This case was referred to Trial Commissioner James F. Davis with directions to make findings of fact and recommendation for conclusions of law under the order of reference and Buie 134(h). The commissioner has done so in an opinion and report filed on October 28, 1970. The plaintiff filed exceptions to the commissioner’s opinion with respect to the rail salvage issue and the vacation pay accrual issue. The defendant filed exceptions to the commissioner’s opinion regarding the issues of donated property depreciation, casualty loss, welded rail, protective facilities, and the Mexican tax credit. Both parties requested review of the commissioner’s opinion. The case has been submitted to the court on oral argument of counsel and the briefs of the parties. Since the court agrees with the commissioner’s opinion, findings of fact and recommended conclusion of law, with certain modifications, as hereinafter set forth, it hereby adopts the same, as modified, as the basis for its judgment in this case.* Therefore, it is concluded the plaintiff is entitled to recover, together with interest as provided by law, on the claims relating to (1) § 1341 computation, (2) excess salvage value, (3) donated property depreciation, (4) casualty loss, (5) welded rail, (6) protective work, and (7) Mexican tax credit, and judgment is entered to that effect. The court further concludes that plaintiff’s recovery shall be subject to the setoffs raised by defendant with respect to (1) rail salvage value, and (2) vacation pay accrual. *271The amount of recovery will be determined in subsequent proceedings under Bule 131 (c).
Commissioner Davis’ opinion, as modified by the court, is as follows:
This is a suit to recover income taxes paid for the year 1955. Plaintiff operates a railroad as a common carrier in interstate commerce, subject to the jurisdiction of the Interstate Commerce Commission. Plaintiff’s petition raised six issues with respect to a claim for refund timely filed with, and ultimately denied by, the District Director of Internal Eevenue at Chicago, Illinois. In its answer and first amended answer, defendant asserted four setoff defenses, pursuant to Missouri Pacific B.R. v. United States, 168 Ct. Cl. 86, 338 P. 2d 668 (1964). Before trial, defendant dropped one of the setoff defenses; and by pretrial stipulation, the parties resolved one issue raised in the petition.1 Trial was held on the remaining eight issues. After trial, defendant conceded that plaintiff is entitled to recover on one other issue raised in the petition.2 Bemaining for resolution, therefore, are seven issues, designated as follows: (a) donated property depreciation; (b) casualty loss; (c) welded rail; (d) rail salvage value; (e) protective work; (f) vacation pay accrual; and (g) Mexican tax credit.
DONATED PROPERTY DEPRECIATION ISSUE
Starting about 1930, plaintiff entered into many agreements with several midwestern states for construction of highway overpasses and underpasses at highway-railroad intersections, and grade-crossing protection equipment, such as flashing-light signals and automatic gates. Though the agreements are not identical, generally plaintiff agreed to perform a large share of the construction work and the states agreed to pay most of the cost. Sometimes, part of the work *272was done by state highway departments. Pursuant to Federal highway aid legislation (particularly acts passed in 1983 and 1944), the Federal Government agreed to pay the governmental share of the cost. Federal funds were allocated to the states to pay for specific construction projects. E.g., the Federal Highway Act of 1944,58 Stat. 839, ch. 626, provided that costs be apportioned between the Government and the railroads, the railroads’ share not to exceed 10 percent.
Most of the agreements did not state expressly whether the respective state governments or plaintiff was to have legal title to the facilities. However, the parties have stipulated that the facilities (jetties, bridges, highway undercrossings and overcrossings, floodlights, flasher signals, and signs) were “contributed” to plaintiff by the states; and this is taken to mean that plaintiff owns them, at least for purposes here material. See Lazarus, infra. In any event, under all the agreements, plaintiff was obligated to maintain and replace as necessary, at its own expense, facilities originally built. The facilities were constructed primarily for the benefit of the public to improve safety and to expedite motor-vehicle traffic flow. The record shows, however, that plaintiff received economic benefits from the facilities, e.g., probable lower accident rates, reduced expenses of operating crossing equipment and, where permitted, higher train speed limits. Plaintiff also received intangible benefits, e.g., goodwill from the community-at-large, which was to plaintiff’s long-term economic advantage.
On February 5; 1943, plaintiff requested permission of the Service to change from retirement to depreciation accounting for road property. On April 23, 1943, the Service responded by letter to plaintiff’s request and enclosed Mimeo 58, entitled “Change from Eetirement to Depreciation Accounting for Eoad Property.” Mimeo 58 set out guidelines under which the changeover in accounting practice by the railroads would be acceptable to the Service and described information to be furnished by the railroads. Plaintiff thereafter furnished to the Service the required information which, in essence, constituted a list of properties subject to depreciation, their cost basis, salvage value, expired life, *273and estimated normal useful life. On September 20, 1944, the Service sent plaintiff a terms letter, incorporating tbe information supplied by plaintiff and some of tbe requirements set out in Mimeo 58. Tbis terms letter was reaffirmed by defendant’s letter of December 14, 1959. Tbe terms letter granted plaintiff permission “to change from retirement to depreciation accounting as of January 1, 1943,” to be effective “upon receipt of a letter agreeing to all tbe terms and conditions set forth herein.” On April 20, 1945, plaintiff accepted tbe terms letter on the condition that “in tbe event that if any of tbe terms and conditions stated in said letter should be changed by statutory amendment, by operation of law, or otherwise,” plaintiff would not be precluded “from tbe benefits of any such changes” and would be “entitled to tbe benefits of any such changes regardless of the acceptance herein contained.”
Mimeo 58 provided in part that “Donated property or contributions or grants in aid of construction from any source must be excluded” from the depreciable base. This statement was not included in the terms letter. However, the schedules of plaintiff’s property, submitted to the Service for which straight line depreciation was requested, did not include the donated property here at issue.
Subsequently, on May 1, 1961, plaintiff submitted to the Service revised schedules for depreciable roadway property and requested the benefit of section 94 of the Retirement-Straight Line Adjustment Act of 1958. The Service responded to plaintiff’s request, noting the terms letter of September 20, 1944, and stating that plaintiff’s revised schedules were acceptable so far as relevant herein. The revised schedules of depreciable property submitted by plaintiff to the Service in 1961 did not include the donated property here at issue.
The issue is whether the facilities are assets properly de-preciable by plaintiff. Plaintiff contends they are property donated to it for use in its business and are depreciable under the rationale of Brown Shoe Co. v. Commissioner, 339 U.S. 583 (1950). Defendant concedes that the facilities are of a character normally subject to allowance for depreciation and *274that to the extent they were paid for by plaintiff, appropriate depreciation deductions are proper. Defendant says, however, that to the extent plaintiff did not pay for the facilities, it cannot depreciate them, relying on Detroit Edison Co. v. Commissioner, 319 U.S. 98 (1943). The parties have agreed to the adjusted tax basis for the facilities in the hands of the plaintiff at the time of acquisition, and the rate of straight-line depreciation applicable, if the court decides that depreciation on the full value is proper. The parties have also agreed that the issue is whether the law of Detroit Edison or Brown Shoe is applicable. Our starting point therefore is a brief discussion of the Detroit Edison and Brown Shoe cases.3
In Detroit Edison, the issue was whether the taxpayer could depreciate the cost of certain electric power lines. The taxpayer, Detroit Edison Company, engaged in the generation of electric power for distribution and sale to the public. It often received applications for service which, in its opinion, would require the construction of power line extensions having a cost not warranted by prospective revenues. Accordingly, the company required the applicants for service to pay the cost of line installations, in part refundable from future revenues collected. The company contended that, to the extent the cost of the lines was not refunded, the customers’ payments were gifts or contributions to the company’s capital and thus were depreciable to it as exhaustible capital assets. The Court held, however, that the payments were not gifts or contributions to the company’s capital but rather were payments for “the price of the service,” i.e., the providing by the company of electrical power to the customers. The Court said “[i]t is enough to say that it overtaxes imagination to regard the farmers and other customers who furnished these funds as makers either of donations or contributions to the Company.” Thus, the taxpayer was not permitted the depreciation deductions it sought.
In Brown Shoe, decided seven years after Detroit Edison, the issue was whether property, including buildings and *275equipment, donated to the taxpayer by community groups, represented “contributions to capital” and were depreciable within the meaning of § 113(a) (8) (B) of the Internal Revenue Code (1939). The property was donated to induce the taxpayer to set up and operate a manufacturing plant in the community. The Court held that the property constituted capital assets in the hands of the taxpayer and was depreciable by it, noting that the “values which the taxpayer received were additions to ‘capital’ as that term has commonly been understood in both business and accounting practice,” and that “contributions to capital may originate with persons having no proprietary interest in the business.” The Court further noted, citing Commissioner v. McKay Products Corp., 178 F. 2d 639, 643 (3d Cir. 1949), that “* * * the assets received * * * are being used by the taxpayer in the operation of its business, * * * will in time wear out, and * * * must eventually be replaced.” Finally, the Court said, in distinguishing the Detroit Edison case, that the “contributions to petitioner were provided by citizens of the respective communities who neither sought nor could have anticipated any direct service or recompense whatever, their only expectation being that such contributions might prove advantageous to the community at large. Under these circumstances the transfers manifested a definite purpose to enlarge the working capital of the company.”
Plaintiff says that the facts here fit Brown Shoe rather than Detroit Edison. Plaintiff notes, among other things, that the facilities in issue are exhaustible assets used in plaintiff’s business, are of a character normally subject to depreciation allowance, and were contributed to plaintiff by state governments whose interest was not to obtain goods or services but rather to benefit the public as a whole. Plaintiff also notes, citing Edwards v. Cuba R.R., 268 U.S. 628, 633 (1925), that the facilities were contributed “unquestionably * * * to increase the capital position of the plaintiff taxpayer for only capital assets were contributed and not monies which could be used for the payment of dividends or expenses ordinarily payable out of earnings or income.” The court concludes that the plaintiff is correct; and although the facts are not *276on all-fours with Brown Shoe, the rationale of that case, rather than Detroit Edison, is controlling.4
Defendant says that plaintiff “has no cost for these facilities and therefore does not stand to lose any portion of its investment by their use or the passage of time * * Plainly, there is no merit to this argument. In Brown Shoe, the taxpayer had no “cost” in the donated property; yet the property was held to be assets depreciable by the taxpayer. Defendant also says that the governmental payments for the facilities “were not intended to be contributions to the capital of the railroad,” but rather were “part of the cost of the state in building its highway system”; and that the facilities are “not related to the production of income but rather to the safety of the local community.” No doubt, the principal purpose of the facilities was to benefit the community-at-large by providing improved safety at railroad-highway intersections. But the fact remains that the facilities enlarged plaintiff’s working capital and were used by plaintiff in its business; and though they may not produce income to the same extent as other railroad property, such as track or freight cars, plaintiff derived economic benefits from them.
