PER CURIAM:

This appeal principally raises two issues: (1) whether the misrepresentations underlying these convictions were not material because no reasonable financial professional would have believed them, and (2) whether the sentences imposed on appellants are procedurally unreasonable. We affirm the District Court's denial of appellants' motions for acquittal, but vacate appellants' sentences and remand for resentencings.1

I. BACKGROUND

This appeal arises out of a conspiracy to defraud a non-existent investor of three billion dollars. In the spring of 2006, defendants John Juncal, James Campbell, Rodney Sampson, and Emerson Corsey were arrested and charged with one count of conspiracy to commit mail fraud and wire fraud in violation of 18 U.S.C. §§ 1341, 1343, and 1349. Over the preceding four months, the defendants had attempted to lure a broker, Thomas Re, into procuring financing for an imaginary Siberian oil pipeline. The two sides exchanged information about possible partners, and the best structure for a loan. But in reality each side duped the other — Juncal and his colleagues had no plans to build a pipeline across the Russian tundra, and Re never represented a hedge fund interested in lending billions to the defendants to do so.

A. The Scheme

Re was the CEO of Universal Lending Group ("ULG"), a small brokerage firm based in Garden City, New York. Re acted as a "salesperson and networker" for the company; he solicited companies looking for financing and his associate, Joseph Bianco, then matched projects with a hedge fund or bank to act as lender. Re was also an informant for the FBI: Before he worked at ULG, Re sold vending machine routes to snack-food companies. On occasion, Re sold distributors a fake route — the distributors would arrive at a location and find no machine — and the FBI eventually caught up with him. In the hopes of receiving a lighter sentence, Re agreed to turn over information to the FBI about others participating in the vending machine scam.

Re encountered the defendants seven months after he began working with the FBI. Re received a call from an acquaintance, Charles Frazier, who told him to look out for a financier named Emerson Corsey — Juncal, Campbell, and Sampson's co-defendant. Corsey called a few days later and explained that he was the Chief Operating Officer of Magnolia International Bank and Trust ("MIBT"), a bank that Corsey described as holding assets of sovereign wealth funds. Corsey said that MIBT wanted to borrow three billion dollars to build a pipeline in Siberia and could offer five billion dollars in U.S. Treasury notes ("T-notes") as collateral for the loan. Re said he would think about it.

Re told his associates about the proposal and, after a brief Google search, they told him the deal "smelled." J. App'x at 516. Re then reported the potential fraud to his FBI handler, who instructed Re to record his conversations with Corsey.

Over the next few months, Re recorded the defendants as they baited him with an escalating series of lies: Corsey explained that MIBT was the central bank for scores of Native American governments, including the Yamasee Indian tribe, a nation with trillions of dollars in assets. Corsey then said that the bank currently represented John Juncal, a "counsel extraordinaire" for the Republic of Buryatia who had been appointed "Vice Premier of Record under the Edict of the Ukase" by the Buryatian government, and the official charged with finding backers for the pipeline. When Re expressed skepticism, Corsey sent him an email with copies of "bond indentures." J. App'x at 1021-22. The indentures bore the seal of the Vice-President of Buryatia, and stated that they were valued at five billion U.S. dollars and secured by T-Notes. The certificates were assigned to "the Great Siberian Pipeline Corporation," which was "a Republic of Wyoming Corporation." Id. The email also contained proof that Juncal controlled the T-notes: Corsey included another certificate signed by Juncal that listed CUSIP numbers for the T-notes (as an expert would later testify, CUSIP numbers are alphanumeric code created by the U.S. Treasury to identify individual bonds).

Prompted by his handler to record statements by more of the defendants, Re asked to speak with Sampson, the purported Chief Financial Officer of MIBT. Sampson laid out the math behind the deal: MIBT's client, John Juncal, would assign the T-notes to Re's client. The T-notes would yield a minimum of 3.9% interest per year and generate over fourteen billion dollars in profit over the course of five years. When Re asked for additional evidence of the T-notes' authenticity, Sampson sent Re an email from an AOL account for MIBT, rather than from MIBT's own domain name. Attached to the email were unformatted files containing copies of the T-notes; Re had to use Word Pad to open those files. Sampson then e-mailed Re a "Letter of Acceptance" that re-explained the financing. The letter stated that, in exchange for a loan of three billion dollars, Re's client would receive T-notes that would yield a 171% return on investment.

Re next pressed Corsey and Sampson to show him physical evidence of the T-notes' existence. And at that point, Corsey put Re in touch with James Campbell, an associate of Juncal's. Campbell explained that Juncal had hidden the bonds in Austria in order to prevent enemies from "basically steal[ing] these from him and bump[ing] him out of the corporation." J. App'x at 1109. Re pretended to buy the excuse, but later asked if he could speak with Juncal directly. Juncal agreed, and sought to assuage Re's fears by comparing his role to the historic secret mission of Christopher Columbus.

