These four cases were consolidated in the Court of Appeals. The facts of one will define the issue present in all.

The taxpayer, Collins, in 1929 purchased 300 shares of stock of the National City Bank of New York which carried certain beneficial interests in stock of the National City Company. The latter company was the seller and the transaction occurred in Minnesota. In 1930 Collins sold 100 shares, sustaining a deductible loss of $41,600.80, which was claimed on his return for that year and allowed. In 1931 he sold another 100 shares, sustaining a deductible loss of $28,163.78, which was claimed in his return and allowed. The remaining 100 shares he retained. He regarded the purchases and sales as closed and completed transactions.

In 1936 Collins learned that the stock had not been registered in compliance with the Minnesota Blue Sky Laws and learned of facts indicating that he had been induced to purchase by fraudulent representations. He filed suit against the seller alleging fraud and failure to register. He asked recission of the entire transaction and offered to return the proceeds of the stock, or an equivalent number of shares plus such interest and dividends as he had received. In 1939 the suit was settled, on a basis which gave him a net recovery of $45,150.63, of which $23,296.45 was allocable to the stock sold in 1930 and $6,454.18 allocable to that sold in 1931. In his return for 1939 he did not report as income any part of the recovery. Throughout that year adjustment of his 1930 and 1931 tax liability was barred by the statute of limitations.

The Commissioner adjusted Collins' 1939 gross income by adding as ordinary gain the recovery attributable to the shares sold, but not that portion of it attributable to the shares unsold. The recovery upon the shares sold was not, however, sufficient to make good the taxpayer's original investment in them. And if the amounts recovered had been added to the proceeds received in 1930 and 1931 they would not have altered Collins' income tax liability for those years, for even if the entire deductions claimed on account of these losses had been disallowed, the returns would still have shown net losses.

Collins sought a redetermination by the Board of Tax Appeals, now the Tax Court. He contended that the recovery of 1939 was in the nature of a return of capital from which he realized no gain and no income either actually or constructively, and that he had received no tax benefit from the loss deductions. In the alternative he argued that if the recovery could be called income at all it was taxable as capital gain. The Commissioner insisted that the entire recovery was taxable as ordinary gain and that it was immaterial whether the taxpayer had obtained any tax benefits from the loss deduction reported in prior years. The Tax Court sustained the taxpayer's contention that he had realized no taxable gain from the recovery.1

The Court of Appeals (133 F.2d 732, 735) concluded that the 'tax benefit theory' applied by the Tax Court 'seems to be an injunction into the law of an equitable principle, found neither in the statutes nor in the regulations.' Because the Tax Court's reasoning was not embodied in any statutory precept, the court held that the Tax Court was not authorized to resort to it in determining whether the recovery should be treated as income or return of capital. It held as matter of law that the recoveries were neither return of capital nor capital gain, but were ordinary income in the year received.2 Questions important to tax administration were involved, conflict was said to exist, and we granted certiorari.3

It is contended that the applicable statutes and regulations properly interpreted forbid the method of calculation followed by the Tax Court. If this were true, the Tax Court's decision would not be 'in accordance with law' and the Court would be empowered to modify or reverse it.4 Whether it is true is a clear-cut question of law and is for decision by the courts.

The court below thought that the Tax Court's decision 'evaded or ignored' the statute of limitation, the provision of the Regulations that 'expenses, liabilities, or deficit of one year cannot be used to reduce the income of a subsequent year,'5 and the principle that recognition of a capital loss presupposes some event of 'realization' which closes the transaction for good. We do not agree. The Tax Court has not attempted to revise liability for earlier years closed by the statute of limitation, nor used any expense, liability, or deficit of a prior year to reduce the income of a subsequent year. It went to prior years only to determine the nature of the recovery, whether return of capital or income. Nor has the Tax Court reopened any closed transaction; it was compelled to determine the very question whether such a recognition of loss had in fact taken place in the prior year as would necessitate calling the recovery in the taxable year income rather than return of capital.

