William Thurston was convicted of conspiring to defraud Medicare of over $5,000,000. Although the recommended sentence under the sentencing guidelines was 60 months’ imprisonment, the district court instead imposed a sentence of three months’ imprisonment. The government appeals, arguing that the sentence is unreasonable. The relevant background is as follows.
In 1998, Thurston was indicted for conspiring to commit Medicare fraud. See 18 U.S.C. § 371. The indictment stemmed from Thurston’s activities in the late 1980s and early 1990s as an executive for Damon Clinical Laboratories (Damon), a corporation that supplied clinical laboratory testing services for health care providers. Medicare reimburses clinical laboratories only for services that are medically necessary for the treatment of the beneficiary’s illness or condition. 42 U.S.C. § 1395(a)(1)(A). Thurston was indicted for conspiring with three others to manipulate Damon’s service options to encourage physicians to order unnecessary tests for Medicare beneficiaries. In particular, the indictment charged Thurston with conspiring to induce physicians to order rarely needed tests for ferritin and apolipoprotein by making them part of a battery of frequently ordered tests and informing physicians, falsely, that the tests did not cost extra.
Prior to trial, the trial judge expressed skepticism about the government’s case. At an April 2000 status conference for codefendant Joseph Isola, Damon’s president, the judge stated that he would be inclined to impose a probationary sentence, even if the government could gain a conviction. Aware that the judge had more than once granted unappealable acquittals in similar cases, the government viewed its likelihood of gaining convictions and long sentences as remote. It therefore offered generous plea agreements to each of the defendants. In exchange for a plea of nolo contendere to a one-count information, the government offered to dismiss the indictment and not to appeal the sentence imposed (which the judge had indicated would not involve imprisonment).
Isola accepted the government’s offer, but Thurston and another codefendant declined.1 Thurston rejected the offer because a guilty plea “would cause him to be shunned in the Mormon church.” At Iso-la’s sentencing hearing, the government did not recommend a sentence, and the judge imposed a three-year term of probation. Subsequently, the trial of Thurston’s remaining codefendant ended when the judge entered a judgment of acquittal pursuant to Fed.R.Crim.P. 29. In December 2001, Thurston was convicted by a jury of conspiring to defraud Medicare by induc*213ing the unnecessary ordering of the ferri-tin test.2
Thurston was first sentenced in May 2002. Applying the sentencing guidelines, the judge determined that the base offense level was six, that Thurston intended to defraud Medicare of more than $5,000,000 (a 14-level enhancement), that the crime involved more than minimal planning (a two-level enhancement), and that Thurston was an organizer or leader of extensive criminal activity (a four-level enhancement). The judge denied Thurston’s request for an acceptance-of-responsibility adjustment and did not rule on the government’s request for an obstruction-of-justice enhancement. The resulting adjusted offense level of 26 established a sentencing range of 63 to 78 months, which was trumped by a 60-month statutory maximum.
The judge departed from the guideline range and imposed a sentence of three months’ imprisonment and twenty-four months’ supervised release. He granted this departure for two reasons. First, he found that imposing a 60-month sentence on Thurston would result in an untoward disparity with the probationary sentence imposed on Isola, whom the court characterized as “the prime architect of the conspiracy.” Second, he concluded that Thurston had demonstrated a record of “extraordinary contributions and service to society, and especially to his religious obligation.”