During his testimony at trial, Re admitted that much of the talk struck him as ridiculous, and that at some point he stopped believing the deal was real. But because of his relationship with the FBI, he continued working with the quartet of pipeline pipe-dreamers. In March, it came time for the deal to close, and Re arranged to meet Campbell and Sampson in New York to sign documents memorializing their agreement. On the day of the closing, the FBI arrested all four defendants; Campbell and Sampson were arrested in New York, Corsey was arrested in Atlanta, and Juncal was arrested at his home, a motel in Whitefish, Montana.

At trial, the government relied almost exclusively on Re's testimony. No defendant took the stand, and the defense attorneys focused their defense on attacking Re's testimony. After a short deliberation, the jury returned a verdict of guilty for all four defendants. The appellants' motions for judgment of acquittal were denied and the cases proceeded to sentencing.

B. The Sentences

In each case, but for the statutory maximum term of imprisonment, the appellants' Sentencing Guidelines recommended range would have been a life sentence, because the offense level was driven off the Guidelines' Sentencing Table by an intended loss amount of three billion dollars. Though each appellant argued that the intended loss grossly overestimated the seriousness of the offense, the District Court disagreed. In March of 2010, the District Court sentenced all four defendants to twenty years' incarceration, the statutory maximum and the Guidelines recommended sentence.

1. Campbell's Sentencing

The District Court gave a very brief explanation of the reasons for Campbell's sentence, and largely adopted the logic and recommendations in his Pre-Sentence Investigation Report ("PSR"). Under the Guidelines applicable in 2010, the PSR calculated Campbell's total offense level to be 47. Under U.S.S.G. § 2B1.1, his base offense level was 7. Thirty points were added because Campbell had intended to defraud an investor of over $400 million. The PSR then added points for four additional enhancements: Campbell received four points because he oversaw five or more co-conspirators, U.S.S.G. § 3B1.1; two levels because he offered fake CUSIPs as an "authentication feature," U.S.S.G. § 2B1.1(b)(10)(A)(ii); two levels because Campbell misrepresented himself as a representative of a government or agency, U.S.S.G. § 2B1.1(b)(8)(A); and two levels because the scheme involved sophisticated means, U.S.S.G. § 2B1.1(b)(9)(C).

At the sentencing hearing, the government withdrew its request for a role enhancement pursuant to U.S.S.G. § 3B1.1. The District Court asked if that change affected the guideline range:

J. App'x at 1373-74. The District Court then inquired why the Guidelines calculation was so far above the statutory maximum. The prosecutor answered:

J. App'x at 1375. The defense then presented its case against enhancements and for downward departures. The District Court rejected each one, explaining its general approach by referencing other recent financial fraud cases:

J. App'x at 1388-89. The District Court then explained the reason for imposing a 20-year sentence:

J. App'x at 1393.

2. Sampson's Sentencing

Sampson's PSR contained an identical offense level score. His attorney objected to all five enhancements, but the District Court only accepted Sampson's arguments on the role enhancement (the government again withdrew its request for the enhancement). As with Campbell, the Court also dismissed the need to recalculate the Guidelines range, explaining that "the maximum is 20 years so it really has no effect."

Sampson was then asked if he wished to make a statement. Sampson said yes, and launched into an impassioned account of the trials of the Yamasee tribe. He stated that the tribe had waged a centuries-long battle against Western encroachment, only signing a treaty with Georgia in 2003. He referred to the people of Buryatia as a "sovereign aboriginal indigenous nation." He then explained the motive of the fraud scheme as "doing something to help the people of Buryatia and the people of the Yamasee nation in order to help as far as education, health, welfare, et cetera, to do for ourselves."

The District Court then sentenced Sampson. Unlike Campbell, Sampson had no prior convictions, but he had been charged with two additional counts of fraud perpetrated while out on bail — he attempted to secure $100 million for a fake oil field, and actually secured $35,000 from another investor through a "lifecycle fraud." But the Court did not mention these distinctions, nor any other details of Sampson's case. Instead, the Court made one curt reference to deterrence:

J. App'x at 1412-13.

3. Juncal's Sentencing

The District Court again adopted the PSR's calculations, including the role enhancement, despite the fact that the government was no longer seeking it. The District Court also found that Juncal had a criminal history category of II. Juncal's only prior convictions were imposed over fifteen years before he tried to sell Re on the Siberian Pipeline, and thus were not countable.2 See U.S.S.G. § 4A1.2e. In his interview with probation, however, Juncal indicated that he had dreamed of building the Siberian pipeline project since the days of "Reagan." The District Court credited this statement, and found that he had commenced planning the Siberian fraud scheme decades before he was caught, and determined his prior convictions were countable. The District Court then calculated the Guidelines in a cursory way:

J. App'x at 1433.

Juncal chose to speak on his own behalf. He explained that he suffered from a crippling bone marrow disease, and he only had a few years to live. "[T]here is no consequence to whether I die in jail or die outside," he lamented. J. App'x at 1438. He then went on to cite his extensive diplomatic service, most of it invented. As with his co-defendants, the District Court imposed the statutory maximum sentence:

J. App'x at 1444. The Court made no mention of its consideration of any of the factors set forth in 18 U.S.C. § 3553(a) other than the Sentencing Guidelines and the need for deterrence.

