concurring:
I join Judge Sack’s opinion in full. I write separately to comment on our decision with respect to Stewart’s sentence. In doing so, I do not wish to express any disagreement with the majority opinion, which has withheld a final judgment on various aspects of Stewart’s sentence in light of the decision to remand. Where I express views on which the majority opinion is silent, I write only for myself.
When a judge of extraordinary ability and a well-earned reputation for exceptional judgment has spent as much time on a case like this, making many decisions of tremendous difficulty — which on review we are all in agreement were correct and wisely done — even the usual deference that the Supreme Court indicated should be paid to the district court’s sentencing must constitute an understatement. This is especially so with respect to any notions as to the substantive reasonableness of a sentence. We may find that there are some procedural or technical matters that warrant more consideration by the district judge, such as the question of whether Stewart committed perjury and the relevance of this in determining an appropriate sentence, given the district court’s treatment of the terrorism enhancement. But for us — who have not been involved in the case and do not know all the backs and forms, some of which may even be best left not fully articulated — to second guess the district court’s judgment seems to me to be precisely what both the Supreme Court and our court sitting en banc in United States v. Cavera, 550 F.3d 180, 194 (2d Cir.2008), have said we should not do. I join the majority opinion because I understand it to avoid second guessing, and because I believe it wisely provides the district court with an opportunity to explain further and perhaps to modify the sentence it has imposed. I write here to explore some of the principal issues on which this panel is divided and to give fuller expression to the importance of having appellate courts appreciate the limited, though still important, institutional role we play, particularly in a case such as this one, where the temptation to go beyond that role is so great.1
*153I.
After calculating Stewart’s sentence under the Guidelines, the district court applied the § 3553(a) factors and concluded that the effect of the terrorism enhancement, “while correct under the guidelines, would result in an unreasonable result” in this particular case. Sent’g Tr. 114. The district court arrived at this conclusion based on, inter alia, (1) what it described as the somewhat atypical nature of Stewart’s case for the imposition of the terrorism enhancement, and (2) the lack of evidence that any victim was harmed as a result of the charged offense.2 Id. at 113. While I agree with the majority opinion that these grounds do not render the terrorism enhancement inapplicable in determining the relevant Guidelines range, I do believe that each ground, if properly articulated, is, as a procedural matter, within the district court’s discretion to consider in its application of the § 3553(a) factors. What is more, I would be extremely reluctant to disturb a district court’s careful effort to look to the unique circumstances of a defendant like Stewart and arrive at an individualized sentence where the Guidelines recommendation is controlled by an undeniably broad enhancement (or reduction) like the terrorism enhancement.
A.
When the terrorism enhancement is applied, it has dramatic consequences on the applicable Guidelines range because it automatically increases both the offense level of a crime and the defendant’s Criminal History Category. In Stewart’s case, for example, the recommended sentence range without the enhancement was 78 to 97 months, while imposition of the enhancement resulted in a recommended sentence of 360 months, the statutory maximum. *154Yet as both the majority opinion and Judge Walker, in his partial dissent, recognize, the terrorism enhancement casts a very broad net. In this case, that breadth was compounded by the fact that the “federal crime of terrorism” for which Stewart was convicted, i.e. the provision of material support, itself covers a wide range of conduct of varying degrees of culpability ranging from the supply of lodging to the contribution of “weapons, lethal substances, [and] explosives.” See 18 U.S.C. § 2389A. When a Guidelines recommendation has such dramatic consequences and yet covers a multitude of sins, unusually broad sentencing discretion in the district court is essential. Indeed, it must be so to comply with the Supreme Court’s remedial holding in United States v. Booker, 543 U.S. 220, 244, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005).
We articulated this precise point in Cavera, recognizing that “some Guidelines enhancements and reductions apply without modulation to a wide range of conduct.” 550 F.3d at 192. We identified as examples several financial crimes where the recommended sentence under the Guidelines varies dramatically according to the money involved, but the culpability of an individual defendant might not be captured accurately by a single variable like financial impact. Id. Perhaps more closely analogous to the terrorism enhancement, we pointed to the Armed Career Criminal Guidelines, under which the recommended sentences for firearms offenses increase sharply if the defendant has a prior conviction for a “crime of violence” — an expansive term that places crimes like attempted burglary of a dwelling under the same umbrella as crimes like murder and rape. See id. (citing U.S.S.G. § 2K2.1(a)). The terrorism enhancement in its breadth of coverage is akin to these examples from Cavera. The majority opinion understands this when it says that the terrorism enhancement “may apply to persons who are culpable in substantially different degrees,” and that the district court “may differentiate between different levels of culpable conduct that nonetheless trigger the same substantial enhancement.” Maj. Op. at 148.
