Affirmed by published opinion. Judge NIEMEYER wrote the opinion, in which Judge TRAXLER and Chief Judge STAMP joined.

OPINION

For his bribery-related convictions, Paul Kinter was sentenced to 46 months imprisonment, a term based on the amount of benefit that a- government contractor received as a result of the bribes rather than the lesser amount of benefit that Kinter personally received from the scheme. It is this lesser amount that Kinter contends is appropriate to consider under U.S.S.G. § 2Cl.l(b)(2)(A). Because the Sentencing Guidelines’ general application principles stated in § IB 1.3(a)(1) instruct that a court’s determination of the amount of “benefit received” must be informed by the scope of Kinter’s activities as well as the reasonably foreseeable activities of persons acting jointly with him, we affirm. In doing so, we also reject Kinter’s argument that the Supreme Court’s recent decision in Apprendi v. New Jersey, 530 U.S. 466, 120 S.Ct. 2348, 147 L.Ed.2d 435 (2000), renders his sentence unconstitutional because it was based in part upon judge-made findings pursuant to the Sentencing Guidelines.

I

When Scott King, an IRS employee, informed Paul Kinter, his former father-in-law, during the summer of 1990 that the IRS planned to consolidate many of its computer maintenance contracts into a single, multimillion dollar contract that it would award to a company certified by the Small Business Administration as a § 8(a) contractor, * the two men decided to sell King’s influence at the IRS in exchange for kickbacks from a yet-to-be-identified company for whom they would obtain the *194 contract. At the time, King was the technical representative for the IRS’s Martins-burg, West Virginia facility and the person on whom the contracting officer, who had authority to award the procurement contract, relied.

With the assistance of a co-conspirator, Mark Nicholas, Kinter eventually located and brought into the scheme Washington Data Systems, Inc., and its subcontractor RGI, Inc. (collectively “Washington Data”), as a § 8(a) contractor. In furtherance of the scheme, Washington Data hired Kinter and Nicholas as “consultants” and agreed to pay them a kickback of approximately 3% of any revenue that Kinter and Nicholas would secure for Washington Data. Kinter and Nicholas were to pay King his share from their amount. Kinter and Nicholas initially paid King approximately $300 per week. After King recommended Washington Data to the contracting officer and Washington Data obtained its first purchase order, King’s payments increased to $500 per week. Washington Data had no previous experience in computer maintenance and would not have received the IRS’s contract but for King’s influence with the contracting officer.

Following the successful completion of that first contract, Washington Data had its “foot in the door” and received approximately 30 short-term and long-term purchase orders from the IRS over the following 5 years. As Kinter had anticipated, the revenues to Washington Data from these contracts exceeded $57 million, generating $9.5 million in profits for Washington Data. Although King ceased to be the technical representative at the IRS’s Mar-tinsburg facility in 1992, Kinter and King continued to receive payments from Washington Data until December 1996. In the aggregate, Kinter received between $340,000 and $350,000 from Washington Data, and he paid a substantial portion of this amount to King.

The grand jury indicted Kinter in December 1998 on charges of conspiracy, in violation of 18 U.S.C. § 371; bribery of a 201(b)(1); and payment of a gratui public official, in violation of 18 t 201(c)(1)(A). Following Kinter’s £ ea to the charges, the district court need Kinter to two concurrent 46-n rms of imprisonment on the briber; nspiracy charges, and one concu -month term on the gratuity charge lculating Kinter’s sentences for the y and conspiracy counts, the dii urt enhanced Kinter’s offense level 1 rels based on the $9.5 million in bei ceived by Washington Data as a res' e bribery scheme. In doing so, the i jected Kinter’s argument that it si ve considered only the $340,001 50,000 amount that Kinter personal! ived. Had the court accepted Kir sition, it would have enhanced Kir :ense level only 8 levels, exposing hi sentencing range of 24-30 monthi isonment. public official, in violation of 18 U.S.C. § 201(b)(1); and payment of a gratuity to a public official, in violation of 18 U.S.C. § 201(c)(1)(A). Following Kinter’s guilty plea to the charges, the district court sentenced Kinter to two concurrent 46-month terms of imprisonment on the bribery and conspiracy charges, and one concurrent 24-month term on the gratuity charge. In calculating Kinter’s sentences for the bribery and conspiracy counts, the district court enhanced Kinter’s offense level by 14 levels based on the $9.5 million in benefits received by Washington Data as a result of the bribery scheme. In doing so, the court rejected Kinter’s argument that it should have considered only the $340,000-to-$350,000 amount that Kinter personally received. Had the court accepted Kinter’s position, it would have enhanced Kinter’s offense level only 8 levels, exposing him to a sentencing range of 24-30 months imprisonment.

