United States v. Lenamond

MEMORANDUM OPINION BUCHMEYER, District Judge. The defendant, Lee Eugene Lenamond, was convicted of income tax evasion for 1973 and 1974. His prosecution was based upon the “bank deposits-cash expenditures” method of proof. 1 Lenamond’s motion for acquittal presents this question: Did the government fail to conduct a full and adequate investigation, and did it fail to follow reasonable leads, concerning the value of Lenamond’s business inventory — and, consequently, his business deductions for “cost of goods sold” 2 — for the years 1973 and 1974? Because the government did not conduct a full and adequate investigation and did not follow reasonable leads, despite inventory figures which were truly astonishing, the bank deposits method of proof was not sufficient. Therefore, the motion for acquittal is granted and Lenamond’s conviction is set aside. The Legal Principles At the conclusion of the government’s case, and again at the end of the evidence, the defendant moved for acquittal. Decision on this motion was reserved, and the case was submitted to the jury. Fed.R. Crim.P. 29(b). Following the return of a jury verdict which found the defendant *854 guilty on both counts of tax evasion, the motion for acquittal was timely renewed. The controlling legal principles concerning the “two traditional indirect methods of proof” used by the government in income tax evasion cases — the net worth analysis and the bank deposits-cash expenditures method — are stated in United States v. Dwoskin, 644 F.2d 418 (5th Cir.1981); United States v. Normile, 587 F.2d 784 (5th Cir.1979); and United States v. Boulet, 577 F.2d 1165 (5th Cir.1978) — and, of course, in Holland v. United States, 348 U.S. 121, 75 S.Ct. 127,99 L.Ed. 150 (1954). As discussed in Dwoskin: “A motion for acquittal must be granted ‘when the evidence is such that a reasonably minded jury must have a reasonable doubt as to the existence of any element of the crime.’ United States v. Slone, 601 F.2d 800, 803 (5th Cir.1979); United States v. Pinner, 561 F.2d 1203, 1207 (5th Cir.1977). In evaluating a claim of insufficient evidence according to this standard, we must consider the evidence in the light most favorable to the government, Glasser v. United States, 315 U.S. 60, 80, 62 S.Ct. 457,469, 86 L.Ed. 680 (1942), resolving reasonable inferences and credibility choices in support of the jury’s verdict, United States v. Henderson, 588 F.2d 157, 161 (5th Cir.1979); United States v. Juarez, 566 F.2d 511, 513 (5th Cir.1978) ... “To prove its case, the government relied upon circumstantial evidence [there, a net worth analysis]. Since circumstantial evidence is to be treated no differently than direct evidence, Holland v. United States, 348 U.S. 121, 140, 75 S.Ct. 127, 137, 99 L.Ed. 150 (1954), the test for judging the sufficiency of the evidence is the same whether the evidence is direct or circumstantial, United States v. Bright, 550 F.2d 240, 242 (5th Cir.1977); United States v. Gomez-Bajos, 507 F.2d 1213, 1221 (5th Cir.), cert. denied, 423 U.S. 826, 96 S.Ct. 41, 46 L.Ed.2d 42 (1975).” (644 F.2d at 420.) However, in Holland, the Supreme Court warned that the net worth method of proof is “so fraught with danger for the innocent that the courts must closely scrutinize its use” (348 U.S. at 125, 75 S.Ct. at 130). This is equally true with respect to the bank deposits-cash expenditures analysis. Accordingly, in Boulet, the Fifth Circuit emphasized that both methods trigger special protections for the accused and particularly careful scrutiny by the courts. 3 “We, therefore, review the record ‘bearing constantly in mind the difficulties that arise when circumstantial evidence as to guilt is the chief weapon of a method that is itself only an approximation.’ Holland v. United States, 1954, 348 U.S. 121, 129, 75 S.Ct. 127, 132, 99 L.Ed. 150. The government must prove a full and adequate investigation in a bank-deposits case just as it must in a net-worth case. Holland v. United States, supra. ‘Such investigation must establish a guarantee of essential accuracy in the circumstantial proof at trial as an element of the government’s burden of proving guilt beyond a reasonable doubt....’ United States v. Slutsky, supra, 487 F.2d [832] at 840 [ (2nd Cir.1973) ].” (577 F.2d at 1168.) As part of this duty to conduct “a full and adequate investigation in a bank deposits case,” the government may not disregard any “explanations of the defendant reasonably susceptible of being checked.” *855 United States v. Boulet, supra 577 F.2d at 1169. As the Supreme Court held in Holland: “... When the Government rests its case solely on the approximations and circumstantial inferences of a net worth computation, the cogency of its proof depends upon its effective negation of reasonable explanations by the taxpayer inconsistent with guilt. Such refutation might fail when the Government does not track down relevant leads furnished by the taxpayer — leads reasonably susceptible