Plaintiffs, who include the Animal Legal Defense Fund ("ALDF"), the Animal Welfare Institute ("AWI"), and three individuals, challenge the United States Department of Agriculture's ("USDA") decision not to adopt a Draft Policy that would have provided guidance to zoos, research facilities, and other regulated entities in how to ensure the psychological well-being of nonhuman primates in order to comply with the federal Animal Welfare Act ("AWA"). Plaintiffs challenge the decision not to adopt the Draft Policy under the Administrative Procedure Act ("APA") as arbitrary and capricious. The district court did not reach the merits of plaintiffs' suit because it determined that the USDA's decision did not constitute reviewable final agency action. We disagree. We hold that at least one of the plaintiffs has standing under Article III of the Constitution, and we conclude that the district court has authority under the APA to review the USDA's decision not to adopt the Draft Policy.

Congress enacted the AWA in 1966 "to insure that animals intended for use in research facilities or for exhibition purposes or for use as pets are provided humane care and treatment . . . ." 7 U.S.C. § 2131(1). As originally enacted, the AWA left research facilities largely unregulated. See, e.g., 7 U.S.C. § 2143(a) (repealed 1985). In 1985, Congress amended the AWA by enacting the Improved Standards for Laboratory Animals Act, Pub.L. No. 99-198, 99 Stat. 1645. This amendment to the AWA instructed the USDA to "promulgate standards to govern the humane handling, care, treatment, and transportation of animals by dealers, research facilities, and exhibitors." 7 U.S.C. § 2143(a)(1). These standards must "include minimal requirements . . . for a physical environment adequate to promote the psychological well-being of primates." Id. § 2143(a)(2)(B). The Secretary proposed a regulation containing these standards in 1989. See Animal Welfare—Standards, 54 Fed.Reg. 10897, 10917 (proposed Mar. 15, 1989).

The 1989 proposed regulation would have imposed a number of detailed "minimum requirements" on regulated entities. These included, for example, a requirement that "nonhuman primates must be housed in primary enclosures with compatible members of the same species or with compatible members of other nonhuman primate species" unless doing so would endanger the animal. 54 Fed.Reg. at 10944. As finally adopted in 1991, however, the regulation, which requires regulated entities to create an "Environmental Enhancement Plan" to benefit nonhuman primates, left the requirements for these "primary enclosures" more vaguely defined. See generally 56 Fed.Reg. 6426 (1991), codified at 9 C.F.R. § 3.81. The social grouping provision of the regulation, for example, contains no specific instruction regarding the pairing or grouping of animals. It leaves regulated entities considerable discretion to house nonhuman primates as they see fit, provided that housing conditions accord with "currently accepted professional standards, as cited in appropriate professional journals or reference guides, and as directed by the attending veterinarian." 9 C.F.R. § 3.81(a).

In 1991, the ALDF challenged § 3.81 in the federal district court for the District of Columbia, arguing that the regulation violated the AWA by failing to impose minimum standards for nonhuman primate conditions of confinement. In a decision later upheld by the D.C. Circuit, the district court concluded that at least one of the plaintiffs had standing to challenge the regulation based on an aesthetic injury he suffered from witnessing the conditions of several nonhuman primates' confinement. This plaintiff had "`been employed and/or worked as a volunteer for various human and animal relief and rescue organizations'" "[f]or his entire adult life." Animal Legal Def. Fund, Inc. v. Glickman, 154 F.3d 426, 429 (D.C.Cir.1998) (en banc) ("Glickman I"). This plaintiff had visited a "game farm" covered by the statute at least nine times, and had seen primates "living under inhumane conditions" at the farm. Id. For example, he saw a large male chimpanzee held in an isolated area. He stated in an affidavit that he knew that chimpanzees "`are very social animals and it upset [him] very much to see'" this chimpanzee "`in isolation from other primates.'" Id. (alteration in original).

The district court struck down § 3.81 as arbitrary and capricious and contrary to law. Animal Legal Def. Fund v. Sec. of Agric., 813 F. Supp. 882, 892 (D.D.C.1993). Reversing this decision in 2000, the D.C. Circuit concluded that "[n]othing in the [AWA's] statutory mandate required greater specificity" than the allegedly vague requirements the regulation imposed. See Animal Legal Def. Fund, Inc. v. Glickman, 204 F.3d 229, 235 (D.C.Cir. 2000) ("Glickman II").

While the challenge to § 3.81 was pending before the D.C. Circuit in what would become Glickman II, the USDA took a survey of USDA inspectors responsible for § 3.81's enforcement. Based on this survey, it published a "Final Report on Environment Enhancement to Promote the Psychological Well-Being of Nonhuman Primates" ("Final Report") on July 15, 1999.1 The Final Report noted that "[a]lmost half the responding employees felt that the criteria in the regulations were not adequate for [regulated] facilities to understand how to meet them and for inspectors to judge if a facility was in compliance." Inspectors complained that § 3.81 provided "few solid criteria" to judge compliance, and a "common refrain" among those surveyed was that "too many enhancement programs consisted of only one or two types of enrichment . . . in an otherwise barren, stimulus-poor environment." Stressing the "urgency of these problems[,]" the Final Report insisted that "[a] strategy had to be developed to fulfill the original intent and language of the Animal Welfare Act . . . ."

On the same day the Final Report was issued, the USDA published—apparently as a response to the report—a "Draft Policy on Environment Enhancement for Nonhuman Primates" in the Federal Register and opened a period for public comment. 64 Fed.Reg. 38145 (July 15, 1999) ("Draft Policy"). In the Draft Policy's preface, the USDA explained why it had not included more detailed requirements in § 3.81 for the treatment of nonhuman primates. See id. at 38146 (declaring that the USDA "intentionally made the regulations regarding promotion of psychological well-being flexible" because "the conditions appropriate for one species do not necessarily apply to another").

However, the preface continued, "after five years of experience enforcing § 3.81," regulated entities "did not necessarily understand how to develop an environment enhancement plan that would adequately promote the psychological well-being of nonhuman primates." 64 Fed.Reg. at 38146. Hence, while the USDA "continue[d] to believe that [§ 3.81's] flexibility . . . is in the best interests of the animals[,]" it also "believe[d] that additional information on how to meet the standards in § 3.81 is necessary." Id. This additional information would remedy the "considerable disagreement" and "confusion" among the public and regulated entities as to what § 3.81 required, and it would give agency inspectors a workable set of enforcement criteria to apply. Id. In other words, the Draft Policy would "clarify[ ] what actions[the USDA] consider[s] necessary in order to comply with" § 3.81. Id.

The Draft Policy listed a number of specific features that a regulated entity could usefully include in an environmental enhancement plan. 64 Fed.Reg. at 38146. In particular, the Draft Policy provided for a safe harbor within which regulated entities could be sure that they were in compliance with the regulation and the AWA. It stated that if regulated entities "develop and follow environmental enhancement plans that are in accordance with the draft policy[,]" they would necessarily "meet the requirements of § 3.81 . . . ." Id. Regulated entities could adopt plans outside the safe harbor, however, so long as the plans complied with the regulation and the AWA. Id.

The Draft Policy remained just that—a draft—for several years. In July 2003, plaintiffs, frustrated that the USDA had failed to take any final action, sued the Secretary and other USDA officials (collectively, the "USDA") in federal district court for the Northern District of California. They contended that the USDA had failed to promulgate standards to "promote the psychological well-being of primates" in violation of the AWA, 7 U.S.C. § 2143(a), and that its failure constituted agency action "unlawfully withheld or unreasonably delayed" in violation of the APA. 5 U.S.C. § 706(1). Plaintiffs sought a declaratory judgment that the USDA was violating the AWA and the APA, as well as an order requiring the USDA to make a final decision with respect to the Draft Policy within 30 days.

In response, the USDA filed a declaration from Dr. Chester Gipson, the Deputy Administrator for Animal Care at the USDA's Animal Plant Health Inspection Service. Dr. Gipson stated that, in his official capacity as Deputy Administrator, he had made an oral announcement at a December 2002 conference of industry and animal protection groups that the "USDA was not going forward with the Draft Policy. . . ." Dr. Gipson declared that he had repeated this oral announcement at a similar March 2003 conference.

The National Association for Biomedical Research ("NABR"), a non-profit organization representing universities and other entities subject to the AWA, intervened in the suit under Rule 24(b). The USDA, joined by the NABR, then moved to dismiss under Rule 12(b)(1), contending, inter alia, that Dr. Gipson's announcements that the agency would not go forward with the Draft Policy rendered the suit moot, and that in any event plaintiffs lacked standing under Article III.

