Frank M. Patti, Sr. appeals the seventy-nine-month sentence he received after he pled guilty to filing a false income tax return in violation of 26 U.S.C. § 7206(1), and conspiring to defraud the United States in violation of 18 U.S.C. § 371. He asserts that the district court abused its discretion by denying his motion for recu-sal and that the district court erred in sentencing him by miscalculating the tax loss and by enhancing his sentence for obstruction of justice. ^[1]

BACKGROUND

On May 31, 2000, a grand jury returned a sealed, eight-count indictment, charging Patti with various tax violations. The case was assigned to Judge Lacey A. Collier, and the trial was scheduled to begin on May 29, 2001. Before the case went to trial, however, Patti filed a motion for recusal pursuant to 28 U.S.C. § 455(a). ^[2] Judge Collier denied that motion on March[*1320] 13, 2001. Although Patti filed supplemental evidence to support his § 455(a) motion after the initial denial, ^[3] Judge Collier filed a supplemental order denying the motion after receiving that evidence.

Thereafter, a setback arose when Patti injured himself in a car accident. The accident brought into question Patti’s competency to stand trial, as he claimed that he had amnesia and could not recall the facts surrounding the alleged tax evasion. As a result of the injuries, Patti was committed to a federal medical center pursuant to 18 U.S.C. § 4241(d)(1). In February of 2002, after a hearing, Judge Collier found that Patti was malingering and that he was competent to stand trial. Thus, a new trial date of April 15, 2002 was established. Patti, however, agreed to plead guilty to filing a false tax return and conspiring to defraud the United States, two of the counts charged in the twenty-four-count second superseding indictment, and his plea was accepted on April 12, 2002. Thereafter, the district court sentenced him to seventy-nine months of imprisonment. This appeal followed.

DISCUSSION

I. Motion for Recusal

Patti asserts that Judge Collier abused his discretion by denying his motion for recusal. The government, however, asserts that we cannot review the denial of Patti’s motion, because Patti waived his right to raise that issue by entering an unconditional guilty plea.

Generally, a voluntary, unconditional guilty plea waives all nonjurisdic-tional defects in the proceedings. United States v. Fairchild, 803 F.2d 1121, 1124 (11th Cir.1986) (per curiam). Although we have addressed several types of claims that have been waived by entering an unconditional guilty plea, see, e.g., United States v. Wai-Keung, 115 F.3d 874, 877 (11th Cir.1997) (per curiam) (holding that a defendant waived his right to appeal the validity of a search); Fairchild, 803 F.2d at 1124 (holding that a defendant waived his right to appeal on the grounds of pros-ecutorial vindictiveness), we have not addressed whether a defendant waives his right to appeal the district court’s denial of a motion for recusal under § 455(a) by entering an unconditional guilty plea. ^[4] Our sister circuits that have considered that issue, however, have come to opposite conclusions, thus creating a circuit split. See United States v. Hoctel, 154 F.3d 506, 507-08 (5th Cir.1998) (holding that a defendant waives his right to appeal the denial); United States v. Gipson, 835 F.2d 1323, 1324-25 (10th Cir.1988) (holding the same); see also United States v. Troxell, 887 F.2d 830, 833 (7th Cir.1989) (holding that the defendant waived her right to appeal the denial of her § 455(a) motion because the “denial of a motion for recusal based on the appearance of impropriety can be challenged only with a writ of mandamus”). But see United States v. Brinkworth, 68 F.3d 633, 637-38 (2d Cir.1995) (holding that a defendant does not waive his right to appeal the denial); United States v. Chantal, 902 F.2d 1018, 1020-21 (1st Cir.1990) (holding the same). For the following reasons, we agree with the Fifth and Tenth Circuits and hold that a defendant waives his right to appeal the denial[*1321] of a § 455(a) motion by entering an unconditional guilty plea.

