concurring.
I fully agree that the Government is entitled to prevail in this case, but I would rest that conclusion on a broader ground than the Court’s opinion might be taken to evince. More especially, I fear that certain distinctions suggested by the Court’s opinion — e. g., between clear and ambiguous statutes, decisions construing statutes for the first time, decisions overruling prior constructions of statutes — may point in the direction of a retroactivity quagmire in civil litigation not unlike that in which the Court has become ensnared in the criminal field. See my dissenting opinion in Desist v. United States, 394 U. S. 244, 256 (1969).
The impulse to make a new decisional rule nonretro-active rests, in civil cases at least, upon the same considerations that lie at the core of stare decisis, namely to avoid jolting the expectations of parties to a transaction. Yet *296once the decision to abandon precedent is made, I see no justification for applying principles determined to be wrong, be they constitutional or otherwise, to litigants who are in or may still come to court. The critical factor in determining when a new decisional rule should be applied to a transaction consummated prior to the decision’s announcement is, in my view, the point at which the transaction has acquired such a degree of finality that the rights of the parties should be considered frozen. Just as in the criminal field the crucial moment is, for most cases, the time when a conviction has become final, see my Desist dissent, supra, so in the civil area that moment should be when the transaction is beyond challenge either because the statute of limitations has run or the rights of the parties have been fixed by litigation and have become res judicata. Any uncertainty engendered by this approach should, I think, be deemed part of the risks of life.
These considerations, I believe, underlie the Court’s holdings in Chicot County Drainage District v. Baxter State Bank, 308 U. S. 371 (1940), where the Court refused to upset a judgment based on a subsequent change in the law, and Cipriano v. City of Houma, 395 U. S. 701 (1969), where we held that municipal bonds, authorized by invalid referenda, would not be subject to challenge “where, under state law, the time for challenging the election result has . . . expired.” 395 U. S., at 706.
To the extent that equitable considerations, for example, “reliance,” are relevant, I would take this into account in the determination of what relief is appropriate in any given case. There are, of course, circumstances when a change in the law will jeopardize an edifice which was reasonably constructed on the foundation of prevailing legal doctrine. Thus, it may be that the law of remedies would permit rescission, for example, but not an award of damages to a party who finds him*297self able to avoid a once-valid contract under new notions of public policy. Cf. Simpson v. Union Oil Co., 377 U. S. 13, 25 (1964). Another instance, though apt to arise infrequently in federal court, would be where certain real property transactions fail to anticipate changes in principles governing land usage, for example, the enforceability of certain kinds of easements or covenants. In such instances it may be appropriate to withhold an equitable remedy and confine an award of damages to a limited period, or the like.* The essential point is that while there is flexibility in the law of remedies, this does not affect the underlying substantive principle that short of a bar of res judicata or statute of limitations, courts should apply the prevailing decisional rule to the cases before them.
On these premises I join the Court’s opinion.
I would not, of course, hold this view of retroactivity binding on state courts and a federal court would, in fact, be obligated to abide by the applicable state rule should a retroactivity question arise in a diversity case.