This appeal followed. This appeal followed.

II

For bribery offenses, the Sentencing Guidelines provide that the sentence shall be enhanced by the greatest of (1) the value of the bribery payment, (2) the “benefit received or to be received” as a result of the bribery payment, or (3) the loss to the government. U .S.S.G. § 201.1(b)(2)(A). If the dollar amount so identified exceeds $2,000, the enhancement is prescribed by the table contained in U.S.S.G. § 2F1.1. See id. That table provides an 8 level enhancement if the dollar amount is more than $200,000, and a 14-level enhancement if the dollar amount is more than $5 million. See id. § 2F1.1(b)(1).

The government contends that the proper measure for determining the enhancement in this case is the “benefit received” by Washington Data—the $9.5 million profit that it received from the IRS contracts, yielding the 14-level enhancement that the district court found in this case. *195 Kinter contends that because he was paid between $340,000 and $350,000, the 8-level enhancement is the correct one. He argues that the § 2C1.1 enhancement contemplates only the amount of benefit that he personally received, not the benefit received by Washington Data.

Because resolution of this issue turns primarily upon the legal interpretation of the Sentencing Guidelines, our standard of review is de novo. See United States v. Nale, 101 F.3d 1000, 1003 (4th Cir.1996); United States v. Jones, 31 F.3d 1304, 1315 (4th Cir.1994).

Kinter’s argument discounts the effect of his crime in a manner that is contrary to the explicit provisions of the Sentencing Guidelines. Section 2C1.1(b)(2) characterizes a “benefit received” as a “specific offense characteristic,” and § lB1.3(a)(l), providing general application principles for the Sentencing Guidelines, instructs that “specific offense characteristics” are to be determined on the basis of

(A) all acts and omissions committed, aided, abetted, counseled, commanded, induced, procured, or willfully caused by the defendant; and

(B) in the case of a jointly undertaken criminal activity (criminal plan, scheme, endeavor, or enterprise undertaken by the defendant in concert with others, whether or not charged as a conspiracy), all reasonably foreseeable acts and omissions of others in furtherance of the jointly undertaken criminal activity,

that occurred during the commission of the offense of conviction, in preparation for that offense, or in the course of attempting to avoid detection or responsibility for that offense....

U.S.S.G. § lB1.3(a)(l) (emphases added). These provisions create two separate bases upon which we conclude that “benefit received,” for the purposes of sentencing Kinter, must include Washiñgton Data’s $9.5 million profit from the IRS contracts. First, Kinter induced, procured, and willfully caused Washington Data to obtain the IRS contracts. As the district court found:

Mr. Nicholas may have been more important than Mr. Kinter in getting to an 8(a) [contractor] that was willing to pay the kickbacks, but Mr. Kinter was no less instrumental in arranging for this three-way relationship to take place.

He not only, in my view, was necessary to that beginning, he participated in it, ... [and he] was intimately involved in the decisions that the co-conspirators made and carried out.

Mr. Kinter was very involved in the ongoing process of receiving the commissions from [Washington Data] and getting them laundered and split up.

These findings alone make the $9.5 million in profits received by Washington Data relevant to the § 2C1.1 calculus through the application of § lB1.3(a)(l)(A).

The district court also found that Kinter and Washington Data undertook the bribery conspiracy jointly and that during the conspiracy Kinter foresaw the scope of the continued course of dealing between the IRS and Washington Data, thus rendering the contractor’s profits includable also through application of § lB1.3(a)(l)(B). The district court stated from the bench,

[T]he initiation of this and the continuing receipt of the commissions [from Washington Data] is all part of the conduct that constitutes the bribery scheme. There was no break, no unforeseen intervening event that stopped the course of this bribery scheme. The fact that IRS continued dealing with [Washington Data] was not unforeseen. It was, in fact, foreseen. It was the purpose of setting up the arrangement.

(Emphasis added). The court found further,

[Washington Data] decided to take [Nicholas and Kinter] up on their offer and benefited mightily from it, so did Mr. Kinter, financially while it lasted, *196 and I don’t think that the $9.5 million or, frankly, over $5 million, the 14 level adjustment overrepresents the seriousness of what happened here.