We indicated in Cavera that when a district court faces such over and under-inclusive Guidelines recommendations and when, after considering the § 3553(a) factors, it promulgates a sentence that varies from that recommendation, the district court’s decision, “if adequately explained, should be reviewed especially deferentially.” 550 F.3d at 192 (emphasis added). Nothing we hold here should be understood to conflict with that principle, or to cast doubt on the district court’s apparent conclusion that there must be much room for discretion under the terrorism enhancement.3
*155While we have raised questions about the closeness of Yousry and Stewart’s sentences in light of Stewart’s seemingly greater level of responsibility and status as a lawyer, and while we would benefit from greater explanation by the district court on this and other issues, it remains the district court’s task, for purposes of sentencing under § 3553(a), to assess where Stewart’s criminal behavior falls within the spectrum of terrorism enhancement culpability. And I would be very reluctant— when and if I had to review a sentence in this case for substantive reasonableness— to find an abuse of discretion in a thoroughly-explained conclusion by the district court that Stewart’s conduct, though undeniably serious, was significantly less serious than that of other defendants subject to the terrorism enhancement.
B.
I am more ambivalent about the degree to which absence of harm is a valid ground on which to mitigate a sentence. I join the majority opinion in withholding judgment on this issue with respect to Stewart. My general view, however, is that while a district court ought to be careful about giving too much weight to a factor like harm that might vary based on events beyond the defendant’s control, we should not preclude a district court from giving lack of harm some weight, even for some crimes of terrorism.
Whether it is fair to assign different levels of culpability in criminal sentencing to the same criminal conduct based on the fortuity of whether harm results has long been a contested question in Anglo-American jurisprudence. See H.L.A. Hast, The Concept of Law 131 (1968) (“Why should the accidental fact that an intended harmful outcome has not occurred be ground for punishing less a criminal who may be equally dangerous and equally wicked?”). But whatever significance the consequences of a defendant’s actions ought to have, it is an inevitable part of human nature — and our law — that we as a society do give consequences considerable weight when we mete out punishment and blame.4 This is deeply entrenched in our legal system. The majority opinion identifies the law of attempts as one generally accepted instantiation of this tendency, Maj. Op. at 139-40, but there are many others — such as crimes of culpable risk creation, like vehicular homicide. And while it is true that material support to terrorism is a complete crime rather than an inchoate one, and so fully punishable even if no further harm results, it simply does not follow that the amount of punishment may not at least in part depend on the harm that occurred. The level of punishment for a completed crime varies all the time based on the amount of harm that has occurred, and the Guidelines themselves often directly embrace such a policy.5
*156Judge Walker suggests terrorism is different, and that, at least in the “very broad heartland of cases,” it is procedural error for the district court to consider absence of harm as relevant to the application of the terrorism enhancement and ultimately to the imposition of a sentence.6 He argues that treating harm as a consideration relevant to sentencing those convicted of terrorism related crimes, such as for material support, would effectively nullify the policy considerations of Congress and the Sentencing Commission, who have not made reductions for lack of harm part of the terrorism enhancement. Even if the Guidelines do not themselves make lack of harm relevant for the application of the terrorism enhancement — and they fail to do so only in the narrow sense that the enhancement does not positively reflect the existence of injury — the Supreme Court has made clear that a district court, which has “greater familiarity with [ ] the individual case and the individual defendant,” may properly decide that sentencing judgments made by the Guidelines fail properly to reflect the § 3553(a) considerations. See Rita v. United States, 551 U.S. 338, 351, 127 S.Ct. 2456, 168 L.Ed.2d 203 (2007).