The 1928 Act provides that 'the Board in redetermining a deficiency in respect of any taxable year shall consider such facts with relation to the taxes for other taxable years as may be necessary correctly to redetermine the amount of such deficiency * * *.'6 The Tax Court's inquiry as to past years was authorized if 'necessary correctly to redetermine' the deficiency. The Tax Court thought in this case that it was necessary; the Court of Appeals apparently thought it was not. This precipitates a question not raised by either counsel as to whether the court is empowered to revise the Tax Court's decision as 'not in accordance with law' because of such a difference of opinion.

With the 1926 Revenue Act, Congress promulgated, and at all times since has maintained, a limitation on the power of courts to review Board of Tax Appeals (now the Tax Court) determinations. '* * * such courts shall have power to affirm or, if the decision of the Board is not in accordance with law, to modify or to reverse the decision of the Board * * *.'7 However, even a casual survey of decisions in tax cases, now over 5,000 in number, will demonstrate that courts including this Court have not paid the scrupulous deference to the tax laws' admonitions of finality which they have to similar provisions in statutes relating to other tribunals.8 After thirty years of income tax history the volume of tax litigation necessary merely for statutory interpretation would seem due to subside. That it shows no sign of diminution suggests that many decisions have no value as precedents because they determine only fact questions peculiar to particular cases. Of course frequent amendment of the statute causes continuing uncertainty and litigation, but all too often amendments are themselves made necessary by court decisions. Increase of potential tax litigation due to more taxpayers and higher rates lends new importance to observance of statutory limitations on review of tax decisions. No other branch of the law touches human activities at so many points. It can never be made simple, but we can try to avoid making it needlessly complex.

It is more difficult to maintain sharp separation of court and administrative functions in tax than in other fields. One reason is that tax cases reach circuit courts of appeals from different sources and do not always call for observance of any administrative sphere of decision. Questions which the Tax Court considers at the instance of one taxpayer may be considered by many district courts at the instance of others.

The Tucker Act authorizes district courts, sitting without jury as courts of claims, to hear suits for recovery of taxes alleged to have been 'erroneously or illegally assessed or collected.'9 District courts also entertain common law actions against collectors to recover taxes erroneously demanded and paid under protest. Trial may be by jury, but waiver of jury is authorized10 and in tax cases jury frequently is waived. In such cases the findings of the court may be either special or general. The scope of review on appeal may be affected by the nature of the proceeding, the kind of findings, and whether the jury was waived under a particular statutory authorization or independently of it.11 The multiplicity and complexity of rules is such that often it is easier to review the whole case on the merits than to decide what part of it is reviewable and under what rule. The reports contain many cases in which the question is passed over without mention.

Another reason why courts have deferred less to the Tax Court than to other administrative tribunals is the manner in which the Tax Court finality was introduced into the law.

The courts have rather strictly observed limitations on their reviewing powers where the limitation came into existence simultaneously with their duty to review administrative action in new fields of regulation. But this was not the history of the tax law. Our modern income tax experience began with the Revenue Act of 1913. The World War soon brought high rates. The law was an innovation, its constitutional aspects were still being debated, interpretation was just beginning, and administrators were inexperienced. The Act provided no administrative review of the Commissioner's determinations. It did not alter the procedure followed under the Civil War income tax by which an aggrieved taxpayer could pay under protest and then sue the Collector to test the correctness of the tax.12 The courts by force of this situation entertained all manner of tax questions, and precedents rapidly established a pattern of judicial thought and action whereby the assessments of income tax were reviewed without much restraint or limitation. Only after that practice became established did administrative review make its appearance in tax matters.