These findings likewise make the full $9.5 million relevant in the § 2C1.1 determination. See United States v. Agostino, 132 F.3d 1183, 1196-97 (7th Cir.1997) (holding that the district court should have considered only personal benefit in its § 2C1.1 calculation because the relevant criminal activity was not “jointly undertaken”); United States v. Pretty, 98 F.3d 1213, 1222 (10th Cir.1996) (discussing the foreseeability of co-conspirators’ profits in the § 2C1.1 context).

In addition to the broad scope of activities made relevant by both subsections (A) and (B) of § lB1.3(a)(l), Kinter’s position is in tension with the Sentencing Guidelines commentary, which provides that “for deterrence purposes, the punishment [for bribery] should be commensurate with the gain to the payer or the recipient of the bribe, whichever is higher.” U.S.S.G. § 2C1.1, cmt. background; see also United States v. Muldoon, 931 F.2d 282, 289 (4th Cir.1991) (citing analogous commentary). Thus, in cases involving a middleman in a bribery scheme, such as Kinter, a court should first determine which party was, in actuality, the payer of the bribe and then calculate the gain to that payer. This analysis differs little in substance from the inquiry required by § lB1.3(a)(l). If Kin-ter had acted within his individual capacity — if, for example, he had bribed the government official without Washington Data’s knowledge — Washington Data’s profits would not have been an accurate measure of the “gain to the payer.” Because, however, Kinter acted as an agent for Washington Data, Washington Data was, in reality, the relevant payer, and Kinter’s sentence was properly enhanced by the gain to Washington Data. See United States v. Bankston, 182 F.3d 296, 317-18 (5th Cir.1999); Agostino, 132 F.3d at 1196-97 n. 5 (citing Muldoon, 931 F.2d at 289). This commentary thus reflects the Sentencing Commission’s informed belief that the repugnance of a middleman’s crime and the magnitude of its effect upon the efficient operation of government are more accurately measured by the ultimate payer’s profits than by the middleman’s relatively scant take from the scheme. See United States v. Gillam, 167 F.3d 1273, 1278-79 (9th Cir.1999).

At bottom, because substantial evidence supported the lower court’s finding that Kinter acted on Washington Data’s behalf, the district court’s inclusion of the $9.5 million in its § 2C1.1 calculus was consistent with the Sentencing Guidelines and its commentary.

Notwithstanding these Sentencing Guidelines’ instructions, Kinter argues skillfully that our decision in United States v. Ellis, 951 F.2d 580 (4th Cir.1992), confines the scope of “benefit received” under § 2C1.1 to the defendant’s personal benefit. We do not agree, however, with this reading of Ellis. In Ellis, the defendant, a limited partner in a dog track, was convicted of paying bribes to West Virginia legislators in return for their votes in support of a bill that increased the profitability of such tracks statewide. Rejecting the government’s argument that the benefit to at least two tracks in the state should have been considered in sentencing the defendant, we stated, “Neither the provision itself nor precedent requires that vicarious liability apply to § 2C1.1 calculations.” Id. at 585. Kinter, interpreting this passage broadly, concludes that Ellis stands for the proposition that “vicarious liability has no applicability in calculating the § 2C1.1 enhancement.”

We were careful in Ellis, however, to limit our holding to its specific facts and distinguish Muldoon, a case in which we stated that the district court should not focus upon the defendant’s personal benefit when making § 2C1.1 determinations. The Ellis decision labeled the Muldoon statement “dictum,” see Ellis, 951 F.2d at 585, and in any event distinguished it, saying,

*197 Muldoon involved discrete government contracts, for which there is a reliable measure of the total benefit received-net profit on the illegally obtained contract. In contrast, calculating the total benefit received by companies who profit from improperly passed legislation is a far less certain endeavor, one that (as this case suggests) is potentially limitless in reach.

Ellis, 951 F.2d at 586 (citing Muldoon, 931 F.2d at 289); see also United States v. Kant, 946 F.2d 267, 269 (4th Cir.1991) (using, without protest from the defendant, corporate benefit as the measure of “benefit received”).