Plaintiffs then amended their complaint to add an alternative claim, contending that Dr. Gipson's oral statements constituted final agency action under 5 U.S.C. § 704. They alleged that this action was arbitrary, capricious, and an abuse of discretion in violation of 5 U.S.C. § 706(2), in that it "fails to articulate a basis for the agency's decision that the [Draft] Policy is no longer `necessary' to fulfill the agency's obligations under the AWA and also fails to address any of the evidence and public comment demonstrating that the policy is necessary to fulfill that statutory mandate." The district court granted the USDA's motion to dismiss the amended complaint. It concluded that the agency had no further duty to act after the D.C. Circuit held in Glickman II that § 3.81 adequately discharged the USDA's obligation under the AWA. Hence, in the view of the district court, the USDA's failure to finalize the Draft Policy did not constitute an "unreasonable delay" in violation of APA § 706(1). The district court also rejected plaintiffs' § 706(2) claim that the refusal to go forward with the Draft Policy had been arbitrary and capricious. It held that Dr. Gipson's oral announcements did not constitute a "definitive policy statement," and consequently were not a reviewable "agency action" within the meaning of § 706(2).

On appeal to us, plaintiffs primarily argue that the decision not to go forward with the Draft Policy was a final "agency action" under § 704, and that the district court erred by dismissing the challenge to Dr. Gipson's statements as unreviewable under § 706(2). In the alternative, they contend that the USDA's failure finally to decide the Draft Policy's fate is an unreasonable delay within the meaning of § 706(1). The USDA and the NABR contend that plaintiffs lack Article III standing and that, if they have standing, their arguments under the APA fail. Because the question of plaintiffs' Article III standing goes to our subject matter jurisdiction, we address it first.

The standing "inquiry involves both constitutional limitations on federal-court jurisdiction and prudential limitations on its exercise." Warth v. Seldin, 422 U.S. 490, 498, 95 S. Ct. 2197, 45 L. Ed. 2d 343 (1975). The USDA and the NABR contend that plaintiffs' standing fails on constitutional grounds.2 In order to satisfy Article III's standing requirements, a plaintiff must show:

(1) it has suffered an injury in fact that is (a) concrete and particularized and (b) actual and imminent, not conjectural or hypothetical; (2) the injury is fairly traceable to the challenged action of the defendant; and (3) it is likely, as opposed to merely speculative, that the injury will be redressed by a favorable decision.

Friends of the Earth, Inc. v. Laidlaw Envtl. Servs., Inc., 528 U.S. 167, 180-81, 120 S. Ct. 693, 145 L. Ed. 2d 610 (2000) (internal quotation omitted). At the motion to dismiss stage, allegations are presumed to be correct, and thus "general factual allegations of injury resulting from the defendant's conduct may suffice" to establish the prerequisites for standing. Lujan v. Defenders of Wildlife, 504 U.S. 555, 561, 112 S. Ct. 2130, 119 L. Ed. 2d 351 (1992).

We hold that at least one plaintiff, Valerie Buchanan, has standing under Article III. Buchanan is a "long-time volunteer and advocate for primates" with experience working with chimpanzees all over the world. Buchanan states that, beginning in 2000, she has formed a "particularly close emotional attachment" to Terry, a chimpanzee housed at the Southern Nevada Zoological-Botanical Park. Buchanan visits Terry "every other month" and states that she "will continue to do so" in the future. Since 1995, Buchanan declares, Terry has been housed alone, with no contact with other nonhuman primates. Terry is "listless" and engages in species-atypical behavior. The only time Buchanan has seen Terry animated is when his former trainer visits. When the trainer leaves, "Terry climbs up in his cage and watches the trainer's car until it exits the parking lot." Based on her expertise, Buchanan believes that Terry's psychological health is deteriorating. She expressed her concern about Terry to the USDA in 2001. The agency responded that, while it "would prefer that he have the companionship of another chimpanzee," the zoo was in compliance with the AWA and § 3.81.

Buchanan complains that Terry's environment and species-atypical behavior injure her interests in observing and enjoying chimpanzees. She states that, if the Draft Policy were finalized, it "would improve the psychological well-being of primates including Terry, thereby reducing the level of abnormal, stereotypic and/or injurious behaviors exhibited by Terry and other primates whom . . . Buchanan enjoys observing and studying in humane conditions." The USDA argues that Buchanan's asserted injury does not satisfy Article III. For the reasons that follow, we disagree.

"Harm to a plaintiff's `aesthetic and environmental well-being' has long been recognized as a cognizable injury" for the purposes of the injury-in-fact requirement. Fund for Animals, Inc. v. Lujan, 962 F.2d 1391, 1396 (9th Cir.1992) (quoting Sierra Club v. Morton, 405 U.S. 727, 734, 92 S. Ct. 1361, 31 L. Ed. 2d 636 (1972)). "[T]he desire to use or observe an animal species, even for purely esthetic purposes, is undeniably a cognizable interest for purpose of standing." Defenders of Wildlife, 504 U.S. at 562-63, 112 S. Ct. 2130. The term "aesthetic" in ordinary usage refers to an artistic sense or a sense of beauty. Our Article III standing jurisprudence gives the term "aesthetic injury" a broader meaning that includes emotional injury. For example, in Glickman I, the D.C. Circuit held that the plaintiff—who had stated that the treatment of the chimpanzee had "upset" him "very much"—had suffered "aesthetic injury." 154 F.3d at 431, 433.

Buchanan's aesthetic injury is virtually indistinguishable from the injury the D.C. Circuit found constitutionally sufficient in Glickman I. Like Buchanan, the Glickman I plaintiff asserted that he had suffered an "`assault on [his] senses'" and had been very upset when he viewed nonhuman primates "living under inhumane conditions" during regular visits to a particular zoo. 154 F.3d at 429-32. Like Buchanan, whose many years of interest in chimpanzees underscore a deep-seated concern for their health, the Glickman I plaintiff had a long-term sustained interest in primate welfare, and had personally visited and— more importantly—continued to visit the animals whose condition caused him aesthetic injury. Id. at 432. Hence, Buchanan, like the plaintiff in Glickman I, has "alleged far more than an abstract, and uncognizable, interest in seeing the law enforced." Id.

The USDA urges us to disagree with the D.C. Circuit's holding in Glickman I. It asks us to restrict Article III aesthetic injury based on animal harm to situations where the diminution or elimination of an animal species causes the alleged harm. This is precisely the argument made by the dissent in Glickman I. See 154 F.3d at 448 (Sentelle, J., dissenting) (the "diminution-of-the-species" limit is necessary to keep the "expanse of standing" from being "bounded only by what a given plaintiff finds to be aesthetically pleasing").

The Glickman I majority expressly rejected this argument. See 154 F.3d at 433 n. 5. It is also inconsistent with our own precedent. In Alaska Wildlife Alliance v. Jensen, 108 F.3d 1065 (9th Cir.1997), we held that a plaintiff who alleged that she saw "`sea lions in the bay with huge trolling lures hanging from their mouths'" had standing based on aesthetic injury to contest the Secretary of the Interior's decision to allow commercial fishing in a national park. Id. at 1068. A plaintiff who "expressed concern over [fishing] vessels' displacement of whales from preferred feeding areas and described how she now plans her visits to the Park to avoid the fishermen's presence" also satisfied the injury-in-fact requirement. Id.; see also Ocean Advocates v. U.S. Army Corps of Eng'rs, 402 F.3d 846, 859-60 (9th Cir.2005) (finding standing based on risk of harm to the environment when the alleged harm would, among other things, "markedly decrease[ ]" the plaintiffs' ability to "observe wildlife").

For purposes of injury-in-fact, a distinction between harm to individual animals, on the one hand, and harm to an animal species through diminution or extinction, on the other, is illogical. The injury at issue is not the animals' but the human observer's. The government fails to explain—as it failed to explain in Glickman I—why knowing of a species' diminution or extinction causes a more acute and objectively demonstrable injury to a human than observing harm to an individual animal. See 154 F.3d at 438. Although we appreciate that aesthetic injury standing is not boundless, we conclude that harm to a nonhuman primate creates a cognizable injury-in-fact for someone with Buchanan's level of engagement with and appreciation for nonhuman primates.

The second prong of the test for Article III standing requires Buchanan to show a causal connection between her injury and the USDA's refusal to finalize the Draft Policy. Put differently, her injury must be "fairly traceable" to the agency's failure to act. Pritikin v. Dep't of Energy, 254 F.3d 791, 797 (9th Cir.2001). Injury that "results from the independent action of some third party not before the court" does not satisfy the causation requirement. Simon v. E. Ky. Welfare Rights Org., 426 U.S. 26, 42, 96 S. Ct. 1917, 48 L. Ed. 2d 450 (1976). But a "chain of causation [may have] more than one link" and still satisfy Article III, so long as the connection between injury and cause is not "hypothetical or tenuous." Nat'l Audubon Soc'y, Inc. v. Davis, 307 F.3d 835, 849 (9th Cir.2002).