Section 455 creates two primary reasons for recusal. See 28 U.S.C. § 455(a)-(b). A judge should recuse himself under § 455(a) when there is an appearance of impropriety. See id. § 455(a). Section 455(a) provides, “Any justice, judge, or magistrate judge of the United States shall disqualify himself in any proceeding in which his impartiality might reasonably be questioned.” Id. “The very purpose of § 455(a) is to promote confidence in the judiciary by avoiding even the appearance of impropriety whenever possible.” Liljeberg v. Health Servs. Acquisition Corp., 486 U.S. 847, 865, 108 S.Ct. 2194, 100 L.Ed.2d 855 (1988). Thus, the standard of review for a § 455(a) motion “is whether an objective, disinterested, lay observer fully informed of the facts underlying the grounds on which recusal was sought would entertain a significant doubt about the judge’s impartiality,” Parker v. Connors Steel Co., 855 F.2d 1510, 1524 (11th Cir.1988), and any doubts must be resolved in favor of recusal, United States v. Kelly, 888 F.2d 732, 745 (11th Cir.1989).

On the contrary, a judge should recuse himself under § 455(b) when any of the specific circumstances set forth in that subsection exist, which show the fact of partiality. 28 U.S.C. § 455(b)(1)-(5). ^[5] For example, a judge should recuse himself “[wjhere he has a personal bias or prejudice concerning a party” or “[w]here in private practice he served as [a] lawyer in the matter in controversy.” Id. § 455(b)(1)-(2). Recusal under this subsection is mandatory, because “the potential for conflicts of interest are readily apparent.” Murray v. Scott, 253 F.3d 1308, 1312 (11th Cir.2001) (internal quotation marks omitted).

Thus, the differences between the two subsections are evident. While subsection (b) sets forth specific circumstances requiring recusal, which establish the fact of partiality, subsection (a) sets forth a general rule requiring recusal in those situations that cannot be categorized neatly, but nevertheless raise concerns about a judge’s impartiality. See 28 U.S.C. § 455(a)-(b). Although both subsections provide reasons for recusal, subsection (b) is stricter than subsection (a) because the need for a judge’s recusal under subsection[*1322] (b) is clear; once it has been established that one of the enumerated circumstances exists, there can be no dispute about the propriety of recusal. Under subsection (a), however, whether recusal is necessary is not as readily apparent. As subsection (a) does not set forth specific circumstances requiring recusal and instead provides a general rule requiring recusal when there is an appearance of impropriety, we must ask “whether an objective, disinterested, lay observer fully informed of the facts underlying the grounds on which recusal was sought would entertain a significant doubt about the judge’s impartiality.” Parker, 855 F.2d at 1524.

Indeed, the treatment of subsections (a) and (b) under § 455(e) further demonstrates the differences between these subsections. Section 455(e) provides, “No justice, judge, or magistrate judge shall accept from the parties ... a waiver of any ground for disqualification enumerated in subsection (b). Where the ground for disqualification arises only under subsection (a), waiver may be accepted provided it is preceded by a full disclosure on the record of the basis for disqualification.” ^[6] 28 U.S.C. § 455(e). Thus, a party can waive recusal under § 455(e) when the motion is brought pursuant to § 455(a), but cannot waive recusal when the motion is brought pursuant to § 455(b).

In light of the differences between subsections (a) and (b), particularly Congress’s express provision for waiver of recusal under subsection (a), we conclude that the “denial of recusal is a pretrial defect which is sublimated within a guilty plea and thereafter unavailable as an issue for appeal.” Gipson, 835 F.2d at 1325. ^[7] Accordingly, Patti waived his right to appeal the denial of his motion for recusal when he entered his unconditional guilty plea.

Moreover, we find it noteworthy that Patti failed to avail himself of the options that were available for obtaining review of the denial of his motion for recusal. Indeed, Patti immediately could have petitioned this Court for a writ of mandamus upon the denial of his motion, or he could have requested permission from Judge Collier to enter a conditional guilty plea in which he reserved the right to appeal the district court’s denial of his motion. Although Patti asserts that neither alternative was viable, ^[8] he did not attempt to[*1323] exercise either option. As a result, we find his arguments regarding the futility of these options unavailing.

