with whom The Chief Justice joins, concurring in part and dissenting in part.
For the reasons given by the Court, I agree that prejudgment interest was impermissibly awarded in this FELA case. Accordingly, I join Parts I and II of its opinion. Because the trial court erroneously gave conclusive effect to a state rule requiring the use of a “total offset” method of calculating present value, I also agree that we must reverse the judgment of the Supreme Court of Pennsylvania upholding that decision. I do not agree, however, that juries must in all circumstances be left free to choose among the total offset rule and alternative methods of accounting for anticipated future inflation.
The majority correctly notes that damages awards in state-court FELA cases must be based on an approximation of present value, and that the jury must be instructed accordingly. Ante, at 339-340. “Although ... [it is] clear that no single method for determining present value is mandated by federal law and that the method of calculating present value should take into account inflation and other sources of wage increases as well as the rate of interest, itt is" equally clear that an utter failure to instruct the jury that present value is the proper measure of a damages award is error.” St. Louis *351Southwestern R. Co. v. Dickerson, 470 U. S. 409, 412 (1985). The reason for this rule is plain: because of the time value of money and the practice of awarding damages in a lump sum, the failure to reduce awards for loss of future wages to their present value would risk the systematic overcompensation of plaintiffs. Ibid.
For reasons that were explained in great detail in this Court’s opinion in Jones & Laughlin Steel Corp. v. Pfeifer, 462 U. S. 523 (1983), the best method for calculating the approximate present value of an award for loss of future earnings could become the subject of reasonable debate in almost any case. Some of the uncertainties arise from the phenomenon of price inflation in our economy, and others arise from the possibility of “productivity gains” that would have increased a particular plaintiff’s real wages over time. Absent congressional action, Pfeifer cautiously declined to impose any one method of accounting for these phenomena on all litigants and all cases. Instead, the Court concluded that in federal bench trials, the judge should be permitted to choose among reasonable methods and should explain the choice. Id., at 548-549, 552-553.
The trial judge in this case believed that the choice among methods of calculating present value was a procedural issue, and she “blindly applied the total offset method [of] Kaczkowski v. Bolubasz, 491 Pa. 561, 421 A. 2d 1027 (1980).” 513 Pa. 86, 100, 518 A. 2d 1171, 1178 (1986) (dissenting opinion below); see also Tr. 577-578. The Supreme Court of Pennsylvania concluded that use of the total offset method was appropriate here because that method had been “deliberately selected [in Kaczkowski] as the rule that most nearly provides an injured claimant with damages to the full extent of the injuries sustained.” 513 Pa., at 92-93, 518 A. 2d, at 1174. Whatever might be said in favor of precluding individual trial courts and juries from reexamining such a rule, we specifically held in Pfeifer that it was improper for a federal trial court to apply the Kaczkowski rule without examining *352its suitability in the particular case before the court. See 462 U. S., at 546, 551-553. The reason for this holding was a reluctance “to select [any] of the many rules [for calculating present value] that have been proposed and establish it for all time as the exclusive method in all federal trials for calculating an award for lost earnings in an inflationary economy.” Id., at 546. State appellate courts are in no better position than this Court to select a single rule for all cases and all time. The reasoning in Pfeifer therefore precludes the automatic application of any method of estimating present value in an FELA case. Although I therefore agree with the Court that the judgment below must be reversed, I do not believe that a jury charge reflecting the Kaczkowski rule would necessarily have been impermissible in this case.
Believing that the choice of a method for estimating present value is always an “essentially factual question,” the majority will permit the trial judge to “recommend to the jury one or more methods” but will not permit the judge to impose “‘rigid mathematical limitation[s].’” Ante, at 341-342. As with many mixed questions of law and fact, one can imagine cases in which the choice between competing methods of calculating present value would turn on disputed factual questions. In such cases, I agree, it would be error to take those questions from the jury. If, for example, the parties offered conflicting evidence as to foreseeable wage increases, attributable to promotions for which the plaintiff would have been eligible but for the injury-that gave rise to the case, a court could not properly impose on the jury a rule under which the market interest rate is deemed equal to price inflation plus all other sources of future wage increases. Cf. Pfeifer, supra, at 544-545 (discussing Beaulieu v. Elliott, 434 P. 2d 665 (Alaska 1967)). The case before us, however, raises no such possibility.
In this case, the trial judge applied the rule set out in Kaczkowski, v. Bolubasz, 491 Pa. 561, 579-583, 421 A. 2d 1027, 1036-1039 (1980), under which the rates of future price *353inflation are presumed equal to the available interest rates. See 513 Pa., at 94-95, 518 A. 2d, at 1175. (Under Kaczkow-ski, it should be noted, the jury rather than the judge estimates the effects of lost “productivity gains,” which depend largely on factors that are specific to the plaintiff and the industry in which the plaintiff worked. See 491 Pa., at 579-580, 421 A. 2d, at 1036-1037.) The relation between future price inflation and interest rates has virtually nothing to do with any particular plaintiff or with the facts of any individual case. The absence of such a connection is illustrated by the present case, in which there is no suggestion that appellant offered any expert testimony or other evidence about future levels of price inflation and interest rates. See Brief for Appellant 34.
