Oswill M. Cummings, Jr., sole proprietor of an intrastate trucking enterprise (Tuscaloosa Motor Express) operating under an Alabama Public Service Commission franchise, appeals1 from an adverse Tax Court decision establishing his federal income tax deficiencies and penalties at approximately $33,000.00 for the taxable years 1944 to 1955. Oswill M. Cummings, Jr., T. C. Memo. 1968-52. See 10 J. Mertens, Law of Fed. Income Taxation § 55.19, p. 4 (Jan. 1969 Supp.). We reverse and remand.

From 1945 through 1958, Cummings personally2 conducted a freight hauling business, operating motor trucks on a regularly scheduled Birmingham and Tuscaloosa route, as well as unscheduled "interline or interchange" hauls3 to other cities both in and out of Alabama.4

In the course of his business at least for the post-1956 period, Cummings maintained separate bookkeeping accounts for the three types of carriage in which he engaged: regularly scheduled, interline, and collect-on-delivery (C.O. D.).5 Cummings' practice was to deposit all collections (scheduled, interline, and C.O.D.) in his business account with City National Bank of Tuscaloosa. He would, thereafter, deduct his own hauling charge and remit by check the balance of each collection to the appropriate carrier, carriers or C.O.D. seller.

In February 1961, IRS agents commenced an investigation of Cummings' operations with a request for his 1959 tax return. Subsequently a request was made for Cummings' books and records for the years 1956 through 1959. There is a dispute still raging on appeal as to what response was made to the initial IRS record request.6 Regardless of which narration is more accurate, the IRS agents successfully marshaled the information they needed for the examination of Cummings' tax affairs from 1956 forward. They secured from his bank information reflecting all deposits to his business7 and personal checking accounts and savings accounts. They seized from the accountant who was preparing the requested annual profit and loss statements Cummings' business records, including cancelled checks, bank statements, depreciation records and tax-related papers.

After completion of the examination of Cummings' tax affairs for tax years 1956-1962, an IRS agent met with Cummings, his attorney and his accountant to discuss proposed adjustments. At this time, the agent informed Cummings and his advisors that an IRS search of its records reflected that he had made no income tax returns for 1944-1955.8 Thereafter, on two separate occasions, the agent requested Cummings' 1944-1955 business records. Standing on his construction of the alleged IRS agreement not to investigate pre-1956 tax years, Cummings' attorney advised the taxpayer not to cooperate with the agents. Consequently, the agents were forced to reconstruct Cummings' 1944-1955 tax deficiencies on the basis of nothing but his total annual deposits reflected in City National Bank records.

IRS Special Agent Mims, who reconstructed Cummings' 1944-1955 income without the aid of any information other than bank records, used the "bank deposits" computation method. Ordinarily, under the "bank deposits" method, the Commissioner is entitled to assume that all money deposited in a taxpayer's account during a given period constitutes taxable income since no starting point exists from which to use the preferred "net worth" method. See Price v. United States, 5 Cir. 1964, 335 F.2d 671, 677; cf. 10 J. Mertens, Law of Fed. Income Taxation § 55.19, pp. 123-124 (1965 rev. ed.). In the Cummings case, however, the peculiar bifurcated investigation of tax years 1956-1962 and 1944-1955 left the IRS in possession of all necessary records to determine actual taxable income for 1956-1962, yet with nothing but bank deposit data for 1954-1955. As a result of this situation, Agent Mims simply applied a 1956-1962 average net profits-to-deposits ratio9 to the 1944-1955 annual bank deposits to determine respective annual deficiencies, as well as the cumulative total liability with appropriate penalties. Because no business records for 1944-1955 were made available to Agent Mims, he testified that no adjustments were made to Cummings' annual bank deposits for those years to eliminate nontaxable items known to be included therein.10

The Tax Court held that Cummings failed to meet his burden of proving that the Commissioner's deficiency determinations for tax years 1944-1955 computed by the "bank deposits" method (without elimination of nontaxable items not specifically documented) was arbitrary and excessive. However, applying the Cohan rule,11 the Tax Court concluded that a ten per cent reduction in annual bank deposits would be a "reasonable" allowance for nontaxable amounts collected for C.O.D. and interline hauls.

