CORRECTION TO INVENTORY METHOD AMOUNTS TO CHANGE IN ACCOUNTING METHOD.

Mark D. Eblen, for respondent.

                                                Taxable (Calendar) Year

Some facts have been stipulated and are so found. The stipulation of facts, with accompanying exhibits, is incorporated herein by this reference. We need find few facts in addition to those stipulated and shall not, therefore, separately set forth our findings of fact. We shall make additional findings of fact as we proceed.

Background

All petitioners except for James A. and Dorothy A. Patterson resided in Kentucky at the time they filed their respective petitions. The Pattersons resided in Florida at the time they filed their petition.

The Huffman Group

The Huffman group of corporations (Huffman group) consists of four members (sometimes, the members): Neil Huffman Nissan, Inc. (Nissan); Neil Huffman Volkswagen, Inc. (Volkswagen); Neil Huffman Enterprises, Inc., d.b.a. Neil Huffman Dodge (Dodge); and Neil Huffman, Inc., d.b.a. Huffman Chrysler Plymouth (Chrysler). The members sell new and used automobiles in Kentucky. At least one of each married pair of petitioners owns stock in one or more of the members. Each of the members has elected to be treated as an S corporation under the provisions of section 1361.

Use of Inventories

The members of the Huffman group all sell merchandise (new and used automobiles). Each, therefore, computes its gross income from sales during a year by subtracting from sales revenue the cost of the goods sold. See sec. 1.61-3(a), Income Tax Regs. Because each is a merchant, each must also use inventories and an accrual method of accounting to determine the cost of the goods sold and to match that cost against sales revenue. See secs. 1.471-1 (merchants must use inventories) and 1.446-1(c)(2)(i) (generally, where inventories necessary, accrual method must be used with regard to purchases and sales), Income Tax Regs. As explained by Stephen F. Gertzman (Gertzman) in his treatise, Federal Tax Accounting, par. 6.02[2], at 6-5 & 6-6, (2d ed. 1993) (cited hereafter as Gertzman par. __, at __), in the case of a merchant that sells a large number of essentially similar or fungible items, the cost of the goods sold during any period is computed in steps, using inventories and an accrual method of accounting, along with various assumptions as to the manner in which the actual costs incurred in acquiring or producing items of inventory are allocated among the items so acquired or produced. To compute the cost of goods sold during a year, the steps are as follows: First, the costs of the items acquired or produced during the year are aggregated. That total is then combined with the aggregate cost of the items on hand at the beginning of the year to produce the total cost of the goods available for sale during the year. That last total is then allocated among items on hand at the end of the year (cost of ending inventory) and items sold during the year (cost of goods sold). The formula for determining cost of goods sold is essentially as follows:

The LIFO Method

For a taxpayer whose ending inventory computed under LIFO reflects the lower prices of antecedent purchases (rather than the higher price of current purchases), the income tax advantage of LIFO is obvious: a reduction in current income, leading, generally, to a reduction in current income tax. The potential for increased gain on account of the allocation of the lower costs of antecedent purchases to ending inventory is not eliminated, however; it is simply deferred until, in time, there is a liquidation of the items to which those lower costs have been allocated. See id. at 7-5. The term "LIFO reserve" refers to the amount by which the FIFO value (e.g., the current replacement cost) of inventory exceeds the LIFO value shown in the accounting records of the taxpayer. See id. par. 7.03[2], at 7-15.⁶ It is a measure of the potential gain in a store of inventoried items on account of the use of the LIFO method.

There is more than one method for computing the value of a LIFO inventory. Id. par. 7.04[1], at 7-30. Nevertheless, all LIFO computational methods involve essentially three determinations: (1) The LIFO inventory must be segmented into groups or "pools" of similar items; (2) a determination must be made as to whether there has been a quantitative change in the inventory of each pool during the period in question, and (3) there must be a determination of the manner in which increments to (i.e., increases in the quantity of) each pool are to be valued. Id. We are here concerned mainly with the third of those determinations.

Two basic LIFO computational methods are permitted by the income tax regulations: the specific goods method, a measure of inventory in terms of physical units of individual items, see sec. 1.472-2, Income Tax Regs., and the dollar-value method, a measure of inventory in terms of dollars, see sec. 1.472-8, Income Tax Regs. Each method is designed to make the three determinations previously identified. Gertzman par. 7.04[1], at 7-30. We are here concerned with the dollar-value method.

