concurring in part and dissenting in part as to the opinion of The Chief Justice.

Although I join in the specific overruling of Commissioner v. Wilcox, 327 U. S. 404 (1946), in The Chief Justice’s opinion, I would affirm this conviction on either of two grounds. I believe that the Court not only devitalized Wilcox, by limiting it to its facts in Rutkin v. United States, 343 U. S. 130 (1952), but that in effect the Court overruled that case sub silentio in Commissioner v. Glenshaw Glass Co., 348 U. S. 426 (1955). Even if that not be true, in my view the proof shows conclusively that petitioner, in willfully failing to correctly report his income, placed no bona fide reliance on Wilcox.

whom

The starting point of any inquiry as to what constitutes taxable income must be the Sixteenth Amendment, which grants Congress the power “to lay and collect taxes on incomes, from whatever source derived . . . .” It has long been settled that Congress’ broad statutory definitions of taxable income were intended' “to use the full measure of [the Sixteenth Amendment’s] taxing power.” Helvering v. Clifford, 309 U. S. 331, 334; Douglas v. Willcuts, 296 U. S. 1, 9. Equally well settled is the principle that the Sixteenth Amendment “is to be taken as written and is not to be extended beyond the meaning clearly indicated by the language used.” Edwards v. Cuba R. Co., 268 U. S. 628, 631.1 The language of the Sixteenth Amendment, as well as our prior controlling decisions, *249compels me to conclude that the question now before us— whether an embezzler receives taxable income at the time of his unlawful taking — must be answered negatively. Since the prevailing opinion reaches an opposite conclusion, I must respectfully dissent from that holding, although I concur in the Court’s judgment reversing petitioner’s conviction. I am convinced that Commissioner v. Wilcox, 327 U. S. 404, which is today overruled, was correctly decided on the basis of every controlling principle used in defining taxable income since the Sixteenth Amendment’s adoption.

The Chief Justice’s opinion, although it correctly recites Wilcox’s holding that “embezzled money does not constitute taxable income to the embezzler in the year of the embezzlement” (emphasis added), fails to explain or to answer the true basis of that holding. Wilcox did not hold that embezzled funds may never constitute taxable income to the embezzler. To the contrary, it expressly recognized that an embezzler may realize a taxable gain to the full extent of the amount taken, if and when it ever becomes his. The applicable test of taxable income, i. e., the “presence of a claim of right to the alleged gain,” of which Wilcox spoke, was but a correlative statement of the factor upon which the decision placed its whole emphasis throughout, namely, the “absence of a definite, unconditional obligation to repay or return [the money].” 327 U. S., at 408. In holding that this test was not met at the time of the embezzlement, the Wilcox opinion repeatedly stressed that the embezzler had no “bona fide legal or equitable claim” to the embezzled funds, ibid.; that the victim never “condoned or forgave the taking of the money and still holds him liable to restore it,” id., at 406; and that the “debtor-creditor relationship was definite and unconditional.” Id., at 409. These statements all express the same basic fact — the fact which is emphasized most strongly in the opinion’s conclusion explaining *250why the embezzler had not yet received taxable income: “Sanctioning a tax under the circumstances before us would serve only to give the United States an unjustified preference as to part of the money which rightfully and completely belongs to the taxpayer’s employer.” Id., at 410. (Emphasis added.)

However, Wilcox plainly stated that “if the unconditional indebtedness is cancelled or retired, taxable income may adhere, under certain circumstances, to the taxpayer.” 327 U. S., at 408. More specifically,.it recognized that had the embezzler’s victim “condoned or forgiven any part of the [indebtedness], the [embezzler] might have been subject to tax liability to that extent,” id., at 410, i. e., in the tax year of such forgiveness.

These statements reflect an understanding of, and regard for, substantive tax law concepts solidly entrenched in our prior decisions. Since our landmark case of United States v. Kirby Lumber Co., 284 U. S. 1, it has been settled that, upon a discharge of indebtedness by an event other than full repayment, the debtor realizes a taxable gain in the year of discharge to the extent of the indebtedness thus extinguished. Such gains are commonly referred to as ones realized through “bargain cancellations” of indebtedness, and it was in this area, and indeed, in Kirby Lumber Co. itself, that the “accession” theory or “economic gain” concept of taxable income, upon which The Chief Justice’s opinion today mistakenly, relies, found its genesis. In that case, the taxpayer, a corporation, had reduced a portion of its debt, with a corresponding gain in assets, by purchasing its bonds in the open market at considerably less than their issue price. Mr. Justice Holmes, who wrote the Court’s opinion, found it unnecessary to state the elementary principle that, so long as the bonds remained a fully enforceable debt obligation of the taxpayer, there could be no taxable gain. However, when the taxpayer retired the debt by purchasing *251the bonds for less than their face, value, it “made a clear [taxable] gain” and “realized within the year an accession to income” in the amount of its bargain. 284 U. S., at 3.

This doctrine has since been reaffirmed and strengthened by us, see, e. g., Helvering v. American Chicle Co., 291 U. S. 426; Commissioner v. Jacobson, 336 U. S. 28, and by the lower federal courts in numerous decisions involving a variety of “bargain cancellations” of indebtedness, as by a creditor’s release condoning or forgiving the indebtedness in whole or in part,2 or by the running of a Statute of Limitations barring the legal enforceability of the obligation.3 In none of these cases has it been suggested that a taxable gain might be realized by the debtor at any time prior to the effective date of discharge, and as Wilcox recognized, there is no rational basis on which to justify such a rule where the debt arises through embezzlement.

An embezzler, like a common thief, acquires not a semblance of right, title, or interest in his plunder, and whether he spends it or not, he is indebted to his victim in the full amount taken as surely as if he had left a signed promissory note at the scene of the crime. Of no consequence from any standpoint is the absence of such formalities as (in the words of the prevailing opinion) “the consensual recognition, express or implied, of an obligation to repay.” The law readily implies whatever “consensual recognition” is needed for the rightful owner to assert an immediately ripe and enforceable obligation of *252repayment against the wrongful taker. These principles are not “attenuated subtleties” but are among the clearest and most easily applied rules of our law. They exist to protect the rights of the innocent victim, and we should accord them full recognition and respect.

The fact that an embezzler’s victim may have less chance of success than other creditors in seeking repayment from his debtor is not a valid reason for us further to diminish his prospects by adopting a rule that would allow the Commissioner of Internal Revenue to assert and enforce a prior federal tax lien against that which “rightfully and completely belongs” to the victim. Commissioner v. Wilcox, supra, at 410. The Chief Justice’s opinion quite understandably expresses much concern for “honest taxpayers,” but it attempts neither to deny nor justify the manifest injury that its holding will inflict on those honest taxpayers, victimized by embezzlers, who will find their claims for recovery subordinated to federal tax liens. Statutory provisions, by which we are bound, clearly and unequivocally accord priority to federal tax liens over the claims of others, including “judgment creditors.”4

*253However, if it later happens that the debtor-creditor relationship between the embezzler and his victim is discharged by something other than full repayment, such as-by the running of a Statute of Limitations against the victim’s claim, or by a release given for less than the full amount owed, the embezzler at that time, but not before, will have made a clear taxable gain and realized “an accession to income” which he will be required under full penalty of the law to report in his federal income tax return for that year. No honest taxpayer could be harmed by this rule.