Of principal importance, under every contract for constructing the facilities, plaintiff was obligated to maintain and replace the facilities at its own expense. This obligation places squarely on plaintiff the economic loss attendant to wear and tear of the property. As noted in the McKay Products case, supra, at 643, cited in Brown Sk oe,
* * * the assets * * * are being used by the taxpayer in the operation of its business. They will in time wear out * * * and must eventually be replaced. Looking as they do toward business continuity, the Internal Revenue Code’s depreciation provisions * * * would seem to envision allowance of a depreciation deduction in situations like this * * *.
Also pertinent is Helvering v. F. & R. Lazarus & Co., 308 U.S. 252 (1939), and cases cited therein, which holds that depreciation deductions go to the party which “bears the burden *277of wear and exhaustion of business property,” irrespective of who may have legal title.
Defendant contends alternatively that, in accordance with the requirements of Mimeo 58 and the terms letter, plaintiff submitted schedules of its depreciable property which excluded the donated property here at issue from basis. The Commissioner of Internal Revenue accepted plaintiff’s proposed basis for its depreciable road property and permitted plaintiff to make the requested accounting change. Included in the terms letter were the amounts to be included in the cost basis of plaintiff’s various road accounts, together with the proviso that “the remaining sum to be recovered through depreciation allowances shall be limited to ithe cost or other basis less the depreciation so accrued * * Thus, says defendant, plaintiff agreed not to take any depreciation deductions based upon donated property.
Defendant also argues that the Retirement-Straight Line Adjustment Act of 1958 provides in section 94(e) that the terms and conditions of the terms letter would be binding on taxpayers such as plaintiff electing the benefits of that Act from the date of the terms letter until the date of such election.
Mimeo 58 provides in relevant part that:
The basis for depreciation shall be the cost of the existing depreciable property to the present taxpayer, determined in accordance with sections 113 and 114(a) of the Internal Revenue Code. * * *.
* * * * *
The basis may include only the investment in property which is actually depreciable. Thus excavations, dredging, expendable small tools, land improvements, land surveySj etc., are not depreciable expenditures, whereas retaining walls, drainage systems, etc., are. Donated property or contributions or grants in aid of construction from any source must be excluded.
# # ifc ifr if:
In view of the fact that it will be impossible for the Bureau to make a detailed investigation of the depreciation basis, the permission letter includes a mutual understanding that the basis may be corrected to conform to the allowable basis under the Internal Revenue Code *278should subsequent investigation disclose errors of cost or valuation. * * *
The terms letter provides in relevant part:
It is mutually understood that this is an agreement in principle and that a detailed investigation of the depreciation basis has not been made by the Bureau, and that the basis may be corrected to conform to the allowable basis under the Internal Revenue Code should investigation disclose errors of cost or valuation. * * *
Defendant argues for too narrow an interpretation of Mimeo 58 and the terms letter. It seems clear that the purpose of the just-quoted provisions was to establish as the basis for depreciation of the road property the basis allowable under 'Sections 118 and 114(a) of the Internal Revenue Code of 1939. The statement “Donated property or contributions or grants in aid of construction from any source must be excluded” was only the opinion of the Service with respect to what constituted the basis allowable under the Code and does not rise to the level of a condition upon which permission to change depreciation methods would be granted.5 Ultimately, the Supreme Court determined that opinion to be erroneous in the case of Brown Shoe Co. v. Commissioner, supra, and held that, pursuant to Section 113(a) (8) (B) of the 1939 Code, a taxpayer was entitled to include in its basis for depreciation certain donated property.
Even assuming that the statement “Donated property or contributions or grants in aid of construction from any source must be excluded” was 'a condition for the changeover, the provision that the basis for depreciation was to be determined in accordance with the Code should take precedence and control. As indicated, the Brown Shoe case, supra, held that certain donated property could be included in the depreciable base under Section 113(a) (8) (B) of the 1939 Code.
Furthermore, the 1944 terms letter itself, which constitutes the only agreement between the parties, says nothing about *279excluding donated property from tbe basis of depreciable property. "While it might be argued and inferred that plaintiff agreed by implication to such condition in the guidelines, it is just as reasonable to infer that plaintiff acquiesced in the condition, without agreeing with it, simply to avoid possible refusal of its requested change in accounting methods.
Also, assuming again that the above statement was a condition for the changeover, that condition must be viewed in light of the law as it existed when the condition was accepted. As stated in Brown Shoe, supra, at 589, 591, it was the position of the Service that Detroit Edison Co. v. Commissioner, supra, settled the question that donated property could not be added to the basis for depreciation. When plaintiff accepted the terms letter in 1945, its acceptance provided that “in the event that if any of the terms and conditions stated in said letter should be changed by statutory amendment, by operation of law, or otherwise,” plaintiff would not be precluded “from the benefits of any such changes” and would be “entitled to the benefit of any such changes regardless of the acceptance herein contained.” The Supreme Court’s decision in Brown Shoe, supra, distinguishing the case of Detroit Edison, supra, and holding that certain donated property could be added to the depreciable base, produced such a change in the conditions of the terms letter and plaintiff is entitled to the benefits thereof.6
In sum, the facilities in issue are exhaustible assets properly depreciable by plaintiff to the full extent of their value.
Casualty Loss Issue
In 1955,13 freight cars owned by plaintiff were destroyed in accidents while on lines of other railroads. The cars were property used in plaintiff’s business and were emergency facilities subject to rapid amortization (60 months) pursuant to § 124A of the 1939 Internal Revenue Code and § 168 of the 1954 Internal Revenue Code. The cars’ cost to plaintiff was *280$76,007.30. At the time of the loss, amortization had accrued to the extent of $36,886.58, the adjusted basis for the cars thus being $39,120.72. As compensation for the loss, plaintiff received $83,186.35, a gain of $44,065.63 over the adjusted basis. The issue is the tax treatment to be accorded such gain.
Plaintiff says the gain should be taxed as a capital gain, pursuant to § 1231 of the 1954 Code which provides in pertinent part:
fi.A s R bbS fl H Í2 >> cj cd £> 3 ■á O ft ft bJO^ 3 3 o J'-g 2 2 -g <S í Si O 5 r, .H £ § is ■ w 'o 1-1. H¿ £ CD CD £>¡30 CD P j -p, * so* 13 ri Hfeg O Pj ri B ó P cq O co P Pj O CD C/2 £5 CD Bvd“ P P p ri 03 ri ^ cd H*co ££ I g s.r CD - S' * £ * ri CD Pj P4 P PJ P pj CQ P ® $ 02 a® s as ^ '-ftg'gS Q“ ® ^2 a o o co ^ loSSg'&S Sri sa.g S.a 3 ° — ft S-ffi ¡3 ffi m m ri P. CQ .. CD g CQ CQ Pj §> m g'isr O 00 ri* ' p 05 ri-, ^ g ® ¡[8 s 5 I' § “ 5i V-S ri- O * § of S e 2 s o '■d ri <£> b! CQ ?LP B CD Pj B CQ P H — * O O H cq }r CQ /H CD H CQ ^ ® m P rP 0 P CQ O a
Defendant, on the other hand, says that the gain should be treated as ordinary income to the extent of the difference between the adjusted basis of the property under rapid amortization and what the adjusted basis would have been under normal depreciation. The parties have stipulated that under straight-line depreciation, the adjusted basis at the time of loss would have been $67,759.94. The difference in bases, therefore, is $28,639.22 ($67,759.94 less $39,120.72). Defendant relies on § 1238 of the 1954 Code which provides:
Gain from the sale or exchange of property, to the extent that the adjusted basis of such property is less than its adjusted basis determined without regard to section 168 (relating to amortization deduction of emergency facilities), shall be considered as gain from the sale or exchange of property which is neither a capital asset nor property described in section 1231. (Emphasis added.)
The parties agree that plaintiff’s gain resulted from involuntary conversion; that § 1231 applies to gains from “involuntary conversion” as well as “sales or exchanges”; but that § 1238 speaks only of gains from “sale or exchange.” That would appear to end the matter in plaintiff’s favor since it has long been held that an involuntary conversion is *281not a sale or exchange. Helvering v. William Flaccus Oak Leather Co., 313 U.S. 247 (1941). Defendant, however, says that § 1238 should be construed to include gains from involuntary conversions, as well as gains from sales or exchanges, because otherwise plaintiff will “be able to take this gain as a capital gain after having taken deductions against ordinary income for the ‘rapid amortization’ under § 168.” Defendant points to two cases, Towanda Textiles, Inc. v. United States, 149 Ct. Cl. 123, 180 F. Supp. 373 (1960), and Kent Mfg. Co. v. Commissioner, 288 F. 2d 812 (4th Cir. 1961), in which § 337 of the 1954 Code, dealing with gains from sales or exchanges by corporations during liquidation, was construed to include gains from involuntary conversions. Defendant says § 1238 should be construed the same way. Despite an appealing logic to defendant’s position, in my view it cannot prevail.
Section 1238 is brief, clear and unambiguous. It was derived substantially unchanged from § 117 (g) (3) of the 1939 Code. The legislative history of § 117 (g) (3), enacted as part of the Revenue Act of 1950, shows that it was intended as a recapture provision to tax as ordinary income gains made upon the voluntary disposition (sale or exchange) of emergency facilities entitled to rapid amortization, to the extent such amortization reduced the basis of the property below normal depreciation. The Congressional committee report7 which discusses and explains § 117(g) (3) speaks only of voluntary conversions, i.e., “sales” and “exchanges,” and gives examples of each. Nothing is said about involuntary conversions. The pertinent Treasury Regulation (Treas. Reg. § 1.1238-1) likewise speaks only of “sales” and “exchanges.” It is thus clear from the statute, the regulations and the legislative history that the purpose of § 117(g) (3) (and later § 1238) was to prevent taxpayers who are entitled to the benefits of rapid amortization of emergency facilities from later disposing voluntarily of those facilities so to convert accelerated amortization deductions from ordinary income into capital gains. Nothing in the legislative history suggests *282that it was also Congress’ purpose to treat gains from involuntary conversions the same way. There is nothing surprising in this since involuntary conversions (unlike sales or exchanges) by their very nature are not events whose timing can be arranged to make inequitable gains after having had the benefits of rapid amortization.