of being checked, which, if true, would establish the taxpayer’s innocence. When the Government fails to show an investigation into the validity of such leads, the trial judge may consider them as true and the Government’s case as insufficient to go to the jury.” (348 U.S. at 135-36, 75 S.Ct. at 135) (emphasis added). However, a full and adequate investigation is required, not a “universal probe.” The government is not required “to perform the impossible” (Dwoskin, 644 F.2d at 423) ... or “to bay down rabbit tracks” and “follow a trail that might have led nowhere” (Normile, 587 F.2d at 786) ... or to conduct a “bacteriophobic search for error” (Normile, 587 F.2d at 787). The government’s duty to investigate and to follow leads does apply to omitted or understated deductions — particularly, in this case, to the defendant’s inventory and his business deductions for cost of goods sold. This is evident from several cases, 4 including United States v. Hall, 650 F.2d 994 (9th Cir.1981), where the Ninth Circuit discussed this very question: “Inventory Cost: Hall and Uranga argue that the Government’s figures did not accurately reflect their inventory cost, and consequently yielded inaccurate figures for ‘cost of goods sold.’ Appellants point out that the cost of goods sold is part of the Government’s calculation of the business’s gross profit and directly related to the income of the business. Thus, inventory valuation becomes an essential part of the Government’s case. “When choosing to proceed against a defendant using the net worth or bank deposits methods of proof, the Government assumes a special responsibility of thoroughness and particularity in its investigation and presentation. Holland, 348 U.S. at 135-36 [72 S.Ct. at 135],.. This responsibility imposes the duty to, inter alia, accurately establish the figures upon which the methods are based, and to reasonably investigate leads which may reveal that the defendants properly reported their income... Here, both this duty to investigate and the duty to establish figures with reasonable certainty are implicated by the Government’s treatment of Hall and Uranga’s inventory valuation. In particular, the Government must show that it had followed through on appellant Hall’s prior notation suggesting that the inventory figures used by the Government were too high. This is a possible explanation for the apparent unreported income and may not be overlooked by the Government.” (650 F.2d at 999-1000) (emphasis added). After a careful review of the evidence in this case (including the evidence concerning the government’s investigation and the defendant’s inventory), and after applying the legal principles just discussed, this Court is convinced that the Holland protections have been violated — and that, consequently, the conviction based upon the bank deposits method of proof must be reversed. The Investigation In July of 1975, Randell Choate, 5 an IRS field agent, began a civil tax investigation *856 of the defendant Lenamond — who owned and operated an auto supply store in a low income area in Dallas, Texas (Choate, p. 136; defendant’s Exhs. 3-18). 6 This was the first fraud investigation conducted by Choate (Choate, p. 61). Over the next 18 months, Choate (sometimes accompanied by other IRS agents) had several meetings 7 with the defendant Lenamond and with the CPA representing Lenamond (Robert Driegert), and completed a bank deposits-cash expenditures analysis. Lenamond was very cooperative in the investigation and, with only one exception, 8 supplied Choate with all of the information he requested — although admittedly Lenamond had no accounting experience or ability, had never “gotten any advice on how to maintain books and records,” and kept “crummy” books (Choate, pp. 119, 122). 9 In June of 1976, the investigation was converted to a joint criminal and civil investigation (Choate, p. 82). On October 6, 1976, Choate held the first criminal investigation meeting with Lenamond — although he did not give any notice of this meeting to the CPA representing Lenamond (Choate, pp. 62-63). Lenamond still continued to cooperate. The investigation concluded in 1976, and it resulted in a recommendation that Lenamond be indicted on income tax evasion charges for 1972,1973, and 1974. Although Lenamond had no substantial assets or expenditures that pointed toward substantial unreported income — he lived in a $13,000 house in Pleasant Grove; he had no fancy clothes, expensive cars, jewelry, stocks or bonds, hidden bank accounts, etc.; and he owed money on a loan from his father (Choate, pp. 150-53) — he did withhold cash from his daily business receipts for personal living expenses, for payment of salaries, and for occasional payments of business expenses (Choate, pp. 4-5). And, Choate’s *857 bank deposits-cash expenditures analysis, indicated that Lenamond’s bank deposits and cash expenditures exceeded his reported gross receipts — specifically, that Lenamond may have had unreported income of $28,631 in 1973 and $29,388 in 1974 (Robert S. Driegert transcript, pp. 33-34; see Govt. Exh. 52). 