Opinion
No. 4705.
July 29, 1953.
Bernard J. Long, Washington, D.C. (Cann Long, Washington, D.C., on the brief), for petitioner.
Helen Goodner, Special Asst. to the Atty. Gen. (H. Brian Holland, Asst. Atty. Gen., and Ellis N. Slack and Harry Baum, Special Assts. to the Atty. Gen., on the brief), for respondent.
Before MAGRUDER, Chief Judge, and WOODBURY and HARTIGAN, Circuit Judges.
In this case the Boston Maine Railroad petitions for review of a decision of the Tax Court of the United States determining that for the calendar year 1941 there is a deficiency in petitioner's income tax in the sum of $937,984.56. The controversy relates to (1) the proper year, and (2) the proper amount, of the deductions which petitioner is entitled to take on account of railroad properties which were acquired by it some time prior to March 1, 1913, at a cost now unknown, such properties having been subsequently abandoned, destroyed, or otherwise put out of useful service in various years from 1932 on, and finally having been formally retired by petitioner in the years 1939, 1940 and 1941. The sole tax year directly involved is 1941; however, because of the application of the net operating loss carry-over provision of the Code, 26 U.S.C.A. § 122, we must also look to the years 1939 and 1940.
Most of the difficulties in this case stem from the fact that the present taxpayer, like many railroads across the United States but unlike most ordinary taxpayers, employs the retirement method of accounting for the costs of worn-out, obsolete, or destroyed equipment. The essential issue — both as to the year and amount of the deductions — is to what extent petitioner's authorized use of this different method of accounting should justify it in making certain deviations from standard tax rules which would be applicable if other methods of accounting were being employed. It will be helpful, therefore, to describe briefly at the outset this retirement method of accounting as distinguished from other accounting techniques designed to accomplish the same purpose, such as, for example, the straight-line depreciation method.
Under straight-line depreciation, the cost of any item of equipment is ratably spread over its estimated useful life, and through these annual depreciation charges it is sought to recoup the initial capital outlay by the time the particular piece of equipment must be withdrawn from service. Thus under this method the net book value of any particular asset becomes less and less as the accumulated depreciation on account of this asset becomes greater and greater. Of course the costs of any capital additions or betterments are added to the ledger value of the particular asset, and hence are in turn depreciated.
Under the retirement system of accounting, however, there are no annual depreciation charges as such, and hence no annual adjustments are made in the book values of the various assets. Only upon the ultimate retirement or replacement with betterments of any particular item of property or equipment is any change made in the investment account; and on that occasion, the full initial book value of the asset (i.e., usually the original cost) is taken out of the capital account, and this sum, diminished by the net salvage proceeds, is then charged to current expense (or in certain cases directly to profit or loss). Meanwhile, any repairs or minor replacements made during the life of the property are charged directly to current expense and are never reflected in the capital account.
In fuller detail, the method as described by the Tax Court works as follows:
(1) If a piece of equipment is merely repaired, or if, though the asset is being replaced, it is a minor item (in absolute value or in its relation to the larger unit of which it is a part), the cost is merely charged to current expense, and the book value of the asset remains unchanged.
(2) If, however, the asset is replaced with betterments, or if the replacement constitutes more than 50 per cent of the unit, then the asset is cleared from the property account through a compensating charge to current expense (or in the case of extraordinarily large items directly to profit and loss), and the cost of the replacement is then charged to the appropriate property account.
(3) Finally, if an asset is retired and not replaced at all, the ledger value is again cleared from the investment account and (less net salvage proceeds) charged directly to profit and loss. Here, however, there is of course no subsequent charge to the investment account since the property has now entirely disappeared.
It will be noted that the procedure is essentially the same for replacements with betterments [(2)] and retirements [(3)]. But since this case is principally concerned with the latter type of transaction, we shall hereafter refer only to it.