The inherent soundness of this rule could not be more clearly demonstrated than as applied to the facts of the case before us. Petitioner, a labor union official, concededly embezzled sums totaling more than $738,000 from the union’s funds, over a period extending from 1951 to 1954. When the shortages were discovered in 1956, the union at once filed civil actions against petitioner to compel repayment. For reasons which need not be detailed here, petitioner effected a settlement agreement with the union on July 30, 1958, whereby, in exchange for releases fully discharging his indebtedness, he repaid to the union the sum of $13,568.50. Accordingly, at least so far as the present record discloses, petitioner clearly realized a taxable gain in the year the releases were executed, to the extent of the difference between the amount taken and the sum restored: However, the Government brought the present action against him, not for his failure to report this gain in his 1958 return, but for his failure to report that he had incurred “income” from — actually indebtedness to — the union in each of the years 1951 through 1954. It is true that the Government brought a criminal evasion prosecution rather than a civil deficiency proceeding against petitioner, but this can in no way alter the substantive tax law rules which alone are determinative of liability in either case.

*254There can be no doubt that until the releases were executed in 1958, petitioner and the union stood in an absolute and unconditional debtor-creditor relationship, and, under all of our relevant decisions, no taxable event could have occurred until the indebtedness was discharged for less than full repayment. Application of the normal rule in such cases will not hinder the efficient and orderly administration of the tax laws, any more than it does in other situations involving “bargain cancellations” of indebtedness. More importantly, it will enhance the creditor’s position by assuring that prior federal tax liens will not attach to the subject of the debt when he seeks to recover it. •

Notwithstanding all of this, The Chief Justice’s opinion concludes that there is no difference between embezzled funds and “gains” from other “illegal sources,” and it points to the fact that Congress, in its 1916 revision of the Income Tax Act, omitted the word “lawful” in describing businesses whose income was to be taxed. The opinion then cites United States v. Sullivan, 274 U. S. 259, in which it was held that, under the revised statute, gains from illicit traffic in liquor must be reported in gross income, since there is no “reason why the fact that a business' is unlawful should exempt it from paying the taxes that if lawful it would have to pay.” Id., at 263. (Emphasis added.) That theory has been the primary basis for taxing “unlawful gains of many kinds” which the prevailing opinion today recites, such as black market profits, gambling proceeds, money derived from the sale of unlawful insurance policies, etc.5 For, even if lawful, the gains from such activities would clearly *255not be exempted from taxation. However, as applied to embezzled funds, the holding in Sullivan contradicts, rather than supports, the Court’s conclusion today. Obviously, embezzlement could never become “lawful” and still retain its character. If “lawful,” it would constitute nothing more than a loan, or possibly a gift, to the “embezzler,” neither of which would produce a taxable gain to him.

There is still another obvious and important distinction between embezzlement and the varieties of illegal activity listed by the prevailing opinion — one which clearly calls for a different tax treatment. Black marketeering, gambling, bribery, graft and like activities generally give rise to no legally enforceable right of restitution — to no debtor-creditor relationship which the law will recognize.6 Condemned either by statute or public policy, or both, such transactions are void ab initio. Since any consideration which may have passed is not legally recoverable, its recipient has realized a taxable gain, an “accession to income,” as clearly as if his “indebtedness” had been discharged by a full, release or by the running of a Statute of Limitations. As we have already shown at length, quite the opposite is true when an embezzlement occurs; for then the victim acquires an immediately ripe and enforceable claim to repayment, and the embezzler assumes a legal debt equal to his acquisition.

To reach the result that it does today, The Chief Justice’s opinion constructs the following theory for defining taxable income:

“When a taxpayer acquires earnings, lawfully or unlawfully, without the consensual recognition, *256express or implied, of an obligation to repay and without restriction as to their disposition, 'he has received income which he is required to return, even though it may still be claimed that he is not entitled to retain the money, and even though he may still be adjudged liable to restore its equivalent.’ North American Oil v. Burnet, supra, at p. 424. In such case, the taxpayer has actual command over the property taxed — the actual benefit for which the tax is paid,’ Corliss v. Bowers, supra. This standard brings wrongful appropriations within the broad sweep of 'gross income’; it excludes loans. When a law-abiding taxpayer mistakenly receives income in one year, which receipt is assailed and found to be invalid in a subsequent year, the taxpayer must nonetheless report the amount as 'gross income’ in the year received. United States v. Lewis, supra; Healy v. Commissioner, supra.”

This novel formula finds no support in our prior decisions, least of all in those which are cited. Corliss v. Bowers, 281 U. S. 376, involved nothing more than an inter vivos trust created by the taxpayer to pay the income to his wife. Since he had reserved the power to alter or abolish the trust at will, its income was taxable to him under the express provisions of § 219 (g), (h) of the Revenue Act of 1924. North American Oil v. Burnet, 286. U. S. 417, is the case which introduced the principle since used to facilitate uniformity and certainty in annual tax accounting procedure, i. e., that a taxpayer must report in gross income, in the year in which received, money or property acquired under a ''claim of right” — a colorable claim of the right to exclusive possession of the money or property. Thus, in its complete form, the sentence in North American Oil from which the above-quoted fragment was extracted reads: “If a taxpayer receives earnings under a claim of right and without *257restriction as to its [sic] disposition, he has received income which he is required to return, even though it may still be claimed that he is not entitled to retain the money, and even though he may still be adjudged liable to restore its equivalent.” Id., at 424. (Emphasis added.) But embezzled funds, like stolen property generally, are not “earnings” in any sense and are held without a vestige of a colorable claim of right; they constitute the principal of a debt. Of no significance whatever is the formality of “consensual recognition, express or implied” of an obligation to repay. By substituting this meaningless abstraction in place of the omitted portion of the North American Oil test of when a receipt constitutes taxable income, the prevailing opinion today goes far beyond overruling Wilcox — it reduces a substantial body of tax law into uncertainty and confusion. The above-cited case of United States v. Lewis, 340 U. S. 590, decided 19 years after North American Oil, demonstrates the truth of this. For there we said:

“The ‘claim of right’ interpretation of the tax laws has long been used to give finality to [the accounting] period, and is now deeply rooted in the federal tax system. . . . We see no reason why the Court should depart from this well-settled interpretation merely because it results in an advantage or disadvantage to a taxpayer.” 340 U. S., at 592.

The same principle was reiterated and applied in Healy v. Commissioner, 345 U. S. 278.