Our driving concern in Ellis was the difficulty inherent in computing the total benefit from the bribe-induced legislation at issue—a vague, indeterminate measure that would have extended to persons and business that had no direct relationship with the defendant. Cf. Gillam, 167 F.3d at 1279 (“In Ellis, the court of appeals simply rejected the government’s cross-appeal from a determination that the § 2C1.1 enhancement for benefit received does not permit the court to take into account benefits received by entities not participants in the bribery scheme”). Significantly, we did not hold in Ellis that “benefit received” never includes corporate benefits; rather, we merely concluded that personal benefit might be the more appropriate measure in a case in which the court could not readily determine the third-party beneficiary’s total profits. And we specifically provided that in factual circumstances involving “discrete government contracts”—such as the IRS contracts in the case before us—the district court need not confine its analysis to the defendant's personal benefit. Ellis, 951 F.2d at 586.

Perhaps more important to the applicability of Ellis to the case before us is the fact that we rendered our decision in Ellis prior to the 1992 amendments to U.S.S.G. § 1B1.3. Under the pre-1992 guideline, specific offense characteristics were determined only on the basis of “all acts and omissions committed or aided and abetted by the defendant.” That provision did not include the language currently found in § lB1.3(a)(l)(A) that encompasses acts and omissions “induced, procured, or willfully caused by the defendant.” See U.S.S.G. Appendix C, amend. 439. More significantly, the pre 1992 version of § 1B1.3 contains no analogue to the current § lB1.3(a)(l)(B), which includes in the determination all reasonably foreseeable acts in furtherance of a jointly undertaken criminal activity. Thus, the 1992 amendments to § 1B1.3 have abrogated the Ellis court’s rejection, in the § 2C1.1 context, of “the principle that a conspirator must answer for his conduct and for all foreseeable consequences of it,” Ellis, 951 F.2d at 585 (citing Pinkerton v. United States, 328 U.S. 640, 66 S.Ct. 1180, 90 L.Ed. 1489 (1946)), at least insofar as the “benefit received” flows to other co-conspirators.

Thus, because the pertinent language in Ellis has no application to the facts of this case and, in any event, has been put into question by subsequent amendments to the Sentencing Guidelines, its holding does not bar us from imputing to Kinter, in accordance with the Sentencing Guidelines, all of the benefits received by Washington Data as a result of his bribery activities. In so holding, we join all of the other circuit courts that have considered the issue, which have uniformly held that when a middleman defendant acts on behalf of a third-party payer of the bribe, the district court may consider the payer’s bribe-generated benefits when calculating the “benefit received” under U.S.S.G. § 2C1.1, as long as those profits were reasonably foreseeable or the result of acts aided, abetted, counseled, commanded, induced, procured, or willfully caused by the defendant. See Bankston, 182 F.3d at 317-18 (5th Cir.); Gillam, 167 F.3d at 1278-79 (9th Cir.); United States v. Muhammad, 120 F.3d 688, 699-700 (7th Cir.1997); Pretty, 98 F.3d at 1222 (10th Cir.); United States v. Ziglin, 964 F.2d 756, 758 n. 3 (8th Cir.1992); cf. United *198 States v. Cohen, 171 F.3d 796, 803 (3d Cir.1999) (interpreting “benefit conferred” language of U.S.S.G. § 2B4.1 as contemplating corporate benefit); United States v. Landers, 68 F.3d 882 (5th Cir.1995) (assuming, without deciding, the same).

Ill

Seizing upon the Ellis court’s discussion of the reliability-of-profit measures, Kinter also contends that the $9.5 million profit figure was not sufficiently reliable to be included in the district court’s calculation of “benefit received.” He argues that the $9.5 million amount represents benefits received by Washington Data not only for contracts to which the bribes were directly related but also subsequent contracts awarded to Washington Data. Kinter’s argument is grounded upon his assertion that he made payments to King only in exchange for King’s recommendation that Washington Data be awarded an initial contract, valued at $950,000. He maintains that after the first contract, a combination of institutional inertia, competent performance by Washington Data, and a lack of other qualified § 8(a) contractors led to the subsequent awards worth $57 million to Washington Data over the course of several years. In other words, he claims that the evidence did not support the sentencing court’s finding that his bribes were the but-for cause of the $9.5 million benefit to Washington Data. The district court’s determination on this mixed question of law and fact is subjected to a standard approximating clear error review when, as here, the issues involved are “essentially factual.” United States v. Daughtrey, 874 F.2d 213, 217 (4th Cir.1989).