Administrative machinery to give consideration to the taxpayer's contentions existed in the Bureau of Internal Revenue from about 1918 but it was subordinate to the Commissioner.13 In 1923, the situation was brought to the attention of Congress by the Secretary of the Treasury, who proposed creation of a Board of Tax Appeals, within the Treasury Department, whose decision was to conclude Government and taxpayer on the question of assessment and leave the taxpayer to pay the tax and then test its validity by suit against the collector.14 Congress responded by creating the Board of Tax Appeals as 'an independent agency in the executive branch of the Government.'15 The Board was to give hearings and notice thereof and 'make a report in writing of its findings of fact and decision in each case.'16 But Congress dealt cautiously with finality for the Board's conclusions, going only so far as to provide that in later proceedings the findings should be 'prima facie evidence of the facts therein stated.'17 So the Board's decisions first came before the courts under a statute which left them free to go into both fact and law questions. Two years later Congress reviewed and commended the work of the new Board,18 increased salaries and lengthened the tenure of its members,19 provided for a direct appeal from the Board's decisions to the circuit courts of appeals or the Court of Appeals of the District of Columbia,20 and enacted the present provision limiting review to questions of law.21

But this restriction upon judicial review of the Board's decisions came only after thirteen years of income tax experience had established a contrary habit. Precedents had accumulated in which courts had laid down many rules of taxation not based on statute but upon their ideas of right accounting or tax practice. It was difficult to shift to a new basis. This Court applied the limitation, but with less emphasis and less forceful resolution of borderline cases in favor of administrative finality than it has employed in reference to other administrative determinations.22

That neglect of the congressional instruction is a fortuitous consequence of this evolution of the Tax Court rather than a deliberate or purposeful judicial policy is the more evident when we consider that every reason ever advanced in support of administrative finality applies to the Tax Court.

The court is independent, and its neutrality is not clouded by prosecuting duties. Its procedures assure fair hearings. Its deliberations are evidenced by careful opinions. All guides to judgment available to judges are habitually consulted and respected. It has established a tradition of freedom from bias and pressures.23 It deals with a subject that is highly specialized and so complex as to be the despair of judges. It is relatively better staffed for its task than is the judiciary.24 Its members not infrequently bring to their task long legislative or administrative experience in their subject. The volume of tax matters flowing through the Tax Court keeps its members abreast of changing statutes, regulations, and Bureau practices, informed as to the background of controversies and aware of the impact of their decisions on both Treasury and taxpayer. Individual cases are disposed of wholly on records publicly made, in adversary proceedings, and the court has no responsibility for previous handling. Tested by every theoretical and practical reason for administrative finality, no administrative decisions are entitled to higher credit in the courts. Consideration of uniform and expeditious tax administrations require that they be given all credit to which they are entitled under the law.

Tax Court decisions are characterized by substantial uniformity. Appeals fan out into courts of appeal of ten circuits and the District of Columbia. This diversification of appellate authority inevitably produces conflict of decision, even if review is limited to questions of law. But conflicts are multiplied by treating as questions of law what really are disputes over proper accounting. The mere number of such questions and the mass of decisions they call forth become a menace to the certainty and good administration of the law.25

To achieve uniformity by resolving such conflicts in the Supreme Court is at best slow, expensive, and unsatisfactory. Students of federal taxation agree that the tax system suffers from delay in getting the final word in judicial review, from retroactivity of the decision when it is obtained, and from the lack of a roundly tax-informed viewpoint of judges.26

Perhaps the chief difficulty in consistent and uniform compliance with the congressional limitation upon court review lies in the want of a certain standard for distinguishing 'questions of law' from 'questions of fact.' This is the test Congress has directed, but its difficulties in practice are well known and have been subject of frequent comment. Its difficulty is reflected in our labeling some questions as 'mixed questions of law and fact'27 and in a great number of opinions distinguishing 'ultimate facts' from evidentiary facts.28

