dissenting:
I dissent. I find unacceptable the majority’s proposals for incorporating the effects of future inflation into damage awards. Those proposals are founded upon a delusive belief in the ability of any court, jury, or expert to avoid utter speculation in attempting to predict future rates of inflation. Today’s holding will inflict upon the district courts burdens that are both complex and time consuming. Those burdens will serve no useful purpose because they will result in estimates of the damages due a plaintiff no fairer or more accurate than other, simpler methods would permit.
I do recognize that the courts must somehow deal with inflation in setting damage awards. I advocate adopting an approach that, as a matter of law, treats the inflation *318rate as totally offsetting the rate at which an award is discounted 1 and that therefore considers neither rate in computing an award. Such an approach, which the majority calls the “Alaska Rule,” is simple to apply and in my judgment more accurate than the formulae the majority advances for dealing with inflation.
I.
The majority, although noting that other methods might be acceptable, presents and endorses two specific proposals for negating inflation’s effect on damage awards.2 The approach that it deems most desirable, in a bare summary, would increase an award by the likely increase in a plaintiff’s wages resulting from all causes, including inflation, and then would discount the award at a rate reflecting the return on a relatively safe investment.3 A second approach would follow what the majority conceives to be the rule announced in Feldman v. Allegheny Airlines, 524 F.2d 384 (2d Cir. 1975). The majority’s treatment of the Feldman Rule is inconsistent4 and may betray a misunderstanding of the Second Circuit approach.5 Essentially, however, the majority would compute the difference between inflation and the discount rates. After consideration of any possibility that the rate of yearly increase in the plaintiff’s wages would differ from the rate of inflation, the award for each future year of plaintiff’s expected life would then be discounted by *319that difference.6 These proposals, the majority asserts, best ensure that a plaintiff receives no more nor less than is his due in a damage award.
I agree that the paramount concern of a court should be ensuring that a damage award is as accurate and fair as is practically possible. I disagree, however, with the majority’s notion that its proposals will ensure that a plaintiff will receive no more nor less than is his due. The majority’s analysis is curiously limited: it involves only computational precision, the determination of which formulae most accurately eliminate the effect of future inflation on a damage award once the rate of inflation is known? I note that even considered on such limited terms the majority’s preferred approach is not beyond objection: simplifications that the majority introduces into its calculations will overstate an award to a plaintiff.7 8 I also think it unfortunate that *320the majority was unable to decide on one specific proposal to mandate. Since the two formulae that the majority endorses will produce different results even if using the same data, the amount of an award will depend arbitrarily on the method of computation chosen.9 Such concerns, however, are not critical to my dissent. The fundamental weakness of the majority’s holding is its failure to consider at all the sheer speculation involved in any prediction of future inflation rates.10 The majority’s willingness to consider only computational precision in gauging the effects of inflation creates a mere illusion of accuracy because the underlying data that the formulae would use are themselves unreliable.
I do firmly maintain that it is quite evident that economic theory cannot predict future inflation rates with any degree of certainty. A survey of the general literature for the past several years illustrates the sorry tale of the repeated confusion, contradictions, and uncertainties of economic forecasts. See, e.g., “Where the Big Economic Models Go Wrong,” Bus. Week 70 (March 30, 1981); “An Amazing Contradiction in Economic Prophecies,” Bus. Week 39 (May 28, 1979); “Inflation Defies the Guidelines,” Bus. Week 32 (February 12, 1979); “Mixed Grades for the Economists,” Bus. Week 30 (December 25, 1978); “Theory Deserts the Forecasters,” Bus. Week 50 (June 29, 1974); Forbes, “Economics Shouldn’t Be a Branch of Mathematics,” Forbes 21 (January 18, 1982); “If We Don’t Know Where We Are, How Can We Tell Where We Are Going,” Forbes 119 (April 2, 1979); “Stockman’s Ladder,” Newsweek 66 (February 9, 1981); “Black-Box Forecasting,” Time 92 (February 23, 1981). Even the Council of Economic Advisors has conceded that “its ability to forecast inflation is at best imperfect.” The Economic Report of the President 1976 20; see also Council of Economic Advisors, Economic Report of the President 1982 215 (admitting federal government cannot fully anticipate the course of the economy); Solow, “The Intelligent Citizen’s Guide to Inflation,” Public Interest 49 (Winter 1975) (suggesting we may be less able to predict price levels 5-10 years into the future than in the past when stable prices were the norm). Moreover, these complaints about economic forecasting generally concern short term forecasts. Over a longer term, events as diverse — and unpredictable — as spending to finance a southeast Asian war, an oil embargo, or world agricultural shortfalls could have a profound impact on future inflation. See W. Branson & J. Litvack, Macroeconomics 408-14 (1976); Council of Economic Advisors, Economic Report of the President 1978 139-42. Cf. Economic Report of the President 1974, supra (“increasingly complex and interdependent world” increases difficulties in forecasting inflation); “Inflation Defies the Guidelines,” supra (discussing variety of factors affecting inflation). *321To attempt to predict the course of inflation for decades into the future is to attempt the impossible, given our present economic knowledge.11 Finally, common experience supports the conclusion that predictions of inflation simply are beyond our capabilities. Older generations can still recall the Great Depression and can remember how unexpected that event was.