10 No action was taken on this recommendation for over three years, and limitations ran on the year 1972. Finally, on April 15, 1980 — the last day of the limitations period for 1973, and almost five years after the investigation started —the defendant Lenamond was indicted for income tax evasion in 1973 and 1974. The indictment charged that additional taxes of $9,796.00 were due in 1973 and that an additional $12,839.80 was due in 1974; at trial, the government reduced these claims to $6,850.82 in 1973 and $9,233.72 in 1974 (see Govt.Exh. 52). The Inventory The only aspect of the government’s bank deposits-cash expenditures investigation that is under attack by the motion for acquittal concerns the inventory at Lenamond’s auto parts store. The defendant’s 1973 income tax return reported a beginning inventory of $40,-884.00 and a year-end inventory of $71,-864.00 (Govt.Exh. 2). The defendant’s 1974 return reported beginning and ending inventories of $71,864.00 and $97,184.00, respectively (GovtExh. 3). However, the defendant’s purchases increased only 2% during this two-year period when, according to the tax returns, inventory increased 188%. And, the combined increase in inventory over the two-year period ($56,350) is just slightly less than the amount of unreported gross receipts ($58,000) claimed by the government’s bank deposits analysis (see GovtExh. 52). The defendant contends that these “astonishing” inventory figures were “arbitrarily arrived at by Mr. Lenamond in ignorance 11 of the tax effect that inflated inventory figures had on his profit. (If ending inventory is artificially inflated, then profits are artificially inflated as well.)”— and that: “The figures themselves, comparing ending inventory which increased 138% [in 1973 and 1974] to purchases which increased 2%, sales which increased 13.7%, and cost of goods sold which decreased 12%, would have lead any reasonable investigator to conclude that something was wrong in the ending inventory figures.” The basic facts concerning IRS Agent Choate’s investigation of the defendant’s inventory — or lack of investigation — are undisputed. Choate determined that the defendant did take inventory yearly, but that he made no itemized inventory list and kept no other inventory records. Accordingly, Choate examined the CPA’s work copy of the defendant’s 1973 return (Choate, pp. 107-110). The ending inventory was first recorded as $43,480 — an increase of less than $3,000 over 1972. However, based upon information given to the CPA by Lenamond, this was changed to $52,278, and then increased a third time, to $71,873. Choate testified: “Q. And those crossed out parts have the effect of increasing Lee’s taxable income dramatically from what had been originally in there, do they not? “A. Yes, sir, it increases it. “Q. How? “A. By causing a larger ending inventory, decrease of goods sold and increases gross profit. “Q. Any accountant knows that? “A. Yes, sir. “Q. What is the effect of having an ending inventory go from 43 to 71 thousand dollars? *858 “A. It would be to increase income by like 28 thousand. “Q. Didn’t it ever occur to you, Mr. Choate, that Lee Lenamond might have been telling you the truth when he said he couldn’t possibly have made that much money that the reason for it was that he had overstated his ending inventory? “A. He never told me that he overstated his ending inventory. “Q. You never looked around for things that might help him only for things that might hurt him, isn’t that true? “A. I just looked at the records he had. “Q. And those records didn’t alert you that something was wrong with the inventory? “A. It looked like he was having problems determining what the inventory was.” (Choate, pp. 108-110) (Emphasis added). 12 Choate also examined the defendant’s tax returns for 1973 and 1974 (Govt.Exhs. 2, 3). They revealed that purchases “remained very constant” during this period, increasing about 2% (Choate, pp. 113-14) — but that the year-end inventory increased 138% (from $40,884 in 1972 to $71,864 in 1973 to $97,184 in 1974). This meant that something was drastically wrong with the inventory figures. 