Under the dollar-value method, once items have been grouped into pools, the next step is to determine whether there has been any change in the quantity of dollars invested in the pools over the year. See Gertzman par. 7.04[3] [b], at 7-44. Those changes are determined by comparing the aggregate base-year cost of the items in a pool at the beginning of the year to the aggregate base-year cost of the items in the pool at the end of the year. See id. par. 7.04[3] [b], at 7-44 to 7-45. If the latter exceeds the former, there has been an increment in the pool; if the former exceeds the latter, there has been a liquidation of all or part of the pool. Id. par. 7.04[3] [b], at 7-45. The base-year cost of an item in a pool is the cost of the item (or what would have been the item's cost if it had been added to the pool) as of the base date. See id. "Base date" is the first day of the first year for which LIFO is adopted. Id. A similar description of the procedure for measuring the change in the size of a pool is found in section 1.472-8(a), Income Tax Regs.

Under any application of the dollar-value method, it is necessary to have a means for computing the base-year costs of the items in a pool and for computing the value of any increment in, or liquidation of, the pool. Gertzman par. 7.04[3] [b], at 7-45. As stated by the regulations, with respect to an increment: "In determining the inventory value for a pool, the increment, if any, is adjusted for changing unit costs or values by reference to a percentage, relative to base-year cost, determined for the pool as a whole." Sec. 1.472-8(a), Income Tax Regs. Three methods for making those computations are authorized by section 1.472-8(e)(1), Income Tax Regs.: the double-extension method, an index method, and a link- chain method. The following Example (1), based on an example in the regulations illustrating the double-extension method,⁸ shows how all three methods work. Example (1) demonstrates the computation of T's ending inventory for year 1.

Example (1): T elects, beginning with calendar year 1, to compute its inventory by use of the dollar-value LIFO method. T creates Pool No. 1 for items A, B, and C. The composition of the inventory for Pool No. 1 at the base date, January 1 of year 1, is as follows:

 Items          Units     Unit Cost     Total Cost

                                  Quantity           Unit

                       Dec. 31, yr. 1,

                                       Ratio(as a

      Dec. 31, yr. 1, inventory at current-year cost     $24,250

The following example, Example (2), continues the facts of Example (1). It is based on the assumption that, as of the beginning of year 1, in addition to electing to compute its inventory by use of the dollar-value LIFO method, T elected to use the link-chain method to compute the base-year and current-year cost of its inventory pools. Example (2) illustrates the computation of T's ending inventory for Pool No. 1 for year 2. An increment in year 2 closing inventory is determined to exist at base-year costs, and a LIFO value is assigned to that increment, using yearly increments in cost, as shown.

Example (2): During year 2, T completely disposes of Item A and purchases Item D, which is properly includible in Pool No. 1. T constructs a prior year unit cost for Item D.

                        Dec. 31, yr. 2,          Dec. 31, yr. 2,

                                        Ratio (as a

      Dec. 31, yr. 2, inventory at current-year cost          $33,500

Example (3):

                           Dec. 31, yr. 3,        Dec. 31, yr. 3,

                                           Ratio of

       Dec. 31, yr. 3, inventory at current-year cost       $28,900

Huffman Group Elections

The parties have stipulated that, prior to the tax years at issue, each member of the Huffman group filed an election to use the link-chain, dollar-value LIFO inventory method (the link-chain method).¹⁰ The parties have further stipulated that those elections were effective for the members as of the close of their taxable years ending as follows: Nissan, June 30, 1979; Volkswagen, Dec. 31, 1979; Dodge and Chrysler, Dec. 31, 1989.

The Accountant's Method

The Huffman group employed an accountant (the accountant) to compute the values of the respective inventories of each member using the link-chain method. The accountant was consistent in his method (the accountant's method) of making those computations each year, for each member, beginning with the year of each member for which it elected the link-chain method (the election year) and continuing thereafter, without exception, until the actions of respondent that led to this litigation (together, and without distinguishing among members, the election and following years). The parties have stipulated that, for each of the election and following years, the accountant omitted a computational step required by section 1.472-8, Income Tax Regs., which addresses the dollar-value method of pricing LIFO inventories. Pursuant to his method, the accountant first determined the items in each dollar-value pool at the end of each year. He then determined the current-year cost of each pool and divided that current-year cost by a cumulative index to determine the base-year cost of the pool. He compared the base-year cost so determined to the base-year cost of the pool as of the beginning of the year. When the end-of-the-year base-year cost exceeded the beginning-of-the-year base-year cost, the accountant determined that there had been an increment to the pool, but he did not multiply the increment by the cumulative index (he failed to "index" the increment) to determine a LIFO value for the increment (sometimes, the accountant's error). He assumed the LIFO value of the increment to be the difference between the end-of-the-year and beginning-of- the-year base-cost of the pool. That assumption led him to conclude that the year-end LIFO value of each pool was its value determined at base-year costs.

