OPINION
Mаnuel Cebollero, the taxpayer, appeals from the tax court’s decision sustaining the *988 Commissioner’s determination of income tax deficiencies for tax years 1982-1984. We now affirm.
I.
The facts are set forth in detail in the opinion below,
Cebollero v. Commissioner,
Appellant and his wife owned a beer and wine store (“Belby”) in Montgomery County, Maryland, which they operated as genеral 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 Cebolle-ro inadequate, the revenue agent attempted to reconstruct Belby’s income independently. Because of the cash nature of Bel-by’s business, Sencer chose a percentagе 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 indicatеd 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 approximаtely 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 *989 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 yеars 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, Cebollеro 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.
II.
Burden of Proof
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,
In
Foster v. Commissioner,
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 сould 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
*990
Belby’s markups constituted “competent and relevant evidence” sufficient to require the tax court to plаce 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,
It is well established that as a general matter, the Commissioner’s determination of deficiency is presumed cоrrect, and the taxpayer bears “the burden of proving it wrong.”
Welch v. Helvering,
290 U.S. Ill, 115,
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,
III.
In a series of decisions beginning with
Stout v. Commissioner,
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 “presump *991 tion of correctness.” 5 See ante. While that decision purported to follow Stout, the language in Foster quoted by the Appellant tends to suppоrt 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 prеsumption 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 nonper-suasion. Foster does not relieve the taxpayer of the burden of proving the government’s assessment wrong by a preponderance of the evidence.
Higginbotham,
IV.
The tax court held that the “burden of going forward” never shifted because Ce- *992 bollero 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,
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 reсeive 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 аffirmatively 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 сhanged 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,
*993 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 Bel-by’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 acсurate list of the prices Bel-by 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,
Accordingly, the decision of the tax court is hereby affirmed in all respects.
AFFIRMED.
Notes
. Cebollero pled guilty to filing a false income tax return for the tax year 1983.
. Belby was required to purchase all of its beer and wine from the Board.
. Cebollero attempted to attack the reconstruction of income with the claim that it failed to account for losses due to spoilage and theft. In response, the agent allowed a deduction of $5000, which Cebollero agreed was a fair figure.
. In
Helvering v. Taylor, supra,
the Court described the stage following proof of arbitrariness as a separate proceeding: ‘‘[t]hen, upon appropriate application that further hearing be had, [the court] should have heard evidence to show whether a fair apportionment might be made and, if so, the correct amount of the tax."
Id.,
.As several courts have noted, the use of the term "presumption” in connection with the taxpayer’s burden of proof has contributed to the confusion in these cases.
See United States v. Pomponio,
. As noted in
Pomponio, supra,
our opinions in
Stout
and
Foster
were the “subject of some criticism."
See, e.g., United States v. Rexach,
. At oral argument, Appellant suggested that because this is a deficiency suit, Higginbotham, which arose in the context of a refund claim, should not govern. We find this contention unpersuasive, in light oí Higginbotham’s specific references to Stout and Foster.
. The tax court’s phraseology is understandable in light of the way in which other Circuits have аddressed the issue. The Third Circuit, for example, held in
Demkowicz v. Commissioner,