13 Choate so testified: “Q. From here to here purchases went up two percent, right? “A. Yes, sir. “Q. Okay. From here to here ending inventory went up 138 percent, can you explain that to the ladies and gentlemen of the jury? “A. No, sir. “Q. There is no explanation for it, is there? “A. I cannot explain it, no, sir. “Q. At the same time the government says his sales were increasing? “A. Yes, sir. “Q. How can his sales be increasing if his purchases are remaining more or less constant and the ending inventory, that is the goods he has left over at the end of the year is increasing a 138 percent? “A. I don’t know, sir. “Q. It can’t happen, can it? “A. It doesn’t appear to, sir. “Q. Something is wrong, isn’t there? “A. Yes, sir. “Q. And something is wrong in Lee’s favor, isn’t it? “A. It would appear to be.” (Choate, pp. 115-16) (emphasis added). Despite this, Choate made no further investigation. He did not ever ask Lenamond if a mistake had been made about the inventory (Choate, p. 119) — even though he knew that Lenamond had no accounting or bookkeeping experience, had not graduated from high school until he was 22, 14 had no professional help regarding his records, had made frequent mistakes on his bookkeeping and bank deposits, had “crummy books,” and had a parts store in such a state of disarray (defendant’s Exhs. 3-18) that “there might be some difficulty” in taking inventory (Choate, pp. 119-22, 163). Choate did ask Lenamond a compound question at the November 5, 1976 meeting: could he “explain why he showed a substantial increase in inventory in 1974 and why his cancelled checks lacked quite a bit equalling his expenses claimed on his return for that year.” (Choate, pp. 29-30, 37-38). However, Lenamond answered only the second part of this question: *859 “A. I believe we asked the question if he had any substantial increase in inventory. “Q. That was November 5th, 1976, right? “A. Yes, sir, I believe so. “Q. And let me read what your memo reflects. ‘Lenamond was again shown his 1974 income tax return which reflected a substantial increase in inventory and purchases about 20 thousand dollars in excess of those substantiated by checks. Lenamond stated the figures shown for purchases on the return is correct and anything not substantiated by checks would be cash purchases.’ He didn’t say a darn thing about his inventory, did he? “A. No, sir. “Q. Y’all didn’t question him about his inventory when he didn’t answer your two part question? You asked him two things, didn’t you? “A. Yes, sir. “Q. You asked him about inventory and purchases, right? “A. Yes, sir. “Q. He answered you about purchases but he forgot to saying anything about inventory? “A. Yes, sir. “Q. And you didn’t ask him, you didn’t follow up, you didn’t want to know, did you? “A. We didn’t ask him any other questions, I don’t believe.” (Choate, pp. 117-18). Choate conceded during testimony that he “just assumed the [inventory] figures from 40 thousand to 71 thousand to 97 thousand dollars were right” — “notwith standing all the indications that [he] had that they were wrong” (Choate, p. 133). And, by letter dated June 11, 1976, Agent Choate reported: “No inventory sheets are kept by Mr. Lenamond, but the inventory seems proper for the two stores and any attempts by him to claim a lesser amount could be overcome.” 15 Even after Lenamond’s attorney, Donald J. Forman, advised IRS attorneys on May 28, 1978 that there was an inventory problem— and that the inventory did not appear to be kept at a fair market value 16 — there was no further investigation. Choate did attempt to explain that he accepted the defendant’s inventory figures because he had no way to check them: the defendant had no inventory records or lists and “without such records it would be impossible to disprove the stated inventory figures.” (Choate, p. 181; Government’s Opposition to Defendant’s Motion for Judgment of Acquittal, pp. 9-10). 17 However, he admitted that he “probably would” have investigated the inventory figures if the situation had been reversed — i.e., if the ending inventory had decreased from $97,-000 to $40,884 in two years — since this would indicate that the taxpayer might be avoiding income taxes by undervaluing inventory: “Q. In connection with the service as an internal revenue agent have you ever had occasion to check a business that had undervalued its ending inventory? “A. I don’t believe I have sir. I can’t remember. “Q. Let’s just say for a second that instead of the inventory increasing from 40 to 71 to 97 it had decreased from say 97 down to 71 to 47. What would the effect of that have been? “A. To reduce the increase or it would — it would increase his cost of goods sold and decrease gross profit. *860 “Q. Wouldn’t you be immediately tipped off that something was wrong and wouldn’t you investigate that? “A. I probably would, I don’t know, sir. “Q. Okay. Because it is your job to determine the correct amount of tax, right? “A. Yes, sir. “Q. If somebody was decreasing their inventory they might not be paying the correct amount of tax, correct? “A. That’s a possibility. “Q. It would at least alert your suspicions? “A. Yes, sir.” (Choate, pp. 178-79). The Holland Consequences Under these facts and under Holland 348 U.S. 121, 75 S.Ct. 127, 99 L.Ed. 150, the government had a duty to conduct a full and adequate investigation and follow leads to determine whether or not Lenamond’s inventory figures were too high. United States v. Hall, 650 F.2d at 999. The inventory figures were truly anomalous: inventory could not increase 138% in two years, while purchases increased only 2% — partieularly at the very time when, according to the government’s bank deposits proof, the defendant’s sales were increasing by 12%. IRS Agent Choate knew this and he knew that something was wrong with the inventory figures (Choate, pp. 115-16). 