Motilall Sudeen was convicted of wire fraud, travel fraud, money laundering and conspiracy offenses and sentenced to 220 months’ imprisonment. He appeals, contending that the district court committed reversible error (1) in severing his trial from that of his co-defendant; (2) in'admitting evidence of an uncharged investment scheme and statements he alleges to be hearsay; (3) in the application of the sentencing guidelines; and (4) by finding certain facts in contravention of United States v. Booker, 543 U.S. 220, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005). We find Booker error with respect to the district court’s use of the 2002, rather than the 2000, guidelines. As to all other claims pertaining to conviction or punishment, we find no reversible error. We therefore affirm the conviction and vacate the sentence and remand for resentencing.
I.
Sudeen sought to build a large urea processing plant in Poplarville, Mississippi, where he ultimately consolidated all his business dealings.1 He formed a corporation, MS Carbamate (“Carbamate”), and acquired land on which to build the plant. He contends that he was “financing” the plant using a variety of investment devices.
The plant was never built, because of alleged regulatory difficulties. As a result, Sudeen breached contracts with investors.
The investment programs Sudeen alleged he was using to finance the plant were, in actuality, constituent frauds in a fairly elaborate Ponzi scheme. Sudeen and his co-conspirators represented to potential investors that their money would be placed in “high yield investment programs” that would generate profits for them at greater-than-market rates of return.2
Sudeen told investors that their principal would remain safely in banks and would be exposed to little or no risk; that the high yield programs involved marquee banks, including the World Bank and the IMF; and that the trading programs were monitored by the federal government. Su-deen periodically mollified investors by returning “dividends” from the programs;3 by encouraging investors to roll over their investments instead of seeking immediate returns; and by reassuring investors that their money had been safely invested and that they would be paid soon.4 Sudeen and his co-conspirators used the funds to make “lulling” payments to encourage fur*387ther investment in the “programs” and for Sudeen’s personal and business expenses.5
Sudeen and Freeman continued to maintain the appearance of safety by issuing investors bogus “Private Placement Agreements” and “Joint Venture Agreements.” They also told investors that Sudeen’s personal wealth guaranteed their investment. Investors were told to purchase Certificates of Deposit from various banks to allow Sudeen to use the credit for loans, the proceeds of which would also be invested. When a given investor demanded proceeds, Sudeen and Freeman would claim that the investor was ineligible because he had failed to comply with fictitious requirements, that the profits were “tied up” by the federal government, or that the returns could not be liquidated from overseas assets.
1. Sudeen was a principal in another Ponzi scheme involving insulin (the “insulin scheme”), for which he and Freeman were not indicted. Sudeen co-mingled funds involved in the two schemes, making lulling payments using resources of one to the other. On this issue the government cites to the record extensively, whereas Sudeen does nothing more than advance speculation.
In February 2002 Sudeen and Freeman were indicted on one count of conspiracy under 18 U.S.C. § 371, fourteen counts of wire fraud under 18 U.S.C. § 1343, two counts of travel fraud under 18 U.S.C. § 2314, and twenty-one counts of money laundering under 18 U.S.C. § 1957. Trial was originally set for May 2002. The district court granted several of the defendants’ motions for continuance.
In December 2002, Sudeen’s cardiologist informed the court that Sudeen was healthy enough to go to trial in January. In January 2003 Sudeen and Freeman again moved for continuance because Freeman’s attorney had a conflict with a case previously scheduled for trial in North Carolina. The government opposed the motion and moved to sever the trials and proceed with Sudeen’s. The district court granted Freeman a continuance, denied a continuance to Sudeen, and granted the government’s motion to sever, thus forcing Sudeen to proceed to trial on January 13.
The jury found Sudeen guilty on thirty-eight of the counts. Applying the 2002 sentencing guidelines, the court calculated an offense level of 37, which produced a guidelines range of 210-262 months’ imprisonment. The court sentenced Sudeen to 220 months.
II.
Sudeen argues that because his and Freeman’s counsel had prepared their defenses jointly, Sudeen was prejudiced by a severance. We disagree.
A.
We review a grant or denial of severance for abuse of discretion. See United States v. Ramirez, 954 F.2d 1035, 1037-38 (5th Cir.1992). A severance is reversible only on a showing of specific compelling prejudice. See United States v. Barnett, 1097 F.3d 138, 144 (5th Cir.1999).6
*388B.