Thurston appealed his conviction and the government cross-appealed the sentence. We affirmed the conviction but remanded for resentencing. See United States v. Thurston, 358 F.3d 51 (1st Cir.2004) (Thurston I). Under the federal sentencing regime then in place, the sentencing court was required to impose a guideline sentence unless a downward departure was permitted.3 See 18 U.S.C. § 3553(b). We concluded that the district court erred by granting Thurston a downward departure so that his sentence would be similar to Isola’s. Thurston, 358 F.3d at 79. We ruled that a departure could not be granted because of the “perceived need to equalize sentencing outcomes for similarly situated co-defendants.... ” Id. We also concluded that the court had improperly granted Thurston a downward departure because of his good works. The guidelines permitted such departures only in cases where the defendant’s good works “were exceptional.” Id. at 78-79. While Thurston had a history of good works, such efforts could not be deemed extraordinary in light of both the nature of the offense and Thurston’s status as a corporate executive with the means to undertake significant charitable endeavors. Id. at 79-80. We therefore remanded with instructions that the guideline sentence of 60-months’ imprisonment be imposed. Id. at 82.
Thurston filed a petition for a writ of certiorari with the United States Supreme Court. While Thurston’s petition was pending, two relevant events occurred. First, the trial judge recused himself because he could not impose the sentence “prescribed by the Court of Appeals.” Second, the Supreme Court decided United States v. Booker, 543 U.S. 220, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005), which replaced the mandatory guideline regime with an advisory system. The Supreme Court subsequently granted Thurston’s petition, vacated this court’s judgment, and *214remanded for further consideration in light of Booker. Thurston v. United States, 543 U.S. 1097, 125 S.Ct. 984, 160 L.Ed.2d 988 (2005). After additional briefing, we remanded for resentencing and ordered that the court consider whether an obstruction-of-justice enhancement should be applied to Thurston’s guideline range.
Following the reassignment of the case to the district court judge who issued the sentence challenged in this appeal, Thurston argued that he should receive the original sentence imposed because “unwarranted disparities between codefendants” and “good works” could be considered post -Booker. The government responded that Thurston should receive the 60-month sentence contemplated in Thurston I.
The district court expeditiously convened a hearing to determine Thurston’s sentence. It began its analysis by accepting the findings made during Thurston’s original sentencing and concluded that an obstruction-of-justice enhancement was not warranted.4 After determining that the recommended guideline sentence was 60 months, the court considered whether a lower sentence was appropriate under the sentencing factors set forth in 18 U.S.C. § 3553(a). Among these factors are the need for the sentence to reflect the seriousness of the offense, to promote respect for the law, to afford adequate deterrence, and to avoid unwarranted sentence disparities among defendants who have similar records and who have been found guilty of similar conduct. Id. § 3553(a)(2),(6).
The court began by finding that the disparity between Isola’s sentence and Thurston’s recommended sentence was unacceptable. As the court explained:
Giving the prime architect of the scheme in this case [Isola] a sentence of probation and giving Mr. Thurston a sentence of 60 months, as the government advocates because he rejected a similar deal and exercised his right to a trial, would greatly injure respect for the law unless the disparity is warranted by some legitimate consideration.
Concluding that there were no material differences between Thurston and Isola other than Thurston’s decision to exercise his right to a trial, the court ruled that this difference called for only a modest increase in Thurston’s sentence: three-months’ imprisonment.
The court next considered whether imposing a three-month sentence would adequately reflect the seriousness of Thur-ston’s crime. It again viewed Isola’s plea agreement and the similar offer to Thur-ston as the benchmark. According to the court, “implicit in the Isola plea agreement, and the same agreement or opportunity being offered to Mr. Thurston, was the government’s view that the probationary sentence was minimally adequate ... to reflect the seriousness of the offense.” The court further explained:
Medicare fraud., is a serious crime. And frequently ... both the Guidelines and the government’s recommendations are too low in white collar cases. However, the government’s plea bargain with Isola and the offer it made to Mr. Thur-ston of ... a probationary sentence are evidence that even a probationary sentence would adequately reflect the seriousness of the offense in this case.
Finally, the court considered whether a three-month sentence would adequately *215deter other potential white collar criminals. It concluded that a short prison term was adequate to serve this goal because “the most significant decision in sending a message to potential white collar criminals is the decision to send the defendant to prison. It’s not so much the amount of time, it’s whether you go away.”
