JON O. NEWMAN, Circuit Judge:

This appeal presents the issue whether the 1982 amendment that changed the method for determining the rate of post-judgment interest may be given any retroactive effect. The practical consequence of the resolution of this somewhat esoteric issue is made abundantly clear in this case, since the additional interest at stake is more than $40 million. Plaintiffs-appellants Litton Corp. and related companies (collectively "Litton") appeal from a July 28, 1983, order of the District Court for the Southern District of New York (William C. Conner, Judge), denying their motion to amend a June 30, 1981, judgment that Litton obtained against defendants-appellees American Telephone Telegraph Co. and related companies (collectively "AT T"). We agree with Judge Conner that the 1982 amendment should not be given retroactive effect and affirm his denial of Litton's motion.

The June 1981 judgment awarded Litton $276,774,729 in treble damages following a jury verdict in its private civil antitrust action against AT T. As amended on July 20, 1981, in a respect not in issue on this appeal, the judgment provided for post-judgment interest at a rate of 9%. The post-judgment interest statute in effect when the judgment was entered specified that interest would run at the rate established by state law. 28 U.S.C. § 1961 (1976). The post-judgment interest rate in New York on June 30, 1981, was 9%. 7B N.Y.Civ.Prac. Law Rules §§ 5003, 5004 (McKinney 1963 Supp. 1983). AT T appealed the amended judgment to this Court on October 28, 1981. On November 19, 1981, the parties entered into a stipulation staying the execution of the amended judgment pending appeal. The stipulation recited the 9% post-judgment interest rate and was embodied in an order of the District Court.

On April 2, 1982, Congress enacted the Federal Courts Improvement Act of 1982 ("FCIA"), Pub.L. No. 97-164, tit. III, § 302(a)(1) and (a)(2), 96 Stat. 25, 55-56, codified at 28 U.S.C. § 1961(a) (1982). The FICA amended 28 U.S.C. § 1961 to provide that, effective October 1, 1982, interest on federal court judgments would be computed at the "coupon issue yield equivalent ... of the average accepted auction price for the last auction of fifty-two week United States Treasury bills settled immediately prior to the date of the judgment." Id. Slightly more than a month after passage of the FCIA, Litton submitted its reply brief to this Court in AT T's appeal of the underlying judgment. Litton did not seek a remand to the District Court for an amendment of the judgment to reflect the interest rate calculation under the FCIA, nor did Litton otherwise seek to preserve the issue for later consideration by the District Court. This Court affirmed the District Court judgment on February 3, 1983, Litton Systems, Inc. v. American Telephone and Telegraph Co., 700 F.2d 785 (2d Cir. 1983), cert. denied, ___ U.S. ___, 104 S.Ct. 984, 79 L.Ed.2d 220 (1984), and issued its mandate on April 29, 1983. On May 18, 1983, more than six months after the effective date of the FCIA, Litton filed a motion in the District Court to amend the final judgment to increase the post-judgment interest rate from 9% to 13.146%, the Treasury bill rate at the time of the entry of judgment. Litton sought to apply the increased rate from June 30, 1981, the date the judgment was entered. In effect, the motion sought an additional $40 million from AT T. On July 28, 1983, Judge Conner denied Litton's motion to amend the final judgment, Litton Systems, Inc. v. American Telephone and Telegraph Co., 568 F.Supp. 507 (S.D.N.Y. 1983). This appeal followed.

Litton urges us to apply retroactively the post-judgment interest provision of the FCIA, which amended section 1961. It relies on the general statement, proclaimed in Bradley v. School Board, 416 U.S. 696, 711, 94 S.Ct. 2006, 2016, 40 L.Ed.2d 476 (1974), that "a court [on direct review] is to apply the law in effect at the time it renders its decision, unless doing so would result in manifest injustice or there is statutory direction or legislative history to the contrary." Litton contends that it is entitled to retroactive application of amended section 1961 because Congress gave no "contrary" indication, such application would not result in manifest injustice, and the judgment of the District Court was pending in this Court on direct review on the effective date of the FCIA.

