United States v. Levinson

OPINION JORDAN, Circuit Judge. Adam Levinson pleaded guilty to one count of wire fraud, in violation of 18 U.S.C. § 1343, and one count of filing a false income tax return, in violation of 26 U.S.C. § 7206(1). During his sentencing hearing, the United States District Court for the District of Delaware granted him a variance from the recommended United States Sentencing Guidelines (“Guidelines”) range of 24 to 30 months of imprisonment and sentenced him to two concurrent 24-month terms of probation, in addition to supervised release, restitution, and a special assessment. The government appeals and argues that the District Court failed to adequately explain the chosen sentence. We agree and will vacate Levinson’s sentence and remand for resentencing. I. Background Levinson was the manager and twenty-percent owner of CoolerSmart, LLC (“CoolerSmart”), a Delaware company that provided filtered drinking water to residences and businesses. CoolerSmart’s majority owner was WaterWorld Ventures, Inc. (“WaterWorld”), a wholly-owned subsidiary of Elkay Manufacturing Company (“Elkay”). Between June 2000 and August 2002, Levinson falsely reported Cool-erSmart’s financial status and operational performance, representing that the company was successful, when in fact it was not. Elkay relied on those reports when it invested millions of additional dollars into CoolerSmart’s business. In 2001, Elkay reviewed CoolerSmart’s operations and finances. In preparation for the review, Levinson implemented a *192 scheme to cover up the falsehoods he had already told about CoolerSmart’s performance. He created a second set of financial and operational books and altered Cool-erSmart’s general ledger and customer lists. He also hired an outside company to enter false customer information into Cool-erSmart’s databases. At his direction, CoolerSmart employees shredded documents, deleted e-mail and other electronic records, created false sales reports, recorded unauthorized expenditures, and falsely reported financial information. Levinson punished employees who refused to participate in his cover-up scheme, and, in at least one instance, he cut an employee’s salary in half and eliminated her benefits. In 2002, Elkay received an anonymous tip warning of misconduct by CoolerS-mart’s management. Elkay responded by hiring an accounting firm to conduct a forensic audit of CoolerSmart. The audit not only uncovered Levinson’s fraudulent reporting practices, it also revealed that Levinson had used over $177,000 of Cool-erSmart’s revenue for his own benefit — to make personal loan payments, to pay for repairs on his home, to take vacations, and to make personal credit card bill payments. Levinson failed to report his use of those funds as additional revenue on his 2000, 2001, or 2002 federal income tax returns. As a result, the government’s aggregate tax loss for those years was approximately $44,000. On June 6, 2006, Levinson was indicted on one count of wire fraud and three counts of filing a false income tax return. He entered into a plea agreement that stated he would plead guilty to the wire fraud count and to one count of filing a false income tax return in the year 2002. Prior to his sentencing hearing, Levinson settled a civil fraud suit brought against him by Elkay by paying Elkay $350,000 and relinquishing his twenty-percent ownership interest in CoolerSmart. However, as to the wire fraud count, the government and Levinson agreed that, for purposes of calculating Levinson’s offense level under § 2Bl.l(b) of the Guidelines and for calculating restitution for the fraud, they would use the amount Levinson took from Cool-erSmart for his personal use, which was $177,289. After adjusting Levinson’s base offense level of 6 upward by ten points to account for the $177,289 loss to CoolerS-mart, his offense level of 16 was further increased by two points pursuant to § 3B1.1 for his role as the organizer, leader, manager, or supervisor of the offense. His adjusted base offense level of 18 was further increased by two points pursuant to § 3B1.3 for his use of special skill in facilitating and concealing his crime, resulting in a final offense level of 20 for the wire fraud count. Levinson’s base offense level for filing the false income tax return was 14, pursuant to § 2T4.1, given that the tax loss was between $30,000 and $80,000. That was adjusted upward by two points under § 2T1.1 for his failure to report more than $10,000 in income derived from criminal activity, resulting in an adjusted offense level of 16. At sentencing, the District Court used the greater of the two offense levels to compute Levinson’s sentence under the Guidelines, as required by § 3D1.3(a). The Court then reduced that offense level of 20 by three levels, in consideration of his acceptance of responsibility, for a total offense level of 17. Because Levinson had no prior