LIPEZ, Circuit Judge.

The United States challenges the sentences imposed on appellees Robert Prosperi and Gregory Stevenson after their conviction of mail fraud, highway project fraud, and conspiracy to defraud the government. Both appellees were employees of Aggregate Industries NE, Inc. ("Aggregate"), a subcontractor that provided concrete for Boston's Central Artery/Tunnel project, popularly known as the "Big Dig." The government charged that over the course of nine years Aggregate knowingly provided concrete that failed to meet project specifications and concealed that failure by creating false documentation purporting to show that the concrete provided complied with the relevant specifications. Several employees of Aggregate, including Prosperi and Stevenson, were convicted of criminal offenses for their roles in the scheme.

At sentencing, the district court calculated the guidelines sentencing range ("GSR") for Prosperi and Stevenson as 87- to 108-months incarceration. Then, explaining fully its rationale for a below-guidelines sentence, the court sentenced Prosperi and Stevenson to six months of home monitoring, three years of probation, and 1,000 hours of community service. The government now appeals, arguing that under Gall v. United States, 552 U.S. 38, 128 S.Ct. 586, 169 L.Ed.2d 445 (2007), the sentences imposed by the district court were substantively unreasonable and that the appellees' crimes warrant incarceration.

We affirm. Although the degree to which the sentences vary from the GSR gives us pause, the district court's explanation ultimately supports the reasonableness of the sentences imposed. The district court emphasized that its finding on the loss amount caused by the crimes, the most significant factor in determining the GSR, was imprecise and did not fairly reflect the defendants' culpability. Hence it would not permit the loss estimate to unduly drive its sentencing decision. Relatedly, it found that there was insufficient evidence to conclude that the defendants' conduct made the Big Dig unsafe in any way or that the defendants profited from the offenses. The court then supplemented these critical findings with consideration of the individual circumstances of the defendants and concluded that probationary sentences were appropriate. We cannot say that it abused its discretion in doing so.

I.

A. The Big Dig's Need for Concrete

Boston's Central Artery/Tunnel project (the "Big Dig"), lasting from 1991 to 2007, was one of the largest public works projects in United States history at the time of its completion. The project entailed replacement of a major elevated highway that passed through central Boston with an underground expressway, as well as the extension of I-90 to Logan Airport. The Massachusetts Highway Department, and later the Massachusetts Turnpike Authority ("MTA"), both agencies of the Commonwealth of Massachusetts (the "Commonwealth"), had primary responsibility for the project. The final cost of the project, approximately $14 billion, was funded jointly by the Commonwealth and the federal government.

Numerous private contractors were hired to help with the construction, and many management responsibilities were delegated to a joint venture between Bechtel Infrastructure and Parsons, Brinkerhoff, Quade & Douglas, Inc. ("B/PB"). These two companies worked as design consultants, performed engineering reviews, and managed much of the construction. Because of the scale and cost of the project, responsibility for construction of various sections was apportioned among different general contractors, and B/PB worked with each in a coordinating role. The general contractors in turn contracted with various sub-contractors. In total, there were approximately 150 individual construction contracts awarded to private companies in connection with the Big Dig. Each of the sub-contractors was bound by the contract specifications and schedules that were set out in general contracts with the Commonwealth.

The Big Dig required approximately 4.2 million cubic yards of concrete, 60 percent of which was provided by Aggregate. For suppliers of concrete, the construction contracts required that: 1) there be a certain mix design, or recipe, for the concrete, based on the intended use; 2) the supplier have plants with an automatic batching system that ensured the proper mixture of each load, or batch, of concrete; 3) the sub-contractor have in place recorders that captured information regarding the mix design, as well as the date and time of batching for each load of concrete, and provided a printout, called a "batch ticket," containing all of the required information; 4) no additional water be added after the concrete mixture was loaded into trucks for delivery; and 5) in most circumstances, the concrete be in place at the construction site within ninety minutes of the time it was mixed and loaded onto trucks.

The batch tickets were especially important because they served as a quality control mechanism. Aggregate's drivers would give the batch tickets to B/PB inspectors as they delivered their loads, and the tickets were left in a holding area close to the placement so that inspectors or field engineers could check on the characteristics of the concrete being placed. In particular, inspectors needed to know the time that the concrete was loaded, the mix design, the volume loaded, and the amount placed.

B. The Fraudulent Scheme

At some point in the mid-1990s, Aggregate began to supply concrete to the Big Dig that failed to meet contract specifications. Of primary concern was Aggregate's practice of topping-off loads of leftover concrete, sometimes of a different mix design, with fresh concrete meeting contract specifications, and providing the entire load as if it were fresh concrete. Loads including leftover concrete were known as "10/9" loads, which was the radio call signal that drivers would use to let the dispatcher know that they had leftover concrete. Historically, Aggregate had provided 10/9 concrete to buyers on private projects, but it initially refrained from doing so on the Big Dig.