What is more, the Court has evidenced profound skepticism toward arguments that certain policy judgments, which require departing from the Guidelines, have implicitly been taken off the table as a *157result of congressional silence or inaction. See Kimbrough v. United States, 552 U.S. 85, 128 S.Ct. 558, 570-73, 169 L.Ed.2d 481 (2007). As the Court explained, it is usually inappropriate to draw inferences from congressional silence on sentencing practices because Congress has shown that, when it wants to, it knows how to direct levels of sentencing in express terms. Id. at 571 (citing 28 U.S.C. § 994(h), which required Sentencing Commission to set Guidelines sentences for recidivist offenders at or near the statutory maximum). As a result, the fact that Congress increased the statutory maximum in 2001 for material support convictions that caused death, see Op. of J. Walker at 175, and did so without saying anything whatever about how a district court may treat harm when issuing a sentence that is less than the applicable maximum, cannot be read to diminish the discretion the district court otherwise has under § 3553(a).7
I also rather doubt that we would be willing to apply consistently a principle that harm is irrelevant in terrorism cases. Consider two potential defendants, both of whom provide funds to different terrorists and have the requisite mens rea to support a conviction under 18 U.S.C. § 2339A. In one case, the terrorist whom the defendant funds attempts to detonate an explosive in a public place, but the explosives do not go off properly and so cause only a few injuries and one death that results, in part also, from inadequate medical treatment. In the second case, the terrorist’s detonation attempt succeeds, blowing up a city bus and causing a major traffic collision that kills or injures hundreds of people. While reasonable minds might differ as to the weight the level of harm should be given in this scenario, surely we would not hold it procedurally unreasonable for the district court to take the amount of harm into account when sentencing the respective defendants and — while sentencing the first severely — use it as a reason to give the second defendant a much greater sentence. Yet if that is so, the converse must hold, because “sentencing discretion is like an elevator in that it must run in both directions.” Cavern, 550 F.3d at 194. To concede, as I think we must, that when hundreds of people are injured or killed rather than just one a district court may take the amount of harm into account and impose a higher sentence, but then to deny the court that same discretion to reach a lower sentence when, through fortuity, no harm results, would manifestly contravene that principle.8
*158II.
I have focused to this point on factors the district court may consider procedurally when sentencing, but this case also illustrates the importance of our adherence as an appellate court to the right procedure for review of district court sentences. In Cavera, we explained that our review involves a two-step process. We first ensure that the district court has not committed procedural error, and only later engage in substantive review to examine whether the district court has rendered a sentence that is one of the “exceptional cases where the trial court’s decision cannot be located within the range of permissible decisions.” 550 F.3d at 189 (internal quotations and emphasis omitted). Following this sequence of review is central to our limited role. Because district courts “have an institutional advantage over appellate courts” when making sentencing decisions, Koon v. United States, 518 U.S. 81, 98, 116 S.Ct. 2035, 135 L.Ed.2d 392 (1996), our job as a reviewing court is mainly to ensure that a district court’s 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. Evaluating the substantive reasonableness of a sentence before we have found it free from procedural defect risks the substitution of our “considered judgment” for that of the district court.
When we identify procedural error, we have recognized the desirability of remanding to the district court to let it correct its mistake and “exercise its discretion anew,” rather than proceeding prematurely to review the sentence for substantive reasonableness. See Cavera, 550 F.3d at 190. While we have not held that this course must necessarily be followed in all instances, a review of our cases makes clear that it is the ordinary and much preferred remedy. See, e.g., United States v. Williams, 558 F.3d 166, 176 (2d Cir.2009) (“In light of our decision to remand, we reject, as premature, [the defendant’s] challenge to the substantive reasonability of his sentence.”); United States v. Williams, 524 F.3d 209, 215-17 (2d Cir.2008) (vacating sentence without reaching issue of whether sentence was substantively reasonable “because we conclude that the district judge committed procedural error”); United States v. Wills, 476 F.3d 103, 111 n. 6 (2d Cir.2007) (“Because we hold that the sentence is procedurally unreasonable, we do not reach the government’s argument that if [the district court had not relied on procedurally erroneous factors,] [Defendant’s] sentence should be deemed unreasonable based on its length alone.”) (emphasis added), abrogated on other grounds by Kimbrough, 128 S.Ct. at 574-75, as recognized in Cavera, 550 F.3d at 191.
