Petitioner-Appellant Jerry S. Payne appeals an adverse decision of the Tax Court, which awarded Respondent-Appellee Commissioner of Internal Revenue ("the government" or "the IRS") $438,722 in delinquent income taxes and penalties for tax years 1987 and 1988, plus interest. As a general rule, the IRS must assess taxes within three years following the date that the return is filed. Here, the IRS did not send Payne a notice of deficiency (an event that tolls the statute of limitations pending assessment) until more than three yearsafter he had filed his return for each of those years. The Tax Court found, nevertheless, that the IRS's collection action was timely under the statutory fraud exception to the three-year statute of limitations.1 As it had to if it were to determine taxpayer fraud, the Tax Court found the government's evidence of fraud to be clear and convincing. But in our clear error review, we see that evidence as weak and equivocal, so that disregarding the statute of limitations cannot be justified on grounds of tax fraud. The judgment of the Tax Court is therefore reversed and judgment rendered in favor of Payne, granting his petition for redetermination of income taxes, penalties, and interest for 1987 and 1988 and holding that the government is barred by the statute of limitations from collecting anything from Payne for those tax years.

Payne is a lawyer. During the years at issue he practiced law, concentrating in litigation. Payne provided extensive legal representation to, and eventually came to own, a corporation called 2618, Inc. ("2618") which, as sole proprietor, operated Caligula XXI, a topless dance club (the "club") in Houston, Texas. Payne also represented Gerard Helmle, one of 2618's two equal shareholders and the club's manager. Among other things, Payne defended Helmle against a criminal charge for possession of illegal drugs. Most of the operable facts of this case arise out of these professional representations.

At the beginning of 1987, Helmle and Leo Kalantzakis each owned one-half of the stock of 2618. As prerequisites to operating a topless dance club in Houston, 2618 needed both (1) a liquor license, technically a Mixed Beverage Permit, from the Texas Alcoholic Beverage Commission (the "TABC"), and (2) a Sexually Oriented Business Permit ("SOB permit") from the City of Houston ("the City"). The SOB permit was required by a Houston ordinance passed in 1986 which provides, inter alia, that one topless dance club cannot operate within 1,000 feet of another. The ordinance also specifies that if two such dance clubs seeking SOB permits are located within 1000 feet of each other, a permit can be issued only to the club that has been in operation longer. As part of his representation of 2618, Payne helped it apply for an SOB permit. The club was located within 1000 feet of a competing topless dance establishment, however, so 2618's application for an SOB permit was denied. The Tax Court recognized that without an SOB permit the club's viability was in serious doubt.

Payne filed suit against the City to force issuance of an SOB permit to 2618. The primary issue in the suit was which club had been in operation longer.

While that suit was proceeding in state district court, Helmle and Kalantzakis, had a falling out. Their dispute resulted in litigation between 2618 and Kalantzakis, in which Payne represented the corporation. Ultimately this matter was settled by Helmle's agreeing to purchase Kalantzakis's stock in 2618, which would leave Helmle as the corporation's sole stockholder.

By this time, Payne had amassed substantial unpaid accounts receivable resulting from his criminal defense of Helmle and his representation of 2618 in several matters. Helmle did not have the financial wherewithal either to fund his purchase of Kalantzakis's stock or to pay Payne's account. The club was Helmle's sole source of income, and his dispute and eventual settlement with Kalantzakis threatened the continued existence of the club. Payne was aware that the club's survival represented his only realistic possibility of ever recovering his fees for legal services rendered to Helmle and to 2618. As neitherHelmle nor 2618 was creditworthy, Payne borrowed $275,000 from Texas Guaranty National Bank then lent that same sum to Helmle, who used these funds to purchase Kalantzakis's stock in 2618.

Payne and Helmle agreed that Helmle would cause 2618 to make monthly payments to Payne so that he, in turn, could make periodic payments of principal and interest on the bank loan. In essence, Payne acted as an intermediary, first in borrowing from the bank and passing the loan proceeds through to his client, and then in receiving funds from his client and immediately disbursing those funds to the bank that had made the loan.

Helmle also agreed to compensate Payne for his increased involvement in the club's operations during this period by paying him a management fee. Payne reported the management fee on his income tax returns for the years in question. He did not, however, report the sums that he received from his clients and then immediately remitted to the lender bank. As to these he took the position that he was a mere accommodation borrower and conduit through which the loan proceeds and repayments passed, not a party in interest to an income-producing transaction.