18 He knew that the defendant “was having problems determining what the inventory was” (Choate, p. 110). He knew that the defendant’s store was in such a state of disarray that “there might be some difficulty” in taking inventory (Choate, p. 163). He knew that the defendant was not sophisticated and had no professional help with his records and “crummy books” (Choate, pp. 119-22). Despite these “leads” 19 — and despite the additional, express “lead” concerning inventory furnished by the defendant’s attorney 20 — the government simply accepted the inventory figures, without any investigation, “notwithstanding all indications ... that they were wrong” (Choate, p. 133). It was possible for the government to investigate the inventory figures, 21 and this would not have been “a rabbit trail leading to nowhere” or “a bacteriophobic search for *861 error (Normile, 587 F.2d 787). Agent Choate could have begun simply by asking the defendant if the figures were too high; he never did so (Choate, p. 119). Nor did Choate interview the employees who helped take the inventory; presumably, both David Lenamond and Ronnie Kelley would have told Choate, as they testified at trial, that the inventory did not increase substantially in either 1973 or 1974. Choate could have learned from the CPA (Murray Hay) that he received such a small amount ($50— 60) for preparing Lenamond’s tax returns, that he did not verify the figures on them. And, since the defendant had cooperated fully in the investigation, Choate could have even requested that an inventory be taken in 1975 or 1976; this could have indicated whether the reported 1973 and 1974 figures were correct. 22 In addition, Choate could have used an indirect method — the industry gross profit percentage — to check Lenamond’s inventory figures. Cf. Bernstein v. Commissioner, 267 F.2d 879 (5th Cir.1959). Specifically, he could have determined the gross profit percentage of similar businesses in the same area and time, with the same merchandising policies — and then applied this percentage to the gross sales of Lenamond in 1973 and 1974 in order to approximate his inventory and his cost of goods sold for these years. 23 The government has used this indirect method in other cases 24 — but did not do so here. 25 At trial, the defendant’s evidence demonstrated how the industry gross profit percentage might have been applied if the Lenamond inventory figures had been investigated. In particular: ... David Lenamond, who had firsthand knowledge of the nature of his father’s business, and Jack Stoller, who had expert knowledge of similar businesses in the same area and time, both testified that the appropriate gross profit percentage for the defendant’s business was approximately 30 percent or less. ... Dr. Kenneth Ferris, the defendant’s expert, applied this 30% gross profit percentage to Lenamond’s sales in 1973 and 1974 (based upon the government’s allegations of gross receipts) in order to estimate the inventory figures and the cost of goods sold deductions. ... Dr. Ferris then testified that, based upon his calculations, the inventory figures were much too high and that, in fact, the defendant overpaid his taxes for 1973 in the amount of $867.10 and underpaid his taxes for 1974 in the amount of $588.24, so that the government owed Lenamond $278.86 for the two year period. 26 *862 Dr. Ferris also testified that the defendant’s gross profit margins for previous years ranged from a low of 21.6% to a high of 24.2% — and thus never reached the 30% “industry gross profit percentage” used in his calculations — while, under the government’s bank deposits calculations reflected in the indictment, Lenamond would have had a profit margin of 42.5% in 1973 and 40.6% in 1974. The government does not dispute the fact that this indirect method could have been used to check the correctness of the Lenamond inventory figures. It argues, however, that this percentage method is merely “an estimation technique dependent on correct gross receipts, among other variables, and does not compute specific inventories.” But the very method of proof upon which the government prosecuted the case (bank deposits-cash expenditures) is an “estimation technique” — and Holland warns lower courts about “the difficulties that arise when circumstantial evidence as to guilt is the chief weapon of a method that is itself only an approximation” (348 U.S. at 129, 75 S.Ct. at 132). United States v. Boulet, supra 577 F.2d at 1168. The government also contends that it was not obligated to use the industry gross profit percentage analysis since the defendant admitted that “all deductions” on his tax returns were correct. Yet, IRS Agent Choate never asked Lenamond if the inventory figures were mistaken — even though he knew that the inventory figures were anomalous and that something was wrong with them. Therefore, any general statement that deductions were correct did not negate the government’s duty to investigate the inventory figures — just as it could have done by various means, including the gross profit percentage method. 