Under the accountant's method, for years in which he determined that there had been an increment to an inventory pool, his failure to index the increment resulted in his understating the yearend LIFO value of the pool (assuming that the cumulative index, expressed as a percent, was greater than 100%), which, in turn, resulted in (1) an unwarranted increase in his computation of the cost of the goods sold from the pool, (2) an understatement of the gross income attributable to those sales, and (3) an overstatement of the LIFO reserve attributable to the pool.¹¹ For years in which he determined that an inventory pool had been liquidated in whole or in part, his past failures to have indexed any increments remaining in the pool at the beginning of the year resulted in his computing too low a cost of goods sold from the pool, which, in turn, resulted in an overstatement of the gross income attributable to those sales. The accountant's error did not result in the permanent omission of any amount of gross income by any member.

The distortion resulting from the accountant's error can be seen in the following example: T, a merchant, elects to compute her LIFO inventory using a dollar-value method and begins her first year under the dollar-value method (year 1) with 100 units of an inventoriable item with a base-year cost of $1.00 a unit. Later that year, after the wholesale price of the item has increased to $2.00 a unit, T purchases 100 units more. Unfortunately, T makes no sales during that year. Applying the accountant's method, nevertheless, T computes a cost of goods sold of $100. She reaches that result by determining the value of her ending inventory (200 units, comprising an opening inventory of 100 units plus an increment of 100 units), at base-year unit cost ($1.00) to be $200 (200 x $1.00). Since the base-year cost of her opening inventory of 100 units is $100, and she purchased 100 units during the year for $200, her cost of goods available for sale is $300, which, after subtracting the value determined for her yearend inventory ($200), results in a cost of goods sold (and a loss) of $100. Assume further that, in the next year (year 2), T decides to liquidate her inventory (200 units) and retire. She sells her inventory in bulk for $300. Her cost of goods sold is her year 2 opening inventory of $200, which results in T realizing a year 2 gain of $100. Of course, T realizes neither a loss in year 1 nor a net gain in year 2. T's failure to index the 100 unit increment included in her year 1 ending inventory distorts her income for both years 1 and 2.¹² The distortion is only matter of timing, however, since the understatement of income in year 1 is rectified by the overstatement of income in year 2. The following table illustrates the distortions:

                                LIFO inventory     LIFO inventory

Respondent's Examination and Adjustments

The following table illustrates respondent's adjustments with respect to Nissan for 1997 through 1999 (all dollar figures in thousands):

                                       1997       1998       1999

Petitioners agree that respondent's calculations of the beginning and ending inventories of each member of the Huffman group are correct.

       Member           1997          1998            1999

In addition, for the earliest year of each member open to adjustment by respondent (the first year in issue), respondent made an additional adjustment under section 481. That adjustment increased the taxable income of the member for that year to reflect the cumulative adjustments to income revealed by respondent's recalculations for all years of the member's up until that year. Those adjustments (the section 481 adjustments) are as follows:

         Member            1997           1998

Change in Method of Accounting

No member of the Huffman group requested respondent's permission to change its method of accounting.

The parties are in agreement that, in computing the LIFO values of the Huffman group's yearend inventories, the accountant employed by the group omitted a computational step required by section 1.472- 8, Income Tax Regs. (addressing the dollar-value method of pricing LIFO inventories). The consequence of the accountant's error was that, generally, he understated the LIFO value of those inventories (which, generally, resulted in an under-reporting of income from sales). Respondent corrected the accountant's error, and petitioners accept respondent's adjustments to the inventories of the members of the Huffman group for all of the years in issue. Petitioners do not accept, however, respondent's determination that, in making those adjustments for the first year in issue of each member, he was implementing a change that he had made in the members' methods of accounting, which necessitated his making additional adjustments for those years pursuant to section 481(a). Petitioners argue that respondent's adjustments were merely the result of his correction of a mathematical error made by the accountant. They point out that, pursuant to section 1.446-1(e)(2)(ii)(b), Income Tax Regs.,¹³ the correction of a mathematical error is explicitly excluded from constituting a change in method of accounting. Because, they argue, there was no change in any member's method of accounting, no section 481 adjustments were warranted. They concede, however, that if section 481 adjustments were warranted, respondent has correctly computed those adjustments. Our sole task is to determine whether the section 481 adjustments were warranted, which requires us to determine whether, in revaluing the members' inventories, respondent corrected a mathematical error or changed the members' methods of accounting for those inventories.