During the time that Kalantzakis owned one-half of the stock of 2618, he had handled the renewals of the corporation's mixed-beverage permit from the TABC. Kalantzakis had apparently developed relationships with high-level personnel at the TABC, which helped expedite the permit renewal process. After Kalantzakis's split with Helmle and Helmle's subsequent purchase of Kalantzakis's stock, however, Kalantzakis was no longer willing to use his relationship with TABC officials for the corporation's benefit. In fact, there are allegations that Kalantzakis lobbied his contacts at the TABC to deny renewal of 2618's mixed-beverage permit. Payne contends that ultimately, through its relationship with Kalantzakis, the TABC learned that criminal drug charges were pending against Helmle. This prompted the head of enforcement for the TABC to determine that, because Helmle was the sole owner of 2618, its mixed-beverage permit should not be renewed.

Payne counseled Helmle that his best solution was to sell the club. Helmle agreed and authorized Payne to find a buyer. Unfortunately for Helmle, though, all potential buyers that Payne contacted lost interest when they discovered that the City had denied the club's application for an SOB license and that the TABC was refusing to renew the club's mixed-beverage permit.

After trying unsuccessfully to preserve any going-concern value that the club might have (apparently at this point, there was little or none), Payne foreclosed on encumbrances of 2618's assets that he held as security for unpaid legal fees. Specifically, Payne foreclosed on the corporation's (1) leasehold interest in the building in which the club operated, (2) furniture, furnishings, fixtures, and leasehold improvements in the building, and (3) right to use the name Caligula XXI. Payne concluded that the assets he foreclosed on had a fair market value of $35,000 and reported this amount as income on his federal income tax return. Payne leased the assets back to 2618 in the hope that the liquor license and SOB permit would be issued, which should make it possible for Helmle or 2618 to pay the remaining legal fees owed to Payne.

After they failed to find a buyer for the club and determined that the reason the TABC would not issue a mixed-beverage permit to 2618 was the criminal charges pending against Helmle, Payne and Helmle embarked on a new strategy to secure a mixed-beverage permit: They entered into a conditional stock-purchase agreement under which Payne agreed to buy all issued and outstanding 2618 stock from Helmle in consideration of Payne's $500,000 note, when and if 2618 secured a mixed-beverage permit. Payne reasoned that when the TABC realized that its issuanceof a permit to the club would terminate Helmle's ownership, the TABC would grant the mixed-beverage permit. Payne negotiated with the TABC for the renewal of the club's permit, but when these negotiations broke down he filed suit against the TABC. The lawsuit was ultimately settled when the TABC agreed to issue the club a mixed-beverage permit. This satisfied the condition precedent in the stock purchase agreement between Helmle and Payne, causing the stock to be transferred to Payne in exchange for his note and making him the sole shareholder of 2618.

Shortly after the stock was transferred to Payne, Helmle agreed to reduce the sum due on Payne's note from $500,000 to $300,000. Even so, Payne never made any payments on the note.

The Tax Court found the credit sale of the stock from Helmle to Payne to be a sham transaction, and reclassified the transfer of the stock from Helmle to Payne as an in-kind payment for past legal services. On appeal, Payne does not contest the Tax Court's characterization of the transaction as a payment in-kind for legal fees. Rather, he contends that the 2618 stock was worthless at the time he received it from Helmle. As such, urges Payne, he was correct in concluding that he need not report receipt of the valueless stock as income from his law practice.

Payne based his conclusion of worthlessness on the specter of the litigation that was then pending between 2618 and the City concerning the denial of the club's SOB permit. Without an SOB permit the club could not operate; and, in Payne's considered professional opinion, 2618's odds of success in that suit were abysmal. Furthermore, the liquor license was the corporation's only significant asset; it had no SOB permit and no longer owned (1) its leasehold interest in the only location where it was licensed to sell liquor, (2) the leasehold improvements needed to conduct the dance club operations, or (3) the trade name under which the club operated. Those assets had long since been lost to Payne through foreclosure, a transaction on which Payne had reported income. In Payne's estimation, these factors combined to render the stock worthless on the date he acquired it.

The IRS prosecuted Payne for criminal tax fraud on facts arising from essentially the same transactions that are at issue in this case ---- and Payne was acquitted. During the pendency of the criminal tax prosecution, Payne filed a civil suit against the IRS for divulging confidential tax information during its criminal investigation. Payne's civil suit against the IRS resulted in a $1.7 million judgment for Payne. The government's appeal of that decision is currently pending before this court.