Preliminarily, we have some doubt whether the quoted statement from Bradley is even applicable to this case. The plaintiff in Bradley sought in the District Court an award of attorney's fees, after prevailing in a school desegregation suit. While an appeal from denial of fees was pending, Congress enacted a statute authorizing fee awards. Civil Rights Attorney's Fees Awards Act of 1976, Pub.L. No. 94-559, § 2, 90 Stat. 2641, codified at 42 U.S.C. § 1988 (1982). Thus, the newly enacted legislation concerned the very legal standard at issue on the appeal. As Justice Blackmun noted:

The question, properly viewed, then, is not simply one relating to the propriety of retroactive application of [the attorney's fee statute] to services rendered prior to its enactment, but rather, one relating to the applicability of that section to a situation where the propriety of a fee award was pending resolution on appeal when the statute became law.

Bradley v. School Board, supra, 416 U.S. at 710, 94 S.Ct. at 2015.

Litton's situation is distinguishable from that of the appellant in Bradley because the appeal that was pending in this Court when the FCIA became effective involved no issue concerning the adequacy of post-judgment interest. Nevertheless, we are reluctant to give Bradley a narrow reading that would make this circumstance decisive. No party to an appeal should be held to a standard that permits consideration of an intervening statute only when issues affected by the statute are already pending on appeal. Such a standard would require either anticipation of statutes not yet enacted or the assertion of frivolous grounds in appeals and cross-appeals in the hope that a new statute might affect their resolution favorably. We agree with the District Court that the Bradley "rule" concerning retroactivity is not limited to cases in which the intervening statute concerns the precise issue already pending before the appellate court.

However, Bradley extends its retroactivity rule only to cases pending on direct appeal at the time a change in the law occurs, 416 U.S. at 710-11, 94 S.Ct. at 2015-16, and there is no unfairness in requiring a party seeking to have a new statute applied to a judgment pending on direct review to alert the appellate court to its contention. Though the Bradley rule does not require clairvoyance, it should not reward reticence. Litton apparently interprets the Bradley rule to mean that whenever a case is pending on direct review at the time a new law is enacted or becomes effective, a party may thereafter secure the benefit of the new law in collateral proceedings in the trial court. We disagree. The Bradley rule is not an invitation to litigants to seek collateral relief; it is an instruction to appellate courts to adjudicate cases pending on direct appeal according to the law then in effect.

Perhaps Litton assumed that it could not press its claim for interest at the increased rate in this Court on the direct appeal since it had not, understandably, included a claim for additional interest in its cross-appeal, which was filed before the FCIA was enacted. Nevertheless, it was open to Litton thereafter to identify the issue on the direct appeal, if only to attempt to preserve its rights to obtain collateral relief in the District Court. Litton could have sought a remand to permit an application to the District court, or could have filed supplemental papers identifying the issue and requesting that the opinion of this Court at least not foreclose recourse to the District Court in order to pursue the interest claim.

Since this Court has not previously spoken authoritatively on the Bradley issue, particularly concerning a party's obligation to identify issues arising under newly enacted statutes while a direct appeal is pending, we are reluctant to rest decision in this case on the fact that Litton did not make any reference to post-judgment interest on direct review in this Court, but instead raised the matter for the first time collaterally in the District Court. Henceforth, however, we will expect any litigant seeking to benefit under Bradley from a new statute enacted while a direct appeal is pending to alert this Court to its contention during the pendency of the appeal. Otherwise, the party will normally be limited to those benefits from newly enacted statutes that are available to be pursued in requests for collateral relief, a narrow category that we do not understand Bradley to have enlarged. See Linkletter v. Walker, 381 U.S. 618, 85 S.Ct. 1731, 14 L.Ed.2d 601 (1965).

We will therefore endeavor to apply the Bradley rule to Litton's claim. In doing so, we find it helpful to have in mind the line of cases in which the rule developed. That line began with Chief Justice Marshall's famous decision in United States v. Schooner Peggy, 5 U.S. (1 Cranch) 103, 2 L.Ed. 49 (1801). The Court there considered whether to apply to a pending appeal of an order condemning a captured ship a treaty, signed and promulgated after the District Court's order, that required restoration of property captured but not "definitively condemned." Chief Justice Marshall first noted that an appellate court usually inquires only "whether a judgment, when rendered, was erroneous or not." Id. at 110. He next stated that "if subsequent to the judgment and before the decision of the appellate court a law intervenes and positively changes the rule which governs, the law must be obeyed...." Id. He then observed, in an important passage not always fully appreciated, that "obey[ing]" the intervening statute did not necessarily mean applying it to the pending case; instead the statute must be construed, and, if only private rights were involved, a retroactive construction was not favored:

It is true that in mere private cases between individuals. a court will and ought to struggle hard against a construction which will, by a retrospective operation, affect the rights of parties, but in great national concerns where individual rights, acquired by war, are sacrificed for national purposes, the contract making the sacrifice ought always to receive a construction conforming to its manifest import.... In such a case the court must decide according to existing laws....