I realize that any estimate of the appropriate damages to award a plaintiff involves substantial speculation. I believe, however, that the degree of speculation involved in prediction of future inflation is far greater than that which we usually require of a jury and that it is not essential to computing an award. Requiring a jury to resolve the amount of future inflation is not a duty similar to that traditionally performed, and usually performed well, by a jury: the finding of facts regarding a particular incident or individuals. Rather, we are demanding that the jury become a seer for the future course of national, or even world, events. Nor is the universe of relevant factors so limited as it is in the usual findings required of a jury. As noted earlier, the factors that may affect inflation are extremely broad in scope and hardly within the practical limits for a jury’s consideration. Finally, most predictions that a jury must make involve findings that can at least be actuarially or statistically estimated. That simply is not the case with estimates of future inflation. Admittedly, of course, a court or jury must at times make some finding concerning a fact not limited to the particular circumstances of a case, as when it determines the effect of future income taxes.12 Even those findings, however, I believe involve less speculation than occurs in predictions of future inflation. We at least can be relatively certain that income taxes will be a fixture of our future. That is not necessarily the case with inflation; the volatility of changes in inflation rates over time alone renders them unique.13 More important, there is not the necessity to speculate on future inflation that exists for findings on, for example, income taxes. No adequate method exists for considering the effect of income taxes on an award without speculating on future taxes. That is not so with inflation. As I note later in this opinion, a related variable, the discount rate, can be used to offset inflation and eliminate the need for its consideration.
The inability of the majority’s proposals adequately to deal with inflation in computing damage awards is not the only sin of those proposals. The majority’s holdings, both in the proposals adopted and in the procedures used for carrying out those proposals, are complex and extremely time consuming. Elaborate expert testimony will be necessary.14 And this testimony will not be limited to future inflation and future returns on investments. A court must also allow evidence of whether a plaintiff’s wages would keep up with or would increase at a rate faster than the inflation rate. By extension it appears that a court must allow separate expert testimony on *322whether any other component of an award that might be affected by inflation, such as an award for future medical expenses, would increase at a rate greater or less than the inflation rate. See note 3, supra. Moreover, the majority’s ambiguous suggestion regarding special verdicts, see majority opinion at note 41, appears to disregard Fed. R. Civ. P. 49(a)15 in removing use of special verdicts from the discretion of the trial court, thereby vastly complicating the task of the jury and of the trial judge.
I conclude that the majority has erected an elaborate superstructure on a foundation of sand. If it were possible to predict the course of inflation over the decades ahead, the majority’s formulae would do so with some precision. Without accurate predictions, however, working through the formulae will be an exercise in futility that will impose serious new burdens on the district courts.
II.