It is difficult to lay down rules as to what should or should not be reviewed in tax cases except in terms so general that their effectiveness in a particular case will depend largely upon the attitude with which the case is approached. However, all that we have said of the finality of administrative determination in other fields is applicable to determinations of the Tax Court. Its decision, of course, must have 'warrant in the record' and a reasonable basis in the law. But 'the judicial function is exhausted when there is found to be a rational basis for the conclusions approved by the administrative body.' Rochester Telephone Corp. v. United States, 307 U.S. 125, 146, 59 S.Ct. 754, 764, 83 L.Ed. 1147; Swayne & Hoyt v. United States, 300 U.S. 297, 304, 57 S.Ct. 478, 481, 81 L.Ed. 659; Mississippi Valley Barge Line Co. v. United States, 292 U.S. 282, 286, 287, 54 S.Ct. 692, 693, 694, 78 L.Ed. 1260; Gray v. Powell, 314 U.S. 402, 412, 62 S.Ct. 326, 332, 86 L.Ed. 301; Helvering v. Clifford, 309 U.S. 331, 336, 60 S.Ct. 554, 557, 84 L.Ed. 788; United States v. Louisville & N.R. Co., 235 U.S. 314, 320, 35 S.Ct. 113, 114, 59 L.Ed. 245; Wilmington Trust Co. v. Helvering, 316 U.S. 164, 168, 62 S.Ct. 984, 986, 86 L.Ed. 1352.

Congress has invested the Tax Court with primary authority for redetermining deficiencies, which constitutes the greater part of tax litigation. This requires it to consider both law and facts. Whatever latitude exists in resolving questions such as those of proper accounting, treating a series of transactions as one for tax purposes, or treating apparently separate ones as single in their tax consequences, exists in the Tax Court and not in the regular courts; when the court cannot separate the elements of a decision so as to identify a clear-cut mistake of law, the decision of the Tax Court must stand. In view of the division of functions between the Tax Court and reviewing courts it is of course the duty of the Tax Court to distinguish with clarity between what it finds as fact and what conclusion it reaches on the law. In deciding law questions courts may properly attach weight to the decision of points of law by an administrative body having special competence to deal with the subject matter. The Tax Court is informed by experience and kept current with tax evolution and needs by the volume and variety of its work. While its decisions may not be binding precedents for courts dealing with similar problems, uniform administration would be promoted by conforming to them where possible.

The Government says that 'the principal question in this case turns on the application of the settled principle that the single year is the unit of taxation.' But the Tax Court was aware of this principle and in no way denied it. Whether an apparently integrated transaction shall be broken up into several separate steps and whether what apparently are several steps shall be synthesized into one whole transaction is frequently a necessary determination in deciding tax consequences.29 Where no statute or regulation controls, the Tax Court's selection of the course to follow is no more reviewable than any other question of fact. Of course we are not here considering the scope of review where constitutional questions are involved. The Tax Court analyzed the basis of the litigation which produced the recovery in this case and the obvious fact that 'regarding the series of transactions as a whole, it is apparent that no gain was actually realized.' It found that the taxpayer had realized no tax benefits from reporting the transaction in separate years. It said the question under these circumstances was whether the amount the taxpayer recovered in 1939 'constitutes taxable income, even though he realized no economic gain.' It concluded that the item should be treated as a return of capital rather than as taxable income. There is no statute law to the contrary, and the administrative rulings in effect at the time tended to support the conclusion.30 It is true that the Board in a well considered opinion reviewed a number of court holdings, but it did so for the purpose of showing that they did not fetter its freedom to reach the decision it thought sound. With this we agree.

Viewing the problem from a different aspect, the Government urges in this Court that although the recovery is capital return, it is taxable in its entirety because taxpayer's basis for the property in question is zero. The argument relies upon § 113(b)(1)(A) of the Internal Revenue Code, 26 U.S.C.A. Int.Rev.Code, § 113(b)(1)(A), which provides for adjusting the basis of property for 'expenditures, receipts, losses, or other items, properly chargeable to capital account.' This provision, it is said, requires that the right to a deduction for a capital loss be treated as a return of capital. Consequently, by deducting in 1930 and 1931 the entire difference between the cost of his stock and the proceeds of the sales, taxpayer reduced his basis to zero. But the statute contains no such fixed rule as the Government would have us read into it. It does not specify the circumstances or manner in which adjustments of the basis are to be made, but merely provides that 'Proper adjustment * * * shall in all cases be made' for the items named if 'properly chargeable to capital account.' What, in the circumstances of this case, was a proper adjustment of the basis was thus purely an accounting problem and therefore a question of fact for the Tax Court to determine. Evidently the Tax Court thought that the previous deductions were not altogether 'properly chargeable to capital account' and that to treat them as an entire recoupment of the value of taxpayer's stock would not have been a 'proper adjustment.' We think there was substantial evidence to support such a conclusion.