II. DISCUSSION

Appellants challenge both their convictions and the sentences imposed as a result of those convictions.

A. Appellants' Motions for Acquittal: Material Misrepresentations

We first turn to appellants' contention that there was insufficient evidence for a reasonable jury to conclude that any misrepresentations they made were material. A defendant challenging the sufficiency of the evidence bears a "heavy burden." United States v. Gaskin, 364 F.3d 438, 459 (2d Cir.2004). We review the evidence "in the light most favorable to the government, crediting every inference that could have been drawn in the government's favor." United States v. Chavez, 549 F.3d 119, 124 (2d Cir.2008). A conviction must be affirmed if "any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt." Jackson v. Virginia, 443 U.S. 307, 319, 99 S.Ct. 2781, 61 L.Ed.2d 560 (1979) (emphasis in the original).

Fraud requires more than deceit. A person can dissemble about many things, but a lie can support a fraud conviction only if it is material, that is, if it would affect a reasonable person's evaluation of a proposal. "In general, a false statement is material if it has a natural tendency to influence, or is capable of influencing, the decision of the decisionmaking body to which it was addressed." Neder v. United States, 527 U.S. 1, 16, 119 S.Ct. 1827, 144 L.Ed.2d 35 (1999) (internal quotation marks and brackets omitted). The appellants argue that this test requires a jury to conduct a subjective inquiry. According to the appellants, Neder requires a jury to assess whether a person with characteristics similar to the targeted victim would have been convinced by the defendants' false statement. And in this case, appellants posit, no reasonable investment professional would have bought the conspirators' absurd story; any broker would have laughed in disbelief the moment he opened an email from a wealthy bank sent from an AOL email address, found doctored copies of T-notes, and learned that a long-disbanded Native American tribe owned them.3 Thus, the argument goes, because no potential victim of this particular fraud would have ever fallen for it, the appellants' lies were not "capable of influencing the decision of [any] decisionmaking body."

We have not directly addressed whether Neder requires something more than an objective assessment of whether a defendant's lies were believable. In a related context, we have held that a defendant is liable for an objectively absurd lie if a subjectively foolish victim believes it. Otherwise "the legality of a defendant's conduct would depend on his fortuitous choice of a gullible victim." United States v. Thomas, 377 F.3d 232, 243 (2d Cir.2004) (citations omitted). Appellants aim to flip this rule so that a defendant is not liable for an objectively absurd lie if a subjectively sophisticated victim would never believe it. But no court has ever extended "the gullible victim" doctrine in this way, and we need not reach the question whether we should.

Taking the facts in the light most favorable to the government, we conclude that a reasonable jury could find that appellants' misrepresentations were material, regardless of whether Re, or others like him, should have heeded them or not. Appellants engaged in an illegal conspiracy to defraud potential investors the moment that Corsey offered Re billions of dollars in fake T-notes as collateral for a large loan. Re testified that he was convinced enough by Corsey's statements to vet the deal with his colleagues. Corsey's misrepresentation, then, actually influenced Re to investigate the proposal. And Re testified that his colleagues quickly performed Google searches to see if Re should explore the deal. Everything that followed that conversation and that initial internet search — the analogy to Christopher Columbus, the promised 171% rate of return, the crudely doctored certificates of authenticity — made Corsey's lie on behalf of the conspirators seem ridiculous in retrospect. But the initial misrepresentation remains material; a reasonable jury could find that a promise of five billion dollars in collateral could persuade Re, his colleagues, and reasonable lenders to make a three-billion-dollar loan.4

B. Sentencing: Review for Reasonableness

We review a district court's sentence for "reasonableness, which is `akin to review for abuse of discretion, under which we consider whether the sentencing judge exceeded the bounds of allowable discretion, committed an error of law in the course of exercising discretion, or made a clearly erroneous finding of fact.'" United States v. Leslie, 658 F.3d 140, 142 (2d Cir.2011) (per curiam) (quoting United States v. Williams, 475 F.3d 468, 474 (2d Cir.2007)). This broad discretion is due to district courts because they are a better positioned institutionally to determine the appropriate sentence in a particular case. United States v. Fernandez, 443 F.3d 19, 27 (2d Cir.2006).

