Stephen Philip Kauffman, Wright, Constable & Skeen, Baltimore, Md., argued for petitioner-appellant.

Murray S. Horwitz, Tax Div., U.S. Dept. of Justice, Washington, D.C., argued (Shirley D. Peterson, Asst. Atty. Gen., Gary R. Allen, William S. Estabrook, Tax Div., U.S. Dept. of Justice, and Abraham N.M. Shashby, Jr., I.R.S., on brief), for respondent-appellee.

Before ERVIN, Chief Judge, PHILLIPS, Circuit Judge, and RAMSEY, Senior United States District Judge for the District of Maryland, sitting by designation.

Manuel Cebollero, the taxpayer, appeals from the tax court's decision sustaining the Commissioner's determination of income tax deficiencies for tax years 1982-1984. We now affirm.

The facts are set forth in detail in the opinion below, Cebollero v. Commissioner, 60 T.C.M. (CCH) 1379, 1990 WL 192757 (1990), and we repeat them only to the extent necessary to resolve the issues raised by this appeal.

Appellant and his wife owned a beer and wine store ("Belby") in Montgomery County, Maryland, which they operated as general partners during the period at issue. Both spouses worked in the store, and Appellant maintained the business records, which included a handwritten cash receipts journal, a handwritten disbursements journal, and cash register tapes. During busy periods, Belby operated with two cash registers, and Appellant has admitted that he willfully understated Belby's gross receipts on its partnership returns by failing to record most of the receipts from the second register. It was this understatement which caused the deficiencies at issue in this case.1

In 1985, Revenue Agent Laura Sencer audited Cebollero's income tax returns and Belby's partnership returns. During an interview on the Belby premises, Cebollero and Sencer reviewed the types of products sold at Belby, the markup applied to these products, and the proportion of Belby's total sales attributable to each product line. In addition, they discussed the manner in which sales were recorded, and Cebollero falsely represented to Sencer that all of Belby's sales had been rung up on a single register. After the interview, Sencer compared the sales reported on the tax returns to Appellant's bank deposits, and was unable to reconcile these amounts. In addition, she discovered unexplained disparities between the amounts reflected on the cash register tapes and those reflected in the cash receipts journals.

Finding the records supplied by Cebollero inadequate, the revenue agent attempted to reconstruct Belby's income independently. Because of the cash nature of Belby's business, Sencer chose a percentage markup method, which in this case involved four basic steps. First, the agent verified Cebollero's stated markups by comparing the prices Belby charged its customers for each brand of beer and wine with the prices it paid to the Montgomery County Liquor Control Board ("the Board") for those products.2 The agent determined the cost to Belby by simply telephoning the Board, and asking for the current prices charged to licensees for the various brands of beer and wine. She then compared the current 1985 wholesale costs obtained from the Board with a current 1985 retail price list supplied by the Appellant, and found that he had accurately represented his markups, as of the time of the audit, to be 22 per cent on cases of beer, 35 percent on six-packs, 25 percent on regular wine, and an average of 50 percent on specialty wines. On direct examination, Sencer testified that Appellant had consistently stated that his 1985 markups were the same as those used from 1982 through 1984.

Second, the agent determined the quantity of beer and wine Belby purchased from the Board by comparing the disbursements journal to the bank statements and the check journal. She found that the total purchases indicated in these records essentially matched those reported in the returns. By adding Belby's purchases to the beginning inventory reported on the tax returns and subtracting its reported ending inventory, she was able to compute the total cost of goods sold for each tax year.

Third, Sencer determined Belby's average sales mix. Appellant represented to Sencer that his total sales mix was approximately 70 percent beer and 30 percent wine. The agent's examination of Belby's records revealed that over the three years in question, the average mix was actually 72.9 percent beer and 27.1 percent wine. Each of these product categories was then divided into two subcategories. According to Appellant, 40 percent of the beer was sold in the form of six-packs, at a 35 percent markup, and the remaining 60 percent in cases, at the 22 percent markup. Similarly, 80 percent of the wine sold was regular wine at a 25 percent markup, with the rest sold at an average markup of 50 percent.