Cummings contends that the Tax Court's allowance of only a ten per cent reduction in bank deposits for C.O.D. and interline collections was based upon the "clearly erroneous" factual finding that Tuscaloosa Motor Express only "occasionally" engaged in C.O.D. and interline hauling; and he argues that he adequately carried his burden to neutralize the presumption of validity attaching to the Commissioner's deficiency determination when he proved that the application of the 1956-1958 net income-to-deposits ratio to the 1944-1955 gross bank deposits was "arbitrary and excessive."12

We note at the outset that our scope of review in an appeal from the Tax Court is limited. See generally 9 J. Mertens, Law of Fed. Income Taxation §§ 51.22 through 51.27, pp. 47-94 (1965 rev. ed.). The Commissioner is required to compute taxable income "under the method of accounting * * * regularly used by the taxpayer," or, if no method was used, "under such method as * * * does clearly reflect income." Int.Rev. Code of 1954, § 446(a, b); 26 U.S.C.A. § 446 (a, b). See Webb v. Commissioner of Internal Revenue, 5 Cir. 1968, 394 F.2d 366, 371; Miller v. Commissioner of Internal Revenue, 5 Cir. 1956, 237 F.2d 830, 838. The Commissioner may use any reasonable computation method to determine the deficiency where the taxpayer fails to maintain or produce adequate records from which actual income might be ascertained. Boyett v. Commissioner of Internal Revenue, 5 Cir. 1953, 204 F.2d 205, 208; Kenney v. Commissioner of Internal Revenue, 5 Cir. 1940, 111 F.2d 374, 375; see Burka v. Commissioner of Internal Revenue, 4 Cir. 1950, 179 F.2d 483, 485. Among the recognized methods where no starting point is apparent is the "bank deposit" reconstruction method, wherein the Commissioner is entitled to assume that all bank deposits are taxable income in the absence of proof to the contrary. Price v. United States, 5 Cir. 1964, 335 F.2d 671, 677; Boyett v. Commissioner, supra; see Mills v. Commissioner of Internal Revenue, 4 Cir. 1968, 399 F.2d 744, 749; Doll v. Glenn, 6 Cir. 1956, 231 F.2d 186, 188; see also 10 J. Mertens, Law of Fed. Income Taxation § 55.19, pp. 123-124 (1965 rev. ed.). A presumption of correctness of the determination exists with the burden of proving otherwise on the taxpayer. Welch v. Helvering, 1933, 290 U.S. 111, 115, 54 S. Ct. 8, 78 L. Ed. 212; Webb v. Commissioner, supra, 394 F.2d at 372; Bryan v. Commissioner of Internal Revenue, 5 Cir. 1954, 209 F.2d 822, 825, cert. denied, 348 U.S. 912, 75 S. Ct. 289, 99 L. Ed. 715; see Easter v. Commissioner of Internal Revenue, 4 Cir. 1964, 338 F.2d 968, 969. Furthermore, if the assessment is right on any theory it must be sustained; it is immaterial that the theory used was wrong. Bernstein v. Commissioner of Internal Revenue, 5 Cir. 1959, 267 F.2d 879, 881; see Blansett v. United States, 8 Cir. 1960, 283 F.2d 474, 478. If the taxpayer believes that the Commissioner's method of computation was unfair or inaccurate, the burden is on him to show such unfairness and inaccuracy. Price v. United States, supra; see United States v. Lease, 2 Cir. 1965, 346 F.2d 696, 699-700 (preponderance of evidence standard required). Likewise, the burden is upon the taxpayer to establish a statutory right to specific deductions and to substantiate his eligibility therefor. See Bishop v. Commissioner of Internal Revenue, 6 Cir. 1965, 342 F.2d 757, 759; Cirillo v. Commissioner of Internal Revenue, 3 Cir. 1963, 314 F.2d 478, 483; Nowland v. Commissioner of Internal Revenue, 4 Cir. 1957, 244 F.2d 450, 453. Yet where the taxpayer has proven that he should be credited with a particular deduction, but he is unable for lack of adequate records to prove with certainty the amount, the Commissioner or the courts make an approximation of what seems reasonably allowable under the circumstances, applying what is generally called the Cohan principle.13 In light of the above guidelines governing tax reconstruction methods, we recognize that we should disturb a Tax Court determination that particular bank deposits represent taxable income only if, under the circumstances of the case, such a conclusion seems speculative and arbitrary rather than rational and fair. See 10 J. Mertens, Law of Fed. Income Taxation § 55.19, p. 124 and cases cited at n. 65 (1964 rev. ed. with Jan. 1969 Supp.). See also Helvering v. Taylor, 1935, 293 U.S. 507, 514, 55 S. Ct. 287, 79 L. Ed. 623. As Judge Goldberg so clearly stated for this Court:

"We recognize that the absence of adequate tax records does not give the Commissioner carte blanche for imposing Draconian absolutes. [Citations omitted.] But such absence does weaken any critique of the Commissioner's methodology. [Citations omitted]."

Webb v. Commissioner, supra, 394 F.2d at 373.

Although we are appalled by Cummings' obvious fraud and by his apparent unwillingness to cooperate with investigating IRS agents by refusing to turn over his 1944-1955 records, we are nevertheless constrained to agree with his contentions.

While we certainly realize that, without any business records for the 1944-1955 tax years, Agent Mims could make no adjustments to eliminate otherwise deductible nontaxable items included in the annual gross bank deposit figure for each tax year, we nevertheless conclude that his application of the 1956-1958 average net profits-to-net deposits ratio to the 1944-1955 annual gross deposits figures was unreasonable under the circumstances. He used as his starting point for determining the 1956-1958 ratio Cummings' gross bank deposits for those years. Yet he chose to use the 1956-1958 net bank deposits figure in his formula. The Tax Court incorporated the Mims ratio into its findings and decision and, therefore, clearly erred.

We conclude that, though weighted as it should be most heavily against the recalcitrant taxpayer and buttressed by the presumption of correctness, Agent Mims' formula adopted by the Tax Court was speculative and arbitrary because it lacked any fair and rational relation to the purpose it sought to achieve: determination of taxable income for 1944-1955, given only gross bank deposits data for those years and complete financial records for the 1956-1962 period. It would appear that a more reasonable formula, although perhaps not the only rational and fair one under the facts of this case, would have been as follows:

               Average Net Profits [1956-1958 or 1956-1962] = ratio.
               Average Gross Bank Deposits [Same period as above]

Such a net profits-to-gross deposits ratio, applied to tax years 1944-1955 for which only gross bank deposit data is available, would match reasonably related financial data for different periods where either adequate or inadequate taxable income information is available. The resulting deficiency determination would be rational and fair rather than speculative and arbitrary.

Likewise, we agree with the Tax Court that the Cohan rule should have been used to approximate a reasonable deduction for interline and C.O.D. collections deposited in Cummings' City National Bank account, only a portion of which represented income to him. We are unable, however, to understand why the Tax Court selected its ten per cent figure. It would appear from the record that Agent Mims actually adjusted the 1956-1962 annual gross bank deposits to eliminate nontaxable portions of interline and C.O.D. collections on the basis of available records for those tax years. Yet neither the findings nor the decision of the Tax Court reflect that any resort was made to such information in establishing the ten per cent allowance.15 We conclude that the Cohan rule, correctly applied, should reflect an allowance figure at least related to, if not relatively consonant with, some reasonable and fair ration of known C.O.D. and interline income-to-collections.16

Recomputation of the 1944-1955 taxable income figures using the recommended average net profits-to-average gross bank deposits ratio and redetermination of the Cohan rule allowance for nontaxable portions of C.O.D. and interline deposits — both predicated upon known and applied information for tax years 1956-1962 — will then establish beyond question that the particular bank deposits method employed actually represented, under the circumstances, a rational and fair determination, rather than a conclusion such as we presently review which seems speculative and arbitrary. See Webb v. Commissioner, supra, 394 F.2d at 373; Commissioner of Internal Revenue v. Welch, 5 Cir. 1965, 345 F.2d 939, 943-944. See also Helvering v. Taylor, supra. See generally 10 J. Mertens, Law of Fed. Income Taxation § 55.19, p. 124 (1964 rev. ed. with Jan. 1969 Supp.).

The decision of the Tax Court is, therefore, reversed and remanded for recomputation of Cummings' 1944-1955 annual tax deficiencies consistent with some "rational and fair" computation standard.

Reversed and remanded.