Justice Breyer
delivered the opinion of the Court.
The Social Security Act authorizes payment of disability insurance benefits and Supplemental Security Income to individuals with disabilities. See 49 Stat. 622, as amended, 42 U. S. C. §401 et seq. (1994 ed. and Supp. V) (Title II disability insurance benefits); §1381 et seq. (Title XVI supplemental security income). For both types of benefits the Act defines the key term “disability” as an
“inability to engage in any substantial gainful activity by reason of any medically determinable physical or mental impairment which can be expected to result in death or which has lasted or can be expected to last for a continuous period of not less than 12 months.” § 423(d)(1)(A) (1994 ed.) (Title II) (emphasis added); accord, § 1382c(a)(3)(A) (1994 ed., Supp. V) (Title XVI).
This case presents two questions about the Social Security Administration’s interpretation of this definition.
First, the Social Security Administration (which we shall call the Agency) reads the term “inability” as including a “12 month” requirement. In its view, the “inability” (to engage in any substantial gainful activity) must last, or must be ex[215]*215pected to last, for at least 12 months. Second, the Agency reads the term “expected to last” as applicable only when the “inability” has not yet lasted 12 months. In the case of a later Agency determination — where the “inability” did not last 12 months — the Agency will automatically assume that the claimant failed to meet the duration requirement. It will not look back to decide hypothetically whether, despite the claimant’s actual return to work before 12 months expired, the “inability” nonetheless might have been expected to last that long.
The Court of Appeals for the Fourth Circuit held both these interpretations of the statute unlawful. We hold, to the contrary, that both fall within the Agency’s lawful interpretive authority. See Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837 (1984). Consequently, we reverse.
I
In 1996 Cleveland Walton, the respondent, applied for both Title II disability insurance benefits and Title XVI Supplemental Security Income. The Agency found that (1) by October 31,1994, Walton had developed a serious mental illness involving both schizophrenia and associated depression; (2) the illness caused him then to lose his job as a full-time teacher; (3) by mid-1995 he began to work again part time as a cashier; and (4) by December 1995 he was working as a cashier full time.
The Agency concluded that Walton’s mental illness had prevented him from engaging in any significant work, i. e., from “engaging] in any substantial gainful activity,” for 11 months — from October 31, 1994 (when he lost his teaching job) until the end of September 1995 (when he earned income sufficient to rise to the level of “substantial gainful activity”). See 20 CFR §§404.1574, 416.974 (2001). And because the statute demanded an “inability to engage in any substantial gainful activity” lasting 12, not 11, months, Walton was not entitled to benefits.
[216]*216Walton sought court review. The District Court affirmed the Agency’s decision, but the Court of Appeals for the Fourth Circuit reversed. Walton v. Apfel, 235 F. 3d 184, 186-187 (2000). The court said that the statute’s 12-month duration requirement modifies the word “impairment,” not the word “inability.” Id., at 189. It added that the statute’s “language ... leaves no doubt” that there is no similar “duration requirement” related to an “inability” (to engage in substantial gainful activity). Ibid. It concluded that, because the statute’s language “speaks clearly” and is “unambiguous,” Walton was entitled to receive benefits despite agency regulations restricting benefits to those unable to work for a 12-month period. Ibid.
The court went on to decide that, in any event, Walton qualified because, prior to Walton’s return to work, one would have “expected” his “inability” to last 12 months. Id., at 189-190. It conceded that the Agency had made Walton’s actual return to work determinative on this point. See 20 CFR §§ 404.1520(b), 1592(d)(2) (2001). But it found unlawful the Agency regulations that gave the Agency the benefit of hindsight — on the ground that they conflicted with the statute’s clear command. 235 F. 3d, at 190.
For either reason, the Fourth Circuit concluded, Walton became “entitled” to Title II benefits no later than April 1995, five months after the onset of his illness. See 42 U.S.C. §§423(a)(l)(D)(i), 423(a)(l)(D)(ii) (providing for a 5-month “waiting period” before a claimant is “entitled” to benefits), 423(c)(2)(A) (1994 ed.). It added that Walton’s later work as a cashier was legally beside the point. That work simply counted as part of a 9-month “trial work period,” which the statute grants to those “entitled” to Title II benefits, and which it permits them to perform without loss of benefits. § 422(c).