Other circuits have expressed a similar preference, and have sometimes stated it in even more categorical terms: first remand to allow a district court to correct procedural errors and only later review for substantive reasonableness.9 This is no formalism. It is consonant with our system of sentencing, which asks the district *159court to reach a complete judgment about the appropriate sentence in light of the factors enumerated in § 3553(a). Procedural errors prevent the district court from properly arriving at such a holistic judgment. And in light of our obligation to defer significantly to a procedurally correct sentence, we should almost always wait until we have such a sentence to review before evaluating overall reasonableness.
Judge Walker argues that judicial efficiency compels us to identify substantive error at the same time we remand for procedural error. No one wants unnecessary appeals. But I fail to see any systemic advantages from issuing advisory opinions on the reasonableness of sentences that we are vacating, and hence that are now no longer really before us. The decision to remand for procedural unreasonableness presupposes that we believe the district court might not have imposed the same sentence but for the procedural defect. See, e.g., United States v. Jass, 569 F.3d 47, 68 (2d Cir.2009) (recognizing that procedural errors in sentencing are subject to harmless error rule). When we remand for procedural flaws, it may seem inviting to comment at the same time on other factors relied upon by the district court that, though procedurally proper, do not, in our judgment, bear the weight the district court attributed to them. It may even be appropriate in dicta to suggest our preliminary view on the matter. But it is not our role to weigh the individual § 3553(a) factors ourselves and, in a holding, to give binding advice to the district court based on our views. See Gall, 128 S.Ct. at 602.
Let me repeat: Substantive reasonableness calls for review of the overall sentence, not individual factors. And § 3553(a) calls for balancing all the relevant factors. As a result, any non-harmless procedural errors necessarily affect the sentencing calculus. Our task is to evaluate (deferentially) the district court’s finished product, and not to grade its individual ingredients. To do this properly, we must almost always wait for the district court to correct the procedural errors that we have identified and to explain its sentence, and only then to consider the substantive reasonableness of the sentence.
I do not mean to suggest an absolutist view on this matter. It may be that there are cases where there is procedural error that — even though it is not harmless — is sufficiently separate from any serious substantive concerns that we have, so as to make it plausible to address the two issues at once. I cannot think of any such cases offhand, however. And, for the reasons well expressed by the majority opinion in its decision to remand, this case is not one of them.
III.
Another, perhaps uncomfortable, issue deserves discussion. Stewart does not appear to have been the only member of Abdel Rahman’s legal team both to agree to abide by the SAMs imposed upon Rah-man and then subsequently to violate them. One of Rahman’s lawyers, Abdeen Jabara, read to Rahman newspaper articles and letters from followers, while another, Ramsey Clark, the former Attorney General of the United States, acknowledged issuing a statement to the media on behalf of Rahman.10 Yet neither Jabara nor Clark was prosecuted for these apparent violations. This does not mean that *160the Government’s decision to prosecute only Stewart was invidious or improper, and I join the majority opinion in rejecting Stewart’s claim of selective prosecution. But though Stewart’s selective prosecution challenge fails, it does not follow that the alleged misconduct of Jabara and Clark— whom the district court may well have decided shared in certain respects the culpable behavior for which Stewart was convicted — is entirely irrelevant to Stewart and to her sentence. I think it possible that it is relevant, and I believe that usually only the district court is positioned to evaluate that relevance.
As to claims of selective prosecution, we have properly recognized that our scope of review is limited, for “the decision as to whether to prosecute generally rests within the broad discretion of the prosecutor.” United States v. Alameh, 341 F.3d 167, 173 (2d Cir.2003). As the Supreme Court has explained, this is because “the decision to prosecute is particularly ill-suited to judicial review.” Wayte v. United States, 470 U.S. 598, 607, 105 S.Ct. 1524, 84 L.Ed.2d 547 (1985). Yet this does not mean that unfettered (even when it is non-invidious) prosecutorial discretion over who gets charged, and for what, is categorically desirable. Quite the contrary: while prosecutorial discretion may be salutary in a wide variety of cases, when left entirely without any controls it will concentrate too much power in a single set of government actors, and they, moreover, may on occasion be subject to political pressure. The result may well be to produce disparities in the way similarly situated people are treated, disparities that our complex, Guidelines-with-district-court-discretion, system has sought to minimize. The district court’s exercise of its sentencing discretion may provide the only effective way to control and diminish unjustified disparities, without operating in the blunt fashion of selective prosecution judicial review. It may reduce improper differences in treatment, without impinging on the executive’s obligation to enforce the law.