Furthermore, the very structure of the statute militates against defendant’s position. Section 1231 speaks of gains from “sales or exchanges” and “involuntary conversion.” Yet § 1238, which qualifies § 1231 and refers expressly to it, speaks only of gains from “sale or exchange.” It therefore can hardly be inferred that Congress intended “sale or exchange” in § 1238 to include “involuntary conversion,” particularly in light of well-established judicial precedent that an involuntary conversion is not a sale or exchange. Flaccus, swpra.
Also pertinent to the construction of § 1238 and Congressional purpose for its enactment are §§ 1245 and 1250 of the 1954 Code, added in 1962 and 1964, respectively. Those sections are recapture provisions, akin to § 1238, but deal with gains made upon the disposition of depreciable property subject to normal depreciation, rather than rapid amortization. In general, §§ 1245 and 1250 provide that gains made upon the disposition of depreciable property, which gains exceed the adjusted basis of the property, shall be taxed as ordinary income rather than capital gains. Sections 1245 and 1250 both speak expressly of gains made upon “involuntary conversion,” as well as “sale or exchange.” The inference therefore is clear that when Congress intends recapture provisions to include gains from involuntary, as well as voluntary, dispositions, it so-provides. Section 1238 does not so-provide; and courts cannot rewrite tax laws, however appealing and logical it may be. See F. W. Fitch, Co. v. United States, 323 U.S. 582 (1945), Shakespeare Co. v. United States, 189 Ct. Cl. 411, 419 F. 2d 839 (1969), cert. denied, 400 U.S. 820 (1970).
Towanda and Kent, relied on by defendant, dealt with § 337 of the 1954 Code and the problem of gain derived by a corporation from an involuntary conversion during a period *283of complete liquidation. Section 337 provides, among other things, that during the 12-month period after a corporation adopts a plan of complete liquidation, “* * * no gain or loss shall be recognized to such corporation from the sale or exchange by it of property within such 12-month period.” In both Towanda and Kent, property of corporations in liquidation was destroyed by fire, and insurance proceeds exceeded the basis of the destroyed property. The Government argued that the gain was taxable to the corporations because § 337 speaks only of gains from “sale or exchange,” not involuntary conversion. Both courts held that the gain was not taxable to the corporation because the purpose of § 337 was to avoid double taxation for gains recognized during corporate liquidation, i.e., taxation first to the corporation, then later to the distributee shareholders. This court said in Towanda at 128,129,180 F. Supp. at 376:
The taxation of the gain derived from an involuntary conversion of the property into cash during liquidation is clearly contrary to the declared purpose of Congress in enacting the section [i.e., § 337], which was to avoid double taxation incident to the liquidation of the corporation, by exempting the corporation from liability for gain derived from the disposition of its capital assets, irrespective of whether or not certain formalities had been observed. Literally, an involuntary conversion is not a sale, but what Congress had in mind was a conversion of a corporation’s capital assets into cash, whether voluntary or involuntary, and the distribution of the cash to the stockholders. * * * The purpose was to exempt the corporation from liability for the tax and to collect the tax from the stockholders alone. This being true, we must hold, in order to carry out the clear purpose of Congress, that an involuntary conversion comes within the intent of Congress when it exempted the corporation from liability íor a tax on the gain derived from a sale of its property m liquidation.
The court in Kent agreed with this court’s reasoning in Towanda, noting that a purpose of § 337 was “to avoid double taxation.”
Defendant invites this court to construe § 1238 similarly to the way § 337 was construed in Towanda and Kent. In my view, this should not be done simply because (a) Towanda *284and Kent dealt with a unique problem, i.e., double taxation flowing from corporation liquidation, and § 337 was construed to carry out “the clear purpose of Congress” of avoiding double taxation, and (b) § 1238, its legislative history and the regulations are clear and unambiguous, and show no Congressional purpose to include gains from involuntary conversions as one of the exceptions to the application of § 1231.
In sum, the gain in question is taxable as a capital gain pursuant to § 1231 of the 1954 Code.
Welded Rail Issue and Bail Salvage Value Issue
These issues revolve around a method of accounting known as retirement-replacement-betterment accounting. Actually, retirement-replacement-betterment accounting is a shorthand term for three separate but interrelated accounting schemes— retirement accounting, replacement accounting and betterment accounting — -by which plaintiff and other railroads determine allowable depreciation for property in track accounts, particularly rail and joint materials (angle bars, bolts and washers). To understand the issues, it is necessary to discuss in some detail the theory and practice of retirement-replacement-betterment accounting.
Retirement accounting
Normally, in computing depreciation for an exhaustible asset, the cost (or other basis) of the asset, less its salvage value, is spread ratably over the asset’s useful life; and through annual depreciation charges, the cost is recouped. This means that the asset’s book value decreases from year to year as depreciation charges accumulate. Capital additions and improvements made to the asset from time to time during its service life are added to the book value and, in turn, are depreciated. The theory is that at t'he end of the asset’s useful life, the accumulated depreciation charges plus salvage value will equal original cost.
Retirement accounting, however, works differently. Bather than making annual adjustments for depreciation, the asset is carried on the books at its full value (usually cost) during its useful life. Then, at the time of retirement from service, *285the book value, diminished by the asset’s salvage value, is charged to current expense. Retirement accounting thus results in deferred depreciation for any given asset. However, over an extended period of time, the depreciation deductions taken under retirement accounting for all assets in the account should closely approximate conventional ratable depreciation methods. As stated in Boston da Maine R.R. v. Commissioner, 206 F. 2d 617, 619 (1st Cir. 1953):
* * * 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. * * *
Retirement accounting is used by the railroads for their track accounts principally because it simplifies the difficult bookkeeping problem of making annual depreciation adjustments for the large number of fungible assets in such accounts, i.e., rail and joint materials, and because it complies with the accounting requirements of the Interstate Commerce Commission as set out in the Uniform System of Accounts for Railroad Companies.
jReplacement accounting
Replacement accounting works hand-in-glove with retirement accounting. It arises when a retired asset, rather than simply being removed from service, is replaced by a like asset. From an accounting standpoint, a replacement transaction comprises two steps: the retirement of one asset and the addition (or replacement) of a like asset. Logically, the accounting treatment for such transaction should entail (a) charging off to current expense the book value of the retired asset, diminished by its salvage value, and (b) capitalizing the cost of the added (or replacing) asset. So treated, the books would tend to reflect, over the life of the account, the *286current value of all assets in service. However, plaintiff does not handle replacement transactions in this fashion. Rather, it takes as a current expense the cost of replacement, diminished by the salvage value of the retired asset, and leaves on the books the value of the retired asset. The theory is that each replacement, when considered in the context of the account as a whole, is a minor transaction, akin to a repair, and should be currently expensed. The result is that, after many years, the composite book value of all assets in service is a residuum of original costs, increased or decreased, as the case may be, by additions and betterments, or retirements without replacement, made from time to time over the life of the account.
On reflection, it can be seen that over a long period of years of no price inflation, it makes no difference from an accounting and depreciation standpoint whether a replacement transaction is handled in one or the other of the two above-noted fashions. To illustrate, assume a section of rail having an original cost of $1,000 is retired after 40 years and is replaced by a like section of rail costing $1,000, and that the salvage value of the retired rail is $200. Under either above-described method of replacement accounting, the book value of the rail in service remains $1,000 and the current charge to expense (representing depreciation) is $800 ($1,000 less $200). In contrast, over a period of substantial price inflation, it makes considerable difference which of the two methods is used. To illustrate, assume the same facts as above except that the cost of replacement, due to price inflation, is $2,000 rather than $1,000. Now, if the accounting transaction is handled by the first method, the charge to operating expense is still $800 ($1,000 less $200), and the $2,000 cost of replacement is capitalized to become the new book value of the rail in service. On the other hand, under plaintiff’s method, the $2,000 is currently expensed, less $200 salvage value for the retired rail, leaving a current charge (representing depreciation) of $1,800; and the book value of the rail in service remains $1,000.
Defendant has no quarrel with the way plaintiff chooses to *287handle replacement aceoimting.8 It conforms to the Interstate Commerce Commission’s requirements for accounting for road property; and, indeed, it is consistent with the theory of replacement accounting in that the cost of replacements of individual assets of a whole account are deemed minor and, for ease of accounting, are currently expensed. A dispute arises, however, over the salvage value to be assigned to rail picked up from service but not disposed of as scrap. This dispute is later discussed in detail.
Betterment aceoimting
Betterment accounting works hand-in-glove with replacement and retirement accounting. It arises in one of two situations: (a) when additional rail is laid where no rail before existed, or (b) where rail laid in replacement is of heavier weight, and thus “better,” than rail being replaced. In the first situation, the current cost of laying additional rail is capitalized, and such cost thereafter remains on the books until the rail is retired. In the second situation, the cost of the “betterment” portion of the rail laid in replacement is capitalized, while the remainder of the cost, reduced by the salvage value of the replaced rail, is currently ex-pensed. To illustrate, if a section of 80-pound rail9 is replaced with 100-pound rail, the cost attributable to the extra 20 pounds is capitalized; and the remaining cost, diminished by the salvage value of the 80-pound rail picked up, is currently expensed. In this way, the cost of additions and improvements is added to the books and remains in the account until retirement.
Welded mil issue
In 1955, as part of a continuing program of track renewal, plaintiff replaced a number of 39-foot lengths of rail with 78-foot lengths of heavier-weight rail. The 39-foot lengths which were replaced were fastened together by angle bars *288and bolts as had been conventionally done in the railroad industry for many years. The new 78-foot lengths were made by welding together two standard 39-foot rails as purchased from steel mills. In turn, the 78-foot lengths were bolted together in the track structure. Thus, in making the renewals, plaintiff substituted two heavier-weight 39-foot lengths of rail having a welded joint for two lighter-weight 39-foot lengths of rail having a bolted joint. Plaintiff accounted for this transaction under the retirement-replacement-betterment method as follows: It charged to current expense the portion of the rail cost and welding cost attributable to the same weight as the rail which was replaced (replacement in kind); it capitalized the portion of the rail cost and welding cost attributable to the weight of the new rail greater than the replaced rail (betterment); and it charged to current expense the book value of the joint materials (angle bars, bolts and washers) which were replaced by the welded joints (retirement). The cost of making welded joints in 1955 was $155,748, of which $140,808 was charged to current expense and $14,940 was capitalized, in accordance with the pro rata apportionment above described.