The supposed conflict between Wilcox and Rutkin, upon which The Chief Justice’s opinion seeks to justify its repudiation of Wilcox,7 has been adequately treated in *258the opinion of Mr. Justice Black, and I agree with him that those cases were fully intended to be, and are, reconcilable, both on their controlling facts and applicable law. If the unnecessarily broad language used in the Rutkin opinion has misled any of the lower federal courts in their understanding of the principles underlying Wilcox, we should clarify their understanding at this time, and continue our adherence to “a prior doctrine more embracing in its scope, intrinsically sounder, and verified by experience.” Helvering v. Hallock, 309 U. S. 106, 119.

announced the judgment of the Court and an opinion in which

The issue before us in this case is whether embezzled funds are to be included in the “gross income” of the embezzler in the year in which the funds are misappro*214priated under § 22 (a) of the Internal Revenue Code of 19391 and § 61 (a) of the Internal Revenue Code of 1954.2

The facts are not in dispute. The petitioner is a union official who, with another person, embezzled in excess of $738,000 during the years 1951 through 1954 from his employer union and from an insurance company with which the union was doing business.3 Petitioner failed to report these amounts in his gross income in those years and was convicted for willfully attempting to evade the federal income tax due for each of the years 1951 through 1954 in violation of § 145 (b) of the Internal Revenue Code of 19394 and § 7201 of the Internal Rev*215enue Code of 1954.5 He was sentenced to a total of three years’ imprisonment. The Court of Appeals affirmed. 273 F. 2d 5. Because of a conflict with this Court’s decision in Commissioner v. Wilcox, 327 U. S. 404, a case whose relevant facts are concededly the same as those in the case now before us, we granted certiorari. 362 U. S. 974.

In Wilcox, the Court held that embezzled money does not constitute taxable income to the embezzler in the year of the embezzlement under § 22 (a) of the internal Revenue Code of 1939, Six years later, this Court held, in Rutkin v. United States, 343 U. S. 130, that extorted money does constitute taxable income to the extortionist in the year that the money is received under § 22 (a) of the Internal Revenue Code of 1939. In Rutkin, the Court did not overrule Wilcox, but stated:

“Wé do not reach in this case the factual situation involved in Commissioner v. Wilcox, 327 U. S. 404. We limit that case to its facts. There embezzled funds were held not to constitute taxable income to the embezzler under § 22 (a).” Id., at 138.6

However, examination of the reasoning used in Rutkin leads us inescapably to the conclusion that Wilcox was thoroughly devitalized.

The basis for the Wilcox decision was “that a taxable gain is conditioned upon (1) the presence of a claim of right to the alleged gain and (2) the absence of a definite, *216unconditional obligation to repay or return that which would otherwise constitute a gain. Without some bona fide legal or equitable claim, even though it be contingent or contested in nature, the taxpayer cannot be said to have received any gain or profit within the reach of §22 (a).” Commissioner v. Wilcox, supra, at p. 408. Since Wilcox embezzled the money, held it “without any semblance of a bona fide claim of right,” ibid., and therefore “was at all times under an unqualified duty and obligation to repay the money to his employer,” ibid., the Court found that the money embezzled was not includible within “gross income.” But, Rutkin’s legal claim was no greater than that of Wilcox. It was specifically found “that petitioner had no basis for his claim . . . and that he obtained it by extortion.” Rutkin v. United States, supra, at p. 135. Both Wilcox and Rutkin obtained the money by means of a criminal act; neither had a bona fide claim of right to the funds.7 Nor was Rutkin’s obligation to repay the extorted money to the victim any less than that of Wilcox. The victim of an extortion, like the victim of an embezzlement, has a right to restitution. Furthermore, it is inconsequential that an embezzler may lack title to the sums he appropriates while an extortionist may gain a voidable title. Questions of federal income taxation are not determined by .such “attenuated subtleties.” Lucas v. Earl, 281 U. S. 111, 114; Corliss v. *217Bowers, 281 U. S. 376, 378. Thus, the fact that Rutkin secured the money with the consent of his victim, Rutkin v. United States, supra, at p. 138, is irrelevant. Likewise unimportant is the fact that the sufferer of an extortion is less likely to seek restitution than one whose funds are embezzled. What is important is that the right to recoupment exists in both situations.

Examination of the relevant cases in the courts of appeals lends credence to our conclusion that the Wilcox rationale was effectively vitiated by this Court’s decision in Rutkin,8 Although this case appears to be the first to arise that is “on all fours” with Wilcox, the lower federal courts, in deference to the undisturbed Wilcox holding, have earnestly endeavored to find distinguishing facts in the cases before them which would enable them to include sundry unlawful gains within “gross income.”9

*218It had been a well-established principle, long before either Rutkin or Wilcox, that unlawful, as well as lawful, gains are comprehended within the term “gross income.” Section II B of the Income Tax Act of 1913 provided that “the net income of a taxable person shall include gains, profits, and income . . . from ... the transaction of any lawful business carried on for gain or profit, or gains or profits and income derived from any source whatever . . . .” (Emphasis supplied.) 38 Stat. 167. When the statute was amended in 1916, the one word “lawful” was omitted. This revealed, we think, the obvious intent of that Congress to tax income derived from both legal and illegal sources, to remove the incongruity of having the gains of the honest laborer taxed and the gains of the dishonest immune. Rutkin v. United States, supra, at p. 138; United States v. Sullivan, 274 U. S. 259, 263. Thereafter, the Court held that gains from illicit traffic in liquor are includible within “gross income.” Ibid. See also Johnson v. United States, 318 U. S. 189; United States v. Johnson, 319 U. S. 503. And, the Court has pointed out, with approval, that there “has been a widespread and settled administrative and judicial recognition of the taxability of unlawful gains of many kinds,” Rutkin v. United States, supra, at p. 137. These include protection payments made to racketeers, ransom payments paid to kidnappers, .bribes, money derived from the sale of unlawful insurance policies, graft, black market gains, funds obtained from the operation of lotteries, income from race track bookmaking and illegal prize fight pictures. Ibid.

The starting point in all cases dealing with the question of the scope of what is included in “gross income” begins with the basic premise that the purpose of Congress was “to use the full measure of its taxing power.” Helvering *219v. Clifford, 309 U. S. 331, 334. And the Court has given a liberal construction to the broad phraseology of the “gross income” definition statutes in recognition of the intention of Congress to tax all gains except those specifically exempted. Commissioner v. Jacobson, 336 U. S. 28, 49; Helvering v. Stockholms Enskilda Bank, 293 U. S. 84, 87-91. The language of § 22 (a) of the 1939 Code, “gains or profits and income derived from any source whatever,” and the more simplified language of § 61 (a) of the 1954 Code, “all income from whatever source derived,” have been held to encompass all “aeces- • sions to wealth, clearly realized, and over which the taxpayers have complete dominion.” Commissioner v. Clenshaw Glass Co., 348 U. S. 426, 431. A gain “constitutes taxable income when its recipient has such control over it that, as a practical matter, he derives readily realizable economic value from it.” Rutkin v. United States, supra, at p. 137. Under these broad principles, we believe that petitioner’s contention, that all unlawful gains are taxable except those resulting from embezzlement, should fail.

