dissenting.
Today’s opinion makes an avulsive change in judicial review of federal administrative action. Whereas previously a reasonable agency application of an ambiguous statutory provision had to be sustained so long as it represented the agency’s authoritative interpretation, henceforth such an application can be set aside unless “it appears that Congress delegated authority to the agency generally to make rules carrying the force of law,” as by giving an agency “power to engage in adjudication or notiee-and-comment rulemaking, or . . . some other [procedure] indicating] comparable congressional intent,” and “the agency interpretation claiming deference was promulgated in the exercise of that authority.” Ante, at 226-227.1 What was previously a general presumption of authority in agencies to resolve ambiguity in the statutes they have been authorized to enforce has been changed to a presumption of no such authority, which must be overcome by affirmative legislative intent to the contrary. And whereas previously, when agency authority to resolve ambiguity did not exist the court was free to give the statute what it considered the best interpretation, henceforth the court must supposedly give the agency view some indeterminate amount of so-called Skidmore deference. Skidmore v. Swift & Co., 323 U. S. 134 (1944). We will be sorting out the consequences of the Mead doctrine, which has today replaced the Chevron doctrine, Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837 (1984), for years to come. I would adhere to our established jurisprudence, *240defer to the reasonable interpretation the Customs Service has given to the statute it is charged with enforcing, and reverse the judgment of the Court of Appeals.
I
Only five years ago, the Court described the Chevron doctrine as follows: “We accord deference to agencies under Chevron . .. because of a presumption that Congress, when it left ambiguity in a statute meant for implementation by an agency, understood that the ambiguity would be resolved, first and foremost, by the agency, and desired the agency (rather than the courts) to possess whatever degree of discretion the ambiguity allows,” Smiley v. Citibank (South Dakota), N. A., 517 U. S. 735, 740-741 (1996) (citing Chevron, supra, at 843-844). Today the Court collapses this doctrine, announcing instead a presumption that agency discretion does not exist unless the statute, expressly or impliedly, says so. While the Court disclaims any hard- and-fast rule for determining the existence of discretion-conferring intent, it asserts that “a very good indicator [is] express congressional authorizations to engage in the process of rulemaking or adjudication that produces regulations or rulings for which deference is claimed,” ante, at 229. Only when agencies act through “adjudieation[,] notice-and-comment rulemaking, or . . . some other [procedure] indicating] comparable congressional intent [whatever that means]” is Chevron deference applicable — because these “relatively formal administrative proeedure[s] [designed] to foster . . . fairness and deliberation” bespeak (according to the Court) congressional willingness to have the agency, rather than the courts, resolve statutory ambiguities. Ante, at 227, 230. Once it is determined that Chevron deference is not in order, the uncertainty is not at an end — and indeed is just beginning. Litigants cannot then assume that the statutory question is one for the courts to determine, accord*241ing to traditional interpretive principles and by their own judicial lights. No, the Court now resurrects, in full force, the pre-Chevron doctrine of Skidmore deference, see Skid-more, supra, whereby “[t]he fair measure of deference to an agency administering its own statute . . . varfies] with circumstances,” including “the degree of the agency’s care, its consistency, formality, and relative expertness, and . . . the persuasiveness of the agency’s position,” ante, at 228 (footnotes omitted). The Court has largely replaced Chevron, in other words, with that test most beloved by a court unwilling to be held to rules (and most feared by litigants who want to know what to expect): th’ol’ “totality of the circumstances” test.
The Court’s new doctrine is neither sound in principle nor sustainable in practice.
A
As to principle: The doctrine of Chevron — that all authoritative agency interpretations of statutes they are charged with administering deserve deference — was rooted in a legal presumption of congressional intent, important to the division of powers between the Second and Third Branches. When, Chevron said, Congress leaves an ambiguity in a statute that is to be administered by an executive agency, it is presumed that Congress meant to give the agency discretion, within the limits of reasonable interpretation, as to how the ambiguity is to be resolved. By committing enforcement of the statute to an agency rather than the courts, Congress committed its initial and primary interpretation to that branch as well.
There is some question whether Chevron was faithful to the text of the Administrative Procedure Act (APA), which it did not even bother to cite.2 But it was in accord with the *242origins of federal-court judicial review. Judicial control of federal executive officers was principally exercised through the prerogative writ of mandamus. See L. Jaffe, Judicial Control of Administrative Action 166, 176-177 (1965). That writ, generally would not issue unless the executive officer was acting plainly beyond the scope of his authority.