We begin our analysis with a review of the sentences' procedural reasonableness. A district court will normally "begin all sentencing proceedings by correctly calculating the applicable Guidelines range," Gall v. United States, 552 U.S. 38, 49, 128 S.Ct. 586, 169 L.Ed.2d 445 (2007), and will then consider the factors listed in 18 U.S.C. § 3553(a) before imposing a final sentence. United States v. Preacely, 628 F.3d 72, 79 (2d Cir.2010). A judge need not utter "robotic incantations" repeating each factor that motivates a sentence. United States v. Crosby, 397 F.3d 103, 113-14 (2d Cir.2005) (abrogated on other grounds). Nonetheless, the judge must explain enough about the sentence for a reviewing court both to understand it and to assure itself that the judge considered the principles enunciated in federal statutes and the Guidelines. Rita v. United States, 551 U.S. 338, 357, 127 S.Ct. 2456, 168 L.Ed.2d 203 (2007). Here, appellants offer two critiques of the way the District Court imposed their sentences. First, they argue that the District Court either miscalculated the Guidelines range or failed to calculate the range entirely. Second, defendants claim that the District Court did not appropriately weigh the factors enunciated in 18 U.S.C. § 3553(a).

The District Court technically may have cut a corner when it glossed over the calculation of appellants' precise adjusted offense level, but that shortcut had no effect on appellants' ultimate Guidelines range. In each case, the amount of intended loss catapulted appellants' offense levels to 37, a number yielding a suggested period of incarceration above the statutory maximum of 20 years. Once the offense level crossed that threshold, any additional enhancement would have had at most an academic impact on the Guidelines range; the recommended term of incarceration would have always reverted back to twenty years. U.S.S.G. § 5G1.1(a) ("[w]here the statutorily authorized maximum sentence is less than the minimum of the applicable guideline range, the statutorily authorized maximum sentence shall be the guideline sentence."). Rather than resolve technical arguments about each enhancement, the District Court chose a more pragmatic approach — it calculated the Guidelines in sufficient detail to know that the resulting sentencing range would inevitably equal 20 years regardless whether certain additional enhancements applied.

The District Court was free to prioritize efficiency over an exercise in futility. In Rita v. United States, the Supreme Court instructed district courts to begin each sentencing evaluation with a proper calculation of the Guidelines range. 551 U.S. at 351, 127 S.Ct. 2456. We have interpreted that command to hold that a district court "commits procedural error where it fails to calculate the Guidelines range." United States v. Cavera, 550 F.3d 180, 190 (2d Cir.2008). Even in cases where courts depart or impose a non-Guidelines sentence, the Guidelines range sets an important benchmark against which to measure an appropriate sentence. For this reason, we have vacated sentences where a district court skipped a step in the Guidelines calculation because it had already decided to impose a non-Guidelines sentence. See United States v. Oliveras, 359 Fed.Appx. 257, 259 (2d Cir.2010) (vacating sentence where district court never addressed career offender enhancement because it intended to reduce criminal history score that overstated severity of defendants' prior convictions). But we have never held that a court commits procedural error where it bypasses a minefield of tricky determinations and still arrives at the correct Guidelines recommended sentencing range. To do so would be the equivalent of docking a test grade even though a student found a quicker route to the right answer. Cf. Cruz v. Miller, 255 F.3d 77, 86 (2d Cir.2001) ("[W]e are determining the reasonableness of the state courts' `decision'... not grading their papers."). Here, the District Court fulfilled its duty to calculate the Sentencing Guidelines sufficiently fully to determine the correct recommended period of incarceration, 20 years, and its particular method of arriving at that determination under the circumstances of these sentencings did not constitute reversible error.5

But the District Court still had an obligation to weigh the factors listed in section 3553(a), and we are not certain that occurred. See Fernandez, 443 F.3d at 28 (declining to hold within-Guidelines sentence presumptively reasonable). Here, appellants' maximum sentences had an air of inevitability. Because appellants' Guidelines ranges prior to application of the section 5G1.1 limit soared past the statutory maximum, the District Court seemed to assume that imposing a statutory maximum sentence reflected a per se reasonable sentence. At Campbell's hearing, the District Court chastened the defense attorney by noting that "even without [an] enhancement there would still be a maximum of a life sentence. Correct?" J. App'x at 1373. After a confusing colloquy with the probation officer and the prosecutor, the Court then correctly noted that as a technical matter the Guidelines range would revert back to the statutory maximum. That unavailable life sentence, however, hung in the air and framed the Court's perfunctory analysis of the section 3553(a) factors. A similar dynamic occurred during Sampson and Juncal's sentencing hearings. See Sampson Hearing (noting that enhancement "really has no effect" because court was forced to ratchet Guidelines range down to a twenty-year sentence); Juncal Hearing (skipping independent Guidelines analysis and imposing statutory maximum). The record, then, is muddied by the District Court's elision of two questions — how to understand the reversions of the Guidelines range to the statutory maximum of 240 months, and whether a statutory maximum sentence was sufficient but not greater than necessary to serve the purposes of sentencing enunciated in section 3553(a).