The Government sought certiorari. It pointed out that the Fourth Circuit’s first holding conflicts with those of other Circuits, compare 235 F. 3d, at 189-190, with Titus v. Sulli[217]*217van, 4 F. 3d 590, 594-595 (CA8 1993), and Alexander v. Richardson, 451 F. 2d 1185 (CA10 1971). It added that the Fourth Circuit’s views were contrary to well-settled law and would create additional Social Security costs of $80 billion over 10 years. We granted the writ. We now reverse.
II
The statutory definition of “disability” has two parts. First, it requires a certain kind of “inability,” namely, an “inability to engage in any substantial gainful activity.” Second, it requires an “impairment,” namely, a “physical or mental impairment,” which provides “reason” for the “inability.” The statute adds that the “impairment” must be one that “has lasted or can be expected to last . . . not less than 12 months.” But what about the “inability”? Must it also last (or be expected to last) for the same amount of time?
The Agency has answered this question in the affirmative. Acting pursuant to statutory rulemaking authority, 42 U. S. C. §§ 405(a) (Title II), 1383(d)(1) (Title XVI), it has promulgated formal regulations that state that a claimant is not disabled “regardless of [his] medical condition,” if he is doing “substantial gainful activity.” 20 CFR § 404.1520(b) (2001). And the Agency has interpreted this regulation to mean that the claimant is not disabled if “within 12 months after the onset of an impairment ... the impairment no longer prevents substantial gainful activity.” 65 Fed. Reg. 42774 (2000). Courts grant an agency’s interpretation of its own regulations considerable legal leeway. Auer v. Robbins, 519 U. S. 452, 461 (1997); Udall v. Tallman, 380 U. S. 1, 16-17 (1965). And no one here denies that the Agency has properly interpreted its own regulation.
Consequently, the legal question before us is whether the Agency’s interpretation of the statute is lawful. This Court has previously said that, if the statute speaks clearly “to the precise question at issue,” we “must give effect to the unambiguously expressed intent of Congress.” Chevron, 467 [218]*218U. S., at 842-843. If, however, the statute “is silent or ambiguous with respect to the specific issue,” we must sustain the Agency’s interpretation if it is “based on a permissible construction” of the Act. Id., at 843. Hence we must decide (1) whether the statute unambiguously forbids the Agency’s interpretation, and, if not, (2) whether the interpretation, for other reasons, exceeds the bounds of the permissible. Ibid.; see also United States v. Mead Corp., 533 U. S. 218, 227 (2001).
First, the statute does not unambiguously forbid the regulation. The Fourth Circuit believed the contrary primarily for a linguistic reason. It pointed out that, linguistically speaking, the statute’s “12-month” phrase modifies only the word “impairment,” not the word “inability.” And to that extent we agree. After all, the statute, in parallel phrasing, uses the words “which can be expected to result in death.” And that structurally parallel phrase makes sense in reference to an “impairment,” but makes no sense in reference to the “inability.”
Nonetheless, this linguistic point is insufficient. It shows that the particular statutory provision says nothing explicitly about the “inability’s” duration. But such silence, after all, normally creates ambiguity. It does not resolve it.
Moreover, a nearby provision of the statute says that an
“individual shall be determined to be under a disability only if his . . . impairment... [is] of such severity that he is not only unable to do his previous work but cannot... engage in any other kind of substantial gainful work which exists in the national economy.” 42 U.S.C. § 423(d)(2)(A) (Title II); accord, § 1382c(a)(3)(B) (Title XVI).