Although it is generally true that “defendants who are indicted together should be tried together,” see United States v. Piaget, 915 F.2d 138, 142 (5th Cir.1990), that generalization says nothing of the legal circumstances that justify deviating from it. Sudeen cites no authority for the standard he advances as the criteria for such deviation: “[Wjhether the Government articulated a specific risk that could be averted only through severance, and it must further query whether alternative means less prejudicial to Sudeen existed to remedy that risk.” Sudeen seems to be fabricating that standard entirely.
Sudeen’s primary argument attempting to show such prejudice involves the short period in which his attorney had to prepare for trial after the severance. Sudeen does not allege prejudice with any specificity. His counsel was ready for trial anyway, because Sudeen would have been deemed the principal had the two defendants been tried jointly, and Sudeen’s and Freeman’s attorneys had been working on a joint defense for many months. Moreover, Sudeen’s attorney was, by all accounts, well prepared for trial.7
III.
Sudeen actually orchestrated another fraudulent venture involving insulin contracts. He does not deny that the funds from those investing in the insulin contracts were used as lulling payments to investors in the fertilizer plant and vice versa.8 The district court therefore found that the funds from the two ventures were commingled and that the evidence regarding the insulin venture was “intrinsic.”
Sudeen argues that the insulin evidence was extrinsic and should have been excluded under Federal Rules of Evidence 403 and 404(b). First, we must decide whether the district court reversibly erred in finding the insulin evidence intrinsic. Second, if we that the evidence was extrinsic, we must determine whether the court erred by failing to exclude the evidence under other evidence rules. We terminate our inquiry at the first stage by finding that the court did not abuse its discretion in treating the evidence as intrinsic.
A.
We review admission of evidence for abuse of discretion. See United States v. Hicks, 389 F.3d 514, 522 (5th Cir.2004). That Sudeen commingled funds between the two sub-schemes is not a contested fact. Intrinsic evidence is generally admissible to allow the jury to “evaluate all the circumstances under which the defendant acted.”9 Evidence is considered intrinsic if it is “inexorably intertwined” with *389evidence used to prove the crime charged. See Navarro, 169 F.3d at 233. Where evidence is intrinsic, it qualifies without reference to rule 404(b), which states generally that “evidence of other crimes, wrongs, or acts is not admissible to prove the character of a person in order to show action in conformity therewith.” See id.10 Citing superior authority and referencing the record far better than Sudeen does, the government overwhelmingly establishes the two propositions necessary for it to prevail on this argument: (1) Funds for the insulin scheme were commingled with funds from the fertilizer scheme, and initial investments in each were used to make lulling payments to investors in the other;11 and (2) such commingling of funds qualifies the insulin scheme as intrinsic evidence.12
B.
Rule 403 states that: “Although relevant, evidence may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury, or by considerations of undue delay, waste of time, or needless presentation of cumulative evidence.”13 The insulin evidence must pass the rule 403 bar without respect to whether the district court admitted it as intrinsic evidence.
The district court noted that if it considered evidence of the insulin scheme to be intrinsic, it would not exclude that material on the ground that it is prejudicial under rule 403.14 This holding is consistent with language in our circuit stating that rule 403 should generally not be used to exclude intrinsic evidence, because intrinsic inculpatory evidence is by its very nature prejudicial.15 The insulin evidence proves both the source of the lulling payments made to Poplarville investors and the destination of funds paid in by those investors. It is highly probative of the existence of a Ponzi scheme, and it is prejudicial only to the extent that it establishes elements of the offense.16
*390IV.
Sudeen further contends that the district court abused its discretion in admitting the out-of-court statements of Earl Gamble and Walter Lauren under Federal Rule of Evidence 801(d)(2)(D) and/or 801(d)(2)(E). Rule 801(d)(2)(D) deems non-hearsay a statement offered against a party made by “the party’s agent or servant concerning a matter within the scope of the agency or employment, made during the existence of the relationship.”
Similarly, rule 801(d)(2)(E) provides that a statement is not hearsay if it is offered against a party and is “a statement by a co-conspirator of a party during the course and in furtherance of the conspiracy. The contents of the statement shall be considered but are not alone sufficient to establish the declarant’s authority!/]”
Sudeen contends that the district court’s finding that Gamble and Lauren were agents or co-conspirators was not supported by substantial evidence.17 He correctly notes that the statements themselves are not sufficient evidence of the relationship to render them admissible. He then asserts that the government offered no additional proof of Gamble’s and Lauren’s roles as co-conspirators or agents of Sudeen. He characterizes them as owners of investment companies who dealt at arm’s length with Sudeen and the potential investors.