If in fact the District Court treated 240 months as the only reasonable sentence (because the statutory maximum prohibited imposition of a true Guidelines sentence closer to life), then it committed procedural error. Even when section 5G1.1(a) ratchets down a range, the lower Guidelines range is still the recommendation, and should be treated as a starting point in a deeper analysis, not a revision that obviates the need to consider further what sentence to impose. See Cavera, 550 F.3d at 189-90 (stating that our review for procedural reasonableness "requires that we be confident that the sentence resulted from the district court's considered judgment as to what was necessary to address the various, often conflicting, purposes of sentencing"). If, however, the District Court understood the effect of section 5G1.1(a), and still chose to impose the maximum sentence after considering the commands of section 3553(a), then the sentence may have been procedurally sound. But the record here is ambiguous, and we have often vacated and remanded sentences where a "the judge's sentencing remarks create ambiguity as to whether the judge correctly understood an available option." United States v. Thorpe, 191 F.3d 339, 342 (2d Cir.1999) (citations omitted); see also United States v. Jones, 531 F.3d 163, 181 (2d Cir.2008) (vacating sentence where "district court's explanation for its decision to impose a Guidelines sentence is ambiguous as to its understanding of an area of law"); United States v. Toohey, 448 F.3d 542, 545 (2d Cir.2006) (counseling that remand is appropriate where record reflects "misunderstanding by a district court as to the statutory requirements and the sentencing range... or misperception about their relevance"). In a case clouded by the possibility of error, we feel it appropriate to give the District Court an opportunity to clarify its thinking.

Remand is also appropriate for two additional reasons. First, the sentencing court gave only a passing mention to any of the section 3553(a) factors. In each case, the District Court justified a maximum sentence by relying almost exclusively on one word — "deterrence." Certainly, a judge need not chronicle the deliberative journey taken to arrive at a sentence, but she must still create enough of a record so that an appellate court can "be confident that the sentence resulted from the district court's considered judgment as to what was necessary to address the various, often conflicting, purposes of sentencing." Cavera, 550 F.3d at 189-90. In other words, just as we do not insist upon "robotic incantations," we require more than a few magic words.

This is especially true in light of the appellants' arguments for lighter sentences. "[W]here the appellant or prosecutor presents nonfrivolous reasons for imposing a different sentence ... the judge will normally go further and explain why he has rejected those arguments." Rita, 551 U.S. at 357, 127 S.Ct. 2456 (2007). Here, appellants' lawyers highlighted significant issues with the intended loss calculation both in their briefs and at sentencing. Given the low risk that any actual loss would result — what hedge fund would fall prey to a purported coalition of Buryatian nationals and Yamasee tribesmen using AOL email accounts to offer five billion dollars in collateral for a loan to build a pipeline across Siberia? — counsel argued that a 30-point mega-enhancement vastly overstated both the seriousness of the offense, and the danger of appellants to their community. The Guidelines acknowledge that potentiality; application note 19(C) to U.S.S.G. § 2B1.1 indicates that a downward departure may be warranted where the offense level resulting from a loss calculation overstates the seriousness of an offense. But the sentencing court never resolved appellants' significant arguments. At Sampson's hearing the District Court did draw a comparison between other financial crimes and this case, but it never resolved the question raised by the appellants — whether treating intended loss like actual loss under all the circumstances of this case leads to a sentence consistent with the dictates of section 3553(a).

In addition, the District Court appears to have assumed that it could incorporate by reference comments made during one appellant's sentencing hearing when conducting another. When Sampson's attorney attempted to argue that the loss was overstated, the District Court referenced a rejoinder made at Campbell's hearing. Even in multi-defendant cases, the court must sentence each defendant separately. Though it should of course avoid disparities among defendants of equal culpability, the court must come to an independent determination of the appropriate punishment for each defendant, and must create a record of that individualized determination. See 18 U.S.C. § 3553(c) (obligating sentencing court to state reasons for imposing a punishment in open court). That did not occur here, and we counsel the District Court to make individualized records on remand.

For these reasons, the appellants' sentences were procedurally unsound and appellants are entitled to resentencing. Although we find procedural error, we decline appellants' request to consider whether their sentences were substantively unreasonable.

II. CONCLUSION

We affirm the appellants' convictions, but vacate their sentences and remand for resentencings consistent with this decision.

UNDERHILL, District Judge, concurring:

Although I agree that appellants' sentences should be vacated and remanded for procedural error, the real problem is that those sentences are shockingly high. For that reason, I would reach the question of substantive reasonableness and would reverse on the merits.