Terry's suffering, caused by the conditions at the zoo where he is confined, gives rise to Buchanan's injury. Obvious causal links connect Terry's suffering (and thus Buchanan's injury) to the USDA's failure to adopt the Draft Policy. The Draft Policy addresses several of the problems at Terry's zoo that, according to Buchanan, render him listless and unhappy. For example, Buchanan complains that Terry's lack of companionship adversely affects his psychological health. The Draft Policy declares that, "[i]n order to address the social needs of nonhuman primates under § 3.81(a), the [zoo's environmental enhancement] plan must provide for each primate of a species known to be social in nature to be housed with other primates whenever possible." 64 Fed.Reg. at 38147 (emphasis added). Although the Draft Policy would have given regulated entities leeway to meet this requirement in different ways, the mandatory language would have eliminated their discretion to house primates in isolation when possible to do otherwise. By contrast, without the Draft Policy clarifying § 3.81's requirements, the USDA could freely give Terry's zoo a clean bill of health when Buchanan complained in 2001. Terry's suffering—and thus Buchanan's injury—fairly traceable to the fact that his zoo can operate in ways that the Draft Policy likely would have restricted or eliminated.

The final requirement for Article III standing is that the relief Buchanan seeks in this suit redress her injury. Causal connection and redressability are two sides of the same coin. The former assesses the relationship between the defendant's conduct and the alleged injury, while the latter asks whether and how the requested relief would help. Allen v. Wright, 468 U.S. 737, 753 n. 19, 104 S. Ct. 3315, 82 L. Ed. 2d 556 (1984). A plaintiff meets the redressability test if it is "likely"—not certain—"that the injury will be redressed by a favorable decision." Laidlaw, 528 U.S. at 181, 120 S. Ct. 693; Defenders of Wildlife, 504 U.S. at 561, 112 S. Ct. 2130. Although "[a] purely speculative favorable outcome will not suffice[,]" Rubin v. City of Santa Monica, 308 F.3d 1008, 1020 (9th Cir.2002) (internal quotation omitted), a "[p]laintiff[ ] need not demonstrate that there is a `guarantee' that [her] injuries will be redressed by a favorable decision." Graham v. Fed. Emergency Mgmt. Agency, 149 F.3d 997, 1003 (9th Cir.1998).

In the Draft Policy's preface, the USDA declared that, for effective application and enforcement of § 3.81, additional information and instruction of the sort contained in the policy was "necessary." 64 Fed.Reg. at 38146. If Buchanan's suit is successful, and the USDA's decision to abandon the Draft Policy is set aside as arbitrary and capricious, the USDA would either have to explain why the Draft Policy is no longer "necessary," or would actually have to adopt it. If the agency cannot provide a satisfactory explanation and ultimately adopts the Draft Policy, the policy's threat of penalty if its safe harbor is ignored is likely to provide substantial incentive to comply with its terms.3

In some standing cases, a court has little more than common sense to guide its determination whether the relief sought will likely redress an injury. That is not so here. We are in the fortunate position of having an unambiguous statement by the regulated entities themselves that the relief Buchanan seeks will likely alleviate Terry's suffering. The NABR—which has intervened as the representative of regulated entities—insists in its brief to us that regulated entities would be effectively compelled to follow the Draft Policy if it were adopted. The NABR made a similar statement in its motion to intervene in the district court: "Any actual change in the challenged regulations, resulting from [Plaintiffs'] action, would directly result in new requirements with which NABR's members must comply under the AWA." (Emphasis added). These words leave little doubt that, at least in the view of the regulated entities, the Draft Policy would require them to take steps that plaintiffs allege would lessen or eliminate their injuries.

The Draft Policy, if adopted, would "alter the legal regime to which" regulated entities' action is subject. Bennett v. Spear, 520 U.S. 154, 169, 117 S. Ct. 1154, 137 L. Ed. 2d 281 (1997). Regulated entities would have a choice—either to follow the Draft Policy's proposed environmental enhancement standards and avail themselves of the safe harbor, or not to follow them and risk penalties for noncompliance. Fear of adverse consequences that motivates otherwise voluntary action can satisfy the redressability prong. See, e.g., McClure v. Ashcroft, 335 F.3d 404, 411 (5th Cir.2003); see also Bennett, 520 U.S. at 169, 117 S. Ct. 1154 (Article III "does not exclude injury produced by determinative or coercive effect upon the action of someone else."). Even if the relief sought serves an "advisory function," the plaintiff establishes redressability if it might have a "powerful coercive effect" on regulated entities. Id.; see also Tozzi v. U.S. Dep't of Health & Human Servs., 271 F.3d 301, 309 (D.C.Cir.2001) (holding that redressability is satisfied when relief would be "highly influential" in shaping a regulated entity's behavior).

There is another argument for redressability, arising from Buchanan's claim that the USDA's withdrawal of the Draft Policy was arbitrary and capricious because it "failed[ ] to articulate a basis for the agency's decision that the [Draft] Policy is no longer `necessary.'" In a similar context— suits involving agencies' alleged failure to comply with the procedural requirements of the National Environmental Policy Act—we have held that redressability is satisfied where procedural compliance "could have influenced" an agency's action. Citizens for Better Forestry v. U.S. Dep't of Agric., 341 F.3d 961, 976 (9th Cir.2003). Because requiring the USDA to offer a rationale for not going forward with the Draft Policy could influence its decision, Buchanan's injury is also redressable on this ground.

We therefore conclude that Buchanan satisfies the redressability requirement for Article III standing.

Because Buchanan has Article III standing, we do not need to consider whether the other individual plaintiffs have it as well. Arlington Heights v. Metro. Hous. Dev. Corp., 429 U.S. 252, 264 & n. 9, 97 S. Ct. 555, 50 L. Ed. 2d 450 (1977). Nor do we need to determine whether the ALDF or the AWI meets the requirements for organizational standing. Laub v. U.S. Dep't of the Interior, 342 F.3d 1080, 1086 (9th Cir.2003).

As indicated by Dr. Gipson's oral announcements, and by the USDA's defense in this case, it is clear that the agency has decided to abandon the Draft Policy. Thus, with the case in its current posture, plaintiffs cannot seek prospective relief under 5 U.S.C. § 706(1) for the agency's unreasonable delay in deciding the Draft Policy's fate. The question remains whether the district court should have reviewed the withdrawal of the Draft Policy as a "final agency action" under the arbitrary and capricious standard of § 706(2).

The APA authorizes federal courts to review a "final agency action," 5 U.S.C. § 704, to determine if it is "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law[.]" 5 U.S.C. § 706(2)(A). The APA "creates a strong presumption of reviewability that can be rebutted only by a clear showing that judicial review would be inappropriate." Nat'l Res. Defense Council, Inc. v. Sec. & Exch. Comm'n, 606 F.2d 1031, 1043 (D.C.Cir.1979) ("NRDC").

The district court determined that it lacked authority to review Dr. Gipson's statements because they did not amount to "definitive policy statements." The issue here, however, is whether the USDA's decision to abandon the Draft Policy, not Dr. Gipson's statements, is reviewable "final agency action" within the meaning of § 704.

The term "agency action" "includes the whole or a part of an agency rule, order, license, sanction, relief, or the equivalent or denial thereof, or failure to act[.]" 5 U.S.C. § 551(13). "[T]he term . . . undoubtedly has a broad sweep[,]" but it "is not so all-encompassing as to authorize [courts] to exercise `judicial review [over] everything done by an administrative agency.'" Indep. Equip. Dealers Ass'n v. Envtl. Prot. Agency, 372 F.3d 420, 427 (D.C.Cir.2004) (quoting Hearst Radio, Inc. v. FCC, 167 F.2d 225, 227 (D.C.Cir.1948)).

"As a general matter," two conditions must be satisfied for an agency action to be deemed "final":

First, the action must mark the consummation of the agency's decisionmaking process . . .—it must not be of a merely tentative or interlocutory nature. And second, the action must be one by which rights or obligations have been determined, or from which legal consequences will flow.

Bennett, 520 U.S. at 177-78, 117 S. Ct. 1154 (1997) (internal quotations omitted). Put differently, an agency action is "final" if it is "definitive" and has a "direct and immediate . . . effect on the day-to-day business" of the party challenging or subjected to the action. Fed. Trade Comm'n v. Standard Oil Co., 449 U.S. 232, 239, 101 S. Ct. 488, 66 L. Ed. 2d 416 (1980) (internal quotations omitted). If an agency action is "binding" on the parties it targets, it is likely "final" for the purposes of judicial review. See 5 Jacob A. Stein et al., Administrative Law § 43.01 (2005).

"Finality" should be interpreted in a "`flexible and pragmatic way.'" Her Majesty the Queen ex rel. Ontario v. Envtl. Prot. Agency, 912 F.2d 1525, 1531 (D.C.Cir.1990) (citing Ciba-Geigy Corp. v. Envtl. Prot. Agency, 801 F.2d 430, 435 (D.C.Cir.1986)). An agency's refusal to describe an action as final is not dispositive of reviewability. Id. Nor does an agency's failure to offer a formal statement explaining its action deprive that action of finality. See id. To the contrary, an action may be reviewed, and set aside as arbitrary and capricious, precisely because the agency has failed "to give a reasoned account of its decision." Fox Tel. Stations, Inc. v. Fed. Communications Comm'n, 280 F.3d 1027, 1045 (D.C.Cir.2002).