criminal record, the Court determined that he fell within Guidelines Criminal History Category I. Therefore, the Court calculated that the recommended sentencing range under the Guidelines was 24 to 30 months imprisonment, two to three years of supervised release, a fine of *193 $5,000 to $354,578 1 restitution in the amount of $177,289, and a $200 special assessment. Levinson moved for a downward departure based on what he claimed was his diminished mental capacity resulting from bipolar disorder. The District Court declined to grant the motion, concluding that Levinson’s actions belied the claim of diminished capacity because his fraud was sophisticated and complex, as were his detailed and extensive efforts to conceal his misrepresentations from Elkay. After determining the recommended sentence for Levinson under the Guidelines, the District Court heard both parties’ arguments regarding the application of the sentencing factors set forth in 18 U.S.C. § 3553(a). 2 The government maintained that a sentence within the Guidelines was appropriate because of the seriousness of the offense, Levinson having engaged in an elaborate, ongoing scheme that violated Elkay’s trust in him as Cool-erSmart’s manager and minority shareholder. The government also pointed out that, as part of his scheme to defraud Elkay, Levinson extorted and manipulated CoolerSmart employees by threatening to demote or fire them and to terminate their benefits if they refused to participate. Finally, the government emphasized that, once Levinson realized that Elkay was about to uncover the fraud, he tried to conceal it while simultaneously negotiating the sale of his CoolerSmart shares to a third party without Elkay’s consent. Levinson argued for a sentence of home or community confinement, as opposed to incarceration. He presented the Court with mitigating factors, expressed in a large number of letters written on his behalf by his family, friends, and members of the community, particularly emphasizing his role as a father, his contributions to society and, in particular, his involvement with a Jewish family service organization. Other mitigating factors he pressed included his prompt payment of $350,000 to El-kay to settle the civil suit, his loss of employment if he were imprisoned, and the financial and emotional toll his imprisonment would have on his young family, especially his learning-disabled son. The Court then considered the § 3553(a) factors. It acknowledged the many letters written on Levinson’s behalf, which urged leniency both for the sake of his family and because he was a good person who had made poor decisions. According to the *194 District Court, those considerations did not distinguish Levinson from other criminals. The District Court did, however, determine that Levinson’s case could be distinguished from other cases involving white-collar crime because his victim was not the public at large but was instead a private business entity that had already settled its civil lawsuit against Levinson. Coming to its conclusion on sentencing, the District Court stated: [Levinson] was a businessman who stole from his business partners for his own benefit. He was and remains, however, a person who contributes to the community, his family and friends. In the end, it seems to me that we have an individual who put the appearance of prosperity above his respect for the law. Balanced against this is the propriety of putting into jail at a substantial cost to the public a nonviolent offender who poses little or no threat to the public and whose crimes had little impact beyond his business partners and his family. ... [W]hen I look at the costs associated with putting someone like Mr. Levinson [in] jail in this day and age compared to the harm he has caused, which has been resolved amicably with his business and which certainly will impose even more harm on his family, I just can’t see that it makes much sense. I just do not. (App. at 38-39.) The Court thus declined to impose a sentence of imprisonment and instead sentenced Levinson to 24 months of probation for the wire fraud count, with a concurrent 24-month term of probation for filing a false income tax return, 100 hours of community service, $177,289 in restitution, and a $200 special assessment. Levinson was also sentenced to a six-month period of home confinement with electronic monitoring. The government then filed this appeal. II. Discussion 3 The government argues that the District Court failed to articulate sufficiently compelling reasons to support its variance from the Guidelines recommendation, and that the Court erred by relying upon clearly erroneous and inconsistent factual findings and on factors unrelated to the sentencing considerations outlined in § 3553(a). According to the government, the District Court committed procedural and substantive errors in imposing Levin-son’s sentence, and