Thus, the district court’s denial of the motion for recusal under § 455(a) is not renewable. We therefore offer no opinion about the merits of Patti’s motion.

II. Sentencing

“We review the district court’s application of the Sentencing Guidelines de novo, and its findings of fact for clear error.” United States v. Wilson, 183 F.3d 1291, 1300 n. 16 (11th Cir.1999). ^[9]

A. Calculation of Loss

We review the district court’s calculation of the amount of loss for clear error. United States v. Renick, 273 F.3d 1009, 1025 (11th Cir.2001) (per curiam).

The district court calculated the tax loss as $4,791,267.18 under the sentencing guidelines and thus determined that Patti’s base offense level was twenty-one. Patti asserts that the district court erred in calculating the tax loss, because it aggregated the corporate and personal tax losses. ^[10]

Whether the tax loss can be calculated by aggregating the corporate and personal tax losses is an issue of first impression in this Circuit. Several of our sister circuits, however, have addressed this issue. The Second and the Seventh Circuits found that it is inappropriate to aggregate the corporate tax loss and the personal tax loss, because doing so overstates the tax revenue lost. See United States v. Martinez-Rios, 143 F.3d 662, 672 (2d Cir.1998); United States v. Harvey, 996 F.2d 919, 920-22 (7th Cir.1993). The Sixth Circuit took a contrary view in United States v. Cseplo, 42 F.3d 360, 364-65 (6th Cir.1994), holding that the guidelines clearly provide for aggregation and that by aggregating the amounts the calculation reflects the seriousness of the harm the defendant caused.

We believe that the aggregation of personal and corporate tax losses is called for under the guidelines. The version of the guidelines that the district court applied in this case provided that “[i]f the offense involves both individual and corporate tax returns, the tax loss is the aggregate tax loss from the offenses taken together.” U.S.S.G. § 2T1.1, cmt. n. 7. Thus, as the Sixth Circuit noted, “[t]he guidelines [we]re very specific about the necessity of aggregating the tax losses.” Cseplo, 42 F.3d at 364. In addition, a[*1324] defendant “ha[s] the opportunity and ability to limit the criminal consequences to one or other of the returns.... By choosing to falsify both returns, [he] ma[kes] the deliberate decision to produce separate harm to the government with respect to both tax liabilities.” Id. at 364-65 (internal quotation marks omitted). Thus, we believe that the defendant should be held accountable for both harms, and we do not believe that the possibility of a lower tax liability had the taxes been reported properly should change that result. See id. at 364-65 & 365 n. 6 (noting that “[i]f [the defendant’s] unorthodox maneuvers resulted in a higher aggregate tax liability than would have existed otherwise, that is a risk [he] chose to run when he elected to break the law”).

Furthermore, the sentencing guidelines were amended in 2001, and the commission adopted the Sixth Circuit’s view. See U.S.S.G. § 2T1.1, cmt. n. 7 (providing that “[i]f the offense involved both individual and corporate tax returns, the tax loss is the aggregate tax loss from the individual tax offense and the corporate tax offense added together” (emphasis added)). Although Patti asserts that the amendment was substantive, we disagree. The commentary provides that it “clarifies the pri- or rule in Application Note 7 of § 2T1.1 [sic] that if the offense involves both individual and corporate tax returns, the tax loss is the aggregate tax loss from the offenses taken together.” U.S.S.G. supp. to app. C 191 (2001) (emphasis added) (internal quotation marks omitted). Thus, this recent amendment further supports our decision to permit the aggregation of corporate and personal tax losses. See United States v. Perulena, 146 F.3d 1332, 1337 n. 11 (11th Cir.1998) (noting “that amendments that do not effect a substantive change, but rather are intended only to clarify the rule adopted by a particular guideline ... constitute strongly persuasive evidence of how the Sentencing Commission originally envisioned that the courts would apply the affected guideline” (alteration in original) (internal quotation marks omitted)).