Id. (emphasis added).

In subsequent cases, the Court applied intervening law to pending cases in a variety of contexts and for a variety of reasons, but these decisions did not depart from Chief Justice Marshall's basic point that, while an intervening law must be obeyed, its application to a judgment pending on direct review is a matter of construction. His point began to be obscured by language in Ziffrin, Inc. v. United States, 318 U.S. 73, 78, 63 S.Ct. 465, 468, 87 L.Ed. 621 (1943), stating, "A change in the law between a nisi prius and an appellate decision requires the appellate court to apply the changed law." What remained unclear was whether the appellate court in "apply[ing]" the new law was only to recognize its potential pertinence to the pending case or was to construe the statute to govern the pending case. The limited holding in Ziffrin did nothing to resolve the uncertainty. By the time of Thorpe v. Housing Authority, 393 U.S. 268, 89 S.Ct. 518, 21 L.Ed.2d 474 (1969), however, it was certainly arguable that the Court was using "apply" in the sense of "determine the outcome" when it said, "The general rule, however, is that an appellate court must apply the law in effect at the time it renders its decision." Id. at 281, 89 S.Ct. at 525 (footnote omitted). Principal reliance was placed on the above-quoted dictum from Ziffrin. Id. at 281 n. 38, 89 S.Ct. at 526 n. 38.

Reviewing the line of cases from Schooner Peggy to Thorpe, Justice Blackmun in Bradley recognized that uncertainty existed, but he saw the uncertainty as involving a choice between canons of construction that created presumptions against or for retroactivity, rather than a choice between using any presumption and giving each statute an individualized construction. As he viewed it, introducing the problem with a pertinent metaphor:

In the wake of Schooner Peggy, however, it remained unclear whether a change in the law occurring while a case was pending on appeal was to be given effect only where, by its terms, the law was to apply to pending cases, as was true of the convention under consideration in Schooner Peggy, or, conversely, whether such a change in the law must be given effect unless there was a clear indication that it was not to apply in pending cases.

416 U.S. at 712-13, 94 S.Ct. at 2016-17 (emphasis in original). He then viewed Thorpe as settling the matter in favor of "the broader reading of Schooner Peggy," id. at 714, 94 S.Ct. at 2017, though Thorpe could perhaps more plausibly have been read as simply a decision that the regulation there involved was fairly to be construed to govern the pending case.

Having opted for a rule of construction that an intervening statute should govern pending cases unless a contrary intention appears, Justice Blackmun then recognized, as Thorpe had, 393 U.S. at 282, 89 S.Ct. at 526, an exception for "manifest injustice." Bradley v. School Board, supra, 416 U.S. at 716-17, 94 S.Ct. at 2018-19. However, in endeavoring to give some content to the category of cases warranting exemption from a general rule of retroactivity for pending cases, he cited Chief Justice Marshall's caution against retroactive construction in suits between private parties. Id. at 717, 94 S.Ct. at 2019. It thus appeared that Chief Justice Marshall's advice that appellate courts should "struggle hard against a construction which will, by a retrospective operation, affect the rights of parties," United States v. Schooner Peggy, supra, 5 U.S. (1 Cranch) at 110, had been dropped from the formulation of a general "rule" concerning retroactivity of intervening statutes, only to reemerge as an important basis for determining when an exception to the general rule should be made.