Of course, having acted in regard to the "whole or a part of an agency rule, order, license, sanction, relief, or the equivalent or denial thereof," an agency remains free to engage in further policymaking activity. But the possibility of future agency action is not sufficient to foreclose review of a definitive action. Otherwise, "review could be deferred indefinitely." Am. Petroleum Inst. v. Envtl. Prot. Agency, 906 F.2d 729, 739-40 (D.C.Cir.1990); see also Fox, 280 F.3d at 1037-38 (rejecting Commission's argument that action "is not final because the agency intends to continue considering the ownership rules").

The classification of an agency's statements and declarations into categories of reviewable and non-reviewable pronouncements is an "imprecise" endeavor that requires a "case-by-case" assessment. Indus. Safety Equip. Ass'n v. Envt'l Prot. Agency, 837 F.2d 1115, 1117 (D.C.Cir. 1988). But there is no doubt that the USDA's abandonment of the Draft Policy is a consummation of its decisionmaking process. There is nothing interlocutory or tentative about Dr. Gipson's statements. The availability of judicial review depends on whether this agency action determined rights or obligations or triggered legal consequences.

B. Legislative Rules, Interpretive Rules, and Policy Statements

In the usual case, a court must determine whether an agency's affirmative act constitutes final agency action for purposes of judicial review. Our task is different, for we must determine whether the USDA's decision to abandon a course of action is reviewable. We divide our analysis into two parts: First, would the Draft Policy, if adopted, have been reviewable? Second, is the USDA's decision not to adopt that Draft Policy reviewable?

1. Reviewability of the Draft Policy if Adopted

A reviewable "agency action" includes an "agency rule." 5 U.S.C. § 551(13). The APA defines "rule" as "the whole or a part of an agency statement of general or particular applicability and future effect designed to implement, interpret, or prescribe law or policy . . . ." 5 U.S.C. § 551(4). Plainly, the Draft Policy is (1) an "agency statement" (2) of "general . . . applicability" and "future effect" (3) designed to "implement" and "interpret" § 3.81 and the AWA. If adopted, it would have constituted a "rule" within the meaning of § 551(4).

The term "rule" may embrace "virtually every statement an agency may make," Avoyelles Sportsmen's League, Inc. v. Marsh, 715 F.2d 897, 908 (5th Cir.1983), but not all agency statements are reviewable. Although the distinction among them is "`enshrouded in considerable smog,'" Am. Bus Ass'n. v. United States, 627 F.2d 525, 529 (D.C.Cir.1980) (quoting Noel v. Chapman, 508 F.2d 1023, 1030 (2d Cir. 1975)), there are essentially three kinds of agency rules: legislative rules, interpretive rules, and policy statements. See Robert A. Anthony, Interpretive Rules, Policy Statements, Guidances, Manuals and the Like—Should Federal Agencies Use Them to Bind the Public?, 41 Duke L.J. 1311, 1315 (1992); see also Attorney General's Manual on the Administrative Procedure Act 30 n. 3 (1947) ("Attorney General's Manual"). The USDA's chosen title for the document at issue here—"Draft Policy"—suggests that it falls into the third category. However, "the agency's own label, while relevant, is not dispositive." Gen. Motors Corp. v. Ruckelshaus, 742 F.2d 1561, 1565 (D.C.Cir.1984). More important for classification purposes are the ends the agency document serves and the effects it will have on regulated entities.

Legislative rules "create new law, rights or duties," are promulgated pursuant to an exercise of delegated power, and require a notice-and-comment period. Ruckelshaus, 742 F.2d at 1565; David v. Donovan, 698 F.2d 1057, 1058 (9th Cir.1983). They indisputably constitute "agency action" and are subject to judicial review, for they determine rights and obligations and have legal consequences. See Gen. Motors Corp. v. Envtl. Prot. Agency, 363 F.3d 442, 448 (D.C.Cir.2004).

Interpretive rules, as defined by the 1947 Attorney General's Manual, are "rules or statements issued by an agency to advise the public of the agency's construction of the statutes and rules which it administers." Attorney General's Manual at 30 n.3. An interpretive rule is issued "to clarify or explain existing law or regulations so as to advise the public of the agency's construction of the rules it administers." Gunderson v. Hood, 268 F.3d 1149, 1154 (9th Cir.2001) (emphasis added). It "spells out a duty fairly encompassed within the regulation that the interpretation purports to construe." Paralyzed Veterans of Am. v. D.C. Arena L.P., 117 F.3d 579, 588 (D.C.Cir.1997).

Interpretive rules lack formal status as law. See Syncor Int'l Corp. v. Shalala, 127 F.3d 90, 94 (D.C.Cir.1997). As a practical matter, however, an interpretive rule may affect the regulatory practices of an agency or the expectations of a regulated entity as to what a law or legislative rule means and how it will be enforced. When an interpretive rule has "a substantial impact on the rights of individuals[,]" its promulgation may constitute final agency action for the purposes of judicial review. Am. Postal Workers Union v. U.S. Postal Serv., 707 F.2d 548, 560 (D.C.Cir.1983); see also Nat'l Ass'n of Home Builders v. Norton, 415 F.3d 8, 15 (D.C.Cir.2005) ("Finality resulting from the practical effect of an ostensibly non-binding agency proclamation is a concept [courts] have recognized in the past."). We and our sister circuits have construed interpretive rules to be judicially reviewable. See, e.g., U.S. West Communications, Inc. v. Hamilton, 224 F.3d 1049, 1055 (9th Cir.2000) (observing that "[w]hat scant case law there is suggests that [reviewable] `final orders' include both interpretive and legislative orders"); Star Enter. v. Envtl. Prot. Agency, 235 F.3d 139, 146 (3d Cir.2000) (describing standard of review for interpretive rule). Cf. Martin v. Occupational Safety & Health Review Comm'n, 499 U.S. 144, 157, 111 S. Ct. 1171, 113 L. Ed. 2d 117 (1991) (observing that interpretive rules are "entitled to some weight on judicial review").

Finally, "general statements of policy" are "statements issued by an agency to advise the public prospectively of the manner in which the agency proposes to exercise a discretionary power." Attorney General's Manual 30 at n.3. In other words, they may tell the public how the agency plans to exercise its enforcement discretion. A policy statement is intended "to allow agencies to announce their tentative intentions for the future . . . without binding themselves." Am. Hosp. Ass'n v. Bowen, 834 F.2d 1037, 1046 (D.C.Cir.1987) (internal quotation omitted) (emphasis added); see also Guardian Fed. Sav. & Loan Ass'n v. Fed. Sav. & Loan Ins. Corp., 589 F.2d 658, 666 (D.C.Cir.1978) ("A general statement of policy . . . does not establish a binding norm. It is not finally determinative of the issues or rights to which it is addressed.") (internal quotations omitted). "[T]he agency retains the discretion and the authority to change its position—even abruptly—in any specific case because a change in its policy does not effect the legal norm." Syncor, 127 F.3d at 94.

General statements of policy are tentative in character: the agency, having announced how it plans to exercise its enforcement discretion, may nonetheless change course in any particular case. This tentative character generally renders general statements of policy insufficiently final or definitive to permit judicial review. Hence, a "typical policy statement" is "not reviewable at all." Tozzi, 271 F.3d at 312 (Silberman, J., concurring); see also Indus. Safety Equip., 837 F.2d at 1119 n. 8 ("Discretionary agency positions are generally best not tested on review until the policy is actually applied . . . ."); Office of Communication of United Church of Christ v. Fed. Communications Comm'n, 826 F.2d 101, 105-06 (D.C.Cir.1987) (nonbinding, prospective quality of policy statement makes it unripe for judicial review).

The Draft Policy, if finalized, would not have amounted to a legislative rule. Its purpose—merely to "clarify" existing obligations prescribed by the AWA and § 3.81—is the most obvious evidence to this effect. 64 Fed.Reg. at 38146. The Draft Policy would have added considerable detail to § 3.81, but it would not have announced any new requirements for regulated entities. To the contrary, even if the Draft Policy were adopted, a regulated entity could still comply with § 3.81 without following its specific recommendations for conditions of captivity.

But neither would the finalized Draft Policy have been a general policy statement—that is, a tentative declaration of how the USDA might have enforced § 3.81 and the AWA. Rather, it is best characterized as a proposed interpretive rule. The Draft Policy would have summarized "what the[USDA] believe[d] must be considered and included" in an environmental enhancement plan. 64 Fed.Reg. at 38146 (emphasis added). It would have provided categorically that entities that follow the specific terms of its safe harbor "will meet the requirements of § 3.81." Id. at 38146 (emphasis added). Compare Am. Bus Ass'n, 627 F.2d at 532 (use of "will" indicates statement is a binding norm), with Guardian Fed., 589 F.2d at 666 (use of term "may" indicates statement is a "general statement of policy"). Such safe harbors, which offer regulated entities a guaranteed way to protect themselves from adverse enforcement, have legal consequence. See Gen. Elec. Co. v. Envtl. Prot. Agency, 290 F.3d 377, 383 (D.C.Cir.2002) ("`In some circumstances, if the language of the document is such that private parties can rely on it as a norm or safe harbor by which to shape their actions, it can be binding as a practical matter.'") (quoting Anthony, Interpretive Rules, 41 Duke L.J. at 1329) (emphasis added); Atlantic Richfield Co. v. Fed. Energy Admin., 556 F.2d 542, 552 (1977) (observing that an agency document that lacks formally binding effect "may have extremely important consequences with respect to the issue of good faith reliance for future acts" and therefore constitute "final agency action").