“The questions mooted before the Secretary and decided by him were whether the fund is a tribal fund, whether the tribe is still existing and whether the distribution of the annuities is to be confined to members of the tribe .... These are all questions of law the solution of which requires a construction of the act of 1889 and other related acts. A reading of these acts shows that they fall short of plainly requiring that any of the questions be answered in the negative and that in some aspects they give color to the affirmative answers of the Secretary. That the construction of the acts insofar as they have a bearing on the first and third questions is sufficiently uncertain to involve the exercise of judgment and discretion is rather plain....
“From what has been said it follows that the case is not one in which mandamus will lie.” t Wilbur v. United States ex rel. Kadrie, 281 U. S. 206, 221-222 (1930).
*243Statutory ambiguities, in other words, were left to reasonable resolution by the Executive.
The basis in principle for today’s new doctrine can be described as follows: The background rule is that ambiguity in legislative instructions to agencies is to be resolved not by the agencies but by the judges. Specific congressional intent to depart from this rule must be found — and while there is no single touchstone for such intent it can generally be found when Congress has authorized the agency to act through (what the Court says is) relatively formal procedures such as informal rulemaking and formal (and informal?) adjudication, and when the agency in fact employs such procedures. The Court’s background rule is contradicted by the origins of judicial review of administrative action. But in addition, the Court’s principal criterion of congressional intent to supplant its background rule seems to me quite implausible. There is no necessary connection between the formality of procedure and the power of the entity administering the procedure to resolve authoritatively questions of law. The most formal of the procedures the Court refers to — -formal adjudication — is modeled after the process used in trial courts, which of course are not generally accorded deference on questions of law. The purpose of such a procedure is to produce a closed record for determination and review of the facts — which implies nothing about the power of the agency subjected to the procedure to resolve authoritatively questions of law.
As for informal rulemaking: While formal adjudication procedures are prescribed (either by statute or by the Constitution), see 5 U. S. C..§§554, 556; Wong Yang Sung v. McGrath, 339 U. S. 33, 50 (1950), informal rulemaking is more typically authorized but not required. Agencies with such authority are free to give guidance through rulemaking, but they may proceed to administer their statute case-by-case, “making law” as they implement their program (not necessarily through formal adjudication). See NLRB v. Bell *244Aerospace Co., 416 U. S. 267, 290-295 (1974); SEC v. Chenery Corp., 332 U. S. 194, 202-203 (1947). Is it likely — or indeed even plausible — that Congress meant, when such an agency chooses rulemaking, to accord the administrators of that agency, and their successors, the flexibility of interpreting the ambiguous statute now one way, and later another; but, when such an agency chooses case-by-case administration, to eliminate all future agency discretion by having that same ambiguity resolved authoritatively (and forever) by the courts?3 Surely that makes no sense. It is also the case that certain significant categories of rules — those involving grant and benefit programs, for example, are exempt from the requirements of informal rulemaking. See 5 U. S. C. § 553(a)(2). Under the Court’s novel theory, when an agency takes advantage of that exemption its rules will be deprived of Chevron deference, i. e., authoritative effect. Was this either the plausible intent of the APA rulemaking exemption, or the plausible intent of the Congress that established the grant or benefit program?
Some decisions that are neither informal rulemaking nor formal adjudication are required to be made personally by a Cabinet Secretary, without any prescribed procedures. See, e. g., United States v. Giordano, 416 U. S. 505, 508 (1974) (involving application of 18 U. S. C. §2516 (1970 ed.), requiring wiretap applications to be authorized by “[t]he Attorney General, or any Assistant Attorney General specially designated by the Attorney General”); D. C. Federation of Civic Assns. v. Volpe, 459 F. 2d 1231, 1248-1249 (CADC 1971) (involving application of 23 U. S. C. § 138 (1970 ed.) requiring the Secretary of Transportation to determine that there is “no feasible and prudent alternative to the use of” publicly owned parkland for a federally funded highway), cert, denied, 405 U. S. 1030 (1972). Is it conceivable that decisions *245specifically committed to these high-level officers are meant to be accorded no deference, while decisions by an administrative law judge left in place without further discretionary agency review, see 5 U. S. C. § 557(b), are authoritative? This seems to me quite absurd, and not at all in accord with any plausible actual intent of Congress.
B
As for the practical effects of the new rule:
1
The principal effect will be protracted confusion. As noted above, the one test for Chevron deference that the Court enunciates is wonderfully imprecise: whether “Congress delegated authority to the agency generally to make rules carrying the force of law,.. . as by .. . adjudication[,] notice-and-comment rulemaking, or . . . some other [procedure] indicating] comparable congressional intent.” But even this description does not do justice to the utter flabbiness of the Court’s criterion, since, in order to maintain the fiction that the new test is really just the old one, applied consistently throughout our case law, the Court must make a virtually open-ended exception to its already imprecise guidance: In the present case, it tells us, the absence of notice-and-comment rulemaking (and “[who knows?] [of] some other [procedure] indieati[ng] comparable congressional intent”) is not enough to decide the question of Chevron deference, “for we have sometimes found reasons for Chevron deference even when no such administrative formality was required and none was afforded.” Ante, at 226-227, 231. The opinion then goes on to consider a grab bag of other factors — including the factor that used to be the sole criterion for Chevron deference: whether the interpretation represented the authoritative position of the agency, see ante, *246at 231-234. It is hard to know what the lower courts are to make of today’s guidance.