The dispute between the parties is whether plaintiff properly accounted for the cost of welding. Defendant says plaintiff should have capitalized the entire welding cost ($155,748), rather than only part of it ($14,940), because a welded joint constitutes a “betterment” over a bolted joint within the context of retirement-replacement-betterment accounting. Defendant relies on Eev. Eul. 67-22, 1967-1 Cum. Bull. 52, which, in essence, holds that under retirement-replacement-betterment accounting, the cost of welded joints, whether laid with new rail or replacement rail, must be capitalized because “welding of rail creates something new or better by substitution or addition of different materials, reduces track renewal cost, prolongs rail and rolling stock life, reduces maintenance costs, and increases the value of the track structure.” The Eevenue Euling notes § 263 of the Internal Eevenue Code (1954) which provides in pertinent part that no deductions shall be allowed for “any amount paid out * * * for permanent improvements or *289betterments made to increase the value of any property or estate * * Plaintiff, on the other hand, contends that under retirement-replacement-betterment accounting, the cost of welding should be currently expensed in full because “the cost of the weld is a part of the cost of the rail itself and should be accounted for on the same basis as is the rail * * Plaintiff also says that welded joints are not betterments over bolted joints because they do not increase the value of the track structure, do not prolong the life of the rail, and do not add new and improved materials to the track system.
Thus framed, it is important to point out that the issue is not whether the cost of replacing bolted joints with welded joints is a repair as opposed to a capital expenditure. The parties agree that the expense is capital in nature. The issue is simply whether welded joints, which replace bolted joints, are such an improvement or betterment to the track that their cost should be capitalized, rather than currently ex-pensed, in accordance with the method of retirement-replacement-betterment accounting used by plaintiff. Defendant agrees that if bolted joints are replaced by bolted joints, the cost of the replacement, less the salvage value of the replaced materials, is chargeable to current expense. Defendant says simply that welded joints are so different from bolted joints and are such an improvement over bolted joints that their cost must be fully capitalized and not charged to current expense until the welded rail is retired or replaced by other welded rail.
The issue boils down to an analysis of the differences between bolted joints and welded joints, both from a cost and technology standpoint. Functionally, the joints do the same job — they hold together the ends of pieces of rail so to make a continuous track structure. Thus, welded joints which replace bolted joints add no new or different function to the rail system. Bolted joints comprise a pair of angle bars which bridge the rail ends across the rails’ web and are fastened to the web by six bolts and washers. In 1955, it cost plaintiff about $10 to install a bolted joint, depending on the weight of the rail.10 Bolted joints require maintenance from time to *290time since they tend to come loose with wear. The record shows that bolt tightening is required the first year after installation and generally each two years thereafter. After extensive use, the bolts and washers may need replacing.
Welded joints, in contrast to bolted joints, are made by fusing together the ends of the rail under intense heat. No material is added to the rail. Bather, about % inch is lost off each rail during the welding process because the rail ends are forced together under pressure, thus creating a bead of raised metal at the weld which must be ground down to make a smooth joint. In 1955, it cost plaintiff $11.83 to make a welded joint. Welded joints do not require maintenance like bolted joints; and in fact the record shows that during the first 12 years of service of welded rail (1955-1967), welded joints required no substantial maintenance at all. However, the record also shows that welded joints are not an unmixed blessing. About 1967, welded joints installed on plaintiff’s main lines in 1955 exhibited serious wear and deterioration known as secondary batter. Secondary batter results from the fact that the fused metal at the weld is harder than the rail on either side of the weld. Thus, the joint wears less rapidly than the main rail. After extensive use and track wear, the metal adjacent either side of the weld tends to dish out as the wheels of railroad cars ride over the weld, i.e., the wheels tend to batter the track downstream of the weld. Secondary batter becomes progressively worse with time and particularly if trains run both directions on the track. Though no maintenance to correct secondary batter was done by plaintiff between 1955 and 1967, the record shows that substantial maintenance will be required soon and that such maintenance will be necessary from time to time during the 40-50 year useful life of the rail. The maintenance will involve grinding down the track and weld to eliminate dished-out areas, or in extreme cases, cropping out the damaged areas and rewelding the rail ends.
Against these facts, it must be concluded that the record does not support defendant’s position that welded joints vis-a-vis bolted joints constitute a “betterment” to the track system in terms of retirement-replacement-betterment ac*291counting. From a purely monetary standpoint, the cost per welded joint is about the same as a bolted joint ($11.83 v. about $10). Though this factor alone is not determinative of the “betterment” question, plaintiff’s investment in welded joints is not substantially greater than if it had used all bolted joints, and thus shows that plaintiff has not made a substantial increase in monetary value in its track system. Defendant, however, points to other factors which it says show that welded joints are “a substantial improvement over the previous method of bolted joints.” It notes that less maintenance is required on welded joints and argues that welded joints prolong the life of the rail and rolling stock. Though, as noted earlier, welded joints require little or no maintenance during early years of use, the problem of secondary batter will ultimately require maintenance of considerable cost. Thus, it cannot be said with any certainty on this record that over the long run, i.e., the 40-50 year useful life of the rail, bolted joints will cost substantially less to maintain and repair than welded joints. As for prolonging the useful life of rail, the record does not show that rail with welded joints will last longer than rail with bolted joints if it be assumed that both types of joints are properly maintained. Useful life of rail is principally a function of the quality and extent of use of the rail itself upon which the type of rail joint, if properly maintained, would appear to have little effect. As for prolonging the life of rolling stock, the evidence is not sufficient to conclude that there has been any significant, or even measurable, reduction in rolling stock maintenance costs since the advent of welded rail, particularly in light of problems of secondary batter which has the same deleterious effects on rolling stock as worn bolted joints. Thus, while the record shows that welded joints result in some advantages over bolted joints, the advantages are not so substantial and track system so improved that welded joints should be considered a betterment over bolted joints for purposes of retirement-replacement-betterment accounting. Plaintiff may therefore charge to current expense the cost of replacing bolted joints with welded joints, in the same proportion as it charges to current expense the cost of replacing the rail itself.
*292Defendant points to another reason why plaintiff should capitalize the cost of replacing bolted joints with welded joints. In accounting for the replacements, plaintiff charged to current expense not only the cost of the welded joints but also the book value of the replaced bolted joint materials (angle bars, bolts and washers). I.e., plaintiff treated the accounting for the replaced joint as if it had been retired without replacement. Defendant says this is improper under replacement accounting as practiced by plaintiff because the asset account must reflect the fact that a joint was replaced, rather than simply retired. Plaintiff, on the other hand, says it is proper to charge to current expense both the welding cost and the book value of the retired bolted joint materials because the bolted joint was retired and the cost of the weld is part of the cost of rail. Clearly, defendant is correct. As held above, the cost of replacing a bolted joint with a welded joint should be accounted for as a replacement, i.e., charged to current expense. Accordingly, such cost must be reduced by the salvage value of the retired bolted joint materials. Otherwise, plaintiff would be getting what is in effect a compounded deduction for a transaction which is in fact a replacement rather than simply a retirement without replacement.
In sum, the proper accounting treatment should be as follows : The cost of welded joints laid in replacement should be charged to current expense (in the proper proportion as above noted), and the charge should be diminished by the salvage value of the replaced bolted joint materials. Thus, the original book value of the joint materials remains in the account; and the accounting transaction is fully consistent with the method used by plaintiff for other replacements.
Rail salvage value issue
To understand this issue, a brief discussion of plaintiff’s track renewal program will be helpful. Over the years and particularly with the advent of heavier rolling stock, plaintiff has sought to upgrade its track system by replacing lighter-weight rail with heavier-weight rail. Typically, the rail on main lines is replaced from time to time with heavier new rail; and the rail picked up off the main lines is used to *293replace lighter rail in secondary or branch lines. In turn, rail picked up from secondary or branch lines, if still in serviceable condition, is used to lay or replace industry, spur or yard tracks. If not in serviceable condition, it is scrapped. Finally, industry, spur or yard rail is replaced when worn out and is sold as scrap. Ordinarily, a section of rail goes through all these steps during its 40-50 year useful life before disposal as scrap.
In accounting for track replacements under tbe Interstate Commerce Commission’s Uniform System of Accounts for Railroad Companies, plaintiff assigns on its books a salvage value to rail picked up. This salvage value is important because, as earlier discussed, the charge to current expense taken under retirement-replacement-betterment accounting for the cost of rail laid in replacement and rail retired without replacement is reduced by the salvage value of the replaced rail in order to arrive at a figure representing allowance for depreciation for Federal income tax purposes for the asset account as a whole. The salvage value assigned by plaintiff to rail picked up depends on whether the rail is to be scrapped or is to be relaid as reusable rail. In 1955 and for many years previous thereto (since about 1919), plaintiff has assigned the values of $17.86 per net ton for rail to be sold as scrap and $22.32 per net ton for reusable rail. Plaintiff does not explain how it arrived at these figures. However, a fair inference from the record is that $17.86 per net ton was the approximate price of scrap in the 1920’s; and $22.32 per net ton was a figure somewhere between the cost of new rail and scrap in the 1920’s, and thus was deemed to be the reasonable value of reusable rail. In 1919, the price of new rail was $35.83 per net ton; during the 1920’s, the price was about $38.
The issue here is the value to be assigned to reusable rail in 1955, to the extent that such rail was relaid as additions or betterments. Defendant says that $22.32 per net ton is unrealistically low because it represents a figure established many years ago when the cost of new rail and the price of scrap were but a fraction of their 1955 values. In 1955, new rail cost $93.70 per net ton and scrap sold for $43.75 per net *294ton. Accordingly, says defendant, reusable rail should be assigned a value somewhere between the price of new rail and the price of scrap, at a figure deemed to be the current fair market value of reusable rail. Defendant argues that such value in 1955 should be $69.40 per net ton, midway between the price of new rail and the price of scrap. Defendant cites and relies on Eev. Eul. 67-145,1967-1 Cum. Bull. 54, which holds that “railroads using the ‘retirement method’ of accounting for depreciation” must value reusable rail at “fair market values.” Defendant also cites and relies on Eev. Proc. 68AL6,1968-2 Cum. Bull. 961, which holds that reusable rail laid in additions or betterments must be valued at a “fair market value somewhere between the new and scrap price for such rail,” and such fair market value is determined “by averaging the new and scrap prices.” Defendant says that it is reasonable to calculate fair market value in this way because (a) there is no established price for reusable rail in the marketplace since reusable rail is not sold by the railroads, due to a shortage of rail, and (b) reusable rail is worth less than new rail and more than scrap, and a value halfway between “has the merit of simplicity, reasonableness, and ease of derivation.”11
Plaintiff objects to defendant’s approach to the valuation problem. While conceding that $22.32 per net ton is unrealistically low, plaintiff contends that the proper value should be the lower of cost or fair market value.12 Plaintiff relies on a letter of technical advice, issued in 1964 by the National Office of the Internal Eevenue Service relating to an audit of plaintiff’s books for 1956-59, which holds that reusable rail should be valued at “the lower of actual costs or current fair market values.” The letter of technical advice was later *295superseded by Eev. Eul. 67-145 and Eev. Proc. 68-46, sufra. In 1955, tlie average investment cost of plaintiff’s rail in place was $48 per net ton, and it is this value which plaintiff says should be assigned to reusable rail. Otherwise, says plaintiff, if fair market value is used under defendant’s formula, the book value of reusable rail laid as additions or betterments is, in effect, written up above cost and results in the taxation of unrealized appreciation.