A.
We review for abuse of discretion the admission of out-of-court statements under rule 801(d)(2)(D) and (E).18 There was corroborating material (evidence other than the statements themselves) demonstrating that Lauren’s and Gamble’s statements were admissible under either subsection.
B.
Lauren received significant commissions for recruiting investors. At least one investor described him as an “associate” of Sudeen’s. Lauren received reports from investors who were recruiting other investors in the high-yield trading program. An employee of MS Carbamate testified that Lauren worked for Sudeen. In light of this corroborating testimony, the district court did not abuse its discretion in admitting Lauren’s statements.
Some of the investors testified that they thought Gamble and Sudeen were business partners. Gamble introduced several potential investors in the high-yield trading program to Sudeen. Another investor described Gamble as an “associate” of Su-deen’s.
Sudeen told one potential investor that Gamble would be her “agent” if she decided to participate in the high-yield program. That same investor wrote a check to Gam*391ble for investment in the program, and that check was later deposited in one of Sudeen’s business accounts. Sudeen informed that investor that Gamble would receive a commission of one percent of her investment principal. When investors inquired as to why they were not receiving the promised returns, Gamble provided explanations for the delays and reassurances that their money was safe. The district court did not abuse its discretion in admitting Gamble’ testimony.
V.
Sudeen contends the court improperly enhanced his sentence for abuse of a position of private trust, arguing that he was not a legitimate trader or investment broker. Section 3B1.3 of the United States Sentencing Guidelines provides for a two-level increase in offense level if “the defendant abused a position of public or private trust, ... in a manner that significantly facilitated the commission or concealment of the offense.”
A.
Even after Booker, we review a district court’s interpretation and application of the guidelines de novo. See United States v. Villegas, 404 F.3d 355, 359 (5th Cir.2005) (per curiam). We thus proceed to review the application of guideline § 3B1.3 here without deference to the district court’s interpretation.19
B.
Sudeen relies on United States v. Eche-varria, 33 F.3d 175 (2d Cir.1994), for the proposition that the district court may enhance a sentence pursuant to § 3B1.3 only where the defendant legitimately occupied a position of trust. Sudeen makes no arguments beyond analogy to that case.
A subsequent Second Circuit case, United States v. Hussey, 254 F.3d 428, 433 (2d Cir.2001), repudiates the reasoning of (but stops short of overruling) Echevarria. That court noted that two other courts of appeals have rejected or criticized Echevarria;20 it also explains that, by adding an application note,21 the United States *392Sentencing Commission rejected Echevar-ria’s, reasoning in November 1998. See id. at 433 n. 3. Most importantly, Reeves, 255 F.3d at 212, affirmed an abuse-of-trust enhancement where the defendants had posed as financial planners and advised their clients to invest in a company owned by a co-defendant. There is no meaningful way to distinguish the facts in that ease from the ones here. Therefore, the district court did not err in its application of § 3B1.3.
VI.
Sudeen claims the district court committed reversible error under Booker, 125 S.Ct. at 738. Specifically, he reasons that the court improperly found facts at sentencing relating to (1) the amount of loss; (2) the number of victims; (3) the use of specific means; and (4) the effective date of the conspiracy’s end. Sudeen’s brief styles the last of these items — the only Booker objection we ultimately believe has merit — as a hybrid of a Booker and an ex post facto violation. It is evident from oral argument and from briefing that neither the government nor Sudeen is entirely sure how the two sets of precedent interact. We conclude that the district court’s application of the 2002 edition of the guidelines — as opposed to the 2000 edition of the same: — is reversible error for the reasons set forth below and in the companion opinion issued contemporaneously herewith, United States v. Freeman, 434 F.3d 369, No. 04-30037, 2005 WL 3525612 (5th Cir. Dec. 23, 2005).
A.
1.
