Pamela Medley complains that the 97-month sentence imposed after our remand for resentencing is harsher than the original 78-month sentence that she appealed. On both occasions the district court sentenced her within the sentencing range computed under the United States Sentencing Guidelines. The first sentence was at the top of the range; the second sentence was at the bottom. The increase in her Guidelines sentencing range resulted from two changes. First, at the sentencing on remand the court used the 2000 version of the Guidelines rather than the 2001 version used at the initial sentencing; on her original appeal Ms. Medley had argued that the 2000 version should have been used. Second, at the sentencing on remand the court correctly1 applied some Guidelines provisions that had been mistakenly omitted or misapplied at the initial sentencing; on the original appeal the Government had noted all but one of those errors and argued that Ms. Medley therefore would not benefit from a remand for resentencing. Ms. Medley contends that an unrebutted presumption of vindictiveness attached to the district court's higher sentence on remand. She also argues that the district court disregarded our mandate to resentence her in light of United States v. Booker, 543 U.S. 220, 125 S. Ct. 738, 160 L. Ed. 2d 621 (2005), because it failed to explain why her initial 78-month sentence was not reasonable. We disagree. There is no basis for a presumption of vindictiveness; and if there were, any presumption was rebutted. And there was no Booker error.

In January 2002 Ms. Medley was convicted on five counts of wire fraud, see 18 U.S.C. § 1343; six counts of mail fraud, see id. § 1341; one count of making a fraudulent claim on the government, see id. § 287; seven counts of money laundering, see id. § 1956(a)(1)(B)(i); two counts of impersonating a government employee, see id. § 912; one count of making false statements, see id. § 1001; and one count of falsely representing her social security number, see 42 U.S.C. § 408(a)(7)(B). The convictions arose out of a complex pattern of deceit that she employed in attempting to obtain government funds intended for victims of the New Mexico Cerro Grande Fire. She was not a victim of the fire.

The probation office applied the 2001 version of the Guidelines in preparing her presentence investigation report (PSR). The PSR, as modified by an addendum, arrived at a total offense level of 24 after it (1) grouped all the conviction counts together under USSG § 3D1.2(d); (2) applied §§ 2S1.1(a)(1) and 2B1.1 to arrive at a base offense level of 6; (3) added 14 levels under § 2B1.1(b)(1)(H) for an intended loss of between $400,000 and $1 million; (4) added 2 levels for money laundering under § 2S1.2(b)(1)(B); and (5) added 2 levels for obstruction of justice under § 3C1.1. It calculated her criminal history as category I, because her only countable prior offense was a 1992 conviction in California. The Guidelines range was therefore 51 to 63 months. The PSR added, however, that her criminal-history category did not adequately reflect her background.

The government filed a motion for an upward departure under the Guidelines on the ground that Ms. Medley's "`criminal history category does not adequately reflect the seriousness of [her] past criminal conduct or the likelihood that [she] will commit other crimes.'" R. Vol. 2 Doc. 110 at 1 (quoting USSG § 4A1.3). It suggested that a criminal-history category of V was the minimum level necessary to protect the public from Ms. Medley. Ms. Medley filed objections to the PSR in which she maintained that she was innocent of all the crimes that the PSR had identified as warranting an upward departure, but she did not dispute the PSR's application of the 2001 Guidelines.

The district court conducted the sentencing hearing on December 17, 2002. It adopted the PSR's factual findings and its Guidelines calculation. It also determined that an upward departure was warranted because a criminal-history category of I did not sufficiently account for Ms. Medley's past criminal conduct. The court found that she had violated court orders in connection with a baseless lawsuit that she had filed against the University of California Medical Center, attempted to drug the attorneys of her opponents in that case, committed fraud in an effort to obtain a New Mexico medical license, and made false statements in an application for public assistance. The court therefore ruled that a criminal-history category of III, leading to a sentencing range of 63 to 78 months' imprisonment, was more representative of Ms. Medley's past criminal conduct. The court imposed a sentence of 78 months.

Ms. Medley appealed her conviction and sentence. Her counsel filed a brief raising nine issues but stating that eight were raised under Anders v. California, 386 U.S. 738, 87 S. Ct. 1396, 18 L. Ed. 2d 493 (1967) (court may grant counsel's motion to withdraw after counsel files a brief that explores possible avenues of appeal and demonstrates that all lack merit). The ninth issue was that the district court had erred in applying the 2001 version of the Guidelines. Application of the 2001 Guidelines rather than the 2000 version, she argued, resulted in an additional three-level increase in the offense level under the amount-of-loss Guideline, USSG § 2B1.1(b)(1) (2000). Ms. Medley also filed a pro se brief asserting that her sentence should be vacated and remanded for resentencing under the 2000 Guidelines.