At this juncture, it should be pointed out that defendant does not challenge the value which plaintiff assigned to scrap in 1955 ($17.86 per net ton) even though the price of scrap in that year was $43.75 per net ton. The reason for this is that plaintiff reports as income the difference between the assigned scrap value and the proceeds of scrap sales so that, as a practical matter, it makes no difference what the assigned value is. Thus, the net effect is that the entire $43.75 per net ton current scrap value goes to reduce the charges to expense for replacements and retirements without replacement. By the same token, if plaintiff had assigned to scrap in 1955 a value higher than $43.75, say for example $50.75, the $7 difference, upon disposition of the scrap, would be charged to current expense, thus reflecting the fact that the decrease to current expense representing the assigned salvage value of scrap had been too high. Furthermore, defendant does not challenge the value plaintiff assigned in 1955 to reusable rail which was picked up and then relaid as replacements in kind. On reflection, it can be seen that the reason for this is that whatever value was assigned washed out of the accounting because the value assigned on pickup, which went to reduce current expense, was the same as the value charged to current expense upon relay, thus a wash. Therefore, defendant’s only concern here, as noted earlier, is the value to be assigned to reusable rail laid as additions or betterments, for this value is capitalized on plaintiff’s books and no corresponding-charge to current expense is made until ultimate retirement.
After careful consideration of all the arguments in the parties’ extensive briefs, I am constrained to conclude that defendant’s position is the correct one. The evidence, including the testimony of expert accounting witnesses, persuades *296me ithat the approach set out in Rev. Rui. 67-145, supra, and Rev. Proc. 68-46, supra, is sound. The issue turns on the manner in which plaintiff employs retirement-replacement-betterment accounting. As discussed earlier, plaintiff charges to expense the current cost of making rail replacements. The aggregate of such costs, reduced by the salvage value of replaced rail, thus represents the allowable depreciation deduction for such transactions for the account as a whole. After many years of an inflationary economy (particularly since World War II) and because rail has a long useful life (40-50 years), the current cost of making replacements (as opposed to retirements without replacement) bears no relationship to the historical cost of the rail in place. Thus, in 1955, plaintiff charged to current expense the $93.70 per net ton cost of laying new rail in replacement, even though the average cost of all rail in place was only $43 per net ton. The symmetry of replacement accounting demands that if the high costs of replacement are to be charged to current expense, then salvage value assigned to reusable rail for relay as additions or betterments must correspondingly be based on current values. Otherwise, the allowable deduction for depreciation for the account as a whole is distorted, will in effect tend to constitute accelerated depreciation, and thus will not reflect a reasonable allowance for “exhaustion, wear and tear (including a reasonable allowance for obsolescence),” as required by § 167 of the Internal Revenue Code (1954).
Plaintiff’s argument that valuing reusable rail at current fair market value rather than historical cost results in taxation of unrealized appreciation is, at first blush, appealing but will not stand analysis for two reasons. First, under retirement-replacement-betterment accounting, as practiced by plaintiff, the value assigned to reusable rail is for the purpose of determining the extent to which the current charge to expense for replacements (and retirements without replacement) should be reduced, so to reflect a reasonable allowance for depreciation. Thus, although the net effect of valuing reusable rail at current fair market value rather than historical cost may result in less after-tax income to plaintiff, it does so not by taxing unrealized appreciation but simply *297by reducing a depreciation deduction to a reasonable level. Secondly, it makes no sense in the context of plaintiff’s retirement-replacement-betterment accounting method to restrict the value assigned to reusable rail laid in additions or better-ments to historical cost when the charges for replacements, which are immediately expensed, bear no relationship to historical cost. Under plaintiff’s theory, reusable rail would be valued in 1955 at $43 per net ton, which is less than the price of scrap ($43.75 per net ton) for that year. This is not reasonable and, as earlier discussed, violates the symmetry of retirement-replacement-betterment accounting.
Furthermore, the requirement of the Interstate Commerce Commission’s Uniform System of Accounts for Eailroad Companies for valuing reusable rail is fully consistent with defendant’s proposition that reusable rail be assigned its current fair market value. Section 10.01-7 (d) of the Uniform System of Accounts for Eailroad Companies, in effect in 1955, defines “value of salvage,” stating that “when such material [e.g., rail] is retained and again used by the carrier [e.g., reusable rail], the value shall he determmed by deducting a fair allowance for depreciation from eurrent prices of the material as new." The plain meaning of this is that reusable rail in 1855 should be valued at somewhere between $93.70, the current price of new rail, and $43.75, the current price of scrap, so that its value represents the price of the material “as new” less “a fair allowance for depreciation.” Defendant’s method of estimating this value, i.e., halfway between the current prices of new rail and scrap, is reasonable and equitable since, on the average, reusable rail picked up by plaintiff from its lines has gone through about half its useful life. Plaintiff argues that new rail prices and scrap prices vary “due to wide fluctuations” and that it is not possible “on a current basis” to ascertain the average value. While the record shows that such prices do vary somewhat, the fluctuations in any given year are not so great as to make determination and use of a reasonable average an undue administrative burden. And, in any event, adjustments can be made at the end of any year to reflect the actual price experience for that year.
*298It is pertinent to note that in 1955, the Interstate Cr merce Commission did not require plaintiff to value • reusable rail in accordance with the express dictates ' §10.01-7(d), supra. Rather, plaintiff was permitted "i assign the value of $22.32 per net ton, a figure which no dor .A was consistent with new rail prices and scrap prices many years ago but clearly is not a proper figure in a year when new rail cost $93.70 per net ton and scrap sold for $43.75 per net ton.13 Whatever may be the reason for the Commission’s acquiescence in a value of $22.32 in 1955, the fact remains that § 10.01-7 (d) of its Uniform System of Accounts for Railroad Companies expressly calls for determining salvage value in a manner wholly consistent with the theory and practice of retirement-replacement-betterment accounting here urged by defendant.14
Plaintiff contends that if it is required to assign to reusable rail laid as additions and betterments in 1955 such rail’s fair market value, then it should be permitted to value scrap at the same figure. Thus, plaintiff argues, scrap should be valued at $69.40 per net ton so that upon disposition at $43.75 per net ton, it sustains a “loss,” which “loss,” says plaintiff, is “an offset against defendant’s offset.” This argument makes no sense within the context of retirement-replacement-betterment accounting as practiced by plaintiff. As earlier noted, the value plaintiff assigns to scrap at the time of pickup is, in reality, an interim figure which is adjusted depending upon the price for which scrap is actually sold. Thus, if scrap is *299valued on pickup at lower-tban-prevailing scrap prices (as it was in 1955), the difference upon disposition is reported as income. If, on the other hand, scrap is valued on pickup at higher-than-prevailing scrap prices, the difference on disposi-sition should be charged to current expense. In either event, the final result must be that scrap is accorded its fair market value which goes to reduce current charges for replacements and retirements without replacement. This is fully consistent with the theory and practice of retirement-replacement-betterment accounting as used by plaintiff and as earlier discussed.
Finally, plaintiff contends that Eev. Eul. 67-145, supra, upon which defendant relies, is not applicable in this case because it was issued in 1967, well after the tax year in suit, and cannot be given retroactive effect. Eather, plaintiff says defendant should be bound by the letter of technical advice issued in 1964 with respect to the audit of plaintiff’s books for 1956-59. There is no merit to this argument. The Commissioner of Internal Bevenue is empowered retroactively to correct mistakes of law in the application of the tax laws to particular taxpayers. Dixon v. United States, 381 U.S. 68 (1965). Rev. Rul. 67-145, supra, supersedes and correctly overrules the letter of technical advice, and there was no abuse of discretion in applying it here in support of the Government’s setoff defense. Furthermore, plaintiff has shown no detrimental change of position in reliance on the letter of technical advice from which it can argue that retroactive application of the Eevenue Buling is inequitable.
To sum up: Plaintiff, in accounting for reusable rail picked up in 1955, must assign to such rail, to the extent it was laid as additions and betterments, its current fair market value in 1955, which equitably and for ease of calculation and administration is halfway between the cost of new rail and the price for scrap in 1955.
PROTECTIVE WORK ISSUE
In 1955, plaintiff carried out 28 construction projects (total cost: $121,026.70) directed to protecting and maintaining its rail embankments near waterways. The work was *300generally of five types: (a) channel diversion, (b) construction of dikes, dams and jetties, (c) bridge grade raising, (d) riprap construction, and (e) bridge and culvert extensions or modifications. The project relating to channel diversion entailed dredging a creek bed to straighten it and divert the flow of water away from plaintiff’s track embankment which was being eroded away. The creek had changed course and was threatening the integrity of the embankment. Six projects related to construction of dikes, dams and jetties. Five of them involved building rock-filled timber cribs, steel jetties and, in one instance, a retaining wall, in streams or rivers to deflect waterflow away from plaintiff’s track embankments. The sixth project involved placing an earth blanket over seepage points in a levee to protect the levee and a railroad bridge during high water of a river. The project relating to grade raising involved raising the end of a railroad bridge to allow passage thereunder of drift material which endangered the bridge during heavy rains. The river which the bridge spanned had silted up, due to upstream construction of a dam, thus raising the river level at times of high water. Two projects related to riprap construction. In both projects, riprap was placed along the banks of a river or creek to slow down and divert waterflow and prevent further erosion which was endangering plaintiff’s rail embankments. Eighteen of the projects related to bridge and culvert extensions. In general, the projects involved modification of existing culverts, trestles, walls, and pipes to alter or divert waterflow which was eroding rail embankments. In some instances, new culverts, trestles, walls, or pipes were installed near existing facilities of a like nature, the purpose of which was to arrest erosion of the embankment areas due to changed waterflow patterns.