2
Another practical effect of today’s opinion will be an artificially induced increase in informal rulemaking. Buy stock in the GPO. Since informal rulemaking and formal adjudication are the only more-or-less safe harbors from the storm that the Court has unleashed; and since formal adjudication is not an option but must be mandated by statute or constitutional command; informal rulemaking — which the Court was once careful to make voluntary unless required by statute, see Bell Aerospace, supra, and Chenery, supra— will now become a virtual necessity. As I have described, the Court’s safe harbor requires not merely that the agency have been given rulemaking authority, but also that the agency have employed rulemaking as the means of resolving the statutory ambiguity. (It is hard to understand why that should be so. Surely the mere conferral of rulemaking authority demonstrates — if one accepts the Court’s logic — a congressional intent to allow the agency to resolve ambiguities. And given that intent, what difference does it make that the agency chooses instead to use another perfectly permissible means for that purpose?) Moreover, the majority’s approach will have a perverse effect on the rules that do emerge, given the principle (which the Court leaves untouched today) that judges must defer to reasonable agency interpretations of their own regulations. See, e. g., United States, v. Cleveland Indians Baseball Co., 532 U. S. 200, 220 (2001) (“We need not decide whether the [informal] Revenue Rulings themselves are entitled to deference[,. . . because] the Rulings simply reflect the agency’s longstanding interpretation of its own regulations”). Agencies will now have high incentive to rush out barebones, ambiguous, rules construing statutory ambiguities, which they can then in turn further clarify through informal rulings entitled to judicial respect.
*2473
Worst of all, the majority’s approach will lead to the ossification of large portions of our statutory law. Where Chevron applies, statutory ambiguities remain ambiguities subject to the agency’s ongoing clarification. They create a space, so to speak, for the exercise of continuing agency discretion. As Chevron itself held, the Environmental Protection Agency can interpret “stationary source” to mean a single smokestack, can later replace that interpretation with the “bubble concept” embracing an entire plant, and if that proves undesirable can return again to the original interpretation. 467 U. S., at 853-859, 865-866. For the indeterminately large number of statutes taken out of Chevron by today’s decision, however, ambiguity (and hence flexibility) will cease with the first judicial resolution. Skidmore deference gives the agency’s current position some vague and uncertain amount of respect, but it does not, like Chevron, leave the matter within the control of the Executive Branch for the future. Once the court has spoken, it becomes unlawful for the agency to take a contradictory position; the statute now says what the court has prescribed. See Neal v. United States, 516 U. S. 284, 295 (1996); Lechmere, Inc. v. NLRB, 502 U. S. 527, 536-537 (1992); Maislin Industries, U. S., Inc. v. Primary Steel, Inc., 497 U. S. 116, 131 (1990). It will be bad enough when this ossification occurs as a result of judicial determination (under today’s new principles) that there is no affirmative indication of congressional intent to “delegate”; but it will be positively bizarre when it occurs simply because of an agency’s failure to act by rulemaking (rather than informal adjudication) before the issue is presented to the.courts.
One might respond that such ossification would not result if the agency were simply to readopt its interpretation, after a court reviewing it under Skidmore had rejected it, by re-promulgating it through one of the Chevron-eligible procedural formats approved by the Court today. Approving this *248procedure would be a landmark abdication of judicial power. It is worlds apart from Chevron proper, where the court does not purport to give the statute a judicial interpretation— except in identifying the scope of the statutory ambiguity, as to which the court’s judgment is final and irreversible. (Under Chevron proper, when the agency’s authoritative interpretation comes within the scope of that ambiguity — and the court therefore approves it — the agency will not be “overruling” the court’s decision when it later decides that a different interpretation (still within the scope of the ambiguity) is preferable.) By contrast, under this view, the reviewing court will not be holding the agency’s authoritative interpretation within the scope of the ambiguity; but will be holding that the agency has not used the “delegation-conferring” procedures, and that the court must therefore interpret the statute on its own — but subject to reversal if and when the agency uses the proper procedures.