Our case law provides support for this approach. For instance, while we have not required a district court to consider sentencing disparities among co-defendants, we have held that district courts are permitted to do so. As we explained, it is “appropriate for a district court, relying on its unique knowledge of the totality of circumstances of a crime and its participants, to impose a sentence that would better reflect the extent to which the participants in a crime are similarly (or dis-similarly) situated and tailor the sentences accordingly.” Wills, 476 F.3d at 110 (internal emphasis omitted). This, as has been recognized, allows district courts to provide a check against certain otherwise unbounded prosecutorial decisions, as to what charges to bring and as to whether to make substantial assistance motions with respect to only some defendants.11 It is not much of an extension to permit the *161district courts to exercise analogous supervision over those decisions as to which prosecutors enjoy the greatest discretion and that result in the greatest disparities: the decisions on whether to bring any charges at all.
There are, of course, many reasons for prosecutors to fail to bring charges or to bring lesser charges than they could have, and some of these reasons are clearly irrelevant to the proper sentence of the person who has been charged and convicted. But other reasons may be relevant because they may suggest arbitrariness and can lead to abuse — such as the political clout of some potential defendants as against others. We as appellate judges are ill-suited to distinguish between relevant and irrelevant reasons in any given case. The same cannot be said, however, for a district court judge who has presided over a whole trial in which the behavior of uncharged or undercharged parties was part and parcel of the discussion.
This does not mean that when a district court issues a sentence that it should articulate its reliance on the prosecution’s decision not to charge (or to undercharge) other parties. I am not sure. While we generally ask a district court to explain the reasons behind its sentence, and indeed indicated in Cavern that “a district court errs if it fails adequately to explain its chosen sentence, and must include an explanation for any deviation from the Guidelines range,” 550 F.3d at 190 (internal quotations omitted), there are some things that are perhaps best left unsaid by the district court, even though their potential relevance is apparent on the record.12 A prosecutor’s decision to charge only some parties rather than others may be one of these things. The district court might quite reasonably want to avoid impugning the reputation of parties who have not been indicted or tried, and hence have not had the opportunity to assert innocence in open court. But that does not mean that a judge who has presided over a full, trial may not have valid reasons to conclude that the failure to charge some potential co-defendants affects, under the § 3553 factors, the propriety of a sentence.13
I am inclined to think that the district court should not be barred from considering the relevance of prosecutorial discretion in a particular case, and that our legal system should take advantage of the district court’s unique position to consider a defendant’s sentence “in its complete relevant context,” Wills, 476 F.3d at 110 (emphasis added). As appellate courts we should therefore also keep this issue in mind when we review a district court’s sentence, and recognize it as a further reason to defer to a district court’s sentencing judgments. In particular, though we may properly ask the district court to explain apparent sentencing anomalies among convicted defendants as both the majority opinion and the partial dissent do here, we should not forget that there might be even greater disparities between *162a defendant and other individuals who were not charged at all.14
IV.
Finally, I would be remiss if I did not follow the majority opinion in observing the fact that all of the acts for which Stewart was convicted occurred before the attacks of September 11.2001, an event that illustrates in particularly excruciating fashion that results do matter to us. It does not diminish the gravity of Stewart’s crimes to take judicial notice of their timing, and to recognize that our attitudes about her conduct have inevitably been influenced by the tragedy of that day. To suggest otherwise, and to ignore that 9/11 has profoundly influenced our retrospective assessment of the culpability of certain actions related to terrorists and terrorist organizations, would be to ignore reality. As the majority opinion says, Stewart herself might well have viewed her actions differently after 9/11 when the dangerousness of terrorism became so palpable, so stark, and — -most important — so proximate. See Maj. Op. at 147-48. We must be careful then in judging Stewart based on lessons that we learned only after her — • very serious — crimes were committed.15 *163In the end, this factor too is part of the district court’s obligation to consider “the nature and circumstances of the offense,” and to impose a sentence that “reflect[s] the seriousness of the offense” and “provide[s] just punishment.” See 18 U.S.C. § 3553(a)(l)-(2).