We have jurisdiction to review decisions of the Tax Court pursuant to I.R.C. § 7482. We review such decisions "in the same manner and to the same extent as decisions of the district courts in civil actions tried without a jury."2 Accordingly, findings of fact are reviewed for clear error and conclusions of law are reviewed de novo.3

This appeal is governed by § 6501 of the Internal Revenue Code. Subsection (a) of § 6501 proclaims the general rule that "the amount of any tax imposed by this title [Title 26, U.S.C.] shall be assessed within 3 years after the return was filed," otherwise any collection effort by the government shall be time barred. As the three-year period set forth in § 6501(a) istolled by the issuance of a statutory notice of deficiency,4 that general rule of limitation can be rephrased to read: Unless a statutory notice of deficiency is sent to the taxpayer within 3 years after the return was filed, the government's collection effort shall be time barred.

In this case, the government sent the statutory notices of deficiency for both 1987 and 1988 more than three years after Payne had filed his returns for those years. Thus, unless an exception to the three-year limitations period is applicable, notices of deficiency were issued too late, and the government is barred from collecting the tax deficiencies, penalties, and interest it now asserts.

The only exception to the general three-year limitations rule of § 6501(a) that is implicated in this appeal is § 6501(c)'s statutory tax fraud exception, which provides: "In the case of a false or fraudulent return with the intent to evade tax, the tax may be assessed, or a proceeding in court for collection of such tax may be begun without assessment, at any time." The burden of proving fraud is on the government.5 To satisfy its burden, the government must prove, by clear and convincing evidence, that at least some portion of the asserted underpayment of tax is the result of fraud.6 If the government carries this high burden with respect to any part of the underpayment, "the entire underpayment shall be treated as attributable to fraud, except with respect to any portion that the taxpayer establishes (by a preponderance of the evidence) is not attributable to fraud."7 We have defined fraud in the following terms: "Fraud implies bad faith, intentional wrongdoing and a sinister motive. It is never imputed or presumed and the court should not sustain findings of fraud upon circumstances which at most create only suspicion."8 Fraud is usually inferred from "conduct, the likely effect of which would be to mislead or conceal."9

As a general rule, the government's determination of a tax deficiency is presumptively correct. A consequence of this presumption is that the taxpayer bears the burden of proving that the government's determination is incorrect or arbitrary.10 We and other courts have held, however, that when the government relies on an exception to the three-year statute of limitations, it bears the burden of proving its entitlement to rely on that exception.11 This means that alone the general presumption of the correctness of the government's deficiency determination cannot serve to establish fraud on the part of the taxpayer; proof of fraud remains the burden of the government. Indeed, to hold otherwise would be to ignore the statute and the related case law that impose on the government the burden of proving fraud by clear and convincing evidence.12 There must be additional evidence, independentof the general presumption of correctness, from which fraudulent intent on the part of the taxpayer can be properly inferred.13

The government asserts that its most compelling evidence of fraud ---- and therefore the evidence most likely to surmount the clear and convincing evidence threshold ---- lies in the 2618 stock transfer from Helmle to Payne. Before the Tax Court, the government introduced an expert report that appraised the stock at $1.14 million as of the date of the transaction. This conclusion was expressly predicated on the expert's assumption that 2618 would continue to operate a topless club "indefinitely." That assumption, however, was directly contrary to the facts as they existed on the date Payne acquired the stock, the only date relevant to the appraisal. At that time, the City was steadfastly refusing to grant 2618 an SOB permit, which all concede was an absolute necessity if the club was to continue operating.

Aware of this flaw in the expert's analysis, the Tax Court "conclude[d] that [the government's] expert's [$1.14 million] valuation for the stock of [2618] should be reduced by a discount of 50 percent to reflect the risks associated with the litigation over the [SOB license]." In essence, the court began by agreeing with the government expert that, with the SOB license, the stock was worth $1.14 million and concluding that, without that license, the stock was worthless. Then simplistically ---- without any analysis or expert evidence of the odds of success in the license litigation ---- the Court arbitrarily split the difference. Consequently, the Tax Court found the stock's value, at the time Payne received it, to be $570,000, exactly half-way between zero and $1.14 million. The Tax Court then jumped directly to its ultimate conclusion that Payne's "failure to report any amount as income in connection with his receipt of the stock was part of [his] fraudulent conduct."

