the defect cured by McLouth’s actual notice оf the VHS model and its chance to comment on it in its own delisting case (which took the form, as noted above, of a “rulemaking”). Of course it is true that defects in an original notice may be cured by an adequate later notice,
Forester v. Consumer Products Safety Commission, 559
F.2d 774, 788 (D.C.Cir.1977), but that curative effect depends on the agency’s mind remaining open enough at the later stage.
Id.
at 788 n. 19;
National Ass’n of Farmworkers Organizations v. Marshall,
III. Remedy
EPA argues that despite the procedural irregularities, McLouth cannot secure reversal of the adverse decision unless it demonstrates that the irregularities caused “specific prejudice,” citing
In re Surface Mining Regulation Litig.,
While we are unwilling to endorse the Fifth Circuit’s blanket rule,
5
we see no factors here militating against remand except for the universally available one that it will consume agency time. McLouth’s temporary exclusion from managing its waste in accordance with the RCRA scheme expired as a matter of law on November 8, 1986,
see
In holding that EPA violated
The availability of this curative option is supported by
While an agency’s open-mindedness in individual proceedings can substitute for a general rulemaking, the record as a whole here convinces us that in ruling on McLouth’s delisting petition the EPA lacked that state of mind with respect to the issues covered by the VHS model. Consequently, if EPA decides to detour around promulgation of a rule in strict accordance with
Notes
. EPA has defined “leachate" as "any liquid, including аny suspended components in the liquid, that has percolated through or drained from hazardous waste."
. The EPA disposes of a delisting petition by
.
Community Nutrition
appears to treat interpretive rules as synonymous with policy statements. The dividing lines drawn by other decisions are far from self-applying.
See, e.g., United Technologies Corp.
v.
EPA,
. EPA makes no claim that the model can be sustained as the product of an adjudication.
Cf. NLRB v. Wyman-Gordon Co.,
. In reaching this broad holding, the Fifth Circuit cited
Braniff Airways v. CAB,