Having taken the position that there is no way to predict future inflation without undue speculation, there remains for me the task of establishing some fair and practicable approach for estimating damage awards. I do not believe that this Circuit should continue to follow Johnson v. Penrod Drilling Co., 510 F.2d 234 (5th Cir.) (en banc), cert. denied, 423 U.S. 839, 96 S.Ct. 69, 46 L.Ed.2d 58 (1975). As the majority notes, Penrod permits an award to be discounted at a rate that includes some prediction of future inflation. Not also allowing some increase in the award to account for inflation unfairly penalizes the plaintiff. Faced with the need to find some alternative method for dealing with the pernicious effects of inflation, I believe that the most desirable approach would be to offset, as a matter of law, the inflation rate with the discount rates.16
The adoption of such a rule would have several advantages over the majority’s plan.17 It would be simple to apply and would require no time consuming expert testimony. Speculation about future inflation would be unnecessary. Since the inflation rate and discount rate are covariant in *323opposite directions,18 the rule would be relatively accurate, certainly much more accurate in my opinion than the majority’s proposals.19 Finally, the effects of inflation would be considered uniformly and consistently for all damage awards. Since the inflation affecting awards will not vary in cases decided at similar times and compensating injuries over similar time periods, such uniform treatment is fairer than an approach in which estimates of inflation vary with each case.20
Having espoused the adoption of an approach that would deal with inflation as a matter of law, I must resolve one further issue, which of two variants of such an approach is preferable. One variant would establish a positive real rate of return on capital, that is, the rate of return that would exist on a relatively safe investment if there were no inflation. Such an approach would mirror that of Feldman v. Allegheny Airlines, 524 F.2d 384 (2d Cir. 1975) (1.5% real return); see also Doca v. Marina Mercante Nicaraguense, 634 F.2d 30 (2d Cir. 1979) (suggesting using 2% real rate of return), cert. denied, 451 U.S. 971, 101 S.Ct. 2049, 68 L.Ed.2d 351 (1981).21 A second variant, the Alaska Rule disparaged by the majority, would treat the inflation and discount rates as offsetting each other totally and therefore cancelling each other out. Such an approach has been advocated or adopted by several courts. E.g., Freeport Sulphur Co. v. S/S HERMOSA, 526 F.2d 300, 308-13 (5th Cir. 1976) (Wisdom, J., specially concurring); Pfeiffer v. Jones & Laughlin Steel Corp., 678 F.2d 453 (3d Cir. 1982); Beaulieu v. Elliott, 434 P.2d 665 (Alaska 1967); Resner v. N. Pac. Ry., 161 Mont. 177, 505 P.2d 86 (1973); Kaczkowski v. Bolubasz, 491 Pa. 561, 421 A.2d 1027 (1980).
I find a Feldman-type approach tempting since theoretically it may be preferable to the Alaska Rule.22 Ultimately, however, I, like Judge Wisdom in his specially concurring opinion in Freeport Sulphur, supra, would adopt the Alaska Rule. A rule establishing a positive real rate of return in practice involves as much guesswork as ruling that the inflation and discount rates offset each other completely — as an examination of the data that the district court in Feldman used in setting a 1.5% real rate will establish. See Feldman v. Allegheny Airlines, 382 F.Supp. 1271 (D.Conn. 1974), aff’d in relevant part, 524 F.2d 384 (2d Cir. 1975).23 Moreover, according to some economists, at least in an inflationary period the inflation component in the discount rate will fail fully to take into account the actual rate of inflation because the future depreciation in the value of money is not entirely foreseen. E.g., I. Fisher, The Theory of Interest 43 (1930); see also Freeport Sulphur Co., supra, 526 F.2d at 310 & n. 7 (Wisdom, J., specially concurring). Such an effect would offset at least partially any positive real rate of return on capital. Even the assumption that there is some *324constant real rate of return is open to question; some economists have suggested that the real rate of return will vary with the inflation rate. J. Keynes, The General Theory of Employment, Interest, and Money 142-43 (1936); Carlson, “Short-Term Interest Rates as Predictors of Inflation: A Comment,” 67 Am. Econ. Rev. 469 (1977); Steindl, “Price Expectations and Interest Rates,” 5 J. Money, Credit & Banking, 939, 948 (1973). Given the uncertainty described above in setting any positive real rate of return, I firmly believe a court should err, if it errs at all, in favor of the plaintiff, the non-culpable party, by establishing a total offset of the inflation and discount rates.24
III.
To summarize, I believe that the majority has sunk into its own Serbonian bog in adopting cumbersome, complicated, apparently illegal (in disregarding Fed.R.Civ.P. 49(a)), and time consuming procedures for the district courts to follow in incorporating the effects of inflation into damage awards. Those procedures cannot be justified by their precision in offsetting the effects of inflation because the data the procedures would use would be the product of pure speculation. A better procedure than that proposed by the majority and far simpler to apply, would, as a matter of law, treat the discount and inflation rates as totally offsetting each other.