For the following reasons we find clearly erroneous the Tax Court's conclusion that Payne's failure to report the stock transfer from Helmle is clear and convincing proof of fraudulent intent. First, we are skeptical of the Tax Court's conclusional finding that, at the time the stock was transferred to Payne, there was a 50 percent chance that 2618 would win the litigation and get the SOB permit. Payne assigned a far smaller chance that this would be the outcome of the suit; and it seems to us that, as the attorney representing 2618 in that ongoing litigation, Payne was in the best position to assess 2618's chances. But even if we were to disregard Payne's opinion as incredible, we cannot disregard the government's failure to adduce evidence, expert or otherwise, on this question. Our review of the record reveals no evidence that we see as probative on this point.

Second, the Tax Court's analysis is predicated on the conclusions of the government's expert as announced in his report. We have examined this report and find that it contains internal flaws not discussed by the Tax Court. For example, the report included the following qualification: "The subject assets, properties or business interests are appraised free and clear of any or all liens or encumbrances . . . ." Based in part on this statement and in part on other indicia in the report, we are convinced that the expert was appraising not only the going concern value of 2618's business with licenses in place, but was also assigning value to the corporation with its lease, leasehold improvements, and trade name in place, specifically, the leasehold interest in the building where the club operated, the furniture, fixtures, equipment, and the leasehold improvements used in nightly operations, and the Caligula XXI name. As we previously noted, though, Payne had already foreclosed on those assets in a separate transaction months earlier, one on which he reported income and to which neither theTax Court nor the government has ascribed fraudulent intent. It is fallacious, therefore, to treat the assets lost by 2618 through foreclosure as contributing to the value of the stock on the date, months later, that it was acquired by Payne in an entirely separate transaction. Even with an SOB permit, how could 2618 operate the club without its trade name, the only location from which it was authorized to conduct its dance and liquor business, and its furniture, fixtures and leasehold improvements?14

Third, the Tax Court's analysis fails to take into account the delay and expense associated with litigating the licensure issue with the City. The government's expert's report qualified its conclusion that if the club were to operate indefinitely its value was $1.14 million; the expert opined that the club's value was $1.14 million "net of any costs associated with removing any impediments preventing operation as a topless club. Such costs would be expected to include legal fees and associated costs." The report went on to explain that two "key parameters need to be assessed with respect" to its projection of value:

[1] the legal avenues available to contest closure under the [SOB] ordinance and [2] the cost of pursuing such remedies. At this time, we believe the estimate of these parameters would necessarily have been speculative as of the valuation date. We have not reviewed information or conducted discussions with individuals who could provide reliable analyses of the relevant legal issues and corresponding costs. As a result, we have not drawn a conclusion with respect to these specific circumstances.

Despite this significant and substantial qualification in the expert's report, the government did not offer evidence regarding these "key parameters" and the Tax Court did not reduce the $1.14 million estimate to reflect the negative effect these factors might have on the price a willing buyer would pay for the stock.

Taken together, the foregoing shortcomings in the Tax Court's analysis and the expert report on which it relied compel us to conclude that the Tax Court's finding on valuation is not supported by the record. Essentially, the Tax Court began with a value that was too high because it ignored the costs of litigating the SOB licensure, and because it included the assets on which Payne had already foreclosed. The court then discounted this inflated figure by 50 percent "to reflect the risks associated with litigation" over the SOB permit. We find no evidence in the record supporting the Tax Court's conclusion that 50 percent was an appropriate discount. The product of the inflated value and the arbitrary ---- and likely excessive ---- odds that the Tax Court assigned to a favorable outcome, i.e., to the issuance of an SOB permit, yield a conclusion as to the value of 2618's stock that we find untenable.

The Tax Court noted that its $570,000 valuation conclusion approximated the amount of Payne's outstanding legal bills with 2618 and Helmle at the time of the stock transfer, and suggested that this "further supported or corroborated" the determination that the stock was worth $570,000 when Payne acquired it. The record indicates, however, that Helmle was not creditworthy at the time of the stock transfer, and had proved himself unable to remit payment for legal services to Payne in a timely fashion, if at all. We agree with the Tax Court that if an attorney and his creditworthy client had arranged an arms-length in-kind payment, the value of the property transferred in paymentshould approximate the value of the legal services for which payment is being made. But this logic breaks down when, as here, the client is not creditworthy and, indeed, has no other assets: An attorney with a substantial account receivable owed by an insolvent client may well have an account receivable with a value of zero. Any such creditor is likely to accept even valueless assets when essentially "writing off" a receivable from an insolvent debtor. We cannot agree with the Tax Court that the amount Helmle owed Payne supports the court's valuation of the 2618 stock.