Bradley creates at most a presumption in favor of retroactive application of new statutes to judgments pending on direct review, but one that may be displaced either by a fair indication that the statute, properly construed, has only prospective effect or by resulting injustice from a retroactive application. We do not understand Bradley to mean, as Litton seems to suggest, that the only cognizable basis for a construction against retroactivity is either explicit statutory language or unequivocal congressional intent. Determination of retroactivity remains a matter of statutory construction, albeit with a presumption in favor of retroactivity as to judgments pending on direct review, contrary to the normal presumption against retroactivity in other circumstances. See United States v. Security Industrial Bank, 459 U.S. 70, 103 S.Ct. 407, 414, 74 L.Ed.2d 235 (1982); Union Pacific Railroad Co. v. Laramie Stock Yards Co., 231 U.S. 190, 199, 34 S.Ct. 101, 102, 58 L.Ed. 179 (1913); United States Fidelity Guaranty Co. v. United States ex rel. Struthers Wells Co., 209 U.S. 306, 314, 28 S.Ct. 537, 539, 52 L.Ed. 804 (1908). With this understanding of Bradley, we turn to the task of construing amended section 1961.

The terms of the new statute favor a construction against retroactivity, even without the "hard struggle" Chief Justice Marshall thought appropriate for cases involving private parties. Rather than simply provide for an increase in the rate of interest, cf. 1981 New York Laws, ch. 258, § 1, codified at 7B N.Y.Civ.Prac.Law and Rules § 5004 (McKinney Supp. 1983) (raising interest rate from 6% to 9%), the amendment to section 1961 establishes a somewhat complicated mechanism for determining the applicable rate and, significantly, ties the rate determination directly to the date of entry of judgment. The statutory direction to calculate a rate based on the average coupon yield for one-year Treasury bills auctioned "immediately prior to the date of the judgment" is a strong indication that Congress expected the new formula to apply only to judgments entered after the effective date of the statute. The legislative history offers no guidance on the issue of retroactivity, leaving the terms of the statute the only indication of the proper construction. We agree with the Seventh and Fourth Circuits that amended section 1961 is not to be construed to apply to judgments entered before its effective date, even though on that date they were pending on direct review. See Merit Insurance Co. v. Leatherby Insurance Co., 728 F.2d 943, 944 (7th Cir. 1984); United States v. Dollar Rent A Car Systems, Inc., 712 F.2d 938, 940 n. 5 (4th Cir. 1983).

Litton contends, alternatively, that if amended section 1961 does not fully apply to judgments pending on direct review on October 1, 1982, so as to provide increased interest from the date of such judgment, it should at least be construed to provide increased interest on such judgments from the effective date of the statute. If the new statute had only increased the rate of post-judgment interest, Litton's fall-back position would be well taken. Statutes raising interest rates on judgments are generally construed to effect an increase in the post-judgment interest accruing on all unpaid judgments, whether or not pending on direct review, with the increase accruing as of the effective date of the statute. See Morley v. Lake Shore Michigan Southern Railway Co., 146 U.S. 162, 168, 13 S.Ct. 54, 56, 36 L.Ed. 925 (1892); Shook Fletcher Insulation Co. v. Central Rigging Contracting Corp., 684 F.2d 1383, 1388-89 (11th Cir. 1982); Spang Industries v. Aetna Casualty Surety Co., 512 F.2d 365, 372 (2d Cir. 1975). However, as we have noted, amended section 1961 does more than accomplish a rate increase; it establishes a new method for computing post-judgment interest, a method tied inextricably to the date the judgment enters. A statute specifying this method cannot sensibly be construed to increase rates on pending judgments as of the effective date of the statute, a conclusion that emerges when one considers how Litton's fall-back position might be implemented.

If amended section 1961 were applied to judgments entered before its effective date but with interest at an increased rate accruing only from the effective date, a choice would have to be made as to the applicable rate. Litton contends that the rate should be the average coupon yield of Treasury bills auctioned just prior to June 30, 1981, the date the judgment entered. That makes no economic sense at all and could not have been intended by Congress. Since Litton's judgment was properly accruing interest at 9% from the date of its entry until at least October 1, 1982, it would be an unwarranted windfall to have it accrue interest thereafter at a rate determined by market conditions 15 months earlier. On the other hand, if the rate accruing from the effective date of the statute was the average coupon yield of Treasury bills auctioned just prior to that effective date, the result would be economically defensible, but hopelessly inconsistent with the terms of the statute, which tie the applicable rate to the date of the judgment. We therefore reject Litton's fall-back position, but see Handgards, Inc. v. Ethicon, Inc., 552 F.Supp. 820 (N.D.Cal. 1982), aff'd, 743 F.2d 1282 (9th Cir. 1984), and construe amended section 1961 to have no application to judgments entered before October 1, 1982.