When a taxpayer acquires earnings, lawfully or unlawfully, without the consensual recognition, express or implied, of an obligation to repay and without restriction as to their disposition, “he has received income which he is required to return, even though it may still be claimed that he is not entitled to retain the, money, and even though he may still be adjudged hable to restore its equivalent.” North American Oil v. Burnet, supra, at p. 424. In such case, the taxpayer has “actual command over the property taxed — the actual benefit for which the tax is paid,” Corliss v. Bowers, supra. This standard brings wrongful appropriations within the broad sweep of “gross income”; it excludes loans. When a law-abiding taxpayer mistakenly receives income in one year, which receipt is assailed and found to be invalid in a subsequent *220year, the taxpayer must nonetheless report the amount as “gross income” in the year received. United States v. Lewis, supra; Healy v. Commissioner, supra. We do not believe that Congress intended to treat a law-breaking taxpayer differently. Just as the honest taxpayer may deduct any amount repaid in the year in which the repayment is made, the Government points out that, “If, when, and to the extent that the victim recovers back the misappropriated funds, there is of course a reduction in the embezzler’s income.” Brief for the United States, p. 24.10

Petitioner contends that the Wilcox rule has been in existence since 1946; that if Congress had intended to change the rule, it would have done so; that there was a general revision of the income tax laws in 1954 without mention of the rule; that a bill to change it11 was introduced in the Eighty-sixth Congress but was not acted upon; that, therefore, we may not change the rule now. But the fact that Congress has remained silent or has re-enacted a statute which we have construed, or that congressional attempts to amend a rule announced by this Court have failed, does not necessarily debar us from re-examining and correcting the Court’s own errors. Girouard v. United States, 328 U. S. 61, 69-70; Helvering v. Hallock, 309 U. S. 106, 119-122. There may have been any number of reasons why Congress acted as it did. Helvering v. Hallock, supra. One of the reasons could well *221be our subsequent decision in Rutkin which has been thought by many to have repudiated Wilcox. Particularly might this be true in light of the decisions of the Courts of Appeals which have been riding a narrow rail between the two cases and further distinguishing them to the disparagement of Wilcox. See notes 8 and 9, supra.

We believe that Wilcox was wrongly decided and we find nothing in congressional history since then to persuade us that Congress intended to legislate the rule. Thus, we believe that we should now correct the error and the confusion resulting from it, certainly if we do so in a manner that will not prejudice those who might have relied on it. Cf. Helvering v. Hallock, supra, at 119. We should not continue to confound confusion, particularly when the result would be to perpetuate the injustice of relieving embezzlers of the duty of paying income taxes on the money they enrich themselves with through theft while honest people pay their taxes on every conceivable type of income.

But, we are dealing here with a felony conviction under statutes which apply to any person who “willfully” fails to account for his tax or who “willfully” attempts to evade his obligation. In Spies v. United States, 317 U. S. 492, 499, the Court said that § 145 (b) of the 1939 Code embodied “the gravest of offenses against the revenues,” and stated that willfulness must therefore include an evil motive and want of justification in view of all the circumstances. Id., at 498. Willfulness “involves a specific intent which must be proven by independent evidence and which cannot be inferred from the mere understatement of income.” Holland v. United States, 348 U. S. 121, 139.

We believe that the element of willfulness could not be proven in a criminal prosecution for failing to include embezzled funds in gross income in the year of misappropriation so long as the statute contained the gloss placed upon it by Wilcox at the time the alleged crime was *222committed. Therefore, we feel that petitioner’s conviction may not stand and that the indictment against him must be dismissed.

Since Mr. Justice Harlan, Mr. Justice Frank- ' eurter, and Mr. Justice Clark agree with us concerning Wilcox, that case is overruled. Mr. Justice Black, Mr. Justice ' Douglas, and Mr. Justice Whittaker believe that petitioner’s conviction must be reversed and the case dismissed for the reasons stated in their opinions.

Accordingly, the judgment of the Court of Appeals is reversed and the case is remanded to the District Court with directions to dismiss the indictment.

It is so ordered.

whom

On February 25, 1946, fifteen years ago, this Court, after mature consideration, and in accordance with what at that time represented the most strongly supported judicial view, held, in an opinion written by Mr. Justice Murphy to which only one Justice dissented, that money secretly taken by an embezzler for his own use did not constitute a taxable gain to him under the federal income tax laws. Commissioner v. Wilcox, 327 U. S. 404. The Treasury Department promptly accepted this ruling in a bulletin declaring that the “mere act of embezzlement does not of itself result in taxable income,” although properly urging that “taxable income may result to the embezzler, depending on the facts in the particular case.”1 *223During the fifteen years since Wilcox was decided, both this Court and Congress, although urged to do so, have declined to change the Wilcox interpretation of statutory “income” with respect to embezzlement. In this case, however, a majority of the Court overrules Wilcox. Only three of the members of the Court who decided the Wilcox case are participating in this case — Mr. Justice Frankfurter, Mr. Justice Douglas, and myself. Mr. Justice Douglas and I dissent from the Court’s action in “overruling” Wilcox and from the prospective way in which this is done. We think Wilcox was sound, when written and is sound now.

I.

We dissent from the way the majority of the Court overrules Wilcox. If the statutory interpretation of “taxable income” in Wilcox is wrong, then James is guilty of violating the tax evasion statute for the trial court’s judgment establishes that he embezzled funds and wilfully refrained from reporting them as income. It appears to us that District Courts are bound to be confused as to what they can do hereafter in tax-evasion cases involving “income” from embezzlements committed prior to this day. Three Justices vote to overrule Wilcox under what we believe to be a questionable formula, at least a new one in the annals of this Court, and say that although failure to report embezzled funds has, despite Wilcox, always been a crime under the statute, people who have violated this law in- the past cannot be prosecuted but people who embezzle funds after this opinion is announced can be prosecuted for failing to report these funds as a “taxable gain.” Three other Justices who vote to overrule Wilcox say that past embezzlers can be prosecuted for the crime of tax evasion although two of those Justices believe the Government must prove that the past embezzler did not commit his crime in reliance on Wilcox. *224Thus, although it was not the law yesterday, it will be the law tomorrow that funds embezzled hereafter are taxable income; and although past embezzlers could not have been prosecuted yesterday, maybe they can and maybe they cannot be prosecuted tomorrow for the crime of tax evasion. (The question of the civil tax liability of past embezzlers is left equally unclear.) We do not challenge the wisdom of those of our Brethren who refuse to make the Court’s new tax evasion crime applicable to past conduct. This would be good governmental policy even though the ex post facto provision of the Constitution has not ordinarily been thought to apply to judicial legislation. Our trouble with this aspect of the Court’s action is that it seems to us to indicate that the Court has passed beyond the interpretation of the tax statute and proceeded substantially to amend it.