. By “discounting” I mean reducing the stream of future payments on an award to its present value. The rate at which an award is discounted will depend on the rate of interest it could earn until the time at which it would be intended to compensate a plaintiff. For example, if an award is intended presently to compensate a plaintiff for $100 lost two years in the future, he should receive now a sum that, if invested and if earning interest, would give him $100 in two years. If the interest rate is 5%, he should receive $90.70 now. See R. Lipsey & P. Steiner, Economics 410-11 (4th ed. 1975); see also majority op. at n. 38 (giving discounting formula).
. The majority informs us neither of who has the prerogative of choosing a particular method nor of which method should be used under particular sets of circumstances. It may intimate that the choice is that of the plaintiffs. See majority op., supra, text at n. 43.
. The majority opinion writes solely in terms of wage increases. For convenience I write in similar terms. I note, however, that inflation would affect components of a damage award other than wages, such as future costs of medical care. I can conceive of no reason to limit consideration of inflation solely to its effect on future wages.
. There are at least two areas of confusion in the majority opinion. First, at times the Court correctly interprets Feldman as establishing a real rate of return on capital from past interest and discount rates and endorses use of that approach. See majority op., text at nn. 30, 48. At other times the majority interprets the Feldman approach as establishing a real rate of return from projected future increases in prices and from returns on safe investments (also presumably based on projections). See id., text at nn. 20-21, 42. Second, at times the majority refers to the court as establishing the factors relevant to the real rate of return, id., text at n. 48, while at other times it refers to the jury as setting the factors from which a real rate of return could be determined, see text at n. 42. It would be wise for the majority at least to give a coherent presentation of the proposals it endorses in order to guide the trial courts.
. The majority does not seem to realize that the “real rate of return” on capital as used in Feldman means the return on investments over the long term assuming no inflation and disregarding short term fluctuations. The majority suggests that the Feldman approach is defective because short term variations “[i]n the dynamic and ever changing world of finance and economics” will render any real rate of return erroneous. There will indeed be such short term fluctuations, but they would not affect the validity of a long term rate. The Feldman district court in fact averaged the short term fluctuations in order to derive a long term rate. See Feldman v. Allegheny Airlines, 382 F.Supp. 1271, 1293-95, 1309-12 (D. Conn. 1974), aff’d in relevant part, 524 F.2d 384 (2d Cir. 1975). Moreover, if there is a constant real rate of return, it can be established by relying on prior years’ evidence. There would be no need, contrary to what the majority at times suggests, see majority op., supra, text at nn. 20-21, 42, for projections of future inflation. The ability to avoid speculation about future inflation is, in fact, one of the prime advantages of the approach in Feldman, see text at nn. 18-19, infra; Doca v. Marina Mercante Nicaraguense, 634 F.2d 30, 39 (2d Cir. 1980), cert. denied, 451 U.S. 971, 101 S.Ct. 2049, 68 L.Ed.2d 351 (1981).
. The majority indicates, at least at times, that a Feldman-type approach should establish a real rate of return as a matter of fact. Other courts have interpreted Feldman as being applicable as a matter of law. See Doca, supra, 634 F.2d at 39 (although elsewhere suggesting that parties could dispute real rate of return); Freeport Sulphur Co. v. S/S HERMOSA, 526 F.2d 300, 310 11 (5th Cir. 1976) (Wisdom, J., specially concurring); see also Note, “Future Inflation, Damages, and the Circuit Courts,” 63 Va. L. Rev. 105, 130 31 (1977). Since a real rate of return could be established, as it was in Feldman. supra, 382 F.Supp. at 1293 95, from statistics of prior years and should not vary with each case, there seems no reason to require finding a real rate of return as a matter of fact and not of law. I discuss adopting such an approach as a matter ot law, infra, see text at nn. 22 24.