One is reminded of Justice Jackson’s words in Chicago & Southern Air Lines, Inc. v. Waterman S. S. Corp., 333 U. S. 103, 113 (1948):
“The court below considered that after it reviewed the Board’s order its judgment would be submitted to the President, that his power to disapprove would apply after as well as before the court acts, and hence that there would be no chance of a deadlock and no conflict of function. But if the President may completely disregard the judgment of the court, it would be only because it is one the courts were not authorized to render. Judgments within the powers vested in courts by the Judiciary Article of the Constitution may not lawfully be revised, overturned or refused faith and credit by another Department of Government.”
I know of no case, in the entire history of the federal courts, in which we have allowed a judicial interpretation of a statute to be set aside by an agency — or have allowed a *249lower court to render an interpretation of a statute subject to correction by an agency. As recently as 1996, we rejected an attempt to do 'precisely that. In Chapman v. United States, 500 U. S. 453 (1991), we had held that the weight of the blotter paper bearing the lysergic acid diethyl-amide (LSD) must be counted for purposes of determining whether the quantity crossed the 10-gram threshold of 21 U. S. C. § 841(b)(1)(A)(v) imposing a minimum sentence of 10 years. At that time the United States Sentencing Commission applied a similar approach under the Sentencing Guidelines, but had taken no position regarding the meaning of the statutory provision. The Commission later changed its Guidelines approach, and, according to the petitioner in Neal v. United States, 516 U. S. 284 (1996), made clear its view that the statute bore that meaning as well. The petitioner argued that we should defer to that new approach. We would have none of it.
“Were we, for argument’s sake, to adopt petitioner’s view that the Commission intended the commentary as an interpretation of § 841(b)(1), and that the last sentence of the commentary states the Commission’s view that the dose-based method is consistent with the term 'mixture or substance’ in the statute, he still would not prevail. The Commission’s dose-based method cannot be squared with Chapman. ... In these circumstances, we need not decide what, if any, deference is owed the Commission in order to reject its alleged contrary interpretation. Once we have determined a statute’s meaning, we adhere to our ruling under the doctrine of stare decisis, and we assess an agency’s later interpretation of the statute against that settled law.” Id., at 294-295 (citations omitted).
There is, in short, no way to avoid the ossification of federal law that today’s opinion sets in motion. What a court says is the law after according Skidmore deference will be the *250law forever, beyond the power of the agency to change even through rulemaking.
4
And finally, the majority’s approach compounds the confusion it creates by breathing new life into the anachronism of Skidmore, which sets forth a sliding scale of deference owed an agency’s interpretation of a statute that is dependent “upon the thoroughness evident in [the agency’s] consideration, the validity of its reasoning, its consistency with earlier and later pronouncements, and all those factors which give it power to persuade, if lacking power to control”; in this way, the appropriate measure of deference will be accorded the “body of experience and informed judgment” that such interpretations often embody, 323 U. S., at 140. Justice Jackson’s eloquence notwithstanding, the rule of. Skidmore deference is an empty truism and a trifling statement of the obvious: A judge should take into account the well-considered views of expert observers.
It was possible to live with the indeterminacy of Skid-more deference in earlier times. But in an era when federal statutory law administered by federal agencies is pervasive, and when the ambiguities (intended or unintended) that those statutes contain are innumerable, totality-of-the-circumstances Skidmore deference is a recipe for uncertainty, unpredictability, and endless litigation. To condemn a vast body of agency action to that regime (all except rule-making, formal (and informal?) adjudication, and whatever else might now and then be included within today’s intentionally vague formulation of affirmative congressional intent to “delegate”) is irresponsible.
II
The Court s pretense that today’s opinion is nothing more than application of our prior case law does not withstand analysis. It is, to be sure, impossible to demonstrate that any of our cases contradicts the rule of decision that the *251Court prescribes, because the Court prescribes none. More precisely, it at one and the same time (1) renders meaningless its newly announced requirement that there be an affirmative congressional intent to have ambiguities resolved by the administering agency, and (2) ensures that no prior decision can possibly be cited which contradicts that requirement, by simply announcing that all prior decisions according Chevron deference exemplify the multifarious ways in which that congressional intent can be manifested: “[A]s significant as notice-and-eomment is in pointing to Chevron authority, the want of that procedure here does not decide the case, for we have sometimes found reasons for Chevron deference even when no such administrative formality was required and none was afforded,” ante, at 230-231.4
*252The principles central to today’s opinion have no antecedent in our jurisprudence. Chevron, the case that the opinion purportedly explicates, made no mention of the “relatively formal administrative procedure[s]," ante, at 230, that the Court today finds the best indication of an affirmative intent by Congress to have ambiguities resolved by the administering agency. Which is not so remarkable, since Chevron made no mention of any need to find such an affirmative intent; it said that in the event of statutory ambiguity agency authority to clarify was to be presumed. And our cases have followed that prescription.
