with whom The Chief Justice and Kennedy join, dissenting.
The question at issue in this case is whether all, or merely some, of the federal prisoners who were convicted of nonviolent offenses and who have successfully completed a Bureau of Prisons (BOP or Bureau) drug treatment program are eligible for a sentence reduction pursuant to 18 U. S. C. § 3621(e)(2)(B). For the reasons outlined below, I believe that Congress has answered that precise question. The statute expressly states that the sentence of every prisoner in that category “may be reduced.” Ibid. The disposition of this case is therefore governed by the first step in the familiar test announced in Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837, 842-843 (1984), for “Congress has directly spoken to the precise question at issue.” Id., at 842.
I
In drafting the statute in question, Congress was faced with a difficult policy choice: whether the commission of particular crimes made certain categories of offenders so dangerous that the costs of offering them early release in return for the successful completion of a drug treatment program outweighed the rewards. The initial drafts of the bill answered that question in the negative and made all federal prisoners eligible for a sentence reduction of up to one year if they successfully completed a drug treatment program. See, e. g., H. R. Rep. No. 103-320, p. 2 (1993). However, the inclusion of those convicted of violent offenses within the category of those eligible for the inducement soon became a fulcrum of criticism for the larger crime bill within *246which the statute was embedded.1 Perhaps as a result of these criticisms,2 the statute ultimately adopted limited the inducement to “prisoner[s] convicted of . . . nonviolent offense[s].” 18 U. S. C. § 8621(e)(2)(B).
Both the text of the statute and the aforementioned history demonstrate that Congress directly addressed the “precise question” of what offenses ought to disqualify prisoners from eligibility for a sentence reduction, and that its unambiguous answer was “violent offenses.” Under the statute as enacted, those who commit crimes of violence are categorically barred from receiving a sentence reduction while those convicted of nonviolent offenses “may” receive such an inducement.
*247The BOP regulation challenged here operates to redefine the set of prisoners categorically ineligible for a sentence reduction, a set unambiguously defined in the text of the statute. It does so by taking a group of prisoners whose offenses the Bureau acknowledges are “nonviolent” within the meaning of the statute3 and imposing the same sanction — categorical ineligibility — upon them as the statute imposes upon violent offenders. In so doing, the Bureau ignores Congress’ express determination that, when evaluating eligibility for a sentence reduction, the salient distinction is the line between violent and nonviolent offenses. By moving this line, the BOP exceeded its authority and sought to exercise its discretion on an issue with regard to which it has none. See, e. g., Chevron, 467 U. S., at 842-843 (“First, always, is the question whether Congress has directly spoken to the precise question at issue. If the intent of Congress is clear, that is the end of the matter; for the court, as well as the agency, must give effect to the unambiguously expressed intent of Congress”); United States v. Haggar Apparel Co., 526 U. S. 380, 392 (1999) (“In the process of considering a regulation in relation to specific factual situations, a court may conclude the regulation is inconsistent with the statutory language or is an unreasonable implementation of it. In those instances, the regulation will not control”).
*248II
I fully agree with the majority that federal prisoners do not become entitled to a sentence reduction upon their successful completion of a drug treatment program; the words "may be reduced” do not mean “shall be reduced.” Nonetheless, while the statute does not entitle any prisoner to a sentence reduction, it does guarantee nonviolent offenders who successfully complete a drug treatment program consideration for such a reduction.
For every nonviolent offender who participates in a drug treatment program, the BOP may be required to make two individualized determinations: (1) whether he or she has successfully completed that program; and (2), if sc, whether his or her preconviction conduct, postconviction conduct, and prospects for rehabilitation justify a sentence reduction. In evaluating whether or not a particular individual is entitled to a sentence reduction, the BOP may give great weight to whichever of these factors it determines to be most relevant. That, however, is a far cry from categorically excluding from consideration prisoners who Congress explicitly intended to obtain such consideration.4
The majority’s concern about the risks and burdens associated with case-by-case decisionmaking in a large number of cases is understandable yet ultimately misguided. In order to fulfill the statute’s requirements, the BOP must already *249evaluate every prisoner seeking the sentence reduction on an individual basis to determine whether that prisoner “successfully completed” his or her drug treatment program. Individualized consideration of the second salient question involves consideration of many of the same personalized factors that go into determining whether a prisoner’s course of drug treatment has been “successful.” To the extent that answering the second question requires consideration of additional factors with a concomitant administrative burden, the costs of such a scheme are, in Congress’ judgment, outweighed by the benefits of encouraging drug treatment and of carefully distinguishing between those prisoners who have earned an early return to their communities and those who require further incarceration.