The decision to use 10/9 concrete on the Big Dig was made by the management of Aggregate's Ready-Mix Concrete Division. Prosperi, who was the General Manager of the Division, and Stevenson, who was its Operations Manager, played a major role in making the decision and enforcing the policy. Once this policy was in place, the practice of using 10/9 concrete became widespread at Aggregate and leftover concrete was sent to the Big Dig on a daily basis. On one occasion, after a group of drivers dumped old concrete, Stevenson told them not to do so again and that they would be disciplined if they did. Dispatchers were required to keep logs of each use of leftover concrete and these logs were provided to Prosperi, Stevenson, and others on a daily basis.

Using leftover concrete in this way creates safety concerns because the concrete poured does not match the intended mix design and may set more quickly than planned because of the increased time between batching and placement. When he first learned that 10/9 concrete was being sent to the Big Dig, Gerard McNally, the head of Quality Control for Aggregate's Ready-Mix Concrete Division, raised concerns about the practice, and particularly about using such concrete in structural elements, such as roof pours. Despite his belief that it was improper, McNally agreed, at Prosperi's request, to consult with dispatchers to ensure that 10/9 loads were of a type that could be safely reused on the Big Dig. As McNally put it, "I realized that that's the way it was going to be, and that as far as I was concerned, a better thing for me to do was to get on board with the idea."

Using the 10/9 logs, investigators were able to determine approximately how many loads of leftover concrete Aggregate provided to the Big Dig. From 1999 to 2004, Aggregate provided approximately 2,638 such loads. Although the records collected went back only to 1999, Aggregate's practice of sending leftover concrete to the Big Dig dated from the mid-1990s. Using the available records to estimate the total number of 10/9 loads Aggregate provided to the Big Dig, the government estimated that Aggregate provided approximately 5,300 loads of 10/9 concrete over the life of the project without the knowledge of project supervisors. These loads amounted to approximately 1% of all of the concrete provided by Aggregate to the Big Dig and 0.6% of all the concrete used in the project.

To conceal the 10/9 practice, Aggregate developed a system that employees acknowledged was "designed to trick the inspector." After fresh concrete was loaded with some portion of leftover concrete, Aggregate employees would create a "dummy ticket" showing that a complete fresh load of concrete, meeting project specifications, had been loaded. To create the dummy ticket, the computer running the batching system was put into demonstration mode. Then the quantity and mix design called for by the specifications was manually entered. After the dummy ticket was printed, a copy was given to the driver to present to inspectors at the Big Dig site. This became a regular procedure on the Big Dig. In addition to this method, the clocks on the batching computers would be set ahead to make it look like the load had been batched later than it actually had, giving the drivers more time to get the concrete to its destination.

Occasionally, inspectors would come to Aggregate's plant to ensure that the proper procedures were being followed. When this happened, the batchmen, the Aggregate employees who loaded each truck, would call the dispatchers using the phrase "city plant" to signal that inspectors were present and no 10/9 loads should be sent out.

Additionally, there were instances when Aggregate ran out of fly ash, an important ingredient in some of the mix designs.1 Aggregate supplied concrete to the Big Dig nonetheless, without informing project supervisors of the absence of fly ash. As with the 10/9 loads, Aggregate would falsify batch tickets to make it appear as if the concrete provided contained the requisite fly ash. Although the problem with the supply of fly ash was intermittent, it became so bad at one point that McNally informed Prosperi that Aggregate had two choices: "either run straight cement [without fly ash] or we stop loading." McNally testified that Prosperi responded "[w]e never stop loading." It is unclear how many loads of concrete without required fly ash were provided to the Big Dig.2

The use of non-specification, and especially leftover, concrete was economically beneficial to Aggregate. Most obviously, leftover concrete that was not re-used would have to be dumped or recycled, an expensive process that Aggregate wished to avoid. It was running out of space to store leftover concrete and was unable to develop efficient alternative uses to deal with the quantities of concrete that were leftover. Dumping leftover concrete was also a time consuming process that pulled drivers and trucks away from their delivery obligations. For all of these reasons, Aggregate re-used as much concrete as possible and minimized the amount that was dumped.

C. Sentencing

After this scheme came to light, Prosperi and Stevenson were charged with: 1) one count of conspiracy to commit mail fraud and make false statements on a highway project; 2) one count of conspiracy to defraud the United States by submitting false claims; 3) eighty-three counts of making, and aiding and abetting the making of, a false statement on a highway project; and 4) fifty counts of mail fraud, and aiding and abetting mail fraud. McNally, John Farrar, Marc Blais, and Keith Thomas, all Aggregate employees, were charged with similar offenses.3 McNally and Thomas pled guilty. After a sixteen-day jury trial, Farrar and Blais were convicted on some counts and acquitted on others. Prosperi and Stevenson were convicted on all counts.

1. The Loss Amount Finding

The amount of loss to be used in calculating the GSR was a hotly contested issue at sentencing. The government and the defendants submitted memoranda describing their positions on the issue of loss, and the court held an evidentiary hearing on the issue.