The government's appellate brief contended that use of the 2001 Guidelines was not plain error. Noting that plain error arises only when the error "seriously affects the fairness, integrity, or public reputation of judicial proceedings," Aplee. Answer Br. in No. 03-2026 at 38; see United States v. Gonzalez-Huerta, 403 F.3d 727, 732 (10th Cir.2005) (en banc), the brief argued that there would be no such effect because Ms. Medley's Guidelines offense level would be higher if calculated correctly on remand. Specifically, the government asserted that her fraud and money-laundering counts had been mistakenly grouped together and that under the 2000 Guidelines her offense level would be increased because of an intended loss of more than $500,000 (under § 2F1.1(b)(1)(K)), more-than-minimal planning (under § 2F1.1(b)(2)), the use of so-phisticated means (under § 2F1.1(b)(6)(C)), the unlawful use of identification to obtain another means of identification (under § 2F1.1(b)(5)(C)(i)), and the grouping rules (§ 3D1.4). It calculated that a criminal-history category of III would place her in a Guidelines sentencing range of 78 to 97 months' imprisonment.

After submission of the original briefs, Ms. Medley sent a letter to this court raising a Booker argument. We remanded for resentencing because Booker, issued after Ms. Medley's original sentencing, had rendered the Guidelines advisory; Ms. Medley had preserved Booker error at sentencing; and the government had not established that any sentencing error was harmless. See United States v. Medley, 130 Fed.Appx. 248, 252 (10th Cir.2005). We said that the remand rendered it unnecessary for us to address Ms. Medley's other sentencing issues. See id.

On remand the government submitted objections not previously made to the PSR, pointing to the mistakes in the original PSR that it had noted in its brief on appeal. One of Ms. Medley's responses to the government's objections was her own objection to "any sentence which is greater than the sentence which was previously imposed." R. Vol. 1 Doc. 130 at 5.

The probation office's second addendum to the PSR tracked the government's objections. It applied the 2000 Guidelines and grouped the money-laundering counts separately from the fraud counts (wire fraud, mail fraud, false claims, impersonation, false statements, and false representation of social security number). It computed Ms. Medley's base offense level for the fraud grouping as 6 under USSG § 2F1.1(a); added 10 levels under § 2F1.1(b)(1)(K) for an intended loss of more than $500,000; added 2 levels under § 2F1.1(b)(2)(A) for more than minimal planning; added 2 levels under § 2F1.1(b)(4)(A) for impersonating a government employee; added 2 levels under § 2F1.1(b)(5)(C)(i) for unauthorized use of identification to obtain another form of identification; added 2 levels under § 2F1.1(b)(6)(C) for use of sophisticated means; and added 2 levels under § 3C1.1 for obstruction of justice. The adjusted offense level for the fraud offenses was 26. It then grouped the money-laundering counts, and under § 2S1.1(a)(2) determined a base-offense level of 20. After applying another two-level adjustment for obstruction of justice under § 3C1.1, it arrived at an adjusted offense level of 22. It then computed the combined-offense level under § 3D1.4, resulting in a total offense level of 28. The only feature of the calculation not noted in the government's prior appellate brief was the upward adjustment for impersonating a government employee. Based on a criminal-history category of I, the applicable Guidelines range was 78 to 97 months' imprisonment. Ms. Medley filed objections to the second addendum, but the probation office rejected them in a third addendum.

The district court resentenced Ms. Medley on November 9, 2005. Her counsel argued vindictiveness, stating that "[i]t is our position that if the court were to increase the sentence, it could be seen as a sign of vindictiveness on the part of the court, or on the part of the prosecution," R. Vol. 12 at 5, and "if the Court were to increase the sentence from what it found appropriate, based on basically the same evidence [as was presented at the first sentencing hearing], then it would be a sign of vindictiveness on the part of the Court." Id. at 39.