We realize that there is a doctrine with wide support to the effect that under some circumstances courts should make their decisions as to what the law is apply only prospectively.2 Objections to such a judicial procedure, however, seem to us to have peculiar force in the field of criminal law. In the first place, a criminal statute that is so ambiguous in scope that an interpretation of it brings about totally unexpected results, thereby subjecting people to penalties and punishments for conduct which they could not know was criminal under existing law, raises serious questions of unconstitutional vagueness.3 Moreover, for a court to interpret a criminal statute in such a way as to make punishment for past conduct under it so unfair and unjust that the interpretation should be given only prospective application seems to us to be the creation of a judicial crime that Congress might not want *225to create. This country has never been sympathetic with judge-created crimes. Their rejection under our Constitution was said to have been “long since settled in public opinion” even as early as 1812 when the question first reached this Court in United States v. Hudson & Goodwin, 7 Cranch 32. In that case this Court emphatically declared that the federal courts have no common-law jurisdiction in criminal cases. They are not “vested with jurisdiction over any particular act done by an individual in supposed violation of the peace and dignity of the sovereign power.” Rather, “[t]he legislative authority of the Union-must first make an act a crime, affix a punishment to it, and declare the Court that shall have jurisdiction of the offence.”4

In our judgment one of the great inherent restraints upon this Court’s departure from the field of interpretation to enter that of lawmaking has been the fact that its judgments could not be limited to prospective application. This Court and in fact all departments of the Government have always heretofore realized that prospective lawmaking is the function of Congress rather than of the courts. We continue to think that this function should be exercised only by Congress under our constitutional system.

II.

We think Wilcox was right when it was decided and is right now. It announced no new, novel doctrine. One need only look at the Government’s briefs in this Court in the Wilcox case to see just how little past judicial support could then be mustered had the Government sought to send Wilcox to jail for his embezzlement under the guise of a tax evasion prosecution. The Government did cite many cases from many courts saying that under the federal income tax law gains are no less taxable because *226they have been acquired by illegal methods. This Court had properly held long before Wilcox that there is no “reason why the fact that a business is unlawful should exempt it from paying the taxes that if lawful it would have to pay.”5 We fully recognized the correctness of that holding in Wilcox:

“Moral turpitude is not a touchstone of taxability. The question, rather, is whether the taxpayer in fact received a statutory gain, profit or benefit. That the taxpayer’s- motive may have been reprehensible or the mode of receipt illegal has no bearing upon the application of § 22 (a).”6

The Court today by implication attributes quite a different meaning or consequence to the Wilcox opinion. One opinion argues at length the “well-established principle . . . that unlawful, as well as lawful, gains are comprehended within the term ‘gross income.’ ” Wilcox did not deny that; we do not deny that. This repeated theme of our Brethren is wholly irrelevant since the Wilcox holding in no way violates the sound principle of treating “gains” of honest and dishonest taxpayers alike. The whole basis of the Wilcox opinion was that an embezzlement is not in itself “gain” or “income” to the embezzler within the tax sense, for the obvious reason that the embezzled property still belongs, and is known to belong, to the rightful owner. It is thus a mistake to argue that petitioner’s contention is “that all unlawful gains are taxable except those resulting from embezzlement.”

As stated in Wilcox, that case was brought to us because of a conflict among the Circuits. The Ninth Circuit in Wilcox had held that embezzled funds were not any more “taxable income” to the embezzler than *227borrowed funds would have been.7 The Fifth Circuit, in McKnight v. Commissioner, had decided the same thing.8 The Eighth Circuit, however, had decided in Kurrle v. Helvering that embezzled funds were taxable income.9 Comparison of the three opinions readily shows that the arguments of the Fifth and Ninth Circuits against tax-ability of such funds were much stronger than the arguments of the Eighth Circuit for such taxability. The whole picture can best be obtained from the court’s opinion in McKnight v. Commissioner, written by Judge Sibley, one of the ablest circuit judges of his time. He recognized that the taxpayer could not rely upon the unlawfulness of his business to defeat taxation if he had made a “gain” in that business. He pointed out, however, that the ordinary embezzler “got no title, void or voidable, to what he took. He was still in possession as he was before, but with a changed purpose. He still had no right nor color of right. He claimed none.”10 Judge Sibley’s opinion went on to point out that the “first takings [of an embezzler] are, indeed, nearly always with the intention of repaying, a sort of unauthorized borrowing. It must be conceded that no gain is realized by borrowing, because of the offsetting obligation.”11 Approaching the matter from a practical standpoint, Judge Sibley also explained that subjecting the embezzled funds to a tax would amount to allowing the United States “a preferential claim for part of the dishonest gain, to the direct loss and detriment of those to whom it ought to be restored.”12 He was not willing to put the owner of *228funds that had been stolen in competition with the United States Treasury Department as to which one should have a preference to get those funds.

It seems to us that Judge Sibley’s argument was then ■ and is now unanswerable. The rightful owner who has entrusted his funds to an employee or agent has troubles enough when those funds are embezzled without having the Federal Government step in with its powerful claim that the embezzlement is a taxable event automatically subjecting part of those funds (still belonging to the owner) to the waiting hands of the Government’s tax gatherer. We say part of the owner’s funds because it is on the supposed “gain” from them that the embezzler is now held to be duty-bound to pay the tax and history ' probably records few Instances of independently wealthy embezzlers who have had honstolen assets available for payment of taxes.

There has been nothing shown to us on any of the occasions when we have considered this problem to indicate that Congress ever intended its income tax laws to be construed as imposing what is in effect a property or excise tax on the rightful owner’s embezzled funds, for which the owner has already once paid income tax when he rightfully acquired them. In our view, the Court today does Congress a grave injustice by assuming that it has imposed this double tax burden upon the victim of an embezzlement merely because someone has stolen his money, particularly when Congress has refused requests that it do so. The owner whose funds have been embezzled has done nothing but entrust an agent with possession of his funds for limited purposes, as many of us have frequent occasion to do in the course of business or personal affairs. Ordinarily the owner is not, and has no reason to be, at all aware of an embezzlement until long after the first misuse occurs. If Congress ever did manifest an intention to select the mere fact of embezzlement *229as the basis for imposing a double tax on the owner, we think a serious question of confiscation in violation of the Fifth Amendment would be raised. All of us know that with the strong lien provisions of the federal income tax law an owner of stolen funds would have a very rocky road to travel before he got back, without paying a good slice to the Federal Government, such funds as an embezzler who had not paid the tax might, perchance, not have dissipated. An illustration of what this could mean to a defrauded employer is shown in this very case by the employer’s loss of some $700,000, upon which the Government claims a tax of $559,000.

It seems to be implied that one reason for overruling Wilcox is that a failure to hold embezzled funds taxable would somehow work havoc with the public revenue or discriminate against “honest” taxpayers and force them to pay more taxes. We believe it would be impossible to substantiate either claim. Embezzlers ordinarily are not rich people against whom judgments, even federal tax. judgments. can be enforced. Judging from the meager settlements that those defrauded were apparently compelled to make with the embezzlers in this very case, it is hard to imagine that the Treasury will be able to collect the more than $500,000 it claims. And certainly the Wilcox case does not seem to have been one in which the Government could have collected any great amount of tax. The employer’s embezzled $11,000 there went up in gambling houses. The scarcity of cases involving alleged taxes due from embezzlers is another indication that the Government cannot expect to make up any treasury deficits with taxes collected from embezzlers and thieves, especially when the cost to the Government of investigations and court proceedings against suspected individuals is considered. And, as already indicated, to the extent that the Government could be successful in collecting some taxes from *230embezzlers, it would most likely do so at the expense of the owner whose money had been stolen.