Six years ago, we unanimously accorded Chevron deference to an interpretation of the National Bank Act, 12 U. S. C. §24 Seventh (1988 ed. and Supp. V), contained in a letter to a private party from a Senior Deputy Comptroller of the Currency. See NationsBank of N. C, N. A. v. Variable Annuity Life Ins. Co., 513 U. S. 251, 255, 257 (1995). We did so because the letter represented (and no one contested) that it set forth the official position of the Comptroller of the Currency, see id., at 263.
Several cases decided virtually in the wake of Chevron, which the Court conveniently ignores, demonstrate that Congress could not (if it was reading our opinions) have acted in reliance on a background assumption that Chevron deference would generally be accorded only to agency interpretations arrived at through formal adjudication, notice-and-comment rulemaking, or other procedures assuring “fairness and deliberation,” ante, at 230. In FDIC v. Philadelphia Gear Corp., 476 U. S. 426, 438-439 (1986), we accorded Chevron deference to the Federal Deposit Insurance Corporation’s interpretation of the statutory term “deposit” reflected in a course of unstructured administrative actions, and gave particular weight to the agency’s “contemporaneous understanding” reflected in the response given by an FDIC official to a question asked at a meeting of FDIC and bank officials. It was clear that the position reflected *253the official position of the agency, and that was enough to command Chevron deference. In Young v. Community Nutrition Institute, 476 U. S. 974 (1986), the statutory ambiguity at issue pertained to a provision that “the Secretary [of Health and Human Services] shall promulgate regulations limiting the quantity [of any poisonous or deleterious substance added to any food] to such extent as he finds necessary for the protection of public health.” The Secretary had regularly interpreted the phrase “to such extent as he finds necessary” as conferring discretion not to issue a rule, rather than merely discretion regarding the quantity that the rule would permit. This interpretation was not, of course, reflected in any formal adjudication, and had not been the subject of any informal rulemaking — it was the Secretary’s understanding consistently applied in the course of the Department’s practice. We accorded it Chevron deference, as unquestionably we should have. And in Mead Cory. v. Tilley, 490 U. S. 714 (1989), a private suit by retirees against their former employer under the Employee Retirement Income Security Act of 1974 (ERISA), we accorded Chevron deference to the Pension Benefit Guaranty Corporation’s interpretation of § 4044(a) of ERISA, 29 U. S. C. § 1344(a) (1982 ed. and Supp. V), that was reflected only in an amicus brief to this Court and in several opinion letters issued without benefit of any prescribed procedures. See 490 U. S., at 722.
I could continue to enumerate cases according Chevron deference to agency interpretations not arrived at through formal proceedings — for example, Pension Benefit Guaranty Corporation v. LTV Corp., 496 U. S. 633, 642-643, 647-648 (1990) (according Chevron deference to the PBGC’s interpretation of the requirements for its restoring a terminated plan under §4047 of ERISA, 29 U. S. C. §1347 (1988 ed.), which interpretation was reflected in nothing more than the agency’s act of issuing a notice of restoration). Suffice it to say that many cases flatly contradict the theory of Chevron set forth in today’s opinion, and with one exception *254not a single case can be found with language that supports the theory. That exception, a very recent one, deserves extended discussion.
In Christensen v. Harris County, 529 U. S. 576 (2000), the Court said the following:
“[W]e confront an interpretation contained in an opinion letter, not one arrived at after, for example, a formal adjudication or notice-and-comment rulemaking. Interpretations such as those in opinion letters — like interpretations contained in policy statements, agency manuals, and enforcement guidelines, all of which lack the force of law — do not warrant Chevron-style deference.” Id., at 587.
This statement was dictum, unnecessary to the Court’s holding. Since the Court went on to find that the Secretary of Labor’s position “ma[de] little sense” given the text and structure of the statute, id., at 585-586, Chevron deference could not have been accorded no matter what the conditions for its application. See 529 U. S., at 591 (Scalia, J., concurring in part and concurring in judgment). It was, moreover, dictum unsupported by the precedent that the Court cited.