. Judge Walker, in his partial dissent, suggests that I believe some judges are infallible and beyond reproach. Far from it; were that my view, I could not join the majority opinion in remanding Stewart’s, sentence. Indeed, I agree completely with the implication in Judge Walker’s statement that all judges, including distinguished appellate ones, can become so preoccupied with a case as to lose perspective. My point is simply that before we assume that a very able district judge has erred substantively in sentencing, we should be especially cautious both in our language and in our judgments, and if we have doubts we should give that judge every opportunity to explain the sentence imposed.
. The district court also found that the terrorism enhancement's increase of Stewart's Criminal History Category from I to VI — the result of the enhancement's "horizontal” component — was “dramatically unreasonable in [her] case” because it "overstate^] the seriousness of [her] past conduct and the likelihood that [she would] repeat the offense.” Sent'g Tr. 113. Because no member of this panel suggests this determination was procedural error, I do not focus on it. It is worth keeping in mind, however, that multiple considerations informed the district court’s view that the terrorism enhancement — which "produced] a guideline range about quadruple the range without that enhancement,” id. at 114 — resulted in a sentencing range for Stewart that was too high. For that reason, some of Judge Walker's contentions — such as that the majority opinion, and my concurrence especially, "erroneously permit[] the district court to eliminate the enhancement altogether primarily because [harm did not result],” Op. of J. Walker at 175, or that the district court "[r]emov[ed]” Stewart from the "terrorism spectrum” because of its view that her crime was atypical, id. at 174 — strike me as misleading. Furthermore, in doing this, and elsewhere in his opinion, Judge Walker seems to label as procedural errors decisions that are normally considered substantive judgments about the amount of weight a particular factor can bear — -judgments that we are to review under a deferential abuse-of-discretion standard after taking into the account “the totality of the circumstances.” See Gall v. United States, 552 U.S. 38, 128 S.Ct. 586, 597, 169 L.Ed.2d 445 (2007).
. Lei me be absolutely clear. Neither the majority opinion nor my concurrence suggests that district courts have greater discretion exclusively for sentencing terrorism defendants. In fact, I believe precisely the opposite. My point is that the terrorism enhancement is, like the examples identified in Cavera, one instance of a Guidelines enhancement that applies "without modulation" to a wide range of conduct, and that we should be especially deferential to a district court’s reasoned decision to vary from such a broad Guidelines recommendation. See Cavera, 550 F.3d at 192.
In contrast, Judge Walker does appear to advocate a separate sentencing jurisprudence for terrorism cases. Judge Walker repeatedly seeks to distinguish terrorism-related crimes from other crimes, and even suggests that, as to harm in terrorism crimes, the sentencing reviews of appellate courts should run in one direction only. See Op. of J. Walker at 175-76. He also states that the wide variety of conduct encompassed by "terrorism support,” unlike the wide variety of conduct covered by other crimes, does not give rise to added district court discretion in sentencing. *155See id. at 174. He attributes this to congressional decisions. With great respect, I have difficulty reading what Congress has mandated as creating these differences. Judge Walker clearly views terrorism-related crimes as in a different category from all other very serious felonies — and he may well be right. It is, however, an error — though a common one among all judges — to attribute one’s own heartfelt and perhaps correct views to the legislators.
. See Sanford H. Kadish, The Criminal Law and the Luck of the Draw, 84 J. Crim. L & Criminology 679, 688 (1994) (“While in principle it's difficult to find good reasons for making desert turn on chance, here’s the rub: most of us do in fact make judgments precisely of this kind.”). See generally Paul H. Robinson & John Darley, Justice, Liability and Blame: Community views and the Criminal Law (1995) (presenting studies suggesting public judgments about criminal culpability turn significantly on the level of harm that results from an action).
. Judge Walker identifies several examples in his opinion, though he reaches a different conclusion about their import. See Op. of J. Walker at 175 n.ll (identifying as examples the increase of the offense level for conspiracy or solicitation to commit murder if the offense results in death, U.S.S.G. § 2A1.5(c)(1), and the increase of the offense level for aggravated assault based on victim’s injuries, Id. § 2A2.2(b)(3)). There are many other examples. See Kate Stith, The Arc of the Pendulum: Judges, Prosecutors, and the Exercise of Discretion, 117 Yale L.J. 1420, 1476 (2008) ("Guidelines sentences for both drug crimes and financial crimes depend heavily on the quantity of harm found by the sentencing judge----’’).