Finally, Sencer computed Belby's gross receipts. For each tax year, she determined the total cost to Belby of each of the four product groups by multiplying the total cost of goods sold by the percentage of sales attributable to each product. She then multiplied the total cost of each group by the verified percentage markup applicable to that group. The resulting figures indicated that Belby's partnership returns had grossly understated its income.3

At trial before the tax court, the parties stipulated that Belby had underreported its income during the years in question, and that Cebollero had knowingly failed to report his proper share of partnership income. The taxpayer challenged only the amount of the deficiencies computed by the Commissioner. The tax court expressly found that the reconstruction of income was reasonable, and upheld the determination of deficiency with minor modifications. On appeal, Cebollero claims that the tax court erred (1) by refusing to shift the burden of proof to the Commissioner, and (2) by admitting evidence of Agent Sencer's telephone call to the Board over his hearsay and relevance objections.

Where the taxpayer fails to keep or produce adequate records from which his income can be determined, the Commissioner "is at liberty to resort to the best procedure available under the circumstances in making his determinations...." Burka v. Commissioner, 179 F.2d 483 (4th Cir.1950). The percentage markup method has long been recognized as an appropriate means of reconstructing the income of cash businesses, including liquor stores. See Webb v. Commissioner, 394 F.2d 366 (5th Cir.1968); Yelding v. Commissioner, 61 T.C.M. (CCH) 3017, 1991 WL 114821 (1991). Appellant does not challenge the use of this method in his case, but maintains that the revenue agent's computations were fundamentally flawed, because they relied on the assumption that the markups in effect in 1985 were the same as those used in 1982-1984. Cebollero claimed that the store's markups actually increased over the period between 1982 and 1984, and that Sencer's use of the 1985 figures therefore overstated the income received in prior years. In support of his assertion that the markups varied, Cebollero testified at trial that he had increased his markups during the relevant period to offset lost sales volume and profits caused by the phasing-in of Maryland's newly raised drinking age. Further, Appellant's cross examination of agent Sencer revealed that, contrary to her testimony on direct examination, Cebollero had previously stated that his markups varied during an interview conducted as part of the criminal investigation.

In Foster v. Commissioner, 391 F.2d 727 (4th Cir.1968), we stated that:

The burden of proof is on the Commissioner to show that the taxpayer received income. This burden is initially satisfied, however, by the fact that the Commissioner's deficiency determination is presumed correct. The burden is thus on the taxpayer to prove the incorrectness of the deficiency determination. This burden is procedural and is met if the taxpayer produces competent and relevant evidence from which it could be found that he did not receive the income alleged in the deficiency notice. If this burden is met, the burden of proof shifts back to the Commissioner to prove the existence and amount of the deficiency.

Id. at 735. (Footnotes omitted). Relying on Foster, Appellant now argues that his evidence on the issue of the variation of Belby's markups constituted "competent and relevant evidence" sufficient to require the tax court to place the burden of proof on the Commissioner. Because the burden-shifting mechanism outlined in Foster, and its "competent and relevant evidence" standard have been frequently misunderstood, some explanation of our approach to the burden of proof in deficiency cases is in order. See United States v. Pomponio, 635 F.2d 293, 297 n. 4 (4th Cir.1980).

It is well established that as a general matter, the Commissioner's determination of deficiency is presumed correct, and the taxpayer bears "the burden of proving it wrong." Welch v. Helvering, 290 U.S. 111, 115, 54 S. Ct. 8, 9, 78 L. Ed. 212 (1933); Tax Ct.R.Prac. & Proc. 142(a). In Helvering v. Taylor, 293 U.S. 507, 515, 55 S. Ct. 287, 291, 79 L. Ed. 623 (1935), the Court refined its formulation of this burden, holding that the taxpayer is entitled to have the determination set aside if he can prove that it is "arbitrary and excessive," regardless of whether he can prove the amount he actually owes. Under the Taylor decision, the determination of the correct amount of tax is an inquiry entirely separate from, and subsequent to, the finding that the Commissioner's determination of deficiency should be rejected as arbitrary and excessive.