The Government relies upon Burnet v. Sanford & Brooks Co., 282 U.S. 359, 51 S.Ct. 150, 75 L.Ed. 383, for the proposition that losses of one year may not offset receipts of another year. But the case suggested its own distinction: 'While (the money received) equalled, and in a loose sense was a return of, expenditures made in performing the contract, still, as the Board of Tax Appeals found, the expenditures were made in defraying the expenses * * *. They were not capital investments, the cost of which, if converted, must first be restored from the proceeds before there is a capital gain taxable as income.' 282 U.S. at pages 363, 364, 51 S.Ct. at page 151, 75 L.Ed. 383. It is also worth noting that the Court affirmed the Board's decision, which had been upset by the circuit court of appeals, and answered, in part, the contention of the circuit court that certain regulations were applicable by saying, '* * * nor on this record do any facts appear tending to support the burden, resting on the taxpayer, of establishing that the Commissioner erred in failing to apply them.' 282 U.S. at pages 366, 367, 51 S.Ct. at page 152, 75 L.Ed. 383.

It is argued on behalf of the Commissioner that the Court should overrule the Board by applying to this question rules of law laid down in decisions on the analogous problem raised by recovery of bad debts charged off without tax benefit in prior years. The court below accepted the argument. However, instead of affording a reason for overruling the Tax Court, the history of the bad debt recovery question illustrates the mischief of overruling the Tax Court in matters of tax accounting. Courts were persuaded to rule as matter of law that bad debt recoveries constitute taxable income, regardless of tax benefit from the charge-off.31 The Tax Court had first made a similar holding,32 but had come to hold to the contrary.33 Substitution of the courts' rule for that of the Tax Court led to such hardships and inequities that the Treasury appealed to Congress to extend relief.34 It did so.35 The Government now argues that by extending legislative relief in bad debt cases Congress recognized that in the absence of specific exemption recoveries are taxable as income. We do not find that significance in the amendment. A specific statutory exception was necessary in bad debt cases only because the courts reversed the Tax Court and established as matter of law a 'theoretically proper' rule which distorted the taxpayer's income. Congress would hardly expect the courts to repeat the same error in another class of cases, as we would do were we to affirm in this case.36

The Government also suggests that 'If the tax benefit rule were judicially adopted the question would then arise of how it should be determined,' and the difficulties of determining tax benefits, it says, create 'an objection in itself to an attempt to adopt such a rule by judicial action.' We are not adopting any rule of tax benefits. We only hold that no statute or regulation having the force of one and no principle of law compels the Tax Court to find taxable income in a transaction where as matter of fact it found no economic gain and no use of the transaction to gain tax benefit. The error of the court below consisted of treating as a rule of law what we think is only a question of proper tax accounting.

There is some difference in the facts of these cases. In two of them the Tax Court sustained deficiencies because it found that the deductions in prior years had offset gross income for those years and therefore concluded that the recoveries must to that extent be treated as taxable gain.37 The taxpayers object that this conclusion disregards certain exemptions and credits which would have been available to offset the increased gross income in the prior years, so that the deductions resulted in no tax savings. In determining whether the recoveries were taxable gain, however, the Tax Court was free to decide for itself what significance it would attach to the previous reduction of taxable income as contrasted with reduction of tax. The statute gives no inkling as to the correctness or incorrectness of the Tax Court's view, and we can find no compelling reason to substitute our judgment. In No. 47 the decision of the Tax Court was upheld by the court below, and in that case the judgment is affirmed. In Nos. 44, 45, and 46, the Court of Appeals reversed the Tax Court, and for the reasons stated its judgments in those cases are reversed.

Reversed.