27 Finally, the government contests the 30% gross profit percentage used the calculations by defendant’s experts. It argues that the jury could have discredited the testimony of David Lenamond and found that 35% was the correct figure 28 — and that cross-examination of Mrs. Sharon Lake Liddy (defendant’s second expert) established that a 35% gross profit margin would still result in “a substantial tax due and owing” by the defendant. This argument, too, is erroneous. The issue is not whether the profit margin was 30% or 35%; it is the total failure of the government to use the industry gross profit percentage to determine if the astonishing inventory figures were too high, or to investigate them in any other manner. And, under Holland, when the government fails to show an adequate investigation, the case should not be submitted to the jury— and, when the government fails to track down reasonable leads, “the trial judge may consider them true and the government’s case as insufficient to go to the jury” (348 U.S. at 135-36, 75 S.Ct. at 135). The Conclusions This is the very type of ease contemplated by Holland. 29 The government should have conducted a full and adequate investigation of the astonishing inventory figures, and it should have followed leads which indicated that the figures were erroneous. It did not do so. Since the investigation was not adequate, this case should not have been submitted to *863 the jury. In addition, since the government did not follow reasonable leads, it is appropriate for this court to accept the defendant’s position that a proper investigation— including the use of the gross profit percentage method — would have revealed that the inventory figures were too high and that, consequently, no substantial taxes were due for 1973 and 1974. Accordingly, the motion for acquittal is granted and the conviction of Lee Eugene Lenamond is set aside. 30 Footnotes [1] “... Under this method, all deposits to the taxpayer’s bank and similar accounts in a single year are added together to determine the gross deposits. An effort is made to identify amounts deposited that are non-taxable, such as gifts, transfers of money between accounts, repayment of loans and cash that the taxpayer had in his possession prior to that year that was deposited in a bank during that year. This process is called ‘purification.’ It results in a figure called net taxable bank deposits. “The government agent then adds the amount of expenditures made in cash, for example, in this case, cash the taxpayer received from fees, did not deposit, but gave to his wife to buy groceries. The total of this amount and net taxable bank deposits is deemed to equal gross income. This is in turn reduced by the applicable deductions and exemptions. The figure arrived at is considered to be ‘corrected taxable income.’ It is then compared with the taxable income reported by the taxpayer on his return.” United States v. Boulet, 577 F.2d 1165, 1167 (5th Cir.1978). [2] A merchant selling goods is entitled to an income tax deduction for his cost of goods sold. Normally, this deduction is computed by adding the beginning inventory to purchases during the year to determine the goods available for sale, and then subtracting the ending inventory. The resulting figure is the “cost of goods sold” and is deducted from the merchant’s gross receipts to arrive at gross profit. Other expenses are then deducted from the merchant’s gross profit to determine the net profit reportable on his income tax return. Internal Revenue Code §§ 61, 62, and 63. [3] In Holland v. United States, 348 U.S. 121, 75 S.Ct. 127, 99 L.Ed. 150 (1954), the Supreme Court discussed a number of “dangers” presented by the net worth method of proof, concluding: “While we cannot say that these pitfalls inherent in the net worth method foreclose its use, they do require the exercise of great care and restraint. The complexity of the problem is such that it cannot be met merely by the application of general rules ... Trial courts should approach these cases in the full realization that the taxpayer may be ensnared in a system which, though difficult for the prosecution to utilize, is equally hard for the defendant to refute ... Appellate courts should review the cases, bearing constantly in mind the difficulties that arise when circumstantial evidence as to guilt is the chief weapon of a method that is itself only an approximation.” (348 U.S. at 129, 75 S.Ct. at 132.) [4] See United States v. Keller, 523 F.2d 1009 (9th Cir.1975); United States v. Shavin, 320 F.2d 308 (7th Cir.1963), cert. denied, 375 U.S. 944, 84 S.Ct. 349, 11 L.Ed.2d 273 (1963); and Beck v. United States, 298 F.2d 622 (9th Cir. 1962), cert. denied, 370 U.S. 919, 82 S.Ct. 1558, 8 L.Ed.2d 499 (1962). See