The government argued that pursuant to Application Note 3(F)(v) to § 2B1.1 of the United States Sentencing Guidelines ("USSG"), the amount of loss should be calculated as the total amount paid by the government for the concrete that failed to meet project specifications, with no credit provided for the value of the goods and services actually provided. Thus, the government took the number of 10/9 loads delivered to the Big Dig between 1999 and 2004, 2,637 — a number that was identified in Aggregate's records — and added to it another 2,700 loads, an estimate of the number of 10/9 loads delivered prior to 1999 and on nights and weekends. Finally, the government also included an estimated 1,200 loads of 10/9 concrete sent to other public construction projects within Massachusetts. In total, these loads included approximately 64,163 yards of non-specification concrete. The government asserted that the average price paid by the government per yard of concrete was $80.90. Accordingly, the government argued that the appropriate loss amount for Prosperi and Stevenson was approximately $5.2 million.4

In contrast, Prosperi and Stevenson argued that there was no loss attributable to them, because the Commonwealth got what it contracted for and expected to receive. In particular, they noted that the MTA certified to federal authorities that post-construction testing showed that the materials used in the project conformed with applicable plans and specifications. The MTA did so after the conduct at issue in this case came to light. Additionally, Prosperi and Stevenson pointed to independent testing commissioned by the government that focused on the areas of the Big Dig believed to have received non-specification concrete. These tests showed that, when completed, those sections of the project met or exceeded standards for concrete strength and permeability. Prosperi and Stevenson also noted that, considered as a whole, little of the total amount of concrete sent to the Big Dig failed to meet the relevant specifications.

From their perspective, the proper measure of loss was the cost of repairs necessary to fix or replace inferior work. Given their claim that the finished product met safety standards and required no repairs, they argued that there was no monetary harm. They also pointed out that there was no actual double-billing for concrete, since the government did not pay for concrete by the load, but instead paid by construction unit (e.g., per foot of completed tunnel) without regard to the amount of concrete actually used. Furthermore, they noted that pecuniary harm is not an element of the crimes charged and that the government did not attempt to prove pecuniary harm during trial.

The district court issued a memorandum explaining its decision on the issue of loss. After summarizing the parties' positions, it stated, "Needless to say, much of this is not helpful. In a case like this it is difficult to apply a mechanical rule of sentencing." United States v. Prosperi, No. 06-10116, at 5, 2010 WL 1816346 (D.Mass. May 6, 2010) (Memorandum and Order on Calculation of Loss for Purposes of Sentencing) ("Loss Order"). The district court noted that part of the rationale for using loss amount to determine the GSR was to eliminate disparities between white- and blue-collar offenders: "One of the goals of the Sentencing Guidelines was to give greater equivalence between penalties for white collar crimes like fraud and violent crimes like robbery. One means chosen by the Sentencing Commission to accomplish this goal was by giving greater weight to the amount of loss involved in a scheme to defraud." Id. at 2. Cognizant of this purpose, it explained:

Id. at 5.

The court adopted the government's loss figure of $5.2 million, stating simply that, "[a]lthough neither the government or the defendants' methodology can be termed precise, I think on the whole the government's method of calculating loss is closest to the mark." Id. However, after making this choice, the district court put the parties on notice that it would not allow the loss figure to drive its sentencing decision, stating that "I do not believe the estimated loss figure — given the nature of the case — has pivotal significance in fashioning an appropriate sentence, something the parties might keep in mind in composing their sentencing arguments and recommendations." Id.

The Probation Department shared the court's concern about the significance of the loss amount. After the court issued its Loss Order, the Probation Department revised the Presentence Reports for both Prosperi and Stevenson in light of the court's conclusion. The final paragraph of the Presentence Reports for both Prosperi and Stevenson notes that "[t]he Court may wish to consider whether the loss in this instance is overstated."

2. The Sentences

The district court's determination of the amount of loss attributable to the defendants was pivotal in the calculation of the applicable GSR. See USSG § 2B1.1. Under the guidelines, both Prosperi and Stevenson were subject to a base offense level of 7. Id. § 2B1.1(a)(1). The loss amount of $5.2 million increased the offense level by 18. Id. § 2B1.1(b)(1). Finally, both Prosperi and Stevenson received an additional increase of four levels for being organizers or leaders in the criminal scheme.5 Id. § 3B1.1(a). Accordingly, both were subject to an adjusted offense level of 29. Given the lack of any prior criminal history, the district court calculated the GSR for both Prosperi and Stevenson as 87- to 108-months imprisonment.