After discussing the above-mentioned § 3553(a) factors, the court sentenced Thurston to essentially the same sentence initially imposed — three months’ imprisonment and 24 months’ supervised release.5 Because Thurston received credit for time served, the sentence did not require Thur-ston to serve any additional prison time.
The government appeals, arguing that the district court placed too much emphasis on the disparity between Isola’s and Thurston’s sentences. The government also argues that the court unreasonably undervalued the seriousness of Thurston’s offense and the need to deter other white collar criminals.
Thurston’s sentencing took place prior to our decision prescribing the post-Booker sentencing process. See United States v. Jimenez-Beltre, 440 F.3d 514, 518-19 (1st Cir.2006) (en banc). Nevertheless, the district court anticipated the correct procedural approach. It calculated the applicable guideline range “before deciding whether to exercise [its] new-found discretion to impose a non-guidelines sentence” based on the § 3553(a) sentencing factors. It also provided a detailed explanation of its reasons for not imposing the recommended guidelines sentence. Id. The dispute in this case centers not on the procedures employed, but on the substantive conclusion.
In reviewing a particular sentence for reasonableness, we stress the need “for a plausible explanation and a defensible overall result.” Id. at 519. We also believe that the guidelines “remain[ ] an important consideration” in sentencing because they “represente ] the only integration of the multiple [sentencing] factors [set forth in § 3553(a) ], often reflect past practice, and [bear] the imprimatur of the [Sentencing Commission], [the] expert agency charged with developing them.” United States v. Smith, 445 F.3d 1, 4 (1st Cir.2006). Thus, we consider the reasonableness of a below-guideline sentence on a sliding scale: “the farther the judge’s sentence departs from the guidelines sentence the more compelling the justification based on the factors in § 3553(a) that the judge must offer.... ” Id. (quoting United States v. Dean, 414 F.3d 725, 729 (7th Cir.2005)).
Here, the district court imposed a sentence 95% below the incarceration period recommended by the guidelines for essentially three reasons: (1) Thurston should receive a sentence close to Isola’s because they are similarly situated codefendants and a wide disparity in their sentences would punish Thurston for electing to go to trial; (2) Thurston’s offense was not terribly serious, as evidenced by the government’s pretrial plea offer of a probationary sentence; and (3) a short prison term serves the interest of general deterrence because white collar criminals typically are more concerned about whether they will be sent to prison than the length of imprisonment. The court approached Thurston’s sentencing with characteristic thoroughness and without the benefit of our recent decisions applying Booker. We are constrained, however, to conclude that these reasons do not warrant the major variance challenged in this case.
We begin by considering the disparity between Isola and Thurston. In the pre-*216Booker era, codefendant disparities could not support a sentence below the guideline range. See United States v. Wogan, 938 F.2d 1446, 1448 (1st Cir.1991). But after Booker “that a factor [was] discouraged or forbidden under the guidelines does not automatically make it irrelevant when a court is weighing the statutory factors apart from the guidelines.” Smith, 445 F.3d at 5. That said, since Booker, we have observed that § 3553(a)(6) — the factor relating to sentencing disparities' — was “almost certainly” enacted by Congress to encourage “a national uniformity focusing upon a common standard and looking to how most cases of the same kind were treated.” United States v. Saez, 444 F.3d 15, 18 (1st Cir.2006); see also United States v. Navedo-Concepcion, 450 F.3d 54, 59 (1st Cir.2006) (“Congress’s concern with disparities was mainly national”); Smith, 445 F.3d at 5 (“Congress’ goal of equality primarily envisions a national norm”). Thus, one possible problem with the court’s rationale was the emphasis on disparities between codefendants without giving significant consideration to encouraging nationwide uniformity in sentencing— the primary focus of § 3553(a)(6). But, even if the emphasis on codefendant disparity was appropriate, Isola and Thurston were not so similarly situated that nearly identical sentences were warranted. See United States v. Ortiz-Torres, 449 F.3d 61, 85 (1st Cir.2006). In our view, the district court undervalued the fact that Isola pleaded nolo contendere.