The threshold for the causation inquiry for § 2C1.1 calculations is relatively low. Cf. United States v. Sapoznik, 161 F.3d 1117, 1119 (7th Cir.1998) (“To show that the bribes benefited the people paying them ... it is enough for the government to show that the bribes facilitated the gambling operations”). In this case, Kinter’s resume and handwritten notes show his expectation that the initial bribe would allow Washington Data to get its “foot in the door” and thereafter produce $60 million in revenues for Washington Data. The district court also cited testimony that Kin-ter remained involved in the scheme long after the first contract, that “[h]is consulting fees continued,” and that “Mr. Kinter was very involved in the ongoing process of receiving the commissions from [Washington Data] and getting them laundered and split up.” Furthermore, there was testimony that Washington Data would not have bid on the contracts absent Kinter’s intervention and that the contracting officer relied upon King’s advice. The record amply supports these factual findings.

Kinter points out that the “foot in the door” evidence should not justify an analysis that would permit the “ ‘benefit received’ to run in perpetuity.” While we agree with this observation, we note that U.S.S.G. § 1B1.3 places an important limitation on this analysis. That section provides that defendants may be held responsible for actions taken in furtherance of a jointly undertaken criminal activity only when those acts are “reasonably foreseeable.” U.S.S.G. § lB1.3(a)(l)(B). Thus, even though we agree that Kinter might not have been held responsible, under a foreseeability standard, for revenues accruing to Washington Data that exceeded $60 million, the evidence supports a conclusion that Kinter expected the illegal conduct to produce up to $60 million in revenues, the amount that yielded $9.5 million in direct benefits to Washington Data. Indeed, as the district court found, “there is a very strong link” between the amount of the bribe paid to King and Washington Data’s $9.5 million profit from the continued contracts with the IRS, the securing of which was “the purpose of setting up the arrangement.”

IV

Finally, we permitted Kinter to file a supplemental brief addressing the rele- *199 vanee to his case of the Supreme Court’s recent decision in Apprendi v. New Jersey, 530 U.S. 466, 120 S.Ct. 2348, 147 L.Ed.2d 435 (2000). Because he failed to raise this issue below, we review the district court’s decision for plain error. See United States v. Hastings, 134 F.3d 235, 239 (4th Cir.1998).

At issue in Apprendi was a New Jersey hate crime statute that provided for the enhancement of a defendant’s sentence if the trial judge found, by a preponderance of the evidence, that the offense was committed with a “purpose to intimidate an individual or group of individuals because of race, color, gender, handicap, religion, sexual orientation or ethnicity.” Apprendi, 120 S.Ct. at 2351 (quoting N.J. Stat. Ann. § 2C:44-3(e) (internal quotation marks omitted)). After Apprendi fired several bullets into the home of an African-American family in a previously all-white neighborhood, he pled guilty before a state court to a firearm possession charge for which the maximum penalty, as established by New Jersey statute, was 10 years imprisonment. See id. at 2352. The trial court, however, found by a preponderance of the evidence that Apprendi’s conduct was “motivated by racial bias” and sentenced him to a 12-year term. Id. After the New Jersey Supreme Court affirmed the judgment, the United States Supreme Court reversed, concluding that “[ojther than the fact of a prior conviction, any fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury, and proved beyond a reasonable doubt.” Id. at 2362-63. Because Apprendi had not admitted his bias in the plea agreement and New Jersey had not proven that bias to a jury beyond a reasonable doubt, the Court concluded that Apprendi could be subjected only to the 10-year maximum punishment for firearm possession — the only crime for which he had waived the process guaranteed to him by the Fifth and Fourteenth Amendments. See id. at 2355-56.

case us, judge determined, under a pre-ponderance standard, that Kinter paid more than one bribe and that Washington Data’s profit was $9.5 million — findings that required the court to impose a sentence of between 46 and 57 months. In the absence of these findings, the maximum punishment allowable under the Sentencing Guidelines for a person standing in Kinter’s shoes would have been ten months. See U.S.S.G. § 2C1.1; id. Ch. 5, Pt. A (Sentencing Table). Kinter therefore contends that Ap-prendi required those two facts to be submitted to a jury and proven beyond a reasonable doubt before they could form the basis of an enhancement of his sentence. This contention essentially boils down to an argument that Apprendi renders much, if not all, of the current sentencing practices under the Sentencing Guidelines unconstitutional. Under Kin-ter’s interpretation of Apprendi, any determination that has the real effect of increasing the maximum punishment to which a defendant is subject under the Sentencing Guidelines must be made by a jury and proven beyond a reasonable doubt.