First, we briefly dispose of Su-deen’s Booker claims involving amount of loss, number of victims, and use of specific means. He concedes that he did not preserve a Sixth Amendment objection to these issues at sentencing, and we review unpreserved Booker arguments for plain error. See United States v. Mares, 402 F.3d 511, 520 (5th Cir.), cert. denied, — U.S. —, 126 S.Ct. 43, 163 L.Ed.2d 76 (2005). The third prong of the plain error analysis requires that, to prove reversible error, Sudeen “demonstrate a probability ‘sufficient to undermine confidence in the outcome.’ ” Id. at 521 (quoting United States v. Dominguez Benitez, 542 U.S. 74, 124, 124 S.Ct. 2333, 159 L.Ed.2d 157 (2004)). Under Mares, “[t]he pertinent question is whether [Sudeen] demonstrated that the sentencing judge — sentencing under an advisory scheme rather than a mandatory one' — would have reached a significantly different result.” Id. Even if all three of these preceding conditions are satisfied, we may exercise our discretion to notice a forfeited error only if that mistake seriously affects “the fairness, integrity, or public reputation of judicial proceedings.” United States v. Cotton, 535 U.S. 625, 631, 122 S.Ct. 1781, 152 L.Ed.2d 860 (2002).
2.
Sudeen attempts to carry his burden by doing nothing more than identifying, for each enhancement, the error and pointing *393out that the district court sentenced him at the low end of the guidelines range. Under the computed range, the district court could have sentenced him to a term of 210 to 262 months; it in fact imposed 220 months. Therefore, Sudeen contends, the court, if it had known of its discretion to do so, likely would have imposed a lesser sentence.
This logic is unpersuasive. A sentence at the low end of the range does not show that the error “must have affected the outcome of the district court proceedings.” United States v. Olano, 507 U.S. 725, 734, 113 S.Ct. 1770, 123 L.Ed.2d 508 (1993). Moreover, in denying Sudeen’s request to sentence him to the minimum guidelines sentence, the district court indicated that it considered Sudeen’s offenses to be very serious.22 Sudeen cannot show a reasonable probability that the court would have imposed a lower sentence under an advisory guidelines regime, so he cannot prevail under a plain error standard of review.
B.
Sudeen asserts that the district court improperly used the 2002 edition of the guidelines instead of the 2000 edition.23 Although we discuss this issue as part of our “Booker error” analysis, the objection actually subdivides into two distinct inquiries, only one of which Booker technically controls: (1) whether the use of the 2002 edition constitutes an independent Booker error and (2) whether the use of the 2002 edition violated the Ex Post Facto Clause (a claim we analyze the same as we would have before Booker).
1.
Sudeen asserts that, under Booker, the district court cannot constitutionally have made factual determinations regarding the end-date of the conspiracy. Sudeen neither admitted that end-date, nor was it found by a jury beyond a reasonable doubt. Nevertheless, that finding plainly increased his sentencing range.
a.
The government argues that Su-deen did not preserve his error under Booker. At sentencing, however, Sudeen’s attorney made the following remarks:
And we would suggest that where like Apprendi the guidelines have [sic], I don’t want to say legal effect, but clearly the practical effect of increasing the statutory maximum for the continuing offense. By having that offense drive a set of numbers that are greater than what you can be sentenced to it [sic] under that offense and which apply to the whole case, it[’]s appropriate to begin looking at is that an issue the jury has to decide.
As we have held in the companion case, Freeman, this language is enough for us to find a mistake preserved for purposes of *394choosing between plain and harmless error analysis.24
b.
There is no dispute that if the conspiracy was proven to extend to a date on or after November 1, 2001, a set of guidelines later than the 2000 version would apply (in an advisory capacity, of course, in the wake of Booker). The indictment states that “[b]eginning in or about March 1997, and continuing to the present [meaning February 28, 2002], ... the defendants ... did knowingly and willfully ... conspire ....” The indictment charges the overt acts under the conspiracy with specificity; the latest such charged act is Sudeen’s promise to pay a particular investor additional money, an act alleged to have occurred “[i]n or about August 2001.”
In its brief on appeal, the government, in an effort to avoid use of the 2000 guidelines, points to proof of several acts occurring on or after November 1, 2001.25 The government also accurately points to the fact that at sentencing, the district court made a finding that the conspiracy continued past November 1, 2001. The flaw in the government’s position, however, is that the procedure it correctly recounts is the very essence of a Booker violation.
The jury was charged in relevant part as follows:
It is not essential that the Government proved that the conspiracy started and ended on those specific dates [i.e., March 1997 through February 2002], Indeed it is sufficient if you find that in fact a conspiracy was formed and that it existed for some time within the period set forth in the Indictment and that at *395least one overt act was committed to further the conspiracy within that period of time.