It follows that, except for the possible adverse effect on rightful owners, the only substantial result that one can foresee from today’s holding is that the Federal Government will, under the guise of a tax evasion charge, prosecute people for a simple embezzlement. But the Constitution grants power to Congress to get revenue not to prosecute local crimes. And if there is any offense which under our dual system of government is a purely local one which the States should handle, it is embezzlement or theft. The Federal Government stands to lose much money by trying to take over prosecution of this type of local offense. It is very doubtful whether the further congestion of federal court dockets to try such local offenses is good for the Nation, the States or the people. Here the embezzler has already pleaded guilty to the crime of embezzlement in a state court, although the record does not show what punishment he has received. Were it not for the novel formula of applying the Court’s new law prospectively, petitioner would have to serve three years in federal prison in addition to his state sentence. This graphically illustrates one of the great dangers of opening up the federal tax statutes, or any others, for use by federal prosecutors against defendants who not only can be but are tried for their crimes in local state courts and punished there. If the people of this country are to be subjected to such double jeopardy and double punishment, despite the constitutional command against double jeopardy, it seems to us it would be far wiser for this Court to wait and let Congress attempt to do it.

III.

The Wilcox case was decided fifteen years ago. Congress has met every year since then. All of us know that the House and Senate Committees responsible for our *231tax laws keep a close watch on judicial rulings interpreting the Internal Revenue Code. Each committee has one or more experts at its constant disposal. It cannot possibly be denied that these committees and these éxperts are, and have been, fully familiar with the Wilcox holding. When Congress is dissatisfied with a tax decision of this Court, it can and frequently does act very quickly to overturn it.13 On one occasion such an overruling enactment was passed by both the House and Senate and signed by the President all within one day' after the decision was rendered by this Court.14 In 1954 Congress, after extended study, completely overhauled and recodified the Internal Revenue Code. The Wilcox holding was left intact. In the Eighty-sixth Congress and in the present Eighty-seventh Congress bills have been introduced to subject embezzled funds to income taxation.15 They have not been passed. This is not an instance when we can say that Congress may have neglected to change the law because it did not know what *232was going on in the courts or because it was not asked to do so, as was the case in Helvering v. Hallock.16 Nor is this a case in which subsequent affirmative congressional action manifested a view inconsistent with our prior decision, as was true in Girouard v. United States.17 What we have here instead is a case in which Congress has not passed bills that have been introduced to make embezzled funds taxable and thereby make failure to report them as income a federal crime. For this Court to hold under such circumstances that the inherent ambiguity of legislative inaction gives the Court license to repudiate the long-standing interpretation of the income tax statute and thereby bring additional conduct within the tax evasion criminal statute seems to us to be flagrantly violative of the almost universally accepted axiom that criminal statutes are narrowly and strictly construed. Our Brethren cite no precedent in which this or any other court in the English-speaking world has so deliberately overruled a long-standing prior interpretation of a statute in order to create a crime which up to that time did not exist.

This Court as well as Congress was fully apprised of the various criticisms made in some Courts of Appeals opinions and elsewhere against the Wilcox holding, yet it has likewise until today steadfastly refused to overrule that holding during these fifteen years. This has been in the face of the fact that the Government expressly urged that we do so in 1955, nine years after Wilcox was decided *233and three years after the decision in Rutkin v. United States, 343 U. S. 130. On that occasion the Court of Appeals for the Second Circuit, speaking through Judge Frank for himself and Judge Medina, had held in the case of J. J. Dix, Inc., v. Commissioner that embezzled funds were not taxable as income, relying wholly on the Wilcox decision.18 Judge Hincks dissented, saying that if the facts of Dix were not enough to distinguish it from Wilcox he would not follow Wilcox. In urging us to grant certiorari, the Government said that the case presented a recurring problem in the administration of the income tax laws. One of the arguments the Government presented for overruling Wilcox, strange as it may seem, was that “[s]everal prosecutions have recently been authorized and are now pending in various District Courts, even though the disputed income in those cases apparently came from embezzlements or closely analogous crimes.”19 And the next to the last sentence of its petition was: “In short, the question whether the proceeds of embezzlement, unlike other illegal income, are to enjoy a preferred tax-exempt status, will continue to perplex the lower courts until it is settled by this Court.”20 We denied certiorari.21 There is surely less reason to repudiate and “devitalize” Wilcox now, six years after the Court, as composed at that time, refused to overrule it.

Of course the rule of stare decisis is not and should not be an inexorable one. This is particularly true with reference to constitutional decisions involving determinations beyond the power of Congress to change, but Congress can and does change statutory interpretations. It *234is perfectly proper and right that it should do so when it believes that this Court’s interpretation of a statute embodies a policy that Congress is against. But Congress has not taken favorable action on bills introduced to overturn our Wilcox holding even after we declined the Government’s request to reverse the identical holding in Dix, the latter having occurred three years after the decision in Rutkin which our Brethren now say may have misled Congress into thinking that we had repudiated the Wilcox holding.

It seems to us that we gave the doctrine of stare decisis its proper scope in our treatment of this Court’s decision in Federal Baseball Club v. National League of Professional Baseball Clubs, 259 U. S. 200. In that case this Court had held for reasons given that professional baseball was not covered by the antitrust acts. Congress was asked through the years to change the law in this respect but declined to do so. In Toolson v. New York Yankees, Inc., 346 U. S. 356, we followed the holding of that case without re-examination of the underlying issues “so far as that decision determines that Congress had no intention of including the business of baseball within the scope of the federal antitrust laws.” Later we were asked to extend the Federal Baseball case and to hold that the business of boxing could not without congressional action be brought within the antitrust laws. We emphatically declined to do so in United States v. International Boxing Club, 348 U. S. 236, nor did we overrule Toolson in that case, despite strong arguments that the reasoning of the Court in the first baseball case was equally applicable to the business of boxing. We said about the proposed exemption of boxing from the antitrust laws that “[tlheir remedy, if they are entitled to one, lies in further resort to Congress.”22 That case and that statement fit this-case precisely. In fact, as we are about to explain, a *235far more meaningful distinction can be made between embezzlement and extortion for purposes of this case than it was possible to make between baseball and boxing for purposes of that case, as Mr. Justice Frankfurter’s dissenting opinion in that case demonstrates.

If the Government wants to prosecute the local crime of embezzlement, ostensibly because of “tax evasion,” it seems clear to us that it should take its request to Congress which has power to pass on it and which has, to date, refused to do what the Government asks us to do in this case.

IV.