Before addressing that question, we shall discuss the relevant provisions of sections 446 and 481.

II. Sections 446 and 481

III. Discussion

To distinguish between error correction and an accounting method change, we must examine both the pertinent Treasury regulation and caselaw.

As we described supra in giving the background of this case, the accountant erred in applying the link-chain method, he did so consistently for each member, beginning in the year the member elected the link-chain method and ending only when respondent found the error, the error resulted in income being under-reported for some (most) years and over-reported for other years, and, if not corrected, the error would not result in the permanent omission of income by the taxpayers. The accountant's error was an error in allocating the cost of goods available for sale during a year between the items sold during the year and the items on hand at the end of the year. Generally, under a system of inventory accounting, the value assigned to the items on hand at the end of one year establishes the value of the items on hand at the beginning of the next year. Consequently, the accountant's error would, if applied consistently (as, in fact, it was), self correct, at least in the sense that, if the error were continued over the life of any inventory pool, the total gain reported on account of the sale of items in the pool would be correct. See, e.g., Wayne Bolt & Nut Co. v. Commissioner, 93 T.C. 500, 509 (1989) (similar conclusion with respect to the income reported through the period in which ending inventory is correctly valued). The accountant's error was, thus, an error in timing. Because it was an error in the proper time for reporting an item of income (gain from sales), the accountant's method was a material item in each member's overall plan of accounting. See sec. 1.446-1(e)(2)(ii)(a), Income Tax Regs. On that ground alone, respondent's change to that method would appear to be a change in a method of accounting, as that expression is used in section 1.446-1(e)(2)(ii)(a), Income Tax Regs. By consistently repeating the same error, the accountant established a pattern, which (although not determinative of) is indicative of a method of accounting. Id.

Nevertheless, section 1.446-1(e)(2)(ii)(b), Income Tax Regs., provides that a change in method of accounting does not include correction of mathematical or posting errors, and petitioners argue that, in correcting the accountant's error, respondent did no more than correct a mathematical or posting error. We have interpreted the term "posting error" to be an error in "'the act of transferring an original entry to a ledger'". Wayne Bolt & Nut Co., v. Commissioner, supra at 510-511 (quoting Black's Law Dictionary 1050 (5th ed. 1979)). That does not describe the accountant's error, and we conclude that the accountant made no posting error. The term "mathematical error" is not, as stated, defined in the regulation, nor have we or any other court defined it for purposes of section 1.446-1(e)(2)(ii)(b), Income Tax Regs. The term does, however, appear in the Internal Revenue Code, principally in section 6213(b), which allows the unrestricted assessment and collection of tax arising out of mathematical or clerical errors. For purposes of section 6213, the term "mathematical or clerical error" is defined by section 6213(g)(2). As pertinent to this case, the definition is "an error in addition, subtraction, multiplication, or division". Sec. 6213(g)(2)(A). Moreover, before Congress provided the specific definition of the term "mathematical or clerical error" found in section 6213(g), Courts generally had limited the scope of the term "mathematical error" for purposes of section 6213(b) and its predecessors to errors in arithmetic. E.g., Farley v. Scanlon, 13 AFTR 2d 932, 933, 64-1 USTC par. 9371 (E.D. N.Y. 1964) (mathematical error "means an error in computing the tax on what the return itself concedes to be income"); Repetti v. Jamison, 131 F. Supp. 626, 628 (N.D. Cal. 1955) ("the term * * * was meant to refer to errors in arithmetic. This opinion is based primarily on the common meaning given to the phrase 'mathematical error,'"). We have no reason to believe that the drafters of section 1.446-1(e)(2)(ii)(b), Income Tax Regs., intended the term "mathematical error" to have any meaning beyond its common meaning, and petitioners have failed to show us that the term has a common meaning different from the common meaning found by the District Court in Repetti; i.e., an error in arithmetic. That definition comports with the scope of the term "posting error", with which the term "mathematical error" is associated in the regulations, and we conclude that the term "mathematical error", as used in section 1.446-1(e)(2)(ii)(b), Income Tax Regs., describes an error in arithmetic; i.e., an error in addition, subtraction, multiplication, or division.