In my view, the loss guideline is fundamentally flawed, and those flaws are magnified where, as here, the entire loss amount consists of intended loss. Even if it were perfect, the loss guideline would prove valueless in this case, because the conduct underlying these convictions is more farcical than dangerous. If substantive review of sentences actually exists other than in theory, it must be undertaken at least occasionally. This would have been an appropriate case in which to do so, because it raises so starkly the problems with the loss guideline. Until this Court weighs in on the merits of the loss guideline, sentences in high-loss cases will remain wildly divergent as some district judges apply the loss guideline unquestioningly while others essentially ignore it. The widespread perception that the loss guideline is broken leaves district judges without meaningful guidance in high-loss cases; that void can only be filled through the common law, which requires that we reach the substantive reasonableness of these sentences.

Ordinarily, we review a sentence to determine whether there was a procedural error in a district court's sentencing before we consider "the substantive reasonableness of the sentence imposed under an abuse-of-discretion standard, taking into account the totality of the circumstances." United States v. Rigas, 583 F.3d 108, 121 (2d Cir.2009) (internal quotation marks omitted). That approach is wholly appropriate in the vast majority of sentencing appeals. Nothing, however, "prevents us from reaching both the procedural and substantive reasonableness of the sentence in the course of an appeal where we find both types of error." United States v. Dorvee, 616 F.3d 174, 182 (2d Cir.2010) (citations omitted). Indeed, addressing both procedural and substantive reasonableness in the same appeal can serve the interest of judicial economy. Id. It is also the only means of addressing systemic problems with the Sentencing Guidelines. In short, "because procedural error and substantive error are permeable concepts... postponing the consideration of substantive reasonableness [can be] a mistake and a missed opportunity." United States v. Stewart, 597 F.3d 514, 515 (2d Cir.2010) (Jacobs, C.J., concurring) (denying rehearing in banc).

Substantive review of sentences provides "a backstop for those few cases that, [even if] procedurally correct, would nonetheless damage the administration of justice because the sentence imposed was shockingly high, shockingly low, or otherwise unsupportable as a matter of law." Rigas, 583 F.3d at 123 (citation and footnote omitted). Not surprisingly, we have only rarely held that a sentence is substantively unreasonable. E.g., Dorvee, 616 F.3d at 181, 184 (vacating "a within-Guide-lines sentence"; "the amount by which a sentence deviates from the applicable Guidelines range is not the measure of how `reasonable' a sentence is. Reasonableness is determined instead by the district court's individualized application of the statutory sentencing factors" (citing Gall v. United States, 552 U.S. 38, 46-47, 128 S.Ct. 586, 169 L.Ed.2d 445 (2007))). In sum, we will set aside a district court's substantive determination of the appropriate sentence only "in exceptional cases where the trial court's decision cannot be located within the range of permissible decisions." United States v. Cavera, 550 F.3d 180, 189 (2d Cir.2008) (citation and internal quotation marks omitted). This is such a case.

The twenty-year sentences imposed on appellants are not merely harsh, they are dramatically more severe than can be justified by the crime the appellants committed. This was a clumsy, almost comical, conspiracy to defraud a non-existent investor of three billion dollars. That scheme never came close to fruition. During his first meeting with Thomas Re, Emerson Corsey described Magnolia International Bank and Trust as the central bank for scores of Native American governments, including the Yamasee Indian tribe, which a Wikipedia search would have revealed as a tribal confederation that was broken up and defeated early in the 18th century. See http://en.wikipedia.org/wiki/Yamasee. It took only a brief Google search for Re and his associates to understand that the proposal "smelled" — which is why the appellants were recorded by Re for months before their arrest. At one point, Corsey provided Re with a certificate signed by John Juncal that listed CUSIP numbers for the T-notes; when Re shared the certificate with his colleagues, they responded by bursting into laughter. Even the terms of the proposed deal itself were laughable: the lender of three billion dollars would, according to the appellants, receive fourteen billion dollars in profit over five years. This scheme amounted to a series of absurd lies piled on top of even more absurd lies. Appellants' conduct was not dangerous because they had absolutely no hope of success.