. Though Judge Walker suggests otherwise, it is not at all unprecedented for a district court to consider lack of harm relevant to sentencing in a terrorism case. Indeed, in a case that Judge Walker cites, the Eleventh Circuit affirmed a district court decision that did just that. See United States v. Garey, 546 F.3d 1359, 1363-64 (11th Cir.2008) (per curiam). In that case, the district court found that the terrorism enhancement applied, but then granted a downward variance based in part on the fact that the defendant had not carried out any violent acts at the time of his apprehension. See United States v. Garey, 383 F.Supp.2d 1374, 1379 (M.D.Ga.2005) ("It is ... troubling that another defendant who carried out a threat to bomb public facilities, injuring and maiming (but not killing) thousands of people, would face the same sentence as this Defendant who did not cause physical injury to a single person."). In upholding the defendant’s sentence as not unreasonable, the Eleventh Circuit specifically noted that the district court had already considered the defendant’s arguments about the lack of actual harm and, on the basis of the § 3553(a) factors, imposed a reasonable sentence below the advisory Guidelines range. See Garey, 546 F.3d at 1364.
Judge Walker cites Garey, and other cases, for the proposition that courts apply the terrorism enhancement even in the absence of harm. Op. of J. Walker at 176 n. 13. That is true, but it is not relevant to the precise issue we face because the district court here asserted that the enhancement applied to Stewart under the Guidelines, and the majority opinion has stated clearly that any decision to the contrary would be error. Maj. Op. at 150-51. It is also incorrect to say that, for all practical purposes, the district court used lack of harm to nullify the sentencing enhancement, because that was not the only consideration relied upon by the district court to support its variance. See supra note 2. The real question is whether, as a procedural matter, lack of harm can support some downward variance in a terrorism case. To this narrow question, *157other courts have said, or at least strongly implied, that the answer is yes.
. Indeed, that Congress saw fit to increase the maximum sentence for material support based solely on whether death results can easily be understood to suggest that Congress thought amount of harm does matter in this context, even if, at times, that harm is largely fortuitous.
. I by no means suggest that a district court may not err by attributing too much importance to results, whether on the upside or the downside. When reviewing a sentence for substantive reasonableness, appellate courts may question whether the presence or absence of actual harm can "bear the weight” attributed to it by a district court. Cavera, 550 F.3d at 191. And while an appellate court must never forget that its review is deferential, see Gall, 128 S.Ct. at 596, it may need to look carefully at a district court’s reasons for giving weight to results, especially given the natural tendency to overvalue consequences. The majority opinion appropriately treats this substantive question as analytically distinct from whether a district court may, as a procedural matter, rely on lack of harm. See Maj. Op. at 139-40 n. 33. The majority opinion also quite properly leaves open, at this time, the question of whether the district court’s actual use of lack of harm as grounds upon which to mitigate Stewart's sentence led to a result that was erroneous. Like the majority opinion, I do not yet express any view as to whether the district court in this case placed too much weight on the ap*158parent lack of harm that resulted from Stewart’s actions. See Part II infra.
. See, e.g., United States v. Delgado-Martinez, 564 F.3d 750, 753 (5th Cir.2009) (holding that where procedural error is identified, "we must remand” unless error is harmless, and that a court should not move on to Gall's second step to review substantive reasonableness of sentence if there is non-harmless procedural error); United States v. Grissom, 525 F.3d 691, 696 (9th Cir.2008) (“[W]e will remand non-harmless procedural errors ... and only proceed to review the substantive reasonableness of procedurally sound sentences.”) (emphasis added).
. See Letter in Support of Lynne Stewart from Abdeen M. Jabara to Judge Koeltl (June 28, 2006), and Letter in Support of Lynne Stewart from Ramsey Clark to Judge Koeltl (Oct. 2, 2006), available in J.App. at 2211, 2334.