the case should be remanded with instructions to impose a sentence within the advisory Guidelines range. We do not perceive all of the flaws alleged by the government, but we do agree that this case needs to be remanded. Although the District Court may ultimately conclude that the sentence now on appeal should, be re-imposed at Levinson’s resentencing, the Court must provide us with enough analysis on the record to permit meaningful appellate review, which it so far has not. A. Roles of District and Appellate Courts By now, the three-step sentencing procedure set forth in United States v. Gunter, 462 F.3d 237, 247 (3d Cir.2006) (“Gun-ter I ”), has become familiar and has been effectively ratified by the Supreme Court’s decision in Gall v. United States, — U.S. -, 128 S.Ct. 586, 169 L.Ed.2d 445 (2007). A district court must begin the process by first calculating the applicable Guidelines range. After that initial calcu *195 lation, the court must then rule on any motions for departure and, if a motion is granted, state how the departure affects the Guidelines calculation. Finally, after allowing the parties an opportunity for argument, the court must consider all of the § 3553(a) factors and determine the appropriate sentence to impose, which may vary from the sentencing range called for by the Guidelines. United States v. Wise, 515 F.3d 207, 216-17 (3d Cir.2008) (citing Gall, 128 S.Ct. at 596-97; Gunter I, 462 F.3d at 247). Our responsibility on appellate review of a criminal sentence is limited yet important: we are to ensure that a substantively reasonable sentence has been imposed in a proeedurally fair way. As directed by the Supreme Court, we take up the procedural review first, looking to see that the district court has committed no significant error by, for example, “failing to calculate (or improperly calculating) the Guidelines range, treating the Guidelines as mandatory, failing to consider the § 3553(a) factors, selecting a sentence based on clearly erroneous facts, or failing to adequately explain the chosen sentence — -including an explanation for any deviation from the Guidelines range.” Gall, 128 S.Ct. at 597. If the sentencing decision passes that first stage of review, we then, at stage two, consider its substantive reasonableness. Id. An estimation of the outer bounds of what is “reasonable” under a given set of circumstances may not always be beyond debate, but the abuse-of-discretion standard by which that estimation must be judged limits the debate and gives district courts broad latitude in sentencing. 4 Obviously, procedural problems may lead to substantive problems, so there are times when a discussion of procedural error will necessarily raise questions about the substantive reasonableness of a sentence. United States v. Goff, 501 F.3d 250, 256 (3d Cir.2007) (“[T]hese ... substantive problems ... are a product of the District Court’s proeedurally flawed approach.”). After all, if one cannot justify a result by the reasons given, that result is, by definition, not a substantively reasonable conclusion to the logical steps provided. But the guidance we get from Gall is to focus in the first instance on the procedural aspect of a sentencing decision. See 128 S.Ct. at 597 (“[The appellate court] must first ensure that the district court committed no significant procedural error.... ”). B. Explanation as Key to Appellate Review One implication of Gall is that we should treat a failure to “adequately explain the chosen sentence — including an explanation for any deviation from the Guidelines range” — as a problem that can typically be addressed by giving the sentencing judge an opportunity to better explain the reasoning behind the decision. It will be a rare case when it is clear that no acceptable reasoning can justify a given sentence. Yet even rightly admired judges may make a decision which we believe is unsupportable, and we are obligated to point that out when it occurs. See Goff, 501 F.3d at 262 (deciding that “a sentence of four months is ... unreasonable in light of the facts and circumstances revealed in the record”); United States v. Pugh, 515 F.3d 1179, 1191 (11th Cir.2008) (“[Gall ] *196 leave[s] no doubt that an appellate court may still overturn a substantively unreasonable sentence, albeit only after examining it through the prism of abuse of discretion, and that appellate review has not been extinguished.”)- In general, however, when we are reviewing a sentence and find ourselves unable to see how the reasons articulated lead to the punishment imposed, we will be focused on trying to obtain a better understanding of the district court’s reasoning. We do not seek to second guess. Given the widely recognized institutional advantages that district courts have in access to and consideration of evidence, we would be foolish to try. See Kimbrough v. United States, — U.S.