The accountant did not make a mathematical error because he did not make an error in arithmetic. He neither divided when he should have multiplied nor multiplied 2 x 2 and found the product to be 5. The accountant erred in that, after deflating the current-year cost of each inventory pool to determine whether, at base-year costs, there had been an increment in the pool, and finding an increment, he failed to multiply the increment by the cumulative index in order to determine the yearend LIFO value of the pool. The accountant reached an erroneous result not because he made a mistake in arithmetic (multiplication) but because he omitted the critical step of multiplication altogether. That kind of error no more lends itself to being classified as an arithmetical (mathematical) error than does the error of the baker who, having intended to double the recipe for a cake he has baked, finds that the cake has only risen half way because he failed to double the measure of baking powder called for by the recipe. Petitioners cannot avoid respondent's section 481 adjustment on the ground that respondent changed no method of accounting because he corrected only mathematical or posting errors.

Nor can petitioners avail themselves of the exceptions in section 1.446-1(e)(2)(ii)(b), Income Tax Regs., specifying that an accounting method change does not include the correction of errors in the computation of tax liability or adjustments not involving the proper time for inclusion of an item of income or the taking of a deduction.

Although section 1.446-1(e)(2)(ii), Income Tax Regs., appears dispositive in respondent's favor, our inquiry does not end there, because courts addressing the issue of whether a change in method of accounting has occurred have not uniformly given consistency and timing considerations the weight given those considerations by the regulations.

Petitioners can find some support for their premise in cases holding that a taxpayer does not change its method of accounting when it does no more than conform to a prior accounting election or some specific requirement of the law. Many of the cases that petitioners rely on, however, were decided before the 1970 revisions to section 1.446-1(e), Income Tax Regs., emphasizing consistency and timing considerations. Petitioners refer us to Thompson-King-Tate, Inc. v. United States, 296 F.2d 290 (6th Cir. 1961), in which the Court of Appeals held that the taxpayer had the right to change its original reporting position and report income from a construction contract in the year the contract was finally completed and accepted because the taxpayer had previously adopted that method for reporting income from construction contracts. Id. at 294. The Court of Appeals emphasized that the taxpayer had "no election" (i.e., choice) but to report the income in the correct year pursuant to the method of accounting it had adopted. Id. at 294, 295. Petitioners also cite N.C. Granite Corp. v. Commissioner, 43 T.C. 149 (1964), and Underhill v. Commissioner, 45 T.C. 489 (1966). In the first case, we said that a taxpayer need not obtain the Commissioner's consent to change its method of accounting "where the law specifically prescribes or proscribes a method of accounting or computation". N.C. Granite Corp. v. Commissioner, supra at 168. In the second case, we held that no timing question was presented (so, therefore, the consent of the Commissioner to change a method of accounting was not required) when, to conform to caselaw, the taxpayer changed its method of recovering its basis in speculative installment notes from a pro-rata recovery method to a method that allowed it to recover all of its basis before it reported any gain. Underhill v. Commissioner, supra at 496. Because those cases were decided before 1970 and do not address the consistency and timing considerations emphasized in section 1.446- 1(e)(2)(ii), Income Tax Regs., their weight is uncertain.

Because the accountant's error in the instant case had precisely the same effect as did the taxpayer's discounting practices in Primo Pants Co. -- viz, it served merely to alter the distribution of a lifetime income among taxable periods -- that case would seem to govern us here, requiring us to conclude that respondent's adjustments to the members' inventories constituted a change in the members' methods of accounting. Petitioners attempt to distinguish Primo Pants Co. and the cases of the Court that follow it, but their reading of those cases is flawed. For example, on brief, petitioners discount the relevance of our holding in Primo Pants Co. because, they suggest: "No contention was made that the undervalued inventory was the result of a mathematical error." On the contrary, our report in Primo Pants Co. states: "Petitioner characterizes the various adjustments to inventory as the mere correction of its application of its lower of cost or market method of valuing inventory." Primo Pants Co. v. Commissioner, 78 T.C. at 714.