While Taylor stopped short of identifying the party upon whom the burden of proof rests after the taxpayer has successfully shown the determination to be arbitrary, the Courts of Appeal have generally placed that burden on the Commissioner. See United States v. Janis, 428 U.S. 433, 442, 96 S. Ct. 3021, 3026, 49 L. Ed. 2d 1046 (1976); Portillo v. Commissioner, 932 F.2d 1128, 1133 (5th Cir.1991); Muserlian v. Commissioner, 932 F.2d 109, 112 (2d Cir.1991); Zuhone v. Commissioner, 883 F.2d 1317, 1325 (7th Cir.1989); Clapp v. Commissioner, 875 F.2d 1396, 1403 (9th Cir.1989). In Higginbotham v. United States, 556 F.2d 1173, 1175 (4th Cir.1977), we expressly adopted this approach, holding that "a taxpayer need only prove by a preponderance of the evidence that the Government's assessment is erroneous, and, once that burden is met, the Government, not the taxpayer, must prove how much the taxpayer actually owes." Id. In other words, in the first phase of a deficiency suit, the issue is the arbitrariness of the Commissioner's determination, and the taxpayer bears the burden of persuasion by a preponderance of the evidence. That burden remains with the taxpayer, and never shifts to the government. If the taxpayer proves that the determination is arbitrary, the presumption of correctness vanishes. The trial then enters a second phase4 in which the issue is the correct amount of the deficiency. The government bears the burden of persuasion during this second phase.

In a series of decisions beginning with Stout v. Commissioner, 273 F.2d 345 (4th Cir.1959) (Haynsworth, J.), this Court developed a separate, subordinate procedural scheme under which the burden of production during the first phase of the suit can shift to the government. In Stout, we held that:

The presumption of correctness is procedural. It transfers to the taxpayer the burden of going forward with evidence, but it disappears in a proceeding to review the assessment when substantial evidence contrary to the Commissioner's finding is introduced. Thereafter, the Tax Court, in such a proceeding, must make its own findings based on the evidence before it, and we may affirm only if the findings of the Tax Court are supported by substantial evidence in the record of that proceeding.

Id. at 350.

Foster v. Commissioner, supra, later revisited the question of the means by which the taxpayer may overcome the "presumption of correctness."5 See ante. While that decision purported to follow Stout, the language in Foster quoted by the Appellant tends to support his argument that a taxpayer automatically proves the incorrectness of the Commissioner's determination, and thereby propels the trial to the ultimate issue of the amount of the deficiency, merely by producing evidence from which it could be found, that he did not receive the alleged income.6 Appellant's argument, however, is inconsistent with Stout, which held that once the taxpayer satisfies his initial procedural burden, the tax court must make its own findings on the issue of arbitrariness, based on all the evidence before it. Appellant's position is also incompatible with Helvering v. Taylor, supra, which affirmatively placed the burden of proof, not merely the burden of production, on the taxpayer.

In Higginbotham v. United States, supra, Chief Judge Haynsworth, the author of Stout, reformulated the approach taken in Foster:

... [t]he discussion of the burden of proof in Foster applies only to the procedural effects of the presumption that an assessment is accurate. Once a taxpayer has introduced evidence sufficient to support a finding that the assessment is wrong, Foster prevents the Government from simply resting on the presumption and requires it to come forward with some evidence to support a conclusion that the assessment is correct in spite of the taxpayer's evidence. But the taxpayer continues to bear the risk of nonpersuasion. Foster does not relieve the taxpayer of the burden of proving the government's assessment wrong by a preponderance of the evidence.

Higginbotham, 556 F.2d at 1176. In effect, Higginbotham revived Stout to the extent that it had been modified by Foster by making clear that the taxpayer's initial production of the required quantum of evidence does not relieve him of the burden of persuasion.7 Higginbotham has not been uniformly interpreted as reaffirming Stout. In Anastasato v. Commissioner, 794 F.2d 884, 887 (3d Cir.1986), the Court cited Stout for the proposition that the ultimate burden as to the amount of a deficiency can "shift" to the Commissioner. The Court erroneously cited Higginbotham, however, as standing for the view that the burden of persuasion remains on the taxpayer even after the "presumption of correctness" has been overcome. It is worth emphasizing that there is no such inconsistency in our cases. In this Circuit, the Commissioner always has the burden of persuasion as to the amount and existence of any deficiency. Before the Commissioner is required to carry that burden, however, the taxpayer first must dispense with the so-called presumption of correctness by carrying his own burden and persuading the court by a preponderance of the evidence that the assessment is arbitrary and excessive. The "shifting" mentioned in both Stout and Higginbotham refers only to the taxpayer's burden of production in the first phase of the trial.