-, 128 S.Ct. 558, 574, 169 L.Ed.2d 481 (2007) (citing as “discrete institutional strengths” the district court’s “superior position to find facts and judge their import under § 3553(a) in each particular case” (internal citation and quotation marks omitted)). In each case, however, we must have an explanation from the district court sufficient for us to see that the particular circumstances of the case have been given meaningful consideration within the parameters of § 3553(a). The Supreme Court has explained that there must be “an individualized assessment based on the facts presented.” Gall, 128 S.Ct. at 597. A necessary corollary of that responsibility is the further obligation to provide sufficient justifications on the record to support the sentencing conclusion. See id. at 597 (“After settling on the appropriate sentence, [the sentencing judge] must adequately explain the chosen sentence to allow for meaningful appellate review and to promote the perception of fair sentencing.”); Rita v. United States, — U.S. -, 127 S.Ct. 2456, 2468, 168 L.Ed.2d 203 (2007) (“The sentencing judge should set forth enough to satisfy the appellate court that he has considered the parties’ arguments and has a reasoned basis for exercising his own legal decision-making authority.”). While the Guidelines are no longer mandatory, United States v. Booker, 543 U.S. 220, 245, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005), and no “extraordinary circumstances” are needed to justify a sentence that varies from their recommended results, Gall, 128 S.Ct. at 595, and while there is no mathematical formula for determining whether a district court’s justifications for a variance are sufficient, id. at 594-95, we nonetheless must be satisfied that, broadly speaking, an adequate justification is provided on the record. In the absence of that, we must remand. 5 *197 We do not pretend that the foregoing observations provide much, if any, guidance. Indeed, we find it difficult to give direction when we are ourselves endeavoring to understand our role in reviewing sentences after Booker, Rita, Gall, and Kimbrough. 6 In the end, though, we think this much is clear about sentencing in the post-Booker era: appellate review, limited though it is by the abuse-of-discretion standard, remains and requires district courts to plainly state the reasoning behind each sentence. Moreover, in deciding on appeal whether the reasons provided by a district court are adequate, the degree that a sentence varies from the recommendation given in the Guidelines matters. See Gall at 594-95 (“In reviewing the reasonableness of a sentence outside the Guidelines range, appellate courts may ... take the degree of variance into account and consider the extent of a deviation from the Guidelines.”). Hence, while we eschew any requirement of direct proportionality, we may look for a more complete explanation to support a sentence that varies from the Guidelines than we will look for when reviewing a sentence that falls within a properly calculated Guidelines range. Cf. United States v. Smalley, 517 F.3d 208, 215 n. 9 (3d Cir.2008) (“Because of the appellate court’s duty to review the sentence for reasonableness, Gall made it clear that ‘failing to adequately explain the chosen sentence-including an explanation for any deviation from the Guidelines’ was procedural error.”). C. The Sentencing at Issue In this case, as is typical in sentencing hearings, the District Court ruled from the bench, following arguments by counsel and the defendant’s allocution. The Court began by correctly calculating the advisory Guidelines range, which, without objection, it determined to be a sentence that included incarceration of 24 to 30 months. It considered, as our precedent requires, Gunter I, 462 F.3d at 247, the defendant’s motion for a downward departure, deciding that his claim of diminished capacity was not persuasive. The Court went on to give a thoughtful explanation of the § 3553(a) sentencing factors it had weighed. In particular, the District Court considered the nature and circumstances of Lev-inson’s crimes, explaining in detail his elaborate fraud and the cover-up scheme he had orchestrated: Mr. Levinson did engage in a detailed, extensive and thorough effort to conceal from Elkay ... numerous fraudulent misrepresentations regarding CoolerS-mart’s performance for more than two years. In order to convince Elkay that CoolerSmart was operating successfully, Mr. Levinson manipulated sales reports and falsified financial information. He hired temporary employees to create a *198 false set of books and sanctioned Cool-erSmart employees who would not help him in his fraudulent scheme. He shredded documents and deleted electronic data. He did such a good job, in fact, of covering up his fraudulent activities that Elkay did not discover them through its routine audits. Only through an anonymous tip was defendant’s conduct finally