Accordingly, the fact of conviction does not necessarily establish that the jury found the existence of any overt acts on or after November 1, 2001. It was only the district court, and not the jury, that found that the conspiracy continued beyond the trigger date for the post-2000 guidelines.26 This is specifically what Booker prohibits: The
actual sentence ... was ... longer than the Guidelines range supported by the jury verdict alone. To reach this sentence, the judge found facts beyond those found by the jury .... “[T]he jury’s verdict alone does not authorize the sentence. The judge acquires that authority only upon finding some additional fact” [quoting Blakely v. Washington, 542 U.S. 296, 305, 124 S.Ct. 2531, 159 L.Ed.2d 403 (2004)].
Booker, 125 S.Ct. at 751.
In short, the indictment charged no specific acts after August 2001, and the jury was told it could find defendants guilty without finding any overt acts on or after November 1, 2001. There is no basis on which, in the wake of Booker, we can infer that any such acts indeed occurred, i.e., that the jury, if asked, would have found them beyond a reasonable doubt.27
c.
Because Sudeen’s sentence was infected with Booker error, and he properly preserved his objection, we must vacate the sentence and remand for resentencing unless we determine that the error was harmless under Federal Rule of Civil Procedure 52(a). See Mares, 402 F.3d at 520 n. 9. “Harmless error is ‘any defect, irregularity, or variance that does not affect substantial rights of the defendant,’ and ‘arises when the mistake fails to prejudice the defendant.’ ”28 Under this standard the government must demonstrate, beyond a reasonable doubt, that the error did not contribute to the sentence that the defendant received. See United States v. Olano, 507 U.S. 725, 734, 113 S.Ct. 1770, 123 L.Ed.2d 508 (1993).
The government does not meet this burden. It points to nothing that would show beyond a reasonable doubt that the court would have imposed the same sentence under an advisory guidelines regime. See Akpan, 407 F.3d at 377.
*3962.
We do not reach Sudeen’s ex post facto claim. Under Akpan, id. at 360 n. 2, we have the authority to leave to the district court the discretion to consider this argument as long as we have already determined there was a reversible Booker violation.
In summary, the judgment of conviction is AFFIRMED. The judgment of sentence is VACATED and REMANDED for resentencing.
. Sudeen characterizes his efforts to acquire the necessary funds for the plant as "legitimate arrangement[s] made with sophisticated investors.” He states that he signed agreements with investors to use their money in "projects” that centered on the plant.
. Sudeen sought to finance Carbamate with the high-yield investment program and what he described to investors as a "private placement secured trading programs.” Sudeen told investors that these programs would yield 20-50% per month and that he would return their principal on the expiration of the investment terms.
. Sudeen also lied about the composition of the banks taking part in the financing. Su-deen and an accomplice, Jerry Freeman, sought to make the trading programs appear legitimate and safe.
. More than fifty people participated in the high-yield trading program, and the victims spent more than $ 17 million.
. Sudeen used the funds, in addition to making lulling payments, to pay Freeman's salary, to remodel Sudeen’s house, support Sudeen’s wife and children, purchase property including luxury goods, and to make credit card payments.
. Under our caselaw, to demonstrate an abuse of discretion as it relates to severance, Sudeen must show "specific and compelling prejudice.” United States v. Mitchell, 777 F.2d 248, 260 (5th Cir.1985). The Mitchell standard requires that we determine abuse of discretion by reference to whether the sever*388ance caused Sudeen clear and compelling prejudice, not whether the district court adequately identified evidence in the record showing that Sudeen would have been prejudiced by joinder.
.In fact, the greatest possibility for prejudice appears to have been that to the detriment of the government, if the district court had allowed defendants to proceed jointly. The trial had been pushed back four times already, and Sudeen's health problems posed a significant potential for further delay.
. In oral argument on appeal, Sudeen's counsel expressly admitted to the co-mingling of funds between the two schemes.
. United States v. Navarro, 169 F.3d 228, 233 (5th Cir.1999); see also United States v. Williams, 900 F.2d 823, 825 (5th Cir.1990) (stating that evidence qualifies as "intrinsic'' "when the evidence of the other act and the evidence of the other crime charged are ‘inextricably intertwined’ or both acts are part of a 'single criminal episode’ or other acts were 'necessary preliminaries' to the crime charged”) (internal citations omitted).
. See 1 Steven A. Saltzburg, Michael M. Martin & Daniel J. Capra, Federal Rules of Evidence Manual § 404.02[11] (LexisNexis 8th ed.2002).