. The majority in fact misestimates the different results produced by independently incorporating the inflation and discount rates by the Feldman approach and by the Alaska Rule. See majority op., text at nn. 32 40. In the example it provides for its approach independently incorporating the inflation and discount rates, it increases the award due the plaintiff by a rate including merit and productivity wage increases. It does not consider merit or productivity increases under either of the latter two approaches. The failure to include such increases is contrary to the majority’s own statements in other parts of its opinion. See majority op., supra, n. 20 & text at n. 31. Under either of the last two approaches, noninflationary wage increases can be determined by looking to the wages commensurate with those increases in the year of the award. The district court in Feldman used such an approach. 382 F.Supp. at 1287; see also majority op., supra, n. 31.
While commenting on the majority’s “helpful hypothetical,” I should emphasize that it is helpful in showing the comparative accuracy of several methods of considering inflation’s effect on damage awards only if we have established a rate of inflation to use in comparing the results. If we cannot establish such a rate, and, as I argue in text, I believe we cannot do meaningless.
. Footnote 36 of the majority opinion, in explaining how to apply the majority’s preferred proposal, assumes a constant yield for each year of an award rather than an award increasing annually at the rate of wage increase. Such an assumption will overcompensate a plaintiff because it will provide him more money than is his due in the early years covered by his award. A plaintiff will, of course, receive less than is his due in the later years of his award, but that will not offset his high early awards, which he could have invested until he needed to use them to make up for receiving less than is his due in the later years. A simple example will demonstrate the point. 1 assume a plaintiff would earn $10,000 originally, increased by 5% each year and that he could invest any amount he receives at 5%. I also assume an award over 10 years. In the table below column A is the year; B is the plaintiff’s income increased by 5% annually; C is the total 10 year award averaged over each year— which is what the majority would award the plaintiff each year; D is the difference between C and B; and finally, E is the cumulative amount the plaintiff would have if he saved (or withdrew from earlier savings) the amount in column D and invested that amount at 5% (or withdrew that amount from investments).
B E
1 $10000 $12578 $2578 $2578
2 10500 12578 2078 4785
3 11025 12578 1553 6577
4 11576 12578 1002 7908
5 12155 12578 423 8726
6 12763 12578 -185 8977
7 13401 12578 -823 8603
8 14071 12578 -1493 7540
9 14774 12578 -2196 5721
10 15513 12578 -2935 3072
As of the last year of his award a plaintiff would have received the entire award that is his due and would have an additional $3072.
At the least the majority’s simplification implicitly concedes the need to sacrifice perfect *320accuracy to the dictates of simplicity and practicability.
. Assuming a discount rate of 6%, an inflation rate of 5%, a thirty year time period for the award, and an award equivalent to a constant $10,000 a year, an approach independently incorporating inflation and discount rates would result in an award of approximately $260,100 while an approach discounting each year’s earnings by the difference between the inflation and discount rates would result in an award of $258,100. See Note, “Future Inflation, Prospective Damages, and the Circuit Courts.” 63 Va. L. Rev. 105, 111 (1977). The $2,000 difference, although not huge, is significant. More important, I find theoretically objectionable the notion that an award should depend merely on the method of its calculation. That objection is particularly apposite where, as here, it is avoidable. Since the majority clearly prefers an approach separately incorporating the inflation and discount rates, there is no reason not to mandate that approach whenever practicable.
. I assume for the purposes of this criticism that the Feldman approach of the majority involves projecting future inflation. Cf. note 4, supra. Although I would find an approach using past inflation to be less objectionable, I ultimately also would reject that approach. See text at nn. 22 24, infra: see also n. 6, supra.
Also, I note that the majority’s first proposal incorporating all future wage increases without separating inflationary increases and then separately discounting would not avoid the need to consider future inflation. Any estimate of future wage increases must, implicitly or explicitly, include projections of future inflation.
. The majority’s further requirement that courts investigate even whether a plaintiff’s wages would have kept up with inflation takes its proposals to an absurd extreme. Given the unpredictability of inflation, there simply is no way of determining whether particular wages would keep up with it.
. Under Norfolk & W. Ry. Co. v. Liepelt, 444 U.S. 490, 100 S.Ct. 755, 62 L.Ed.2d 689 (1980), a jury must consider income taxes in making an award of damages.
. I note that for significant portions of this country’s history, roughly the late 19th century after the Civil War and the period between the World Wars, the dominant trend in prices was deflationary. See Solow, supra, Public Interest at 36-39. Future inflation may not be so inevitable as the majority presumes.