In other words, the statute, in the two provisions, specifies that the “impairment” must last 12 months and also be severe enough to prevent the claimant from engaging in virtually any “substantial gainful work.” The statute, we con[219]*219cede, nowhere explicitly says that the “impairment” must be that severe (i e., severe enough to prevent “substantial gainful work”) for 12 months. But that is a fair inference from the language. See Brief for AARP et al. as Amici Curiae 13 (conceding that an impairment must remain of “disabling severity” for 12 months). At the very least the statute is ambiguous in that respect. And, if so, then it is an equally fair inference that the “inability” must last 12 months. That is because the latter statement (i. e., that the claimant must be unable to “engage in any substantial gainful activity” for a year) is the virtual equivalent of the former statement (i. e., that the “impairment” must remain severe enough to prevent the claimant from engaging in “substantial gainful work” for a year). It simply rephrases the same point in a slightly different \yay.
Second, the Agency’s construction is “permissible.” The interpretation makes considerable sense in terms of the statute’s basic objectives. The statute demands some duration requirement. No one claims that the statute would permit an individual with a chronic illness — say, high blood pressure — to qualify for benefits if that illness, while itself lasting for a year, were to permit a claimant to return to work after only a week, or perhaps even a day, away from the job. The Agency’s interpretation supplies a duration requirement, which the statute demands, while doing so in a way that consistently reconciles the statutory “impairment” and “inability” language.
In addition, the Agency’s regulations reflect the Agency’s own longstanding interpretation. See Social Security Ruling 82-52, p. 106 (cum. ed. 1982) (“In considering ‘duration,’ it is the inability to engage in [substantial gainful activity] that must last the required 12-month period”); Disability Insurance State Manual §316 (Sept. 9, 1965), Government Lodging, Tab C, §316 (“Duration of impairment refers to that period of time during which an individual is continuously unable to engage in substantial gainful activity because [220]*220of” an impairment); OASI Disability Insurance Letter No. 39 (Jan. 22, 1957), id., Tab A, p. 1 (duration requirement refers to the “expected duration of the medical impairment” at a “level of severity sufficient to preclude” substantial gainful activity”). And this Court will normally accord particular deference to an agency interpretation of “longstanding” duration. North Haven Bd. of Ed. v. Bell, 456 U. S. 512, 522, n. 12 (1982).
Finally, Congress has frequently amended or reenacted the relevant provisions without change. E. g., Social Security Amendments of 1965, § 303(a)(1), 79 Stat. 366; see also S. Rep. No. 404, 89th Cong., 1st Sess., pt. I, pp. 98-99 (1965) (“[T]he committee’s bill . . . provide[s] for the payment of disability benefits for an insured worker who has been or can be expected to be totally disabled throughout a continuous period of 12 calendar months” (emphasis added)); id., at 98 (rejecting effort to provide benefits to those with “short-term, temporary disabilities],” defined as inability to work for six months); H. R. Rep. No. 92-231, p. 56 (1971) (“No benefit is payable, however, unless the disability is expected to last (or has lasted) at least 12 consecutive months” (emphasis added)); S. Rep. No. 744, 90th Cong., 1st Sess., 49 (1967) (“The committee also believes . . . that an individual who does substantial gainful work despite an impairment or impairments that otherwise might be considered disabling is not disabled for purposes of establishing a period of disability”). These circumstances provide further evidence — if more is needed — that Congress intended the Agency’s interpretation, or at least understood the interpretation as statutorily permissible. Commodity Futures Trading Comm’n v. Schor, 478 U. S. 833, 845-846 (1986).
Walton points in reply to Title II language stating that a claimant who is “under a disability . . . shall be entitled to a . . . benefit . . . beginning with the first month after” a “waiting period” of “five consecutive calendar months . . . throughout which” he “has been under a disa[221]*221bility.” 42 U. S. C. §§423(a)(l)(D)(i), 423(c)(2)(A). He adds that this 5-month “waiting period” assures a lengthy period of time during which the applicant (who must be “under a disability” throughout) has been unable to work. And it thereby provides ironclad protection against the claimant who suffers a chronic, but only briefly disabling, disease, such as the claimant who suffers high blood pressure in our earlier example. See supra, at 219. This claim does not help Walton, however, for it shows, at most, that the Agency might have chosen other reasonable time periods — a matter not disputed. Regardless, Walton’s “waiting period” argument could work only in respect to Title II, not Title XVI. Title XVI has no waiting period, though it uses identical definitional language. And Walton does not explain why we should interpret the same statutory words differently in closely related contexts. See Department of Revenue of Ore. v. ACF Industries, Inc., 510 U. S. 332, 342 (1994) (“ ‘[Identical words used in different parts of the same act are intended to have the same meaning’ ” (quoting Sorenson v. Secretary of Treasury, 475 U. S. 851, 860 (1986) (some internal quotation marks omitted)).