■ Our Brethren advance as a reason for overruling Wilcox the 1952 decision in Rutkin v. United States, which was decided three years before we denied certiorari in the Dix case. They say that “the reasoning used in Rutkin leads us inescapably to the conclusion that Wilcox was thoroughly devitalized.” This follows, to some extent, the statement in the Government’s brief that “Wilcox and Rutkin cannot be reconciled on the basis of asserted technical differences between the extortionist and the embezzler. . . . The proper course, we submit, ... is to recognize that the Wilcox rationale was rejected in Rutkin, is unsound, and can no longer be regarded as having vitality. Embezzled funds represent taxable gains.”23

There is no doubt that some of the reasoning in the Rutkin opinion rejected some of the reasoning in the Wilcox opinion. But this is true only with respect to the broad general standards formulated in the two cases, and such standards of course cannot be accepted as universal panaceas to be mechanically applied to solve all the concrete problems in cases like these. Moreover, the Rutkin opinion expressly purported not to overrule Wilcox and *236specifically said that Wilcox was still to govern cases fitting its facts, clearly meaning embezzlement cases.24 And the Government had not asked in Butkin that Wilcox be overruled. Its argument was that Wilcox was “inapplicable” to the facts in the Butkin record. The Government’s brief went on to emphasize that the record in Wilcox showed only the bare receipt of money wholly belonging to another, while Rutkin had received the money “as a result of a bilateral agreement” and, as the Court of Appeals had pointed out, “with a 'semblance of a bona fide claim of right’, a conclusion fully substantiated by the testimony of both the petitioner and the Government witness Reinfeld.”25 The Government went on to distinguish Butkin further by pointing out that there was “not the slightest hint in the record” that Rutkin ever had an obligation to repay the funds he took.

After this Court was persuaded by the Government in Butkin to accept its distinctions between Butkin and Wilcox, it seems rather odd to have the Government now contend that the two cases are irreconcilable. While we disagreed, we can understand why the majority in But-*237kin drew the distinctions it did. Although the victim of either embezzlement or extortion ordinarily has a legal right to restitution, the extortion victim, like a blackmail victim, can in a sense be charged with complicity in bringing about the taxable event in that he knowingly surrendered the funds to the extortionist, sometimes in payment of an actual obligation. Unlike the victim of an ordinary theft, he generally knows who has taken the property from him and he consents to the taking though under duress; and unlike most victims of embezzlement, he is able to report the taking to law enforcement officers during the taxable year and his failure to do so might be considered a kind of continuing consent to the extortionist’s dominion over the property. The longer he acquiesces the less likely it becomes that the extortion victim ever will demand restitution;26 but once the victim of an embezzlement finds out that his property has been stolen, he most likely will immediately make efforts to get it back. Thus, although we still think Rutkin was wrongly decided for the reasons expressed in the dissenting opinion in that case, we can understand the argument for application of a sort of caveat emptor rule to persons who submit to blackmail or extortion, since it is far from certain that they will ever expose themselves by seeking repayment of what they paid out. The distinctions between crimes like embezzlement and crimes like blackmail and extortion, therefore, are not merely *238technical, legalistic “attenuated subtleties” for purposes of this decision, but are differences based upon practicalities such as often underlie the distinctions that have been developed in our law.

In departing from both the Wilcox and Rutkin decisions today, our Brethren offer no persuasive reasons to prove that their judgment in overruling Wilcox is better than that of the Justices who decided that case. It contributes nothing new to the analysis of this problem to say repeatedly that the dishonest man must be subject to taxation just as the honest. As already said, Chief Justice Stone and the others sitting with him on the Wilcox Court fully accepted that general principle and we do still. Applying it here, we would say the embezzler should be treated just like the law-abiding, honest borrower who has obtained the owner’s consent to his use of the money.27 It *239would be unthinkable to tax the borrower on his “gain” of the borrowed funds and thereby substantially impair the lender’s chance of ever recovering the debt. The injury that the Government would inflict on the lender by making the borrower less able to repay the loan surely would not be adequately compensated by telling the lender that he can take a tax deduction for the loss, and it is equally small comfort to the embezzlement victim for the Government, after taking part of his property as a tax on the embezzler, to tell the victim that he can take a deduction for his loss if he has any income against which to offset the deduction. There is, of course, one outstanding distinction between a borrower and an embezzler, and that is that the embezzler uses the funds without the owner’s consent. This distinction can be of no importance for purposes of taxability of the funds, however, because as a matter of common sense it suggests that there is, if anything, less reason to tax the embezzler than the borrower. But if this distinction is to be the reason why the embezzlement must be taxed just as “the gains of the honest laborer,” then the use of this slogan in this case is laid bare as no more than a means of imposing a second punishment for the crime of embezzlement without regard to revenue considerations, the effect on the rightful owner, or the proper role of this Court when asked to overrule a criminal statutory precedent. The double jeopardy implications would seem obvious,28 *240and discussion of the serious inadvisability for other reasons of thus injecting the Federal Government into local law enforcement can be found in the dissenting opinion in Rutkin.

We regret very much that it seems to be implied that . the writer of the Rutkin opinion and those who agreed to it intended to overrule Wilcox when it is manifest that the language the Court used in Rutkin was meant to leave precisely the opposite impression. We are sure that our Brethren at that time did not intend to mislead the public, and it would be hard to imagine why they said what they did in the Rutkin opinion had they not specifically considered and rejected the possibility of overruling Wilcox then and there. We think it is unjustifiable to say nine years after Rutkin that it “devitalized” or “repudiated” the Wilcox holding when the Rutkin opinion said explicitly that Wilcox is still the rule as to embezzlement. Congress has seen fit to let both decisions stand, and we think the present Court should do the same.

V.

Even if we were to join with our Brethren in accepting the Government’s present contention that Wilcox and Rutkin cannot both stand, we would disagree as to which of the two decisions should now be repudiated. This is true not only because we would feel less inhibition about narrowing rather than broadening the reach of a previously construed criminal statute. Regardless of such considerations, our conviction that the Rutkin case was wrongly decided in this Court remains undiminished and has been further substantiated by the subsequent events in that controversy, which show all the more clearly the deplorable consequences that can- result when federal courts subject people who violate state criminal laws to *241a double or treble prosecution for the state crime under the guise of attempted enforcement of federal tax laws.29

For the foregoing reasons, as well as the reasons stated in Mr. Justice Whittaker’s opinion, we would reaffirm our holding in Commissioner v. Wilcox, reverse this judgment and direct that the case be dismissed.

whom

I fully agree with so much of The Chief Justice’s opinion as dispatches Wilcox to a final demise. But as to the disposition of this case, I think that rather than an outright reversal, which his opinion proposes, the reversal should be for a new trial.

*242I share the view that it would be inequitable to sustain this conviction when by virtue of the Rutkin-Wilcox dilemma it might reasonably have been thought by one in petitioner’s position that no tax was due in respect of embezzled moneys. For as is pointed out, Rutkin did not expressly overrule Wilcox, but instead merely confined it “to its facts.” Having now concluded that Wilcox was wrongly decided originally, the problem in this case thus becomes one of how to overrule Wilcox “in a manner that will not prejudice those who might have relied on it.” Ante, p. 221.

It is argued, in reliance on Spies v. United States, 317 U. S. 492, and Holland v. United States, 348 U. S. 121, that so long as Wilcox remained on the books the element of “willfulness” required in prosecutions of this kind1 “could not be proven,” and hence, that the conviction of this petitioner fails without more. This would mean, I take it, that no future prosecution or past conviction involving tax derelictions of this nature, occurring during the Wilcox period, may be brought or allowed to stand. I cannot agree to such a disposition, which, in my view, is warranted by neither principle nor authority and would carry mischievous implications for the future.