also United States v. Fowler, 605 F.2d 181 (5th Cir.1979). [5] A transcript of Choate’s testimony at trial has been prepared; it will be cited in this opinion as “Choate, pp._” (although these page *856 numbers will, of course, differ from those in the transcript on appeal). [6] For a period of time beginning in 1971, the defendant owned a second auto supply store in Plano, Texas, which was operated by his son, Nolan Lenamond, but which never made a profit. [7] The meetings included those held on March 17, 1976, on May 19, 1976, on October 6, 1976, on November 5, 1976 and on November 23, 1976. Choate prepared memoranda of these meetings and, when called to testify at trial, had little independent recollection of the meetings, and had to refresh his memory by reading from the memoranda (Choate, pp. 67-68). [8] Although not relevant to this motion for acquittal, the government did contend that Lenamond promised to provide a “spiral notebook," but never did, and that this constituted evidence of wilfulness. This issue was hotly contested, and was fairly presented to the jury, but Choate did concede that Lenamond “stated consistently that he couldn’t find this spiral notebook,” and that Choate never had “any evidence or any facts .. that [Lenamond] destroyed it” (Choate, pp. 123, 163). It was also undisputed that the defendant offered several boxes full of purchase tickets, cash tickets, sales invoices, etc. — containing the same information that had been recorded in the spiral notebook — but that Choate chose not to examine these because it would be too time consuming. In particular, Choate testified: “Q. You know if you had gone over the spiral notebook there would be nothing in the spiral notebook that wasn’t in the seven boxes of purchases or sale tickets, don’t you? “A. There shouldn’t be. “Q. All right. You never even bothered to look at the sales tickets, did you? “A. I couldn’t. “Q. Why couldn’t you? “A. Because of the time restraints of trying to look at each ticket. “Q. Mr. Choate, the Government is trying to send this man to jail for ten years, isn’t that worth enough of your time to look at those tickets? “A. It wasn’t that time because I wasn’t in the criminal investigation.” (Choate, pp. 104-05). [9] Lenamond did not testify at trial. However, Dr. Murray — a clinical psychologist who specialized in learning disability disorders and who had examined the defendant — testified that Lenamond had a 6th grade reading comprehensive level and a 7th or 8th grade math comprehensive level; that he made “systematic errors” in simple arithmetic; and that, “in his opinion, Lenamond could not complete IRS tax forms anymore than any 11-12 year old.” Lenamond, who was 22 when he graduated from high school, paid a CPA (Murray Hay) $50-60 to prepare his 1972 and 1973 returns and paid a bookkeeper (Loretta Holbrook) $25-30 to prepare his 1974 returns. [10] These figures were approximations. In the indictment, they were changed to alleged unreported income of $26,997 in 1973 and $31,016 in 1974. At trial, the government’s proof claimed unreported income of $29,838 in 1973 and $31,205 in 1974 (Govt.Exh. 52). [11] The defendant did not testify at trial. His son, Nolan Lenamond — who ran the Plano store and who helped the defendant take inventory — testified that he did not know “that having too high an inventory was going to result in too high of profits.” [12] The CPA who made these different entries on the work copy of the 1973 return testified that Lenamond “said he reworked his inventory very carefully and they made a mistake and this was the correct figure” — but the CPA never asked to see the inventory records and never checked the inventory himself (Murray Hay, pp. 19-22). [13] And, it drastically reduced the defendant’s business deduction for cost of goods sold, and increased his income taxes, as shown by this partial table used during Choate’s cross-examination: 1972 1973 1974 Purchases $182,000 $178,000 $187,000 Ending Inventory 41,000 72,000 97,000 Cost of Goods Sold 183,000 146,700 161,000 [14] See footnote 9. [15] Affidavit of Howard A. Weinberger, dated June 19, 1980 (submitted with Defendant’s Motion to Set Aside Jury Verdict and Enter Judgement of Acquittal). [16] The evidence established that some of the inventory had been valued at the retail sales price or the jobber sales price, instead of cost, but that this was corrected before the 1973 return was filed (see footnote 12 and testimony of Murray Hay and Nolan Lenamond). [17] Another IRS agent who testified at trial as an expert witness (Oldham) agreed that something was wrong with the inventory figures, but said there were “no leads he could have followed.” [18] The defendant’s expert witness, Dr. Kenneth Ferris, testified that any auditor or accountant who did not know that the Lenamond inventory figures were erroneous would have been negligent — and that any reasonable accountant would have investigated the inventory figures as part of an audit. [19] The