It is important to note immediately that the final charge to expense upon the retirement of a piece of equipment is not to be understood as an effort to make up for the failure to take any prior yearly "depreciation" on this particular item during its useful life. Rather the underlying theory of the retirement method is that the charges to expense on account of all the items retired or replaced in any particular year are taken as a rough equivalent of what would be a proper depreciation allowance for all the working assets of the company for that year. The assumption is that once the system is functioning normally and the retirements are staggered fairly regularly, the charges to expense on account of equipment wearing out or otherwise disappearing from service are spread out and stabilized, and hence will approximate the results under straight-line depreciation. At the same time it is thought that this method will eliminate the not inconsiderable bookkeeping problem of making annual depreciation adjustments on account of each and every asset owned by an extensive railroad. Thus, under both methods, the total charges to expense on account of any particular asset — to the extent that such a segregated concept makes any sense at all under the averaging approach of the retirement system — are the same; also, if the retirements occur fairly regularly, the total annual charge on account of equipment becoming unserviceable should eventually tend to become roughly equivalent under both methods. However, under the retirement system the total capital account (i.e., the book value of the assets) is always higher, since under that system no adjustments are made in this account until an item is actually retired.
With these preliminary notions in mind, we turn to the first issue, which can be briefly stated as follows: In what year should petitioner be required to take its tax deduction for equipment which it retired in the years 1939, 1940 and 1941, but which was for the most part destroyed, abandoned or otherwise put out of active service in years prior to 1939? The government maintains the deduction should be taken in the year of loss; the taxpayer, on the other hand, says it should be taken in the year of actual retirement.
In order fully to understand the treatment for which petitioner contends, it is necessary to examine in some detail the procedure which it has evolved over the years for dealing with retirements. The initial proposals relating to the repair, replacement, abandonment or retirement of petitioner's roadway equipment usually originate in its engineering department. Where an expenditure of more than $1500 is contemplated, the executive committee of petitioner's board of directors must first pass on the proposal; in the remaining cases, petitioner's president is authorized to approve the project. In addition, where any roadway properties constituting a line of railroad are sought to be temporarily abandoned or permanently retired, the permission of the Interstate Commerce Commission must first be obtained. In some cases state regulatory agencies must also be consulted. When the requisite authorization has been obtained both from the governing regulatory agency and from petitioner's management, the contemplated work is commenced. Petitioner's accounting department then assigns an AFE (Authority for Expenditure) number to the particular project and all charges pertaining thereto are collected in this manner. When the work is finally completed, the appropriate bookkeeping entries are made. Thus, for example, where a retirement has been authorized, upon the completion of the demolition the accounting department will charge current expense with the listed book value of the particular asset, decreased by the value of any material salvaged, and increased by the costs of demolition. (These latter figures are reported to the accounting department by the division engineer in charge and the valuation engineer.)
It was this final bookkeeping entry, usually occurring in the year of actual demolition (except where in the normal course of bookkeeping procedures the entry was not made until the beginning of the following year), which petitioner considered conclusive as to the year in which the various retirements occurred. Of course in the very nature of the indicated procedure the year of physical demolition and retirement on petitioner's books of a particular asset might often post-date by several years the temporary abandonment or perhaps the partial destruction by casualty of that asset. By way of example, we refer to a typical item in the footnote. The other items are discussed in detail in the lengthy opinion of the Tax Court, 16 T.C. 1517.
AFE 18913 — The Lawrence Engine House.
This property was partly damaged by fire on May 2, 1937. The division engineer recommended removal and retirement on February 8, 1940. This was approved by petitioner's president on March 1, 1940. The work of demolition was begun in June and completed in October of 1940. Petitioner claimed that this property was retired in 1940; the Tax Court held that it was retired prior to 1939 apparently on the ground of the fire in 1937.