The Spies and Holland cases, which are said to support outright reversal, stand for no more than that where, as here, a criminal tax statute makes “willfulness” an element of the offense, the Government must prove an “evil motive and want of justification in view of all the financial circumstances” on the part of the defendant, in failing to do what was required of him. While I agree that in the present case this made germane on the issue of willfulness the petitioner’s reliance or nonreliance on the *243continued vitality of the Wilcox doctrine,2 I can find nothing in Spies or Holland which justifies the view that the mere existence of Wilcox suffices alone to vitiate petitioner’s conviction as a matter of law. If, as appears to. have been the case, there was erroneous failure to take that factor into account at the trial on the issue of willfulness, the most that should happen is that petitioner should be given a new trial. This indeed is what Spies and Holland affirmatively indicate as the right solution of the problem this case presents. In Spies, it was said (at pp. 499-500):

“. . . By way of illustration, and not by way of limitation, we would think affirmative willful attempt may be inferred from conduct such as keeping a double set of books, making false entries or alterations, or false invoices or documents, destruction of books or records, concealment of assets or covering up sources of income, handling of one’s affairs to avoid making the records usual in transactions of the kind, and any conduct, the likely effect of which would be to mislead or to conceal. If the tax-evasion motive plays any part in such conduct the offense may be made out even though the conduct may also serve other purposes such as concealment of other crime.

“In this case there are several items of evidence apart from the default in filing the return and paying the tax which the Governmént claims will support an inference of willful attempt to evade or *244defeat the tax. These go to establish that petitioner insisted that certain income be paid to him in cash, transferred it to his own bank by armored car, deposited it, not in his own name but in the names of others of his family, and kept inadequate and misleading records. Petitioner claims other motives animated him in these matters. We intimate no opinion. Such inferences are for the jury. If on proper submission the jury found these acts, taken togther with willful failure to file a return and willful failure to pay the tax, to constitute a willful attempt to evade or defeat the tax, we would consider conviction of a felony sustainable.” To the same effect, see Holland, supra, at p. 139.

In the case at hand, the evidence of devious financial arrangements might well support the inference that petitioner’s purpose was not only to commit the embezzlement but also to secrete and immunize his gains from what he considered to be his tax liabilities in respect of those gains. The District Court, as the trier of the facts (there having been no jury), found that petitioner’s acts were “willful and were done in a knowing and conscious attempt to evade and defeat” his tax obligations. But since it does not appear that petitioner’s possible reliance on the Wilcox doctrine was considered below, Spies and Holland make it appropriate for us to send the case back for a new trial. They do not support foreclosing the Government from even undertaking to prove that the petitioner’s conduct was “willful” in this respect.

An outright reversal is equally unsound on principle. I take it that our decisions, in the tax and any other field for that matter relate back to the actual transactions with which they are concerned, and that that is only the normal concomitant of the fact- that we do not sit as an administrative agency making rulings for the future, but rather adjudicate actual controversies as *245to'rights and liabilities under the laws of the United States. There can be, I think, two justifications for barring a prosecution of this petitioner in the unusual circumstances presented here: (1) that by reason of Rutkin having formally left intact the Wilcox doctrine, petitioner did not have due warning of his possible criminal liability; and (2) that the Court, in making new “law” in Butkin, should, like the legislature, not impose criminal liability ex post facto.

As to the first consideration, where the defendant is charged in a case like this, with having “willfully” violated the law, I believe that both reason and authority require no more than that the trier of fact be instructed that it must take into account in determining the defendant’s “evil motive and want of justification,” Spies v. United States, 317 U. S., at 498, his possible reliance on Wilcox, which not until now has this Court explicitly stated was wrongly decided. As far as fairness to this petitioner is concerned, I do not see why that is not amply accorded by the disposition which Spies itself exemplifies. See p. 243, supra. On the other hand, if the trier of fact, properly instructed, finds that the petitioner did not act in bona fide reliance on Wilcox, but deliberately refused to report income and pay taxes thereon knowing of his obligation to do so and not relying on any exception in the circumstances, I do not see why even the strictest definition of the element of “willfulness” would not have been satisfied. Willfulness goes to motive, and the quality of a particular defendant’s motive would not seem to be affected by the fact that another taxpayer similarly situated had a different motive.

An altogether analogous situation was presented in United States v. Murdock, 290 U. S. 389. In that case the respondent had been convicted of willfully failing to supply information to the Bureau of Internal Revenue in that he relied on the possibility of state prosecution as *246justifying his invoking the federal privilege against self-incrimination. The Court said in that case:

“. . . He whose conduct is defined as criminal is one who ‘willfully’ fails to pay the tax, to make a return, to keep the required records, or to supply the needed information. Congress did not intend that a person, by-reason of a bona fide misunderstanding as to his liability for the tax, . . . should become a criminal by his mere failure to measure up to the prescribed standard of conduct. . . .

“It follows that the respondent was entitled to the charge he requested with respect to his good faith and actual belief. Not until this Court pronounced judgment in United States v. Murdock, 284 U. S. 141, had it been definitely settled that one under examination in a federal tribunal could hot refuse to answer on account of probable incrimination under state law. The question was involved, but not . decided, in Ballman v. Fagin, 200 U. S. 186, 195, and specifically reserved in Vajtauer v. Comm’r of Immigration, 273 U. S. 103, 113. The trial court could not, therefore, properly tell the jury the defendant’s assertion of 'the privilege was so unreasonable and ill founded as to exhibit bad faith and establish willful wrongdoing. This was the effect of the instructions given'. We think the Circuit Court' of Appeals correctly upheld the respondent’s right to have the question of absence of evil motive submitted to the jury . . . .” (Emphasis-supplied.)

It would seem that precisely the same disposition is in order in this case. Nor'do I think that distinctions in terms of the nature of the defendant’s legal misapprehension, its degree, its justifiability, or its source are either warranted or would be manageable as a basis for deciding future cases.

*247Coming now to the other possible rationale for barring the prosecution of this petitioner, it might be argued that petitioner at the time he failed to make his return was not under any misapprehension as to the law, but indeed that at the time and under the decisions of this Court his view of the law was entirely correct. The argument not only seems to beg the question, but raises further questions as to the civil liability of one situated in the circumstances of this petitioner. Petitioner’s obligation here derived not from the decisions of this or any other court, but from the Act of Congress imposing the tax. It is hard to see what further point is being made, once it is conceded that'petitioner, if he was misled by the decisions of this Court, is entitled to plead in defense that misconception. Only in the most metaphorical sense has the law changed: the decisions of this Court have changed, and the decisions of a court interpreting the acts of a legislature have never been subject to the same limitations which are imposed on legislatures themselves, United States Constitution, Art. I, §§ 9, 10, forbidding them to make any ex post facto law3 and in the case of States to impair the obli*248gation of a contract. Ross v. Oregon, 227 U. S. 150; New Orleans Waterworks Co. v. Louisiana Sugar Refining Co., 125 U. S. 18.

The proper disposition of this case, in my. view, is to treat as plain error, Fed. Rules Crim. Proc. 52 (b), the failure of the trial court as trier of fact to consider whatever misapprehension may have existed in the mind of the petitioner as to the applicable law, in determining whether the Government had proved that petitioner’s conduct had been willful as required by the statute. On that basis I would send the case back for a new trial.