duty to investigate “leads” is merely part of the government’s duty to conduct a full and adequate investigation. Thus, even if the matters just discussed were not considered “leads” under Holland, the government still had a duty — because of them — to investigate the Lenamond inventory figures. See United States v. Boulet, 577 F.2d 1165, 1168 (5th Cir.1978) (the “government must prove a full and adequate investigation in a bank deposits case just as it must in a net worth case”); United States v. Hall, 650 F.2d 994, 999 (9th Cir.1981) (the government had both the “duty to investigate and the duty to establish figures with reasonable certainty” because of the taxpayer’s notation that the inventory figures were too high). [20] Contrary to the government’s arguments, the attorney’s statements about the inventory did constitute a “lead” under Holland (see testimony of IRS Agent Douglas Fortney) ... it was “timely furnished” since the statement was made more than two years before trial ... and, as discussed above, it was a “lead reasonably susceptible of being checked.” Moreover, the government had a duty to investigate the astonishing inventory figures irrespective of whether the attorney’s statements constituted a “lead.” See footnote 19. [21] IRS Agents are, in fact, instructed to investigate inventory figures in net worth cases where the taxpayer claims the figures on the return are too high. Paragraph 424.9(4) of the Special Agent’s Handbook, 5 CCH Internal Revenue Manual, states: “... To resolve this, the investigating officers should try and corroborate the inventory figures shown on the taxpayer’s returns by admissions of the taxpayer, statements of employees who took the inventory, copies of inventory records, etc.” Although this instruction applies to the value of inventory in net worth cases, it is just as relevant to the value of inventory in a bank deposits-cash expenditures investigation. United States v. Boulet, 577 F.2d 1165, 1168 (5th Cir.1978). [22] The government contends that “there was no reasonable way [it] could have calculated defendant’s ending inventories for the years 1973 and 1974, without his assistance." Yet, since the government did not investigate the inventory figures, it never asked for the defendant’s assistance — even though he had cooperated throughout the investigation. See footnote 8. [23] Dr. Kenneth Ferris, the defendant’s expert, described this indirect method — and testified that it was appropriate to use it to compute cost of goods sold when inventory records are missing or inaccurate. [24] Apparently, this method was used in the trial in United States v. Normile, 587 F.2d 784 (5th Cir.1979) (Defendant’s Response to Government’s Opposition to Post-Trial Motions, p. 4). And, paragraph 427.11 of the Special Agent’s Handbook, 5 CCH Internal Revenue Manual, states that although the gross profit percentage method “is not a prime method of proof and by itself would be of very little value in criminal cases,” this percentage method “is very useful for test checking; for corroborating the results obtained by some other means of proof such as ... bank deposits ...” (emphasis added) [25] If the situation had been reversed — i.e., if the Lenamond inventory had decreased from $97,000 to $40,884 in two years — the government could have used the industry gross percentage profit method to determine if the inventory had been undervalued in order to avoid taxes (Choate, pp. 178-7.'). Cf. Thor Power Tool Co. v. Commissioner, 439 U.S. 522, 528, 99 S.Ct. 773, 779, 58 L.Ed.2d 785 (1979). [26] Defendant also presented similar testimony and calculations from another expert witness, Mrs. Sharon Lake Liddy (who had served as an IRS field agent for 5 years). [27] A taxpayer may introduce evidence of his own improperly computed cost of goods sold to show both a lack of a tax deficiency and lack of willfulness in civil and criminal fraud cases. See Jenkins v. United States, 313 F.2d 624 (5th Cir.1963); Lee v. United States, 466 F.2d 11 (5th Cir.1972); and United States v. Cramer, 447 F.2d 210 (2d Cir.1971). [28] The 35% figure is not based upon a comparison of similar businesses, but upon “the relationship of the cost of goods sold and gross receipts reported on the 1973 return.” Thus, it would be reliable only if the anomalous inventory figures are correct. It was undisputed that the defendant’s gross profit margin in pri- or years never exceeded 24.2% (testimony of Dr. Kenneth Ferris). [29] Compare the investigation in this case with those discussed in United States v. Boulet, supra 577 F.2d 1165; United States v. Slutsky, 487 F.2d 832 (C.A. 2, 1973), cert. denied, 416 U.S. 937, 94 S.Ct. 1937, 40 L.Ed.2d 287 (1974); and United States v. Normile, supra 587 F.2d 784. [30] Because of this action, it is not necessary to consider the two additional matters raised by the defendant — that a new trial should be granted because of possible jury misconduct and because of alleged errors in the jury charge.