“Plea bargaining flows from the mutuality of advantage to defendants and prosecutors, each with his own reasons for wanting to avoid trial.” Bordenkircher v. Hayes, 434 U.S. 357, 363, 98 S.Ct. 663, 54 L.Ed.2d 604 (1978). In order to gain a conviction without risking an unsuccessful trial, prosecutors frequently discount the sentence they offer in a plea bargain by the probability, of loss. See Albert W. Aschuler, The Prosecutor’s Role in Plea Bargaining, 36 U. Chi. L.Rev. 50, 58-60 (1968) (stating that “the overwhelming majority of prosecutors view the strength or weakness of the state’s case as the most important factor in the task of bargaining”); Stephanos Bibas, Plea Bargaining Outside the Shadow of Trial, 117 Harv. L.Rev. 2463, 2470 (1994) (“The strength of the prosecution’s case is the most important factor” in plea bargaining).
The record supports the government’s argument that this is why Isola and Thur-ston were offered such favorable plea agreements. As noted above, the trial judge expressed doubts about the government’s case and an inclination to impose a probationary sentence even if the government gained a conviction. Moreover, the government believed that the trial judge had a history of granting acquittals in similar cases.
Thurston had the opportunity to accept this offer but declined. Thurston’s decision to go to trial entailed the risk of a stiffer post-conviction punishment. See Yeje-Cabrera, 430 F.3d 1, 25 (1st Cir.2005). This is so because “after trial, the factors that may have indicated leniency as a consideration for the guilty plea are no longer present.” Alabama v. Smith, 490 U.S. 794, 801, 109 S.Ct. 2201, 104 L.Ed.2d 865 (1989). Once Thurston elected to go trial, he was no longer situated similarly to Isola. The primary factor that led to Iso-la’s light sentence — the government’s concern that it would not prevail at trial — was no longer present.
The district court declined to consider this distinction material because, in its view, imposing a substantially harsher penalty on Thurston would impose a burden on Thurston’s exercise of his constitutional right to a jury trial.6 The law does *217not support this conclusion. “The fact that the defendant who pleads gets a benefit over those who go to trial and are convicted is a necessary artifact of any plea bargaining regime. The law long ago determined that there was nothing ... illegal about any burden on trial rights caused by such a differential.” Yeje-Cabrera, 430 F.3d at 25. In short, there was too much emphasis placed on Isola’s sentence in determining the appropriate imprisonment period for Thurston. Cf. Navedo-Concepcion, 450 F.3d 54, 2006 WL 1575573, at *5. (concluding that the district court was reasonable in sentencing a defendant who went to trial to a longer sentence than codefendants who pleaded guilty because “defendants who plead guilty often get ... lower sentences”).
We turn next to whether a three-month sentence adequately reflected the seriousness of Thurston’s crime. In considering this sentencing factor, the district court indicated that it agreed with our observation in Thurston I that “[h]ealth care fraud is a serious crime and the federal interest in combating it is powerful.” 358 F.3d at 81. Nevertheless, the court concluded that a three-month sentence was appropriate because it represented the government’s view of the seriousness of the offense. To reach this conclusion, the court relied on a sentencing guideline policy statement requiring that plea negotiation practices be conducted to “promote the statutory purposes of sentencing prescribed in 18 U.S.C. § 3553(a).” U.S.S.G. Part B, Intro. Comment. The court interpreted this policy statement to mean that, whenever the government offers a plea bargain, it is implicitly certifying that the proposed sentence adequately reflects the seriousness of the offense.