disclosed. It’s also apparent from the record that Mr. Levinson engaged in this complex course of conduct not just to keep his job and his company going[;] he used over $175,000 of the ill-gotten proceeds to lease a luxury car for his personal use, to travel to the beach, to Florida, ... for personal expenses with unauthorized company money. (App. at 35-36.) The Court then turned to Levinson’s personal history and characteristics: With respect to Mr. Levinson’s history and his characteristics, by all accounts, he is engaging, energetic and productive, as a member of the community and as a family member and friend. Likewise, the criminal activity which brings us to court today necessarily demanded these same characteristics. (Id. at 36.) In considering Levinson’s personal circumstances, the Court specifically acknowledged the many letters of support from Levinson’s friends and family and their pleas for leniency. The Court stated that the characterization of Levinson in the letters as a good person who just happened to make some poor decisions did not make him stand out, because most defendants are not bad people and “like Mr. Levinson have made poor choices, motivated by a variety of ills.” (Id. at 37.) The Court also acknowledged the accuracy of the letter-writers’ concern that Levin-son’s family would suffer if he were imprisoned. That Levinson’s family “would suffer most if he goes to jail,” did not distinguish him from other white-collar criminals sentenced to terms of imprisonment, the Court said, because “[t]hat sad outcome is endemic to most criminal cases, as most defendants have families who suffer because their loved one wasn’t thinking about them during the criminal activity.” (Id.) However, as previously noted, the District Court did say that Levinson’s case could be distinguished on the basis that Levinson’s crimes did not inflict financial harm on the public. The Court stated: It is important to note, however, that Mr. Levinson’s victim was a private business entity, [whose] principals have resolved their dispute for a sum of money. In other words, Mr. Levinson did not harm the public from a financial point of view. (Id. at 37.) The Court asserted that only Elkay, the corporation with which Levin-son had co-owned CoolerSmart, had been harmed, and that the corporation’s losses had already been addressed by a $350,000 civil settlement paid by Levinson. (Id. at 37-38.) The Court’s peroration bears repeating: [W]hen I look at the costs associated with putting someone like Mr. Levinson [in] jail in this day and age compared to the harm he has caused, which has been resolved amicably with his business and which certainly will impose even more harm on his family, I just can’t see that it makes any sense. I just do not. (Id. at 38-39.) The Court concluded that it would “reject the advisory guideline range of imprisonment” and order probation. (Id. at 39-40.) The government raises several issues regarding the Court’s reasoning and the resulting sentence. We think it sufficient to focus on two. *199 1. Clearly Erroneous Factual Foundation First, the government argues that the District Court based its decision on the clearly erroneous premise that Levinson had inflicted no financial harm on the public. There were two counts of conviction, the fraud count, which the Court discussed at some length, and the tax count, which, the government says, the Court entirely ignored. According to the government, since Levinson admitted guilt for filing a false tax return, and since the District Court had concluded that the tax loss associated with that count exceeded $44,000, the Court erred in basing Levinson’s sentence on the assertion that “Mr. Levinson did not harm the public from a financial point of view.” (Id. at 37.) Levinson endeavors to justify the District Court’s comments about private versus public harm by saying that there in fact is no public harm since, “with restitution of the wire fraud proceeds to the victim ..., the Defendant would be entitled to a deduction and thus owe the Government no net taxes by virtue of the 'evasion.’” (Appellee’s Br. at 3.) We are compelled to conclude that the government has much the better of those positions. Even if Levinson were correct as to the technical feasibility of the deduction he says he plans to claim on future tax returns, and we make no comment on that at all, his argument is still flawed. It attributes to the District Court a reason that is nowhere stated or even implied in the Court’s discussion of the case; there is simply nothing to indicate that the Court had in mind a “defraud now, deduct later” rationale for bypassing comment on the tax count of conviction. More importantly, however, Levinson’s argument fails to address the principal problem identified by the government, which is that there is no explanation by the District Court for how this case can be said to entail purely private harm when there is a tax fraud conviction involving a specific dollar loss to the United States Treasury. We thus agree with the government that the District Court rested its sentencing decision on an unsound factual foundation. See Concrete Pipe & Prods. of Cal., Inc. v. Constr. Laborers Pension Trust for S. Cal., 508 U.S. 602, 622, 113 S.Ct. 2264, 124 L.Ed.2d 539 (1993) (“A finding is clearly erroneous when although there is evidence to support it, the reviewing body on the entire evidence is left with the definite and firm conviction that a mistake has been committed.”) (internal quotation marks and citation omitted). 