In short, the USDA would have bound itself to a particular interpretation of § 3.81 for enforcement purposes had it adopted the Draft Policy. Such restrictions on an agency's regulatory discretion are features of interpretive rules, not general policy statements. See Syncor, 127 F.3d at 94 ("The primary distinction between . . . any rule . . . and a general statement of policy, then, turns on whether an agency intends to bind itself to a particular legal position.").

We conclude that the Draft Policy, if adopted, would have been judicially reviewable as an interpretive rule from which legal consequence flowed. Our conclusion is buttressed by the Draft Policy's publication in the Federal Register, and the fact that the USDA opened a public comment period for it. Both of these procedural attributes would have strengthened the Draft Policy's claim to binding authority had it been adopted. See Molycorp, Inc. v. Envtl. Prot. Agency, 197 F.3d 543, 545 (D.C.Cir.1999); Rank v. Nimmo, 677 F.2d 692, 698 (9th Cir.1982). Moreover, a record assembled through a notice-and-comment period enables judicial review by providing courts with materials necessary to evaluate the rule's sufficiency. See Int'l Union, UMW v. Mine Safety & Health Admin., 407 F.3d 1250, 1259 (D.C.Cir.2005).

2. Reviewability of the Withdrawal of the Draft Policy

We now turn to the second part of our inquiry—whether the USDA's abandonment of the Draft Policy is judicially reviewable. This part of the inquiry raises two interrelated questions. First, did the abandonment of a Draft Policy have legal consequences or determine rights or obligations? Second, does it make a difference that the USDA had no legal obligation to propose or adopt the Draft Policy in the first place?

a. Legal Consequences or Determination of Rights or Obligations

The abandonment of a proposed rule such as the Draft Policy leaves the status quo in place. But maintaining the status quo has legal consequences. Our decision in Defenders of Wildlife v. Norton, 258 F.3d 1136 (9th Cir.2001), illustrates this point. The Endangered Species Act ("ESA") requires that the Secretary of the Interior designate as "endangered" or "threatened" any species at risk of extinction. 16 U.S.C. § 1532(6). The Secretary withdrew a proposed rule that would have designated the flat-tailed horned lizard an endangered species, thereby leaving the status quo undisturbed. 258 F.3d at 1138-39. The consequences of leaving the status quo undisturbed were obvious. Because it denied the flat-tailed horned lizard the protections afforded to endangered species by the ESA, the Secretary's decision, if incorrect, exposed the lizard to risk of extinction. We therefore reviewed the Secretary's decision.

It is clear that the Draft Policy would have determined rights and obligations of the regulated entities. A regulated entity in doubt about its obligations under § 3.81 and the AWA would have known, after the adoption of the Draft Policy, precisely what it could have done to ensure that it fulfilled its obligations. Compliance with the safe harbor requirements would have protected against adverse enforcement actions by the USDA. Indeed, as we have been told by the intervenor NABR, the Draft Policy would have had the practical consequence of strongly encouraging, perhaps even requiring, regulated entities to comply with its terms. Thus, it clearly would have had a legal consequence if adopted.

The converse is also true. Without the Draft Policy, there is no safe harbor. As the Final Report indicated, the practical effect is that regulated entities have little incentive to change their behavior. Because the abandonment of the Draft Policy means that regulated entities will not change their behavior, the status quo remains undisturbed. The consequence is that harm to chimpanzees like Terry continues, and aesthetic injury to individuals like Buchanan also continues.

Of course, the USDA may promulgate other Draft Policies in the future, as Dr. Gipson has suggested in his declaration it intends to do. For now, however, the agency's last word is that the Draft Policy is not necessary, with the effect that the Draft Policy has been abandoned, and that the status quo remains unchanged. See Fox, 280 F.3d at 1038.

b. USDA Had No Obligation to Act in the First Place

The conclusion that the failure to adopt the Draft Policy had legal consequence is not necessarily the end of the inquiry. The USDA relies on the D.C. Circuit's decision in Glickman II, which held that § 3.81 adequately discharged the USDA's rulemaking obligation under the AWA. Because it had no statutory duty to propose the Draft Policy in the first place, the USDA argues, its decision to abandon the policy is unreviewable. We disagree.

We do not question the conclusion of the D.C. Circuit in Glickman I. Because § 3.81 passed muster standing on its own, the USDA's decision to propose the Draft Policy was not compelled by the AWA. "An agency's discretionary decision not to regulate a given activity is inevitably based, in large measure, on factors not inherently susceptible to judicial resolution, e.g., internal management considerations as to budget and personnel; evaluations of its own competence; weighing of competing policies within a broad statutory framework." NRDC, 606 F.2d at 1046. Moreover, "even if an agency considers a particular problem worthy of regulation, it may determine for reasons lying within its special expertise that the time for action has not yet arrived." Id. Judicial second-guessing of such decisions triggers concerns of over-reaching, particularly when the agency has already deemed the regulation unworthy of adoption. Id. at 1045.

Nonetheless, courts have reviewed, albeit highly deferentially, the withdrawal of proposed rules like the Draft Policy. An example is Center for Auto Safety v. National Highway Traffic Safety Administration, 710 F.2d 842 (D.C.Cir.1983). A statute provided that the National Highway and Traffic Safety Administration ("NHTSA") "`may by rule ... amend the average fuel economy standard ... to a level which [it] determines is the maximum feasible average fuel economy level ....'" Id. at 844 (quoting 15 U.S.C. § 2002(a)(4)) (emphasis added). The agency refused to finalize a proposed regulation that would have tightened fuel economy standards for certain cars. Noting the reasons the NHTSA gave for withdrawing its proposed regulation, the court concluded that

[t]hese statements accompanying the withdrawal of the [proposed rule] clearly interpret the relevant statute and indicate NHTSA's policy regarding the exercise of discretion granted to it by that legislative enactment. Given our prior decisions construing the Administrative Procedure Act's broad definition of "rule," we are compelled to conclude that NHTSA has prescribed a rule sufficient to grant this court jurisdiction ....

Id. at 846. In other words, the retreat to the status quo as the agency's policy was reviewable, even though the statute at issue, by using the term "may," allowed the agency the choice whether to engage in rulemaking in the first place.

The D.C. Circuit's decision in Professional Drivers Council v. Bureau of Motor Carrier Safety, 706 F.2d 1216 (D.C.Cir. 1983), is also instructive. There, the agency withdrew a proposed amendment to an existing rule that specified hours-of-service requirements for commercial truck drivers. The decision to withdraw the proposed rule did not take place against a legal obligation to act:

This is not a situation where the agency has shirked its statutory duty by refusing to regulate. The agency has regulated the field. In this instance, the agency's statutory authority to regulate is permissive. The statute provides that the agency "may establish reasonable requirements." The permissive nature of the statute implies broad agency discretion in selecting the appropriate manner of regulation.

Id. at 1221. The D.C. Circuit recognized the limited scope of judicial inquiry into this sort of agency decisionmaking:

The circumscribed scope of this review is dictated by both the nature of the administrative proceeding (informal rulemaking) and by the nature of the ultimate decision (not to promulgate rules). The record in an informal rulemaking proceeding is a less than fertile ground for judicial review .... Furthermore, rulemaking is an inherently policy-oriented process and the agency must be accorded considerable deference in evaluating information presented and reaching decisions based upon its expertise. Our review is also circumscribed by the fact that the agency decided not to promulgate new rules in an area already heavily regulated.

Id. at 1220-21 (internal quotation omitted) (emphasis added). Nonetheless, the court in the end reaffirmed its authority to exercise judicial review. It gave effect to the agency's interest in having relatively unfettered power to decide whether to adopt rules not by abandoning its reviewing authority altogether but by exercising a highly deferential standard of review. See id.

Finally, International Union, United Mine Workers of America v. United States Department of Labor demonstrates that once an agency has embarked on a course of action, even a discretionary course of action, a court may review the agency's failure to explain its subsequent abandonment of that course of action. There, the D.C. Circuit reviewed the withdrawal of a proposed Air Quality rule by the Mine Safety and Health Administration ("MSHA") after a public hearing and comment period. 358 F.3d at 41-42. The D.C. Circuit explained that while the MSHA was under no obligation to adopt the proposed rule, "or, for that matter, any rule," it was not "free to terminate the rulemaking for no reason whatsoever." Id. at 43-44 (quotation marks omitted). The D.C. Circuit reviewed the withdrawal of the proposed rule under the APA's "deferential `arbitrary and capricious' standard," stating that it would "give more deference to an agency's decision to withdraw a proposed rule than we give to its decision to promulgate a new rule or to rescind an existing one." Id. at 43. Despite this heightened deference, the D.C. Circuit held that the withdrawal was arbitrary and capricious because the MSHA's explanations—that "there was a `change in agency priorities,'" the record was stale, and a recent Eleventh Circuit opinion cast doubt on the proposed rule—were inadequate. Id. at 44-45.