Perhaps the Tax Court concluded that, because the $500,000 note that Payne gave to Helmle in exchange for the stock was a sham, it constituted evidence of fraudulent intent. And, if the court did so conclude, it may well have been correct. But any fraud associated with that element of the transaction was just as likely if not more likely directed at some other party ---- e.g., the City or the TABC15 ---- as at the IRS. Consequently, even if we assume arguendo that the sham credit sale might constitute clear and convincing evidence of fraud, it is not clear and convincing evidence of tax fraud.

There is no direct evidence in the record of any deceptive or evasive conduct by Payne. The government argues, nevertheless, that Payne's fraudulent intent could be inferred from his failure to report income stemming from the stock transfer. To infer fraud from this transaction, though, one first has to accept the conclusion that the stock had value when Payne received it, a conclusion about which we are dubious.

Even if we assume for purposes of argument that the Tax Court did not clearly err when it determined that the stock had substantial value at the time Payne received it, we are nevertheless left with the firm conviction that the court clearly erred when it concluded that this transaction and Payne's omission of its value from his tax return constitute clear and convincing evidence of fraud. "The fraud meant is actual, intentional wrongdoing, and the intent required is the specific purpose to evade a tax believed to be owing."16 Payne's explanation is that he believed the stock to be worthless, and we find his explanation to be plausible ---- at least as plausible as the government's competing explanation that (1) the stock had substantial value when Payne received it, and (2) that Payne knew this and thus believed he owed tax but did not report it.

The question before us is whether the Tax Court committed clear error when it found that the government proved the fact of fraud by clear and convincing evidence. "A finding is 'clearly erroneous' when although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed."17 Judged against this standard, we find clearly erroneous the Tax Court's determination that the government proved fraud by clear and convincing evidence.

As we observed, the government asserts that Payne's return position on the stock transfer presents its strongest case for tax fraud. Consequently, if the evidence about that transaction fails to surmount the clear and convincing evidentiary hurdle, then a fortiori the evidence about other transactions in which Payne was involved, proffered by the government to support its contention of fraudulent intent, must also fail. As the possibility nevertheless remains that the cumulative effect of all of the government's evidence could constitute clear and convincing evidence of tax fraud, we have closely scrutinized the entire record in search of evidence of fraud that might enhance the evidence that we have discussed. Our review of the record only serves to reinforce our conclusion that the Tax Court clearly erred in finding that the government proved fraud by clear and convincing evidence. Accordingly, we are constrained to reverse the Tax Court's determination that Payne filed his tax returns with fraudulent intent. Consequently, the statutory fraud exception to the three-year statute of limitations is not available to the government. As the government's deficiency notices for 1987 and 1988 were not duly furnished within the applicable three-year period under the statute of limitations provided in § 6501(a), collection of additional taxes, penalties, and interest for those years is time barred. The Tax Court therefore erred reversibly in applying the statutory fraud exception of § 6501(c) and denying Payne's petition for redetermination of taxes, penalties, and interest assessed pursuant to those tardy notices of deficiency.

The expansive record in this case certainly demonstrates that Payne has no acumen for keeping orderly records of his financial dealings; and we sympathize with the government and the Tax Court for the difficulty they faced in reconstructing Payne's financial affairs and then attempting to determine their tax consequences. In addition, we are aware that, in some cases, poor record keeping has been deemed indicative of fraud. But, as there is little else in this record to suggest that Payne had direct fraudulent intent, his deficiency in record keeping is not sufficient to sustain the government's burden of proving fraud to the required degree.

At bottom, the competing contentions of the parties are obvious. Payne insists that, for the years in question, his allowable deductions exceeded his taxable income and, believing that he would not owe any taxes, he paid little attention to the preparation of his returns. In contrast, the government urges that when Payne prepared and filed his returns, he did so with the intention of understating his income tax liability. Despite our painstaking review of the record, we are unable to determine which of these competing positions more closely comports with reality. We are able to determine from our record review, however, that the government has failed to support its version with evidence any more convincing than the evidence that Payne has adduced in support of his version. This evidentiary equipoise results in a draw, leaving us with the firm conviction that the government has failed to carry its burden of proving fraud by the heightened clear and convincing standard. We hold, therefore, that the Tax Court erred reversibly in allowing the statutory fraud exception to prevail over the three-year statute of limitations.

The judgment of the Tax Court is reversed for the foregoing reasons and judgment rendered in favor of Payne, granting his petition for redetermination and holding that the government is time barred from collecting additional taxes, penalties, and interest from Payne for his tax years of 1987 and 1988.

REVERSED and RENDERED.