The Christensen majority followed its above-quoted dictum with a string citation of three cases, none of which sustains its point. In Reno v. Koray, 515 U. S. 50 (1995), we had no occasion to consider what level of deference was owed the Bureau of Prisons’ interpretation of 18 U. S. C. § 3585(b) set forth in an internal agency guideline, because our opinion made clear that we would have independently arrived at the same interpretation on our own, see 515 U. S., at 57-60. And although part of one sentence in Koray might be read to suggest that the Bureau’s “Program Statement]” should be accorded a measure of deference less than that mandated by Chevron, this aside is ultimately inconclusive, *255since the sentence ends by observing that the statement was “a ‘permissible construction of the statute’ ” under Chevron. 515 U. S., at 61 (quoting Chevron, 467 U. S., at 843). In the second case cited, EEOC v. Arabian American Oil Co., 499 U. S. 244 (1991), it was again unnecessary to our holding whether the agency’s interpretation of the statute warranted Chevron deference, since the “longstanding . . . ‘canon of [statutory] construction’ ” disfavoring extraterritoriality, 499 U. S., at 248, would have required the same result even if Chevron applied. See 499 U. S., at 260 (Scalia, J., concurring in part and concurring in judgment). While the opinion did purport to accord the Equal Employment Opportunity Commission’s informally promulgated interpretation only Skidmore deference, it did so because the Court thought itself bound by its pre-Chevron, EEOC-specific decision in General Elec. Co. v. Gilbert, 429 U. S. 125 (1976), which noted that “‘Congress, in enacting Title VII, did not’” intend to give the EEOC substantive authority to resolve statutory ambiguities, Arabian American Oil, supra, at 257 (quoting Gilbert, supra, at 141). Lastly, in Martin v. Occupational Safety and Health Review Comm’n, 499 U. S. 144 (1991), the question of the level of deference owed the Secretary of Labor’s interpretation of the Occupational Safety and Health Act of 1970, 84 Stat. 1590, as amended, 29 U. S. C. § 651 et seq., was neither presented by the case nor considered in our opinion. The only question before the Court was which of two competing interpretations of 29 CFR §1910.1029 (1990) — the Secretary’s or the Occupational Safety and Health Review Commission’s — should have been deferred to by the court below. See 499 U. S., at 150. The dicta the Christensen Court cited, 529 U. S., at 587 (citing 499 U. S., at 157), opined on the measure of deference owed the Secretary’s interpretation, not of the statute, but of his own regulations, see generally Manning, Constitutional Structure *256and Judicial Deference to Agency Interpretations of Agency Rules, 96 Colum. L. Rev. 612 (1996).
To make matters worse, the arguments marshaled by Christensen in support of its dictum — its observation that “interpretations contained in policy statements, agency manuals, and enforcement guidelines, all . . . lack the force of law,” and its citation of 1 K. Davis & R. Pierce, Administrative Law Treatise § 3.5 (3d ed. 1994), 529 U. S., at 587 — are not only unpersuasive but bear scant resemblance to the reasoning of today’s opinion. Davis and Pierce, and Professor Robert Anthony upon whom they rely, see Anthony, Which Agency Interpretations Should Bind Citizens and the Courts?, 7 Yale J. on Reg. 1 (1990), do indeed set forth the argument I have criticized above, that congressional authorization of informal rulemaking or formal (and perhaps even informal) adjudication somehow bespeaks a congressional intent to “delegate” power to resolve statutory ambiguities. But their analysis does not permit the broad add-ons that the Court’s opinion contains — “some other [procedure] indicating] comparable congressional intent,” ante, at 227, and “we have sometimes found reasons for Chevron deference even when no such administrative formality was required and none was afforded,” ante, at 231.
III
To decide the present case, I would adhere to the original formulation of Chevron. “ ‘The power of an administrative agency to administer a congressionally created ... program necessarily requires the formulation of policy and the making of rules to fill any gap left, implicitly or explicitly, by Congress,’ ” 467 U. S., at 843 (quoting Morton v. Ruiz, 415 U. S. 199, 231 (1974)). We accordingly presume — and our precedents have made clear to Congress that we presume— that, absent some clear textual indication to the contrary, “Congress, when it left ambiguity in a statute meant for im*257plementation by an agency, understood that the ambiguity would be resolved, first and foremost, by the agency, and desired the agency (rather than the courts) to possess whatever degree of discretion the ambiguity allows,” Smiley, 517 U. S., at 740-741 (citing Chevron, supra, at 843-844). Chevron sets forth an across-the-board presumption, which operates as a background rule of law against which Congress legislates: Ambiguity means Congress intended agency discretion. Any resolution of the ambiguity by the administering agency that is authoritative — that represents the official position of the agency — must be accepted by the courts if it is reasonable.