In his original deficiency notice, the Commissioner challenged only one particular item — the Franklin Tilton Railroad retirement — as having been deducted in the wrong year. However, by affirmative allegation in his answer in the Tax Court, the Commissioner challenged a number of other items: 10 for 1939, 16 for 1940, and 13 for 1941. As to all of these, the Commissioner maintained that the actual loss, destruction or abandonment in fact occurred some time prior to 1939 and that therefore the deductions were improper. The Tax Court sustained the Commissioner on most of these items, though holding as to some of those affirmatively alleged by the Commissioner in his answer that the Commissioner had not met the burden of proof imposed on him. While the Tax Court recognized that petitioner had employed the retirement method of accounting, the Court felt that this made no difference. It stated that under § 23(f) a loss is deductible only in the year sustained, and held that "[That] year is * * * the year of retirement, * * * the year of destruction or abandonment". 16 T.C. at 1554. Employing the doctrine of the cases arising under § 23(f) to the effect that the year of loss is a question of fact to be determined from all the circumstances, the Tax Court then went through a painstaking analysis and, as previously indicated, found that since most of the items had been partially destroyed or abandoned from active service prior to 1939 they were deductible only in those earlier years; the fact that frequently neither the company nor the ICC had yet authorized their retirement, or that often the actual demolition was not commenced until after 1939, was apparently deemed irrelevant. These conclusions as to the applicable law taxpayer seeks to have us review.
"§ 23. Deductions from gross income.
"In computing net income there shall be allowed as deductions: * * *
"(f) Losses by corporations. In the case of a corporation, losses sustained during the taxable year and not compensated for by insurance or otherwise."
Petitioner begins by conceding that it is clearly settled that under § 23(f) and the applicable cases a loss is deductible only in the year when it actually occurs, which question is a question of fact for the trial court to determine in the first instance. Boehm v. Commissioner, 1945, 326 U.S. 287, 66 S.Ct. 120, 90 L.Ed. 78; Mine Hill Schuylkill Haven R.R. Co. v. Smith, 3 Cir., 1950, 184 F.2d 422, certiorari denied, 1951, 340 U.S. 932, 71 S.Ct. 496, 95 L.Ed. 673. However, petitioner strenuously contends that as a result of its employment of the retirement accounting method this case is not governed by that subsection of the law. Rather it contends that the applicable section is 23( l) (the general section dealing with depreciation), which provides that "[there shall be allowed as a deduction] A reasonable allowance for the exhaustion, wear and tear (including a reasonable allowance for obsolescence) — (1) of property used in the trade or business * * *"; and that under this subsection no particular year of deduction is indicated, the only requirements being the general reasonableness of the deduction and appropriate consistency in the accounting method followed, neither of which the government has here challenged.
So far as we have been able to discover, this precise question has never before been considered by any court other than the Tax Court below, which, it seems to us, did not appear to appreciate the full import of petitioner's argument. For the reasons hereafter indicated, we have come to the definite conclusion that petitioner's point is well-taken.
It will be recalled that the cardinal principle of the retirement method is that no adjustment is made in the book value of an asset until it is actually retired or replaced with betterments. Only in that year is the entire book value charged off to expense. If a piece of property, therefore, is damaged by fire, let us say, this casualty is not then recognized; for the accounting treatment to be accorded to this event cannot be determined until it is decided whether the property will be repaired or replaced in kind (in which case the cost of repair or replacement will be charged directly to expense and the ledger value of the original asset remains unchanged), or whether it will be replaced with betterments or retired (in which case the ledger value of the whole asset is charged off to expense, and the cost of the replacement, if any, is charged to the investment account). The same uncertainty applies to abandonments.
It is clear, therefore, that the retirement method simply does not allow for the recognition of casualties or other temporary interruption in useful service. The theory, it must always be remembered, is that the aggregate of the retirements, replacements and repairs occurring in a particular year is taken as a rough equivalent of the wear and tear, the obsolescence, and the casualty losses occurring in all the properties of the company in that same year. Indeed, as has already been pointed out, if the method is to be properly applied, it is quite impossible to take any appropriate recognition of a fire loss until it is decided whether the property in question is to be repaired, replaced with or without betterments, or retired. Cf. Commissioner v. Harwick, 5 Cir., 1950, 184 F.2d 835, holding that even under § 23(f) an exception to the usual rule is made where it is uncertain whether and to what extent there will be insurance coverage for the loss; the loss is held not to be sustained until the question of insurance reimbursement is finally determined. See also Charles F. Jeffrey, 12 T.C.M. 534 (May 20, 1953).