Restricting itself to the evidence that had been before it at the first sentencing, the district court again granted an upward departure, assigning Ms. Medley a criminal-history category of III. It also adopted the offense-level calculations set forth in the second addendum to the PSR, resulting in a Guidelines range of 97 to 121 months' imprisonment. It stated that this range was sufficient to satisfy the purposes of sentencing, noted that it had considered the 18 U.S.C. § 3553(a) factors, and sentenced Ms. Medley to 97 months' incarceration. The court then addressed her vindictiveness argument:

With regard to the argument that was made [about] vindictiveness, I must address that.... I have ... review[ed] your transcript, and look[ed] at your case. I do believe that you have received the benefit of a truly individualized Booker analysis.... I am always very careful, and I've been very careful in your case. It doesn't bother me that you appealed. Please don't ever think that. As a matter of fact, I look at the appellate process as a safety net.... We just handle too many criminal cases in one day, and I thank God there are other people looking over our shoulder, and making sure that we don't make a mistake so that someone that shouldn't be in jail as long as we have sentenced them to jail.... I'm glad somebody else is looking at this.... I do believe that the enhancements that were brought up for the first time during this round of sentencings are very much justified by your actions in this case, and by the facts in this case, and by the facts that existed at the time of the original sentencing.

R. Vol. 12 at 54-57.

Ms. Medley contends that the increase in her sentence on remand must be set aside because (1) the government has not rebutted the presumption that the increase resulted from vindictive retaliation for her first appeal and (2) the sentencing court did not comply with Booker because it failed to explain why the original sentence was not reasonable.

Although the government contends that Ms. Medley did not preserve her vindictiveness argument below, we need not resolve the matter. Assuming that we should apply de novo review, see United States v. Gurule, 461 F.3d 1238, 1246 (10th Cir.2006) (constitutional challenges to a sentence are reviewed de novo), we affirm.

At Ms. Medley's resentencing the district court imposed a higher sentence based on a recalculation of her Guidelines sentencing range. She claims that this increase violates the doctrine set forth in North Carolina v. Pearce, 395 U.S. 711, 89 S. Ct. 2072, 23 L. Ed. 2d 656 (1969). The Supreme Court stated in that case that to "assure the absence of [vindictiveness]" that would deter defendants from challenging their convictions, the Due Process Clause requires that "whenever a judge imposes a more severe sentence upon a defendant after a new trial, the reasons for his doing so must affirmatively appear." Id. at 726, 89 S. Ct. 2072. Ms. Medley argues that the higher sentence after remand creates a presumption of vindictiveness that the government has not rebutted.

Later decisions, however, have narrowed Pearce. In Texas v. McCullough, 475 U.S. 134, 106 S. Ct. 976, 89 L. Ed. 2d 104 (1986), the jury had sentenced the defendant to 20 years' imprisonment, but after a retrial the judge sentenced him to 50 years. See id. at 136, 106 S. Ct. 976. The judge explained that the higher sentence was based on new evidence at the second trial showing that the defendant had played a much larger role in the crime than was evident at the first trial. See id. The Supreme Court held that no presumption of vindictiveness arose because the judge who imposed the new sentence had also been the judge who had granted the defendant a new trial after the original sentence. See id. at 138-39, 106 S. Ct. 976. It further held that any presumption of vindictiveness was overcome by the judge's explanation for the increase in sentence. See id. at 141. McCullough recognized that the explanation did not come within the Court's prior language "permit[ting] `a sentencing authority to justify an increased sentence by affirmatively identifying relevant conduct or events that occurred subsequent to the original sentencing proceedings.'" Id. (brackets omitted) (quoting Wasman v. United States, 468 U.S. 559, 572, 104 S. Ct. 3217, 82 L. Ed. 2d 424 (1984)). But, it explained, "[t]his language ... was never intended to describe exhaustively all of the possible circumstances in which a sentence increase could be justified. Restricting justifications for a sentence increase to only `events that occurred subsequent to the original sentencing proceedings' could in some circumstances lead to absurd results." Id.

Alabama v. Smith, 490 U.S. 794, 109 S. Ct. 2201, 104 L. Ed. 2d 865 (1989), then reduced the Pearce doctrine to its essential core. The Court held that the defendant has the burden "to prove [that] actual vindictiveness," id. at 799-800, 109 S. Ct. 2201, caused the higher sentence and that a presumption of vindictiveness arises only in circumstances "in which there is a reasonable likelihood that the increase in sentence is the product of actual vindictiveness on the part of the sentencing authority," id. at 799, 109 S. Ct. 2201 (citation and internal quotation marks omitted).