At the sentencing hearing, after noting the GSR, the district court emphasized that it did not believe that the GSR, driven largely by the loss estimate, accurately reflected the defendants' culpability: "As far as I am concerned, the presentence report and the Guidelines calculation, which we all recognize are advisory, are influenced by the difficulty of assigning an accurate loss value to the case, which is the critical element around which the Guidelines are structured." The court observed that it settled on a loss figure "[a]s a formal matter," and picked the government's formula because "[it] was probably as good as any." It added that "I would note parenthetically that without the 18-point escalator, one would be looking at an 8- to 14-month range with a Zone C alternative sentence available to the Court if the Guidelines, again, applied in a mandatory fashion."6

Having made these comments, the court asked to hear from the parties. The government described a lack of accountability generally among the companies and individuals who worked on the Big Dig, as well as an alleged lack of remorse on the part of Prosperi.7 It also suggested that there may be long-term maintenance issues with the project because of the defendants' conduct. The government attempted to tie the defendants' conduct to a broader corporate culture of corruption, invoking Bernie Madoff, the BP oil spill, and the financial crisis.8

Prosperi's counsel argued that the loss amount overstated the defendants' culpability. He noted the numerous letters from Prosperi's family, friends, and business associates attesting to Prosperi's good character and role in the community. Additionally, he reminded the court that Prosperi's wife suffered from a terminal cancer and that Prosperi was an important caregiver for her.

Stevenson's counsel reminded the court that tests performed after the scheme came to light indicated that the finished structures ultimately met project specifications. He also argued that there was no profit motive animating Stevenson's conduct. Finally, he called the court's attention to Stevenson's positive community activities, and his care for his disabled daughter and elderly parents. Subsequently, Prosperi and Stevenson each briefly spoke on their own behalf. They each apologized for their actions and asked the court for leniency.

The district court then explained its reasoning in arriving at the sentences imposed. It acknowledged the numerous letters it had received from the defendants' family, friends, and community members, and noted that it found them "sincere, supportive, and, I'm sure, an accurate portrayal of the defendants' lives." It also observed that if any of the letter writers came to Prosperi and Stevenson seeking advice in a similar situation, the two defendants "would have been the first to say, `This is morally wrong, you shouldn't do this, this should not be the choice that you should make.' Why they made that wrong choice for themselves, as I say, is the piece of the puzzle that I find hardest to answer." The court went on to state:

Responding to the government's invocation of the financial crisis, the BP oil spill, and Bernie Madoff, the district court explained:

The court added that "I really cannot sentence a culture. I have to sentence the defendants as human beings, and the real choice in this case is what are the punishment alternatives."

Balancing the 18 U.S.C. § 3553(a) factors, the court noted that, "There is one benefit, and only one, that I see in this case to incarceration, and that is the sanction of deterrence that a sentence [of incarceration] would pose for others." As for the issue of incarceration and specific deterrence, the court added that it saw no risk of recidivism on the part of either Prosperi or Stevenson. It went on to note that "[i]ncarceration will incur a large cost to taxpayers, and an even larger personal cost in Mr. Prosperi's case to his ill wife and, to some degree, to Mr. Stevenson's family, as I recognize that they both play important roles as caregivers and caretakers in their families."

The district court concluded by noting, "I have given perhaps more reflection to this than perhaps any but one or two other sentences I have had to impose, and I have come to the conclusion that an alternative to incarceration is the appropriate sentence in this case." It added that, "I think [the defendants' conduct] was wrong, and I have made it clear that I think it was wrong; but I do think my decision is the correct one given all of the factors that are at play." Accordingly, despite the GSR, neither Prosperi nor Stevenson was sentenced to any term of imprisonment. Rather, each was sentenced to six months home monitoring, three years probation, 1,000 hours of community service, and a modest fine.9

In its Statement of Reasons supporting Prosperi's sentence,10 the district court explained:

The Statement of Reasons for Stevenson was identical, except it substituted, for the line regarding Prosperi's terminally ill wife, a statement that "the sentence imposed will allow the defendant to meet his family obligations."

II.

We review the reasonableness of a sentence, whether inside or outside the guidelines range, "under a deferential abuse-of-discretion standard." Gall, 552 U.S. at 41, 128 S.Ct. 586. There are two parts to this inquiry. First, we ask whether the district court committed a procedural error in selecting a sentence. Id. at 51, 128 S.Ct. 586. This may include "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." Id. If an appellant makes no claim of procedural error, as is the case here, we limit our review to the substantive reasonableness of the sentence. See United States v. Martin, 520 F.3d 87, 92 (1st Cir.2008). "When conducting this review,... [we] take into account the totality of the circumstances, including the extent of any variance from the Guidelines range." Gall, 552 U.S. at 51, 128 S.Ct. 586.

Following the Court's decision in Gall, we have noted that sentencing "necessitates a case-by-case approach, the hallmark of which is flexibility." Martin, 520 F.3d at 91. Accordingly, "a sentencing court should not consider itself constrained by the guidelines to the extent that there are sound, case-specific reasons for deviating from them." Id. However, if a court chooses to impose a sentence outside of the guidelines range, "[t]he court's reasons for deviation should typically be rooted either in the nature and circumstances of the offense or the characteristics of the offender; must add up to a plausible rationale; and must justify a variance of the magnitude in question." Id. From this, it follows that "a major departure should be supported by a more significant justification than a minor one." Gall, 552 U.S. at 50, 128 S.Ct. 586.