Under these circumstances, we think that the government's attempt to apply § 23(f), dealing largely with casualty losses, to a retirement system taxpayer results in an undesirable hybridization of the retirement method; for not only are casualty losses and other losses of useful value not recognized as such under the retirement system, but also the rule has always been, as to cases falling under § 23(f), that the deduction shall be equal to the actual loss sustained, whereas under retirement accounting the deduction should be equal to the entire original basis. The flexible provisions of § 23( l), on the other hand, seem to be admirably adapted to the retirement system which, after all, is but another method of accounting for "exhaustion, wear and tear". In this connection, we think it not inappropriate to point out that in a sizable organization like petitioner's, it is especially important and desirable to work out some reasonably certain procedure for dealing in an orderly manner with the variety of situations which may arise, thereby avoiding the chaos which might well result from treating each retirement on an ad hoc basis. We think that petitioner has developed such a procedure, and now should be allowed to pursue it consistently.
In sum, since the essence of retirement accounting is that the ledger value of an asset is charged off only when that item is retired, it seems clear that the deduction should be taken in the year in which the property is in fact retired. This does not mean the year in which the property is partly destroyed; nor does it mean the year in which, let us say, it is decided to suspend operations over a particular line temporarily. Instead it means the year in which, as a result of the consent of the regulatory agencies involved and the authorization in the ordinary course of business of the management authorities in charge, the property is actually authorized to be retired from service and is accordingly dismantled. Generally this will also be the year in which petitioner makes the charge-off on its books, and for reasons of certainty, and in the absence of indications of bad faith or abuse, it would seem as though this year should be considered determinative. Of course it is quite clear, as the government hastens to point out, that normally a taxpayer's bookkeeping operations cannot determine its tax consequences. Yet the record of the instant case indicates that taxpayer has been diligent and businesslike in making the accounting adjustments after the demolition work has been completed, and indeed the government does not seem to charge the taxpayer with any improprieties in its accounting methods. To be sure, because of the normal pressure of work in a big organization there were a number of items with respect to which the demolition work was completed at the end of one year and the accounting entries were not made until the beginning of the next year. But these occasional discrepancies, we think, are well within the spirit if not the letter of the Commissioner's own regulations.
Treas. Reg. 111, § 29.43-2: "The expenses, liabilities, or deficit of one year cannot be used to reduce the income of a subsequent year. A taxpayer has the right to deduct all authorized allowances, and it follows that if he does not within any year deduct certain of his expenses, losses, interest, taxes, or other charges, he cannot deduct them from the income of the next or any succeeding year. It is recognized, however, that particularly in a going business of any magnitude there are certain overlapping items both of income and deduction, and so long as these overlapping items do not materially distort the income they may be included in the year in which the taxpayer, pursuant to a consistent policy, takes them into his accounts." (Italics added.)
The government also urges that no significance can be attached to the requirement that the ICC must first give its permission before certain properties can be retired, for it is well-settled that the requirements of a regulatory agency are not determinative for tax purposes where these requirements conflict with some particular provision in the Internal Revenue Code. Old Colony R.R. Co. v. Commissioner, 1932, 284 U.S. 552, 562, 52 S.Ct. 211, 76 L.Ed. 484; Mine Hill Schuylkill Haven R.R. Co. v. Smith, supra; Kansas City Southern Ry. Co. v. Commissioner, 8 Cir., 1931, 52 F.2d 372, 377. But under the foregoing interpretation there is no conflict between the requirements of the ICC and the provisions of the Internal Revenue Code. On the contrary, petitioner's adherence to the requirements of the governing regulatory agencies lends an element of consistency and soundness to its accounting practices.