2. Inadequate Explanation for Variance The second issue raised by the government is the adequacy of the District Court’s explanation for varying downward from a sentence including 24 to 30 months imprisonment to a sentence with no imprisonment. Again, we agree that the record is problematic, and, even if there were no issue regarding the unmentioned tax conviction, we would need to return this case to the District Court for a more complete explanation of its decision to significantly vary from the recommended sentence of imprisonment. That conclusion rests on two concerns. One is the lack of explanation of how this defendant or his crimes differ in any way that warrants the special leniency shown. The other concern, which is related to the first, is that, if there is no real distinction between Levinson and other white-collar defendants, then the District Court appears to have rested its decision on a policy disagreement with the Guidelines, which it did not articulate or explain. As to the first concern, the Court’s sentencing decision expressly declares that Levinson does not differ from typical defendants, either in regard to his general *200 character or the effect that his incarceration would have on his family. We can imagine some surprise in the courtroom, then, as the direction shifted from what appeared to be a conclusion that Levin-son’s case would naturally fall within the Guidelines to the very different conclusion that a probationary sentence was warranted. While the Court did make the already noted comments about public versus private harm, we do not understand it to have been saying that that distinction made Levinson personally different from other defendants. Rather, those comments appear to bear on the policy disagreement that we will turn to shortly. To the extent the Court commented on the personal characteristics of Levinson and the particulars of his crimes, all it said was that Levinson was not atypical, so there was, in that regard, no adequate explanation for the relatively wide variance between the applicable Guidelines range and the sentence the Court pronounced. See Gall, 128 S.Ct. at 597 (“We find it uncontroversial that a major departure should be supported by a more significant justification than a minor one.”); cf. United States v. Howe, 543 F.3d 128, -, 2008 WL 4251920, [citation forthcoming] (3d Cir.2008) (upholding downward variance after reviewing reasons advanced by district court to justify the variance). Nor did the Court identify any other basis for varying, except for its observation that the costs of incarcerating a non-violent offender who had already paid some restitution to the victim left the Court believing that a prison sentence “makes no sense.” (App. at 39.) Those comments about cost, and the earlier comments about public and private harm, make it appear that the Court left the realm of weighing the particulars of this case and entered into a consideration of general penal policy. Policy considerations are not off-limits in sentencing, see Kimbrough, 128 S.Ct. at 570 (quoting government concession that “as a general matter, ‘courts may vary [from Guidelines ranges] based solely on policy considerations, including disagreements with the Guidelines’”), but care must be taken in reaching the conclusion that the District Court did here, because the public versus private harm distinction is not without nuance and because considering costs of incarceration will likely contravene very deliberate policy choices embedded in the Guidelines. The identity of the principal victim as a private corporation should not necessarily lead to the conclusion that there has been no harm to the public. In the broadest sense, all crime involves public harm, since crimes are crimes for the very reason that they involve a violation of public norms and require something more than the correction of a private wrong. In a narrower sense, crimes against private entities can and do inflict public harm, as, for example, when a publicly traded corporation has been defrauded and there is consequent damage to public trust in our nation’s capital markets. That does not mean that private versus public harm is a meaningless distinction. The Guidelines themselves take account of that distinction at times. See USSG § 2A6.1(b)(4) (directing offense level be increased by four levels if offense resulted in substantial disruption of public functions or services); USSG § 5K2.14 (providing for upward departure “[i]f national security, public health, or safety was significantly endangered” by the defendant). 