The Apprendi Court, however, did not paint with the broad brush that Kinter now offers us. On the contrary, the majority opinion explicitly limited its holding to factual determinations “that increase[ ] the penalty for a crime beyond the prescribed statutory maximum.” Apprendi, 120 S.Ct. at 2362-63 (emphasis added). The outcome of our case thus turns upon the definition of “prescribed statutory maximum.” If the district court’s factual findings did not result in an enhancement that exceeded that maximum, Apprendi is irrelevant. The government contends, and all of the Courts of Appeals to have considered the issue have thus far agreed, that to find the “prescribed statutory maximum” as contemplated in Apprendi, one need only look to the language of the statute *200 criminalizing the offense, and no further. See United States v. Nealy, 232 F.3d 825, 829 n. 3 (11th Cir.2000); United States v. Doggett, 230 F.3d 160, 165 (5th Cir.2000); Talbott v. Indiana, 226 F.3d 866, 869 (7th Cir.2000). If we followed that conclusion, we would have to reject Kinter’s Apprendi claim, as the statutes at issue in this case — 18 U.S.C. §§ 201(b)(1) and 371-permit sentences of up to 15 and 5 years, respectively, which are well in excess of the 46-month concurrent sentences that Kinter received.

Although we ultimately agree with the conclusion reached by these other Courts of Appeals, Kinter’s argument is not without support, and the issue is sufficiently complex to warrant a brief discussion here. After all, the Apprendi dissenters expressed their fear that Apprendi would eventually stand for the principle that “a defendant is entitled to have a jury decide, by proof beyond a reasonable doubt, every fact relevant to the determination of [his] sentence under a determinate-sentencing scheme” — a fear that the majority did little to allay. Apprendi, 120 S.Ct. at 2393-94 (O’Connor, J., dissenting). Moreover, claims such as Kinter’s are indeed covered by the holding of Apprendi if the relevant “prescribed statutory maximum” is found in the Sentencing Guidelines rather than on the face of the relevant substantive statute.

And though we reject it here, there is at least a colorable argument that the Sentencing Guidelines do provide that maximum. As Justice Thomas noted in his concurring opinion in Apprendi, “the Guidelines ‘have the force and effect of laws.’ ” Apprendi, 120 S.Ct. at 2380 n. 11 (Thomas, J., concurring) (quoting Mistretta v. United States, 488 U.S. 361, 413, 109 S.Ct. 647, 102 L.Ed.2d 714 (1989) (Scalia, J., dissenting)). Moreover, because the máximums set by the Sentencing Guidelines may not be exceeded by sentencing judges, they are legally binding enactments in a manner nearly indistinguishable from congressionally enacted criminal statutes. If, for example, the district court in this case had sentenced Kinter to 59 months imprisonment on the bribery charge, we would have been required to vacate it because the court would have disregarded the maximum 57-month penalty for Kinter’s crime prescribed by law (i.e., by the Sentencing Guidelines) — even though the 59-month sentence would have been well below the 15-year maximum established by 18 U.S.C. § 201(b)(1). See Mistretta, 488 U.S. at 413, 109 S.Ct. 647 (Scalia, J., dissenting). Moreover, the sentencing ranges promulgated by the Commission — including their máximums and mínimums — are incorporated into the federal statutes by 18 U.S.C. § 3553(b).

If this analysis were correct, Apprendi would indeed work a water-shed change upon the federal courts’ current sentencing practices. District courts would no longer be permitted to make factual determinations that had the effect, in any real sense, of enhancing the defendant’s sentence, and the Sentencing Guidelines would thus be rendered essentially useless, insidiously undermining the constitutional seal of approval bestowed upon the Sentencing Commission by the Supreme Court in Mistretta. Cf. Apprendi, 120 S.Ct. at 2391-95 (O’Connor, J., dissenting) (warning that the majority opinion would have precisely this effect). But cf. id. at 2400 (Breyer, J., dissenting) (noting that the majority had expressed “no constitutional objection” to the Guidelines).