The tax court held that the "burden of going forward" never shifted because Cebollero failed to produce evidence "establishing" the incorrectness of the determination. Appellant has attacked this ruling as impermissibly stringent in light of Foster.

The tax court indicated its understanding of the burden of production as follows: "where the notice of deficiency is found to be arbitrary, the burden of going forward with the evidence is shifted to respondent." Cebollero v. Commissioner, 60 T.C.M. (CCH) at 1384. This statement, combined with the court's citation to Taylor, makes it clear that the court used the phrase "burden of going forward" in reference to the taxpayer's burden of persuasion on the issue of the arbitrariness, and not in reference to the burden of production in Higginbotham. A review of the trial transcript and the tax court's opinion as a whole, however, reveals that aside from this semantic misunderstanding, the tax court properly allocated the burdens of proof.8

At the close of Appellant's evidence, counsel argued that he had "established initially that based on the record the notice of deficiency appears to be arbitrary and excessive." The court then proceeded to hear the government's evidence on the manner in which it determined the deficiencies. As discussed above, we have described the evidence required to satisfy the taxpayer's burden of production in Stout as "substantial evidence contrary to the Commissioner's finding;" in Foster as "competent and relevant evidence from which it could be found that he did not receive the income alleged;" and in Higginbotham as "evidence sufficient to support a finding that the assessment is wrong." Each of these statements simply requires the Commissioner to come forward with evidence of the reasonableness of its determination once the taxpayer comes forward with evidence which, if believed, could support a finding that the determination is arbitrary and excessive. Only if the court affirmatively finds, in light of both parties' evidence, that the taxpayer has proven the determination arbitrary, does the trial proceed to the second question of the existence and amount of the deficiency. Appellant received the full benefit of the shifting burden of production when the tax court required the government to proceed with its case.

The tax court's holding that neither "the burden of proof," nor the "burden of going forward," shifted to the government, amounts to a finding that Cebollero did not carry his burden of persuasion at the close of the evidence. The tax court was required to find facts based on conflicting testimony. In deciding that Cebollero had failed to meet his burden of production, the tax court expressly found that the taxpayer's own uncorroborated, self-serving trial testimony that he changed his markups was not as credible as the statement he made to revenue agent Sencer during the initial interview that the markups remained constant. The tax court's assessment of the credibility of witnesses and its findings of fact are entitled to substantial deference and "are not subject to reversal unless clearly erroneous." Gatling v. Commissioner, 286 F.2d 139, 141 (4th Cir.1961). The Supreme Court has held that "[w]here there are two permissible views of the evidence, the factfinder's choice between them cannot be clearly erroneous." Anderson v. Bessemer City, 470 U.S. 564, 574, 105 S. Ct. 1504, 1511, 84 L. Ed. 2d 518 (1985). We therefore affirm the tax court's finding that the Commissioner's determination was reasonable, and its decision not to require the government to present further proof on the amount of the deficiency.V.

Appellant also claims that the tax court erred by considering evidence of the revenue agent's testimony regarding her telephone conversation with Board. We find his hearsay and relevance arguments to be without merit.

Agent Sencer testified extensively about the methods she used to reconstruct Belby's income, which included an independent verification of the markup figures supplied by Cebollero. She testified that she made the call to the Board for the purpose of obtaining an accurate list of the prices Belby paid for the goods it sold. The tax court admitted this testimony, after careful consideration, for the limited purpose of showing the reasonableness of her method, and not the accuracy of information.

It is well recognized that a determination of deficiency "may often rest on hearsay or other inadmissible evidence," and that the court need not examine this underlying evidence in determining whether a deficiency notice is arbitrary and excessive. See Jackson v. Commissioner, 73 T.C. 394, 400 (1979); Zuhone v. Commissioner, 883 F.2d 1317, 1325 (7th Cir.1989). The issue before the tax court was not the amount owed but rather the reasonableness of the determination, and the government bore the burden of production. The fact of the telephone conversation was therefore relevant without regard to the truth of the statements made by the Board, and it was not improper for the tax court to consider the testimony for this limited purpose.

Accordingly, the decision of the tax court is hereby affirmed in all respects.

AFFIRMED.