Nothing in the statute at issue here displays an intent to modify the background presumption on which Chevron deference is based. The Court points, ante, at 233, n. 16, to 28 U. S. C. § 2640(a), which provides that, in reviewing the ruling by the Customs Service, the Court of International Trade (CIT) “shall make its determinations upon the basis of the record made before the court.” But records are made to determine the facts, not the law. All this provision means is that new evidence may be introduced at the CIT stage; it says nothing about whether the CIT must respect the Customs Service’s authoritative interpretation of the law. More significant than § 2640(a), insofar as the CIT’s obligation to defer to the Customs Service’s legal interpretations is concerned, is § 2639(a)(1), which requires the CIT to accord a “presumfption of] correctness]” to the Customs Service’s decision. Another provision cited by the Court, ante, at 233, n. 16, is §2638, which provides that the CIT, “by rule, may consider any new ground in support” of the challenge to the Customs Service’s ruling. Once again, it is impossible to see how this has any connection to the degree of deference the CIT must accord the Customs Service’s interpretation of its statute. Such “new ground[s]” may be intervening or newly discovered facts, or some intervening *258law or regulation that might render the Customs Service’s ruling unsound.5
There is no doubt that the Customs Service’s interpretation represents the authoritative view of the agency. Although the actual ruling letter was signed by only the Director of the Commercial Rulings Branch of Customs Headquarters’ Office of Regulations and Rulings, see Pet. for Cert. 47a, the Solicitor General of the United States has filed a brief, cosigned by the General Counsel of the Department of the Treasury, that represents the position set forth in the ruling letter to be the official position of the Customs Service. Cf. Christensen, 529 U. S., at 591 (Scalia, J., concurring in part and concurring in judgment). No one contends that it is merely a “post hoc rationalization]” or an “agency litigating positio[n] wholly unsupported by regulations, rulings, or administrative practice,” Bowen v. Georgetown Univ. Hospital, 488 U. S. 204, 212 (1988).6
*259There is also no doubt that the Customs Service’s interpretation is a reasonable one, whether or not judges would consider it the best. I will not belabor this point, since the Court evidently agrees: An interpretation that was unreasonable would not merit the remand that the Court decrees for consideration of Skidmore deference.
IV
Finally, and least importantly, even were I to accept the Court’s revised version of Chevron as a correct statement *260of the law, I would still accord deference to the tariff classification ruling at issue in this case. For the case is indistinguishable, in that regard, from NationsBank of N. C., N. A. v. Variable Annuity Life Ins. Co., 513 U. S. 251 (1995), which the Court acknowledges as an instance in which Chevron deference is warranted notwithstanding the absence of formal adjudication, notice-and-comment rulemaking, or comparable “administrative formality,” ante, at 231. Here, as in NationsBank, there is a tradition of great deference to the opinions of the agency head, ante, at 231, n. 13. Just two Terms ago, we observed:
“As early as 1809, Chief Justice Marshall noted in a customs case that ‘[i]f the question had been doubtful, the court would have respected the uniform construction which it is understood has been given by the treasury department of the United States upon similar questions.’ United States v. Vowell, 5 Cranch 368, 372. See also P. Reed, The Role of Federal Courts in U. S. Customs & International Trade Law 289 (1997) (‘Consistent with the Chevron methodology, and as has long been the rule in customs cases, customs regulations are sustained if they represent reasonable interpretations of the statute’); cf. Zenith Radio Corp. v. United States, 437 U. S. 443, 450 (1978) (deferring to the Treasury Department’s ‘longstanding and consistent administrative interpretation’ of the countervailing duty provision of the Tariff Act.” United States v. Haggar Apparel Co., 526 U. S. 380, 393 (1999).
And here, as in NationsBank, the agency interpretation in question is officially that of the agency head. Consequently, even on the Court’s own terms, the Customs ruling at issue in this case should be given Chevron deference.
*261For the reasons stated, I respectfully dissent from the Court’s judgment. I would uphold the Customs Service’s construction of Subheading 4820.10.20 of thé Harmonized Tariff Schedule of the United States, 19 U. S. C. § 1202, and would reverse the contrary decision of the Court of Appeals. I dissent even more vigorously from the reasoning that produces the Court’s judgment, and that makes today’s decision one of the most significant opinions ever rendered by the Court dealing with the judicial review of administrative action. Its consequences will be enormous, and almost uniformly bad.
It is not entirely clear whether the formulation newly minted by the Court today extends to both formal and informal adjudication, or simply the former. Cf., e. g., ante, at 230.
Title 5 U. S. C. §706 provides that, in reviewing agency action, the court shall “decide all relevant questions of law” — which would seem to mean that all statutory ambiguities are to be resolved judicially. See *242Anthony, The Supreme Court and the APA: Sometimes They Just Don’t Get It, 10 Am. U. Admin. L. J. 1, 9-11 (1996). It could be argued, however, that the legal presumption identified by Chevron left as the only “questio[n] of law” whether the agency’s interpretation had gone beyond the scope of discretion that the statutory ambiguity conferred. Today’s opinion, of course, is no more observant of the APA’s text than Chevron was — and indeed is even more difficult to reconcile with it. Since the opinion relies upon actual congressional intent to suspend § 706, rather than upon a legal presumption against which § 706 was presumably enacted, it runs head-on into the provision of the APA which specifies that the Act's requirements (including the requirement that judges shall “decide all relevant questions of law”) cannot be amended except expressly. See §559.