In general, although the government does not really concern itself with the application of § 23( l) and appears to be content to rely on the established law under § 23(f), the main thrust of its argument seems to be that it would be inconceivable in effect to let the taxpayer itself determine when and how it would take these deductions. But this objection, to the extent that it has any merit at all, is one that is inherent in the retirement method; and it is now well-established both by the cases, see, e.g., Chicago North Western Ry. Co. v. Commissioner, 7 Cir., 1940, 114 F.2d 882, certiorari denied, 1941, 312 U.S. 692, 61 S.Ct. 711, 85 L.Ed 1128, and by the administrative blessing under the Bureau of Internal Revenue's Bulletin "F", that this is an acceptable method of accounting for tax purposes. Moreover, it should be repeated that the government has not claimed that petitioner is incorrectly applying the retirement method; nor has it charged petitioner with an abuse of the permissible business discretion which is an inherent element of that system. Under these circumstances there appears to be no valid reason why once a taxpayer has been authorized to use the retirement system he should not be allowed to apply it to its logical fullness even though this may conflict at points with standard practices which would be applicable if some other method of accounting had been used.
This brings us to the second issue in the case — the amount of the deductions which petitioner should get on account of those properties, retired in 1939, 1940 or 1941, which it had acquired at an unknown cost some time prior to March 1, 1913.
Prior to 1913 petitioner did not keep very precise records of its assets. It had simply one lump sum investment account representing railroad properties newly purchased or constructed. The account was by no means up to date, because on the one hand some of the items whose cost was included in it had since been retired without any adjustment in the account having been made; and on the other hand a number of minor items had been purchased whose cost was not reflected in the account as they had been charged directly to expense.
In 1914 the ICC first began to require petitioner to conform to a detailed system of accounts, and as to all equipment subsequently acquired petitioner followed this new system. However, the past lump sum investment account was still kept on its books. At about the same time Congress directed the ICC to make a detailed physical inventory of the properties owned by all railroads, determining with respect thereto (a) the original cost, (b) the cost of reproduction new as of June 30, 1914, and (c) the cost of reproduction new less sustained depreciation. See Physical Valuation of Property Act, 37 Stat. 701, 49 U.S.C.A. § 19a. The Commission immediately set about this task, and after considerable litigation the reproduction cost new of petitioner's entire properties was finally established at $125,325,554, and the cost of reproduction new less sustained depreciation was determined to be $98,837,926. At this same time petitioner's lump sum asset account for properties acquired prior to 1914 stood at $124,778,152. However, in view of the considerations previously advanced — that the ICC figure represented an actual physical inventory of the railroad's properties, while the company's book figure was an obsolete one, — the close resemblance between these two figures could be said to be more or less coincidental. With respect to the third figure which the Commission was directed to determine — the actual cost of the equipment which the railroad had on hand as of June 30, 1914 — the Commission made no finding, explaining that this was impossible "owing to the insufficiency of the [petitioner's] accounting records."
When the petitioner sought to retire some of these properties acquired prior to March 1, 1913, it became important to determine their tax basis so that petitioner might avail itself of the deduction allowed by the Internal Revenue Code. Section 23( l), as we have seen, authorizes a reasonable allowance for exhaustion, wear and tear, and obsolescence; and § 114(a) states that "The basis upon which exhaustion, wear and tear, and obsolescence are to be allowed * * * shall be the adjusted basis provided in section 113(b)". See also Treasury Regulations 111, § 29.23( l)-4. Section 113(b), in turn, defines the adjusted basis as the basis under § 113(a) — i.e., cost, in this case — plus or minus various indicated adjustments. The only adjustment here in issue is that required by subsection 113(b) (1)(C) "for exhaustion, wear and tear, obsolescence, amortization, and depletion [prior to March 1, 1913], to the extent sustained;" the basic question being whether that subsection also applies to a taxpayer using the retirement system, in view of the fact that such a taxpayer does not make any adjustments for accrued depreciation but charges off the entire initial cost in the year of retirement.
Even if, as the government contends, the deduction came under § 23(f) dealing with corporate losses, the net result at this point would still be the same, since § 23(i) says that the amount of the loss under § 23(f) shall be the adjusted basis as provided in § 113(b).