Our construction of amended section 1961 is reenforced by consideration of the criteria identified in Bradley to assess "manifest injustice." The three factors to consider in deciding whether an injustice will result from a ruling in favor of retroactivity are "(a) the nature and identity of the parties, (b) the nature of their rights, and (c) the nature of the impact of the change in law upon those rights." Bradley, supra, 416 U.S. at 717, 94 S.Ct. at 2019. Each of these factors counsels against retroactive application of the FCIA in this case.

The suit is one between private parties. Nevertheless, Litton suggests that it is not within the category of "mere private cases" Chief Justice Marshall contemplated in United States v. Schooner Peggy, supra, 5 U.S. (1 Cranch) at 110. Whatever consideration is appropriate for antitrust plaintiffs as private attorneys general bringing suits that serve the national interest, there is little doubt that national interests are not affected by the outcome of this collateral action to secure an increase in the rate of post-judgment interest.

The second factor is the nature of the rights affected by retroactive application of the change in federal law. The Supreme Court noted in Bradley that it had "refused to apply an intervening change to a pending action where it has concluded that to do so would infringe upon or deprive a person of a right that had matured or become unconditional." 416 U.S. at 720, 94 S.Ct. at 2020 (citations omitted). AT T had a right, grounded on the prior version of section 1961, to rely on the continued accrual of interest at the 9% rate, at least until the effective date of the new statute. In addition, the stipulation between the parties, embodied in a separate order of the District Court, provided a basis for believing that the 9% interest rate would continue to apply throughout the pendency of appellate review.

The third factor is the impact that retroactive application of the change in law would have upon existing rights. The essence of this inquiry is whether the change in law would "impose an additional or unforeseeable obligation" upon the disadvantaged party. Bradley, supra, 416 U.S. at 721, 94 S.Ct. at 2021. In Bradley the Supreme Court concluded that the third factor weighed against a finding of injustice because there was no indication that the new statutory obligation, if known, would have caused the defendant to alter its conduct. Id. The same cannot be said here. If AT T had had reason to believe that it would be required, pending appeal, to pay a significantly higher rate of interest on the judgment against it, it might have paid the judgment immediately, rather than incur the additional liability, and sought return of its money if it prevailed. Indeed, Litton's arguments that the FCIA rate is closer to AT T's cost of capital only increases the likelihood that AT T would have acted differently had it been able to anticipate the change in law.

The claim in this case may instructively be compared with a somewhat similar situation in Brady v. Chemical Construction Corp., 740 F.2d 195 (2d Cir. 1984), involving an increase in the applicable rate of pre-judgment interest. We noted in Brady that the increased prejudgment rate had been applied to the damages from the effective date of the statute increasing the rate to the date of the judgment. 740 F.2d at 199 n. 6. That award of interest, however, could not have affected the conduct of the defendant. The prejudgment interest was awarded to compensate the plaintiff for loss of the use of the damages ultimately awarded between the date of the loss and the date of the judgment. The defendant did not know it was obligated to pay damages to the plaintiff until a judgment entered. The increase in the rate of prejudgment interest simply imposed an additional consequence upon conduct that had already occurred — the defendants' breach of fidelity bonds. With post-judgment interest, however, the defendant, knowing the amount of money it is obligated to pay, is entitled to rely on the existing rate of interest in deciding whether to pay the judgment pending appeal.

These considerations indicate that it would be manifestly unjust to accept Litton's basic claim and apply an increased rate of post-judgment interest retroactively from the date of entry of the judgment. Admittedly, there would not be unfairness in applying an increased rate only from the effective date of the new statute if it was reasonably clear to a judgment debtor that a new rate became applicable to the judgment at that date. But, as we have pointed out, that intermediate position is so plainly contrary to the statutory scheme that Congress selected in amending section 1961 that we cannot accept it, and it would be unjust to expect a judgment debtor to have so understood the statute. Applying the principles of Bradley to construe the statute and to assess whether retroactivity is manifestly unjust, we conclude that amended section 1961 applies only to judgments entered after October 1, 1982.

The order of the District Court is affirmed.