The D.C. Circuit has offered two helpful criteria to determine when the withdrawal of a voluntarily proposed rule may be amenable to at least a "minimal level of judicial scrutiny." NRDC, 606 F.2d at 1047. The agency must have "held a rulemaking proceeding," and it must have "compiled a record narrowly focused on the particular rules suggested but not adopted." Id. In NRDC, the D.C. Circuit explained that the concern that a court may improperly oblige an agency to spend resources defending a withdrawn rule is diminished when the agency has undertaken extensive rulemaking proceedings before it abandons a proposed rule. Id. at 1045-46. Such activity "evidences [the agency's] view that the proposals are sufficiently meritorious to warrant further investigation ...." Id. at 1046. Having solicited public participation, the prospect of judicial review would also force the agency to give due regard to citizen input. Id. Moreover, concerns over a court's lack of institutional competence are lessened to the degree that the pre-withdrawal rulemaking proceeding has been extensive and has thereby provided more data to the reviewing court. Id. at 1047.

In essence, the criteria help to distinguish between two classes of cases, one ordinarily reviewable and the other not reviewable. In the first class, an agency undertakes an action—for example, promulgation of a Draft Policy—but subsequently changes course and abandons that action. Properly applied, the criteria ensure that a sufficient record exists for a court to identify a discrete change of course, and to review the change of course to determine whether it was arbitrary and capricious. Id. at 1046-47; see also Fox, 280 F.3d at 1037 ("There is no question [an agency] determination not to repeal or to modify a rule, after giving notice of and receiving comment upon a proposal to do so, is a final agency action subject to judicial review.").

In the second class of cases, an agency simply declines to take any action at all. Absent a particular statutory mandate that the agency act, these cases are considered presumptively unreviewable in part because courts lack a "focus for judicial review ... to determine whether the agency exceeded its statutory powers." Heckler v. Chaney, 470 U.S. 821, 832, 105 S. Ct. 1649, 84 L. Ed. 2d 714 (1985); see also id. at 834, 838, 105 S. Ct. 1649. Courts have no means to choose between the "infinite number of rules that an agency could adopt in its discretion," NRDC, 606 F.2d at 1046, other than undertaking the "complicated balancing of a number of factors" ordinarily considered to be "peculiarly within [the agency's] expertise." Chaney, 470 U.S. at 831, 105 S. Ct. 1649. The D.C. Circuit's analysis in NRDC is fully consistent with the Supreme Court's later decision in Chaney, for NRDC concerned an abandonment of a course of action whereas Chaney concerned a mere failure to act.

We endorse and apply the D.C. Circuit's criteria in NRDC. Both are satisfied here, and thus this case falls within the class of reviewable cases. The USDA gave notice by publishing the Draft Policy in the Federal Register on July 15, 1999, and it solicited comments from the public. Plaintiffs allege that, by the time the comment period ended on October 13, 1999, see 66 Fed. Reg. 61327, 61364 (Dec. 3, 2001), the agency had received more than 200 sets of comments. Moreover, the comments solicited were to address the "content of the draft policy," and thus focused narrowly on the proposed rule. 64 Fed.Reg. at 38147.

The USDA's decision to abandon the Draft Policy is definitive and final. By leaving the status quo in place, the USDA gives regulated entities no incentive to adopt environmental enhancement plans that comply with a set of standards the agency at one time deemed "necessary" for the effective implementation of § 3.81. The USDA held rulemaking proceedings before it abandoned the Draft Policy and therefore assembled a record amenable to judicial review. We therefore conclude that the abandonment of the Draft Policy constitutes a final agency action for the purposes of § 706(2).

We conclude that plaintiff Buchanan has standing under Article III based on the aesthetic injury she has described. We further conclude that the USDA's withdrawal of the Draft Policy constitutes final agency action that is reviewable pursuant to § 706(2) of the APA. We therefore reverse the district court's dismissal of this case. On remand, the district court shall determine whether the USDA's withdrawal of the Draft Policy was arbitrary and capricious in violation of § 706(2) of the APA.

REVERSED and REMANDED.

In holding that we can review withdrawal of proposed regulations an agency had no duty to adopt, my colleagues overlook the sea-change in administrative law wrought by Heckler v. Chaney, 470 U.S. 821, 831-32, 105 S. Ct. 1649, 84 L. Ed. 2d 714 (1985), which held that we have no authority to review an agency's discretionary decision not to act. See The Supreme Court—Leading Cases: Judicial Review of Agency Inaction, 99 Harv. L.Rev. 264, 269 (1985) (hereafter "Judicial Review of Agency Inaction") ("Chaney represents a significant departure from the trend of recent administrative law."). Failure to appreciate the fundamental distinction between agency action and inaction permeates every aspect of the majority's analysis, starting with the threshold question of standing.

1. The majority wastes much toner trying to show that USDA's Draft Policy at issue here would, if adopted, ameliorate the conditions plaintiffs complain about. Maj. op. at 834-836. Even if this were true, which it is not, see pp. 845-847 infra, it doesn't matter. In determining whether plaintiffs have standing, the question is not whether the agency's proposed action would have benefitted them, but whether this lawsuit will do so. The Draft Policy only matters if the court can order, or at least move the agency towards, its adoption.

Certainly we cannot order USDA to adopt the Draft Policy.1 No court I am aware of has ever ordered an agency to adopt a particular regulation, and for good reason.2 Adoption of regulations is an executive function that no court has authority to perform. Even where the agency is required by law to act, and a court finds that it failed to do so, the court cannot tell the agency what regulation to adopt; rather, it must remand for further action. See, e.g., Envtl. Def. Ctr., Inc. v. EPA, 344 F.3d 832, 863 (9th Cir.2003); Defenders of Wildlife v. Norton, 258 F.3d 1136, 1146-47 (9th Cir.2001). In such cases it can at least be said that the court nudged the agency towards a result favorable to the plaintiffs by correcting the agency's misinterpretation of the statute and thus narrowing its discretion. But the agency here has no duty to act; it fully discharged its statutory responsibility by adopting the regulations approved by the DC Circuit. See Animal Legal Def. Fund, Inc. v. Glickman, 204 F.3d 229, 235 (D.C.Cir.2000) ("Glickman II"). The most we can do is tell the agency that its reason for rejecting the Draft Policy is insufficient. See maj. op. at 844. This won't ensure that the agency will adopt the Draft Policy or narrow its discretion in any meaningful way. The agency is free to open up the proceedings for further comment, adopt a different policy, or again reject the policy, this time giving better reasons.3

I suppose that plaintiffs could then bring another lawsuit and we could, once again, disapprove the agency's reasons. But no matter how often the kabuki is repeated, it can never give plaintiffs what they want—an order that materially limits the agency's discretion to reject the Draft Policy. This is because the agency has no legal duty to act, and there is thus no law we can bring to bear in narrowing its discretion. The agency can refuse to adopt the Draft Policy again and again, citing reasons of policy and convenience that we may not second-guess. See pp. 847-850 infra. Plaintiffs are pushing a long and very limp string, and no matter how hard they try, they can never get the object at the far end to move.

2. Let's assume, nevertheless, that plaintiffs could obtain an order forcing the agency to adopt the Draft Policy. Would the policy, if adopted, redress their injuries, as that term is understood in our standing jurisprudence? See Allen v. Wright, 468 U.S. 737, 751, 104 S. Ct. 3315, 82 L. Ed. 2d 556 (1984). Plaintiffs have the burden to show they have standing and cannot rely on "[s]peculative inferences ... to connect their injury to the challenged actions." Simon v. E. Ky. Welfare Rights Org., 426 U.S. 26, 45, 96 S. Ct. 1917, 48 L. Ed. 2d 450 (1976).

A favorable result that hinges on a third party's response is generally too speculative to support standing. In Linda R.S. v. Richard D., 410 U.S. 614, 618, 93 S. Ct. 1146, 35 L. Ed. 2d 536 (1973), for instance, the Supreme Court held that a mother didn't have standing to enforce a statute that imprisoned dead-beat dads; while such prosecution might have resulted in future child-support payments, the result was "speculative." Likewise, Allen v. Wright held that plaintiffs had no standing to challenge the tax exempt status of segregated private schools because there was no assurance the schools would desegregate out of fear of losing their exemptions. 468 U.S. at 758-59, 104 S. Ct. 3315. In applying these precedents we have held that an injury is not redressed if any benefit to plaintiffs is "conditioned on the discretionary behavior of any third party." Fernandez v. Brock, 840 F.2d 622, 627 (9th Cir.1988).