Sentencing is “primarily a judicial function.” United States v. Mistretta, 488 U.S. 361, 390, 109 S.Ct. 647, 102 L.Ed.2d 714 (1989). This was the law both before and after the promulgation of the sentencing guidelines. See U.S.S.G. Part B, Intro (“[SJentencing is a judicial function.... This is a reaffirmation of a pre-guidelines practice.”); S.Rep. 98-225 at 159, reprinted in 1984 U.S.C.C.A.N. 3342 (stating that placing the Sentencing Commission in the judicial branch was based on the understanding that “sentencing should remain primarily a judicial function”). Sentencing remains a judicial function post-Booker. See Jimenez-Beltre, 440 F.3d at 518-19.
Judicial independence in sentencing is reflected in the rule governing the acceptance of pleas. Under Fed.R.Crim.P. 11(c)(3)(a), a sentencing court is empowered to reject any plea agreement which binds the court to impose a specific sentence if the court deems the sentence inappropriate. See C. Wright, Federal Practice & Procedure, § 175.1 at n. 38 (3d ed.1999) (citing cases in which courts declined to accept plea agreements because of a disagreement with the agreed-to sentence). Post-Booker cases also emphasize the sentencing court’s obligation independently to consider the § 3553(a) factors in arriving at the appropriate sentence. See United States v. Buchanan, 449 F.3d 731, 740 (6th Cir.2006) (stating that courts must exercise “independent judgment” in imposing sentence); United States v. Cooper, 437 F.3d 324, 329 (3d Cir.2006) (stating that courts must give “meaningful consideration to the § 3553(a) factors”).
The government’s lenient pretrial offer to Thurston did not relieve the district court of its independent obligation to assess the seriousness of Thurston’s offense. The court suggested that the government undervalued the seriousness of Thurston’s crime through its pretrial offer and then *218mistakenly deferred to its perception of the government’s pretrial position. See United States v. Lazenby, 439 F.3d 928, 934 (8th Cir.2006) (“Under the Sentencing Reform Act and Booker sentencing discretion rests in the final analysis with the sentencing judge, not with the prosecution.”); United States v. Escobar Noble, 653 F.2d 34, 36 (1st Cir.1981) (“It is the particular function of the court, not the prosecutor, to say the last word about the justice of a sentence.”). While the government’s pretrial position might be relevant, a sentencing court ought not simply defer to the government’s pretrial position, without taking into account the court’s own expressed misgivings about the government’s view or the government’s explanation for its change in position.
Finally, we address the district court’s deterrence rationale. As set forth above, the court concluded that a sentence imposing substantial prison time on Thurston would not promote general deterrence because white collar defendants typically are more concerned about whether they will go to prison than with the actual length of imprisonment.
This rationale is problematic in light of our recent post-Booker pronouncements. “The clear import of the statutory [sentencing] framework is to preserve Congress’s authority over sentencing policy and to guarantee that the exercise of judicial discretion over sentencing decisions be based on case-specific circumstances, not on general, across-the-board policy considerations.” United States v. Pho, 433 F.3d 53, 62 (1st Cir.2006). Accordingly, post-Booker, “general disagreement with the broad-based policies enunciated by Congress or the Commission” do not justify imposing a sentence that is below the recommended guideline range. Id. at 65.
“One of the goals of the entire guidelines regime was to minimize discrepancies in the treatment of ‘white collar’ and ‘blue collar’ offenses.” Thurston I, 358 F.3d at 80. From the outset of the guideline regime, the Sentencing Commission determined that the penalties for white collar crime had to be increased because of the inequity in the pre-guideline sentencing practice of “punishing economic crimes less severely than other apparently equivalent behavior.” U.S.S.G., ch. 1, pt. A, § 3. In recent years, Congress has expressed concern that punishments for white collar offenses were too lenient. In 2002, Congress enacted the White Collar Crime Penalty Enhancements Act which, inter alia, instructed the Sentencing Commission to consider whether the sentencing “guidelines and policy statements ... are sufficient to deter ... [white collar] offenses” and to “modify the sentencing guidelines” accordingly. See Pub.L. No. 107-204, § 905(b)(1),(2). In response, the Commission increased the penalties and enhancements for certain white collar offenses. See 68 Fed.Reg. 2615-01 (2003); see also U.S.S.G. § 2B1.1 (2001) (setting forth the 2001 economic crime sentencing reforms which increased the guideline penalties and enhancements for white collar offenses).