Walton also asks us to disregard the Agency’s interpretation of its formal regulations on the ground that the Agency only recently enacted those regulations, perhaps in response to this litigation. We have previously rejected similar arguments. Smiley v. Citibank (South Dakota), N. A., 517 U. S. 735, 741 (1996); United States v. Morton, 467 U. S. 822, 835-836, n. 21 (1984).
Regardless, the Agency’s interpretation is one of long standing. See supra, at 220. And the fact that the Agency previously reached its interpretation through means less formal than “notice and comment” rulemaking, see 5 U. S. C. §553, does not automatically deprive that interpretation of the judicial deference otherwise its due. Cf. Chevron, 467 U. S., at 843 (stating, without delineation of means, that the “ ‘power of an administrative agency to administer a congres-[222]*222sionally created ... program necessarily requires the formulation of policy’ ” (quoting Morton v. Ruiz, 415 U. S. 199, 231 (1974))). If this Court’s opinion in Christensen v. Harris County, 529 U. S. 576 (2000), suggested an absolute rule to the contrary, our later opinion in United States v. Mead Corp., 533 U. S. 218 (2001), denied the suggestion. Id., at 230-231 (“[T]he want of” notice and comment “does not decide the case”). Indeed, Mead pointed to instances in which the Court has applied Chevron deference to agency interpretations that did not emerge out of notice-and-comment rule-making. 533 U. S., at 230-231 (citing NationsBank of N. C., N. A. v. Variable Annuity Life Ins. Co., 513 U. S. 251, 256-257 (1995)). It indicated that whether a court should give such deference depends in significant part upon the interpretive method used and the nature of the question at issue. 533 U. S., at 229-231. And it discussed at length why Chevron did not require deference in the circumstances there present — a discussion that would have been superfluous had the presence or absence of notice-and-comment rulemaking been dispositive. 533 U. S., at 231-234.
In this case, the interstitial nature of the legal question, the related expertise of the Agency, the importance of the question to administration of the statute, the complexity of that administration, and the careful consideration the Agency has given the question over a long period of time all indicate that Chevron provides the appropriate legal lens through which to view the legality of the Agency interpretation here at issue. See United States v. Mead Corp., supra; cf. also 1 K. Davis & R. Pierce, Administrative Law Treatise §§ 1.7, 3.3 (3d ed. 1994).
For these reasons, we find the Agency’s interpretation lawful.
Ill
Walton’s second claim is more complex. For purposes of making that claim, Walton assumes what we have just decided, namely, that the statute’s “12 month” duration require[223]*223ments apply to both the “impairment” and the “inability” to work requirements. Walton also concedes that he returned to work after 11 months. But Walton claims that his work from month 11 to month 12 does not count against him because it is part of a “trial work” period that the statute grants to those “entitled” to Title II benefits. See 42 U. S. C. § 422(c). And Walton adds, he was “entitle[d]” to benefits because — even though he returned to work after 11 months — his “impairment” and his “inability” to work were nonetheless “expected to last” for at least “12 months” before he returned to work.
To illustrate Walton’s argument, we simplify the actual circumstances. We imagine: (1) On January 1, Year One, Walton developed (a) a severe impairment, which (b) made him unable to work; (2) Eleven (not twelve) months later, on December 1, Year One, Walton returned to work; (3) On July 1, Year Two, the Agency adjudicated, and denied, Walton’s claim for benefits. Walton argues that, even though he returned to work after 11 months, had the Agency looked at the matter, not ex post, but as if it were looking prior to his return to work, the Agency would have had to conclude that both his “impairment” and his “inability” to work “can be expected to last for a continuous period of not less than 12 months.” § 423(d)(1)(A). He consequently satisfied the 12-month duration requirement and became “entitled” to benefits before he returned to work; he was in turn entitled to a “trial work” period; and his subsequent work as a cashier, being “trial work,” should not count against him.