. Even if each direction were not specifically proven (i.e., the state specifically identified only a particular transaction where money flowed from one scheme to another, and not vice versa), the co-mingling of funds obviously justifies the’ inference that money actually flowed both ways.
. The district court alternatively ruled, in an in limine hearing, that the insulin evidence was admissible under rule 404(b), which states that prior (or other) wrongs or acts are admissible as "proof of motive, opportunity, intent, preparation, plan, knowledge, identity, or absence of mistake or accident.” See United States v. Dula, 989 F.2d 772, 777 (5th Cir.1993) ("Evidence of an uncharged offense arising out of the same transaction or series of transactions as the charged offense is not an 'extrinsic' offense within the meaning of [rule] 404(b) and is therefore not barred by this rule.”) We do not reach this issue.
. Fed.R.Evid. 403; see also 1 Saltzburg et al., supra, § 403.02[16], at 403-37 (stating that rule 403 "is one of exclusion of otherwise admissible evidence”).
. Specifically, the court stated: "It's not confusing if it's the same, if it's advancing one, it’s not misleading, it’s not prejudicial any more than the indictment[,] charge or something [that] is prejudicial in the sense that it’s offensive but it’s part of the crime.”
. See United States v. Powers, 168 F.3d 741, 749 (5th Cir.1999) ("all probative evidence is by its very nature prejudicial”) (internal citations omitted); United States v. Leahy, 82 F.3d 624, 637 (5th Cir.1996) (stating that rule 403 should be used sparingly and only where the prejudicial effect substantially outweighs the evidence’s probative value).
. Sudeen contends that the court inappropriately failed to give a limiting instruction as to the purposes of the insulin evidence. First, we see no reason why such an instruction is *390necessary if we deem the evidence intrinsic. Second, the record indicates that the jury instructions quite candidly addressed the limited character of "similar acts.”'
. There is some discrepancy between the briefs as to which of the two subsections applies to Gamble and which to Lauren. The government asserts that Gamble's statements were admitted under the agency exception of subsection (d)(2)(D) and that Lauren's were admitted under either (D) or (E). Whether the relationship among these three men is one of agency or co-conspiracy, however, need not be decided for us to rule on this issue. For a general discussion of agent versus co-conspirator admissions, see 4 Saltzburg et al., supra, § 801.02[6][f]-[g].
. See United States v. Solis, 299 F.3d 420, 443 (5th Cir.2002) (stating standard with respect to Fed.R.Evid 801(d)(2)(E)); United States v. Cent. Gulf Lines, Inc., 974 F.2d 621 (5th Cir.1992) (same with respect to Fed.R.Evid. 801(d)(2)(D)).
.The government cites United States v. Reeves, 255 F.3d 208, 212 (5th Cir.2001), for the proposition that the panel "review[s] the application of the guideline to the facts for clear error.” This reasoning should be distinguished from our more recent holding in Vil-legas. The apparent discrepancy is nonetheless something we feel compelled to address, even though the standard-of-review issue makes little difference in the ultimate outcome here, because Sudeen's challenge would fail even under our de novo scrutiny.
We assess this issue much as did the Second Circuit:
Section 3B1.3 of the Sentencing Guidelines provides for a two-level enhancement if the defendant "abused a position of public or private trust, or used a special skill, in a manner that significantly facilitated the commission or concealment of the offense.” U.S.S.G. § 3B1.3 (1998). Whether a defendant occupies a "position of trust” within the meaning of this provision is viewed from the perspective of the victim, and is a question of law, which we review de novo. See United States v. Wright, 160 F.3d 905, 910 (2d Cir.1998). Whether a defendant abused a position of trust in a manner that "significantly facilitated the commission or concealment of the offense” is a question of fact, which we review for clear error. See United States v. Hirsch, 239 F.3d 221, 227 (2d Cir.2001).
United States v. Hussey, 254 F.3d 428, 431 (2d Cir.2001). We endorse this distinction.
. See Hussey, 254 F.3d at 432 n. 3; see, e.g., United States v. Gill, 99 F.3d 484, 489 (1st Cir.1996); United States v. Barnes, 125 F.3d 1287, 1292 (9th Cir.1997).