Gall emphasizes the "very broad" discretion afforded sentencing courts and the deference accorded their sentencing decisions. United States v. Innarelli, 524 F.3d 286, 292 (1st Cir.2008); see also United States v. Taylor, 532 F.3d 68, 70 (1st Cir. 2008) ("[O]ur review of substantive reasonableness is highly deferential."); Martin, 520 F.3d at 98 ("Under [United States v.] Booker [543 U.S. 220, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005),] and Gall, there is a heavy emphasis on a sentencing court's informed discretion."). This deference is founded on several "institutional advantages" possessed by the district court, including "a superior coign of vantage, greater familiarity with the individual case, the opportunity to see and hear the principals and the testimony at first hand, and the cumulative experience garnered through the sheer number of district court sentencing proceedings that take place day by day." Id. at 92 (citing Gall, 552 U.S. at 50-53, 128 S.Ct. 586). We have explained that "once the GSR is properly calculated, `sentencing becomes a judgment call' for the court, and the court may construct a sentence varying from the GSR `based on a complex of factors whose interplay and precise weight cannot even be precisely described.'" Innarelli, 524 F.3d at 292 (quoting Martin, 520 F.3d at 92).

Thus, "[t]he fact that the appellate court might reasonably have concluded that a different sentence was appropriate is insufficient to justify reversal of the district court," Gall, 552 U.S. at 51, 128 S.Ct. 586, and a district court's "choice of emphasis" when considering relevant factors is not a ground for vacating a sentence, United States v. Zapata, 589 F.3d 475, 488 (1st Cir.2009). Ultimately, "[t]here is no single reasonable sentence in any particular case but, rather, a universe of reasonable outcomes," United States v. Walker, 665 F.3d 212, 234 (1st Cir.2011), and "[w]e generally respect the district court's sentence as long as the court has provided a plausible explanation, and the overall result is defensible," Innarelli, 524 F.3d at 292.

III.

Guided by this legal framework, we address the government's challenges to the substantive reasonableness of the district court's sentences. The heart of the government's argument is its repeated observation that the probationary sentences imposed are an eighty-seven-month (100%) variance from the bottom of the applicable guidelines range, and the related assertion that the court's rationale for such a dramatic variance is not plausible. As noted, the court observed at sentencing that "without the 18-point escalator [due to the loss estimate], one would be looking at an 8- to 14-month range with a Zone C alternative sentence available to the Court if the Guidelines, again, applied in a mandatory fashion." The government vigorously disputes the court's view that the loss amount here was an unfair proxy for culpability. Accordingly, our evaluation of the government's challenge to the sentences turns in large measure on whether we find that the court has offered a plausible explanation for its treatment of the loss amount.

A. The Loss Amount

The court regarded the loss determination set forth in its Loss Order, although required by the guidelines, as an uncertain figure. The government paid for the Big Dig by construction unit (e.g., per foot of completed tunnel). It did not directly pay for concrete by the load. Thus, any estimate of the price paid for the 10/9 concrete provided by Aggregate would be imprecise. Additionally, the MTA certified to federal authorities that the materials used in the construction of the Big Dig conformed with applicable plans and specifications, even after the conduct at issue in this case came to light. This certification calls into question whether there was any actual monetary loss to the government. Most importantly, independent testing commissioned by the government, and focusing on the areas of the project believed to have received non-specification concrete, showed that those sections of the project met or exceeded standards for concrete strength and permeability. None of the testing of samples taken from the Big Dig indicated a need for repairs or specific concerns regarding durability.

Presumably, the court relied on Application Note 3(F)(v) to USSG § 2B1.1 in determining the loss amount. That note provides that, "[i]n a case involving a scheme in which ... goods for which regulatory approval by a government agency was required but not obtained, or was obtained by fraud, loss shall include the amount paid for the property, services or goods..., with no credit provided for the value of those items or services." USSG § 2B1.1 cmt. n.3(F)(v). Strict application of this advisory application note in fashioning a sentence, without further analysis, would treat this case identically to one in which the defendants had provided no usable concrete in the 10/9 loads. While this is the approach that the application note recommends, the court was reluctant to use it in this case, given the MTA's certification that the finished product met project specifications.

It is true that the evidence on the durability issue was mixed. The same expert firm that performed the referenced tests also concluded that Aggregate's practice of using leftover concrete, and providing concrete without fly ash, could potentially have serious durability consequences. At the evidentiary hearing on loss, the government's expert witness stated that, despite the fact that samples taken from the Big Dig passed relevant tests, "I think I feel pretty confident there will be some long-term issues." In particular, he stated that "I do believe there will be long-term premature deterioration," but noted that any such deterioration would be the product of a combination of factors, with the use of non-specification concrete being only one. The expert also acknowledged that the permeability tests that he relied upon to reach this conclusion were not tests of samples that were taken from the Big Dig itself, but instead were efforts to create concrete similar to those samples by replicating the conditions under which 10/9 concrete was used. In the face of this contradictory evidence, the district court concluded that any failure of the project was speculative and unlikely. The government has not challenged this factual finding.