Railroad embankments have an indefinite service life. Unless attacked by outside forces, they remain in serviceable condition so long as the railroad operates trains over its rail lines. Embankments near waterways are built originally by plaintiff using good engineering practices to provide permanent protection against water erosion. However, unforeseen and unforeseeable changes in waterflow patterns near the *301embankments sometimes require remedial steps to arrest erosion. It is not possible to predict when remedial steps will be necessary. Plaintiff is engaged in a continuous program of inspection and evaluation of damage to its embankments. When necessary to correct deterioration, the steps taken are intended to arrest the problem and prevent a premature end to the service life of the embankment. The nature of the corrective work done depends upon the severity of damage and the possibility of continued damage. The work is done in the simplest and cheapest manner under the circumstances to insure an end to the particular erosion problem and to keep the rail line in safe operating condition.
Plaintiff contends that the costs of the projects here in issue are deductible as “ordinary and necessary” business expenses pursuant to § 162 of the 1954 Internal Revenue Code. Plaintiff says “the work done was in the nature of repairs and maintenance” in order to “maintain the * * * property [i.en the rail lines and embankments] in an ordinary efficient operating condition or to prevent the useful life from being shortened due to unusual circumstances.” Plaintiff points to Treas. Reg. § 1.162-4 which says in pertinent part that the cost of repairs “which neither materially add to the value of the property nor appreciably prolong its life but keep it in an ordinary efficient operating condition, may be deducted as an expense * * *.”
Defendant, on the other hand, says that the costs in issue must be capitalized pursuant to § 263 of the 1954 Code which provides in part that no deductions shall be allowed for “* * * any amount paid out for new buildings or for permanent improvements or betterments made to increase the value of any property or estate.” Defendant also relies on Treas. Reg. § 1.263 (a)-1 which provides in part that no deduction shall be allowed for “* * * amounts paid or incurred (1) to add to the value, or substantially prolong the useful life, of property owned by the taxpayer such as plant or equipment, or (2) to adapt property to a new or different use.” Defendant says that each project here in issue consists of “a new facility or addition of a permanent nature,” and points particularly to the retaining walls and culverts as examples of *302a “capital asset or improvement, representing new and permanent construction with a useful life extending beyond the year in which the expenditure was incurred.”
Whether expenditures are currently deductible or must be capitalized depends on the purpose for which the expenditures are made. As stated in Griffin & Co. v. United States, 182 Ct. Cl. 436, 456, 389 F. 2d 802, 814 (1968) :
* * * If their purpose was merely to repair property and thus keep it in good condition during its probable useful life, the expenditures were deductible as business expenses. Illinois Merchants Trust Co., 4 B.T.A. 103, 106 (1926). On the other hand, if the purpose of the expenditures in question was to improve property and thus increase its value, the expenditures were capital in nature and were not deductible as business expenses. Duffy v. Central R.R., 268 U.S. 55, 62-63 (1925). The question of purpose is, of course, a factual issue. Russell Box Co. v. Commissioner, 208 F. 2d 452, 454 (1st Cir. 1953).
The record here supports plaintiff. The construction work, though differing in detail from project to project, all was undertaken as remedial steps to correct problems of water erosion which threatened to render unserviceable already existing railroad embankments, some of which had been in place for as long as 60 years. The value and original service life of the embankments were not increased by the corrective work done; nor were the embankments adapted to any new or different use. Plainfleld-Union Water Co. v. Commissioner, 39 T.C. 333 (1962). Bather, the work simply arrested and corrected a condition of deterioration which threatened a premature end to the otherwise indefinite service life of the embankments. American Bemberg Corp. v. Commissioner, 10 T.C. 361 (1948), aff'd, 177 F. 2d 200 (6th Cir. 1949); Midland Empire Packing Co. v. Commissioner, 14 T.C. 635 (1950). The fact that some of the construction work may have been relatively permanent and had a useful life of more than one year is not controlling. Connecticut Light & Power Co. v. United States, 156 Ct. Cl. 304, 299 F. 2d 259 (1962). Nor is it controlling that the work restored the embankments to their pre-erosion condition and added value as compared with the situation immediately prior to the work. The proper test, as stated in Oberman Mfg. Co. v. Commissioner, 47 T.C. *303471 (1967), is “whether the expenditure materially enhances the value, use, life expectancy, strength or capacity as compared with the status of the asset piior to the condition necessitating the expenditure.”
This court’s decision in Kansas City Southern R.R. v. United States, 125 Ct. Cl. 287, 112 F. Supp. 164 (1953), is fully in point.15 There, the integrity of railroad track beds was threatened by water pockets which developed in the subsurface beneath the beds. The problem was corrected by driving into the ground at the end of each railroad tie wooden poles six inches in diameter, at a cost of $116,548.77. This court held that such cost was deductible as a current expense because the problems created by water pockets “appear casually, and, as they appear, they present a problem of maintaining an existing completed track in a safe and economical operating condition.” The court further pointed out that the fact that the poles themselves had an anticipated useful life of many years was not significant because the “railroad track, and the parts of it where the poles were driven were no more useful than the other parts which had not needed this work.” The rationale of the Kansas Gity case is applicable here. The record shows that the erosion which threatened the integrity of plaintiff’s embankments “appears casually” at times and places which plaintiff cannot predict and was due to changed waterflow conditions. The erosion problems were thus corrected as they appeared, to protect and maintain the embankments in a safe operating condition. This is not a case where, at the time of original construction of the embankments, it was possible to foresee and make permanent preparation for all contingencies of erosion damage. In fact, plaintiff’s embankments were constructed originally under sound engineering practices to protect against waterflow as it then existed; and the construction was done with the expectation that the embankments would last indefinitely. Were it otherwise, the costs here in issue might properly be considered as capital expenditures.
At the oral argument in this case, defendant argued that the instant situation is distinguishable from Kansas City *304Southern R.R. v. United States, supra, because plaintiff set up on its books separate accounts for the cost of the protective work bere ait issue in such a manner as to indicate that those costs were treated as capital improvements, and therefore as depreciable assets rather than as repairs or maintenance expenses.
Plaintiff acknowledges that the cost for the protective work at issue was placed in Interstate Commerce Commission accounts 3 and 6 which are capital accounts. However, it is by now well-settled that bookkeeping entries, while in some circumstances of evidential value, are not determinative for Federal tax purposes. That decision must rest upon the actual facts. Doyle v. Mitchell Brothers Co., 247 U.S. 179, 187 (1918); Helvering v. Midland Mut. Life Ins. Co., 300 U.S. 216, 223 (1937); A. Kreamer, Inc. v. United States, 66 Ct. Cl. 308, 316 (1928) ; Allen v. Commissioner, 117 F. 2d 364, 368 (1st Cir. 1941); Commissioner v. North Jersey Title Ins. Co., 79 F. 2d 492, 493 (3d Cir. 1935); Sitterding v. Commissioner, 80 F. 2d 939, 941 (4th Cir. 1936); Commissioner v. Landers, Corp., 210 F. 2d 188, 192 (6th Cir. 1954); Northwestern States Portland Cement Co. v. Huston, 126 F. 2d 196, 199 (8th Cir. 1942); Terminal Investment Co., 2 T.C. 1004, 1013 (1943). Any doubt that the rule is otherwise in the instant situation because these accounts were established by the Interstate Commerce Commission was early dispelled in Old Colony R.R. v. Commissioner, 284 U.S. 552 (1932), where the Supreme Court stated:
* * * Moreover, the rules of accounting enforced upon a carrier by the Interstate Commerce Commission are not binding upon the Commissioner, nor may he resort to the rules of that body, made for other purposes, for the determination of tax liability under the revenue acts. * * * [Id. at 562.]
Accord, Connecticut Light & Power Co. v. United States, 156 Ct. Cl. 304, 310, 299 F. 2d 259, 263 (1962); New York Central R.R. v. Commissioner, 79 F. 2d 247, 252 (2d Cir. 1935), cert. denied, 296 U.S. 653; Mine Hill & Schuylkill Haven R.R. v. Smith, 184 F. 2d 422, 426 (3d Cir. 1950), cert. denied, 340 U.S. 932 (1951); Helvering v. Edison Bros. *305Stores, 133 F. 2d 575, 579 (8th. Cir. 1943), cert. denied, 319 U.S. 752.
Defendant also argues that Mimeo 58, the lists of depre-ciable property submitted by plaintiff, the terms letter, plaintiff’s acceptance thereof, and the later revised schedules demonstrate that the I.C.C. accounts were expressly established for tax purposes. The above discussion with respect to the donated property issue is equally applicable to this argument by defendant. The purpose of the schedules of depreciable property, Mimeo 58, and the terms letter was to provide that the basis for depreciation of the road property was to be determined in accordance with the Internal Revenue Code. Even assuming that the statement “Donated property or contributions or grants in aid of construction from any source must be excluded” was a condition independent of the just-stated purpose, plaintiff’s acceptance thereof was conditional and plaintiff is entitled to the benefits of the changes in the law by the Supreme Court’s decision in Brown Shoe Co. v. Commissioner, supra, as discussed above.
This rationale also disposes of the first and third grounds put forth by defendant for distinguishing the instant case from Kansas City Southern R.R. v. United States, supra. That is to say that (1) the item of protective work involved in Kansas Gity Southern B.B. v. United States, supra, was not included in the itemized list of depreciable properties so that it retained its character as an expense; and (3) plaintiff agreed that the I.C.C. accounts would be binding for Federal tax purposes.
Defendant also says that the Kansas Gity Southern B.B. case does not apply because when it was decided, the I.C.C. regulations no longer required the capitalization of the items that were in issue there, but permitted them to be expenses. As already indicated, however, the accounting procedure established by a regulatory agency is not determinative of Federal tax liability. Each case must rest upon the actual facts. This was the court’s approach in the Kansas Gity Southern B.B. case, as shown by the following discussion:
Our problem is to determine whether the pole-driving work was maintenance work or capital improvement work. If, at the time of the original construction of the *306railroad, it had been possible to foresee where the water pockets would appear, and the poles had then been driven, their cost would, of course, have been a capital cost. But the water pockets develop and appear casually, and, as they appear, they present a problem of maintaining an existing completed track in a safe and economical operating condition. If in particular locations it should be found that ordinary, railroad ties were subject to extraordinary deterioration because of soil conditions or insects, and the ties were replaced with others of a type which would withstand these attacks better and longer, that replacement would, we think, be maintenance. The fact that the replacements, once made, would be good for many years, would not seem to be significant. When a building or a machine is repaired, it is not unusual that the repaired portion is better than and will outlast the parts that have not yet needed repairs. In the instant case the railroad track, after the poles were driven, was still just a railroad track, and the parts of it where the poles were driven were no more useful than the other parts which had not needed this work. [125 Ct. Cl. at 289,112 F. Supp. at 165-66.]