The Agency’s regulations plainly reject this view of the statute. They say, “You are not entitled to a trial work period” if “you perform work ... within 12 months of the onset of the impairment(s)... and before the date of any notice of determination or decision finding ... you ... disabled.” 20 CFR § 404.1592(d)(2) (2001). This regulation means that the Agency, deciding before the end of Year One, might have found that Walton’s impairment (or inability to work) “can [224]*224be expected to last” for 12 months. But the Agency, deciding after Year One in which Walton in fact returned to work, would not ask whether his impairment (or inability to work) could have been expected to last 12 months.
The legal question is whether this Agency regulation is consistent with the statute. The Court of Appeals, accepting Walton’s view, concluded that it is not. It said that the Agency’s rules — permitting the use of hindsight when reviewing claims — are inconsistent with the statute’s plain language, 235 F. 3d, at 191. And, here, other courts have agreed. See Salamalekis v. Commissioner of Soc. Sec., 221 F. 3d 828 (CA6 2000); Newton v. Chater, 92 F. 3d 688 (CA8 1996); Walker v. Secretary of Health and Human Servs., 943 F. 2d 1257 (CA10 1991); McDonald v. Bowen, 818 F. 2d 559 (CA7 1986).
Nonetheless, we believe that Agency regulation is lawful. See Chevron, supra, at 843. The statute is ambiguous. It says nothing about how the Agency, when it adjudicates a matter after Year One, is to treat an earlier return to work. Its language “can be expected to last” 12 months, 42 U. S. C. § 423(d)(1)(A), simply does not say as of what time the law measures the “expectation.” Indeed, from a linguistic perspective, the phrase “can be expected” foresees a decision-maker who is looking into the future, not a decisionmaker who is in the future, looking back into the past in order to see what then “was,” “could be,” or “could have been” expected. And read in context, the purpose of the phrase “can be expected to last” might be one of permitting the Agency to award benefits before 12 months have expired, not one of denying the Agency the benefit of hindsight. See 65 Fed. Reg., at 42780; cf. also S. Rep. No. 404, at 99.
At the same time, the Agency’s regulation seems a reasonable, hence permissible, interpretation of the statute. In effect it treats a pre-Agency-decision actual return to work, e. g., Walton’s return in December Year One, as if it were determinative of the expectation question. With Year Two’s hindsight, Walton’s “inability” to work “can” not “be [225]*225expected to last 12 months.” And use of that hindsight avoids the need for the Year Two decisionmaker in effect to answer a highly unwieldy question in what grammarians might call the pluperfect future tense.
Of course, administrators and judges are capable of answering hypothetical questions of this kind. But here the question concerns what must be a eontrary-to-faet speculation about the future. It is a speculation that, however often raised, would rarely prove easy to resolve. And the statute’s purpose does not demand its resolution. Indeed, one might ask why, other things being equal, a claimant who returns to work too early ordinarily to qualify for benefits nonetheless should qualify if but only if that return was a kind of medical surprise. Of course, as Walton says, such a rule would help encourage (or at least not discourage) a claimant’s early return to work. See generally S. Rep. No. 1856,86th Cong., 2d Sess., 15-16 (1960). But the statute does not demand that the Agency make of this desirable end an overriding interpretive principle. And the Agency has recognized and addressed the problem of work disincentives in other ways. See, e. g., 20 CFR §§ 404.1574(c), 404.1575(d) (2001).
The statute’s complexity, the vast number of claims that it engenders, and the consequent need for agency expertise and administrative experience lead us to read the statute as delegating to the Agency considerable authority to fill in, through interpretation, matters of detail related to its administration. See Schweiker v. Gray Panthers, 453 U. S. 34, 43-44 (1981). The interpretation at issue here is such a matter. The statute’s language is ambiguous. And the Agency’s interpretation is reasonable.
We conclude that the Agency’s regulation is lawful.
* * *
The judgment of the Fourth Circuit is
Reversed.