Our decision in United States v. Rourke, 984 F.2d 1063 (10th Cir.1992), is instructive. Rourke's original sentence of 13 years' incarceration should have included a term of three years' parole. See id. at 1065. After the sentencing hearing and imposition of sentence, Rourke noticed that the district court had failed to include the parole term. See id. He brought this error to the district court's attention, arguing that his term of imprisonment should be reduced by three years once the parole term was added. See id. The court rectified its error by adding a three-year parole term to the defendant's 13-year sentence. See id. Because the sentencing judge had explained that the increase was to "correct an inadvertent omission and to comply with the statutory mandate," we held that there was no merit to the defendant's vindictiveness claim. Id. at 1066.

The circumstances of this case hardly suggest "a reasonable likelihood that the increase in [Ms. Medley's] sentence [was] the product of actual vindictiveness." Smith, 490 U.S. at 799, 109 S. Ct. 2201 (citation and internal quotation marks omitted). On the contrary, just as the judge in Rourke was correcting an error, the district court here was complying with the mandate of Booker to compute correctly the Guidelines sentencing range and then consider that range in exercising discretion under 18 U.S.C. § 3553. See United States v. Kristl, 437 F.3d 1050, 1053-55 (10th Cir.2006). Moreover, even were we to impose a presumption of vindictiveness, the presumption was clearly rebutted by the explanation for the changes in the Guidelines calculation.

On appeal Ms. Medley does not challenge any of the Guidelines calculations, nor does she suggest that she can prove actual vindictiveness. Thus, we hold that the increase in her sentence did not deny her due process.

Ms. Medley also contends that the district court did not follow our mandate that she be resentenced in light of Booker, because 18 U.S.C. § 3553, as construed by Booker, required the court to explain why the prior, vacated sentence was not reasonable. (She does not contend that her new sentence is itself unreasonable.) We review this contention de novo. See United States v. Acosta-Olivas, 71 F.3d 375, 377 (10th Cir.1995) (district court's interpretation of a statute or the Guidelines is reviewed de novo).

We discern no error. Our mandate after the first appeal directed the district court to resentence her in light of Booker, which made the Guidelines advisory, thereby giving the judge discretion in imposing sentence. See Medley, 130 Fed.Appx. at 252. As we have recently explained, there may be a substantial range of reasonable sentences for a defendant. See United States v. Begay, 470 F.3d 964, 975 (10th Cir.2006). The sentencing court can properly select any term within that range and need not explain why it did not choose another reasonable sentence.

Ms. Medley's sentence is AFFIRMED.

McCONNELL, J., concurring.

I agree with the majority that the circumstances of this case do not suggest "actual vindictiveness" on the part of the sentencing judge. I also agree with the majority's analysis of the narrow legal question raised by the higher sentence on remand and therefore join the opinion. But there are reasons to think that our current system of sentencing has a systematic bias in favor of higher sentences on remand from successful appeals, even successful appeals by the defendant. If so, this presents troubling questions of fairness and possibly even of due process, wholly apart from any "actual vindictiveness" on the part of district judges.

In any complex system—and the Sentencing Guidelines surely qualify—there is a not-insubstantial chance that those who administer the system will make mistakes, overlooking potentially applicable adjustments and enhancements. Odds are that most of these mistakes are to the benefit of the defendant, because the vast majority of Guidelines provisions relate to upward enhancements rather than downward adjustments. If a potentially applicable Guidelines provision is overlooked, most likely that means the defendant will receive a lower sentence than if the calculation were performed flawlessly. Add to this the possibility that the defendant, who is usually intimately familiar with his own history in a way the prosecution is not, will be more assiduous in fly-specking the presentence report (PSR) than prosecutors.

In this case, for example, the original PSR mistakenly grouped the defendant's fraud and money-laundering offenses and overlooked applicable enhancements based on intended loss, more than minimal planning, use of sophisticated means, unlawful use of identification, obstruction of justice, and impersonating a government employee. These oversights cumulatively resulted in an advisory Guidelines range nineteen months lower than it should have been (measuring from the bottom of the range). During the initial sentencing, the government either did not notice or chose not to object to these errors, and the district court adopted the erroneous PSR calculations in determining the advisory Guidelines range. Unless this case is aberrational, there must be many similar cases where defendants receive the benefit of a less-than-perfect Guidelines calculation.