Although the court noted the I.C.C. accounting procedure, it did so only as support for the determination it had made on the facts and not as authority therefor. Defendant has failed to distinguish the instant case from Kansas Oity Southern B.B., supra, which, as found by the trial commissioner, is fully in point.
In sum, the construction projects were in the nature of repairs and maintenance, rather than capital improvements, and their cost is currently deductible.
YaoatioN Pay Accrual Issue
In 1955, plaintiff accrued and took as a deduction on its Federal income tax return an amount estimated to be its liability for vested vacation pay earned by its employees in 1955 and payable in 1956. Because plaintiff did not know how much it would in fact pay out in 1956 for vacation pay liability accrued in 1955, it had to estimate its liability on the basis of past experience.
Prior to 1955, plaintiff used the following method to estimate its vacation pay liability accrued during a given year: *307Taking tbe year 1952 as an example, plaintiff first determined the actual payments made to its employees in 1952 (for vacation pay liability accrued in 1951). The amount of actual payments made in 1952 was then adjusted for prospective salary changes, changes in employees’ vacation brackets, or prospective changes in employment status; and the adjusted figure became the estimated liability for 1952 (to be paid out in 1953) and formed the basis of the deduction taken on plaintiff’s 1952 tax return (filed in 1953).
Starting in 1955, plaintiff changed its method of estimating accrued vacation pay liability. Although plaintiff maintained its books of account on the basis of the calendar year, it closed its books for vacation pay liability purposes (for some unexplained reason) at the end of the first 10 months of 1955. Thus, when estimating its accrued liability for 1955, plaintiff did not use as a basis the actual payments made in 1955 but rather used the actual payments made during the first 10 months of 1955, plus an estimate of the payments made during the last 2 months of 1955. The estimate for the last 2 months of 1955 was based on payment experience in 1954, for which plaintiff determined that about 20 percent of the vacation payments were made during the last 2 months of that year. Plaintiff assumed that the same percentage would apply in 1955. In fact, the actual payments made in 1955 were $4,926,897, whereas plaintiff based its deduction on a figure of $5,233,624.70, a difference and overstatement of $306,727.70. Plaintiff arrived at the $5,233,624.70 figure by summing up $4,372.649.93, the actual payments made for the first 10 months of 1955, and $860,974.77, the amount estimated to be the payments for the last 2 months of 1955, based on 1954 experience.
Defendant says that the method used by plaintiff in 1955 was improper and that plaintiff should have used the method it had consistently employed prior to 1955. Had plaintiff done so, its deduction for 1955 would have been $306,727.70 less than the deduction actually taken. Plaintiff concedes that in 1955 it changed its method of estimating accrued liability but argues that the deduction taken was nonetheless “reasonable” because “the accrual deduction claimed for 1955 *308exceeded by 4% the amount actually paid in 1956 for vacation pay earned and vested in 1955.” Plaintiff points to a rule-of-thumb used by the Internal Eevenue Service which permits a tolerance of 4 percent between accrual estimates and actual payments.
At the outset, it is noted that plaintiff is wrong in stating that “the accrual deduction claimed for 1955 exceeded by 4% the amount actually paid in 1956.” In fact, the difference was about 5.3 percent (finding 36). In any event, plaintiff cannot prevail for more fundamental reasons, discussed below.
Section 461 of the 1954 Internal Eevenue Code and Treas. Eeg. § 1.461-1 (a) (2) set out rules for determining the proper year and amount in which deductions may be taken. The Eegulations state that taxpayers on the accrual method of accounting must deduct expenses in “the taxable year in which all the events have occurred which determine the fact of liability and the amount thereof can be determined with reasonable accuracy.” Furthermore, the Eegulations state:
Where a deduction is properly accrued on the basis of a computation made with reasonable accuracy and the exact amount is subsequently determined in a later taxable year, the difference-, if any, between such amounts shall be taken into account for the later taxable year in which such determination is made. (Emphasis added)
The requirement that estimates be “made with reasonable accuracy” means that estimates must be based on the best available information. The best available information for estimating accrued vacation pay liability in any given year is the actual payments made during that year for vacations which accrued the year before; and prior to 1955, it was on such basis that plaintiff estimated its accrued liability. Plaintiff has made no persuasive showing why it changed from this method for the year 1955. A letter of explanation to the Internal Eevenue Service dated June 8, 1962, simply stated that accruals for vacation pay liability beginning “in 1954” were made “on the books of account, making it necessary to base the accrual on information available at the time the books are closed.” There was no explanation why the books were closed for vacation pay liability purposes at the end of 10 months when for all other purposes, plaintiff’s books are *309maintained on the basis of the calendar year. The only reasonable inference is that at the time plaintiff’s tax return for 1955 was prepared (in early 1956), the actual payment figures for the last 2 months of 1955 had not been summed up or simply were ignored, though they obviously were available and clearly should have been used as plaintiff had done in earlier years.
The 4-percent tolerance rule used by the Internal Eevenue Service has no application here, even if plaintiff came within it. As correctly stated by defendant, that rule “does not justify a deduction that is four percent more than the proper amount merely because the taxpayer, for reasons not explained in the record, wishes to change its method of estimating accruals, and thereby achieve an increased deduction on the accrual basis.” In no event can the 4-percent rule be used to justify estimated accruals based on information other than the best available.
Accordingly, the deduction taken by plaintiff for accrued vacation pay liability for 1955 was overstated by $306,727.70, and defendant is entitled to the appropriate offset against any recovery by plaintiff.
MexioaN Tax Ceedit Issue
The issue is whether plaintiff can take a tax credit against Federal income taxes for taxes paid to Mexico in 1955. Plaintiff says it is entitled to a credit of $74,767.97 pursuant to §§ 901-904 of the 1954 Internal Eevenue Code. Defendant challenges plaintiff’s right to the tax credit. The Internal Eevenue Service held that plaintiff was entitled to a deduction from gross income for such tax but not a tax credit. The factual underpinning of this issue is in all material respects identical to Missouri-Illinois R.R. v. United States, 180 Ct. Cl. 1179, 381 F. 2d 1001 (1967), and Missouri Pacific R.R. v. United States, 301 F. Supp. 839 (E.D. Mo. 1967) aff'd in part, rerid in fart and remanded, 411 F. 2d 327 (8th Cir. 1969), cert. denied, 396 U.S. 1037 (1970). In both cases, it was held that the taxpayer was entitled to the claimed tax credit. Nevertheless, defendant again seeks to litigate the issue.
*310The defendant contends that no tax “was imposed on plaintiff by the Government of Mexico.” This argument is not persuasive. The facts show that on January 1, 1954, a new Mexican federal income tax law became effective. This statute was entitled “Ley del Impuesto Sobre la Renta,” which translated into English means “Income Tax Law on Rents.” Schedule VI of this statute imposed a tax on income derived from:
XIV. — Rentals, prizes, royalties and retributions of all kinds which are received as owners or holders of personal property or of rolling stock from the persons to whom they grant the use or exploitation of the same without transferring the ownership thereof.
Under Article 129th, the lessor of rolling stock is entitled to certain deductions. However, paragraph XIV provides that those taxpayers to which paragraph III of Article 6th refers, who receive income for any of the reasons set forth in paragraph XIV, shall pay the tax in Schedule I. The plaintiff is a taxpayer within paragraph III of Article 6th. Schedule I provides in Article 26th that the basis of the tax in said schedule shall be the taxable profit which is the difference between the income received by the taxpayer during a period and the deductions authorized by the law.
This law clearly imposed a tax on income received by plaintiff as per diem charges paid by Mexican railroads to the plaintiff for use of plaintiff’s railroad cars on Mexican railroad lines in Mexico in 1955 (the taxable year involved in this case).
The defendant argues further that “no tax was paid by the plaintiff to the Government of Mexico”. It is concluded that this argument is without merit.
Article 28th of Schedule I of the above statute provides that when because of the nature and characteristics of the operations carried out by the taxpayer it is not possible, by ordinary procedures, to determine with exactness the taxable income, the department of the treasury may enter into agreements for the determination of the tax base. The Mexican treasury department considered that it was difficult to determine the taxable income of the foreign companies leasing rolling stock into Mexico and to check and verify the ex*311penses and taxable income of these companies. Therefore, the Mexican Government decided to enter into agreements, under Article 28 of the income tax law, for the payment of the tax.
On December 31,1954, an agreement was executed by representatives of United States and Canadian railroads, of the Government of Mexico and of the Government-owned railways of Mexico, with respect to the payment of the tax on the income received by the United States and Canadian railroads from Mexican railroads. This agreement recited that certain United States and Canadian railroads (including the plaintiff herein) received income from sources located in the Republic of Mexico from the rental of rolling stock to the National Railways of Mexico and the Mexican Railway, which income was held to be subject to the Mexican income tax law. The agreement thereafter provided that the income received from the rental of roll ing .stock while in the Republic of Mexico, by the United States and Canadian railroads which do not operate in Mexico through agencies or branches, was subject to payment of Mexican income tax under Schedule VI of the income tax law; that the Mexican railways were authorized and obligated to retain the tax for which the United States and Canadian railroads were liable, and to pay to the Mexican treasury the amount so retained; that the amounts retained shall be equal to the tax due; that the payment of the amounts so retained, to the Mexican Government, would be in full satisfaction of the obligations of the United States and Canadian railroads for Mexican income taxes on the rolling stock rentals; that each United States and Canadian railroad was obligated to file an income tax return setting forth all income from the Mexican railroads for the rental of rolling stock and the total Mexican income tax due thereon, and that the department of the treasury and public credit of the Republic of Mexico agreed that the retention of the difference between the basic per diem rate and the Mexican per diem rate would fully satisfy the obligations of each of the United States and Canadian railroads for Mexican income taxes on rolling stock rentals. The agreement was executed by representatives of United States and Canadian railroads, the Government of Mexico and the National Railways of Mexico and Mexican Railway, relating *312to the manner of payment of the income tax obligations of the United States and Canadian railroads (including the plaintiff). A similar agreement was executed July 18, 1955, covering rentals received from some of the privately owned railroads in Mexico. These privately owned lines constitute a very small part of the railroads in Mexico, with the main railroads being agencies of the Mexican Government. All of the agreements were retroactive to January 1954.
Pertinent clauses of the agreement between the American railroads (including plaintiff), Canadian railroads and the Minister of Finance and Public Credit (on behalf of the Government of Mexico) and the General Manager of the Government-owned Mexican railroads are as follows:
CLAUSES:
1. The revenue received by American and Canadian Railroads not operating in Mexico through agencies or branches derived from rental of rolling stock while _ in the Republic of Mexico, is subject to the payment of income tax, as per Item VI.