7 But to say that the *201 Guidelines’ policies on incarceration are inapplicable to white-collar crime directed at private entities would be an extraordinary assertion of judicial power, impermissible in the abstract, 8 and, even when confined to a specific case, certainly requiring a thorough explanation. The Sentencing Commission recommended terms of imprisonment for economic crimes like Levinson’s because of its concern that sentencing for white-collar crime had been ineffectual. See USSG § 1A1.1, Ch. 1, Pt. A.4(d) (2006) (“Under [pre-Guidelines] sentencing practice, courts sentence[d] to probation an inappropriately high percentage of offenders guilty of certain economic crimes, such as theft, tax evasion, antitrust offenses, insider trading, fraud, and embezzlement, that in the Commission’s view are ‘serious.’ ”). In addition, it has been noted that probationary sentences for white-collar crime raise concerns of sentencing disparities according to socio-economic class. See United States v. Mueffelman, 470 F.3d 33, 40 (1st Cir.2006) (“Restitution is desirable but so is the deterrence of white-collar crime (of central concern to Congress), the minimization of discrepancies between white- and blue-collar offenses, and limits on the ability of those with money or earning potential to buy their way out of jail.”). Presumably, the Commission was aware of the costs of incarceration when it made its judgment that white-collar criminals generally should be sent to prison. To use the Commission’s characterization, white-collar crimes such as wire fraud and tax fraud are “serious,” and typically will warrant serious punishment, including prison time. Thus, if a district court wants to vary from the Guidelines for a reason that is contrary to the Commission’s stated position, it must explain why the general policy should not apply in the particular case before it. See United States v. Gunter, 527 F.3d 282, 286 (3d Cir.2008) (district court is “free to disagree with the policy underlying the crack/powder ratio as applied to that particular defendant and make an appropriate downward variance in its sentence.... [But][t]here must be meaningful consideration of the § 3553(a) factors and the particular circumstances of the case before a variance is made.” (emphasis omitted)). That was not done here. Instead, the District Court simply said that it had reviewed its past sentencing decisions and found prison appropriate when some public, as opposed to private, harm had been inflicted. That statement is not enough to tell us why the Guidelines, which, even *202 after Booker, remain “the starting point and initial benchmark” for sentencing, Gall, 128 S.Ct. at 596, should not apply to Levinson. It may be that the District Court has reasons we have not understood for varying widely from the recommended Guidelines range of sentences in this case. We do not say that a sentence of probation would be, on this record, plainly outside the boundaries of permissible discretion. We hold only that the justifications given for the sentence are inadequate for us to recognize them as reflecting a proper exercise of discretion. III. Conclusion Accordingly, we will vacate the sentence and remand the case for further proceedings in accordance with this opinion. Footnotes [1] The upper end of the fine range was defined as twice the gross gain or loss from the crimes, 18 U.S.C. § 3571(d), and the Court used the $177,289 loss as the basis for its fine calculation. [2] Section 3553(a) provides that a court is to consider the following factors when imposing a sentence: (1) the nature and circumstances of the offense and the history and characteristics of the defendant; (2) the need for the sentence imposed (A) to reflect the seriousness of the offense, to promote respect for the law, and to provide just punishment for the offense; (B) to afford adequate deterrence to criminal conduct; (C) to protect the public from further crimes of the defendant; and (D) to provide the defendant with needed educational or vocational training, medical care, or other correctional treatment in the most effective manner; (3) the kinds of sentences available; (4) the kinds of sentence and the sentencing range established for ... the applicable category of offense committed by the applicable category of defendant as set forth in the guidelines ...; (5) any pertinent policy statement ... issued by the sentencing commission ...; (6) the need to avoid unwarranted sentence disparities among defendants with similar records who have been found guilty of similar conduct; and (7) the need