. The majority acknowledges that some courts have allowed a factfinder to use its common experience as well as expert testimony in determining the rate of future inflation, see majority op., supra, text at n. 12; Riha v. Jasper Blackburn Corp., 516 F.2d 840, 845 (8th Cir. 1975), but it never indicates to the district courts whether it approves of reliance on common experience. Given its repeated references to the use of expert testimony, however, the majority seems implicitly to disapprove of such reliance.
. Rule 49(a) states that a court may require special verdicts. The former Fifth Circuit has held that use of such verdicts or interrogatories is within the discretion of the district court. E. g., Central Progressive Bank v. Fireman's Fund Ins. Co., 658 F.2d 377, 381 (5th Cir. 1981); Jones v. Miles, 656 F.2d 103, 106 n. 4 (5th Cir. 1981); Miley v. Oppenheimer & Co., Inc., 637 F. 2d 318, 334 (5th Cir. 1981); see also 5A J. Moore & J. Lucas, Moore's Federal Practice § 49.03[ 1J, at 49 10 — 49 13 (1982) (“[T]he court has complete discretion as to whether a special or general verdict is to be returned.... [T]his should not be reviewable, except, perhaps, for gross abuse, which can rarely be shown.”) (footnote omitted). The majority apparently has chosen to ignore Rule 49(a) and to make use of special verdicts non-discretionary. See majority op., supra, nn. 41 & 42.
. An alternative approach that would have the advantage of simplicity would be to allow the jury merely to rely on its common experience in finding future inflation rates. See Bach v. Penn Central Transp. Co., 502 F.2d 1117, 1122 23 (6th Cir. 1974). However, I do not believe such an approach would be desirable. If experts cannot forecast inflation, a factfinder can hardly be expected to use more than pure guesswork in its findings.
. I recognize that in Norfolk & W. Ry. Co. v. Liepelt, 444 U.S. 490, 495, 100 S.Ct. 755, 758, 62 L.Ed.2d 689 (1980), the Supreme Court disapproved of the exclusion of one factor in an award, income taxes, to offset a rule on attorneys’ fees. That disapproval is inapplicable here. The factors considered in Liepelt were not remotely related to each other. By contrast, the factors at issue here are closely interrelated and opposite in their effect on damage awards.
Judge Wisdom, in his specially concurring opinion in Freeport Sulphur Co. v. S/S HERMOSA, 526 F.2d 300, 311-12 (5th Cir. 1976), noted that Chesapeake and Ohio Ry. v. Kelly, 241 U.S. 485, 36 S.Ct. 630, 60 L.Ed. 1117 (1916), requires discounting lost future earnings to present value in F.E.L.A. cases. In advocating that the Fifth Circuit adopt the Alaska Rule for incorporating the effects of inflation into damage awards, which would eliminate discounting, he suggested that Kelly’s holding could be limited to F.E.L.A. cases. I am not convinced that Kelly need literally control even in the F.E.L.A. context. In offsetting the discount rate by the inflation rate, a court would implicitly be considering the effect of discounting on an award and would be complying with the spirit of the mandate in Kelly.
. The discount rate includes a component incorporating expectations about future price levels. R. Lipsey & P. Steiner, Economics 417 (4th ed. 1975). The higher the expected inflation rate (and the rate at which an award would be increased to account for inflation) is the higher the discount rate (and the rate at which an award will be reduced) will be.
. A court should, of course, still consider noninflationary increases in wages. It could do so by considering the wages commensurate at the time of the award with the increases the plaintiff could be expected to receive in the future due to, for example, promotions or higher productivity. See note 7, supra.
. Judicial estimates of future inflation rates have in the past varied widely. One commentator has found estimates varying between 13a % and 5‘/2%. Note, supra, 63 Va. L. Rev. at n. 32 & n. 180.
. As I stated earlier, I believe that the approach of Feldman would be better adopted as a matter of law, not of fact. See note 6, supra.
. Theoretically the real rate of return should be positive to account for the opportunity cost of the foregone use of capital.
. The district court in Feldman found that the real rate of interest varied between - 2.9% and + 2.0% on short term treasury bills and varied between -8.9% and + 3.7% on long term federal notes. 382 F.Supp. at 1309, 1312; see also Note, supra, 63 Va. L. Rev. at 131.
. See Note, supra, 63 Va. L. Rev. at 132.