Taxpayers on other occasions have brought Korn Indus., Inc. to our attention. See, e.g., Superior Coach of Fla., Inc. v. Commissioner, 80 T.C. at 912 (facts before us distinguishable from those in Korn Indus., Inc.); Wayne Bolt & Nut Co. v. Commissioner, supra at 511 (similar). In Superior Coach, we noted that some commentators had pointed out that the good-faith exception seemingly created by Korn Indus., Inc. appears to be without statutory authorization. Superior Coach, Inc. v. Commissioner, supra at 914 n.5. Indeed, assuming that consistently made accounting errors are generally inadvertent (i.e., made in good faith), an inadvertence-based exception to the general rule (that the consistent treatment of an item amounts to a method of accounting) would seem to swallow that general rule. We need not resolve that conundrum today, because, as in the past, the facts before us are distinguishable from those in Korn Indus., Inc. v. United States, supra.¹⁶ Unlike in Korn Indus., Inc., the accountant's error in failing properly to apply the link chain method neither was an interruption in a history of proper application of that method nor was it restricted to only a portion of the costs to be taken into account in valuing inventories. The facts of Korn Indus., Inc. are distinguishable.

Evans v. Commissioner, supra, is a Memorandum Opinion of this Court, and memorandum opinions are not binding. See, e.g., Dunaway v. Commissioner, 124 T.C. 80, 87 (2005). Moreover, the conclusion we expressed in Evans, that the taxpayer merely misapplied the cash method, appears to contradict an example in the regulations interpreting section 481. Example (2), in section 1.4461(e)(3)(iii), Income Tax Regs., involves a taxpayer who consistently reports its income and expenses on an accrual method of accounting except for real estate taxes, which it reports on the cash method of accounting. The example concludes that a change in the treatment of real estate taxes from the cash method of accounting to an accrual method of accounting is a change in method of accounting because the change is a change in the treatment of a material item in the taxpayer's overall accounting practice. Finally, it is doubtful that intent plays a significant role in determining whether a taxpayer has adopted a method of accounting. See Buyers Home Warranty Co. v. Commissioner, T.C. Memo. 1998-98; see also Johnson v. Commissioner, 108 T.C. 448, 494 (1997) ("If the change affects the amount of taxable income for 2 or more taxable years without altering the taxpayer's lifetime taxable income, then it is strictly a matter of timing and constitutes a change in method of accounting."), affd. in part and revd. in part 184 F.3d 786 (8th Cir. 1999).

We agree with respondent that the facts of Gimbel Bros., Inc. and Standard Oil Co. are distinguishable from those now before us. The parties have stipulated that, for each member, for the election and following years (i.e., for 10 or 20 years), the accountant omitted a computational step required by the regulations governing the dollar-value method of pricing LIFO inventories. We agree with respondent that the members may, individually, have elected the link-chain method, but no member adopted it until respondent made his corrections. That alone distinguishes the facts before us from those in Gimbel Bros., Inc. and Standard Oil, Co., where the errors were committed in the context of a broader compliance with the taxpayer's proper method of accounting. Moreover, although stipulated by the parties, it is questionable whether all four of the members actually elected to use the link- chain method to value their respective inventories.¹⁸Gimbel Bros., Inc. and Standard Oil Co. are distinguished.

The notion that a taxpayer does not change its method of accounting when it merely conforms to a prescribed (but ignored) method of accounting is contradicted by at least one example in section 1.446-1(e)(2)(ii)(c), Income Tax Regs. Sec. 1.446- 1(e)(2)(ii)(c), Example (1), Income Tax Regs., addresses a merchant (a jeweler) erroneously reporting income from sales on the cash method of accounting. As discussed supra under the heading "Use of Inventories", inventories and an accrual method of accounting are required when the sale of merchandise is an income-producing factor. The example holds that a change from the cash method to the accrual method is a change in the merchant's overall plan of accounting and thus is a change in method of accounting. Moreover, the notion is also inconsistent with the more recent view of the courts that a taxpayer needs the Commissioner's consent to change from an erroneous to a correct method of accounting. See, e.g., Wayne Bolt & Nut Co. v. Commissioner, 93 T.C. at 511 ("A change in method of accounting occurs even when there is a change from an incorrect to a correct method of accounting."), and other cases noted in Convergent Techs., Inc. v. Commissioner, T.C. Memo. 1995 320. There are also three examples in section 1.446-1(e)(2)(iii)(c), Income Tax Regs., holding that an impermissible method of accounting is a method of accounting a change from which requires the consent of the Commissioner: Examples (6), (7), and (8). We question whether there is vitality to the notion that a taxpayer conforming to a required but theretofore ignored method of accounting does not change its method of accounting by so conforming.