The government presented evidence that Aggregate management and staff received bonuses based on the company's profitability. The fraudulent scheme served to increase the company's profitability, and hence it may have had an effect on year-end bonuses. However, the evidence presented by the government indicated that bonuses were introduced by the company in 1999, at least three years after Aggregate began to send 10/9 concrete to the Big Dig. Accordingly, the prospect of a larger year-end bonus could not have been the motivation to enter into this scheme or to continue it for its first three years. Furthermore, there was no evidence that Prosperi and Stevenson received any bonuses even after the policy was in place, much less larger bonuses due to the fraudulent scheme.

On the basis of this record, the court found that the defendants did not seek to enrich themselves personally and did not personally benefit from the scheme. The court added that these findings distinguished them from typical white-collar defendants. Again, the government has not challenged the court's factual findings on this issue.

In summary, several factors played into the court's decision to treat the loss amount as an unfair proxy for culpability. The loss amount finding itself was necessarily imprecise. In light of the MTA's certification that the finished product met project specifications, there was value in the concrete provided. Testing of samples taken from the Big Dig showed no need for repairs, and any link between the conduct of the defendants and a future failure of the Artery Project was speculative and unlikely. Finally, Prosperi and Stevenson did not personally benefit from the scheme. Given these findings and considerations, the district court offered a plausible explanation of its refusal to allow the loss estimate to control its sentencing determination.

B. Corporate Culture at Aggregate

In its Loss Order, the court noted, "What appears to have been at play was a corporate culture in which pressure, much of it self-generated, was exerted on defendants to perform service for the short-term benefit of the organization without heed to the moral consequences or public harm." Loss Order at 5. Similarly, at sentencing, the district court stated, "[I]t is not clear to me why these defendants were necessarily plucked out to be the `poster children' — if I may use the phrase — for a larger corporate culture that I agree was morally lazy, and so focused on the short term that it became heedless to the consequences or impacts of the behavior that it encouraged." Finally, in the Statement of Reasons for both Prosperi and Stevenson, the court made a third reference to corporate culture, noting, "There is no evidence that the defendant intended to enrich himself personally or intended to harm the CA/T [Central Artery/Tunnel] project or taxpaying public in any specific sense. Instead, the defendant was part of a corporate culture that did not consider moral consequences or public harm." The government makes much of those "corporate culture" statements, arguing that "[b]y ascribing blame for the offenses of conviction to the purported evils of the `corporate culture' rather than to the individual defendants themselves, the district court both absolved Prosperi and Stevenson of responsibility for their criminal actions and sentenced them based on the `straw man' of corporate culture rather than on Prosperi and Stevenson themselves."

We read the court's "corporate culture" statements differently. The reference to Aggregate's corporate culture in the Statement of Reasons comes immediately after the district court distinguished Prosperi and Stevenson from other white-collar criminals by noting that they were not motivated by personal enrichment. Thus, the reference to corporate culture was an attempt to identify an alternative motive — a single-minded interest in the success of their company, and not to absolve the defendants of all responsibility. Even in referring to Aggregate's corporate culture, the court did not fail to take note of Prosperi and Stevenson's role in creating that culture. In its Loss Order, the court noted that the pressure on Prosperi and Stevenson produced by Aggregate's corporate culture was largely self-generated. Thus, the court acknowledged that Prosperi and Stevenson each bore some responsibility for creating the corporate culture that it condemned.

Also, Prosperi and Stevenson did receive more substantial sentences than their co-defendants. As noted, they were each sentenced to six months of home detention, three years of probation, and 1,000 hours of community service. In contrast, the other defendants, who held lesser positions within Aggregate's Ready-Mix Concrete Division, received lesser sentences. Farrar received three months home detention, three years probation, and 750 hours community service; Blais received three months home detention, two years of probation, and 250 hours of community service; McNally received no home detention, eighteen months probation, and 200 hours of community service; and Thomas received no home detention, one year probation, and 125 hours of community service. Accordingly, the sentences imposed on Prosperi and Stevenson do reflect their higher positions within the corporation and their greater culpability.

C. Corporate Culture Generally

The government argues that the district court misapprehended the government's position when it considered Prosperi and Stevenson to be "singled out" for prosecution. In support of this argument, the government cites the district court's statement at sentencing that "[i]t is tempting but ultimately, I think, an abuse of my power as a sentencing judge to hold these defendants responsible for all of the excesses of modern corporate ills." However, this statement is a response to the government's sentencing argument, in which it stated:

Later in its argument, discussing the letters submitted on the defendants' behalf, the government added:

Given these arguments, the court understandably felt that the government was asking it to consider the "excesses of modern corporate ills" in sentencing the defendants. The court refused to do so, focusing instead on the actions of the defendants and their personal circumstances.

D. Intent to Harm

In its Statement of Reasons, the court stated that "[t]here is no evidence that the defendant[s] ... intended to harm the CA/T project or taxpaying public in any specific sense." The government argues that the district court's reliance on this finding does not justify the sentences imposed. In particular, the government notes that they knowingly committed fraud, and many fraud defendants do not specifically intend to harm their victims. It points out that a person need not intend to harm a victim to be held responsible for the foreseeable consequences of his actions.