to provide restitution to any victims of the offense. [3] The District Court had subject matter jurisdiction over this case pursuant to 18 U.S.C. § 3231. We have jurisdiction over the government's appeal pursuant to 18 U.S.C. § 3742 and 28 U.S.C. § 1291. [4] Although we use an abuse-of-discretion standard when reviewing a district court's decision, "the amount of deference we give will depend on the type of procedural error asserted on appeal." Wise, 515 F.3d at 217. Thus, we will conclude that the district court abused its discretion regarding factual findings only if those findings are clearly erroneous. Id. In contrast, if the procedural error is purely legal, we do not defer to the district court. Id. [5] In Rita, the Supreme Court emphasized the importance of district courts providing sufficient explanations of their sentencing decisions, and the latitude those courts have in deciding what is sufficient in a given case: The sentencing judge should set forth enough to satisfy the appellate court that he has considered the parties’ arguments and has a reasoned basis for exercising his own legal decisionmaking authority. Nonetheless, when a judge decides simply to apply the Guidelines to a particular case, doing so will not necessarily require lengthy explanation. Circumstances may well make clear that the judge rests his decision upon the Commission's own reasoning that the Guidelines sentence is a proper sentence (in terms of § 3553(a) and other congressional mandates) in the typical case, and that the judge has found that the case before him is typical. Unless a party contests the Guidelines sentence generally under § 3553(a)— that is argues that the Guidelines reflect an unsound judgment, or, for example, that they do not generally treat certain defendant characteristics in the proper way — or argues for departure, the judge normally need say no more.... Where the defendant or prosecutor presents nonfrivolous reasons for imposing a different sentence, however, the judge will normally go further and explain why he has rejected those arguments. Sometimes the circumstances will call for a brief explana *197 tion; sometimes they will call for a lengthier explanation. Where the judge imposes a sentence outside the Guidelines, the judge will explain why he has done so.... By articulating reasons, even if brief, the sentencing judge not only assures reviewing courts (and the public) that the sentencing process is a reasoned process but also helps that process evolve. 127 S.Ct. at 2468-69 (citations omitted). [6] As an example of the challenge, there are somewhat mixed messages that can be drawn from Gall. On the one hand, we are told that proportionality between the extent of a variance and the extent of the justification for the variance is not required, 128 S.Ct. at 595 (rejecting an approach “that uses the percentage of a departure as the standard for determining the strength of the justifications required for a specific sentence”), while, on the other hand, we are advised that a major variance "should be supported by a more significant justification than a minor one,” id. at 597. [7] Note, however, that at other points the Guidelines equate public and private harm, as in § 3B1.3, which directs that an offense level be increased by two points if a defendant has abused a position of either public or private trust. Significantly, the District Court in this *201 case applied the increase called for in § 3B1.3, although the reason it gave was that "the defendant abused a special skill to facilitate the commission or concealment of the instant offense." (App. at 10.) We tend to agree with the conclusion in the Presentence Report that the increase is also applicable because Levinson abused a position of private trust. See Presentence Report at 12 (“[T]his enhancement is applicable to the offense, because Mr. Levinson occupied a fiduciary position in CoolerSmart, and used that position to commit and conceal the offense."). [8] Kimbrough made it clear that district courts may weigh the applicability of general Guidelines policies as part of the individualized sentencing determination made to satisfy § 3553(a)’s requirement that the sentence for a specific defendant be "sufficient, but not greater than necessary,” to comply with the statute's aims. 128 S.Ct. at 570. However, nothing in Kimbrough or in our own jurisprudence leaves a district court free to state its own general sentencing policies in contravention of the Guidelines. Cf. United States v. Ricks, 494 F.3d 394, 403 (3d Cir.2007) ("[A] district court may ... view the sentencing disparity [imposed by the crack vs. powder cocaine Guidelines] as too vast. However, it must do so as applied to the particular defendant that appears before the court. In terms of sentencing process, a court must give its reasons for why it views the ratio as too harsh when applied to the defendant.”).