When the taxpayer was preparing its 1939, 1940 and 1941 returns involving the retirement deductions on account of numerous items of property acquired prior to March 1, 1913, it was faced not only with this question of the applicability of § 113(b)(1)(C), but also with its inability to determine the original cost of the items, as required by § 113(a). Under the circumstances it did the best it could by taking its lump sum book cost figure and using the ICC inventory values to break this figure down into its component parts. The Commissioner objected to taxpayer's making any use whatsoever of its antiquated book cost figure and in his deficiency notice determined that the ICC reproduction cost new figures should be utilized exclusively. As this made very little difference, the taxpayer had no objections to this substitution. However, the Commissioner also insisted that the ICC reproduction cost new figures must be adjusted downward for sustained pre-1913 depreciation, apparently proceeding on the assumption that § 113(b)(1)(C) required this; furthermore, he sought to disallow those parts of the deductions representing engineering costs, for example, the cost of laying out a particular railroad line. It was as to the propriety of requiring these two deductions that the taxpayer petitioned the Tax Court.
Since, as has already been pointed out, the actual cost of these items was unknown, the Commissioner in making this determination was proceeding under the rule of Cohan v. Commissioner, 2 Cir., 1930, 39 F.2d 540, 543, which lays down the general principle that where some deduction is clearly warranted but its amount cannot be properly ascertained, the Commissioner shall make the best estimate he can.
With respect to the engineering costs the taxpayer was sustained by the Tax Court, and though the government initially sought to have us review this determination its cross petition has since been dismissed upon motion. Hence we need not concern ourselves further with this issue. With respect to the deduction for pre-1913 depreciation, however, the Tax Court, though agreeing with the taxpayer that § 113(b)(1)(C) has no application to a person using the retirement method, sustained the Commissioner on another ground. It pointed out that while the proper basis of the deduction should be "cost" undiminished by any pre-1913 depreciation, the taxpayer had not shown that the figure ultimately arrived at by the Commissioner was less than "cost". In effect, thus, the Tax Court said that since the taxpayer had the burden of proving the Commissioner's determination erroneous, and since it could not establish the "cost" of the items involved, the Commissioner's determination, no matter how arrived at, would be allowed to stand.
As concerns the holding that § 113(b)(1)(C) has no applicability here, we are in entire agreement with the Tax Court, and shall not add much to what it has already said. In coming to this conclusion, the Tax Court relied on a recent decision in the Second Circuit, Commissioner v. Union Pacific R.R. Co., 2 Cir., 1951, 188 F.2d 950, where Judge Learned Hand, after carefully analyzing this question, concluded that to read the statute literally and apply it to a retirement system taxpayer would do severe violence to the principles of retirement accounting, and would result not in straight-line depreciation or retirement accounting, but some "unjust and illogical hybrid." 188 F.2d at page 951. This is so because under retirement accounting no deductions are ever taken for sustained depreciation until an asset is actually retired, and in that year the entire unadjusted book value is charged off. If petitioner were now required to take a deduction for sustained pre-1913 depreciation as to assets retired in, say, 1939, its allowance for the wear and tear sustained in all its properties in that year would be inadequate. The legislative history, moreover, bears out the fact that the seemingly mandatory language of § 113(b)(1)(C) was intended solely to distinguish between pre- and post-1913 depreciation, and there is no indication that Congress contemplated the problem which would arise if that section were applied to a retirement basis taxpayer. H.Rep. No. 1, 69th Cong., 1st Sess. 5 (1925), reprinted in 1939-1 Cum.Bull. 318 (Part 2). Hence, as Judge Hand has recently said in another connection, what we must do "is to project ourselves, as best we can, into the position of those who uttered the words, and to impute to them how they would have dealt with the concrete occasion." United States v. Klinger, 2 Cir., 1952, 199 F.2d 645, 648, affirmed by an equally divided court, 345 U.S. 921, 73 S.Ct. 779. We think that for the reasons already expressed, it is desirable in so far as possible to preserve the integrity of the retirement method and that this requires us to hold with the Tax Court that § 113(b)(1)(C) has no applicability here.
But as already indicated the Tax Court, while agreeing that the Commissioner's subtraction of pre-1913 depreciation under § 113(b)(1)(C) was improper, nevertheless sustained the Commissioner because petitioner, who had the burden of proof, had failed to show that the unadjusted basis ( i.e., the cost) of the assets here in question exceeded the figure ultimately arrived at by the Commissioner as to the amount of the deduction to which the petitioner was entitled. Since it is agreed by all parties that the actual cost of the properties here in dispute is in fact unknown and unknowable, this is tantamount to saying that the Commissioner's determination — no matter how erroneously arrived at — must stand.