A single factor — loss, specifically intended loss — drove the Guidelines calculation, and a single section 3553(a) factor — deterrence — provided the basis for accepting the Guidelines recommended sentence. A district court may not presume that a sentence within the Guidelines range is reasonable. Cavera, 550 F.3d at 189. Although the District Court used the intended loss amount correctly for purposes of calculating the Sentencing Guidelines range, the Court erred by failing to dramatically discount that calculation when weighing the section 3553(a) factors against the totality of the circumstances of this case. This conspiracy to defraud involved no actual loss, no probable loss, and no victim. The scheme was treated as sophisticated, but could be more accurately described as a comedic plot outline for a "Three Stooges" episode. Because the plan was farcical, the use of intended loss as a proxy for seriousness of the crime was wholly arbitrary: the seriousness of this conduct did not turn on the amount of intended loss any more than would the seriousness of a scheme to sell the Brooklyn Bridge turn on whether the sale price was set at three thousand dollars, three million dollars, or three billion dollars. By relying unquestioningly on the amount of the intended loss, the District Court treated this pathetic crime as a multi-billion dollar fraud — that is, one of the most serious frauds in the history of the federal courts. See United States v. Parris, 573 F.Supp.2d 744, 753 (E.D.N.Y. 2008) (collecting sentences for major securities fraud convictions).

The error of accepting intended loss as a proxy for the seriousness of this crime was "compounded by the fact that the district court was working with a Guideline that is fundamentally different from most and that, unless applied with great care, can lead to unreasonable sentences that are inconsistent with what § 3553 requires." Dorvee, 616 F.3d at 184. The loss guideline, like the child pornography guideline at issue in Dorvee, was not developed by the Sentencing Commission using an empirical approach based on data about past sentencing practices. As such, district judges can and should exercise their discretion when deciding whether or not to follow the sentencing advice that guideline provides. See Kimbrough v. United States, 552 U.S. 85, 109-10, 128 S.Ct. 558, 169 L.Ed.2d 481 (2007); see also United States v. Diaz, No. 11-Cr.-00821-2(JG), 2013 WL 322243 at *3-7 (E.D.N.Y. Jan. 28, 2013) (expressing disagreement with Guidelines ranges for drug trafficking offenses because they are not based on empirical data and national experience).

The fraud guideline was initially set forth in Guideline section 2F1.1. The Sentencing Commission set the original 1987 Guidelines for economic offenses higher than historical sentences in order to further the deterrence and just punishment goals of sentencing. In 1989, in response to the savings and loan crisis, Congress passed legislation increasing the maximum penalties for financial fraud offenses and directing the Sentencing Commission to include specific offense characteristic enhancements in the fraud guideline. See Robert S. Bennett, et al., Taking it to the Banks: The Use of the Criminal Process to Regulate Financial Institutions, 109 BANKING L.J. 28, 28-34 (1992). In 2001, the Sentencing Commission amended the Guidelines to combine the fraud, theft and embezzlement, and property destruction guidelines into a single guideline, section 2B1.1. That change was accompanied by the publication of a new loss table that had the effect of increasing offense level calculations, especially for high-dollar-value crimes. See U.S. SENTENCING COMM'N, REPORT TO THE CONGRESS: INCREASED PENALTIES UNDER THE SARBANES-OXLEY ACT OF 2002 at 7 (2003). Most recently, the fraud guideline was amended in 2003 in response to Congressional directives in the Sarbanes-Oxley Act. Id. Those amendments included further changes to the loss table that added offense level points in the highest loss cases. U.S.S.G. app. C amend. 647 (Nov. 1, 2003). The three sets of amendments to the loss table of the fraud guideline alone have effectively multiplied several times the recommended sentence applicable in 1987 for large-loss frauds, which itself was set higher than historic sentences. Each of the three increases in the recommended Guideline ranges for fraud crimes was directed by Congress, without the benefit of empirical study of actual fraud sentences by the Sentencing Commission.

The history of bracket inflation directed by Congress renders the loss guideline fundamentally flawed, especially as loss amounts climb. The higher the loss amount, the more distorted is the guideline's advice to sentencing judges. As a well-known sentencing commentator has put it, "For the small class of defendants... convicted of fraud offenses associated with very large guidelines loss calculations, the guidelines now are divorced both from the objectives of Section 3553(a) and, frankly, from common sense. Accordingly, the guidelines calculations in such cases are of diminished value to sentencing judges." Frank O. Bowman, III, Sentencing High-Loss Corporate Insider Frauds After Booker, 20 FED. SENT'G REP. 167, 168 (2008).

When the Guidelines range zooms off the sentencing table, sentencing judges are discouraged from undertaking close examination of the circumstances of the offense and the background and characteristics of the offender. That certainly happened here. But the low marginal utility of the guideline in this very high intended loss case should have prompted greater, not lesser, reliance on the section 3553(a) factors other than the Guidelines. As one District Court applying the loss guideline put it, "Where the Sentencing Guidelines provide reasonable guidance, they are of considerable help to any judge in fashioning a sentence that is fair, just, and reasonable. But where, as here, the calculations under the guidelines have so run amok that they are patently absurd on their face, a Court is forced to place greater reliance on the more general considerations set forth in section 3553(a), as carefully applied to the particular circumstances of the case and of the human being who will bear the consequences." United States v. Adelson, 441 F.Supp.2d 506, 515 (S.D.N.Y.2006), aff'd mem., 301 Fed.Appx. 93 (2d Cir.2008).