Of course, if the third party has no choice, its response is not discretionary and may provide a basis for standing. But this is only so if the regulation in question is sufficiently coercive. Maj. op. at 835-836. The Draft Policy here is not coercive; it is, as the majority recognizes, a "safe harbor." See id. at 830, 831, 835, 836, 840, 841. Entities that house primates need not follow the Draft Policy; they may lawfully choose alternate means of statutory compliance. If fear of incarceration in Linda R.S. and loss of tax benefits in Allen v. Wright were insufficiently coercive to give plaintiffs there standing, I don't see how a wholly permissive regime—one that expressly allows for alternate means of compliance—can give these plaintiffs standing.

The majority mistakenly relies on the representation of the intervenor, the National Association for Biomedical Research (NABR). NABR takes this view, however, based on its interpretation of the Draft Policy as mandatory—a position the majority correctly rejects. Maj. op. at 839. Once we have determined that the Draft Policy is merely a safe harbor, NABR's members will be free, just like everyone else, to abide by the policy or not, as they see fit. Certainly, nothing NABR has said in this litigation can bind its members to follow the policy, should it be implemented. Because there is no way of knowing whether any particular facility will comply, plaintiffs cannot show that they will benefit from the policy because the facilities they are concerned with may not.

What intervenor says doesn't matter anyhow, because it cannot speak—nor does it purport to speak—for all facilities that house primates. "NABR is an association of more than 300 universities, colleges, medical schools, veterinary schools, research institutes, research companies, and pharmaceutical manufacturers, as well as individual biomedical researchers, engaged in the care and use of animals in biomedical research." Yet the single example on which the majority bases standing involves Terry, a chimpanzee housed in a zoo. Zoos differ from the research-industry entities that make up NABR's membership. They're typically cash-strapped non-profit organizations with significant public support in the local community. They need not fear that sensitive and expensive research projects will be disrupted if they fall out of compliance. Their risk calculus might be quite different from that of organizations represented by NABR. The Southern Nevada Zoological-Botanical Park where Terry is housed, a facility already certified by USDA as compliant, may well choose not to comply with the Draft Policy.

But let's assume, for the sake of argument, that Terry's zoo did feel it had to follow the Draft Policy. I don't see how we can be sure that such compliance would ameliorate Buchanan's injury. The gravamen of her complaint is that Terry is lonely because he is housed by himself. The majority concludes that compliance with the Draft Policy would alleviate Terry's suffering, pointing to the following language: "[i]n order to address the social needs of nonhuman primates under § 3.81(a), the [zoo's environmental enhancement] plan must provide for each primate of a species known to be social in nature to be housed with other primates whenever possible." Maj. op. at 834 (quoting Draft Policy on Environmental Enhancement for Nonhuman Primates, 64 Fed.Reg. 38,145, 38,147 (proposed July 15, 1999)) (alterations in original). The majority zooms in on "must" but gives no effect to "whenever possible." It also overlooks the language that exempts facilities from this requirement if there is "[d]ocumented unavailability of compatible individuals." 64 Fed.Reg. at 38,147. In other words, the zoo must only house Terry with another chimpanzee if it has another chimpanzee to house him with. Buchanan could easily have alleged that Terry's zoo had a compatible primate, but she didn't. For all we know—for all the record reflects—Terry is the only chimp at the Southern Nevada Zoological-Botanical Park, or the zoo may have other chimps that can't be moved without disruption to themselves or their current companions. Even if the zoo believed it needed to comply with the Draft Policy, it could do so by documenting the unavailability of another compatible animal. To conclude that Buchanan has established standing is to make speculation the hallmark of redressability.4

1. The majority concludes that the agency's failure to adopt the Draft Policy is reviewable because the Draft Policy itself, if adopted, would have been reviewable. Maj. op. at 840-841. This doesn't follow; adoption and nonadoption of regulations are asymmetrical events. Regulations change the law, and thus can sharply affect the legal interests of private parties. Failure to adopt regulations leaves rights and responsibilities unchanged; it leaves the status quo in place. See Kennecott Utah Copper v. U.S Dep't of the Interior, 88 F.3d 1191, 1207 (D.C.Cir.1996) ("Because [the agency's] decision to withdraw the [proposed regulation] did not alter substantive legal obligations under previously published regulations, the agency's decision to withdraw the document did not constitute a `regulation' within the meaning of § 113 of CERCLA."). Because parties are not entitled to an agency's discretionary regulation, they can't claim to be aggrieved when the agency fails to adopt it.

When an agency adopts regulations, a court can judge whether those regulations are consistent with the agency's grant of authority (under the broad limits of Chevron USA Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 844-45, 104 S. Ct. 2778, 81 L. Ed. 2d 694 (1984)), and whether they constitute an abuse of discretion. There is, in the words of the Supreme Court, "law to apply." Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 410, 91 S. Ct. 814, 28 L. Ed. 2d 136 (1971). But an agency may choose not to adopt discretionary regulations for a variety of reasons, many of which a court can't review: a change in policy; a lack of enforcement resources; a lack of scientific expertise to address the problem at this time; a change in direction based on a determination that the problem is better addressed some other way. All of these matters are, in the words of the Administrative Procedure Act, "committed to agency discretion by law." 5 U.S.C. § 701(a)(2). There is nothing for a court to review.

There was a time when the Supreme Court (like the majority here) disdained the distinction between agency action and inaction. See, e.g., Rochester Tel. Corp. v. United States, 307 U.S. 125, 143, 59 S. Ct. 754, 83 L. Ed. 1147 (1939). But the Court turned its back on this caselaw in Heckler v. Chaney, 470 U.S. 821, 105 S. Ct. 1649, 84 L. Ed. 2d 714 (1985), which held that an agency's decision not to act is, under ordinary circumstances, unreviewable.

Chaney involved death row inmates who claimed that the FDA had shirked its duty by failing to prevent states from using drugs for lethal injections. According to the inmates, these drugs had not been proven "safe and effective" for use in human executions, as required by the applicable statute. Id. at 824, 105 S. Ct. 1649. In rejecting this contention, the Court closely examined the text of the APA, and concluded that an agency's failure to act is insulated from judicial review in two circumstances. First, under 5 U.S.C. § 701(a)(1), an agency decision may not be judicially reviewed "when Congress has expressed an intent to preclude judicial review." 470 U.S. at 830, 105 S. Ct. 1649. Neither here nor in Chaney is this provision applicable. But the Court went on to hold that judicial review is also precluded under 5 U.S.C. § 701(a)(2) where "no judicially manageable standards are available for judging how and when an agency should exercise its discretion." 470 U.S. at 830, 105 S. Ct. 1649. The Court held that an agency's discretionary decision not to enforce a statute—even though it has the authority to do so—is unreviewable.

In reaching this conclusion, the Court noted that the agency's decision not to exercise its enforcement powers is "generally committed to [its] absolute discretion." Id. at 831, 105 S. Ct. 1649. This, the Court said, "is attributable in no small part to the general unsuitability for judicial review of agency decisions to refuse enforcement." Id. "The reasons for this general unsuitability," the Court continued, "are many." Id. Among them, the Court listed the following:

[T]he agency must not only assess whether a violation has occurred, but whether agency resources are best spent on this violation or another, whether the agency is likely to succeed if it acts, whether the particular enforcement action requested best fits the agency's overall policies, and, indeed, whether the agency has enough resources to undertake the action at all....

In addition to these administrative concerns, we note that when an agency refuses to act it generally does not exercise its coercive power over an individual's liberty or property rights, and thus does not infringe upon areas that courts often are called upon to protect. Similarly, when an agency does act to enforce, that action itself provides a focus for judicial review, inasmuch as the agency must have exercised its power in some manner. The action at least can be reviewed to determine whether the agency exceeded its statutory power.

Id. at 831-32, 105 S. Ct. 1649.

While Chaney spoke directly to an agency's enforcement actions, its teachings sweep more broadly. Regulations implicate precisely the same concerns addressed in Chaney; one can properly view the adoption—or non-adoption—of regulations as an enforcement decision.5 In adopting regulations, the agency implements the mandate of its governing statute by coercing the conduct of private parties. Regulations channel the agency's enforcement efforts and commit agency resources to a particular interpretation of the statute. A decision to promulgate regulations, just like a decision to enforce, amounts to an exercise of agency authority, which "at least can be reviewed to determine whether the agency exceeded its statutory powers." Id. at 832, 105 S. Ct. 1649. Failure to adopt discretionary regulations, by contrast, may reflect considerations besides the applicable statute, such as a change in policy, a lack of resources, a decision to pursue a different course of action or a decision to gather additional information before acting. Such failure to adopt cannot normally be tested against the applicable statute, because it may be based on non-statutory considerations of policy and convenience that are within the agency's inherent authority.