2. The National Railways of Mexico and Mexican Railway, with authority provided for in Article 201 of the Income Tax Law, are — compelled to withhold the tax accrued by the American and Canadian-Railroads on payments derived from rental of rolling stock while in the Republic of Mexico. The withholding is to be made on the total of the tax accrued and, therefore, the American and Canadian Railroads are exempt from the obligation decreed in Article 169 of — Rulings of the Income Tax Law, which in essence is the payment of the tax through the cancellation of stamps on receipts issued to — the National Railways of Mexico and Mexican Railway for taxes derived from the rental of rolling stock while in the Republic of Mexico.
3. The National Railways of Mexico and Mexican Railway agree to pay without discounts to each of the American and Canadian Railroads listed on enclosure “A” signed by each party and included in this — Agreement, the basic Per-Diem rate in effect at the time it is due, to retain the difference between the basic Per-Diem rate and the Mexican Per-Diem rate, and to pay the amounts withheld to the Mexican Government as the total obligations of the American and Canadian Railroads, in accordance with the Income Tax Law of Mexico regarding the rental of rolling stock.
*3134. The American and Canadian Eailroads are to file with the Ministry of Finance and Public Credit during the month of January of each year a declaration of Income Tax earned the previous year, wherein all of the revenues received by the Mexican Eailroads through the rental of rolling stock are to be mentioned and the total of the tax due on same.
5. It is agreed that the declarations of Income Tax filed individually by the American and Canadian Eail-roads listed on enclosure “A”, will include a statement of the total of the taxes mentioned as earned and withheld by the National Eailways of Mexico and Mexican Eailway, in accordance with Articles 201 and 28 of the Income Tax Law of Mexico. It is also agreed that the revenue for the rental of rolling stock mentioned in said declarations will be computed at the basis of the Mexican Per-Diem rate.
6. The Ministry of Finance and Public Credit of the Eepublic of Mexico agrees that the withholding of the difference between the Basic Per-Diem rate and the Mexican Per-Diem rate mentioned in — Clause 3, will cover completely all obligations of each of the American and Canadian Eailroads listed on enclosure “A” of the Income Tax Law of Mexico on rental of rolling stock. The Ministry agrees, in addition, to furnish to each of the American and Canadian Eailroads through a receipt or other form, a statement certifying said obligations.
7. The Ministry of Finance and Public credit will grant the National Eailways of Mexico and the Mexican Eailway for as long as — this obligation of payment of tax on rental of rolling stock exists, a subsidy equal to the amount of taxes withheld.
The tax withheld by the Mexican railroads which were parties to the tax agreements was computed, in accordance with such agreements, as the difference between the basic.per diem rate and the Mexican per diem rate. In the case of railroads not parties to the agreements, the tax was 10 percent of the gross rental.
In the absence of any tax agreements, specific provisions of the Mexican income tax statute provided for the withholding of 10 percent of the gross rentals. In such instances, while the rental income would also be taxed under Schedule I of the Mexican income tax law, deductions would be permitted (such as salaries, overhead, depreciation, and professional services).
*314The Mexican railroads which were not parties to the tax agreements were obligated to pay the amount of the tax withheld from the gross rentals earned, directly to the Mexican treasury department.
The Mexican income tax law does not provide for a subsidy to the Mexican railroads, but this law does hold those obligated to withhold amounts as tax, liable at all times jointly with the taxpayer for payment of the tax.
The Association 'of American Railroads (AAR) sets per diem rental rates for freight cars which are operated off the line owning the car and on other railroad lines. The freight car per diem rate for 1955 was $2.40, which was the amount that one United States or Canadian railroad paid another United States or Canadian railroad for each calendar day a freight car owned by the second railroad remained on the first railroad’s lines. The per diem rate paid by the Mexican railroads in 1955 for use of American freight cars was $3.40, of which $1.00 was withheld by the Mexican railroads as payment for tax due the Mexican Government and $2.40 was paid to the American railroad whose cars were used. For 1955, plaintiff claims payment of taxes to Mexico of $85,-656.87, for which it claims a tax credit of $74,767.97 under §§ 901-904 of the Internal Revenue Code for 1954.
It is concluded that the tax of $85,656.87 imposed by the Mexican income tax law upon the freight rentals earned by the plaintiff while its railroad cars were in Mexico in 1955, and which tax was withheld, by the Mexican railroads for the Mexican treasury, was a payment by the plaintiff of such taxes to the Mexican Government. Furthermore, it was a foreign income tax within the meaning of the Internal Revenue Code of the United States. The fact that the payment was made through a system of withholding and debits and credits in accordance with the accounting and fiscal policies and procedures of the Government of Mexico does not detract from its Status as a payment of the taxes due by the plaintiff. The system is much like that used in the United States where our government withholds income tax from the salaries of its employees. Such amounts are later credited to the income tax due by the taxpayer through a debit and credit system between the Internal Revenue Service and the *315Treasury. The taxpayer never sees the withheld, funds, never has them in his possession and never personally pays the actual dollars to anyone. Yet, he is considered as having paid the tax to the extent of the withheld funds. This is simply the modem and logical way of handling such payments that has been in vogue in this country for many years. Under these circumstances, it is surprising that the Internal Beve-nue Service has questioned the withholding procedure by which the plaintiff paid its income taxes to the Mexican Government in 1955. The procedure established by the Mexican statute and the above agreement made unnecessary the conversion of Mexican pesos into dollars and the payment of same to plaintiff as rent on the cars and the subsequent conversion of the dollars to pesos and their payment by plaintiff to the Mexican treasury. Even if such payments had been made entirely in pesos in Mexico, such a procedure would have been more complicated than that which was used. In any event, the Government of Mexico could adopt any procedure it desired, and it did so. We have no right to complain.
The fact that the Mexican Government later granted subsidies to the Mexican railroads equal to the plaintiff’s tax payments does not change the fact that the plaintiff made the payments. This is wholly an internal Mexican affair that is of no concern to us. As a matter of fact, most of the railroads were owned and operated by the Mexican Government. Under these circumstances, the subsidies were no doubt accomplished by debits and credits within the government. But here again, we are not concerned with what Mexico did in this regard.
The plaintiff is entitled to the tax credit of $74,167.97 on the $85,656.87 foreign income tax it paid to Mexico in 1955.16 See Missouri Pacific R.R. v. United States, 183 Ct. Cl. 168, 392 F. 2d 592 (1968).
The dissenting opinion of Davis, Judge, in which Laramorh, Senior Judge, and Ddrfee, Senior Judge, join in part, and in which Nichols, Judge, joins, and the dissenting opinion of Nichols, Judge, follow the opinion of the trial commissioner which has been adopted by the court.
Defendant says in its brief that plaintiff “is entitled to the recovery claimed under * * * [the excess salvage} issue.” Thus, no findings on, or discussion of, that issue are required.
For excellent and detailed analyses of Detroit Mdiaon and Brown Shoe, as well as related cases, see Freeman and Speiller, Tax Consequences of Subsidies to Induce Business Location, 9 Tax L.R. 255 (1953-54) ; also, Note Tax Consequences of Non-Shareholder Contributions to Corporate Capital, 66 Yale U.J. 1085 (1956-57).
It has been suggested that Brown Shoe “In effect” overruled Detroit Edison. See the Freeman and SpelUer article cited in n. 3, at 262. In any event, Detroit Edison does not apply here.
Similarly, plaintiff’s failure to include the donated property here at issue in the schedules of plaintiff’s property for which straight line depreciation was requested and in the later revised schedules may reflect no more than plaintiff’s apparent opinion at those times that such property was not depreciable.
In passing, it Is noted that § 362(e) (1) of the 1954 Code effectively overruled the Brown Shoe case with respect to property contributed to a corporation by a nonshareholder after June 22, 1954. It provides that the basis of such property shall be zero. Since the property here in question was all contributed to plaintiff prior to June 22, 1954, § 362(c) (1) does not apply.
H.E. Rep. No. 3124, 81st Cong., 2d Sess. (1950-2 Cum. Boil. 580, 586). See also S. Rep. No. 23T5, 81st Cong., 2d Sess. (1950-2 Cum. Bull. 483, 560).
The Internal Revenue Service accepts replacement accounting, as well as retirement and betterment accounting, as proper methods for determining depreciation under § 167 of the 1954 Internal Revenue Code. However, replacement accounting has been criticized and was the subject of a Congressional study in 1957. See Report op Committee on Government Operations, H.R. Rep. No. 1167, 85th Cong., 1st «ess. (1957).
I.8., rail weighing 80 pounds per linear yard.
Bolted Joints for 112-pound rail cost $9.81 each ; for 129-pound rail, $10.39.
Plaintiff; installs reusable rail for some of Its Industrial customers as side track; and in 1955, it charged such customers about 60 percent of the price of comparable niew rail. While this tends to establish, a fair market price on the traditional basis of a willing buyer and willing seller, it is not a truly reliable price since it includes installation cost and is subject to rebate depending on the later volume of use to which the track is subjected. Defendant, therefore, bases its fair market value of reusable rail on the formula, above noted, rather than the 60 percent figure.
Plaintiff’s reply brief says that “plaintiff recognizes that its previously assigned value of * * * [$22.32 per net ton] was arbitrary * * *” and that "plaintiff is willing to accept a value * * * of the lower of actual cost or fair market value * *
The railroads were not uniform in the values they assigned to reusable rail in 1955. The record shows that values ranged from §63.34 per net ton down to values lower than §22.32 per net ton. (Finding 27.)
At trial, a former chief of the section of accounting of the Interstate Commerce Commission testified that the interpretation and application of § 10.01-7 (d) is governed by the “major principle” that in no case shall “salvage value exceed the original cost.” Despite the testimony, nowhere in the Commission’s uniform System of Account® for Railroad Companies does there purport to be a rule stating such “major principle” and the witness could point to none. No doubt, such “major principle” has to do with the fact that the Interstate Commerce Commission, a rate-making regulatory agency, is obliged to see that the railroads do not overstate the book value of their assets as a whole. Otherwise, unjustified rate increases might be urged. Nevertheless, § 10.01-7 (d) is clear on its face, its meaning is plain and unambiguous, and it remains a rule currently in force (¶ 18 under “Regulations Prescribed” in the 1968 edition of the Uniform System of Accounts for Railroad Companies).
The Kansas City Southern ease is followed in Rey. Rol. 54-356, 1954-2 Cum. Bull. 82, and Rey. Rnl. 58-278, 1958-1 Cum. Bull. 92.
The defendant has conceded that, If the tax was Imposed and paid, the plaintiff is entitled to such credit under Sections 901 — 904 of the Internal Revenue Code of 1934.