Thus, the district court properly calculated the tax loss. ^[11]

B. Obstruction of Justice

We review the district court’s determination that the defendant’s conduct warrants an obstruction of justice enhancement for clear error. United States v. Garcia, 208 F.3d 1258, 1261 (11th Cir.2000), vacated on other grounds by 531 U.S. 1062, 121 S.Ct. 750, 148 L.Ed.2d 653 (2001). The district court should increase a defendant’s base offense level by two if

the defendant willfully obstructed or impeded, or attempted to obstruct or impede, the administration of justice during the course of the investigation, prosecution, or sentencing of the instant offense of conviction, and ... the obstructive conduct related to (i) the defendant’s offense of conviction and any relevant conduct; or (ii) a closely related offense.

U.S.S.G. § 3C1.1. ^[12]

Here, there were three grounds upon which the district court relied to enhance[*1325] Patti’s sentence. As we find that two of those grounds independently support the obstruction enhancement, we decline to address the third. ^[13]

First, the district court found that Patti feigned amnesia and noted that his malingering postponed the trial for a year and forced the government to waste time and resources in evaluating his competency. Patti asserts that the district court erred in basing the enhancement upon its determination that he feigned amnesia in his competency hearing, because such a ruling will chill defendants’ willingness to raise competency issues. ^[14] We, however, disagree and, like the Fifth Circuit, find that

applying the obstruction enhancement to defendants who willfully feign incompetency in order to avoid trial and punishment does not unconstitutionally chill a defendant’s right to seek a competency hearing. While a criminal defendant possesses a constitutional right to a competency hearing if a bona fide doubt exists as to his competency, he surely does not have the right to create a doubt as to his competency or to increase the chances that he will be found incompetent by feigning mental illness.... [This ruling does not] put defense counsel to the Hobson’s choice of forgoing competency hearings for a client who may well be incompetent ... or requesting such hearings and exposing the client to the risk of a § 3C1.1 enhancement if he is ultimately found competent. Counsel should warn his client that feigning incompetency, whether to create doubt as to his competency so as to prod his attorney into requesting competency hearings or to convince the court that he cannot stand trial, will trigger a § 3C1.1 enhancement.

See United States v. Greer, 158 F.3d 228, 237-38 (5th Cir.1998).

Second, the district court found that Patti was involved in the attempted arson at his accountant’s office, which was undertaken to destroy documents relevant to his tax fraud. Patti asserts that there was insufficient evidence to connect him to the attempted arson and that the district court relied upon unreliable hearsay in determining that he was involved in the arson. Although Patti asserts that the district court based its ruling upon unreliable hearsay, one of the coconspirators in that arson testified at the sentencing hearing; his testimony about the statements of fellow coconspirators did not constitute hearsay. ^[15] See Fed.R.Evid. 801(d)(2)(E) (“A statement is not hearsay if ... [t]he statement is offered against a party and is ... a statement by a coconspirator of a party during the course and in furtherance of the conspiracy.”). Moreover, even if this testimony were consid[*1326] ered hearsay, the law of this Circuit clearly provides that reliable hearsay can be considered at sentencing. Wilson, 183 F.3d at 1301 (“A court may consider any information (including hearsay), regardless of its admissibility at trial, in determining whether factors exist that would enhance a defendant’s sentence....”). The district court may rely upon such evidence “as long as the evidence has sufficient indicia of reliability, the court makes explicit findings of fact as to credibility, and the defendant has an opportunity to rebut the evidence.” United States v. Anderton, 136 F.3d 747, 751 (11th Cir.1998) (per curiam). As the district court found the coconspirator’s testimony credible and there was sufficient indicia of reliability, the district court properly relied upon the coconspirator’s testimony about the arson.

Thus, the district court did not err in enhancing Patti’s sentence under section 3C1.1.

CONCLUSION

Accordingly, we AFFIRM.