It follows that, if those who administer the system have occasion to scour the Guidelines a second time with respect to a particular defendant, they often will discover reasons why the advisory Guidelines range should be increased. Whenever a sentence is appealed and reversed, such an occasion will present itself. That is what happened here: on appeal, and then after remand, the government re-examined the PSR calculations, objected to the errors, and obtained a higher sentence. This suggests that, after reversal on appeal, many defendants will end up worse off even though they were the victors in the appellate court.

The prospect that the defendant will end up worse off is not primarily a product of vindictiveness, but of complexity. For the same reason that manual recounts of punched ballots in Florida in 2000 tended to produce more valid votes and that audits of tax returns tend to discover additional tax liability, a redo of a Guidelines calculation is likely to produce a higher range the second time around. I am reminded of my tenth-grade algebra teacher, who had a standing policy that she would correct exam-grading errors only if the complaining student consented to have the entire exam regraded. I never felt confident enough to run the risk. Defendants must frequently feel the same.

If these speculations are valid, defendants may be wary of appealing lest, on remand, the probation office and the prosecution revisit the Sentencing Guidelines calculations and calculate a higher range. This opportunity for sentence recalculation is effectively a tax—payable in months or even years of additional prison time—on taking an appeal. This strikes me as unfair to the defendant and bad for the system. Appeals serve an important function, and we should not create disincentives for criminal defendants to appeal when they have meritorious grounds for doing so.

At this point, however, the existence of such a "tax" on appeals is purely speculative. The Sentencing Commission, which is statutorily charged with keeping relevant statistics for the purpose of improving sentencing, 28 U.S.C. § 995(a), does not collect or analyze data on changes in the length of sentencing after remand. I urge it to do so. If my hunch that there are systematic tendencies toward increases is correct, it would also be useful to know whether these are the product of recalculations by probation officers, objections by the prosecution, or exercises of Booker discretion by district courts. Empirical study of these matters by academic researchers would be illuminating.

One possible remedy to this problem, if it turns out there really is a problem, is to apply principles of "law of the case" to sentencing determinations. In other words, the prosecution would be required to raise any objections to the PSR in the initial sentencing; if the initial sentence is reversed, the remand would be limited to correcting the error identified on appeal and would not extend to de novo recalculation of the entire advisory Guidelines range.

This is similar to the way appeals work in other legal contexts. As a general rule, "once a court decides an issue, the same issue may not be relitigated in subsequent proceedings in the same case." Grigsby v. Barnhart, 294 F.3d 1215, 1218 (10th Cir. 2002) (quoting Ute Indian Tribe of the Uintah and Ouray Reservation v. Utah, 114 F.3d 1513, 1520 (10th Cir.1997)). According to this doctrine, lower court rulings not contested on appeal are abandoned—they are assimilated into the law of the case—and may not be revisited upon remand absent compelling circumstances. See Demarest v. Price, 130 F.3d 922, 942 n. 9 (10th Cir.1997).

There is no doctrinal reason why sentencing calculations cannot be subjected to this rule. These calculations are expressly found by the trial court based on evidence presented to it. Tellingly, we do not allow trial courts to revisit other legal or factual findings on remand when, as in the case of sentencing calculations, the predicate evidence for those findings was available at the time of the original hearing. See United States v. Monsisvais, 946 F.2d 114, 117 (10th Cir.1991). If the government recommends a miscalculated sentence at the trial level, and fails to challenge or otherwise correct its mistakes on appeal, law-of-the-case doctrine suggests they should not be allowed to take a second try on remand.

Even if such an approach were not adopted wholesale, individual appellate panels might consider phrasing their remands in sentencing decisions more narrowly. Not every remand for re-sentencing need be de novo. Appellate courts are entitled to direct lower courts to address only specific issues on remand. See United States v. Webb, 98 F.3d 585, 587-88 (10th Cir.1996); United States v. Smith, 116 F.3d 857, 859 (10th Cir.1997) (holding that on remand "the district court is free to reconsider the sentencing package de novo unless the appellate court has specifically limited the district court's discretion on remand").

Finally, district courts would do well to take the possibility of an "appeal tax" into account in their exercise of Booker sentencing discretion. Title 18 U.S.C. § 3553(a)(6) instructs courts to avoid unwarranted sentence disparities among defendants with similar records who have been found guilty of similar conduct. If it is true that defendants whose PSRs have been recalculated as a result of appeal and remand systematically receive higher sentences than those who do not, this provision suggests that district courts should take notice and do something to counter it.