We conclude, however, that the Sentencing Guidelines pass muster under the Apprendi Court’s conception of due process for reasons that closely parallel the principles animating the Mistretta Court’s separation-of-powers-based decision. Mistretta made clear that the Sentencing Commission and its Sentencing Guidelines enjoy a unique constitutional status. See Apprendi, 120 S.Ct. at 2380 n. 11 (Thom *201 as, J., concurring). Although Congress delegated significant legislative power to the Commission to “promulgate sentencing guidelines for every federal criminal offense,” Mistretta, 488 U.S. at 371, 109 S.Ct. 647, the substance of that delegation is decidedly nonlegislative in character. The Sentencing Guidelines’ function—to “impose sentences within the broad limits established by Congress,” id. at 396, 109 S.Ct. 647—is, the Mistretta Court explained, “clearly attendant to a central element of the historically acknowledged mission of the Judicial Branch,” id. at 391, 109 S.Ct. 647 (comparing the Guidelines to the Federal Rules of Criminal and Civil Procedure). In other words, the Commission’s act of establishing sentencing ranges in the Guidelines is categorically different from the legislative act of setting a maximum penalty in a substantive criminal statute. The Commission’s task is not the fundamentally legislative one of creating and delineating hundreds of new crimes through the Guidelines. See id. at 373 n. 7, 109 S.Ct. 647. Instead, the Sentencing Guidelines are merely “a constitutional mechanism for channeling the discretion that a sentencing court would otherwise enjoy” in the absence of the Guidelines. United States v. Mack, 229 F.3d 226, 244 (3d Cir.2000) (Becker, C.J., concurring).

This characterization of the Sentencing Guidelines is extremely significant because the Supreme Court, in both Apprendi and its precursor, Jones v. United States, 526 U.S. 227, 119 S.Ct. 1215, 143 L.Ed.2d 311 (1999), explained that there is no constitutional infirmity in a trial court’s use of facts proven only by a preponderance of the evidence when exercising the wide discretion to “impos[e] a judgment unthin the range prescribed by statute.” Apprendi, 120 S.Ct. at 2358; see also Jones, 526 U.S. at 248, 119 S.Ct. 1215 (“It is not, of course, that anyone today would claim that every fact with a bearing on sentencing must be found by a jury; we have resolved that general issue and have no intention of questioning its resolution”); cf. United States v. Watts, 519 U.S. 148, 156, 117 S.Ct. 633, 136 L.Ed.2d 554 (1997) (per curiam) (holding that, in general, the use of a preponderance standard at sentencing satisfies due process). There is, accordingly, no due process infirmity in a district court’s use of similarly-proven facts to determine a sentence under the constitutionally unique process established by the Commission, which is itself part of the judicial branch, given that the Sentencing Guidelines merely serve to regulate discretion long entrusted to the sentencing judge. See Mistretta, 488 U.S. at 396, 109 S.Ct. 647 (stating that the Guidelines “do no more than fetter the discretion of sentencing judges to do what they have done for generations—impose sentences within the broad limits established by Congress”). Because Apprendi does not apply to a judge’s exercise of sentencing discretion within a statutory range, the current practice of judicial factfinding under the Guidelines is not subject to the Apprendi requirements—at least so long as that factfinding does not enhance a defendant’s sentence beyond the maximum term specified in the substantive statute.

The Sentencing Guidelines do not create crimes. They merely guide the discretion of district courts in determining sentences within a legislatively—determined range, and this discretion has been entrusted to the federal courts “[f]rom the beginning of the Republic.” Apprendi, 120 S.Ct. at 2358 n. 9 (quoting K. Stith and J. Cabranes, Fear of Judging: Sentencing Guidelines in the Federal Courts 9 (1998) (internal quotation marks omitted)). Accordingly, we agree with the conclusions reached by the Fifth, Seventh, and Eleventh circuits that the relevant “maximum” under Apprendi is found on the face of the statute rather than in the Sentencing Guidelines. See Nealy, 232 F.3d at 829 n. 3; Doggett, 230 F.3d at 165; Talbott, 226 F.3d at 869 (7th Cir.); cf. Apprendi, 120 S.Ct. at 2366 n. 21 (citing, with seeming approval, language from Edwards v. United States, 523 U.S. 511, 515, 118 S.Ct. 1475, 140 L.Ed.2d 703 (1998), indicating *202 that the statute itself provides the relevant maximum). Because the sentencing enhancements at issue in this case did not extend Kinter’s sentence beyond the máxi-mums prescribed for his offenses by the substantive provisions of the United States Code, the government was not required to submit to a jury and prove beyond a reasonable doubt the facts relevant to those enhancements.

For this reason and the others given above, the judgment of the district court is

AFFIRMED.