That is a fair summary of an important legal proposition. For its part, however, the court apparently thought it was more relevant to the sentencing decision in this case that the government had failed to establish that any particular harm had resulted, or would result, from the defendants' actions.11 In its brief, the government asserts, without any citation to the record, that the court accepted the fact that the defendants' fraud "will adversely affect large portions of the CA/T in the future." In fact, the court found that the evidence presented did not indicate a likelihood of a future adverse effect. The court characterized any future harm as "speculative" and stated that "I heard no evidence that that is likely."

Moreover, in distinguishing the defendants from other white-collar fraud defendants, the court emphasized the absence of an intent to harm the Big Dig, or any direct intent on the part of the defendants to enrich themselves. This distinction remains relevant in considering the cases cited by the government on appeal, in which probationary sentences for fraud defendants have been vacated for substantive unreasonableness. In United States v. Livesay, 587 F.3d 1274 (11th Cir.2009), which involved the billion dollar HealthSouth fraud, the defendant was part of "an illegal scheme to artificially inflate HealthSouth's earnings and to falsely report HealthSouth's financial condition." Id. at 1276. In particular, the defendant "instructed HealthSouth's accounting staff to manipulate various accounts" to produce a pre-determined earnings per share, and also participated in preparing SEC filings that he knew to be materially misstated. Id. In vacating the probationary sentence originally imposed, the Eleventh Circuit emphasized that the fraud left "victims too numerous to be counted," id. at 1278, including innocent shareholders who were "bilk[ed]" out of over a billion dollars, id. at 1279. Furthermore, the court noted that the defendant substantially enriched himself by his conduct.12 Id.

In another case relied upon heavily by the government, United States v. Cutler, 520 F.3d 136 (2d Cir.2008), the Second Circuit reversed as substantively unreasonable the sentences of two defendants convicted of bank and tax fraud. One defendant had been sentenced to one year and one day of incarceration, and the other to three years of probation. In rejecting these sentences, the Second Circuit relied on several of the same arguments that the government makes here, finding that the district court erred in departing based on family circumstances and in deciding that the loss overstated the defendant's culpability. Id. at 163. However, in a case decided nine months after Cutler, the Second Circuit, sitting en banc and citing Cutler, stated that, "[t]o the extent that our prior cases may be read to imply a more searching form of substantive review, we today depart from that understanding." United States v. Cavera, 550 F.3d 180, 189 (2d Cir.2008). The Cutler decision is not helpful to the government's position.

E. Deterrence

The government argues that the sentences imposed are contrary to Congress's stated policy of increasing sentences for white-collar offenders to provide an adequate general deterrent. As the government observes, Congress has noted that deterrence is "particularly important in the area of white collar crime." S.Rep. No. 98-225, at 76 (1983), reprinted in 1984 U.S.C.C.A.N. 3182, 3259. We have previously emphasized the importance of general deterrence in white-collar crime. See United States v. Mueffelman, 470 F.3d 33, 40 (1st Cir.2006) (stating importance of "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").

By statute, the USSG must be "entirely neutral as to the ... socioeconomic status of offenders." 28 U.S.C. § 994(d). It is impermissible for a court to impose a lighter sentence on white-collar defendants than on blue-collar defendants because it reasons that white-collar offenders suffer greater reputational harm or have more to lose by conviction. See USSG § 5H1.2 (stating that "[e]ducation and vocational skills are not ordinarily relevant in determining whether a departure is warranted"); id. § 5H1.5 ("Employment record is not ordinarily relevant in determining whether a departure is warranted."); id. § 5H1.10 (stating that socioeconomic status is not relevant in determining a sentence).

We see no indication in the record that the court failed to observe these directives. In explaining its view that the sentences imposed provided an adequate deterrent, the court noted:

This reasoning applies equally well to defendants convicted of white-collar and blue-collar crimes. Also, we understand the court's comments on the burdens of the criminal process to be a comment on the specific deterrence of these defendants from any future criminal conduct.

Additionally, the court understood and credited the argument that incarceration increases the deterrent effect of a sentence on others. It weighed this benefit of incarceration against the costs of incarceration:

With this explanation, the district court fulfilled its obligation to consider the importance of general deterrence in fashioning its sentences. It decided for the reasons given that the other interests at stake made a non-incarcerative sentence appropriate in this case. It rejected the view that the interest in general deterrence could only be served by incarceration.

F. Personal Circumstances

Finally, the government argues that the personal circumstances of Prosperi and Stevenson do not distinguish them from other similarly situated defendants and do not justify the downward variance in this case. Under the guidelines, departures from the GSR based on family circumstances are "ordinarily" inappropriate save under stringent conditions. USSG § 5H1.6. However, post-Booker, a judge may vary from the GSR, disagreeing with details or even major premises, see Kimbrough v. United States, 552 U.S. 85, 101, 128 S.Ct. 558, 169 L.Ed.2d 481 (2007) ("[A]s a general matter, courts may vary from Guidelines ranges based solely on policy considerations, including disagreements with the Guidelines."), but the variance must be reasonable and, in almost all cases where a defendant has a family, some hardship and disadvantage to them will result wherever incarceration is part of the sentence.