The majority’s, worry that individualized decisionmaking might lead to “favoritism, disunity, and inconsistency” is similarly misplaced. Ante, at 244. To suggest that deci-sionmaking must be individualized is not to imply that it must also be standardless. If the Court today invalidated the regulation in question, its decision would not preclude the BOP from adopting a uniform set of criteria for consideration in evaluating applications for sentence reductions. Nor would it necessarily preclude the Bureau from giving dispositive weight to certain postconviction criteria or near-dispositive weight to preconviction criteria. Cf. Heckler v. Campbell, 461 U. S. 458, 467 (1983). The Bureau would remain free to structure its decisionmaking in any way it saw fit as long as in so doing it did not contravene policy decisions explicitly made by the statute’s drafters. As Congress has already addressed preinearceration conduct in § 3621(e)(2)(B), the Bureau may not categorically exclude a prisoner not convicted of a violent offense from consideration for early release on the basis of such conduct without exceeding the limits of its discretion.
Accordingly, I respectfully dissent.
Throughout 1993 and 1994, Republican leaders gave numerous speeches contrasting their proposed crime bill and the administration’s. One contrast repeatedly stressed was that the Republican bill set aside more money for prison construction while the Democratic bill allocated greater funds to drug treatment. This difference allegedly reflected differing views as to how society should deal with violent criminals. To this end, Republican leaders repeatedly criticized the inclusion of violent criminals in the sentence reduction provision. See, e. g., 139 Cong. Rec. 27209 (1994) (remarks of Sen. Hatch) (“Their treatment allows all Federal prisoners, including the most violent, to have their sentences reduced, if you will, at the Bureau of Prisons’ discretion if they complete a drug treatment program. Boy, I can see where everybody is going to do that. You can imagine the sincerity of that”); 139 Cong. Rec. 27460 (1993) (remarks of Sen. Hatch) (“The Democratic crime bill actually permits the Bureau of Prisons to decrease the sentence of Federal inmates — violent offenders included — who complete drug treatment programs. Their bill also proposes that States be given grant money which can be used to implement home confinement and other alternative sanctions for violent offenders”).
The House initially approved a version of the bill that would have extended the inducement to all federal prisoners. The Senate, where the criticism of the inclusion of violent offenders was more pronounced, see n. 1, supra, limited the provision to nonviolent offenders. The Conference Committee accepted the Senate’s limitation. H. R. Conf. Rep. No. 103-711, p. 381 (1994).
The BOP regulation challenged here treats an otherwise nonviolent offense where a gun was carried as a “nonviolent offense” within the meaning of the statute, and the case was argued on that assumption. As the majority notes, ante, at 233-236, the BOP initially attempted to classify such crimes as violent offenses, but receded when the Courts of Appeals divided over the validity of such a construction. The question over which the Courts of Appeals initially divided is not before us today. If it were, the arguments raised by both.sides would be quite different, with the debate likely focusing on whether “nonviolent offense” is best understood as a term of art or in relation to a more colloquial understanding of violence.
This Court’s decision in INS v. Yueh-Shaio Yang, 519 U. S. 26 (1996), relied upon by the majority, ante, at 242-248, is not to the contrary. Yueh-Shaio Yang did not involve an effort by an administrative agency to categorically exclude from consideration for a benefit a particular class of individuals because of a characteristic considered and rejected by Congress as a basis for categorical exclusion. Rather, that case involved the related yet distinct question whether such a characteristic may be given any weight by the agency in making an individualized case-by-case determination whether to grant the benefit to a particular individual. If the issue in this case were whether the BOP could even consider the nature of the offense in determining whether to grant a particular sentence reduction, Yueh-Shaio Yang would be relevant to our analysis.