The failure to adopt regulations, like failure to enforce, may be reviewed in certain narrow circumstances. The first, and most common, is where the agency has a legal duty to act. Id. at 832-35, 105 S. Ct. 1649. The typical such case is where Congress has directed the agency to take a particular regulatory action—such as listing species as endangered. See Defenders of Wildlife v. Norton, 258 F.3d 1136, 1137-38 (9th Cir.2001); see also Greater L.A. Council on Deafness, Inc. v. Baldrige, 827 F.2d 1353, 1361 (9th Cir.1987). The second is where the agency refuses to act based solely on a belief that it lacks jurisdiction. Chaney, 470 U.S. at 833 n. 4, 105 S. Ct. 1649; see, e.g., Transpacific Westbound Rate Agreement v. Fed. Mar. Comm'n, 938 F.2d 1025, 1026-27 (9th Cir.1991) ("The Commission's order declined jurisdiction over parts of shipping agreements that regulate wholly foreign transportation.... Instead, the Commission held that it had jurisdiction only over the United States-foreign portion of the agreements."). The third is where the agency's refusal to act violates constitutional rights. Chaney, 470 U.S. at 838, 105 S. Ct. 1649. Finally, a court might have jurisdiction where "the agency has `consciously and expressly adopted a general policy' that is so extreme as to amount to an abdication of its statutory responsibilities." Id. at 833 n. 4, 105 S. Ct. 1649 (quoting Adams v. Richardson, 480 F.2d 1159 (D.C.Cir.1973) (en banc)).

None of these exceptions apply. There is certainly no claim that the agency believed it lacked jurisdiction to adopt the Draft Policy. Nor is there any suggestion that the agency abdicated its statutory responsibilities or violated constitutional rights. And, most significantly, we know for a fact that the agency had no statutory duty to act because the DC Circuit found the regulations the agency had adopted sufficient. Glickman II, 204 F.3d at 235. The agency need do no more. Adoption of the Draft Policy was an entirely discretionary decision, one the agency could choose to eschew based on a variety of non-legal considerations.

The majority remands for a determination "whether the USDA's withdrawal of the Draft Policy was arbitrary and capricious in violation of § 706(2) of the APA," maj. op. at 844, but does not say how that determination can be made. The implication seems to be that review should be on the record of the rulemaking proceedings. Id. at 40. But it's not clear why this should be so. Absent a statutory duty to act, an agency need not adopt regulations, even if all public comments submitted favor them (which is certainly not the case here). The agency may decide not to adopt regulations because of a change in administrations, or some other change in policy.6 Or, the agency may decide to withdraw a proposed regulation in response to a change in law.7 The agency might also choose a different path.8 An agency's decision not to proceed with a regulation—just like a decision not to enforce—can properly be based on a variety of judicially unreviewable considerations.

The majority's implication that withdrawal of the policy must be justified by the public comments is alarming. It locks an agency into a course of action—a course it cannot abandon, unless it be for reasons a court finds sufficient on the administrative record. This not only improperly constrains agency discretion under Chaney, it discourages agencies from proposing discretionary regulations, lest they be stuck with them if they cannot convince a federal court that the record supports abandonment. If this was ever the law, it's certainly not the law today.

The majority relies on several cases from the DC Circuit, see maj. op. at 841-844, but those cases have been overtaken by Chaney.9 Indeed, the case on which my colleagues principally rely, Natural Resources Defense Council, Inc. v. SEC, 606 F.2d 1031 (D.C.Cir.1979) ("NRDC"), is listed in Justice Marshall's concurrence as repudiated by the majority.10 Chaney, 470 U.S. at 850 n. 7, 105 S. Ct. 1649 (Marshall, J., concurring); see also Judicial Review of Agency Inaction, supra, at 264 n. 1 (listing NRDC among the cases inconsistent with Chaney). Justice Marshall strongly disagreed that agency refusals to act are presumptively non-reviewable, but wrote only for himself; eight Justices joined in the opinion.

2. Even on its own terms, the majority is wrong. Assuming standing, assuming reviewability, assuming the cases the majority relies on are still good law, judicial review is still not justified. The majority extracts from the DC Circuit cases two requirements for reviewing non-adoption of a regulation and finds both satisfied. Maj. op. 842-844. In fact, there are two additional requirements. First, the agency's regulatory process must have come to an end. See Ctr. for Auto Safety v. Nat'l Highway Traffic Safety Admin., 710 F.2d 842, 847-48 (D.C.Cir.1983) (action not ripe if the agency "has left open the possibility that the agency will initiate new rulemaking proceedings at some point in the future"). Second, the agency must have "explained in detail its reason for not adopting those rules." NRDC, 606 F.2d at 1047. Neither requirement is satisfied here.

a. It is perfectly clear that the agency's decision-making process has not come to rest. Dr. Gipson's announcement that the rulemaking had been terminated adds as follows: "USDA and NIH had agreed to develop best management practice guidelines concerning nonhuman primates." The agency has not abandoned its effort to give further guidance as to the management of primates. Rather, it has shifted focus toward other means of implementing that objective.

In Center for Auto Safety, on which the majority relies, the agency terminated rulemaking, but "left open the possibility that the agency [would] initiate new rulemaking proceedings at some point in the future." 710 F.2d at 847 (emphasis added). Because the agency still "ha[d] a chance ... to amend" its existing regulations, the DC Circuit concluded that judicial review of the refusal to do so at that time would have "entangl[ed] [the courts] in abstract disagreements over administrative policies." Id. at 848 (quoting Abbott Labs. v. Gardner, 387 U.S. 136, 148, 87 S. Ct. 1507, 18 L. Ed. 2d 681 (1967)). Here we have not merely a possibility that the agency will resume the regulatory process at some unspecified future time; we know it's doing so now. If review was premature in Center for Auto Safety, it is previable here.11

The majority's ruling also conflicts with Ecology Center, Inc. v. United States Forest Service, 192 F.3d 922, 926 (9th Cir. 1999), which held that the monitoring of forest activity was not reviewable. The Forest Service there published reports that were not interlocutory or tentative, but the monitoring itself was one of the "steps leading to an agency decision, rather than the final action itself." Id. at 925. Thus, it did not "mark the consummation of the agency's decision making process," id., and review was premature. Here, Dr. Gipson's declaration is not the end point of the agency's decision making, but only a step along the way. It is no more reviewable than the reports in Ecology Center.

Review of interlocutory agency statements hopelessly entangles the judiciary in the administrative process. The agency here cannot have two policies on the same issue: It can either adopt something like the Draft Policy, or it can adopt best management practice guidelines like the ones it is now considering. By allowing this lawsuit to proceed, the majority leaves open the possibility that the agency will be forced or coaxed into adopting the Draft Policy.12 Even if the agency prevails before the district court, further agency proceedings will be delayed at least until the next appeal—which may be years away. Until this lawsuit is over, the agency cannot implement best practice guidelines, which may take the agency in a different direction from the Draft Policy. The majority thus interferes with the agency's ongoing consideration of a live issue.

b. The majority ignores an additional requirement in the DC Circuit case law—that the agency "explain[] in detail its reasons for not adopting those rules." NRDC, 606 F.2d at 1047.

In the cases the majority cites, see maj. op. at 841-844, the agencies all provided reasons a court could evaluate. NRDC, the cornerstone of the majority's argument, involved "lengthy explanatory statements," id. at 1039, that the DC Circuit repeatedly referenced in affirming the agency's decision. Id. at 1058-59, 1061-62. In Professional Drivers Council v. Bureau of Motor Carrier Safety, 706 F.2d 1216 (D.C.Cir.1983), the agency published its reasons for terminating its rulemaking, id. at 1220, and the DC Circuit upheld the agency by reviewing the "relevant, permissible factors" upon which the agency relied. Id. at 1222. In Center for Auto Safety, the agency also published its reasons for withdrawing the policy, 710 F.2d at 844, but the court never reached the issue. See p. 851 supra. And in International Union, the agency provided three reasons, which the court then reviewed. 358 F.3d at 44.

By contrast, the agency here hasn't provided a detailed explanation as to why it abandoned the Draft Policy, or any reason at all. We have only Dr. Gipson's announcement that "the USDA was not going forward with the Draft Policy on Environmental Enhancement for Nonhuman Primates." Without an explanation for the agency's refusal to adopt the Draft Policy, it is impossible to determine whether its decision is arbitrary or capricious. Even if we assume that the agency did so for reasons contained in the record, we don't know what those reasons are. Judicial review would thus "have an undesirably abstract and hypothetical quality." NRDC, 606 F.2d at 1047.

What are the district judge and the parties to do on remand? The judge can certainly review the record for comments supporting and opposing the Draft Policy, but how is he to determine which of them are legitimate and which aren't? What weight is he to give the various arguments? Considering and weighing public comments, then deciding whether to adopt, reject or modify a proposed rule is a quintessentially administrative function. It calls for the kind of expertise a district judge does not have and the kind of policy judgment a district judge is not competent to make. It is simply not a judicial function. Reviewing an agency decision where there is no agency decision to review is much like making ham and eggs without ham or eggs. It is precisely for this reason that the DC Circuit cases on which the majority relies require an explicit and detailed statement of reasons from the agency as a condition for judicial review. By ignoring this crucial requirement, the majority sends the parties and the district court on a fool's errand.

* * *

The majority expands the law of standing beyond recognition. It unmoors administrative law from sound principles of judicial review, and insinuates the federal courts into sensitive policy judgments that are the exclusive province of the Executive Branch. It ignores the teachings of the Supreme Court and misapplies the precedents it relies on. It will cause no end of mischief. Count me out.