In this case, it is impossible to describe the sentences imposed on appellants as substantively reasonable. In my view, none of the section 3553(a) factors, singly or in combination, can justify these shockingly high punishments, which are far greater than necessary 19 to punish or deter appellants' conduct. The District Court provided "no reason why the maximum sentence of incarceration was required to deter [the appellants] and offenders with similar history and characteristics." Dorvee, 616 F.3d at 184. The bare assertion of the need to deter, unconnected to the background and characteristics of a defendant or the nature and circumstances of a crime, provides only superficial support for a sentence of imprisonment. Here, the factor of deterrence simply cannot "bear the weight assigned it under the totality of circumstances in the case." Cavera, 550 F.3d at 191. None of the appellants had ever served a significant term of imprisonment and two of the three had spent no time in prison whatsoever. This raised the question whether a statutory maximum sentence is necessary to deter future wrongdoing. See United States v. Mishoe, 241 F.3d 214, 220 (2d Cir.2001) (counseling district courts to consider previous terms of incarcerations and calibrating new sentences to reflect prior time behind bars). That is especially true here because appellants are older than most defendants and, accordingly, can be expected to have a lower risk of recidivism than most. See U.S. Sentencing Commission, MEASURING RECIDIVISM: THE CRIMINAL HISTORY COMPUTATION OF THE FEDERAL SENTENCING GUIDELINES 16 (2004). Indeed, one of the appellants will surely die in prison, if required to serve a sentence anywhere close to twenty years; although an effective life sentence, by definition, provides complete deterrence, imposing such a sentence for the conduct underlying this conviction would be senseless.

The absence of any actual loss whatsoever and especially the absence of a victim significantly undercut any argument that this crime was particularly serious. Outside the context of Sentencing Guidelines calculations, intended loss is always less serious than actual loss, so its value as a proxy for seriousness of a crime must be carefully examined. And not all actual loss is equally serious. A fraud that results in the loss of even a few thousand dollars by an elderly or sick person who, as a result of the loss, becomes unable to afford the necessities of life or medical care is much more serious than a fraud that results in ten or a hundred times that loss by a large corporation able to absorb the financial consequences without a need to close plants, fire employees, or even declare the loss as material in public financial reports. Simply put, contrary to the assumption underlying the loss guideline, not all dollars of loss are fungible.

Moreover, the convictions in this case were for conspiracy, which proscribes an agreement to commit fraud and punishes that agreement the same whether or not a fraud was actually committed. "[W]e punish unconsummated efforts to cause harm as `attempts' or `conspiracies' (albeit usually less severely than completed crimes) so long as the would-be-perpetrator has come close enough to success that we can be confident his malignant designs were real and not mere fantasy, and thus that his conduct was morally blameworthy." Frank O. Bowman, III, Coping with `Loss': A Re-examination of Sentencing Federal Economic Crimes Under the Guidelines, 51 VAND. L.REV. 461, 559 (1998). The circumstances of these convictions put them very close to the boundary of mere fantasy (or perhaps the boundary of mental competence) and the sentences should have reflected that fact. As with most attempt crimes and unconsummated conspiracies, the actual loss here was zero. The wrongfulness of the appellants' conduct does not turn in any meaningful sense on the amount that the conspirators sought to obtain; all other things equal, an effort to secure $3 billion for construction of an imaginary pipeline is not 100 times as serious as an effort to secure $30 million for construction of an imaginary factory. Accordingly, factors other than intended loss become critical in distinguishing the seriousness of unconsummated criminal conduct.

District judges are not permitted to treat the Sentencing Guidelines as reasonable. Cavera, 550 F.3d at 189. The corollary of that proposition is that district judges have an obligation to consider whether a sentence other than a Guidelines sentence would be sufficient, but not greater than necessary, to serve the purposes of sentencing. 18 U.S.C. § 3553(a). Accordingly, district judges have an obligation to consider whether to depart from the Guidelines sentencing range or to impose a non-Guidelines sentence in every case. That duty will frequently require judges applying the loss guideline to evaluate whether the calculated Guidelines range substantially overstates the seriousness of a crime. See U.S.S.G. § 2B1.1 application note 19(C) ("There may be cases in which the offense level determined under this guideline substantially overstates the seriousness of the offense. In such cases, a downward departure may be warranted."). That duty went unfulfilled in this case, and the result was the imposition of shockingly high sentences on the appellants.

The Court of Appeals missed an opportunity in this case to provide much needed guidance to district judges who must apply the misguided loss guideline. Thankfully, the District Court will have the opportunity at resentencing to undertake the difficult task of weighing all of the section 3553(a) factors under the circumstances of these cases to reach sentences that are sufficient, but not greater than necessary, to serve the purposes of sentencing.