Appellant’s proposed jury instruction, moreover, made no mention of inflation, and thus appears on its face to have been potentially far more misleading than the Kaczkowski-type instruction actually given. Compare App. 68a-69a with Tr. 701-702. The majority does not say, and I do not think one could properly say, that the trial judge erred by refusing to give appellant’s proposed jury instruction. Nor, as this Court’s Pfeifer opinion amply demonstrates, is the prediction of future inflation .and interest rates a matter that can properly be referred to common experience or common sense. The majority nonetheless concludes that the jury in this case should have been left free to choose between a reasonable rule like the one adopted in Kaczkowski and such other methods as may have occurred to the jurors. I cannot agree that federal law requires a trial judge to permit the jury to speculate about something so uncertain and so disconnected from any evidence in the case at hand.
The majority suggests that its decision to the contrary is supported by Chesapeake & Ohio R. Co. v. Kelly, 241 U. S. 485 (1916), and Louisville & Nashville R. Co. v. Holloway, 246 U. S. 525 (1918). The majority does not, and I believe could not, say that the result reached today is compelled by *354those decisions. Kelly held only that it is error for a court to take no steps to ensure that an FELA award is reduced to its approximate present value. Although the Court assumed that juries would calculate the reduction, it did not say or imply that the choice of a method for making that calculation must always be left entirely to the jury. Similarly, Holloway held that a party was not entitled to a jury instruction requiring that an award be discounted by the legal rate of interest in effect at that time. Choosing the legal rate of interest as the discount rate in an FELA action would be arbitrary because there is no reason to suppose that it takes account of inflation or of the reasonable investment opportunities actually available to a prevailing plaintiff. See generally Jones & Laughlin Steel Corp. v. Pfeifer, 462 U. S., at 537-539. As Pfeifer indicates, however, the “total offset” rule at issue in this case is not arbitrary and would not constitute reversible error if employed by a judge in a bench trial, even if the parties had not introduced evidence to support it. See id., at 546-547, 551-553. Thus, although Holloway rejected a party’s claim that it was entitled to a particular “rigid mathematical limitatio[n]” on a jury’s discretion, it did not say or imply that a judge may never impose any such limitation. Compare Holloway, supra, at 528, with ante, at 340-341.
Although it was error to apply Kaczkowski automatically in this case, I see no reason to prohibit the trial court in an FELA case from employing a rebuttable presumption in favor of the Kaczkowski rule. That rule is reasonable, and we indicated in Pfeifer that it could be used in a bench trial so long as the judge chose it deliberately. 462 U. S., at 551-553. Unless one of the parties shows by evidence or argument that some other rule or method of estimating present value may be more appropriate, I simply see no basis for turning a jury loose to speculate about the complex and tech*355nical matters that underlie any method of estimating present value.*
In the case before us, appellant offered a jury instruction on the calculation of present value that was inconsistent with Kaczkowski. App. 68a-69a. It has not been suggested, however, that appellant offered evidence to show that present value could be estimated more accurately by some method other than the one adopted in Kaczkowski; nor is there any suggestion that appellant was denied an opportunity to present such evidence. See Tr. 577; Brief for Appellant 34. Assuming that the record would not support such suggestions (a question that I would leave for the Supreme Court of Pennsylvania to address on remand), I would require only that the trial judge make a “deliberate choice” between the Kaczkowski rule and others that may be suggested. See Pfeifer, supra, at 552-553. If a reasonable choice were made in favor of the Kaczkowski rule, I would permit the jury’s verdict to be reinstated.
Despite the majority’s suggestion to the contrary, ante, at 341, Pfeifer did not imply that the choice between the Kaczkowski rule and some competing method is necessarily a jury question. First, the decision in Pfeifer dealt with cases arising under the Longshoremen’s and Harbor Workers’ Compensation Act, where the trier of fact is a federal judge. 462 U. S., at 547. For that reason, we had no occasion in Pfeifer to decide which of the many questions that may arise in the calculation of present value may be resolved by the judge in a jury case. Second, the Pfeifer opinion did not consistently say that the method of calculating present value must be chosen by the “trier of fact.” See, e. g., id., at 548-549 (“[W]e do not believe a trial court adopting [a ‘real interest rate’] approach . . . should be reversed if it adopts a rate between 1 and 3% and explains its choice”). Third, Pfeifer’s expressed expectation that a trial court would explain its choice of a particular discount rate suggests that the decision is not always an “essentially factual question.” See ante, at 342. Judges are not ordinarily expected to give explanations for their findings of fact, and juries ordinarily do not give explanations at all.