The district court is not alone in viewing long prison sentences as unnecessary to deter white collar crimes. See, e.g., United States v. Yeaman, 248 F.3d 223, 238 (3d Cir.2001) (Nygaard, J., dissenting in part); A. Mitchell Polinsky & Steven Shavell, On the Disutility and Discounting of Imprisonment & the Theory of Deterrence, 28 J. Leg. Stud. 1, 12 (1999). But, as expressed through the recommended guideline offense levels, this view is not shared by “Congress or the Commission, as its agent.” Pho, 433 F.3d at 65.
A court may sentence below the guidelines because the guideline sentence appears unreasonable in the “particular *219case[],” Jimenez-Beltre, 440 F.3d at 518, but not because of “general disagreement with broad-based policies enunciated by Congress or the Commission.” Pho, 433 F.3d at 65. Here, the court pointed to no facts concerning Thurston’s offense or individual circumstances indicating that a three-month sentence would adequately deter similar crimes. Cf. United States v. Rattoballi 452 F.3d 127, 136 (2d Cir.2006) (concluding that a below-guideline sentence for a white collar defendant was unreasonable where the district court imposed a lenient sentence based on characteristics common to all white collar defendants). Therefore, this was not an appropriate ground for reducing Thur-ston’s sentence.
As demonstrated, the reasons provided do not support Thurston’s receiving a sentence 95% below the guideline recommendation. Typically this conclusion would end the discussion, and we would remand for resentencing. See Smith, 445 F.3d at 7. But this case is atypical. Some of the conduct underlying Thurston’s conviction is almost two decades old, and Thurston’s sentence now has been vacated twice. We are concerned that vacating Thurston’s sentence without providing further guidance might lead to another contentious district court proceeding and a third appeal. This would benefit no one.
Two district judges have concluded that a sentence below the guideline recommendation is warranted in this case. While we think that there are plausible reasons to support a below-guideline sentence, these reasons do not permit the extreme variance that has been granted.
One ground supporting a below-guideline sentence could be that the intended loss attributed to Thurston overvalued the seriousness of his offense. This was a basis for a downward departure under the guidelines, see United States v. Ravelo, 370 F.3d 266, 275 (2d Cir.2004), and remains relevant post-Booker, see United States v. McBride, 434 F.3d 470, 477 (6th Cir.2006).
The trial judge found that Thurston intended a loss of over $5,000,000. “Much of Thurston’s guideline sentence range ... was driven” by this determination. Thurston I, 358 F.3d at 68. We concluded in Thurston I that the intended loss finding was sustainable, but believed that the issue presented a somewhat close question. Id. at 69. Unconstrained by the mandatory guideline regime, a sentencing court plausibly could conclude that a $5,000,0000 intended loss finding, with its resultant 14-level increase in the offense level, leads to a modest overstatement of the seriousness of Thurston’s crime.
Another possible reason for a below-guideline sentence is Thurston’s documented record of good works. This was one of the bases on which Thurston received a downward departure in his first sentencing. We vacated that departure because a good works departure was impermissible unless the good works were “exceptional.” Thurston I, 358 F.3d at 79. While we concluded that Thurston’s endeavors were not “exceptional,” we did acknowledge that “[sjave for his crime, Thurston ... lived a creditworthy life.” Thurston I, 358 F.3d at 79. Post-Booker, a sentencing judge “could well conclude that a ... discouraged factor[ ] did not quite justify a departure from the guidelines ... but might justify a somewhat shorter sentence under a reasonableness standard.” United States v. Bradley, 426 F.3d 54, 55-56 (1st Cir.2005). Under this more flexible approach, Thurston’s good works could justify a somewhat shorter sentence.