The taxpayer seems to have several answers to this. First, it says that it does not have the burden of proving "cost" since the Commissioner used the ICC valuations in lieu of cost, which procedure the taxpayer did not challenge; that it contested only the adjustments to "cost" for pre-1913 depreciation and engineering expenses. The Commissioner in turn replies that he never determined the statutory "cost"; indeed that the ICC found that this figure could not be determined and that he merely used the ICC reproduction cost new figures as a "starting point" in his calculation. Moreover he says that even if he had wanted to use the ICC figures in place of the statutory requirement of "cost", he would have been forbidden from doing so by the clear command of the statute.
The actual indications in the record on this point are inconclusive. There is much to substantiate the taxpayer's argument that the Commissioner in fact did not seek to use the ICC figures as "cost". Indeed it is difficult to see how the Commissioner could have made the 113(b)(1)(C) adjustment if he had not already determined the statutory cost. But there is also some indication to the contrary.
Be that as it may, we do not think the taxpayer needs to rest its position on such a technical ground. There certainly is no doubt that the Commissioner made the adjustment for pre-1913 depreciation pursuant to what he conceived to be the statutory mandate in § 113(b)(1)(C). If this statutory provision is now held to be inapplicable as a matter of law, there seems to be no sound reason why the Commissioner's erroneous subtraction of pre-1913 depreciation should nevertheless be sustained solely on the ground that the taxpayer for lack of the necessary cost records cannot show the inaccuracy of the end result. Surely the requirement imposing the initial burden of proof on the taxpayer cannot go this far. The mere fact that the taxpayer is unable to prove the exact figures of original cost should not entitle the Commissioner, under the shelter of the rule as to the burden of proof, to use any and all means, no matter how obviously illogical and incorrect, to arrive at a substitute figure. Such a rule would in a situation like this leave the taxpayer wholly at the whim and caprice of the government.
The government also suggests that the deduction was made in order to arrive at a better estimate of cost. But whether or not the ICC figures represented the best estimate of cost — and there is some persuasive evidence that they did — there certainly was no rational basis for reducing these reproduction cost new figures by depreciation sustained prior to 1913 in order to arrive at a better estimate of actual cost.
It might be argued that this is all right, since it is the taxpayer's own doing which has got it into this dilemma. That argument might have some force in a situation where the taxpayer's lack of adequate records is the result of carelessness or failure to take proper precautions. But there is no evidence here that it was customary prior to 1913 for railroads to keep detailed cost records; indeed the very fact that Congress directed the ICC to make the indicated valuations is some evidence to the contrary.
To be sure, it might be that if the Commissioner had simply stated a final figure for the amount of the allowable deduction without seeking to justify it, the taxpayer in view of its lack of records would have been helpless to disprove it; also perhaps it is true that the government could have disallowed any deduction at all. See Burnet v. Houston, 1931, 283 U.S. 223, 228, 51 S.Ct. 413, 75 L.Ed. 991. But see Cohan v. Commissioner, note 6 supra, which states a now well-accepted principle of tax administration. In any event, here the Commissioner has commendably sought rationally to explain how he arrived at the final amount of the allowable deduction. We think that it is not asking too much, then, to hold the government to its professed standard of rationality by striking down any arbitrary and erroneous steps in the calculation. As applied to the present situation, we are unpersuaded by the government's reliance on such cases as Helvering v. Gowran, 1937, 302 U.S. 238, 245, 58 S.Ct. 154, 82 L.Ed. 224, holding that a tax determination by the Commissioner which is correct in result, though arrived at by wrong reasoning, should be sustained.
Accordingly we hold that the amount of the various deductions should be based on the ICC reproduction cost new figures which the government itself has proposed, but that these figures should not be diminished on account of sustained pre-1913 depreciation.
The decision of the Tax Court is vacated, and the case is remanded to that Court for further proceedings in accordance with this opinion.