Consider a taxpayer that elects a method of accounting and, for some time, adheres to the method (thereby adopting it). The taxpayer then, for some time, deviates from the method before, again, adhering to it. The notion that the taxpayer did not change its method of accounting when it either, first, deviated from the method or, thereafter, adhered to the method is a notion that is narrower than the previously described notion, and it is one we have supported. See, e.g., Evans v. Commissioner, T.C. Memo. 1988 228. We have not, however, been consistent in holding that a taxpayer does not change its method of accounting when it does no more than adhere to a method adopted pursuant to a prior accounting election. See, e.g., Sunoco, Inc. & Subs. v. Commissioner, T.C. Memo. 2004-29 (retroactive attempt to change treatment of certain mining expenses would be change in method of accounting, and not mere correction of error, where taxpayer had knowingly and consistently, albeit improperly, capitalized and amortized expenses that should have been included in taxpayer's cost of goods sold); Handy Andy T.V. & Appliances, Inc. v. Commissioner, T.C. Memo. 1983-713 (specifically finding that an impermissible change in accrual methodology was a change in method of accounting and that reversion to original methodology was a second change in method of accounting, warranting a section 481 adjustment). Indeed, in First Natl. Bank of Gainesville v. Commissioner, 88 T.C. 1069 (1987), a transferee liability case, the transferee argued that the transferor's alteration of a LIFO inventory valuation procedure constituted the correction of an accounting error and not a change in method of accounting. We held that, although the alteration in question may have constituted the correction of an error, it also constituted a change in method of accounting pursuant to section 472(e). Id. at 1085. We added: "Where the correction of an error results in a change in accounting method, the requirements of section 446(e) are applicable." Id.

Our inconsistency in holding that a taxpayer does not change its method of accounting when it does no more than conform to a prior accounting election is not necessarily inconsistent with section 1.446-1(e)(2)(ii)(a), Income Tax Regs. That is because, generally, pursuant to section 1.446-1(e)(2)(ii)(a), Income Tax Regs., it is the consistent treatment of an item involving a question of timing that establishes such treatment as a method of accounting. Therefore, a short-lived deviation from an already established method of accounting need not be viewed as establishing a new method of accounting.¹⁹ If not so viewed, neither the deviation from, nor the subsequent adherence to, the method of accounting would be a change in method of accounting. The question, of course, is what is short-lived. The Commissioner's position is that consistency is established for purposes of section 1.446- 1(e)(2)(ii)(a), Income Tax Regs., by the same treatment of a material item in two or more consecutively filed returns. Rev. Proc. 2002-18, 2002-1 C.B. 678. We have said something similar. Johnson v. Commissioner, supra at 494. We need not today determine how long is short. Here, even if we were to assume that the members elected the link-chain method and adopted it, see supra pp. 46-48, no member deviated from the link-chain method for less than 10 years. That is not a short-lived deviation.

IV. Conclusion

For the first year in issue of each member, respondent's revaluation of the member's inventory constituted a change in the member's method of accounting. Therefore, respondent's section 481(a) adjustments are permissible. Each petitioner owning shares of stock in any member of the Huffman group must take into account his or her share of the section 481 adjustments. We need decide no other issue.

To reflect the foregoing,

Decisions will be entered for respondent.

² Hereafter, all section references are to the Internal Revenue Code of 1986, as amended and in effect for the years in issue.

³ FIFO is authorized by sec. 1.471-2(d), Income Tax Regs., and LIFO is authorized by sec. 472.

⁴ The following example is based on an example in Gertzman, Federal Tax Accounting, par. 7.02, at 7-4 (2d. ed. 1993) (cited hereafter as Gertzman par. __, at __):

Example: Assume that, in its first year of operation, a retailer acquires identical products at the following times and costs:

        Date         Number        Unit Cost         Total

Under FIFO, the ending inventory would be deemed to cost $12.68 (consisting of a layer of 10 units at $1.06 a unit and a layer of 2 units at $1.04 a unit). The balance of the cost of goods available for sale, $38.82, would be allocated to the 38 units sold and would be deemed the cost of goods sold.

Under LIFO, the ending inventory would be deemed to cost $12.04 (consisting of a layer of 10 units at $1.00 a unit and a layer of 2 units at $1.02 a unit). The balance of the cost of goods available for sale, $39.46, would be allocated to the 38 units sold and would be deemed the cost of goods sold.

⁵ In the example supra in note 4, the use of LIFO instead of FIFO increased the cost of goods sold by $0.64 (from $38.82 to $39.46). That $0.64 represents the inflation that had occurred during the year in the cost of the 12 items that remained on hand at the end of the year ((10 units x increase in price of $0.06 a unit) + (2 units x increase in price of $0.02 a unit)).