Isola’s probationary sentence also could justify a somewhat reduced sentence. While Isola and Thurston were not similarly situated, we acknowledge the district court’s concern that too wide a divergence *220between the sentence imposed on Isola and Thurston could negatively affect the public’s “respect for the law.” In administering the criminal justice system, it is important to “preserve the appearance of justice.” Cullen v. United States, 194 F.Sd 401, 408 (2d Cir.1999). A sentencing court could plausibly conclude that extremely divergent sentences would undermine the accepted notion that similar conduct should be punished in a somewhat similar manner. Cf. Lazenby, 439 F.3d at 934 (recognizing that extreme disparities in the sentences imposed on coconspirators could “fail[] to promote respect for the law”).
Balanced against these grounds for leniency are reasons for imposing a stiff penalty. Thurston committed a serious offense warranting substantial punishment. Medicare “fraud is a serious crime ... [which] affects the financial integrity of [a] program[ ] meant to aid tens of millions of people in need of health care. Every dollar lost to fraud is a dollar that could have provided medical care to the elderly or the disabled.” Thurston I, 358 F.3d at 81. Moreover, the recommended guidelines sentence is 60 months and, for the reasons articulated above, this variable remains an “important consideration” in identifying a reasonable sentence. Jimenez-Beltre, 440 F.3d at 518; supra at 9-10.
An appellate court is not well-suited to determine the appropriate sentence for a defendant. See Smith, 445 F.3d at 7 (“Framing a ... sentence is, in the first instance, the responsibility of the district judge”). But post-Booker, appellate courts are tasked with drawing the line separating reasonable from unreasonable sentences. See Booker, 543 U.S. at 263, 125 S.Ct. 738 (stating that “appellate judges will prove capable” of applying the “reasonableness” standard under advisory guidelines). Having reviewed the record, the recommended guideline sentence, and the § 3553(a) factors, we conclude that a sentence of fewer than 36 months’ imprisonment would fail reasonableness review in the present circumstances. See United States v. Moreland, 437 F.3d 424, 437 (4th Cir.2006) (vacating sentence imposed as unreasonable under Booker and remanding with instructions setting a floor for resen-tencing).
In setting the minimum sentence that could withstand reasonableness review, we are not endorsing 36 months as the correct sentence for Thurston. Indeed, given the seriousness of Thurston’s offense and his role as an organizer and lead implementer of the fraud, this court would be inclined to impose a sentence at or near the guideline recommendation if it were acting as the sentencing court. But our role here is only to determine the minimum sentence that could be considered reasonable on the present record. Accordingly, we emphasize that in remanding for resentencing, we are not ordering that a 36-month sentence be imposed. The district court is free to solicit further argument and evidence from the parties, but, absent an extraordinary development, the court must impose a sentence of no less than 36-months’ imprisonment.
Vacated and remanded.
. In the interim, the fourth codefendant died.
. The judge entered a Rule 29 judgment of acquittal on the apolipoprotein charge.
. At that time, we reviewed decisions to grant downward departures de novo. See 18 U.S.C. 3742(e), as amended by Pub.L. No. 108-21, § 401.
. Thurston argues that the district court was wrong to accept the enhancements imposed in the first sentencing on the ground that Booker raises the standard of proof for imposing guidelines enhancements from a preponderance of the evidence to beyond a reasonable doubt. We have rejected this argument. E.g., United States v. Yeje-Cabrera, 430 F.3d 1, 17 (1st Cir.2005); United States v. Antonakopoulos, 399 F.3d 68, 80 (1st Cir.2005).
. The court added a $25,000 fine which was not part of the initial sentence.
. The court did conclude that a small increase in Thurston’s sentence was appropriate be*217cause his decision to go to trial caused additional expenditures by the government.