Here, whether or not squarely within the exception set forth in the guidelines, the circumstances of Prosperi's family are atypical and powerful, both in justifying a variance and in the home confinement actually chosen. At the time of sentencing, Prosperi's wife was battling terminal cancer. She submitted a letter to the district court stating that, "I depend on my husband for almost everything. He is my caregiver, my love and he is irreplaceable. I need him by my side." Similarly, her sister submitted a letter stating that, "I fear that without [Prosperi] as her caregiver, her optimism and hope will be diminished and will have a devastating impact in her ongoing battle."

Most significantly, the doctor treating Prosperi's wife submitted a letter to the court stating that her survival was "due, in no small part to the ... remarkable care and dedication of her husband" and that "[s]he certainly would not be alive today without his attentiveness and his capacity to recognize when she is in trouble." The doctor added that "Mr. Prosperi's support has been a critical factor in keeping [his wife] alive" and "I am quite concerned about [her] ability to function without her husband."

While Stevenson presented evidence that he too served as a caretaker for members of his family, his circumstances are not as compelling. Stevenson played an integral role in the on-going care for his adopted, badly disabled daughter, and said that he was the primary caregiver for his elderly parents. Also, there were numerous letters detailing Stevenson's charitable work and support for friends and neighbors. Seemingly, his disabled daughter is now adult and living in a group home, but there was evidence that the family hoped to be able to bring their daughter, who will need care indefinitely, to a facility near to the family home, allowing Stevenson to play a role in providing care.

Furthermore, after concluding that Prosperi would not receive a sentence of incarceration, the court was entitled to take this fact into consideration in fashioning Stevenson's sentence. See United States v. Tejeda, 481 F.3d 44, 60 (1st Cir. 2007) ("[A] district court may consider disparities among co-defendants in determining a sentence."). Stevenson was a subordinate of Prosperi and seemingly participated in the fraudulent actions under the superintendence of his superior — not an excuse but a factor that a judge might reasonably think argues against a higher sentence, especially when for both men the family needs are poignant beyond the ordinary.

We have been clear that, post-Booker, "[a] district court ... may take idiosyncratic family circumstances into account, at least to some extent, in fashioning a variant sentence." Martin, 520 F.3d at 93. Although policy statements issued by the Sentencing Commission are relevant in determining the type and degree of idiosyncracy necessary to support a given variance, they are not decisive. Id. Here, for the reasons stated, the particular circumstances of both Prosperi and Stevenson were a permissible factor for the court to consider in imposing its variant sentences.

IV.

As we said at the outset of this opinion, the degree to which the sentences challenged in this appeal vary from the GSR has given us pause. We are mindful of how rare it is to encounter a variance of this magnitude. See United States v. Negroni, 638 F.3d 434, 446 (3d Cir.2011) (noting, in a case involving a 70- to 87-month GSR and a probationary sentence, that "[t]he parties have not identified any case, and we have not found one, in which an appellate court upheld a probationary sentence that so significantly varied from the Guidelines range"). One can easily argue that home confinement remains an unreasonably shallow sentence for a serious and deliberate crime which had the potential to cause large monetary loss and even physical harm to others. Many judges would have imposed prison sentences in this case even though no actual loss or harm was established, save possibly to public confidence.

That said, "while the extent of the difference between a particular sentence and the recommended Guidelines range is surely relevant, courts of appeals must review all sentences — whether inside, just outside, or significantly outside the Guidelines range — under a deferential abuse-of-discretion standard." Gall, 552 U.S. at 41, 128 S.Ct. 586. As we have previously observed, "Gall teaches that it is error to allow the dramatic nature of variance to unduly influence our review for substantive reasonableness." United States v. Thurston, 544 F.3d 22, 25 (1st Cir.2008). We have acknowledged that even when we believe that a § 3553(a) goal is not met by a sentence, we must consider the totality of the circumstances, and in particular whether the sentence sacrifices that goal to satisfy other legitimate competing interests of the sentencing regime. Id. (finding three-month sentence reasonable despite 63- to 78-month GSR).

In this case, the district court carefully explained its sentencing decisions. Most significantly, the court explained why the estimated loss amount was an unfair proxy for culpability, and why it should not drive the sentencing process. Importantly, it also found that there was insufficient evidence to conclude that the defendants' conduct compromised the structural integrity of the Big Dig, or that they sought to enrich themselves. Coupled with the individual circumstances of the defendants, these findings provided a "plausible explanation [for the sentences], and the overall result is defensible." Innarelli, 524 F.3d at 292.

For the foregoing reasons, the judgment of the district court is affirmed.

So ordered.