. See Kate Stith & José A. Cabranes, Fear of Judging: Sentencing Guidelines in the Federal Courts 140-42 (1998) (articulating concern, pre-Booker, that “exercise of broad prosecutorial authority over sentencing within a system that severely limits the sentencing discretion of federal judges means that the power of prosecutors is not subject to the traditional checks and balances that help prevent abuse of that power" and advocating system where judges have the "countervailing discretionary authority to restrain prosecutorial power”) (emphasis in original); Ryan Scott Reynolds, Note, Equal Justice under the Law: Post-Booker, Should Federal Judges Be Able to Depart from the Federal Sentencing Guidelines to Remedy Disparity Between Codefendant’s Sentences?, 109 Colum. L.Rev. 538, 564-66 (2009) (contending that the emerging rule among circuit courts that allows district courts to consider co-defendant disparity is desirable because it mitigates negative effects of excessive prosecutorial power over sentencing).
. Cf. Guido Calabresi, A Common Law for the Age of Statutes 172-81 (1982) (discussing tradeoffs between open recognition of a doctrine and greater opacity, where the latter might be useful in preventing a doctrine’s abuse).
. Significantly, our system of sentencing allows the district court to consider uncharged conduct by the defendant so as to ensure that the sentence is based upon "the real conduct that underlies the crime of conviction.” Booker, 543 U.S. at 250, 125 S.Ct. 738 (Remedial Op., Breyer, /.). One of the principal reasons for this, as expressed by the Court, is to prevent prosecutors, when they make charging decisions, from "exercis[ing] a power the Sentencing Act vested in judges.” Id. at 257, 125 S.Ct. 738.
. At footnote 20 post, Judge Walker describes as academic my discussion of the possible relation between the sentence imposed on the defendant before us, and the absence of any charges brought against the other lawyers originally involved in this case — lawyers, some of whose acts could surely be described in language as powerful as that used by Judge Walker with respect to the defendant’s misdeeds. Whatever may be said of my views, they are not academic. They reflect instead the very practical consequences of: (a) the necessary absence of any judicial control over prosecutorial decisions as to whom to charge, decisions which, as Judge Walker rightly notes, are "exclusive and absolute” and "insulated from judicial review,” Op. of J. Walker at 184 n. 20; (b) the inevitable tendency of human beings, and hence also of "prosecutors acting in good faith,” id., to take easy rather than difficult (and possibly politically costly) actions. (It is perhaps worth noting again in this regard that one of the uncharged lawyers in this case was formerly the chief prosecutor of the United States, while the defendant is described by the district court as having “represented the poor, the disadvantaged and the unpopular,” Sent'g Tr. 115); (c) the fact that without "initiating [any] inquisitorial foray into the prosecutor's office,” Op. of J. Walker at 184 n. 20, a district court can get a pretty good read on the likely relative behavior of those involved in the complex set of actions that led to a particular trial, including those participants not charged; and (d) that the district court is given by law the direct task of determining what is "just punishment” and, in doing so, to "avoid unwarranted ... disparities,” 18 U.S.C. § 3553(a).
. On various occasions Judge Walker, in his partial dissent, makes reference to much higher sentences given by other federal courts to those who have aided terrorists, including in cases in which no harm occurred. See Op. of J. Walker at 166 n. 4, 176 n. 13. It is no small matter, however, that the overwhelming majority of the cases Judge Walker cites involved post-9/11 actions by the defendants. See, e.g., United States v. Khan, 309 F.Supp.2d 789, 796 (E.D.Va.2004) (describing activities of Randall Royer and co-defendants and indicating "[t]he indictment alleges that ... preparations culminated in [co-defendants and their co-conspirators] attending a terrorist and jihad training camp after September 11, 2001, with the intent to proceed to Afghanistan and fight for the Taliban and Al-Qaeda against United States troops”).
This is not the only difference between Stewart and the defendants sentenced for material support crimes in the cases Judge Walker mentions — leaving aside one case, relied on by Judge Walker, where the convictions were subsequently vacated on appeal. See United States v. Al-Moayad, 545 F.3d 139 (2d Cir. 2008) (vacating convictions of Al-Moayad and his co-defendant Mohsen Zayed due to serious evidentiary errors). Indeed, in one instance where the district court imposed a 180-month sentence, the court explicitly found that, for a variety of reasons, the defendant was not similarly situated to Stewart. See United States v. Aref, NO.04-CR-402, 2007 WL 804814, at *7 (N.D.N.Y. Mar. 14, 2007).