⁶ In the example supra note 4, assuming LIFO, the LIFO reserve at the end of the year would be $0.64, calculated as follows:

      FIFO value (current replacement cost)

Assume that T (a manufacturer) began operations a number of years ago with 4 pounds of item A that cost $0.10 a pound. Its total inventory was thus valued at $0.40. Normal operations require the taxpayer to purchase and consume 4 pounds of A each year. The LIFO value of its closing inventory would, thus, have remained $0.40 notwithstanding that the cost of A increased to $0.50 a pound in the interim. Assume further, that, because of technical advantages, an equal quantity of item B may now be used in lieu of item A. The current price of B is $0.40 a pound, and, because of the price advantage of B over A ($0.10), T, this year, purchases 4 pounds of B and consumes its remaining stock of A. Like A, B has a base-year cost of $0.10. Under those facts, if T follows the dollar-value method with a single inventory pool that includes both items A and B, its cost of goods sold and ending inventory will be as follows:

 Quantitative change in base-year cost of inventory:

⁸ Sec. 1.472-8(e)(2)(v), Example (1), Income Tax Regs.

⁹ The computational procedures for the link-chain method are described by the Commissioner in Rev. Proc. 97-36, sec. 2.04(1)(c) and (d), 1997-2 C.B. 450, 451.

¹⁰ The parties have attached documentation to the stipulation of facts evidencing those elections. The documentation is inconsistent with the described elections with respect to (1) Neil Huffman Enterprises, Inc., d.b.a. Neil Huffman Dodge, and (2) Neil Huffman, Inc., d.b.a. Huffman Chrysler Plymouth, in that it indicates that those corporations elected to adopt "an index method as provided in [sec. 1.472-8(e)(1), Income Tax Regs., * * * which] will be developed by double extending * * * a representative portion of inventory at beginning of year cost and current cost." Such an index method is distinct from the link-chain method purportedly adopted. We address the significance of that fact infra in sec. III.C.3.b.iii of this report.

¹¹ The yearend LIFO value of the pool was understated because, even under the LIFO method, inventory cannot be carried at a cost lower than the actual cost of purchasing the inventory. Cf. Fox Chevrolet, Inc. v. Commissioner, 76 T.C. 708, 732 n.15 (1981).

¹² For the 100 units purchased during year 1, the index would be 200%, reflecting the doubling during the year in the unit cost of the inventoriable item.

¹³ In citing sec. 1.446-1(e)(2)(ii)(a) and (b), Income Tax Regs., we refer to that section as in effect before its revision by T.D. 9105, 2001-4 C.B. 419, 423, which replaced much of the content of that section with the substantially similar content of sec. 1.446-1T(e)(2)(ii)(a) and (b), Temporary Income Tax Regs., 69 Fed. Reg. 42 (Jan. 2, 2004).

¹⁴ Sec. 481(a) provides:

¹⁶ Though adhering to the holding of its predecessor in Korn Indus., Inc. v. United States, 209 Ct. Cl. 559 (1976), see Diebold, Inc. v. United States, 16 Cl. Ct. 193, 204 n.9 (1989), affd. 891 F.2d 1579 (Fed. Cir. 1989), the U.S. Claims Court (now the U.S. Ct. of Fed. Claims) has emphasized the primacy of consistency and timing in establishing a method of accounting. See Diebold, Inc. v. United States, supra.

¹⁷Gimbel Bros., Inc. v. United States, 210 Ct. Cl. 17, 535 F.2d 14 (1976), like Korn Indus., Inc. v. United States, supra, was decided by the Court of Claims and is therefore not binding upon us. Further, the former case analyzes and applies regulations in effect before 1970. We nevertheless include the case in this discussion because it was decided following the issuance in 1970 of the regulations in effect for the instant case.

¹⁸ See supra note 10.

¹⁹ Nor in reaching that conclusion would a court have to find that the taxpayer committed a posting or mathematical error. See sec. 1.446-1(e)(2)(ii)(b), Income Tax Regs.

²⁰ It is worth mentioning that the use of price indexes in applying the dollar-value method is a matter to which Congress in sec. 472(f) and the Secretary of the Treasury in his regulations, see, e.g., sec. 1.472-8(e)(3), and revenue procedures have devoted attention. Among the latter are Rev. Proc. 97-36, 1997-2 C.B. 450, and Rev. Proc. 97-37, 1997-2 C.B. 455 The first of those revenue procedures describes the adoption of the "Alternative LIFO Method" (a dollar-value link-chain method for retailers of autos and light-duty trucks) as a change in method of accounting. The second likewise describes the inventory price index computation (IPIC) method.