. Application note 2 to U.S.S.G. § 3B1.3 now provides:
Th[e] adjustment ... also applies in a case in which the defendant provides sufficient indicia to the victim that the defendant le*392gitimately holds a position of private or public trust when, in fact, the defendant does not. For example, the adjustment applies in the case of a defendant who (A) perpetrates a financial fraud by leading an investor to believe that the defendant is a legitimate investment broker; or (B) perpetrates a fraud by representing falsely to a patient or employer that the defendant is a licensed physician. In making the misrepresentation, the defendant assumes a position of trust, relative to the victim, that provides the defendant with the same opportunity to commit a difficult-to-detect crime that the defendant would have if the position were held legitimately.
. Specifically, the district court stated:
I’ve heard countless stories ... that people lost their whole life savings or their future. And they entrusted it to you. They gave you their life savings.
It wasn't for many of them given to you for the urea plant, it was given to you to invest in the high interest scheme that you told them about. And there was no such program ... You just used their money, used Peter’s money to pay Paul ... And these victims didn't do anything wrong and you just abused them.
. The government posits that the court actually used the 2003 version, which was in effect on the date of sentencing. The difference is irrelevant, because § 2S1.1 is the same in both. Use of the 2000 version, as urged, would have yielded a substantially lower range.
. Not only does this language resemble that which we found sufficient for preserving error in Freeman, but it is more specific than that which we held to satisfy the error preservation requirements in United States v. Akpan, 407 F.3d 360 (5th Cir.2005) ("Although [the defendant] never explicitly mentioned the Sixth Amendment, Apprendi, or Blakely until his Rule 28(j) letter, we are satisfied that his objection adequately apprised the district court that he was raising a Sixth Amendment objection ....'').
. According to the government's brief,
Freeman continued to receive biweekly salary payments of $1,450 through January 17, 2002, and $9,000 was wired to co-conspirator Walter Lauren at an account in Switzerland as late as February 27, 2002. Mortgage payments on the Poplarville property using funds from investors were made until January 15, 2002. In addition, lulling payments of $10,000, $2,000, and $2,500 were made to investors Frank Gunn, Kenneth Breaux and Sheran Frickey, respectively, on December 12, 2001. In November 2001, Sudeen promised that he would give a bank guarantee to investor Mattias Baumeler. In February 2002 Baumeler met with Sudeen in Switzerland, and Su-deen promised that he would remit all overdue profits within two weeks.
Moreover, Alice Celestin testified at trial and sentencing that every 120 days she and her husband "rolled over” their principal and purported interest payments into a new contract. When she met with Freeman on July 1, 2001, and signed a fifth contract, she advised him that she was going to need $54,000 back in November. When she didn't receive the money, she telephoned Freeman frequently. In December 2001 Freeman called her and said that he had both good news and bad news: she was getting money, but it was only $10,000. They met the next day and he gave her three separate checks totaling $10,000. She testified that at that time she still believed that she had funds invested in insulin. She continued calling Freeman and during their last conversation in February 2002 he said that her funds had "two more banks to clear.” Freeman’s misrepresentations plainly lulled Mrs. Celestin into the continued belief that her funds were safely invested and that the promises made at the time of her initial investment would be fulfilled.
(Record citations and footnote omitted.)
. The government relies entirely on the fact that the district court found that the conspiracy continued past November 1, 2001. At no point does the government even attempt to argue that we may infer that the jury made any such finding, much less that it did so beyond a reasonable doubt.
. It is true that under the law of this circuit, "[o]rdinarily, a defendant is presumed to continue involvement in a conspiracy unless that defendant makes a 'substantia! affirmative showing of withdrawal, abandonment, or defeat of the conspiratorial purpose.” United States v. Puig-Infante, 19 F.3d 929, 945 (5th Cir.1994) (citation and internal quotation marks omitted). Accord United States v. Schorovsky, 202 F.3d 727, 729 (5th Cir.2000). These authorities, although sound, address a situation entirely different from the one presented here; they involve multi-person conspiracies in which the defendant claims he tried to withdraw from the conspiracy that was continued by his co-conspirators. Here, there were no acts found by a jury after the trigger date, so there is no jury-found conspiracy at all that existed on or after that date.
In other words, those authorities demonstrate that the absence of an overt withdrawal can extend the operative dates of an alleged withdrawing defendant’s vicarious liability to the end of the conspiracy. Those authorities do not suggest, however, that the absence of an overt withdrawal extends the length of the conspiracy itself.
. Akpan, 407 F.3d at 376-77 (quoting rule 52(a); United States v. Munoz, 150 F.3d 401, 413 (5th Cir.1998)).