See infra, at 247-250.
As a sole, teasing example of those “sometimes” the Court cites NationsBank of N. C., N. A v. Variable Annuity Life Ins. Co., 513 U. S. 251 (1995), explaining in a footnote that our “longstanding precedent” evinced a tradition of great deference to the “‘deliberative conclusions’” of the Comptroller of the Currency as to the meaning of the banking laws the Comptroller is charged with enforcing. Ante, at 231, n. 13. How it is that a tradition of great judicial deference to the agency head provides affirmative indication of congressional intent to delegate authority to resolve statutory ambiguities challenges the intellect and the imagination. If the point is that Congress must have been aware of that tradition of great deference when it enacted the law at issue, the same could be said of the Customs Service, and indeed of all agencies. See, e. g., 4 K. Davis, Administrative Law Treatise §30.08, pp. 237-238 (1958) (describing the “great weight” accorded the “determination^]” of the Federal Trade Commission (quoting FTC v. Cement Institute, 333 U. S. 683, 720 (1948)); Report of the Attorney General’s Committee on Administrative Procedure, S. Doc. No. 8, 77th Cong., 1st Sess., 90-91 (1941). Indeed, since our opinion in Chevron Congress must have been aware that we would defer to all authoritative agency resolutions of statutory ambiguities. Needless to say, NationsBank itself makes no mention of any such affirmative indication, because it was never the law. The many other cases that contradict the Court’s new rule will presumably be explained,- like NationsBank, as other “modes” of displaying affirmative congressional intent. If a tradition of judicial deference can be called that with a straight face, what cannot be?
The Court also states that “[i]t is hard to imagine” that Congress would have intended courts to defer to classification rulings since “the scheme for CIT review includes a provision that treats classification rulings on par with the Secretary’s rulings on ‘valuation, rate of duty, marking, restricted merchandise, entry requirements, drawbacks, vessel repairs, or similar matters,’ ” ante, at 232-233 (quoting 28 U. S. C. § 1581(h), and citing § 2639(b)). I fail to see why this is hard to imagine at all. If anything, the fact that “the scheme for CIT review . .. treats classification rulings on par with the Secretary’s rulings on” such important matters as “ ‘valuation, rate of duty,... restricted merchandise [and] entry requirements,’ ” ante, at 232-233, which often require interpretation of the Nation’s customs and tariff statutes, only strengthens the case for according Chevron deference to whatever statutory interpretations (as opposed to factual determinations) such rulings embody. In other words, the Court’s point is wrong — indeed, the Court’s point cuts deeply into its own case — unless the Court believes that the Secretary’s personal rulings on the legal criteria for imposing particular rates of duty, or for determining restricted merchandise, are entitled to no deference.
The Court’s parting shot, that “there would have to be something wrong with a standard that accorded the status of substantive law to every one of 10,000 ‘official’ customs classifications rulings turned out each year from over 46 offices placed around the country at the Nation’s entryways,” ante, at 238, n. 19, misses the mark. I do not disagree. The *259“authoritativeness” of an agency interpretation does not turn upon whether it has been enunciated by someone who is actually employed by the agency. It must represent the judgment of central agency management, approved at the highest levels. I would find that condition to have been satisfied when, a ruling having been attacked in court, the general counsel of the agency has determined that it should be defended. If one thinks that that does not impart sufficient authoritativeness, then surely the line has been crossed when, as here, the General Counsel of the agency and the Solicitor General of the United States have assured this Court that the position represents the agency’s authoritative view. (Contrary to the Court’s suggestion, there would be nothing bizarre about the fact that this latter approach would entitle the ruling to deference here, though it would not have been entitled to deference in the lower courts. Affirmation of the official agency position before this court — if that is thought necessary — is no different from the agency’s issuing a new rule after the Court of Appeals determination. It establishes a new legal basis for the decision, which this Court must take into account (or remand for that purpose), even though the Court of Appeals could not. See Thorpe v. Housing Authority of Durham, 393 U. S. 268, 282 (1969); see also United States v. Schooner Peggy, 1 Cranch 103 (1801).)
The authoritativeness of the agency ruling may not be a bright-line standard — but it is infinitely brighter than the line the Court asks us to draw today, between a statute such as the one at issue in NationsBank that (according to the Court) does display an “affirmative intent” to “delegate” interpretive authority, and innumerable indistinguishable statutes that (according to the Court) do not. And, most important of all, it is a line that focuses attention on the right question: not whether Congress “affirmatively intended” to delegate interpretive authority (if it entrusted administration of the statute to an agency, it did, because that is how our system works); but whether it is truly the agency’s considered view, or just the opinions of some underlings, that are at issue.